Idea Transcript
Long-term investment outlook
UK Edition June 2017
The value of investments, and the income from them, can go down as well as up and investors may get back less than the amount invested. Past performance is not a guide to future returns. Important information For professional investors and financial advisers only – not for use by retail investors The above document is strictly for information purposes only and should not be considered as an offer, investment recommendation, or solicitation, to deal in any of the investments or funds mentioned herein and does not constitute investment research as defined under EU Directive 2003/125/EC. Aberdeen Asset Managers Limited (‘Aberdeen’) does not warrant the accuracy, adequacy or completeness of the information and materials contained in this document and expressly disclaims liability for errors or omissions in such information and materials. Any research or analysis used in the preparation of this document has been procured by Aberdeen for its own use and may have been acted on for its own purpose. The results thus obtained are made available only coincidentally and the information is not guaranteed as to its accuracy. Some of the information in this document may contain projections or other forward looking statements regarding future events or future financial performance of countries, markets or companies. These statements are only predictions and actual events or results may differ materially. The reader must make their own assessment of the relevance, accuracy and adequacy of the information contained in this document and make such independent investigations, as they may consider necessary or appropriate for the purpose of such assessment. Any opinion or estimate contained in this document is made on a general basis and is not to be relied on by the reader as advice. Neither Aberdeen nor any of its employees, associated group companies or agents have given any consideration to nor have they or any of them made any investigation of the investment objectives, financial situation or particular need of the reader, any specific person or group of persons. Accordingly, no warranty whatsoever is given and no liability whatsoever is accepted for any loss arising whether directly or indirectly as a result of the reader, any person or group of persons acting on any information, opinion or estimate contained in this document. Aberdeen reserves the right to make changes and corrections to any information in this document at any time, without notice. The MSCI information may only be used for your internal use, may not be reproduced or redisseminated in any form and may not be used as a basis for or a component of any financial instruments or products or indices. None of the MSCI information is intended to constitute investment advice or a recommendation to make (or refrain from making) any kind of investment decision and may not be relied on as such. Historical data and analysis, should not be taken as an indication or guarantee of any future performance analysis forecast or prediction. The MSCI information is provided on an ‘as is’ basis and the user of this information assumes the entire risk of any use made of this information. MSCI, each of its affiliates and each other person involved in or related to compiling, computing or creating any MSCI information (collectively, the ‘MSCI’ Parties) expressly disclaims all warranties (including without limitation, any warranties of originality, accuracy, completeness, timeliness, non-infringement, merchantability and fitness for a particular purpose) with respect to this information. Without limiting any of the foregoing, in no event shall any MSCI Party have any liability for any direct, indirect, special, incidental, punitive, consequential (including, without limitation, lost profits) or any other damages (www.msci.com). FTSE International Limited (‘FTSE’) © FTSE 2017. ‘FTSE®’ is a trade mark of the London Stock Exchange Group companies and is used by FTSE International Limited under licence. RAFI® is a registered trademark of Research Affiliates, LLC. All rights in the FTSE indices and/or FTSE ratings vest in FTSE and/or its licensors. Neither FTSE nor its licensors accept any liability for any errors or omissions in the FTSE indices and/or FTSE ratings or underlying data. No further distribution of FTSE Data is permitted without FTSE’s express written consent. Issued by Aberdeen Asset Managers Limited. Authorised and regulated by the Financial Conduct Authority in the United Kingdom.
Contents Foreword 04 Introduction 05 Report team
09
Chapter 1: Strategic asset allocation in a low interest-rate environment
10
Chapter 2: Long-term growth
20
Chapter 3: Trends in inflation, interest rates and currencies
30
Chapter 4: Regional outlook
38
Chapter 5: Equities
48
Chapter 6: Government bonds
60
Chapter 7: Credit
68
Chapter 8: Emerging-market sovereign debt
74
Chapter 9: Listed alternatives and property
78
Appendix: Expected returns by base currency
84
aberdeen-asset.co.uk
03
Foreword Asset allocation decisions are amongst the most important investors make. A poor combination of assets and strategies can result in annual returns two or three percentage points per year lower than an optimal choice. Compounded over decades, this can easily knock a third off the value of a pension.
Andrew McCaffery Group Head of Solutions, Aberdeen
“With bond yields near alltime lows, growth sluggish and equity valuations starting to look stretched, alternative assets deserve a much larger allocation in portfolios.”
Aberdeen is well known for its equity expertise, but over the last decade we have successfully extended our asset management capabilities across the full range of asset classes. As of March 2017 more than two thirds of our assets under management are invested in asset classes other than equities. Alongside equities and bonds, we now manage substantial assets in infrastructure, private equity, property, real assets, emerging market debt, hedge funds and loans, to name a few. The biggest new area is our multi-asset solutions business with over £90bn of assets under management1. This business is founded on the ability to conduct strategic asset allocation for clients. We have developed a substantial team of research and asset allocation specialists to do this work. This report is an indication of the detailed thinking and advice they provide. Perhaps the most striking conclusion of this report is the growing importance of alternative assets in asset allocation today. The traditional ‘balanced’ equity/bond approach to asset allocation is no longer the best option for many investors. With bond yields near all-time lows, growth sluggish and equity valuations starting to look stretched, alternative assets deserve a much larger allocation in portfolios. We expect assets like emerging-market debt or infrastructure to offer returns that are much higher than government bonds, but with risks that are only modestly higher; or to put it another way, returns that are nearly the same as equities, but with much lower risk profiles. While we can offer no guarantees, more aggressively diversified portfolios that include combinations of these alternative assets may offer significantly higher risk-adjusted returns than conventional equity-bond portfolios. We think this suggests a secular shift in portfolio allocation. The report demonstrates that today’s slower growth and low interest rates are the result of deep structural changes to the world economy that are likely to be with us for some time. In this environment, we suspect investors will continue to benefit from alternative sources of diversification, without the low returns on offer from developed-market government bonds. Finally, I’d highlight one more trend: previously our multi-asset clients have preferred to invest largely in liquid public markets – and this investment outlook reflects this focus. However, appetites are changing. More clients are prepared to sacrifice liquidity in favour of the higher returns available from unlisted, ‘private’ alternative assets. These assets are one of the most rapidly developing areas in Aberdeen’s investment capability. Future iterations of this report will expand to cover the special challenges and opportunities posed by private alternative asset classes.
Source: Aberdeen Asset Management PLC, Interim Report and Accounts 2017.
1
04
Foreword
Introduction This long-term investment outlook report sets out the economic and asset class views that inform our strategic asset allocation work for clients. In our view, successful asset allocation is about selecting the combination of assets that best meet our clients’ investment objectives, given the economic circumstances. These last few words are vital: economic circumstances can change in ways that have a dramatic impact on asset class returns. Often investors seem to forget this, assuming instead that the returns received in the past are a reliable guide to the future. This can be a mistake. Craig Mackenzie Senior Investment Strategist Lead Author
For example, if you were to assume that UK government bonds will deliver returns close to their 20-year average of 6%, you would make a substantial allocation to this class. But current economic circumstances are very different to those of the late 1990s and early 2000s. Policy interest rates in the UK are still near zero. 10-year gilt yields are only around 1%. This means that a 1% expected return is more realistic, suggesting a very low allocation to government bonds and, as we discuss in Chapter 1, it makes sense to look elsewhere for equity diversification. Long-term economic trends
“Strategic Asset Allocation must start with a view of where the economy is now, and how it might evolve over the long term.”
Historical investment returns reflect yesterday’s economic circumstances and are not a good guide to the future. So, for us, Strategic Asset Allocation must start with a view of where the economy is now, and how it might evolve over the long term. This requires an understanding of the forces that drive secular economic trends. We think about these forces on two distinct time horizons: long term and medium term. Our long-term horizon is five to 10 years in the future. On this horizon, our assumptions about economic developments are driven by secular factors: trends in demography, productivity, inflation and interest rates. How will ageing populations affect economic growth? How will today’s weak business investment affect future productivity? Will the global savings glut persist in depressing equilibrium interest rates? On this horizon, we also assume that many variables will revert to their long-term equilibrium values – for example, we assume equity valuations, credit spreads and exchange rates revert to fair value equilibrium levels.
“What economic policies will the Trump administration actually pass? Will China have a hard landing?”
These long-term factors are discussed primarily in Chapters 3 and 4, but we also consider fair value equilibrium values in our discussion of each asset class later in the report. We also consider a medium-term time horizon of three to five years. Over this horizon, markets are driven more by the familiar pattern of recession and recovery associated with the business cycle – together with related credit and policy interest rate cycles. This horizon is considered in Chapter 4. Business cycles are often regionally specific, so Chapter 4 considers the outlook for each region, as well as more idiosyncratic questions: What economic policies will the Trump administration actually pass? Will China have a hard landing? What will be the outcome of Brexit for the UK economy? Will Italy remain in the Eurozone?
aberdeen-asset.co.uk
05
Introduction continued Fig. A: Risk and return for UK investors Local
GBP
GBP Hedged
Asset
Local Currency
3Y
5Y
10Y
3Y
5Y
10Y
3Y
5Y
10Y
3Y 5Y Volatility Sharpe Ratio
UK Equities
GBP
5.5
5.5
6.1
5.5
5.5
6.1
5.5
5.5
6.1
17.2
0.28
US Equities
USD
7.2
5.5
5.3
8.7
4.6
3.5
5.9
4.2
4.1
16.8
0.21
Europe ex UK Equities
EUR
6.2
5.5
5.2
6.9
5.7
5.0
7.0
6.2
6.0
20.4
0.27
Japan Equities
JPY
4.6
4.7
4.6
5.0
5.1
5.1
5.1
5.4
5.9
22.1
0.22
Pacific ex Japan Equities
AUD
5.6
6.3
6.6
4.4
4.7
5.0
3.8
4.5
5.1
17.7
0.22
Emerging Markets Equities
Various
5.9
6.1
6.3
4.8
4.8
5.0
n/a
n/a
n/a
n/a
n/a
Global Equities
Various
n/a
n/a
n/a
7.4
4.9
4.3
5.8
4.7
4.8
16.7
0.25
UK Gilts (All Maturity)
GBP
1.5
0.9
1.1
1.5
0.9
1.1
1.5
0.9
1.1
5.9
0.05
UK Inflation-Linked Gilts
GBP
0.2
0.5
1.3
0.2
0.5
1.3
0.2
0.5
1.3
10.8
-0.01
US Treasuries (All Maturity)
USD
1.8
1.9
2.3
3.2
1.0
0.6
0.6
0.6
1.2
4.3
-0.01
Euro Govt Bonds (All Maturity)
EUR
0.8
0.6
1.1
1.4
0.8
0.9
1.5
1.3
1.9
4.0
0.17
Japanese Govt Bonds (All Maturity)
JPY
0.1
0.0
-0.1
0.5
0.3
0.4
0.6
0.7
1.2
3.7
0.01
Global DM Govt Bonds
Various
n/a
n/a
n/a
1.8
0.8
0.7
0.9
0.8
1.4
3.3
0.06
UK IG Bonds
GBP
2.8
2.4
2.8
2.8
2.4
2.8
2.8
2.4
2.8
7.5
0.23
US IG Bonds
USD
3.4
3.4
3.8
4.8
2.5
2.0
2.2
2.1
2.6
6.3
0.23
Euro IG Bonds
EUR
1.7
1.8
2.3
2.3
2.0
2.1
2.4
2.5
3.1
4.7
0.40
Global IG Bonds
Various
n/a
n/a
n/a
4.1
2.4
2.1
2.3
2.2
2.7
5.8
0.27
US High Yield Bonds
USD
4.6
4.8
5.3
6.0
3.8
3.5
3.4
3.4
4.0
9.8
0.29
Europe High Yield Bonds
EUR
2.5
2.4
3.0
3.2
2.6
2.8
3.3
3.1
3.8
12.3
0.20
EM Debt (Hard)
USD
3.5
4.9
5.5
4.9
4.0
3.7
2.3
3.5
4.2
9.0
0.32
EM Debt (Local)
Various
5.6
6.3
6.5
4.4
5.5
5.8
n/a
n/a
n/a
n/a
n/a
Senior Secured Loans
USD
5.8
5.4
5.5
7.2
4.4
3.7
4.6
4.0
4.2
8.7
0.39
Insurance Linked Securities
USD
3.9
4.3
5.1
5.3
3.4
3.3
2.6
3.0
3.8
8.3
0.28
UK Commercial Property
GBP
3.0
3.8
5.9
3.0
3.8
5.9
3.0
3.8
5.9
13.8
0.23
US Commercial Property
USD
5.1
5.3
5.8
6.6
4.4
3.9
3.9
4.0
4.5
13.4
0.25
Europe Commercial Property
EUR
2.7
4.3
5.9
3.4
4.4
5.7
3.5
5.0
6.7
13.7
0.32
Global Commercial Property
Various
n/a
n/a
n/a
5.1
4.5
4.8
3.9
4.5
5.4
10.0
0.39
UK REITs
GBP
4.2
4.6
6.0
4.2
4.6
6.0
4.2
4.6
6.0
23.5
0.17
Global Private Equity
USD
11.3
10.8
10.7
12.8
9.8
8.7
10.0
9.4
9.3
14.2
0.62
Global Private Infrastructure
USD
10.2
10.2
10.2
11.7
9.2
8.3
9.0
8.8
8.9
10.1
0.81
Global Private Real Assets
USD
5.9
6.0
6.4
7.3
5.0
4.5
4.6
4.6
5.1
19.0
0.21
UK Infrastructure Social
GBP
5.3
5.2
5.2
5.3
5.2
5.2
5.3
5.2
5.2
11.3
0.40
UK Infrastructure Renewables
GBP
5.9
5.8
5.8
5.9
5.8
5.8
5.9
5.8
5.8
11.3
0.46
Alternative Risk Premia
USD
5.5
6.0
6.8
7.0
5.0
5.0
4.3
4.6
5.5
10.6
0.37
Hedge Funds
USD
3.5
4.0
4.8
4.9
3.0
3.0
2.3
2.6
3.5
7.1
0.28
ABS - Mezzanine
USD
5.0
5.5
6.3
6.4
4.5
4.4
3.8
4.1
5.0
9.0
0.38
Commodity Futures
USD
1.5
2.0
2.7
2.8
1.0
0.9
0.3
0.6
1.5
21.6
0.00
UK Cash 3M LIBOR
GBP
0.5
0.8
1.7
0.5
0.8
1.7
0.5
0.8
1.7
1.1
0.14
USD Cash 3M LIBOR
USD
1.9
2.3
3.1
3.3
1.4
1.3
0.7
1.0
1.9
1.1
0.34
EUR Cash 3M LIBOR
EUR
0.0
0.3
1.1
0.7
0.5
0.9
0.7
1.0
1.8
1.1
0.35
Source: Aberdeen, March, 2017. Note: Volatility and Sharpe ratio refers to GBP Hedged. Return projections are estimates and provide no guarantee of future results. DM = developed market, IG = investment grade, EM = emerging market, REITs = real estate investment trust, ABS = asset-backed security, LIBOR = London interbank offered rate.
06
Introduction
Fig. B: Asset summary view Asset class
Summary view
Equities
Our long-term view is that equity returns, at around 5% pa, will be weaker than they have been in recent years. Valuations in developed markets are now elevated and our long-term growth forecasts are for sluggish growth. In the near term our favoured region is Europe, where we think the business cycle has further to run and profit margins more room for expansion. Though, if the Trump administration is able to get its corporate tax cuts through Congress, the US may again prove attractive. Emerging markets are relatively inexpensive. However, the risks of a material slowdown in China suggest allocations to India and other countries may offer better returns through a combination of strong domestic growth and low exposure to China risk.
Government bonds
While bonds yields are higher than they were in 2016, they remain near historical lows. Low starting yields make for low long-term bond returns. So our bond return forecasts remain extremely low – with the global average return of only 1.4% pa. In the short term we expect higher short-dated yields in the US, but we don’t expect as much movement at longer maturities, partly due to ongoing QE in Europe and Japan. The big picture is one of structurally low equilibrium interest rates, which suggest that low bond returns look here to stay. It’s hard to pick a preferred region. Returns in Europe and Japan are particularly low. The US is marginally more attractive – though less so after currency hedging costs are taken into account.
Corporate credit
The US credit cycle is now mature. Yet US credit spreads are still tight, offering a small premium for bearing credit risk. In addition, the (gently) rising government-bond yields we expect will erode returns, particularly for longer-duration investment-grade bonds. Our returns forecasts are modest (2.7% investment grade (IG), 4% high yield (HY)) – and a lot lower than a year ago, when spreads were much higher. Our favoured credit asset is senior secured loans (4.2%). Loans have floating rate coupons, so are insulated from rising interest rates, and offer relatively high yields.
Emerging-market debt (EMD)
Emerging-market government bonds are a relatively attractive asset class – particularly the localcurrency variety. Yields are currently at the high end of their range (7%) offering strong income return. With one or two exceptions, most of the emerging market economies covered by the standard EMD local-currency index are in good shape, with solid growth, controlled inflation and low government debt levels. Currency risk is an issue for local currency EMD, but on average, local currency is now fair value, so for long-term investors there is no longer a structural currency risk.
Property
Property has delivered strong returns for several years. Relative to government bonds, the yields offered by property are attractive in many markets – the exception is the US, where the gap has closed. Retail property is a major component of property indices. The existential threat to this segment from internet retail is a structural risk for property investors and suggests a more selective approach to the asset class. UK property also must contend with Brexit risk, and its risks to the UK economy and especially the London office market.
Listed alternatives
Returns from listed alternative assets are particularly compelling for investors facing low bond yields and mediocre expected equity returns. We like infrastructure, asset-backed securities and insurancelinked securities. These assets offer diversification from equities, relatively low risk and higher returns than developed market government bonds. Though it is important to note that with these asset classes, the choice of vehicle makes a big difference to investment outcomes.
Currencies
The dollar is now expensive using standard equilibrium exchange rate models. The gap in interest rates between the US and other developed market economies will keep it strong in the near term, but over the long term we expect negative currency returns for European investors in US assets. GBP, EUR and JPY are all cheap on an equilibrium exchange rate basis. We expect them to strengthen in the long term. Sterling’s fate is heavily linked to the eventual form of Brexit. A disorderly Brexit could see sterling weaken further. EM currencies are, on average cheap versus the dollar and around fair value versus other developed currencies.
Source: Aberdeen Asset Management, March 2017. Note: Expected returns are expressed hedged to GBP (except EMD which is unhedged) for a 5-year horizon.
aberdeen-asset.co.uk
07
Introduction continued The factors that drive asset class returns Macroeconomic factors are important for many asset classes, but they are by no means the only factors that drive returns. The second part of this report considers each asset class individually and discusses the specific factors which drive expected returns in each case. Understanding the factors which drive returns for each asset class is vital in forming our investment expectations. For example, profit margins are a key driver of equity returns. In the US these are currently at very high levels by historical standards. If you think profit margins must soon revert to their long-term average, you are likely to assume very low expected returns. If, on the other hand, like us, you think they will remain elevated (see Chapter 5), then you will be more optimistic. Each asset class has its own distinctive risk factors, but also shares factors in common with other asset classes. This brings us to the other main reason why it is important to understand the factors which drive returns: diversification. Two asset classes may look very different, but if they are exposed to the same risk factors, their returns are likely to be correlated, particularly in troubled markets. One of the most basic goals of asset allocation is to achieve the considerable risk-return benefits that are available from effective diversification. If diversification proves to be a mirage, because asset classes are exposed to the same underlying risk factor, then we will have failed. Understanding the common and differentiated exposure of each asset class to risk factors is therefore a central preoccupation of our work.
“At today’s prices, are we being sufficiently rewarded for bearing risk?” Valuation is another key question we ask when we consider risks for each asset class. At today’s prices, are we being sufficiently rewarded for bearing risk? Valuation depends both on expectations for future cash flows, and a consideration of the rate used to discount these cash flows. How will risk-free rates behave in the future? What risk premia are reasonable? These are questions we address in considering the fair value for each asset class.
08
Introduction
Our asset class chapters, Chapter 5 to Chapter 9, discuss each of the asset class’s, return drivers and our assumptions about future trends. We also discuss the methods we use to model expected returns. A very brief summary of our current views on each asset class is provided in the table (Fig. B).
“We are increasingly allocating to less familiar asset classes like emergingmarket government debt, infrastructure, loans, and insurance-linked securities.” This report covers the traditional liquid listed asset classes that have dominated our clients’ portfolios for decades – equities, government bonds, corporate credit and property. But, as we discuss in Chapter 1, we are increasingly allocating to less familiar asset classes like emerging-market government debt, infrastructure, loans, and insurance-linked securities. This report covers these newer asset classes. Our focus in this report remains on liquid asset classes. Illiquid assets like direct infrastructure, private equity and private debt, are increasingly important and we plan to include them in future versions of this report.
Dr. Craig Mackenzie Senior Investment Strategist Head of Strategic Asset Allocation Research
Report team This report was a team effort. Craig Mackenzie, who leads Aberdeen’s Strategic Asset Allocation research, was the report’s main author, but the credit belongs to a much wider group of individuals. The report’s economics chapters have benefited considerably from the insights of the economics group in the Economic and Thematic Research team (Lucy O’Carroll, Paul Diggle, Luke Bartholomew, Marta Palka). Chapter 4 (regional economic outlook) was based heavily on their forecasts. Thomas Laskey provided material for Chapters 5, 6 and 7 on fixed-income, and Jennifer Mernagh and Alex Tempier for Chapter 8 on alternatives. Alex is also responsible for marshalling the datasets used for the report’s charts and for managing the infrastructure used for returns forecasts and portfolio risk modelling.
We have benefited from conversations with asset class specialists in Aberdeen’s Equities, Fixed-Income, Alternatives and Property teams, as well as the insights of Aberdeen’s Tactical Asset Allocation group, and our Portfolio Engineering colleagues. Richard Dunbar, Head of the Economic and Thematic Research team, and colleagues Jodie Stewart and Caroline Armstrong have provided invaluable support throughout the project. Grateful thanks are due to Chris Hill and others in Aberdeen’s investment writing and production teams, who sharpened our prose. Finally, we acknowledge the invaluable experience of working with our innovative Diversified Multi-Asset team led by Mike Brooks. Mike has helped us to take the road less travelled in our search for investment returns.
aberdeen-asset.co.uk
09
Chapter
01 10 document_name
Strategic asset allocation in a low interest-rate environment
• Traditional asset allocation relied on low risk to diversify equity risk • Very low bond yields mean this approach may deliver poor returns • Better risk-adjusted returns are available for investors who diversify across a wider range of less familiar asset classes.
aberdeen-asset.co.uk11
Strategic asset allocation in a low interest-rate environment In a world of slower growth and low bond yields, conventional approaches to investment will increasingly struggle to fulfil investors’ long-term goals. With the main developed-market government bond index yielding around 1%, conventional bonds no longer offer a meaningful income return at low risk. Similarly, with a sluggish global economy and valuations starting to be a little stretched, in our view equities look unlikely to match the high rates of capital growth they have delivered in the recent past.
“The standard ‘balanced’ equity-bond portfolio mix is unlikely to be the best option for most investors.”
To get the ball rolling, this chapter summarises some of the key themes in this report and gives an overview of our favoured asset allocation approach for typical investors.
In this environment, investors are not well served by traditional asset allocation approaches. The standard ‘balanced’ equity-bond portfolio mix is unlikely to be the best option for most investors, for reasons that we explain below.
There is no sugar-coating it; we think investment returns from conventional balanced portfolios are likely to be on the low side in the future. Over the last 20 years, for a UK investor, global equity returns have averaged 8% and UK government bonds 6.5%3. A classic 60:40 balanced portfolio would have returned over 7% per year. Over the next decade, our modelling suggests 3.5% per year is a realistic expectation.
We think many investors will do better by diversifying their portfolios more widely, including a range of less familiar asset classes – for example, emerging-market debt, leveraged loans, infrastructure funds and insurance linked securities. It may offer a material improvement in expected returns, with less risk. Over the next 10 years, we expect the return from a 60:40 equity-bond portfolio to be only around 3.5% with an expected volatility of 11%; whereas our favoured, highly-diversified portfolio targets a return of around 5% per annum with an expected volatility of just 7.5%1. Needless to say, adopting a more ambitious diversification strategy that incorporates unfamiliar sources of return brings its own challenges: it requires an expert grasp of each asset class and solid understanding of the economic factors which drive their returns; as well as a sophisticated ability to monitor and model risk both for individual assets and for portfolios as a whole. With over £90bn of multi-asset funds under management2, Aberdeen has been able to develop a strong capability to support clients in navigating this unfamiliar terrain. We have built a set of sophisticated tools for risk modelling and return forecasting. Our strategic asset allocation research draws on an experienced Economic and Thematic Research team, a specialised Portfolio Engineering group, and a large array of asset class specialists across Aberdeen’s global business. As one of the largest active managers in Europe, we have considerable breadth of expertise to draw on.
Compounded over many years, the difference between 7% and 3.5% per year has a substantial impact on capital accumulation and retirement income. By way of illustration, if someone saves £10,000 a year for 30 years and reinvests investment income, a 7% investment return will result in a capital sum of £950,000, which in turn might provide an annual retirement income of around £38,000 (assuming a realistic 4% annuity rate). By contrast, a lower investment return of 3.5% per annum results in a total capital sum of £520,000, and a retirement income of just £21,000. To achieve the same level of income (£38,000), with an investment return of just 3.5%, one would need to save £18,000 per year, nearly twice as much as before. Needless to say, increasing personal saving by 80% would, to put it mildly, be challenging for many in today’s workforce, who must also pay-off student debt and save for a down-payment on a house. Is there any alternative? We have to accept that the last few decades have been a golden era for investors, delivering returns that are unlikely to be repeated. A 7% annual return is probably too ambitious a goal. But including a much wider range of assets and diversifying more aggressively provides a route for investors to do significantly better than they could if they relied only on equities and low-risk bonds: 5% per year, perhaps, rather than 3.5%.
This report provides a summary of our research and our investment outlook for each of the main asset classes used in the multi-asset portfolios advised by Aberdeen. It also includes a survey of the structural and cyclical economic forces which shape the investment landscape today. We hope this helps our clients with long-term planning, and also with their strategic asset allocation between asset classes.
We should add that, while this is a simple example based on an individual investor, the basic challenge is not very different for pension funds, insurance companies and other institutional investors. Under conventional investment strategies, low returns from equities and low-risk bonds may require higher contributions from scheme sponsors, or renegotiation of future commitments to safeguard scheme solvency.
Predicting returns is not easy. We know that our forecasts are likely to be imperfect. It is important that our clients do not simply take our returns expectations at face value, but also understand the thinking behind them so that they appreciate our underlying assumptions and world view. That is the aim of this report.
This is not news to the investment industry, which has been grappling with the prospect of lower future returns for several years now. To some extent the problem has been masked by the high returns achieved in recent years. These high returns are unlikely to continue. In today’s investment environment, past returns really are not a good guide to the future, and the challenge is likely to become more acute.
Figures are provided at the end of this chapter. Target returns are no guarantee of future results. 2 Aberdeen Asset Management PLC, Interim Report and Accounts 2017. 1
12
Aiming for the best outcome in a low-return environment
Strategic asset allocation in a low interest-rate environment
Annualised returns of FTSE All World total return index and BoAML Gilt index 31/12/1996-31/12/2016.
3
The structural causes of lower expected returns
Fig 1.1: 10-year average real GDP, historical and projected
Why do we expect lower returns from conventional asset classes in the future? Equity returns will likely be lower due to a combination of sluggish global economic growth and the fact that, on most measures, equity valuations are now stretched, particularly in the US.
1987-1996 1997-2006 2007-2016 2017-2026
“We expect low yields to persist for a decade or more, due to structurally low ‘equilibrium’ real interest rates.” In our view, bond returns will be lower, partly as a mathematical function of the low yields they offer today. But this is not a temporary blip: we expect low yields to persist for a decade or more, due to structurally low ‘equilibrium’ real interest rates. The problem is not one of low government bond returns for a year or two which can be endured; it is low bond returns for a decade or more. This requires a strategic response. We explore the secular trends that cause slower economic growth and lower interest rates in more detail in Chapter 2 and Chapter 3. For now a summary will suffice.
Slower trend growth At a high level, the long-term economic growth rate is determined by two main factors: the number of people employed in the economy, and how productive they are. Economic growth is faster if you add more workers or increase the output they generate per hour worked; it is slower if the workforce shrinks or productivity growth slows. For the last 100 years or so, the labour force in most large economies has been growing a steady rate of one or two percent a year. But declining fertility rates in recent decades mean that populations in many economies are now growing much more slowly. In fact, in many countries in Europe and East Asia, including China, the labour force is actually shrinking, or soon will be. The tailwind of rapid population growth has now become a headwind. Slower trend growth is the likely result. Similarly, in the first two thirds of the 20th century, productivity improved at a rapid pace in advanced economies. But around 1970, productivity growth slowed significantly. There was a brief burst of faster growth in the late 1990s and early 2000s, but productivity growth has slowed again since the financial crisis, and is now slower than it was even in the 1970s. There is much debate about why this slowdown has happened and what might happen next (see Chapter 2). In short, there are some reasons to hope that productivity growth will rise from its current abysmal level of 0.5% per year, but it is unrealistic to expect it to compensate fully for stagnation in the labour force. The combination of a slower labour force growth and relatively modest rates of productivity growth results in lower rates of trend GDP growth than we have been used to.
“Slower labour force growth and relatively modest rates of productivity growth results in lower rates of trend GDP growth.”
US
3.0
3.3
1.3
1.7
UK
2.5
3.0
1.1
1.9
Germany
2.3
1.5
1.2
1.0
Japan
3.3
1.0
0.5
0.7
China
10.2
9.5
9.0
4.4
World
4.4
3.9
3.4
3.4
Source: Aberdeen, Oxford Economics, April 2017. Note: Percentage changes YoY. Real GDP calculated in local currency except for World in PPP USD (projections start in 2016 for World).
China, Trump and other factors Another important factor affecting the global economy in the medium-term (the next five years or so) is China. China’s economy is now vying with that of the US to be the biggest in the world and it is still growing strongly - at over 6% per year. As a result, China alone is responsible for around one third of global GDP growth. But China’s current growth rate of nearly 7% is unlikely to be sustained for much longer. It is based on rapid credit-financed growth of investment in property, infrastructure and other fixed assets. History suggests that periods of exceptionally fast credit growth are usually followed by periods of much slower growth as the economy reduces leverage and balance sheets are repaired4. China is unlikely to prove an exception to this rule.
“Markedly slower growth in China will prove a challenge for the rest of the world.” China has many strengths. This adjustment process will, if successful, lay the foundation for more sustainable growth in the future. But markedly slower growth in China will prove a challenge for the rest of the world - particularly for those commodity economies most dependent on Chinese demand. Global growth rates will likely be rather lower as a result. The economic policies of the Trump administration could also have a big effect on the global economy in the medium term, in either direction. Planned tax cuts and reforms to corporate tax and regulation, coupled with new infrastructure investment, could raise the US productivity rate, allowing faster and more sustainable economic growth than we expect. Though optimism about the tax cut agenda has waned as the Trump administration has failed to establish legislative momentum. Of course, China and the US are not the only regions that might affect global growth in the medium term. We review the outlook for all regions in Chapter 4.
Maliszewski et al. (2016) Resolving China’s Corporate Debt Problem, International Monetary Fund.
4
aberdeen-asset.co.uk
13
Strategic asset allocation in a low interest-rate environment continued Lower interest rates
This ‘zero lower bound’ problem is a reason why the recovery from the financial crisis has been so painfully slow. With low equilibrium real interest rates likely to be an ongoing feature of the economic landscape, this zero lower bound problem may reoccur, resulting in more sluggish recoveries in the years to come, and producing average growth rates that fall short of their already modest potential.
Fig. 1.2: 10-year nominal yields (%) 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 -1
Disappointing economies beget unstable politics
1990 US
1994 Japan
1998 France
2002 Germany
2006 Italy
2010
2014
UK
Source: Oxford Economics, February 2017. Note: Chart shows yield on 10 year maturity government bond in each country.
Lower growth is not the only challenge facing investors. Many of the same structural trends that are depressing economic growth are also depressing interest rates. For economists, the equilibrium real interest rate is the price that balances the markets for savings and investment. Ageing baby-boomers and rapidly emerging economies are generating high levels of saving. On the other hand, slower growth, less appetite for business and public sector investment, and lower costs of capital equipment, together mean that there is not enough demand to use these savings for capital investment projects. As a result we have a ‘savings glut’ which depresses equilibrium interest rates (see Chapter 3). This helps explain the long-term decline in global interest rates from the 1980s to today’s historically low levels. Although interest rates may rise marginally as the global economy strengthens, these factors will keep bond yields low for many years.
“While government bonds may continue to provide some diversification, they no longer provide much of a return.” For decades investors have relied on government bonds to provide a low risk source of income, and growth-oriented investors have used them as the principal means to diversify their risk from equities. While government bonds may continue to provide some diversification, they no longer provide much of a return. Gilts have provided an average return of 6.5% over the last 20 years. The return we expect over the next 10 years is between 1 and 2%.
Secular stagnation While low interest rates are a problem for investors in their own right, they are also a problem for the wider economy. When faced with a recession, central banks reduce interest rates aggressively to stimulate recovery. In previous recessions, central banks have reduced interest rates by three or four percentage points. But when central bank interest rates are at or near zero, there is little room for this kind of stimulus. Of course, there are alternatives such as quantitative easing and negative interest rates, but there are worries that these are less effective and create their own problems.
14
Strategic asset allocation in a low interest-rate environment
The recent success of anti-establishment political parties seems to owe something to the disappointing economic performance of modern economies. The idea that a substantial section of the population has been ‘left behind’ may have been a contributor to the result of the Brexit referendum and Donald Trump’s election. Persistent economic problems in Europe are contributors to the popularity of the French National Front, Italy’s Five Star Movement and populist parties in other countries.
“The fear from an investment perspective, is that anti-establishment parties will pursue populist economic policies which further undermine growth rates.” Whether this is a problem depends on one’s political point of view. The fear from an investment perspective, is that anti-establishment parties will pursue populist economic policies which further undermine growth rates. For example, anti-globalisation sentiment may result in a surge in protectionism and a decline in global trade, resulting in even slower growth.
The need for a policy response Today’s politically volatile mix of low growth, income inequality, and disruptive economic change makes for an uncertain context for long-term investors. There is a compelling need for more robust and effective policies directed at promoting faster, more inclusive growth. In some areas there is already wide-spread agreement about solutions, but a lack of political support for implementation. For example, the IMF and the OECD have advocated largescale public infrastructure spending that increases productivity, underpins aggregate demand and reduces the savings glut, while paying for itself. But despite some positive rhetoric, infrastructure programmes of sufficient scale still seem a long way off.
“With labour forces growing more slowly, faster productivity growth is needed to compensate.” In other areas, there is still work to be done to identify effective policy solutions. Productivity is a case in point. With labour forces growing more slowly, faster productivity growth is needed to compensate. Instead, productivity growth is lower than ever. But there is little consensus about how to solve the productivity
‘puzzle’. The same goes for the ‘left behind’. Rapid economic change tends to create winners and losers. Modern economies have not been effective at mitigating the pain for losers and re-employing them in productive ways. Failure to tackle these problems will likely result in lower growth and more unstable politics. Among other things, this is not a good environment to generate high long-term investment returns. (Returns that will provide for the retirement incomes of an ageing population.) For this reason, if for no other, Aberdeen is strongly supportive of efforts to develop policies that enable faster, more inclusive growth – starting with a large-scale global programme of infrastructure investment.
Elevated valuations? Another potential hurdle facing investors is the fact that asset prices have appreciated significantly in recent years and most assets classes are now looking on the expensive side relative to their historical levels. Since the financial crisis equity returns have been impressive. The FTSE All World Index is up 180% or 14% per year. Some of this performance is down to earnings growth, but returns have been boosted by expansion in valuation multiples as the chart below (Fig. 1.3) shows. Fig. 1.3: (%) Historical equity valuations (%) Returns 50
40
As we explained previously, we expect interest rates to remain unusually low, so these high historical valuations may persist for some time. We think it is prudent to assume that valuations will revert to their historic levels in the longer term, but we do not see today’s valuation as an immediate threat for investors.
A more diversified investment strategy In summary, we can expect slower growth in the future than we experienced in the past, together with persistently low interest rates. On a 10 year horizon, we expect equity returns of 4.5-5.5% rather than the 30 year average of 7-8%. Our expected government bond returns range from 0-2% compared to a longterm average of 6-7%. This means that we might expect a traditional balanced investment portfolio to deliver a return of around 3.5% versus a 7% average over the last 20 years. While the fundamental economic facts of life mean that we should expect somewhat lower returns than we have enjoyed in the past, we think that investors can do better than the meagre returns we expect from traditional balanced portfolios. Our favoured approach to asset allocation is different from the traditional approach in two areas. Firstly, it replaces low-return government and investment-grade corporate bonds with higherreturn diversifiers such as listed infrastructure and higher-yielding fixed-income. Secondly, it reduces the equity weight in the portfolio. Equities have a relatively high level of risk, but, given lower growth prospects and somewhat elevated valuations, their expected returns are no longer as competitive as they once were. Equities remain our largest single asset class exposure, but the level is closer to 30% than the 60% associated with a traditional balanced portfolio. The difference is made up by alternative assets with lower risk but similar returns.
30 20 10 0 -10 -20 -30 -40 -50 2001 US
2004 2007 UK Europe ex UK
2010
2013
2016
Source: Bloomberg, Aberdeen, February, 2017. Note: Chart shows the level of Aberdeen’s equity valuation basket versus its 15 year average. 0% is ‘fair value’.
Equities are no longer cheap on a historical basis anywhere, and valuations are somewhat stretched in the US and Europe (see chart Fig. 1.3). High valuations today normally mean lower returns tomorrow. The story is similar for other assets like property and infrastructure.
“Investors are still being rewarded for bearing risk.” But this is not the whole story. While most assets look expensive relative to their historical valuation levels, their valuations generally look much less stretched compared to government bonds. The ‘risk premium’ earned by investors over risk-free interest rates is still fairly substantial for equities and many other risk-assets. Investors are still being rewarded for bearing risk.
Alternative diversifiers Prior to the 2008 financial crisis, investors had invested in a variety of unconventional asset classes that they thought offered them diversification from equities. In the event, they were disappointed to discover during the 2008-09 equity bear market this diversification was largely illusory. These ‘diversifiers’ were highly correlated with equities. Armed with this experience, investors are rightly now rather more sceptical about apparent claims to diversification. We do extensive research to test these claims are robust.
“The asset classes that provide the most robust diversification from equities are those whose underlying cashflows are insensitive to the health of the economy.”
aberdeen-asset.co.uk
15
Strategic asset allocation in a low interest-rate environment continued The asset classes that provide the most robust diversification from equities are those whose underlying cashflows are insensitive to the health of the economy. Equity risk is driven primarily by market expectations of volatility in the economy. The worst equity returns tend to occur during economic recessions. So, a more reliable way to diversify equity risk is to invest in assets that are insensitive to changes in the economy. Infrastructure is an example of an economically insensitive asset. Infrastructure funds hold portfolios of public-private partnership hospitals, schools and renewable energy infrastructure: wind and solar farms. The cashflows from these assets are derived from longterm, government-backed contracts and subsidies. This makes them relatively insensitive to changes in the economy. Their correlation with equities is therefore low. As the chart below (Fig. 1.4) shows, these assets do not tend to sell-off when equity markets experience strongly negative returns. Observe how well social infrastructure has fared when equities and real estate have experienced large falls in value.
Catastrophe bonds are another good example. These bonds are issued by insurance companies who have, for example, written insurance policies on properties in Florida that may be damaged by a severe hurricane. In return for sharing the risk of loss from a bad hurricane season, catastrophe bond investors are paid an annual insurance premium. These investments generate an attractive return in most years, but lose money in the event of a severe hurricane season. The key point is that these risk events are unlikely to occur at the same time as stock market crashes, so the returns from catastrophe bonds show low correlation with equity returns. They are good diversifiers. Returns are currently on the low side for the most popular catastrophe bonds, given strong investor interest and lack of major risk events that push up premiums, but remain attractive in the more sophisticated vehicles. Other examples of alternative diversifiers are asset-backed securities, emerging-market government bonds, alternative risk premia, senior secured loans, and certain kinds of property.
Fig. 1.4:inReturns periodsequity of negative equity Returns periods ofinnegative sentiment (%) sentiment (%) 20 10 0 -10 -20 -30 -40 -50 -60 -70
Oct 07 - Mar 09 Global equities
Apr 10 - Jun 10 Global REIT
Jul 11 - Oct 11
Mar 12 - Jun 12
Jul 15 - Sep 15
Dec 15 - Feb 16
Social infrastructure
Source: Aberdeen, Bloomberg, March 2017. Note: Calculated as the total return (share price plus gross dividends) over the period. Social infrastructure is based on HICL Infrastructure. Past performance provides no guarantee of future results.
Higher returns at lower risk As the chart overleaf shows (Fig. 1.5), many of our diversifying assets also potentially offer returns that are nearly as high as equities, but with much lower risk. For example, another asset class that we currently advise many of our clients to hold in their multi-asset portfolios is emergingmarket government bonds (or EM debt for short). While this asset class has had a chequered history, many EM governments have learned from past mistakes. Government debt levels are generally rather lower than in developed markets and EM debt is increasingly issued in local currency and for the countries included in the benchmark EM debt indices, credit risk is low. We particularly like EM debt that is issued in the country’s domestic currency, primarily because yields are rather higher. The downside is that there is a risk of currency depreciation. Our long-term fair value currency models suggest that EM currencies are no longer expensive compared to a developed-market currency basket, so there is reduced structural depreciation risk. And while a full currency hedging strategy is not realistic for EM local currency debt, there are ways to mitigate currency risk. 16
Strategic asset allocation in a low interest-rate environment
“EM local currency debt offers an equity-type return for around half the volatility.” The volatility of EM bonds ranges between 7 and 13% depending on base currency. This compares to around 18% for developedmarket equities and 30% for EM equities. With expected returns about the same level as equities, EM local currency debt may offer an equity-type return for around half the volatility. EM debt also has a somewhat lower correlation with equities, so there is an additional diversification benefit too. Other examples, which offer equity-like returns with less risk are senior secured loans, asset backed securities and listed infrastructure. As a large, global, active asset management company, Aberdeen manages dedicated funds for emerging-market debt, leveraged loans, infrastructure, smart beta and other alternative assets. This is helpful when assessing the risk and return characteristics of these less familiar asset classes. This report devotes significant space to these asset classes in Chapters 6 and 7.
Fig. 1.5: Expected asset risk and return (%) 7 Europe ex UK Equities
6
Emerging Markets Sovereign Debt (Local)* UK Listed Social Infrastructure
Expected return pa
5
Emerging Markets Sovereign Debt (Hard)
3 2
Japan Equities Emerging Markets Equities*
Pacific ex Japan Equities US Equities UK Commercial Property Global High Yield Bonds
Senior Secured Loans
4
UK Equities
Insurance Linked Securities Global IG Bonds Euro Govt Bonds
1
UK Gilts US Govt Bonds
GBP Cash
0 0
2
4
Equities
Rates
6 Credit
8 Alternatives
Commodity Futures 10
12
14 16 18 Expected Volatility
20
22
24
26
28
30
32
Cash
* GBP Unhedged. Source: Aberdeen, March 2017. Note: Returns are over five years (GBP hedged) on a per annum basis. Return projections are estimates and provide no guarantee of future results.
Equity preferences
Adding illiquid assets
Lower expected returns mean that equities currently receive a lower proportion of asset allocation than they have in previous years. However, they remain the single biggest asset class in most of our multi-asset portfolios. But rather than 50-60% of a conventional ‘balanced’ portfolio, they comprise roughly one third of our diversified multi-asset portfolios.
Most investors rely exclusively on assets that are listed on markets but, as we discus in Chapter 7, higher returns are often available from unlisted or privately held assets like private equity, private infrastructure, direct property and private debt. Private assets typically offer higher returns than their listed versions because investors receive an ‘illiquidity premium’ in compensation for losing the ability to release their capital at short notice. This premium typically adds one or two percent to returns.
Within equity regions, our current preference is for European equities – the European business cycle is picking up steam and has much more room to run than the US cycle. Japan faces the headwind of slower growth, and the US is now rather expensive. It is possible that the US might do better, if the Trump administration manages to introduce meaningful reforms. This is still uncertain. We also like some EM markets, but given our concerns about risks of a China slowdown, we prefer markets with less exposure to Chinese demand. We also like ‘alternative risk premia’ and ‘smart-beta’ strategies in equities. There is a large amount of academic evidence that equity portfolios that exploit market anomalies such as low volatility, value, quality and momentum can offer more attractive risk-return characteristics.
“We allocate a share of equity exposure to these quantitative equity strategies.” We allocate a share of equity exposure to these quantitative equity strategies, primarily with the goal of reducing risk. Our favoured smart-beta equity strategy has a volatility around 15% lower than a global equity index, with a higher expected return.
Investors sometimes mistakenly believe that the fact that private assets are illiquid and must be held for five to 10 years, means that they get no cash return in the short-term. Naturally enough, this would be unattractive for investors who require income in the near term. In fact, many private assets offer a substantial income return during the period they are held. For example, private infrastructure funds might pay their investors a dividend of 5-6% per annum. In the past, accessing private assets has been the preserve of large institutional investors who can become limited partners in a private equity fund. However, it is increasingly possible to access some of the higher returns available from private assets through more liquid vehicles. In the UK, many investment trusts hold private assets and are accessible to individual investors. Infrastructure investment trusts are a good example. Investment trusts do not require investors to lock up their money, even though the underlying assets are themselves illiquid. They do this by allowing their shares to trade on the stock market with the share price at a premium or a discount to net asset value. In the following example below we exclude these assets from our opportunity set, given many investors are unable to hold illiquid assets; including them would add another percentage point or so to annual expected returns.
aberdeen-asset.co.uk
17
Strategic asset allocation in a low interest-rate environment continued Fig. 1.6: Better results are possible from a more radically diversified portfolio Asset Class
Traditional balanced
Global Equities
Modern balanced
Weight %
Expected return %
Weight %
Expected return %
60
4.8
50
4.8
Global Equities Low Volatility
Highly diversified Weight % 30
Expected return % 4.8
Volatility %
Sharpe Ratio
16.7
0.20
14.2
0.23
Global DM Govt Bonds
25
1.4
15
1.4
3.3
(0.03)
Global IG Bonds
15
2.7
15
2.7
5.8
0.22
10
4.1
Global High Yield Bonds
10
4.1
11.8
0.22
EM Debt (Local)
15
6.3
8.4
0.57
Senior Secured Loans
10
4.2
8.7
0.32
ABS - Mezzanine
5
5.0
9.0
0.39
Insurance Linked Securities
5
3.8
8.3
0.28
5
5.4
10.0
0.39
10
5.1
8.4
0.43
Hedge Funds
5
3.5
4.0
0.52
Other Diversifiers
5
5.5
10.6
0.38
*
Global Commercial Property
10
5.4
UK Infrastructure Social
Portfolio Total Return (%)
3.6
3.9
4.9
Portfolio Volatility (%)
10.8
10.3
7.4
Portfolio Sharpe Ratio
0.19
0.23
0.45
* Unhedged, expressed against a basket of developed-market currencies. Source: Aberdeen, March 2017. Note: Returns and volatilities are in percent and based on our 10-year horizon (GBP hedged on a per annum basis) standard forecasts for market benchmark, they include no assumption of additional returns from active manager skill. Returns are gross of fees and does not reflect investment management fees. Had such fees been deducted, returns would have been lower. Expected return is not an indication of future results.
Aberdeen Solutions manages a wide range of portfolios to meet a variety of different needs. The diversified growth portfolio above is a typical example of our favoured multi-asset approach. This kind of portfolio won’t suit everyone, but aims to meet the needs of investors who previously invested using a typical balanced equitybond approach. The portfolio has no allocation to government bonds or investment-grade credit, a lower exposure to equities, and a much wider range of diversified sources of returns.
“The addition of each diversifier to the portfolio lowers the portfolio risk.” One attractive feature of the diversifying asset classes that we hold in our portfolios is that their returns not only have a low correlation with equities, but they also have a low correlation with each other. The addition of each diversifier to the portfolio lowers the portfolio risk. Given their individual returns are reasonably attractive, this means it may be possible to achieve a lower-risk portfolio with the same level of expected returns offered by higher-risk equities.
We believe, that in an environment of very low government bond yields and sluggish global growth, this portfolio may offer rather better return prospects than traditional balanced strategies, while preserving the defensive characteristics that, in the past, were provided by developed-market government bonds. This approach also has the benefit of a natural bias towards assets with reliable cashflows and hence an annual income stream to investors of 4.5%. Our 10-year annualised expected return forecast for the portfolio above is around 5% - roughly the same as equities. But risk is less than half that of equities. Using our long-term risk model, portfolio volatility is estimated to be under 7%, little more than that offered by government bonds (volatility of 4%), nearly half that of a 60:40 equity-bond portfolio (13%), and well under half of that offered by equities on their own (17%).
“More aggressive diversification strategies may deliver equity-like returns with much lower risk.” Our conclusion is that our more aggressive diversification strategies may deliver equity-like returns with much lower risk.
18
Strategic asset allocation in a low interest-rate environment
aberdeen-asset.co.uk
19
Chapter
02 20 document_name
Long-term growth
• Ageing and shrinking working-age populations will result in lower rates of global growth • Sluggish productivity growth should improve, but not by enough to offset worsening demographics • Low equilibrium interest rates make it harder for central bankers to stimulate economies • Poor economics makes for volatile politics, more concerted growth-oriented policies are needed.
aberdeen-asset.co.uk21
Long-term growth Introduction In forming our views about asset class returns, we start by reviewing the long-term economic context in which those returns will be generated. As this chapter explains, today this is dominated by marked changes in demographic trends around the world, together with hard questions about the future rate of productivity growth, and the extent to which we face ‘secular stagnation’. Our conclusion is that economic growth will be lower in the future than it was in the decades prior to the financial crisis, but there are reasons for hoping that it won’t be quite as bad as the doomsayers predict. In the long run, returns from equities and many other risk assets are driven by economic growth. Growth in corporate revenues and earnings are the key driver of equity returns. For equity markets as a whole, this is ultimately derived from growth in the economy. Economic growth is not the whole story – for equities, trends in profit margins, share issuance, and valuation also matter – but over the long term economic growth is the main engine of returns. Judging how fast economies are likely to grow is therefore a fundamental starting point for our return forecasts. We distinguish between two timeframes. Over the long term (5-10 years and beyond), our growth forecasts are driven mainly by our view on the long-term growth potential of economies. This is determined by structural trends in demographics and labour productivity. In the short-to-medium term (1-5 years), growth may deviate from potential over the business cycle. This is a key determinant of medium-term returns. Equity bear markets tend to coincide with recessions, and equity bull markets with economic recoveries. Chapter 4 discusses our view on the state of the business cycle and the medium-term prospects for regional economies. In this chapter, we consider the long-term drivers of growth, starting with demographic factors.
Labour-force growth
Fig. Ageing populations (share above Aging2.1: populations (share above 65 years old) (%)65 years old) (%) 35
30 25 20 15 10 5
1960 US
2020
2030
2040
Fig. 2.2: Developed markets’ growth labour-force growth (%) Developed markets, labour-force (%) 2.5
“Much of the world is in the early stages of a major demographic transition, which is likely to result in a global economy that grows more slowly.”
-0.5
Long-term growth
2010
As a result of this shift in working-age population growth, it is highly likely that the labour force in these countries will either shrink, or grow more slowly than in the past. This will significantly reduce the rate of potential GDP growth.
2.0
Populations are ageing as people live longer and the ‘baby boomers’ reach retirement. Fertility rates are falling as people get wealthier and cultures change: fewer children are being born. Working-age populations in much of the developed world and in large parts of East Asia are shrinking, or soon will be. In countries
1970 1980 1990 2000 Japan Germany
Source: Aberdeen, OECD, Oxford Economics, February 2017. Note: Population ages 65 and above (% of total).
One of the most important factors driving long-term economic growth is the change in the size of an economy’s labour force. Crudely speaking, the faster the growth in the number of workers, the faster the economy can grow. As we discuss below, changes to productivity are also important, but demographic change is likely to be a key factor in the years ahead.
The great demographic transition Much of the world is in the early stages of a major demographic transition, which is likely to result in a global economy that grows more slowly.
22
where working-age populations may not shrink outright, such as the US and the UK, they are set to grow much more slowly. This shift is unprecedented in modern history.
1.5 1.0 0.5 0.0
-1.0 -1.5 1981 1986 1991 1996 2001 2006 2011 2016 2021 2026 2031 2036 Japan
Germany
US
UK
Source: Oxford Economics, February 2017. Note: percentage change in YoY labour-force total.
Japan was the first country to experience this transition, with demographic changes resulting in labour-force growth turning negative in the late 1990s and bouncing around zero ever since. Japan’s subsequent experience is informative. As expected, economic growth has slowed significantly. This has not been good for company revenue growth or equity investment returns, as we would expect.
The working-age population in certain European countries is also poised to start to shrink over the next decade. The decline is concentrated in Germany and some southern European countries, with France and the UK doing better. This suggests a significantly slower potential growth rate for the Eurozone than seen pre-crisis. East Asia is following Europe, although the change is rather more dramatic for this region, which until fairly recently experienced much faster working-age population growth. In China, the change has a lot to do with its recently abandoned ‘one child’ policy. But globally it is a function of growing wealth and women joining the workforce, both of which tend to result in smaller families. This is particularly important when thinking about future returns for emerging-market equities. The most popular equity-market benchmarks have very high exposure to countries where the labour force is expected to shrink – Chinese, Korean and Taiwanese companies make up over 50% of the MSCI EM equity index. Fig. Asia labour-force Asia 2.3: labour-force growth (%) growth (%) 5
4 3 2
Immigration Immigration is an important factor and a major reason why working-age populations are forecast to continue growing in the US and UK. However, tolerance for immigration can change. The recent electoral success of populist anti-immigration movements suggests that tolerance appears to be declining in many places, potentially reducing the positive future impact of immigration on labour-force growth. In the long term, given the rapidly growing populations in Africa and the extent of persistent income inequality between countries, it seems likely that migration pressures are only likely to increase. Perhaps Europe will eventually find a way to come to terms with African immigration as a solution to its demographic problems? But in the current environment, it seems unwise to make any ambitious assumptions on this front. Labour-force participation rate The size of the labour-force also depends on what proportion of the working-age population participates in the workforce. In the long run, participation is driven by a combination of cultural and policy changes.
1 0 -1 -2
It is the nature of demographics that these trends are already determined to some extent: the workforce of 20 years’ time has already been born. However, there are some ways that countries can compensate for lower birth rates – by allowing more immigration, increasing labour-force participation amongst the working-age population, and encouraging older people to remain at least partly attached to the workforce.
1981 1986 1991 1996 2001 2006 2011 2016 2021 2026 2031 2036 India
China
South Korea
Source: Oxford Economics, February 2017. Note: percentage change in YoY labour-force total.
“Labour forces will not fall in all countries. There are some stand-out exceptions among emerging markets, especially India.” Labour forces will not fall in all countries. There are some standout exceptions among emerging markets, especially India, Indonesia, and much of sub-Saharan Africa. We expect this stronger population growth to lead to more rapid GDP growth in these countries, which may be of interest to investors. However, it is notable that these rapidly growing economies currently form a very small part of the global equity index. Despite their rapid rate of growth, this will change only gradually. Demography is not necessarily destiny If these trends develop as projected and if productivity fails to come to the rescue, then economic growth in much of the world will be lower in the future.
“Countries can compensate for lower birth rates – by allowing more immigration, increasing labour-force participation and encouraging older people to remain in the workforce.”
For example, one of the biggest changes in the last 100 years has been the rapidly increasing participation of women in the paid labour force. This rapid change has given a major one-off boost to potential GDP growth in many countries. While this obviously cannot be repeated, some economies still have fairly low rates of female labour-force participation. For example, in Italy and Japan there is some scope to achieve faster growth by encouraging greater female participation. As the chart (Fig. 2.4) shows, Germany has been particularly successful at offsetting its demographic challenges by increasingly labour-force participation, mainly by encouraging more women to join the work force. Fig. 2.4: Diverging participation (%) Diverging labour-forcelabour-force participation (%) 80 78 76 74 72 70 2000 UK
2002 US
2004 Germany
2006
2008
2010
2012
Japan
Source: OECD, February 2017. Note: Labour-force participation rate (15-64 year-olds).
aberdeen-asset.co.uk
23
Long-term growth continued The size of the labour force is also dependent on the age at which people choose to retire. Given longer expected life spans and the decreasingly physical nature of work, people may be able to work longer. On the other hand, governments are raising the age at which state benefits and pensions are paid. This can delay the effects of demographic change, though it is not always popular. Labour force participation can also be a function of the business cycle, in particular if the economy suffers an extended period of high unemployment. Years of failing to find employment often causes people to give up. The US is a case in point: in the last 15 years, as the chart below (Fig. 2.4) shows, US labour-force participation rates have fallen significantly. This represents a loss of several million potential employees. The country would enjoy faster growth if these workers were enabled to re-join the workforce.
“US labour-force participation rates have fallen significantly.”
Economies grow either because the total number of hours worked increases or because the output produced each hour (labour productivity) increases. Demographic change largely accounts for the former, as average hours worked is relatively stable and depends on some of the same factors. But what are the prospects for productivity growth? Fig. 2.5: Slowdown of US labour productivity growth (%) 4
3 2 1 0
1968
1976
1984
US Recession Indicator (NBER) 10 Year moving average
1992
2000
2008
2016
Productivity growth
Source: Conference Board (TED), November 2016. Note: Growth of Labour Productivity per hour worked, percent change.
“In recent years, productivity growth has been dismal, averaging just 0.4% since 2010.” As the chart (Fig 2.5) shows, productivity growth in the US has varied significantly over time. Between the end of the second world war and 1970, labour productivity grew at 2.6% per year.
24
Long-term growth
Fig. 2.6: Slowdown of OECD labour(%) productivity (%) Slowdown of OECD labour productivity 4
3 2 1 0 -1
-3
Productivity
1960
The productivity slowdown is not confined to the US. Similar trends are observed in all advanced economies.
-2
Part of Donald Trump’s successful pitch to the ‘Rustbelt’ was that he would ‘bring jobs back’. It seems unlikely that his focus on renegotiating trade agreements is the best way to achieve this. More promising are his intentions to reform the tax system and invest in infrastructure. More effective programmes to enable workers to develop news skills and, perhaps, to move to places where the jobs are more plentiful would also be helpful.
-1 1952
Since then, with the exception of a brief period starting in the late 1990s, it grew at only 1.7%. In recent years, productivity growth has been dismal, averaging just 0.4% since 2010.
1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Japan
UK
US
Euro Area (19 countries)
Source: OECD, March 2017. Note: annual change in output per hour worked.
Should we assume productivity growth remains at the current dismal levels? Would a return to long-term average growth rates be a reasonable forecast, or should we consider forecasting even higher growth to take productivity levels back to pre-crisis trends?
“The high levels of productivity growth achieved in the post-second world war period were truly exceptional and unlikely to be repeated.” Historical productivity in the US Perhaps the best known recent work on this question is Robert Gordon’s analysis of the history of American growth. This suggests that the high levels of productivity growth achieved in the postsecond world war period were truly exceptional and unlikely to be repeated. This period brought the mass adoption of new technologies – in particular, electricity and electric motors. Electrification was a truly revolutionary technology that had widespread effects on the economy. In particular, there was radical improvement to the productivity of factory production lines, turbocharged by massive government-funded capital investment in factory modernisation during the second world War that was then turned to peacetime use. In addition, there was rapid improvement in the education of the workforce as a much greater proportion of the population went to college or university – including a vastly greater percentage of the female population. These factors allowed an extraordinary pace of improvement in productivity. However, this pace of improvement has not been sustained. Once most of the population with the aptitude for university level education are receiving it, it is much harder to increase further rapid gains in the population’s educational attainment. The scale of war-time government-funded capital investment in industrial infrastructure is also unlikely to be repeated.
This does not mean the end of innovation or that we will never see faster productivity growth - with artificial intelligence around the corner, this seems very unlikely - but it is a valid reason for not basing a productivity forecast on the long-term average. The digital revolution The brief return to faster productivity growth in the late 1990s and early 2000s is worth highlighting. It was largely driven by the universal adoption of computers and computer networks in the business community, as well as the extremely rapid improvement in telecommunications, computer processing power and software that took place during this period. A sector-by-sector analysis of productivity growth shows that the sectors where productivity improved fastest were those that use information technology most intensively.
“The biggest productivity improvements typically come during the relatively short period when a technology goes from early commercialisation to full-scale mass adoption.” However, once businesses were fully computerised, the subsequent productivity growth rate slowed. The biggest productivity improvements typically come during the relatively short period when a technology goes from early commercialisation to full-scale mass adoption. For example, we may see a period of rapid change in the next 20 years as trucking fleets are upgraded with driverless technologies. This will likely boost productivity growth, but once it is complete productivity growth rates will slow again. This may be the pattern for the future, as it was in much of the past. We may see periods of slow productivity growth punctuated by bursts of faster growth during phases of mass-adoption of new technologies. The emergence of driverless vehicles, the replacement of routine clerical jobs with machine-learning algorithms, and the development of new genetic engineering techniques might all significantly improve productivity at some point in the future, perhaps substantially. But after the one-off improvement in the productivity level, growth slows again. Gordon makes the point that the effects of most technologies are quite local in their economic impact. Electrification was an exception to this, and computerisation to some extent. However, driverless vehicles will primarily affect the trucking, logistics and taxi driving sectors, which are worth less than 5% of the economy. While output per person may rise dramatically in these sectors, the effect on average national productivity is likely to be small.
“Historically, there has been much faster productivity growth in manufacturing than in service sectors.” This highlights a wider problem. Historically, there has been much faster productivity growth in manufacturing than in service sectors. The latter often depends on human interaction (e.g. haircuts, restaurant service, elderly care). Although driverless taxis are an
exception, in general it is hard to make productivity improvements in these industries. A one-hour massage takes an hour’s work, by definition. As manufacturing shrinks as a proportion of the total economy, it is consequently harder to achieve rapid productivity growth. The service sector accounts for 80% of GDP in the US today, compared with less than 50% in 1950. Manufacturing has steadily contributed less, declining from nearly 40% to 12% over this period. As a result, productivity-enhancing innovations must take a different form than in the past, and the scope for improvement might be permanently reduced. Mismeasurement Some analysts have suggested the slowdown in productivity growth since the early 2000s is a mirage resulting from mismeasurement. One argument is that the ongoing improvement in the quality of technological products is being underestimated, which results in inflation being overstated. If inflation is overstated, true real output (output growth after subtracting inflation) and productivity (real output per hour worked) would be higher than the reported statistics suggest. For example, domestic broadband speeds are 10 times faster than a decade ago, but the cost is more or less the same. You are getting far more bandwidth for your money, so on a quality-adjusted basis inflation in broadband prices has been highly negative. An influential Brookings Institute paper1 finds that inflation in the digital economy is indeed being understated. The level of output and productivity should be higher. However, the paper notes that this effect was even greater with technological innovations in the past. Arguably, relative to the past, current productivity growth rates are even worse than we thought. It is also sometimes argued that GDP-based output measures fail to capture the large benefits of modern technologies. Today we spend far more time browsing the internet in our homes and on our smartphones than we used to. Most of us consider this to be a huge benefit, but because much of the content we consume is offered for zero cost (e.g. you don’t pay to search Google), it is not recognised by GDP data. GDP may therefore be dramatically underreporting the welfare benefits from the new technologies. This seems plausible, but it is hardly new. Many of the vast benefits of electric lighting, air-conditioning, central heating and plumbing did not show up in the GDP numbers and were therefore similarly undercounted. Perhaps the biggest problem with this argument, at least for our purposes, is that investors do not get returns from gains in welfare. Capital invested must generate financial profits, so investors get returns from increases in market output. Internet services, even if valuable to their users, do not generate returns for investors unless they are effectively monetised, or allow businesses to be more productive in other ways. The commercial value of our internet browsing is modest and concentrated in the revenues of a very small number of companies, notably Google and Facebook. http://www.frbsf.org/economic-research/publications/economic-letter/2017/february/ does-growing-mismeasurement-explain-disappointing-productivity/
1
aberdeen-asset.co.uk
25
Long-term growth continued Post-crisis slowdown In the last five or so years productivity growth in the US and elsewhere has been particularly dismal. Although the factors above may explain part of this, failing to achieve anything like the 1.5% annual rate that predominated from the 1970s to the early 2000s suggests further explanation is required. There is a lot of disagreement about the underlying causes. Declining skill levels One contributor to the recent poor performance is the accelerated loss of skilled workers from the labour force, as baby boomers retire and are replaced by younger workers that are less skilled and productive. More of these workers are therefore required to fulfil the same duties. This is consistent with the OECD finding that, for the first time in the 240-year history of the US, the next generation will be less educated than the preceding generation. This trend is likely to endure for at least the next decade, continuing to depress productivity. Low investment levels Another culprit is the relatively low level of business investment in recent years. As shown in the chart below, investment has fallen well below the long-run average in the US, and especially so when depreciation of the capital stock is taken into account. Fig. 2.7: US gross and net investment (%)
19 1947-1990 Averages
9 4
0 -1 1950 1956 1962 1968 1974 1980 1986 1992 1998 2004 2010 2016 US Recession Indicator (NBER) Net Domestic Investment
Fig. 2.8: Composition of US businessgrowth productivity growth (%) Composition of US business productivity (%) 2.5
2.3
2.0 1.5 0.9
1.0 0.5 0.0 -0.5
1948-2007 Total factor productivity Labour composition
2010-2015 Capital intensity Total productivity growth
Source: Aberdeen, US Bureau of Labor Statistics, March 2017. Note: Private Non-Farm Business Sector (Excluding Government Enterprises). In percentage points, average annual rate. Chart shows the average productivity growth between WW2 and the financial crisis compared with the average since 2010. In both cases, the subcomponents are shown.
This is more of a problem in some countries than others. Business investment in the UK, for example, has been truly awful – and the country’s productivity performance has followed suit.
24
14
Workers produce more per hour when equipped with the latest tools and technologies. Some studies suggest that as much as two thirds of the recent productivity slowdown is as a result of slower growth in business investment, as demonstrated by the reduction in capital intensity growth in the chart (Fig. 2.8) below2.
Gross Domestic Invesment
Source: Federal Reserve Bank of St. Louis (FRED), Thomson Reuters Datastream, February 2017. Note: Gross/Net Private Domestic Investment, as a percentage of GDP, Seasonally Adjusted.
“Some studies suggest that as much as two thirds of the recent productivity slowdown is as a result of slower growth in business investment.”
Economists have various theories about the reasons for the slowdown in business investment. One argument is that there is less competition in the business community and few start-ups, reducing economic dynamism. A lack of competition tends to result in less efficient use of resources, and lower productivity. Another suggestion is that, with unemployment high and wages low for much of the period since the financial crisis, companies have substituted cheap labour for capital and held back on making investments. It is also possible that lower business investment is simply a function of lower growth expectations. Standard ‘accelerator’ theories of business investment suggest that investment levels are a function of growth expectations. The lower the rate of growth expected, the less management will be keen to invest. Sluggish growth and lower global growth expectations thereby create a vicious circle. If this is part of the cause of unusually low productivity, policies that encourage capital expenditures could enable a recovery in productivity growth. The current Republican tax plan proposes various changes to encourage business investment. If implemented, it is possible we may see some progress in the US. However, given the demographic trends discussed long-term growth expectations are likely to remain depressed, so incentives for business investment will most likely remain low despite policy changes.
Furman, J (2015) Productivity growth in advanced economies. Peterson Institute. https:// obamawhitehouse.archives.gov/sites/default/files/docs/20150709_productivity_ advanced_economies_piie.pdf
2
26
Long-term growth
“In leading companies, productivity is growing as fast as ever. The problem is that a large disparity has emerged between the leading companies and the rest.” Weak diffusion of productivity through the economy The issue of poor productivity growth is explored in a recent IMF report3, which highlights the large gap between the most productive companies and the least. In leading companies, productivity is growing as fast as ever. The problem is that a large disparity has emerged between the leading companies and the rest. There are various theories about why this has happened. One is an extension of the business investment argument described previously. Laggard companies have failed to invest sufficiently in R&D. There is also a more damning argument that companies have become too short-termist, with executives focusing too much on hitting the near-term targets that trigger their share option packages, rather than working for longer-term success.
The biggest near-term cause for optimism is the possibility of higher levels of business investment, particularly in the US, as wages start to rise and companies invest to improve productivity in response, perhaps aided by Trump-administration tax cuts. Addressing low productivity is one of the most important challenges for economic policy makers. Faster productivity growth is our best hope for offsetting slow or negative labour-force growth. If we are unable to improve on today’s dismal productivity growth, western economies will see very low levels of GDP growth, with negative consequences for incomes and investment returns. The somewhat eclectic and tentative discussion above demonstrates that the causes of the productivity slowdown are still not fully understood. There is an urgent need to remedy this lack of knowledge, and to develop and implement a more robust set of policy solutions. In the meantime, we forecast only modest productivity growth in developed countries, somewhat faster than the current dismal levels, but rather lower than the faster growth before the crisis.
There may also be wider problems with the uptake of new business practices if laggard companies do not have the right skills to implement them. There is clear evidence from the UK that productivity growth demonstrates strong regional patterns. Productivity is 60% higher in companies based in London than those in Northern Ireland, and there is a large divergence between skills levels in these two regions.4
Emerging economy catch up The discussion of productivity above has been about the rate of improvement in productivity in advanced economies which are at or near the technological frontier. They are already implementing the most modern technologies widely across their economies. In contrast, emerging economies can make large gains in productivity as they catch up by developing better infrastructure, modern industries, better education and move people from low productivity rural agricultural jobs to industrial jobs in cities.
Another possible reason for slower post-crisis productivity is the lack of the ‘creative destruction’ in which capitalism thrives. Since the financial crisis, many companies have been unable to repay their debts. But instead of letting them fail, banks, particularly in Europe have preferred to keep them on life support rather than writing off the debts. The trouble is that ‘zombie’ companies like this can’t invest in developing their business because banks will not make new loans to them, but their ongoing existence limits the ability of their more dynamic competitors to thrive.
In the last 50 years Japan, South Korea, and most significantly China have grown exceptionally quickly though dramatic and sustained increases in the productivity of their economies. However, catch up is not inevitable. Not all emerging economies have followed South Korea and Japan in reaching advanced economy productivity levels. A large number have become mired in what is known as the ‘middle income trap’, where productivity growth slows and catch-up stalls. In line with the disappointing productivity story elsewhere, productivity growth in emerging economies has slowed down too in the last decade.5
Conclusion Technological progress is alive and well and we can expect periods of more rapid productivity growth in the future. Artificial intelligence looks set to offer large – if rather disruptive – gains in productivity across a wide range of business sectors. However, it seems unlikely that these benefits will appear at scale in the next few years. For example, it may take a decade or two of experimentation and regulatory change before we see the mass adoption of driverless commercial vehicles.
Having said that, productivity growth remains higher in emerging markets than developed markets. Some countries are still showing strong growth with positive implications for GDP.
In addition, the current barriers to productivity growth described above – lack of diffusion, low levels of capex, the loss of experienced older workers – will not be quickly resolved. So it does seem reasonably safe to conclude that even if productivity growth does return to something approaching the rapid pace of the 20th century, this is not imminent. Adler et al. (2017) Gone with the headwinds: global productivity headwinds. IMF. https:// www.imf.org/en/Publications/Staff-Discussion-Notes/Issues/2017/04/03/Gone-withthe-Headwinds-Global-Productivity-44758 4 CBI (2016) Unlocking regional growth. CBI. http://www.cbi.org.uk/index.cfm/_api/render/ file/?method=inline&fileID=9AF06398-223D-4214-B96F1AD8A2FE4CC8 3
Haldane, A (2017) Productivity puzzles. Bank of England. http://www.bankofengland. co.uk/publications/Pages/speeches/2017/968.aspx
5
aberdeen-asset.co.uk
27
Long-term growth continued Secular stagnation and other factors In typical business cycles recessions are rapidly followed by strong recoveries, during which growth rises above the economy’s long-term potential growth rate, making up for some or all of the ground lost during the recession. Total output goes back to the full productive capacity of the economy.
In the past it has largely made sense to ignore the business cycle in long-term economic forecasts. Over a 10 year time horizon, the slower growth in the downturn is offset by the faster growth in the recovery, and the economy will on average grow in line with improvements in the overall productive capacity. But this assumption does not seem to be playing out during the postfinancial crisis period and many economists are concerned the current period of stagnation might be a persistent problem.
“The post-financial crisis business cycle is very unusual in that growth did not quickly recover “A world in which aggregate demand does not to take output back to the previous trend.” rebound quickly from recessions is one where As the charts (Fig. 2.9 and Fig. 2.10) below show, the post-financial the global economy spends more time below crisis business cycle is very unusual in that growth did not quickly potential than above.” recover to take output back to the previous trend. There was no making up for lost ground. Households, companies and even governments have continued trying to save too much and spend too little – ‘aggregate demand’ has been weak. As a consequence it has taken a long time to reduce unemployment to pre-crisis levels. In Europe, this still has not happened eight years on, though progress continues to be made. Fig. 2.9: Eurozone potential GDP pre-financial crisis versus outcome
A world in which aggregate demand does not rebound quickly from recessions is one where the global economy spends more time below potential than above – with lower than potential growth on average. This is particularly unappealing given potential growth itself is expected be fairly low for the reasons discussed in the previous sections. This is the scenario US economist Larry Summers describes in his revival of the idea of ‘secular stagnation’.
13,500
The extent to which this concern is justified depends a lot on understanding why the recovery has been so weak since the last recession.
USD PPP (bn) 14,500
12,500
Balance sheet recessions It could be argued that the recovery has been so muted because we have been recovering from a ‘balance sheet’ recession. This describes a world where, prompted by an economic downturn, companies and households reduce their debt levels and strengthen their balance sheets. While doing so, they must save more and consume less, weakening demand. Similarly, banks must cleanse their balance sheets of bad loans. During this process, they are less willing to extend new credit. The process of balance sheet repair takes time, hence the extended period of slow demand growth.6
11,500 10,500 9,500 8,500 7,500 6,500 1980
1984
1988
1992
Eurozone real GDP
1996
2000
2004
2008
2012
Trend (1980-2007)
Fig. 2.10: US potential GDP pre-financial crisis versus outcome USD PPP (bn) 20,500 18,500 16,500 14,500 12,500 10,500 8,500 6,500 1980 1984 1988 US real GDP
1992
1996 2000 2004 2008 2012
2016
The positive aspect of this line of argument is that once balance sheets are repaired the economy will eventually get back to normal. If true, we should still expect faster growth to make up for at least part of the ground lost since the financial crisis. However, global debt levels are now higher than they were before the financial crisis. This is mainly because emerging economies, particularly China, have considerably added to their debt levels. But even in regions like the US, where there was initially meaningful deleveraging, debt levels have started to rise again as firms and households start to borrow more. This suggests that the period of widespread deleveraging has finished, and we must conclude that this has not been associated with a significant acceleration in growth, at least so far. Balance sheet strength is not the only factor at work.
Trend (1980-2007)
Source (both charts): Aberdeen, Oxford Economics, March 2017. Note: Trendlines consists of polynomial regression of 1980-2007. Real GDP is in USD PPP billion (in 2010) exchange rate.
6
28
Long-term growth
Koo, R (2014) Balance sheet recession is the reason for ‘secular stagnation’. VoxEU. http://voxeu.org/article/balance-sheet-recession-reason-secular-stagnation
The zero lower bound A very different kind of problem is that interest rates are currently near zero, as Larry Summers points out in his discussions of secular stagnation.
“The issue with very low interest rates is that they make it much harder for central banks to stimulate the economy in a recession.” The issue with very low interest rates is that they make it much harder for central banks to stimulate the economy in a recession. One of the main reasons why there is typically a strong rebound following a recession is that central banks reduce interest rates by several percentage points, providing a strong monetary stimulus that encourages faster growth. If nominal interest rates are near zero, central banks cannot reduce interest rates by much. Nominal interest rates cannot go much below zero or investors will hold cash instead. Central bankers have developed other tools, such as forward guidance on interest rates and quantitative easing, and have experimented with negative interest rates. But there are concerns that these policies are less effective – as recent history perhaps demonstrates – and may have unwelcome financial stability side effects. As we discuss in the next chapter, today’s low interest rates are caused by demographic and other structural factors. As a result, they are likely to be with us for an extended period of time, and central banks are likely to face further episodes where they cannot cut interest rates by enough to bring output up to potential. As such, Summers fears an extended period of secular stagnation where recessions are followed by weak and protracted recoveries. Hysteresis A further related problem is ‘hysteresis’. In the absence of sufficient policy support to bring demand up to potential output, the economy faces a persistent lack of demand. “Hysteresis” is a process by which a persistent shortfall in demand has a negative effect on the productive capacity of the economy. The idea makes sense, as people who remain unemployed for a long time may drop out of the labour force and fall behind on skills, while companies invest less in new projects and equipment, reducing the productive capacity of the economy. Hysteresis means that failure to stimulate demand sufficiently – for example, because of the zero lower bound – means that slow recoveries do not merely mean lower growth in the short-term, but they also reduce growth potential in the long-run.
“Recent experience suggests that the weak economic performance of developed economies is politically destabilising.”
Weak economies lead to disruptive politics Recent experience suggests that the weak economic performance of developed economies is politically destabilising. Slow economic growth tends to result in low growth in wages. In addition, growth may not be spread evenly. Some regions do better than others, and the rich may fare better than the middle classes. People who feel ‘left behind’ by the economic policies of conventional political parties, unsurprisingly, vote for populist, anti-establishment politicians who offer radical solutions. The economic policies prescribed by populists will not necessarily increase incomes. Policies such as drastically reducing immigration or walking away from free trade agreements may end up reducing growth in incomes even further.
“The structural problems with productivity and structurally low interest rates described previously are unlikely to be solved by business-as-usual policies.” This is not to dismiss the need for radical solutions. The structural problems with productivity and structurally low interest rates described previously are unlikely to be solved by business-as-usual policies. Ambitious new policies are likely to be needed. For example, many economic policy makers have suggested the world embarks on large-scale infrastructure investment to boost demand, capital investment and productivity. Given that sensible infrastructure investment pays for itself, this should be a high priority. Conclusions Demographic trends are not destiny, but they are hard to change. We can be confident the demographic transition we have described means that trend growth is likely to be slower than in the past in most countries. This is embedded in our long-term growth forecasts. It is much harder to be confident about productivity forecasts. Our base case is that productivity growth will improve from its current levels, though not to the high rates of growth seen before the financial crisis. We think these reflected an unusually fast period of technological change, not a sustainable long-term trend. There is a downside risk that we do not manage to achieve these faster growth rates, and we are left with something like the trend of the last few years. This would be a very disappointing outcome, and would result in growth projections even lower than those we forecast. There is also considerable uncertainty about the secular stagnation question. We do think equilibrium interest rates are likely to remain unusually low. If this is right, then the zero lower bound is likely to be a persistent barrier to effective central bank economic stimulus. This may result in persistently below-trend growth. For the time being we see this as a downside risk rather than our base case.
aberdeen-asset.co.uk
29
Chapter
03 30 document_name
Trends in inflation, interest rates and currencies
• Today’s ultra-low interest rates are driven by structural trends that will not reverse in the near future • Central banks have failed to hit their inflation target post-financial crisis. We expect more success in the next few years • The dollar is expensive on historical measures, but is unlikely to revert to equilibrium while interest rate differentials persist.
aberdeen-asset.co.uk31
Trends in inflation, interest rates and currencies Long-term inflation If growth is the most important factor determining the valuation of equities, inflation is arguably the most important factor for government bonds. High inflation reduces the real value of the fixed cash flows available from (nominal) government bonds. If inflation is expected, the price of bonds falls and the yield rises.
Trends in interest rates
After the period of high inflation of the 1970s, the inflation rate has fallen dramatically and is much more predictable. This is a direct result of the decision by most central banks to make price stability a central goal of monetary policy. (By stable prices central bankers typically mean an inflation rate of around 2%.) Central banks have also been granted independence from government control. This has made it harder for governments to stimulate inflation for political purposes (e.g. reducing interest rates to boost the economy in the run up to a general election).
Perhaps the single most important issue for strategic asset allocation today is the extremely low level of interest rates. Low government bond yields mean low returns for investors relying on government and other investment-grade bonds for their returns. Bond yields also affect the discount rate that investors in equities and other risk assets use to value their holdings. Low yields mean, for example, that US equities are less expensive than they seem. The biggest question for investors today is whether today’s ultra-low yields will persist over the long term.
In developed economies, the strong track record of central banks means that, as a starting point, it is reasonable to assume that inflation in the long term will be around the central bank target of 2%.
The extraordinary decline in bond yields There is no question that global bond yields are very low. This is not just a post-financial crisis phenomenon – it is part of a 30-year trend. Average bond yields have fallen 4.5 percentage points since the 1980s1. It is also a global phenomenon, occurring across developed and developing markets. Japan used to be an outlier, but the rest of the world has caught up. This is a truly exceptional state of affairs. According to Bank of England research, nominal interest rates may be lower now than they have ever been in 5,000 years2.
There are a few comments to make though. First, despite their heroic efforts central bankers have largely failed to achieve their inflation targets for most of the period following the financial crisis. In part, this was because of the very strong deflationary pressures exerted by persistently weak demand as economies struggled to grow. Also, when interest rates are at or near 0% it is hard for central banks to stimulate economies using conventional tools. We discussed this ‘zero lower bound’ problem in the previous chapter: central banks cannot push nominal interest rates much below zero which limits their ability to stimulate the economy in a slump. While inflation is now at target levels in the US and UK, it remains below target in Europe and Japan. We may yet find ourselves returning to an environment where inflation is persistently below target. If so, an average inflation assumption of 2% may be optimistic. On the other hand, several influential economists have argued that central banks should target a higher inflation rate to avoid exactly these problems. If central bankers were to achieve an average inflation rate of, say, 4% we would be less likely to find ourselves stuck at the zero lower bound. (There are several variants of this argument including inflation level targeting and nominal GDP targeting – but they all result in higher inflation.) One advantage of a higher inflation rate is that central banks would be able to reduce “real” interest rates further. A 0% nominal interest rate and an inflation rate of 2% translates into a -2% real interest rate. But a zero nominal rate with inflation at 4% translates into a -4% real rate. A shift in central bank inflation-targeting policy would result in higher inflation forecasts. Lastly, there is a scenario where long-term inflation might move higher. In many countries, government debt levels are very high. Weak economic growth and ageing populations mean that the fiscal position may even worsen significantly over time. If so, politicians might, at some point, find it attractive to reduce the real value of this debt by allowing rather higher inflation rates. The coincidence of interests of politicians wanting to inflate away the debt and central bankers trying to escape the zero-lower-bound problem, means a shift in the inflation regime is a small, but material, risk for long-term investors to consider. 32
“The single most important issue for strategic asset allocation today is the extremely low level of interest rates.”
Trends in inflation, interest rates and currencies
Fig. 1.2: 10-year nominal yields (%) 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 -1
1990 US
1994 Japan
1998 France
2002 Germany
2006 Italy
2010
2014
UK
Source: Oxford Economics, February 2017. Note: Chart shows yield on 10 year maturity government bond in each country.
A slow rebound from the financial crisis or something deeper? For several years following the financial crisis, it seemed reasonable to assume that low interest rates were just a short-term hangover from an unusually deep ‘balance sheet recession’; this would soon dissipate and interest rates would bounce back to levels closer to historical averages. This is consistent with the conventional view which suggests that nominal interest rates should approximate to nominal GDP over the long term. So if you expect inflation to be 2% and real GDP to be 2.5%, nominal interest rates should be 4.5% – and so a return to higher levels of interest rates should be expected. Rachel and Smith (2015). Haldane (2105) Stuck. Bank of England. http://www.bankofengland.co.uk/publications/ Documents/speeches/2015/speech828.pdf.
1 2
Hamilton J et al. (2015). The Equilibrium Read Funds Rate: Past, Present and Future. US Monetary Policy Forum. http://econweb.ucsd.edu/~jhamilto/USMPF_2015.pdf
3
But is this just the central banks’ fault for persisting with low interest rate policies and quantitative easing? We do not think so. It is true that central bank policy has depressed interest rates – but this confuses cause and effect. Central banks would be delighted to be in a position to raise rates to more normal levels, but they can only tighten monetary policy when justified by rising growth and inflation. If they tighten too early they are likely to push the economy into recession and deflation – and this, in turn, would require even looser monetary policy in the future. There is a growing acknowledgement that there are deeper forces at work. The bond market seems to have accepted this view. Over the last few years it has progressively lowered its expectation (expressed in bond market forward yields) of future long-term interest rates. The projected yield on the 10-year US Treasury bond five years forward has fallen from 5.4% in January 2010 to 3.2% in April 2017. The Federal Reserve has also, it seems, accepted this new reality. Its ‘dot plot’ view of the long-term level for interest rates has fallen in a similar way. What is driving this change in perspective? Central bankers and bond markets believe that interest rates are not just influenced by expectations about growth and inflation, but also by a more mysterious variable that they call the equilibrium real rate of interest (also known as the “natural” or “neutral” rate, or R* for short)4. Equilibrium interest rates Economists imagine a global market for savings and investment where would-be suppliers of savings meet businesses and governments who require funds for investment. The resultant market price is the rate of interest capital investors need to pay to reward savers for access to their capital. In other words, the equilibrium real interest rate is the rate necessary to balance desired savings with desired investment when the economy is operating at capacity and inflation is stable5.
“The more demand there is for saving, and the less demand for capital investment, the lower the equilibrium interest rate.” The more demand there is for saving, and the less demand for capital investment, the lower the equilibrium interest rate. The following chart (Fig. 3.2) provides a stylised illustration of how various drivers of increased global saving, and a falling demand for capital investment have conspired to reduce global rates.
Federal Reserve chair Janet Yellen recently referred to equilibrium interest rates no fewer than 25 times in a single speech. http://blogs.ft.com/gavyndavies/2015/04/02/yellenshoots-for-equilibrium-interest-rates/. 5 The concept was first used by Swedish economist Knut Wicksell in the 19th century, but is also a key part of the modern New Keynesian models used by many central banks today. The canonical model is Woodford, M (2003) Interest and Prices: Foundations of a Theory of Monetary Policy. Princeton University Press.
Fig. 3.2: Global long-term neutral rate: savings and investments shift (%) 1
6
2
3
5 World real interest rate
Recent research suggests that in fact there is little evidence for this conventional view. Empirically, nominal rates do not converge with nominal GDP3. And the fact is that eight years after the financial crisis, policy interest rates are still near zero in many developed economies even though nominal GDP is over 3%.
4 Investment schedule
3 2 1 0
6
Saving schedule
5 10
15
20
25
30
35
4 40
Global saving and investment as a share of global GDP Source: Bank of England, Aberdeen, February 2017. Note: The chart shows the level of % of GDP saved and the % invested on the x axis and the interest rate on the y axis. It plots savings/investment level and interest rates over the last 20 years. While the intersection between savings/investment levels and interest rates can be measured, the desired savings and desired investment curves cannot be observed and can only be hypothesised. Illustrative shifts: (1) Demographics, (2) Rising inequality, (3) Global savings glut, (4) Relative price of capital, (5) Public investment, (6) Spreads.
This way of theorising about interest rates gained prominence before the financial crisis. Ben Bernanke, who later became chair of the Federal Reserve, used it to explain the persistence of low long-term interest rates, even while the Fed was raising policy rates. He suggested that there was a ‘savings glut’ – that is to say, an excess of global desired saving over global desired investment6. More recently Larry Summers, former US Treasury Secretary, argued that we may be facing an extended period of ‘secular stagnation’ because the equilibrium real interest rate may be stuck below zero, as discussed in the previous chapter7. Global savings factors Demographic change People’s savings and investment behaviour changes over their life cycle. The working-age population saves more than either the very young and the very old. Over the last 30 years the working-age population has grown relative to the non-working age population. As fertility rates have fallen, the younger cohort has fallen in size relative to the working-age cohort, the old age cohort has started to grow but not enough to compensate. Another factor is the average age of the working population. People save more in the second half of their working lives than in the first. The average age of the working-age population has risen over the last 30 years, again increasing supply of savings. Retirement ages are also rising, again increasing the size of the working-age population. This demographic shift means that global desire to save has been rising.
4
http://www.federalreserve.gov/boarddocs/speeches/2005/200503102/ In a speech to the National Association of Business Economics in February 2014, http://larrysummers.com/wp-content/uploads/2014/06/NABE-speech-Lawrence-H.Summers1.pdf
6 7
aberdeen-asset.co.uk
33
Trends in inflation, interest rates and currencies continued
Emerging markets’ savings After the painful experience of the Asian crisis in the 1990s, many emerging economies have chosen to build large foreign currency reserves to prevent a repeat of the currency crises that occurred then. This build-up of reserves is a form of saving. At the same time, a steady decade-long rise in oil prices meant that oil exporting nations experienced substantial budget surpluses, adding to global savings – though this pressure has waned recently.
“The cost of the capital goods required for investment projects has fallen dramatically in recent decades.” Investment factors Lower cost of investment The cost of the capital goods required for investment projects has fallen dramatically in recent decades – by 30% since the 1980s9. This is as a result of cheaper information and communications technology. Empirical observation suggests a sectoral shift in economies away from capital-intensive sectors to sectors that need less capital. For example, many of world’s most valuable technology companies have become very large with very little capital investment. Lower public investment Public investment in infrastructure has been on a long-term downward trend in developed economies for decades. This trend has not been helped by the sharp rise in government debt-to-GDP ratios during the financial crisis, and the subsequent austerity policies introduced to bring them back down. In this environment it has proved hard to make a successful case for increasing borrowing for infrastructure investment, even though the interest rate on government borrowing is at all-time low. Higher risk premia Companies, who provide much of the demand for investment capital, do not in practice make investment decisions based on the risk-free interest rate. They need to consider their overall cost of capital: the risk-free rate combined with the risk premium they must pay for equity or credit. Risk premia vary significantly over time and can offset or amplify effects from changes to the risk-free rate. International Monetary Fund modelling suggests that risk premia have risen since the 1990s – perhaps in part due to the lower and more uncertain growth expectations discussed in the previous chapter. Higher risk premia have a downward impact on desired investment for any given risk-free rate.
Rachel and Smith (2015). Rachel and Smith (2015).
8
9
34
Trends in inflation, interest rates and currencies
Lower growth expectations A final reason for low equilibrium interest is lower growth expectations. This only explains the downward trend in the period since the crisis. Before this point growth expectations were stable. After the crisis, it has become clear that demographic change and lower productivity growth mean that potential GDP is somewhat lower than it was before the crisis. This reduces the amount of investment required by the economy to meet growing demand. Fig. 3.3: Explaining the decline in global real interest rates 0 -50 -100 Basis points
Rising inequality Another important factor is rising inequality. The top 1% of the income distribution (the richest) save over 40% of their income compared to an average of less than 10% for the rest8. The greater proportion of national income that goes to the rich, the more saving there will be. In much of the world the share of income going to the richest has increased markedly since the 1980s.
-150 -200 -250 -300 -350 -400
1980-2015 Lower public Investment Low cost of capital investment EM government saving Rising inequality Higher risk premia Demographics Growth
Source: Aberdeen, Bank of England, February 2017.
“Are low interest rates now a permanent state of affairs?” The future Are low interest rates now a permanent state of affairs? That looks unlikely. Eventually, many of the demographic and structural changes described previously will reverse. However, this does not look likely to happen any time soon. We may need to wait till the late 2020s before equilibrium nominal rates are back above 4%10. But the timing of this is highly uncertain. It all depends on the shifting forces that determine the balance between demand for saving and the future of demand for investment. Demographic trends Given all the talk about the ageing population, it is tempting to think that the dependency ratio is likely to reverse soon – with large numbers of retired people creating an ever greater burden on the working population – resulting in much less disposable income being available for saving. This is expected to happen eventually, but the process is likely to be protracted. First, in an effort to offset the fiscal implications of ageing populations, governments are raising the retirement age. This temporarily increases the size of the working-age population. Second, the duration of retirement increases as people live longer. They will need to save more, and start saving earlier, to ensure sufficient income in retirement. (Low interest rates make this harder.) This also is likely to delay the time at which savings rates start to fall again.
10
Equilibrium nominal rates are, roughly speaking, equilibrium real rates plus expected inflation.
Income inequality Income inequality is driven by many factors. If technology means many middle income jobs are automated there could be an even stronger bifurcation in the labour market with more income inequality. Similarly, if the economy creates more winner-takes-all business models with high barriers to entry, then we may see the spoils concentrated in the hands of the lucky few. But inequality is also a public policy choice. It results from taxation policy, public education policy and levels of investment in urban development. The kind of frustration that motivates some supporters of Brexit and of Donald Trump could result in a realignment of policy priorities and a reversal of current trends. This is very uncertain. On balance, it seems unlikely that we will see a great deal of progress in this area in the near future. Global capital balances A major source of Bernanke’s ‘savings glut’ was the growth in EM foreign exchange reserves. This seems to have peaked and has fallen back somewhat, particularly in China. The decline may continue a little, though it is likely that most countries will wish to maintain much larger reserves than they held in the 1990s.
“China, Germany, Japan and a few other countries continue to have large surpluses of savings over domestic investment.” Aside from foreign exchange reserves, it remains the case that China, Germany, Japan and a few other countries continue to have large surpluses of savings over domestic investment, which are being exported to the rest of the world in the form of current account surpluses. This source of excess savings has been identified as a key problem by the Trump administration, so there may be pressure for action. Though the causes of these savings surpluses are deep-seated and not easily resolved. As China rebalances its economy by reducing its domestic investment rate, its surplus could even rise substantially if its savings rate does not fall in parallel.
However, with the possible exception of some as yet uncertain commitments by Donald Trump, this is more talk than action. It is also important to note that the great surge in public investment spending seen in China in the last decade is coming to an end as the economy begins to rebalance. It seems most likely that public investment will not increase. Higher risk premia There is some evidence that risk premia are related to macroeconomic uncertainty. Low risk premia in the late 1990s corresponds to rising productivity growth, economic stability and optimism about the future (‘the great moderation’). This contrasts with higher observed risk premia in the 1970s and post-financial crisis period, where growth uncertainty has been elevated. If the global economy continues to recover, we might see some mean reversion in risk premia (with a positive impact on equilibrium levels). But if, say, a hard landing in China were to push the world into another recession, we might see a continuation of the downward impact on rates. Low growth It seems likely that the lower growth rate since the financial crisis will be persistent. It is driven by demographic change and lower productivity growth. While productivity may well improve, it looks unlikely – at least in the near term – to reach the high levels of the 1990s and so will not compensate for the slower growth arising from the shrinking working-age populations in many countries. Accordingly, little improvement is expected. On the other hand, there is no reason to think growth expectations will weaken significantly from here. Conclusion With the exception of the modest rebound from demographic trends of +30bps or so, most of the future projections made by the Bank of England appear to be for little or no change: equilibrium rates will be only slightly higher (perhaps 1%) over the next 15 years. Fig. 3.4: The past versus the future – real interest rates remain close to today’s levels 100 0
Basis points
However, there is one important counterargument. Old people tend to consume much more than young people – particularly when you consider health and care services in old age. A dependency ratio with an equal proportion of old and young dependents is not the same as one with more old people. There will be less net saving in the latter group. It is possible that this could mean that savings rates start to decline rather sooner.
-200 -300 -400
Lower cost of investment The technological factors that have driven down the cost of investment are not diminishing. Further downward pressure on the cost of investment is likely. Public investment There is a growing concern that monetary policy is unlikely to be able to continue on its current course and that a turn to fiscal policy is likely, with infrastructure spending being a favoured focus.
-100
-500
1980-2015
2015-2020
2020-2030
Unexplained Lower public investment Low cost of capital investment EM government saving Rising inequality Higher risk premia Demographics Growth Change in global real rates Source: Aberdeen, Bank of England, February 2017. Note: Chart decomposes changes in equilibrium real rates in the past and two future periods. While rates do not continue to fall in the future, they only grow very slowly (circa. 20bps per decade from current levels.
Borio, C (2017) Secular stagnation of financial cycle drag. Bank of International Settlements, www.bis.org/speeches/sp170307.pdf.
11
aberdeen-asset.co.uk
35
Trends in inflation, interest rates and currencies continued We emphasise that analysis of the various factors that combine to drive equilibrium interest rates is highly uncertain and the estimates for the precise value of this rate are uncertain. So it is important to accept that there is a range of possible values. Perhaps +1% to -1% is a reasonable assumption at the moment. It is also worth noting that this is an equilibrium concept; in the short term, interest rates can go higher or lower than their equilibrium value as the business cycle progresses.
“Too much easing during recessions and not enough tightening in recoveries, has resulted in a massive rise in global debt levels.” We also note that some economists have alternative explanations for the secular decline in interest rates. The Bank of International Settlements – the central bankers’ central bank – argues that today’s low rates have more to do with the asymmetric response of central banks to previous business cycles11. Too much easing during recessions and not enough tightening in recoveries, has resulted in a massive rise in global debt levels over the last few decades, and large scale allocation of capital to unproductive activities (e.g. to property speculation rather than productive investment). This argument is intriguing – and implies a different model of interest rate determination than the loanable funds model described previously. However, it is controversial, not least because it raises fundamental questions about the current monetary policy orthodoxy. For our current purposes, this debt-based view does not much change our conclusion. Overcoming the financial drag that the Bank of International Settlements believes is holding down interest rates will take a long time. In the meantime equilibrium interest rates will remain low.
“Interest rates will tend to remain at a low level for an extended period of time.” Despite concerns about inaccuracy and incompleteness, we think analysis of equilibrium real rates is useful. One of the most striking features of economic life in the last 20 years has been the decline in interest rates. This approach explains most of this decline in an intuitively plausible way by appealing to empirically observable data. It gives us fairly strong reasons for thinking that on average interest rates will tend to remain at a low level for an extended period of time, perhaps more than 10 years. However, lower-for-longer does not mean lower-for-ever. Eventually interest rates may rise as spending – for example on healthcare – by larger aging populations outweighs the saving of the smaller working age population.
Currency returns Investors today are more than ever inclined to invest globally. An important factor affecting the returns they get is the behaviour of the exchange rate between their home currency
Borio, C (2017) Secular stagnation of financial cycle drag. Bank of International Settlements, www.bis.org/speeches/sp170307.pdf.
11
36
Trends in inflation, interest rates and currencies
and the currencies in which their foreign assets are priced. Currency movements can be large, often larger than the local currency return of the assets concerned. It is therefore crucial to consider currency when investing. In our forecasts, we express expected returns in both the local currencies of the assets we forecast and in the various ‘base currencies’ of our clients. For UK investors with sterling as the base currency, we incorporate our expectations for changes in the value of sterling versus the other currencies in which our investments are denominated. For example, for US equities we offer a ‘local’ dollar expected return and a sterling expected return. If we expect the US dollar to appreciate with respect to sterling, then the local return will be higher than the return expressed sterling terms. To illustrate this, the table below (Fig. 3.5) shows our 10 year expected returns for regional equities expressed in local, sterling and sterling-hedged terms. (We explain returns from currency hedging at the end of this section.) Fig. 3.5: Global equity returns - currency impact Local currency GBP GBP Hedged 10Y 10Y 10Y UK Equities
6.1
6.1
6.1
US Equities
5.3
3.5
4.1
Europe ex UK Equities
5.2
5.0
6.0
Japan Equities
4.6
5.1
5.9
Pacific ex Japan Equities
6.6
5.0
5.1
Emerging Markets Equities
6.3
5.0
N/A
Source: Aberdeen, March 2017. Note: Returns are in percentage, per annum. Projections are estimates and provide no guarantee of future results.
Currency movements are hard to predict, particularly in the short term. But in the longer term currency ‘fair-value’ models can provide a useful guide. Roughly 90% of the time, exchange rates move within 20% of their fair value level. This means when currencies are outside this range – either extremely cheap or extremely expensive – you can be reasonably confident that they will eventually revert to fair value. There are various ways of measuring what a ‘fair’ exchange rate might be. The Fundamental Equilibrium Exchange Rate approach estimates the exchange rate that would theoretically be required to bring the two countries’ current account surpluses and deficits into balance when both are operating at full employment. Another approach is to use econometric models based on factors that have predicted interest rate movements in the past to predict the future path of exchange rates. These factors typically include trends in relative productivity growth, terms of trade and net external investment position. The Behavioural Equilibrium Exchange Rate is the result of one such method. We use an average of equilibrium exchange rate metrics to assess whether currencies are mis-priced. For our long-term forecasts, we assume that over 10 years, currencies will return to their fair value levels.
Fig. 3.6: Relative currency long-term valuations (%) 40 30
Expensive
20 10 0 -10
Cheap
-20
TRY PLN SEK HUF RON CZK MXN NOK JPY EUR GBP THB MYR CAD INR CNY TWD CLP KRW SGD IDR AUD ILS USD CHF HKD NZD PHP PEN ZAR RUB BRL COP
-30
Source: Aberdeen, consensus BEER and FEER model estimates, March 2017. Note: Provides misalignment from long-term fair value estimates. Orange dots show average of BEER (Behavioural Equilibrium Exchange Rate) and FEER (Fundamental Equilibrium Exchange Rate) models. Blue line shows the difference between the FEER and BEER model estimates. Where BEER is derived from econometric estimates of real currency appreciation/depreciation based on underlying trends in currency fundamentals. FEER is derived from an econometric approach to estimating current account positioning.
However, at lower levels of over- or under-valuation (when currencies are less than 20% away from their equilibrium level) these fair value metrics do not offer a strong signal. It is clear from history (see following chart, Fig. 3.7) that currencies can trend significantly in one direction or another for several years before reverting to fair value. Fig. 3.7: Trade Weighted US Dollar Index 150 140 130 120
Conversely the euro and the yen are historically cheap versus the dollar but widening interest rate differentials will keep them weak in the short term. Sterling is also fundamentally cheap versus the dollar, but the future path is highly Brexit dependent. It is possible that we end up with a rather harder Brexit than is currently expected, in which case sterling may decline further. EM currencies are less cheap versus USD than most DM currencies. On a three-year horizon we assume a weakening bias for EM currencies given expected dollar strength, together with downside risks of Chinese renminbi depreciation and the general risk that China slows down faster than expected. Returns from currency hedging Currency risks can be mitigated via currency hedging. The can be achieved in various ways, but the simplest method is to use foreign exchange futures contracts. The process of currency hedging also generates a positive or negative return itself. So, in addition to forecasts for local currency and ‘base currency’ returns, we also provide returns forecasts that include adjustments to take account of hedging costs (see our expected returns table at the start of this report, Fig. A). Currency futures contracts are priced based on interest rate differentials. This reflects the no-arbitrage principle: with efficient markets you should not be able to systematically exploit currency differences by borrowing in one currency to lend in another. There are exceptions to this rule, but it provides a guide. When local rates are higher than overseas rates, the hedge return is positive and vice versa. Our assumed hedging return approximately equals this interest rate differential. Fig. 3.8: GBP hedging cost (%)
110 100
10Y USD
(1.2)
80
EUR
0.8
70
JPY
1.3
AUD
(1.5)
90
60 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 2013 2017 Source: Federal Reserve Bank of St. Louis (FRED), February 2017. Note: Trade Weighted US Dollar Index: major currencies, Index Mar 1973=100, weekly, not seasonally adjusted
Source: Aberdeen, March 2017. Note: Costs are in percentage, per annum. Positive cost refers to a gain from hedging.
This can result from temporary differences in monetary policy or business cycles. For this reason we sometimes take a short-term view on currencies that is different to the long-term mean reversion assumption.
“We think the US dollar is expensive versus its fair value, but we assume USD will continue to be strong over the next couple of years at least.” This is the case at the moment. We think the US dollar is expensive versus its fair value, but we assume USD will continue to be strong over the next couple of years at least, driven by the significant interest rate differential that we expect to exist between the US and most other developed economies. This strength will eventually fade, but not in the near term. aberdeen-asset.co.uk
37
Chapter
04 38 document_name
Regional outlook
• The world has been enjoying a synchronised cyclical upswing, driven partly by a stimulus-driven rebound in China and partly by renewed optimism about US growth • Our base case is that the upswing continues in the near term, though there are clouds on the horizon • China must eventually rein in credit growth. Doing so will likely result in slower growth for China and related economies • Uncertainty about Trump’s tax cuts makes it hard to predict the US outlook.
aberdeen-asset.co.uk39
Regional outlook
“Recessions are one of the few events that have the power to reduce multi-year equity market return.” The business cycle is important when forecasting on this time horizon because recessions are one of the few events that have the power to reduce multi-year equity market returns substantially below their long-term trend level. If we think the risk of recession is high, we are likely to want to lower our exposure to economically sensitive assets like equities. Conversely, brave investors who are willing to buy equities during the recession are often rewarded with high returns as they benefit from the rapidly rising prices that normally accompany the rebound. To produce our economic forecasts for this medium-term horizon, we use a state-of-the-art economic modelling tool. This is a variant of the ‘dynamic stochastic general equilibrium’ models that are commonly used by economic forecasters. We acknowledge that the global economy is a complex system that any model can only capture imperfectly. To combat this difficulty, we consider a variety of scenarios for future economic prospects. Each one is based on a set of plausible, scenario-specific starting assumptions. We then run the model to see what it tells us about potential outcomes. We also make use of other modelling approaches. For example, we have econometric recession-risk models which look at the behaviour of economic and market variables that have predicted recessions successfully in the past. We take the output of our various models and use our judgement about the most likely ‘central case’ scenario as well as the probability of the various alternate scenarios. A key step is to decide whether risks are evenly balanced around the central case, or tilted upwards or to the downside.
“Global indicators of macroeconomic momentum suggest that the global growth rate has improved in the last six months.” Our central case Global indicators of macroeconomic momentum suggest that the global growth rate has improved in the last six months. Our central forecast is that this improvement will continue in the next year or two. This is driven by a major fiscal and credit stimulus in China, and expectations of stronger growth in the US, partly on the recent improving trend but also on expectations of modest fiscal stimulus. We also expect Europe’s slow recovery to continue. However we
40
Regional outlook
expect the UK to slow due to both Brexit uncertainty and the impact on real incomes of higher inflation following sterling’s post-Brexit-referendum depreciation. Our central forecast for emerging economies is for slightly better performance. This is anchored by the return to faster growth in China and boosted by recoveries in Brazil and Russia.
“We foresee a mild cyclical upswing against a background of structurally slower growth.” Inflationary pressures are also likely to be a little stronger as output gaps close – particularly in the US – and as China reduces overcapacity in its commodity sectors. So nominal GDP growth – which is the fundamental driver of corporate revenue growth – should be stronger too. It’s worth emphasising that this positive medium-term view on growth is only positive relative to the modest standards of recent global growth. The long-term picture described in Chapter 2 remains in place. We foresee a mild cyclical upswing against a background of structurally slower growth.
Uncertainty and scenarios While this view describes what we consider to be the most likely outcome, there is an unusually high degree of uncertainty surrounding our forecast. We have considered a range of other scenarios in which outcomes are in some cases worse and in others better. Fig. 4.1: Non-inflationary growth (%) 3 Non-inflationary growth
1 0 -1
-4
Trump stimulus
European political risk Secular stagnation strikes back
China hard landing
Downturn -5 -6 -5
-4
Europe pulls together
Bond market sell-off
China reins in credit
-2 -3
Upturn Global productivity rebound
2
GDP
The previous chapter explained that our long-term forecasts are based on secular trends in demographic and other economic factors. Our medium-term (1-5 year) view is driven more by the current performance of each regional economy and of the phase of their business cycles – the regular cycles of recessions and recoveries that occurs in most economies.
Hard Trump -3 -2 -1 0 Consumer price index (CPI)
1
Stagflation 2 3
Source: Aberdeen, March 2017. Note: Size of bubbles reflects probability of scenarios.
Of these the most important scenarios are those relating to the economic policies of the Trump administration and China’s management of its twin debt and rebalancing challenges. These scenarios are most important because: they affect the world’s two largest economies; outcomes that diverge from our central scenario are quite likely; and the likelihood of these scenarios is heavily influenced by policy choices rather than slow-moving economic trends.
China downside scenarios China’s growth in recent years has been heavily dependent on very rapid credit growth – particularly in the corporate sector. The investments financed by new credit are showing progressively lower returns. In the long term, credit cannot grow faster than debt-servicing capacity; so sooner or later China will have to rein in its credit growth. A recent IMF study1 shows that countries that have experienced similar rapid credit expansions have subsequently endured extended periods of slower growth and in some cases a financial crisis as well. Fig. 4.2: Private credit as a percentage of GDP 220 190 160 130 100 70 40 1980
1986 Japan
1992 Thailand
1998 Spain
2004
2010
2016
China
Source: Bloomberg, as at 13 March 2017. Note: Calculated as credit to private non-financial sector (adjusted for breaks).
5Y Credit to GDP ratio change (%)
80 70 Thailand
Spain
50 40
20 10 -8
-6
-4
Rebalancing and controlling credit growth are inter-related. Much credit growth arises from attempts to stimulate fixed-asset investment growth. Reducing credit growth and reducing fixed-asset investment go hand in hand. The challenge is to enable consumption to grow fast enough to offset slower fixed-asset investment growth and the impact of deleveraging. China’s consumption sector has been growing quickly in recent years, but it is unrealistic to expect consumption growth to be high enough to fully offset lower fixed-asset investment growth while the economy continues to grow at its current 6-7% rate. We need to prepare for the possibility that China undergoes a period of markedly slower growth as it rebalances.
“China has accounted for around one third of global growth during the last five years.”
It is hard to know exactly when China will act to restrain credit growth and accelerate its rebalancing process. In order to preserve stability before the important 19th Party Congress in October 2017, it is likely that policies to rein in debt are likely to be postponed. But the longer they are postponed the greater the risks.
Japan
30
0 -10
Rebalancing Debt problems are not the only challenge facing China. The country is also attempting to rebalance its economy from a model where growth is driven by fixed-asset investment (investment in construction, industrial development and infrastructure), to one driven by consumption.
Wider implications China has accounted for around one third of global growth during the last five years. Were China to grow significantly more slowly – at 3 to 4% say, rather than the current 6 to 7% – this would have significant knock-on implications for global growth. Our scenario work suggests emerging-market commodity exporters, and Asian economies most exposed to China, would be most affected, but the implications would be global.
Fig. 4.3: Credit booms are normally followed by post-boom GDP slowdown
60
“The challenge is to enable consumption to grow fast enough to offset slower fixed-asset investment growth.”
-2
0
2
Difference in 5Y avg. annual GDP growth pre vs. post Source: IMF, Aberdeen, 2016. Note: Dots show difference in average GDP growth pre-credit boom versus post-credit boom. In most cases GDP was much slower in the post-boom period.
This suggests that it is likely that China will face a similar slowdown. Though it seems quite possible that China will be able to avoid a financial crisis, particularly if it begins the process of addressing the debt problems in the corporate sector sooner rather than later. Unlike many other economies in the IMF study, China’s banking system remains under state control, so a cascade of banking collapses may be avoidable.
4 Using our scenario modelling, we have explored some of the consequences of either a managed slowdown or a more disruptive hard-landing and the economic implications are negative both for China and the world. On a three-to-five year horizon we think the balance of risks for China are to the downside. As we will see in subsequent chapters, this has implications for our views on returns for the most China-exposed asset classes.
Maliszewski, et al (2016) Resolving China’s Corporate Debt Problem http://www.imf.org/external/pubs/ft/wp/2016/wp16203.pdf.
1
aberdeen-asset.co.uk
41
Regional outlook continued Trump scenarios The other major area of uncertainty for our medium-term view is the nature and impact of the economic policies of the Trump administration. The Trump administration has set out ambitious goals for cutting taxes, deregulating business, and increasing infrastructure investment. On the other hand, it has suggested it may cancel or renegotiate trade deals, impose tariffs and restrict immigration. These diverse policies could have major implications for the US and global economy and for asset class returns.
“It is still not clear whether, or to what extent, the Trump agenda will be implemented.” It is still not clear whether, or to what extent, the Trump agenda will be implemented. On the positive side, so far, the White House has sounded more constructive on trade than expected. So perhaps fears of increasing protectionism have been overdone.
“The differences between fiscally conservative and moderate Republicans may be too fundamental to allow them to find common ground.” On the other hand, while the administration has set out an outline for an ambitious tax reform plan, it is not clear whether the Republicans will be able to assemble a majority to pass the reform agenda through Congress. The lack of success so far in repealing Obamacare suggests that the differences between fiscally conservative and moderate Republicans may be too fundamental to allow them to find common ground. What policies are we uncertain about? There is also a fair bit of uncertainty about detail. The White House budget outline and the Republican plan are different in key respects. This is important, because the extent of the eventual fiscal stimulus depends partly on where the cuts fall and on whether they will be offset by tax increases elsewhere. Congress look likely to favour a budget that is balanced in the relatively short term. Trump’s emphasis is stimulus. Trump has also announced his intention to implement a large-scale infrastructure programme. Will it receive congressional support? Corporate tax cuts Both sides promise corporate tax cuts. This should boost corporate post-tax earnings. But if, in order to be fiscally neutral, the tax reform agenda combines tax cuts with the closure of loopholes and restrictions on tax deductions, then there will be winners and losers, and the overall benefit to the corporate sector may be less clear cut.
“The evidence of the first 100 days suggests that the administration’s bark may be worse than its bite on trade.” Trade Trump promised a more assertive trade policy and a renegotiation of trade agreements. He has already withdrawn from the Trans Pacific Partnership negotiations and started the process of renegotiating NAFTA. Paul Ryan proposes to introduce “border adjustment” as part of the tax reform package. There has also been talk about unilateral trade sanctions, especially with China, though less so of late. A surge in US protectionism would be negative for global trade and growth. But the evidence of the first 100 days suggests that the administration’s bark may be worse than its bite on trade. Uncertain consequences of policy changes Even if we knew what policies to expect, there are big questions about their economic consequences. A stimulus may result in faster growth – but how much faster? It is unusual for the government to embark on a fiscal stimulus late in the business cycle when unemployment is already low. The stimulus might result in inflationary pressure and a more rapid rise in policy interest rates, offsetting the effect of the stimulus, or in a worst-case scenario, triggering a recession.
“The stimulus may result in higher corporate investment and, so, rising productivity.” On the other hand, it is possible that the stimulus may result in higher corporate investment and, so, rising productivity – allowing growth to accelerate without adding inflationary pressure. It may also result in further improvement in the unusually low level of labour force participation, with similar non-inflationary consequences. Until we have more clarity on the shape of Trump administration policies, there will be a large amount of uncertainty about which of these outcomes to expect. At present, we guess that the risks are reasonably balanced on either side of our central case. We have explored some of these possible outcomes using our scenario modelling tools and they result in strongly different results for the US and global economies. Fig. 4.4: GDP performance under scenarios* 2016 2017 2018 2019 2020 2021 Hard Trump US
1.6
1.4
1.0
0.0
0.0
0.8
World
2.3
2.2
1.9
2.0
2.2
2.7
US
1.6
2.3
3.8
2.3
1.8
1.8
World
2.3
2.7
3.5
3.1
2.7
2.7
Trump stimulus
Average year-on-year % change; GDP growth at 2010 prices and market exchange rates. Source: Aberdeen Asset Management, Oxford Economics, March 2017. Note: Projections are estimates and provide no guarantee of future results. *
42
Regional outlook
Fig. 4.5: CPI inflation performances under scenarios 2016 2017 2018 2019 2020 2021 Hard Trump
100
US
1.3
2.6
2.3
2.4
3.0
2.9
World
2.7
3.1
2.8
2.5
2.6
2.6
Trump stimulus
90 80 70 60
US
1.3
2.9
3.1
2.1
1.9
1.8
World
2.7
3.2
3.6
3.2
3.0
3.0
Source: Aberdeen Asset Management, Oxford Economics, March 2017. Note: Projections are estimates and provide no guarantee of future results.
50 40 30 20 10
Recession risk model History suggests that a US recession is the risk event with the most power to negatively impact returns on most risk assets over a medium-term forecasting period. Many other factors can affect returns in the short term. And in the long term (e.g. 10 years) even big events like recessions tend not to affect average returns very strongly. But, as the chart below (Fig. 4.6) shows, on a three-year horizon, recessions are very significant. Fig. 4.6: S&P 500 per annum returns over 3Y and 10Y horizons (%) 30
0 2010
2011 6 Months
2012 1 Year
2013 2 Years
2014 3 Years
2015
2016
4 Years
5 Years
Source: Aberdeen, Thomson Reuters Datastream, March 2017.
After several years of signalling very low recession risk, in mid-2015 our model (see Fig. 4.7) started to suggest higher risk. This peaked in early 2016 at around 45%, but has subsequently fallen back to suggest lower risks. The recent positive change was driven by improving corporate profits, higher labour force participation, and improvements in the Conference Board’s leading economic indicator. Normally this indicator trends continually in one direction or the other, but there have been periods in the past when it pauses for months or years. We appear to be currently experiencing a pause in the trend. The model led us to increase the downward skew in our view of risk in 2016, but its more positive signals today support a more neutral view on the balance of risk.
20 10 0 -10 -20 1953
Current risk levels Fig. 4.7: Aberdeen Solutions medium-term recession risk model US economy (%)
Regional views 1962 1971 1980 US recession indicator (NBER) 3 year pa return
1989 1998 2007 10 year pa return
2016
Sources: Aberdeen, Bloomberg as at 23 February 2017. Note: US recession indicator (monthly), as defined by The National Bureau of Economic Research. Note: Past performance provide no guarantee of future results.
To forecast recession risk we use an econometric regression model which uses historical predictors of past recessions (CPI inflation; the output gap; credit growth, corporate profit growth and residential investment all relative to GDP; mortgage servicing costs; Conference Board leading indicator; and a financial market conditions index) as independent variables.
“A high recession risk would normally lead us to shift the balance of risk in our strategic asset allocation to the downside.” A high recession risk would normally lead us to shift the balance of risk in our strategic asset allocation to the downside and vice versa.
Fig. 4.8: Real GDP growth (annual % change) 2016 2017 2018 2019 Developed Markets
1.7
1.9
2.1
1.7
Emerging Markets
4.1
4.5
4.9
4.8
Source: Aberdeen, Oxford Economics, as at 24 April 2017. Note: GDP, PPP exchange rate, real (USD Millions: 2010 prices), percentage change YoY. Projections are estimates and provide no guarantee of future result.
Fig. 4.9: Real GDP growth (annual % change) 2016 2017 2018 2019 World
3.0
3.4
3.7
3.5
US
1.6
2.1
2.6
1.7
Eurozone
1.7
1.6
1.4
1.4
UK
1.8
1.9
1.3
1.6
Japan
1.0
1.3
1.3
0.9
China
6.7
6.3
5.9
5.7
Source: Aberdeen, Oxford Economics, March 2017. Note: GDP measure based on 2005 purchasing power parities, a technique used to determine the relative value of different currencies. Projections are estimates and provide no guarantee of future result.
aberdeen-asset.co.uk
43
Regional outlook continued US The US economy’s slow recovery from the financial crisis is nearly complete, the large output gap following the crisis is nearly closed and unemployment is approaching its cyclical low. After a sluggish year of growth in 2016 – only 1.6% – there appears to be some acceleration going into 2017. We forecast growth of 2.1% for the year as a whole. Confidence has improved, retail sales have posted decent gains, and employment growth is maintaining momentum. It is possible this growth may be accelerated further by Trump’s stimulus – our upside stimulus scenario suggests growth of 3.81% might be possible in 2018, though our base case is much lower at 2.6%. This illustrates the fact that the great unknown at this point is exactly what kind of economic policy the Trump administration will deliver and how it will affect an economy that is already running at full speed. During the election campaign Trump talked about boosting growth in the US economy to a heady 3% or 4%, compared to trend growth of a little less than 2%, and promised fiscal stimulus and infrastructure spending to achieve it.
“Trump’s promised fiscal stimulus may serve mainly to increase inflationary pressures.” But if there is, as we suspect, little slack remaining, Trump’s promised fiscal stimulus may serve mainly to increase inflationary pressures, thus accelerating the rise of growth-restraining interest rates and pushing the dollar higher. This will not end up boosting growth by much.
“Carefully crafted corporate tax cuts and deregulation may result in a surge in corporate investment.” Optimists hope that carefully crafted corporate tax cuts and deregulation may result in a surge in corporate investment and a substantial increase in productivity growth. This would enable faster growth without triggering inflationary pressures. This upside scenario could result in a significant boost in growth prospects and would be very welcome. However, at this late stage in the business cycle, we wonder how likely it is that companies will engage in a major capital investment surge. Business investment has been rather muted in recent years. Will this suddenly change? Draft Republican tax plans suggest that capital investment may receive rather more favourable tax treatment than in the past, and perhaps this will be sufficient. But a degree of scepticism is justified. There is also some hope that a tight labour market may tempt some of the large number of workers who have left the labour market since the financial crisis to re-join. It is true that labourforce participation is unusually low. However, most of the leavers are baby-boomers who have retired, and disabled workers. So the scope here is perhaps more modest than it initially seems.
44
Regional outlook
On the other hand, Trump promises to deport illegal immigrant workers in large numbers, and to allow in fewer legal migrants. These policies will serve to tighten the labour market, provoking inflationary pressure and an interest rate response. Our conclusion at this stage is that Trump’s stimulus is likely to have some benefit in terms of growth, but this will be rather modest – in the region of 0.5% per year rather than the 1% or 2% surge he has suggested. With inflation pressures remaining reasonably contained in the near term, we expect the Federal Reserve to continue raising interest rates at a moderate pace. The Fed raised interest rates at the end of 2016 and again in March 2017, and may raise interest rates again once or twice in 2017, and three times in 2018.
“Trump’s protectionism, if translated to policy, is a risk to our growth projections.” Trump’s protectionism, if translated to policy, is a risk to our growth projections. Calls for unilateral tariffs on China and Mexico would have repercussions. Domestic demand may weaken amid heightened uncertainty and rising consumer price inflation. Meanwhile, exports would be likely to falter under the burden of increased trade barriers (in retaliation for the imposition of US tariffs), a stronger dollar and slowing global trade. In this scenario, we would expect to see growth to slow to 0% in 2019.
UK Prior to 23 June 2016, the UK economy had been humming along nicely. While there have been some impacts of the Brexit referendum result on inflation, overall GDP growth has remained steady, and for the last calendar year came in at a relatively healthy 2.1%.
“Brexit is expected to have a significant negative economic effects in coming years.” However, Brexit is expected to have significant negative economic effects in coming years. Inflation is likely to rise significantly from recent levels following sterling’s depreciation, from 1.2% now to around 3% by mid-2017. This will weigh on real disposable incomes and therefore consumption growth. With surveys suggesting that post-referendum uncertainty has caused some firms to postpone capital spending, we also expect business investment to make less of a contribution to growth. Though a decent contribution to growth from higher exports because of a more competitive currency should partly offset these effects.
“We expect the Bank of England to look through this spike in inflation.” Despite the rapid rise in prices, we do not expect sterling to depreciate significantly from this point, making this jump in inflation a one-off event. We expect the Bank of England to look through this spike in inflation and to keep rates on hold until growth prospects improve.
We expect interest rates to remain at 0.25% until well into 2019, though with no further quantitative easing once the current programmes of gilt and corporate bond purchases have been completed. Longer term Over the longer term, the eventual shape of Brexit will have a big impact on the economy. The nature of Brexit is even more uncertain after the outcome of the June election. A hard Brexit seemed on the cards after Theresa May ruled out membership of the single market or the customs union. But it is now possible that a minority Conservative government may take a more conciliatory approach, and a softer Brexit may result. From an economic point of view, there are two main dimensions to Brexit uncertainty: the extent of trade and migration barriers at the end of the process and the duration and nature of any transition period.
“The nature of Brexit is even more uncertain after the outcome of the June election.” At one end of the spectrum we might yet end up remaining within the single market as part of a Norway-style arrangement. At the other, the UK leaves Europe at the end of the Article 50 process chaotically, with no trade deal and falls back to World Trade Organisation rules. Estimates of the economic effect of a hard Brexit cover a wide range; with GDP between -2.5% and -9% lower than the pre-Brexit forecast by 20302. The uniqueness of this event and its attendant uncertainties explain the wide range of estimates. Timing is a key issue for investors, ‘no deal’ at the end of the Article 50 process would create a chaotic and immediate economic shock. On the other hand, an extended period in which Britain remains a member of the single market while trade negotiations are concluded would defer and mitigate the economic impact of Brexit, allowing more time for the economy to adjust. With all this uncertainty, it is very difficult to forecast the future of the UK economy with confidence. While our central case is moderately optimistic about UK growth over the next 10 years, the balance of risk is tilted to the downside. This has implications for our forecasts for sterling, gilts, UK equities and property.
“The Eurozone economy has gained momentum during the last 12 months.” Europe ex UK Modest growth to continue The Eurozone economy has gained momentum during the last 12 months. As we discussed in Chapter 2, shrinking labour forces in many Eurozone economies mean that growth is structurally lower than the pre-crisis period. High government debt levels in southern Europe and imbalances within the Eurozone are also headwinds for growth. We expect long-term potential growth rates of only around 1% per year. The short term looks a little better. The glacial recovery from the financial crisis means that there is still lot of slack in many European labour markets, particularly in southern Europe. So the Eurozone economy can hope to enjoy a period of modestly faster growth without triggering inflationary pressure. This is what we expect, with growth of 1.6% in 2017 and 1.4% in 2018 – slightly faster than trend.
“We expect consumer price inflation to increase to 1.5% in 2017 – still a little below the ECB’s target level.” Inflation at last While we think the continuing output gap should keep inflation below European Central Bank (ECB) target levels. We think prices have escaped from the mild deflation of the last few years. They will be boosted in the short term by the temporary effect of rising oil prices since the low in early 2016. We expect consumer price inflation to increase to 1.5% in 2017 – still a little below the ECB’s target level. However, household spending should prove resilient, with improvements in the labour market outweighing the effects of price rises. Thereafter, as the output gap closes, we think CPI inflation will tick up gradually, reaching the target by 2020. Weak euro helps exporters The US Federal Reserve looks set to continue to raise policy interest rates while the ECB maintains policy rates near zero. This should continue to keep the euro relatively weak versus the dollar, strengthening export competitiveness for European companies. Political risks After Trump and Brexit there were understandable anxieties that 2017 elections in France would add to the roster of antiestablishment political victories. This fate was avoided, though the fact that hard left and hard right candidates collectively received over 40% of the vote, suggests that the fundamental risks remain.
HM Treasury (2016) The Long-term economic impact of EU membership and the alternatives https://www.gov.uk/government/uploads/system/uploads/attachment_data/ file/517415/treasury_analysis_economic_impact_of_eu_membership_web.pdf and Begg et al (2016). The economic impact of Brexit: jobs, growth and the public finances https://www.lse.ac.uk/europeanInstitute/LSE-Commission/Hearing-11---The-impact-ofBrexit-on-jobs-and-economic-growth-sumary.pdf.
2
aberdeen-asset.co.uk
45
Regional outlook continued
“Perhaps the biggest populist threat is the risk “Emerging markets are very likely to grow that anti-euro parties will secure a majority in significantly faster than developed markets the 2018 Italian elections.” in both the medium and long term – 4% versus 2% in the next few years.” In the near term, perhaps the biggest populist threat is the risk that anti-euro parties will secure a majority in the 2018 Italian elections. Italy’s economy is still weak; its banking sector suffers from high levels of non-performing loans; and government debt levels remain high. Euro membership is blamed for making it impossible for Italy to regain competitiveness via currency depreciation. So far the European recovery has not been derailed by politics – except in the UK. Nevertheless, after the shock of Brexit and Trump, we would be foolish to be too confident of the signals opinion polls are currently sending.
Japan Japan’s economy is benefiting from a number of factors, including rising consumer demand, infrastructure spending and a weaker yen. GDP growth in 2016 rose to 1.0% and we predict 1.3% for 2017 and 0.8% in 2018. That figure, while low, is around the level we forecast for Japan’s long-term trend growth, given the demographic headwinds it faces (see Chapter 2). In the short term, the biggest boost is being provided by the depreciation of the yen, which has fallen by around 10% against the dollar since Trump’s election victory. It appears likely there is further weakness to come. We forecast the yen could fall to around 117.5 versus the dollar by the end of 2017 and remain at a similarly weak level throughout 2018. This is good news for exports, which we predict will grow by 0.7% in 2017 and 1.3% in 2018. Will inflation finally re-emerge? Demand in Japan is structurally depressed by the very high savings rate in the corporate sector. This, together with weak wage growth, has persistently depressed inflation. It is possible that this may start to turn around. Energy prices and import costs are set to rise, driven by global commodity prices moves and yen weakness. However, with the labour market tightening, wage growth may also finally start to increase. History suggests that we should not be too optimistic. Our central case is that inflation will remain low by global standards, increasing from -0.1% in 2016 to 0.6% in 2017 and 0.8% in 2018. But deflationary concerns should dissipate. Against this backdrop, we think it likely that the Bank of Japan (BoJ) will continue to defend the 10-year government bond yield target of “around 0%” for now. On the other hand, we remain alert to the possibility that quantitative easing will finally have its intended effect and inflation might surprise on the upside, allowing the BoJ finally to end QE. On the downside, the main risk comes from a potential China slowdown. China accounts for almost 20% of Japan’s exports. A large downturn in the Chinese economy would have major global implications and could therefore hurt Japan further, as the yen strengthens due to its reputation as a ‘haven’ currency.
46
Regional outlook
Emerging markets Emerging markets (EM) account for over 80% of the world’s population and nearly half of global GDP growth. On that basis, it may seem unjust to compress discussion of the many large countries in this region to a single section in this review. However, the reality is that from an investment perspective, the total value of emerging-market equities and bonds remains relatively modest as a share of the global total – under 15% of MSCI World equity index for example. This investor-oriented perspective also drives our focus within EM in this report. The most popular regional equity index – the MSCI emerging-market equity index – is highly exposed to East Asia (60% of the total), with rather less exposure to Latin America and European EMs. Emerging markets are very likely to grow significantly faster than developed markets in both the medium and long term – 4% versus 2% in the next few years. This is driven mainly by faster productivity growth as countries ‘catch-up’ with productivity levels of advanced economies. In countries like India and Nigeria, population growth is also a big factor. However, this rate of growth will be rather slower in the next 10 years than the previous decade as China’s economy rebalances and population growth slows, particularly in East Asia. We expect EM growth to pick up speed in the near term, partly because Russia and Brazil are rebounding from their recessions, but more importantly because of China’s large fiscal and credit stimulus. We forecast EM growth of 4.1% for 2017 rising to 4.5% in 2018, up from 3.4% last year. Inflation has also picked up a bit, particularly in producer prices. But we expect it to remain contained.
“The biggest question for EM is how soon and how quickly China acts to address its credit and rebalancing problems.” The biggest question for EM is how soon and how quickly China acts to address its credit and rebalancing problems discussed at the beginning of this chapter. The pain would likely be felt most in the EM economies most exposed to China’s demand for commodities, whose prices would fall due to a sustained slowdown in China’s fixed-asset investment growth.
The possible impact of ‘bad’ Trump Another risk for our central case is that the Trump administration may adopt a protectionist trade policy. Under such a scenario, Trump could impose tariffs of up to 45% and 35% on Chinese and Mexican goods respectively. Subsequently, as US consumers switch away from the higher-priced Mexican and Chinese imports, the US would then levy 20% tariffs on South Korea and Taiwan. A trade war could then ensue, with Mexico, China, South Korea and Taiwan all retaliating with similar tariffs on US exports. Under this scenario our EM growth forecasts would be revised down notably, with China growth falling to 4%. Dollar strength may also be a problem for EM economies. Higher interest rates in the US, particularly if coupled with protectionist policies, may lead the dollar higher. This will create stress for EM companies and countries with substantial dollardenominated debt. Avoiding the risks EM economies are by no means equal in their exposure to the risks of a China slowdown or Trump trade policy. India, for example, is relatively less exposed to either of these risks. The economy continues to develop rapidly, with a growing share of global GDP. The Reserve Bank of India has been steadily building its credibility, and the appointment of Urjit Patel as Governor signalled the commitment to inflation-targeting and adherence to monetary policy discipline. Prime Minister Narendra Modi has set his sights on what he wants to achieve, and reform is well underway. While investors would like to see progress speed up, a quick fix is not what is required. Genuine progress will inevitably take time. Brazil continues to emerge from the depths of its recession, and 2016 has been a particularly strong year for Brazilian equities. Inflation has also fallen to a near two-year low while the current account deficit continues to narrow. A Trump presidency will not trouble Brazil as much as its Mexican neighbour either, due to the closed nature of the economy and the fact Brazil does not trade as much with the US. The South American powerhouse remains rather more exposed to China risks. We expect the Brazilian economy to exit the recession next year, although the pace of recovery will be slow and there are still some tough reforms for President Michael Temer to pass – not aided by yet another corruption scandal.
aberdeen-asset.co.uk
47
Chapter
05 48 document_name
Equities
• Slower earnings growth and somewhat stretched valuations in many markets mean equity returns are likely be lower than in recent years • Europe is the most promising region in the near term, given a stronger cycle and more headroom for earnings growth • Strong opportunities also exist in EM, though we prefer regions with lower exposure to commodity demand from a slowing China
aberdeen-asset.co.uk49
Equities Introduction
We arrive at our forecast for long-term equity returns by breaking returns down into their component parts: earnings per share growth (from growth in corporate revenues per share and profit margins), valuation mean-reversion, and dividend income; and then forming a view of the likely trends for these components. The following chart provides an indication of the relationship between these components and the sources which inform our views on trends. Fig. 5.1: Equity forecasting model Total return = Capital return + Income
Capital return Earnings per share (EPS) growth x Valuation change
Dividend per share EPS x Pay-out ratio
Valuation change Valuation future
EPS growth Revenue growth x Profit margin growth
Valuation now
Change in share-count growth
Valuation metrics 12-month forward P/E, P/Sales, P/B, divident yield 7-year mean-reversion assumptions
Revenue growth Nominal GDP adjusted for international revenues
Our forecasts operate at two levels: asset class – enabling clients to make informed allocation decisions between equities, government bonds and other classes; and regional – allowing them to make geographical allocation decisions. For some investors, this is as simple as the choice between their domestic market and international equities; for others, it is a more detailed decision about balancing US versus Japan versus Europe versus emerging markets. In both cases, the allocation decision is informed not only by our returns forecast, but also by our view of the relative risk of different assets and the correlation of their returns with one another. For example, emerging-market equity returns are more volatile than developed markets, so higher expected returns are needed to justify taking more risk. Similarly, Japanese equities are typically less tightly correlated with other regions, offering greater diversification. We use a software ‘optimisation’ tool to select the best combination of assets given our expectations of return, risk and correlation.
“In the last 50 years there have been decadelong periods with very low EPS growth and other periods where growth has been rapid.1” Source: Aberdeen, Thomson Reuters Datastream, March 2017.
1
50
Equities
Profit margin Weak trend reversion assumptions with sector input
Pay-out ratio Future trend in pay-out ratios
Share-count Use index divisors as proxy for net dilution
The drivers of equity returns Earnings per share Over the long term earnings per share (EPS) growth is a large and variable source of return for equity investors. In the last 50 years there have been decade-long periods with very low EPS growth and other periods where growth has been rapid. To understand how EPS changes over time, it is useful to consider its three components. Earnings are a function of revenues multiplied by profit margins. Earnings per share is, unsurprisingly, earnings divided by the number of shares issued, the ‘share count’. Revenues Economic growth is a good guide to corporate revenue or sales growth. Gross domestic product measures the total output of an economy. In market economies the majority of this output is composed of the sales companies make to households. The longterm correlation between growth and revenues is high. As we saw in Chapter 2, demographic factors mean that economic growth in most regions is likely to be significantly lower in the future than it has been in the past. This suggests that revenue growth will likely be a lower. There are two adjustments we need to make to GDP before we can use it as the basis for revenue forecasts. GDP is expressed in real terms, after inflation has been subtracted. But corporate revenues are expressed in nominal terms (without subtracting inflation). So we use nominal GDP (real GDP + expected inflation) as the basis for our sales forecasts.
Another problem is that a portion of corporate revenues in any country come from sales overseas. Overseas GDP growth might be slower or faster than domestic growth. For example, GDP growth in Japan is expected to be extremely low, but many Japanese companies derive a large share of their revenue from sales in faster-growing Asian countries. We need to adjust for this.
“GDP growth in Japan is expected to be extremely low, but many Japanese companies derive a large share of their revenue from sales in faster-growing Asian countries. ”
For each regional market index, we adjust our revenue growth assumptions to reflect the nominal GDP growth rates in the countries from which companies generate their revenues. In some countries this makes little difference but in others the difference is large. The London stock market is a case in point: around 40% of FTSE100 corporate revenues come from the UK, 25% from the US, 15% from the rest of Europe, and 20% from the rest of the world. The following table shows our assumptions for each region.
Fig. 5.2: International revenues by source (%) UK
US
Europe ex UK
Japan
Developed Asia
Emerging Markets
UK equity
42.9
25.1
14.3
4.6
2.5
10.6
US equity
2.0
80.0
6.3
2.5
1.4
7.9
14.2
25.7
45.1
4.1
2.2
8.7
Japan equity
2.5
28.0
7.9
45.6
4.4
11.6
Pacific ex Japan equity
1.7
10.2
5.4
12.6
53.4
16.8
Emerging Markets equity
1.4
8.1
4.4
4.3
2.3
79.5
Europe ex UK equity
Source: Aberdeen, HSBC, Oxford Economics, March 2017. Note: Geographic breakdown of revenues by exposure to regional MSCI equity indices.
“Changes to the level of profit margins has often been the biggest source of long-term variation in earnings.”
Fig. 5.3: US after-tax corporate profits to GDP 0.12 0.10 0.08
Profit margins Changes to the level of profit margins has often been the biggest source of long-term variation in earnings. Profit margins vary significantly over time. They show a strong pattern of expansion during economic recoveries and contraction before, and during, recessions. More significantly, in many regions they also show an upward trend over the last 20 years, particularly in the US. The chart below, showing profits as a share of GDP (Fig. 5.3), provides an indication of these patterns. It has often been assumed that profit margins revert to their long-term average over the course of the business cycle. The idea of mean reversion is appealing. If companies become extremely profitable this will attract new entrants and profit margins will be driven back down through increased competition, or so the argument goes. There is also some empirical evidence for this. In the US, for much of the period between 1950 and 2000 profits seemed to mean revert as a share of GDP.
“Whether the upward trend in margins continues, plateaus or reverts to a lower mean, is a key question for equity investors.”
0.06 0.04 0.02 0.00 1950 1956 1962 1968 1974 1980 1986 1992 1998 2004 2010 2016 US recession indicator (NBER)
CPATAX/GDP
Source: Federal Reserve Bank of St. Louis (FRED), Thomson Reuters Datastream, March 2017. Note: Corporate profits after tax with inventory valuation adjustment (iva) and capital consumption adjustment (CCAdj)/gross domestic product, bil. of $/bil. of $, quarterly, seasonally adjusted annual rate
However, as the chart shows (Fig. 5.3), since 2000 the picture has been rather different. In last 20 years, profit margins have increased substantially in the US and several other countries. Although there has still been some cyclicality, with margins collapsing during recessions, the main trend has been steadily upward. Whether the upward trend in margins continues, plateaus or reverts to a lower mean, is a key question for equity investors. Expected returns would be dramatically lower if margins were to mean-revert to their post-1950 average. We think this is very unlikely in the foreseeable future. Average profit margins are driven by structural trends in economies as well as compositional changes to stock market indices. These can over-ride pressures for mean reversion.
aberdeen-asset.co.uk
51
Equities continued Structural trends in margins The winner-takes-all nature of the technology sector is one important factor for profit margins. Many modern technologies have network effects which create a natural monopoly for incumbents. Microsoft with its Windows platform, Google with search, Apple with its high quality ecosystem, and Facebook with its social network all have high profit margins arising in part from the high barriers to entry. These quasi-global monopolies will not last forever, but they may last for a long time and collectively, they represent a different economic structure to the pre-globalisation, pre-internet period, where these kinds of network-monopoly opportunities didn’t exist to the same extent. They act as a barrier to the mean-reversion logic described above. Potential new entrants may be attracted by Google’s high profits, but they know that trying to compete with Google’s dominant position in internet search is a near impossible task. The result is that Google’s profit margins are not eroded by competition. Eventually a new technology will likely come along, disrupting Google’s position, but this seems a distant prospect.
“Most of the long-term growth trend in margins has come from just two sectors: technology and banking.” For these kinds of companies profit margins are structurally higher. If we decompose US profit margins by sector, it is clear that most of the long-term growth trend in margins has come from just two sectors: technology and banking, both of which have high barriers to entry. While post-financial crisis regulation has somewhat weakened the profits available in the banking sector, margins are again high relative to other sectors. And margins in the tech have made up the difference. Fig. 5.4: S&P 500 Index historical sector earnings as a % of sales 12 10
6 4 2 0 1990
1994
1998
2002
2006
2010
2014
US recession indicator (NBER) Utilities Telecommunications Industrials Materials Health Care Energy Consumer Staples Consumer Discretionary Information Technology Real Estate Financials Sum Source: Aberdeen, Bloomberg, Siblis Research, March 2017. Note: Calculated as a function of index earnings/sales multiplied by index weighting. Past performance provide no guarantee of future results.
It is also worth noting how average profit margins relate to the changing composition of the index. As banks and tech companies have risen as a share of the revenues of the US index, their high profit margins have assumed a greater weight. In 1980, the finance and IT sectors accounted for about 10% of the revenues of the S&P 500; by 2017, they were responsible for around 25% of revenues and a whopping 40% of profits. These sectors now have 52
Share count Investors often focus on forecasting earnings growth without thinking much about changes in the share count. But the number of shares in issue is an important consideration, particularly when thinking about returns at the market-index level.
“By reducing the share count, buy-backs increase the value of the remaining shares and therefore boost investor returns.” Since the 1980s, companies in the US have increasingly used share buy-backs as a means of returning cash to shareholders. By reducing the share count, buy-backs increase the value of the remaining shares and therefore boost investor returns. The total share count of the US market has fallen in recent years as a result, boosting aggregate earnings per share. In most markets around the world, share count tends to increase over the long term. There are two main sources of new shares. Firstly, listed companies may raise capital via issuing new shares, so-called ‘rights issues’. This dilutes the claims on earnings of existing shareholders. The second is initial public offerings (IPOs), where private companies raise capital by issuing shares for sale on the public markets. This does not dilute the rights of investors in the companies already listed on the index. However, it does mean that a gap opens up between the total earnings of the index and earnings per share. As companies join the index via IPOs, the index’s total earnings rise, but the addition of newly listed shares means that total earnings per share may not. For every market we assess the level of buy-backs and new share issuance to calculate a net change to share count. In the US there has been a net fall in share count of around 1% per year (which, roughly speaking, translates into a 1% boost to earnings per share). In other developed markets, it is the other way around; with net share issuance increasing at about 1% per year, reducing EPS growth by roughly this amount.
8
-2
much higher profit margins than average and this has significantly raised the average profit margins of the index. It is unlikely that these sectors will shrink any time soon. Indeed, with higher interest rates in the US and eventually elsewhere, we might expect banking profits to rise.
Equities
“Rapid economic growth in Asia has been associated with equally rapid rates of company formation.” This issue is most significant for emerging markets, particularly those in Asia. Rapid economic growth in Asia has been associated with equally rapid rates of company formation and public listing. In many respects, more new companies are good for these economies: they draw in new capital investment and, as a result, drive faster economic growth. The problem is that investors sometimes see this rapid growth and assume that the value of their investments will grow equally fast. But they should remember that they have no claim on much of
this faster growth. It is not generated organically by their existing investment, but instead by new capital provided by investors in the newly listed companies. China’s rapid growth in the 1990s provides a good illustration. From 1992 to 2002 the total market capitalisation of S&P/IFC China equity index rose from $18bn to $681bn, an annualised growth rate of 39%. You might imagine that investors in China were made very rich indeed. Not so. During this period the number of companies listed in China rose from 52 to 1296; or an increase of 32% per year. In other words, the growth in the total capital value of the market was not primarily from an increase in the value of existing shares. Instead it came from the addition of newly issued shares and the capital provided by the investors that purchased them. In fact, during the period in question, the S&P/IFC China equity index grew by only 3.5% per year1. Rapid economic growth is often driven in part by fast growth in the capital base, and so does not result in equally rapidly appreciation in share prices. It is not just IPOs: rights issues can also be more of a problem for emerging markets. Emerging markets tend to have less developed banking sectors and bond markets, so companies have tended to make more use of equity issuance to finance capital investment.
Though this has gradually changed as their capital markets deepen. It is worth emphasising that this issue is more of a headwind for market-wide return forecasts than it is for active investors with concentrated portfolios. By ensuring stock selection is focused on companies with strong protections for minority shareholders, active investors can mitigate dilution from rights issues. Another simple way to illustrate this dilution point is by showing the gap between growth in earnings (or profits) and the growth in earnings per share. As a comparison of the charts below (Fig. 5.5 and Fig. 5.6) shows, earnings grow much faster than earnings per share in emerging markets and Asian indices. This is what you would expect with high levels of IPOs and rights issues. The reverse is true in the US, where EPS has grown a little faster than earnings. Again what you would expect when there are net buy-backs. We take account of the change in share count when forecasting regional equity returns. Helpfully, stock market index companies incorporate net share issuance when calculating their index prices. This means dividing market capitalisation by market price gives a reasonable estimate of the rate of change of share count. We look at the trend for each market and extrapolate to the future. We assume that as emerging economies mature the dilution rate falls somewhat.
Fig. 5.5: US and emerging-market equities net income
Fig. 5.6: US and emerging-markets equities EPS
3,000
600
2,500
500
2,000
400
1,500
300
1,000
200
500
100
0 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 Emerging Markets
0 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
US
Emerging Markets
US
Source: Aberdeen, Thomson Reuters Datastream, March 2017. Note: Rebased to 100 at 1 January, 1996.
Fig. 5.7: US, UK, Europe ex UK equity share dilution 14,000
600
12,000 10,000 8,000 6,000
7,000
1,400,000
500
6,000
1,200,000
400
5,000
1,000,000
300
4,000
80,0000
3,000
60,0000
2,000
40,0000
1,000
20,0000
200
4,000
100
2,000 0 1976 US
1986
1996
Europe ex UK
2006
2016
Fig. 5.8: Japan, Pacific ex Japan, emerging-markets equity share dilution
0 2026
UK (RHS)
0 1976
1986
Pacific ex Japan
1996
2006
Emerging Markets
2016
0 2026
Japan (RHS)
Source: Aberdeen, Thomson Reuters Datastream, March 2017. Note: Share dilution is obtained through the market divisor (index market value divided by price). Projections are estimates and provide no guarantee of future result.
Speidell, L et. al (2005) Dilution is a drag…The impact of financings in emerging markets.
1
aberdeen-asset.co.uk
53
Equities continued Equity valuation Share prices often move independently of earnings. This is because the premium investors require for bearing equity risk changes over time, in line with investor risk aversion. When investors are enthusiastic and risk aversion is low, they are prepared to pay high prices for future earnings. During the ‘dot. com’ boom around 2000, valuation multiples reached all-time highs. But euphoria doesn’t last forever, and as risk aversion increases, valuation multipliers decline. These shifts in sentiment can be sudden and dramatic – as often happens at the start of a recession. Or they can be slow and long lasting. For example, during the 1980s, equity investors enjoyed an extended period of bumper returns, as valuation multiples expanded from historical lows in the 1970s towards more normal levels. Equity prices rose much faster than both GDP and corporate earnings during this period as a result.
“Over long time horizons realised returns are strongly and reliably negatively correlated with valuation levels.” Valuation is one of the most useful metrics for Strategic asset allocation. Over long time horizons realised returns are strongly and reliably negatively correlated with valuation levels. Fig. 5.9: S&P 500 Index price-to-earnings correlations and ratios
“Strategic asset allocation adds value by routinely rebalancing portfolios away from assets that have become expensive to those that offer more value.” Metrics How do you assess whether equities are expensive or cheap? There are many methods for answering this question and no uniquely authoritative answer. We look at key valuation ratios: price/forward earnings (P/E), price/book value, price/sales and dividend yield for each regional stock market index. Respectively, these compare current share prices to levels of profits, value of assets, sales and dividend pay-outs. Each ratio has its own strengths and weaknesses, but collectively we think they give us a better picture than any of them provides individually.
“Emerging markets, Japan and Pacific regions are around fair value.” We compare each ratio’s current level to its long-term average. The chart below (Fig. 5.10) shows how the above basket of valuation metrics has varied against its 15- year average. The central 0% line indicates ‘fair value’. Currently, most developed world equity markets are somewhat expensive on this basis – particularly the US. Emerging markets, Japan and Pacific regions are around fair value.
P/E (Forward 12M EPS)
P/E (Trailing 12M EPS)
Returns next year
(0.33)
(0.31)
Returns in next 3 years
(0.57)
(0.41)
Returns in next 5 years
(0.59)
(0.42)
30
Returns in next 10 years
(0.86)
(0.74)
20
Average historical P/E
16.7
16.5
10
Current P/E
18.3
21.8
Source: Aberdeen, Bloomberg, as at 13 March 2017. Notes: Correlations measure relationship between P/E ratio and future total returns (including gross dividends), where a more negative correlation indicates stronger mean reversion. Forward 12M is calculated from Bloomberg broker estimates and consensus, which may exclude one-time extraordinary gains/losses. Trailing 12M is calculated using EPS before extraordinary items. Past performance provide no guarantee of future results.
“If you buy when equities are expensive you are very likely to get a low return over the next decade.”
Fig. 5.10: US, UK, Europe ex UK equity valuations (%) 50 40
0 -10 -20 -30 -40 2001
2007 UK
2010 2013 Europe ex UK
2016
2019
2022
2025
2022
2025
Fig. 5.11: Japan, Pacific ex Japan, Emerging Markets equity valuations (%) 50 40
In other words, if you buy when equities are expensive you are very likely to get a low return over the next decade. If you buy when they are cheap, you are more likely to get a higher return.
30
This correlation is stronger over longer periods. As the table indicates, it is not material after one year, it starts to come through after three years and has a major impact on returns after five years.
0
Strategic asset allocation adds value by routinely rebalancing portfolios away from assets that have become expensive to those that offer more value.
2004 US
20 10 -10 -20 -30 -40 2001
2004 Japan
2007 2010 Pacific ex Japan
2013 2016 2019 Emerging Markets
Source: Aberdeen, Bloomberg, March 2017. Note: Chart shows the level of Aberdeen’s equity valuation basket versus its 15-year average.
54
Equities
Reversion to what mean? We must also consider which long-term average to use. Unfortunately, as the chart below (Fig. 5.12) shows, the average for these metrics itself varies over long time periods. The period from the 1980s to the early 2000s demonstrates that valuation ratios can march upwards or downwards for extended periods of time. Fig. 5.12: S&P 500 historical price/earnings (%) 30 25
Fig. 5.13: Global equity returns 3Y 5.5 7.2 6.2 4.6 5.6 5.9
UK equities US equities Europe ex UK equities Japan equities Pacific ex Japan equities Emerging Markets equities
5Y 5.5 5.5 5.5 4.7 6.3 6.1
10Y 6.1 5.3 5.2 4.6 6.6 6.3
Source: Aberdeen, January 2017. Note: Returns are in local currency and in percentage, per annum. Return projections are estimates and provide no guarantee of future results.
20 15 10 5 1954
Regional equity outlook
1961
1968
1975
1982
US recession indicator (NBER)
1989
1996
2003
2010
2017
P/E ratio
Source: Bloomberg, March 2017.
There are mean-reversion cycles within this period, but these are overwhelmed by the long-term trend. Why did this happen? Equities fell out of favour during the 1970s as very high inflation and extremely high risk-free interest rates made them relatively unattractive. Collectively, investors reduced their exposure to equities, preferring other asset classes. With less money chasing the same flow of earnings, the P/E ratio fell from 19.4 in 1965 to 8.2 in 1978. As central banks brought inflation under control and interest rates began a 30 year downward trend to unprecedented lows, equities came back into favour and asset allocation slowly reversed. This drove the trend in margin expansion and with it the exceptional period of high returns. It is notable that much of the gain for investors during this period came from rising valuations, not from improvement in corporate earnings. These long-term swings in investor sentiment towards equities mean that we cannot simply use a very long-term (e.g. 50-year) average P/E and assume equities will revert to it. It is misleading to compare two periods when the opportunity cost of holding equities is significantly different. Mean reversion should be to a level which reflects the current regime: rather higher than the 1970s, but lower than the irrational exuberance of the technology-bubble period of 2000. This is a matter of educated guess work rather than science. We use a 15-year period for mean reversion (hence the 15-year average used in the chart above - Fig. 5.12). This includes the low valuations during the financial crisis as well the current period of exceptionally low risk-free interest rates and the somewhat higher valuations they justify.
Our base case is that equity returns will be rather lower than their long-term historical average. For the US, we expect a return of 5.3% per annum versus a 50-year average of 6.8%. This is partly due to the lower global economic growth expectations described in Chapter 2. Worsening demographics, mediocre productivity growth and structurally weak demand make for an extended period of sluggish growth. But it also reflects the fact that developed-market valuations are now a little stretched. While we don’t expect a reversion to fair value in the short-term, this weighs on our long-term return forecast. Japan, Pacific and emerging markets offer more value. UK equities Fig. 5.14: UK return factors (%) 8 7 6 5
3.9
5.5
4 3
3.4
2.7
2
1.6
1 0
Revenue Share Margin EPS growth count expansion change Increase
Decrease
Valuation Price Dividend Total return return income return
Subcomponent sum
Total return
Source: Aberdeen, March 2017. Note: Calculated in local currency terms on a three-year per annum basis.
As discussed above, the UK equity market is unusual in that its largest companies are highly international in nature and derive most of their revenues from overseas operations: 43% UK, 25% US, 14% Europe, and 18% for the rest of the world. Our revenue growth assumptions reflect this mix. It means that the UK equity market is much less sensitive to UK growth than most other equity markets are to their domestic economies. The international nature of UK companies means they are less exposed to current worries about Brexit’s effects on the UK economy than one might guess.
aberdeen-asset.co.uk
55
Equities continued
“The 15% fall in sterling value after Brexit resulted in a substantial boost to reported earnings.” Having said that, UK corporate revenues are received in foreign currencies but reported in sterling. This means that the changes to the exchange rate associated with Brexit worries do affect UK share prices. So far this has been beneficial. The 15% fall in sterling value after Brexit resulted in a substantial boost to reported earnings. But these defensive features have their limits. If we do end up with a hard Brexit and if, as many economists expect, this were to do significant harm to the UK economy in the medium term, there would be some negative effects for UK equities. UK profit margins have been very depressed in recent years, due mainly to the high concentration of oil and mining stocks in the UK and the collapse of commodity prices. Commodity prices have revived somewhat and the companies have cut costs aggressively, so we anticipate a rebound in margins, supporting returns. Collectively, UK companies issue slightly more shares than they buy back, so this net issuance acts as a slight drag on EPS. Nevertheless, overall, we assume EPS growth is solid if unspectacular.
“It is uncertain whether the Republican majority in Congress will be able to secure the votes necessary for its tax-cutting agenda.” It is uncertain whether the Republican majority in Congress will be able to secure the votes necessary for its tax-cutting agenda. If they are successful, these tax cuts should enable faster growth in net profit margins, offsetting the negative effects on margins of higher wages and interest costs. We make some allowance for this, resulting in a higher short-term forecast. But the failure, so far, to repeal Obamacare reduces our confidence that this outcome will be achieved. Another major source of earnings per share growth in recent years has been share buy-backs. We think that this is likely to continue – and could even be boosted if tax regulations encourage a repatriation of foreign earnings.
Our valuation metrics suggest UK companies are a little expensive compared to the recent historical average. Our standard mean-reversion assumption somewhat reduces our projections for returns.
Europe ex UK Fig. 5.16: Europe ex UK return factors (%) 8
UK companies pay relatively high dividends, and we expect this to continue. Overall, we think UK equities offer solid returns over the next three years, and slightly stronger returns on a 10-year horizon.
7 6 4
1
7
0
5 4
2.0
7.1
7.2
5.2
4.9
3.5
3.1
Revenue Share Margin EPS growth count expansion change Increase
Decrease
Valuation Price Dividend Total return return income return
Subcomponent sum
Total return
Source: Aberdeen, March 2017. Note: Calculated in local currency terms on a three-year per annum basis.
3 2 1 0
6.2
4.4
2
8 6
3.1 3.1
5 3
US equities Fig. 5.15: US return factors (%)
Revenue Share Margin EPS growth count expansion change Increase
Decrease
Valuation Price Dividend Total return return income return
Subcomponent sum
Total return
Source: Aberdeen, March 2017. Note: Calculated in local currency terms on a three-year per annum basis.
Our base case for US equity returns is 5.3% over 10 years. This return is lower than long-term historical averages, reflecting our expectation that US growth will be rather lower than its long-term average due to slower growth in the labour force and mediocre productivity growth (see Chapter 2); and also due to the fact that valuations are on the high side relative to history.
56
Profit margins are not far from their historical highs. But, as we discuss previously, this is mainly a result of the structurally robust profitability of the technology sector, so it is sustainable. As the business cycle matures, faster wage growth and higher interest costs tend to lead to the erosion of margins, but these factors could be offset by the large corporate tax cuts proposed by the Trump administration.
Equities
Over the longer term, our revenue growth expectations for European companies are held back by lower nominal growth projections. Ageing populations and structurally high unemployment in southern Europe mean that companies cannot expect rapid domestic growth and maintaining inflation may also be a recurrent problem. This hurts revenue growth.
“Europe’s recovery is strengthening and has further to run than that of the US.” However, in the short-term, Europe’s recovery is strengthening and has further to run than that of the US. In Europe, unemployment and the output gap are still high and therefore there is more scope for growth to exceed potential.
Fears that Europe’s clutch of elections will result in another populist victory have so far been confounded. But with a potential election in Italy in the next 12 months, it is too early to remove these concerns from the table. It is also worth noting that European companies derive substantial international revenues, so can gain from more rapid growth elsewhere. We think this will boost their growth somewhat.
“The most positive aspect of our European equity forecast is the potential for a rebound in profit margins.” The most positive aspect of our European equity forecast is the potential for a rebound in profit margins. Unlike the US, European companies have not seen profit margins rebound much since the financial crisis. Various factors – a slower recovery, lower inflation, and a poorly performing banking sector – have held margins down. However, we expect some improvement, at least in the short-term. In particular, the expected pick-up in inflation should help in this regard: if producer-prices rise faster than labour costs, profit margins will receive a short-term boost. European companies collectively issue slightly more shares than they buy-back, so this net issuance is a slight drag on earnings per share (EPS). Nevertheless, overall, we assume EPS growth is solid if unspectacular. Our basket of valuation metrics suggest European companies are trading a little above fair value, suggesting a modest negative return from valuation mean reversion. European dividend pay-out ratios are unusually high at around 70% of earnings. It is doubtful that such high levels are sustainable over the long term, but with capital expenditure running at fairly low levels, higher pay-outs are more affordable. In the short-term, we expect earnings to rise higher, allowing pay-out ratios to fall while maintaining dividends around current levels. Japan Fig. 5.17: Japan return factors (%) 8 7 6 5 2.0
4 3 2
2.9
2.9 1.9
1 0
Revenue Share Margin EPS growth count expansion change Increase
Decrease
4.6
companies have increasingly looked overseas for growth and this does provide some offset to the otherwise extremely low picture for revenue growth.
“In Japan, profit margins are extremely high compared to recent history.” On the other hand, in Japan, profit margins are extremely high compared to recent history. This is partly as a result of Abenomics and the push for reforms to corporate governance, meaning there is more pressure to improve return on equity. Current levels seem sustainable. We assume that margins will remain at or near this high plateau, but we see no strong reason to expect more progress upward. Japanese companies, like their counterparts in Europe, issue more shares at the index level than they buy back, diluting returns a little. Given the large amounts of cash that Japanese companies hold on their balance sheets, there is a possibility that at some point this will change and companies might use buy-backs or higher dividends to disperse it to shareholders.
“Japanese companies hold large amounts of cash due in part to their painful experience in the 1990s, when retrenching banks refused to lend to them.” However, Japanese companies hold large amounts of cash due in part to their painful experience in the 1990s, when retrenching banks refused to lend to them. The Japanese corporate memory is long, and prudential cash holdings look set to remain on corporate balance sheets rather than finding their way to shareholders, pockets. So we project net share issuance and dividend pay-outs remaining around current levels in Japan. Overall, therefore, our EPS growth expectations for Japanese companies is the lowest of all regions. Given that many of the factors involved are structural, it is hard to have much hope for a big change here. The biggest support for Japanese equity returns comes from valuation. Japanese companies are cheaper than other regions using our valuation basket. We expect some modest positive valuation return. Nevertheless, whether you look on a three-year or a 10-year horizon, this region is our least favoured from a returns point of view.
2.6
Valuation Price Dividend Total return return income return
Subcomponent sum
Total return
Source: Aberdeen, March 2017. Note: Calculated in local currency terms on a three-year per annum basis.
The Japanese economy is the most advanced in the demographic transition described in Chapter 2 and its growth prospects are the lowest in the world, severely limiting revenue growth expectations for domestically-oriented firms. Unsurprisingly, Japanese aberdeen-asset.co.uk
57
Equities continued Emerging markets Fig. 5.18: Emerging-markets return factors (%) 8 7 6 5
2.4
5.3
5.9
4 3.6
3
3.4
2 1 0
Revenue Share Margin EPS growth count expansion change Increase
Decrease
Valuation Price Dividend Total return return income return
Subcomponent sum
Total return
Source: Aberdeen, March 2017. Note: Calculated in local currency terms on a three-year per annum basis.
“10 years ago emerging markets saw doubledigit nominal GDP growth. In recent years, growth has fallen to mid-single digits.” In the first decade of this century, emerging-market equities strongly outperformed developed markets. The picture over the last five years has been more mixed. The main reason for the change is the decline in nominal growth rates. 10 years ago emerging markets saw double-digit nominal GDP growth. In recent years, growth has fallen to mid-single digits. Growth is still likely to be rather faster than for developed economies, but the gap is much smaller.
58
Equities
Perhaps the biggest question for EM equity investors is what happens in China. China is by far the largest EM economy, and many other EM economies are heavily exposed to Chinese demand. As we discussed in Chapter 4, the resolution of China’s debt problem may result in a period of significantly slower economic growth. There is much uncertainty about the timing and extent of this slowdown, but our base case is for a material slowdown in China’s growth rate, reducing our expected returns.
“Some EM economies – India is a good example – have strong domestic sources of growth and low exposure to China.” This does not mean EM returns will be low everywhere. Some EM economies – India is a good example – have strong domestic sources of growth and low exposure to China. India has a low weight in the MSCI EM index which we forecast here, but active investors can design a portfolio that maximises exposure to countries with strong domestic growth stories and little exposure to China risk. Another challenge for MSCI EM equity index returns is the gap between rapid growth in earnings and slower growth in earnings per share in the region (see Fig. 5.6). As we discussed previously, high levels of share issuance in many EM economies dilute existing investors, claim on earnings growth – hence the negative red bar in the previous chart (Fig. 5.18). Active investors can mitigate this problem, but our forecast is for the MSCI EM index, so this is a drag on our forecast for index returns.
aberdeen-asset.co.uk
59
Chapter
06 60 document_name
Government bonds
• Government bond yields are low on long historical comparisons, and consequently we expect very low returns • Structural imbalances between desired saving and investment in the economy will hold yields down • As the Fed raises interest rates, US Treasury yields will therefore rise only gradually. UK and German yields will follow but remain below the US.
aberdeen-asset.co.uk61
Government bonds Developed-market government bonds have long been the mainstay of low-risk investment portfolios, and the most important diversifier for equities for higher-risk investors. However, after an historic 30 year bull market, bond returns from here are likely to be a lot lower than in the past – even without assuming a bond market crash. This will be a major challenge for asset allocators. Calculating government bond returns
Fig. 6.1: Roll returns can be significant 2.0
The return on a typical government bond has two main components: an income component and a capital return component.
The capital return is a little more complicated. It is a function of the change in the yield and the duration of the bond. Duration is a bond-market concept related to the maturity of a bond. Duration measures the sensitivity of the price of the bond to changes in the bond’s yield. The longer the duration of a bond (i.e. the longer you have to wait for the bulk of the cash payments), the more sensitive the price is to changes in yield. This concept is useful because it makes it easy to calculate the capital return of a bond, simply by multiplying the duration of the bond by the change in yield. In the upside-down world of bond investment, yields rise as the price of a bond falls, and vice versa. So a fall in yield gives a positive capital return and a rise in yield gives a negative capital return. For example, the 10-year maturity US government bond currently has duration of around nine years. This means that if the yield on the bond rises from 2.5% to 3.5%, this one percentage point movement in yields, multiplied by the duration (1% x 9) results in a capital loss of 9%. For an individual bond, the return can be approximated by simply adding the duration-related capital return to the income return. However, for our strategic asset allocation views, we need to estimate the returns from whole bond indices. How do bond indices differ from individual bonds? Thanks to a never-ending process of short-dated bonds maturing and new, longer dated bonds being issued, the average maturity of a bond index is roughly constant. Time passing shortens the maturity of each individual bond, but bond indices stay more or less unchanged. This introduces another form of return, known as ‘roll’. Usually, longer-maturity bonds have a higher yield than shorter maturity bonds, reflecting greater uncertainty over longer horizons. As time passes and maturity reduces, an individual bond “rolls down the yield curve”, so the yield falls and the price rises. Old bonds maturing and new bonds being issued is equivalent to investors selling high-priced bonds and buying low-priced bonds, and consequently increasing returns.
62
Government bonds
Yield (%)
The income component for a bond can be approximated by the yield-to-maturity (in the following we refer to this simply as the ‘yield’). This is the annualised return an investor will make on an individual bond if they buy today and hold it until it matures.
1.5 1.0 0.5 0.0 -0.5 -1.0
0
5
10
German yield curve
15 Maturity
20
25
30
Return from yield and roll
Source: Aberdeen, Bloomberg, as at 2 March, 2017. Note: For illustrative purposes, maturity in years. Returns based on assumption of unchanged yield curve over one year. Projections are estimates and provide no guarantee of future results.
For bond indices, income and capital returns can be approximated in the same way as for individual bonds, using the average yield and average duration. We include the three sources of returns in our calculations for bond indices. It is worth noting that, despite yields moving up and down in the short term, income returns tend to offset capital returns for bond indices over the long term. In the previous US government bond example, the 2.5% per year income has risen to 3.5%. Over time this higher income gradually offsets the 9% capital loss. As a rough rule of thumb, on a 10-year horizon, the yield at the point you buy the bond index is a fairly good first approximation of the annualised expected return you can hope to achieve, plus a little extra for roll. With government bond index yields today near 0.4% in Europe, 1.9% in the UK, and 2.5% in the US, we can be fairly confident in expecting government bond returns to be low compared with their history.
Forecasting bond yields In order to estimate expected returns using the methods described above we need to forecast how bond yields will change over time. Given that for strategic asset allocation purposes we need estimated expected returns for bond indices, we need to forecast the full yield curve at each of our three, five and 10 year forecasting horizons. To do this we focus on two parts of the curve: yields on short-term bonds (which are primarily driven by the central bank policy interest rate), and the yield on the 10-year maturity (10Y) bond. We use a simple mathematical function to estimate yields for intermediate points on the curve and to extrapolate to yields of bonds with maturities of more than 10 years.
Our forecasts for policy rates in different countries are linked to our inflation and growth forecasts for these countries and our understanding of central bank reaction functions. For example, if we see developments that could push inflation above a central bank’s 2% target, we will forecast a rising path for policy rates. If we think output gaps will grow or the risks of recession high are, our expectations of policy rates will be lower. The yield on a 10Y bond can be decomposed into expectations for inflation and expectations for real (after inflation is removed) policy rates over the 10-year period, as well as a risk premium reflecting uncertainty about these two items, known as the ‘term premium’. Inflation expectations depend primarily on the credibility of central bank targets and perceived changes to their reaction functions. Future real policy rates depend on both the current macroeconomic situation, as seen in output gaps, as well as on long-term structural factors which affect the balance between savings and investment, such as demographics and productivity expectations (see discussion in Chapter 2). Term premium is the compensation investors receive for bearing the risk that policy rates and inflation differ from their expectations over a long horizon. The term premium is dependent on various factors, including for example, the extent of supply and demand for bonds and the extent of market risk aversion. To forecast the 10Y yield we use two methods. First we look at bond market forward yields, which can be derived from the yield curve. These effectively represent the best collective market view of future bond yields. History suggests that these are not always particularly accurate, but over the last five years they have been rather more accurate than consensus economist and central bank estimates. We also use an econometric model of bond yields. Our model is the result of a multiple regression of policy rate expectations for a year ahead, market-implied inflation expectations (so-called ‘breakeven’ inflation), short-term growth expectations in the form of purchasing manager indices, and the size of central bank balance sheets. By inserting our expected values for each of these variables in each country for different points in the future, we can use the model to forecast yields. We then take an average of market forwards and our model to generate a forecast for 10Y bond yields. Fig. 6.2: 10Y US Treasuries yield – model forecast versus actual (%) 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Actual
Model
Model Standard Error
Source: Aberdeen, Bloomberg, March 2017. Note: Projections are estimates and provide no guarantee of future results.
Asymmetric risk, but no bond market crash? Before discussing our detailed views on yields it is worth making some more general observations. First, with bond yields at unprecedented low levels, there is potentially an asymmetric risk facing bond investors, particularly in Europe and Japan.
“Although there are no firm rules, there are limits to how far bond yields can fall.” Although there are no firm rules, there are limits to how far bond yields can fall. Until recently it was assumed that bond yields could not drop below zero because investors would sell bonds and hold cash instead. This was tested by negative interest rate policies in Japan and Europe. It turns out that investors are prepared to continue to hold bonds with slightly negative yields. There are various reasons for this, likely relating to safety and diversification, as well as issues with getting hold of or storing large volumes of bank notes. However, it remains the case that the downside for yields is limited. This means that, within some margin of error, there is a maximum possible capital return that investors can hope to receive. This is the difference between the current yield and some slightly negative number, say -0.5%. So if, for example, yields are at 1% and duration is eight years, the maximum capital return from a downward movement in yields is 1.5% x 8 = 12%. On the other hand, the maximum level to which yields might rise has less obvious limits. If yields were to return to their pre-crisis level, we might expect a rise of 3% from current levels resulting in a 24% (3% x 8) capital loss. In economies where there already is a negative rates policy – Europe and Japan – the return profile is in some ways even more asymmetric. If yields are already negative, you might only hope to pick up an upside return in the low single digits. There appears little reward for a significant downside risk.
“The imbalance between high desired saving and low desired investment in the global economy should hold rates down.” That said, we do not think the risk of a large sell-off is very high. Bond yields are near all-time lows for good reason. Demographic and other structural factors have depressed the equilibrium real interest rate, especially in those regions where yields are currently negative. In the absence of a major change in public policy the imbalance between high desired saving and low desired investment in the global economy should hold rates down. This comfort is not as reliable as we might like. Equilibrium interest rates are not observable and so cannot be measured. They might not be as low as we think. There are counter arguments to the view that they will remain low. For example, retiring baby boomers may spend more than we expect in their retirement, particularly on health care. It should also be emphasised that equilibrium interest rates are, by definition, an equilibrium concept. The economy can depart from equilibrium for periods of time. Rates could rise above their equilibrium level.
aberdeen-asset.co.uk
63
Government bonds continued A more practical reason why bond yields are likely to remain near historically low levels in the short term is the persistence of output gaps in Europe and Japan, which continue to prompt central bank bond purchases. This ‘quantitative easing’ (QE) is supporting the bond market, keeping yields low. Of course, at some point in the next few years it is possible this central bank buying will stop. Our econometric yield model behaves very differently if QE stops – yields become much more sensitive to inflation. So the comfort QE provides bond investors may well be time-limited. On the other hand, there is always a risk that the current phase of economic expansion could come to an end unexpectedly. A recession may come before Europe and Japan have finished with QE. This is not a happy thought, but it is another reason for not expecting a rapid return to a world of high government bond yields.
High debt levels The fact that the global debt stock is now at all-time highs may also serve to limit rising interest rates. In the US, while mortgage debt is below its financial crisis peak, consumer debt and corporate debt have both risen significantly. The economy is consequently highly sensitive to broad ‘financial conditions’ – a measure for how easy it is to access credit – of which interest rates are an important part. The higher the level of debt, the more the interest cost burden will increase as interest rates rise. HSBC estimate that the late 2016 rise in US rates increased interest costs from around 4% of GDP to 6% of GDP. Based on the prospective interest rates suggested by bond markets, this will rise to 8% of GDP by 2019.
“Central banks will have to move very gradually when the time comes for interest rate hikes to avoid tightening financial conditions more than is warranted.” Were rates to increase faster, we would return to a level of interest costs we saw during the financial crisis, which could create strain in the economy. Many companies already have low levels of interest cover, so fairly modest rises in interest rates could trigger defaults and significant economic stress. As a result, central banks will have to move very gradually when the time comes for interest rate hikes to avoid tightening financial conditions more than is warranted, so the level of rates may remain low.
Regional policy rates We do not see much to change our view that equilibrium interest rates will remain low for some time, making it unlikely that interest rates will rise to the levels considered normal before the financial crisis. That does not mean they will not rise at all, though.
64
Government bonds
“The speed at which rates are increased will be much slower than in previous cycles.” US rates US unemployment is now fairly low and, as a result, wage growth is rising. After the unprecedented period of near-zero policy rates, the Federal Reserve has embarked on a path of tighter monetary policy. The Fed has been at pains to suggest that the speed at which rates are increased will be much slower than in previous cycles, although on current forecasts the pace of rate rises is set to increase in 2017. Over the last few years, the Fed has been too optimistic about the strength of the economy and the resulting future path of interest rates. Policy has therefore perhaps been set too tight. The famous ‘dot plot’ forecast consistently expected rates to rise, while in fact they stayed flat as wage growth remained slow. Bond market investors, on the other hand, have tended to expect lower policy rates through this period, possibly reflecting a less optimistic economic outlook. However, for the first time in many years the bond market now more or less agrees with the Fed’s dot plot. Our forecast follows a similar path. We expect the Fed to raise policy rates at a rate of two to three 25-basis-point hikes per year over the next three years, taking rates to 2.5-3% by the end of 2019. We may also see the Fed’s balance sheet shrink as the Fed stops buying bonds to replace maturing securities in its portfolio. This could add upward pressure on yields along the curve. This path of policy rate rises is slow compared to previous tightening cycles. The slower rate of ascent is partly a function of the Fed’s belief that the equilibrium real interest rate is structurally depressed. The tightness of monetary policy is partly a function of the gap between this equilibrium rate and the current real policy rate. Low equilibrium rates call for lower policy rates. Furthermore, the unbalanced nature of global business cycles means that a bigger difference between interest rates in US dollars and other currencies results in a stronger dollar, which also serves to tighten financial conditions. There are risks to both sides of this forecast. On the upside, the proposed Trump stimulus may result in higher inflationary pressure and faster rate rises in response. Normally the government undertakes fiscal stimulus when unemployment is high, not when it has reached low levels. With output gaps nearly closed, inflationary pressure may force the Fed to raise rates, thereby neutralising some of the effects of the stimulus. Other protectionist Trump policies could also act to increase inflationary pressure. Immigration of workers from Mexico, offshoring manufacturing and general openness to international trade has likely served to hold down wages in the US. If Trump succeeds in his aim of reducing immigration and offshoring, we may expect the economy to run into supply constraints sooner.
Second, how much inflation we get from a stimulus depends partly on whether the supply side of the economy can be improved by Trump’s broader policies. If higher wages coupled with deregulation and tax cuts on capital expenditure promote more productivityenhancing business investment, there could be scope for faster growth without increasing inflationary pressure. Third is the question of how much the labour force participation rate will respond to rising employment demand. The participation rate fell after the crisis and has barely recovered. If Trump succeeds in his goal of creating jobs though a stimulus package, perhaps discouraged workers will re-engage with the labour market and participation will rise again, reducing inflationary pressure.
“A shock big enough to cause a recession will come eventually. When it does, the Fed will likely bring rates back down towards zero.” Fourth is the risk that this business cycle will turn down. It is already the third longest in post-war history. A shock big enough to cause a recession will come eventually. When it does, the Fed will likely bring rates back down towards zero. One possible scenario is that anticipated Trump stimulus prompts the Fed to tighten monetary policy too sharply, provoking an early recession. US 10 year yields While we expect policy rates to rise significantly – to 2.5-3% or so over the next three years – we do not expect a parallel rise at the long end of the curve. Our forecast for 10 year yields is a little below 3% in 3 years and only 3.5% in 10 years.
“10-year yields failed to rise much despite an increase in policy rates of several percentage points.” In some respects we expect to see a pattern similar to the period prior to the financial crisis. The Fed Chairman at the time, Alan Greenspan, called it a ‘conundrum’. He was puzzled why 10-year yields failed to rise much despite an increase in policy rates of several percentage points. The answer suggested by Ben Bernanke was that, due to a global ‘savings glut’, equilibrium rates are structurally depressed. This theory has been expanded on significantly since the crisis, as seen in our discussion of why equilibrium rates are set to remain low for an extended period.
Fig. 6.3: US Treasury 10Y yield forecast 3.7 3.5 3.3 Yield (%)
There are four main risks to the downside for Fed rates. First, deficit hawks in Congress may limit the extent of the fiscal stimulus. Having protested aggressively against large deficits over the last eight years, it would be a significant change in tone from House Republicans to pass large tax cuts without offsetting spending cuts.
3.1 2.9 2.7 2.5
0
1
2
Market forwards
3
4 5 6 Years in future
7
8
9
10
AAM Projection
Source: Aberdeen, Bloomberg, as at 27 March, 2017. Note: 10 Year refers to maturity of the bond. Projections are estimates and provide no guarantee of future results.
UK gilts The impact of Brexit on the economy is a key uncertainty for gilt yields. Brexit remains likely to result in weaker GDP growth in the near term; in particular as higher import costs reduce real disposable income. The robust state of the economy in the immediate period after Brexit may suggest that economic forecasters were too pessimistic in expecting a sharp slowdown or even a recession in 2017 as a result of the Brexit vote. Nevertheless, we still expect slower growth during the period of Brexit negotiation, which will likely mean an extension to the period of near-zero policy interest rates. This is in spite of a rise in inflation, mainly the result of sterling depreciation, which the Bank of England is likely to see as a one-off event and not a justification for raising rates. It now looks likely that the UK will end up with a fairly “hard” Brexit, exiting the single market and the customs union, though hoping to replace membership with a bespoke free trade and customs deal. It is highly uncertain how successful the government will be in these goals or how long it will take.
“Falling out of the single market without a trade deal – the ‘cliff edge’ – would likely lead to an extended period of slow growth.” What happens in the period following 2019 when the Article 50 process is scheduled to complete depends a lot on whether Britain has secured a deal to maintain something like the current arrangements on a transitional basis. The prime minister now appears willing to make transitional arrangements, having previously claimed that full arrangements could be reached within the two year negotiation period. Falling out of the single market without a trade deal – the ‘cliff edge’ – would likely lead to an extended period of slow growth or a recession. A transitional arrangement would allow for a much smoother path for growth and potentially even rising policy rates.
aberdeen-asset.co.uk
65
Government bonds continued Longer maturity bonds may be more sensitive to inflation than the Bank of England. Much of the rise in 10 year bond yields in 2016 reflected expectations of higher inflation. But if, as we expect, the current inflationary pressure subsides once sterling finds a stable level, our model suggests 10 year yields will not rise much further. A lot depends on the response of the economy to Brexit uncertainty and whether the UK is able to avoid a ‘cliff edge’ exit. Fig. 6.4: UK gilt 10Y yield forecast 2.9 2.7
Yield (%)
2.5 2.3 2.1 1.9 1.7 1.5
0
1
2
3
Market forwards
4 5 6 Years in future
7
8
9
10
AAM Projection
“Political risks could derail the euro, if voters turn forcefully against the common currency.” However, political risks could derail the euro, if voters turn forcefully against the common currency. For example, were Beppe Grillo and the Five Star Movement to win a majority in the next Italian election on an anti-euro platform this risk would increase significantly. Currently the electoral arithmetic makes this look unlikely, but there are many political risks threatening the established order in the Eurozone, and after surprises in 2016 it would be unwise to write them off. It is also important to note that Europe is vulnerable to external shocks. A recession in the US or a sharper slowdown in China might drag Europe back into recession, with renewed ECB stimulus and lower rates the most likely outcome. On the other, hand faster growth in Europe coupled with higher rates in the US may result in an earlier taper by the ECB and a steepening of the yield curve. Fig. 6.5: German 10Y yield forecast 1.8
Source: Aberdeen, Bloomberg, as at 27 March, 2017. Note: 10 Year refers to maturity of the bond. Projections are estimates and provide no guarantee of future results.
Europe continues to grow a little faster than potential. It still has a large output gap and high unemployment, which maintains downward pressure on inflation. Core consumer price index (CPI) is currently still at only 0.8%, well below the central bank’s target. For this reason the European Central Bank (ECB) is likely to remain in easing mode for the next few years. However, the scope for the current QE process to keep buying government bonds may reach limits by 2018. Concerns have also been raised about the damage done to the financial sector by extremely low interest rates, although ECB studies have largely dismissed these. We therefore expect some tapering in QE over the next two years. This could result in higher yields at the long end of the curve, even though we expect policy rates to remain below zero. Yields on government bonds from the more fragile southern European economies incorporate a risk premium or spread above German yields to reflect higher risks. As we have seen with the recent volatility surrounding the Italian constitutional referendum and French presidential election, there is potential for European government bond spreads to widen on perceived risk to the future of the Eurozone, or if individual countries’ economies are faring poorly. The willingness of the ECB to act as a “lender of last resort” to governments is the key defence against threats to the Eurozone from economic risks, and despite doubts about this creeping in when Greece was in acute difficulty in 2015, it is generally thought to be reliable, especially for larger member states.
66
Government bonds
1.4 Yield (%)
Eurozone government bonds
1.6
1.2 1.0 0.8 0.6 0.4
0
1
2
Market forwards
3
4 5 6 Years in future
7
8
9
10
AAM Projection
Source: Aberdeen, Bloomberg, as at 27 March 2017. Note: 10 Year refers to maturity of the bond. Projections are estimates and provide no guarantee of future results.
“Enthusiasm about a major regime shift in Japan towards permanently higher inflation expectations has been pared back.” Japanese government bonds Enthusiasm about a major regime shift in Japan towards permanently higher inflation expectations has been pared back over the last couple of years, despite unemployment falling to low levels. The Bank of Japan’s continued failure to create domestic inflation despite extensive and inventive monetary policy stimulus measures raises questions about the ability of monetary policy makers to create inflation when nominal interest rates are at the zero lower bound. After serial failure it is not easy to be optimistic that inflation will rise to target levels in the near future. The corporate sector continues to save a large share of its income and wage growth remains low. Until this changes it seems likely that the Bank of Japan will remain in easing mode with policy rates at or below zero.
The Bank of Japan’s (BoJ) policy of targeting the yield on the 10 year Japanese government bond makes it relatively easy to forecast yields up to this maturity. We assume the BoJ hits its target, though we might expect a steeper curve from this point as expectations of inflation build. This policy of targeting 10 year yields makes assessing risks around the target tricky. Scope for further downward movement in yields relies on the BoJ lowering the target in a further attempt to promote inflation. It is also possible that the current set of policies starts to work and delivers a sustainable rise in inflation, allowing the BoJ to raise or even remove its 10 year yield target resulting in a significant bond sell-off.
Yield (%)
Fig. 6.6: Japan 10Y yield forecast
But the central message is that bond returns everywhere are much lower than they have been in the past. This creates an incentive for investors to look for other sources of income and diversification from equities. Fig. 6.7: Government bond returns 3Y
5Y
10Y
UK Gilts
1.5
0.9
1.1
UK Index Linked Gilts
0.2
0.5
1.3
US Government Bonds
1.8
1.9
2.3
US Index Linked
1.5
1.9
2.6
Euro Government Bonds
0.8
0.6
1.1
(0.4)
(0.1)
0.6
1.5
Euro Index Linked
1.2
Japanese Government Bonds
0.1
0.0
(0.1)
Australian Government Bonds
1.9
2.1
2.6
Australian Index Linked Government Bonds
1.9
2.3
2.9
0.9 0.6
Source: Aberdeen, March 2017. Note: Returns are in local currency and in percentage, per annum. Projections are estimates and provide no guarantee of future returns.
0.3 0.0
0
1
2
Market forwards
3
4 5 6 Years in future
7
8
9
10
AAM Projection
Source: Aberdeen, Bloomberg, as at 27 March 2017. Note: 10 Year refers to maturity of the bond. Projections are estimates and provide no guarantee of future results.
Return expectations Bond returns are a function of yields. So the yield assumptions above drive our regional bond forecasts. In the US we expect the Fed to raise short-term rates by approaching 2% over the next three years, but while we expect the 10Y yield to rise, our forecast is that it will rise only modestly (30bps or so). In the UK, Europe and Japan, short-term interest rates are unlikely to rise much, if at all, over the next three years, but we expect 10Y yields to rise a little – by a similar amount to the US.
“Modestly rising yields result in a small amount of capital loss for all these markets, but this is offset by the slightly higher income higher yields provide.” Modestly rising yields result in a small amount of capital loss for all these markets, but this is offset by the slightly higher income higher yields provide. In the US, yields are currently higher than those in other regions and this gap will remain. This means that there is more income to offset the capital loss and returns will be slightly higher than elsewhere.
aberdeen-asset.co.uk
67
Chapter
07 68 document_name
Credit
• Our view on the credit cycle drives our corporate bond forecasts, in combination with long-term spread reversion to average levels • The latest indicators suggest a fairly neutral view on US credit, but Trump’s tax policies could create a more favourable environment • The risk of a spike in spreads may be rising, and default rates are elevated • Loans may be attractive in a rising rate world.
aberdeen-asset.co.uk69
Credit Corporate bonds In an age where expected returns on government bonds are unusually low, investors have been attracted to corporate bonds for the higher yields they offer. These higher yields are earned by investors in return for bearing the risk that companies will default, or that the risk of default will increase. Investment-grade versus high-yield bonds Corporate bonds have different degrees of credit risk. Bonds issued by financially strong companies have high credit ratings (e.g. AA) and normally carry low credit risk. Their default rates are very low (less than 1%) and, as a result, their spread over government bonds is also low. US ‘investment-grade’ (IG) bonds currently have an average spread of 120 basis points and a yield of 3.3%. Bonds issued by weaker companies have lower credit ratings (e.g. BB+ or lower), higher default rates and higher credit spreads. These bonds are not eligible for inclusion in the investment-grade indices and find a place in the ‘high-yield’ (HY) index. The US highyield index currently has an average spread of 410 basis points and a yield of 6.3%.1 Drivers of corporate bond returns As with government bonds, corporate bond returns can be separated into income and capital return components. Income return approximates to the bond’s yield to maturity, while capital return results from changes to the bond’s yield and is calculated by multiplying the duration of the bond by the change in yield. However, unlike government bonds, corporate bonds also have a degree of credit risk: the risk that the bond issuer will default on its debt. This risk manifests in credit spreads and losses from defaults. Credit spreads Bond investors think about corporate bond yields as composed of the underlying government bond ‘risk-free’ yield and a ‘credit spread’. The spread is also known as the credit risk premium, and represents the additional return earned by investors as compensation for taking the risk that bonds will default.
“Yields on corporate bonds can therefore change for two reasons: a change to market views on government bonds yields or a change in market perceptions of credit risk.” The credit cycle Fig. 7.1: Stages of the credit cycle
Yields on corporate bonds can therefore change for two reasons: a change to market views on government bonds yields or a change in market perceptions of credit risk. Spread widening results in a negative capital return, and vice versa. Default loss Over time a small fraction of bonds will end up defaulting on payments. Defaults are losses which reduce returns directly. Normally some of the value of defaulting bonds is eventually recovered, so we estimate both a default rate and a recovery rate for each of the bond indices we forecast, and combine these to give a loss rate. For the US high-yield index the long-term average default rate is around 4% and the recovery rate is a little below 50%, giving an expected loss rate of around 2% which is subtracted from our forecasts for long-term bond returns.2 For investment-grade bond indices, defaults are highly unusual because high-rated bonds will typically be downgraded and fall out of the index before default occurs. As such, defaults have little direct bearing on our returns forecasts. However, bonds falling out of the index introduces another factor that reduces returns in investment-grade bond indices. Bonds will typically ‘price in’ bad news before being removed from the index. So the index return is lower than the average yield would suggest – income returns are offset by losses from ratings downgrades. Over long horizons this may cut the return from the credit spread by something in the region of half, although this is highly variable. Maturity differences Investment-grade bonds tend to have longer maturities than highyield bonds. This is important because it means the duration of the indices differs significantly. The combination of lower credit spreads and higher duration means that investment-grade bond returns are more affected by changes to government bond yields and less influenced by credit risk, and vice versa for high-yield bonds. Bonds in different regions have different combinations of these risks. UK investment-grade bonds tend to have longer maturities so they are more influenced by duration risk than US investmentgrade bonds.
High growth
Rebound • Widening of profit margins from cost reductions • Growth of free cash flow • Reduction of leverage
Expansion • Company debt levels increase • Pick-up in speculative activity • Rise in volatility Rising leverage
Declining leverage Restructuring • Firms focus on cost cutting, generating cash and surviving • Debt is repaid by companies • Balance sheet health improved
Slowdown • Credit contraction from fall in asset values or recession • Weaker earnings constrain deleveraging
Low growth Source: Aberdeen, February 2017. Source: Moody’s (Annual default study: corporate default and recovery, 1920-2016), February 2017.
2
Source: Bloomberg, Bank of America Merrill Lynch.
1
70
Credit
“The credit cycle is one of the most important factors when thinking about credit risk. It describes the expansion and contraction of access to credit over time.” The credit cycle is one of the most important factors when thinking about credit risk. It describes the expansion and contraction of access to credit over time, and is more or less aligned with the business cycle. A stylised credit cycle proceeds through some distinct stages. At the beginning of the cycle, in the period following a recession, many of the weakest companies have defaulted on their bonds and have been removed from the bond index. Companies that survived the recession have reduced leverage on their balance sheet. Corporate bond issuers tend to be in fairly good financial shape and bond indices have reduced exposure to vulnerable companies. As a result, default rates are low and credit spreads are tight. This benign environment may last several years, but as the cycle develops and memory of the recession fades, companies start to become increasingly enthusiastic about their prospects. They tend to borrow more, allowing ‘financial excesses’ to build in the system. Investors start to become nervous about credit risk and spreads may begin to rise. As the cycle matures central banks raise interest rates, making funding costs higher. Corporate earnings are often weaker as companies face higher borrowing costs and margins are squeezed. Weaker companies, now highly leveraged, begin to struggle. Default rates start creeping up and credit spreads rise further. This process can become self-reinforcing. Higher interest rates and wider credit spreads make it harder for companies to refinance when their existing bonds mature, increasing default risks. Normally at this point spreads widen and there is a surge in defaults as the cycle comes to an end, possibly culminating in a recession and preparing the ground for a new cycle.
“The end-of-cycle spike in yields can have a dramatic effect on investment returns.” The end-of-cycle spike in yields can have a dramatic effect on investment returns. Recall that the capital return on corporate bonds is a function of the change in yield and the default loss. Putting some indicative numbers to this, high-yield credit spreads could widen from 3% to 10% or more over the course of the cycle. Default rates might also rise from 3% to 10%. This translates roughly into a capital loss from default of 5% (10% default rate x 50% recovery rate) and a capital loss from widening spreads of 28% (duration 4 years x 7% change in yield). This would result in a 33% loss by the end of the cycle, only partially offset by higher income from the 7% wider spreads. The good news for high-yield credit investors is that as spreads fall back to normal levels after the economy picks up and lenders become more comfortable taking credit risk, they enjoy a symmetrical positive capital return. Buy-and-hold investors suffer
permanent losses from the higher default rate, but this is offset by higher income. In the end long-term investors may not be permanently affected, but nervous investors who sell when spreads are wide suffer most. Our view on corporate bond returns Over the long term (a 10-year horizon) we assume credit spreads revert to their long-term average level. This can have important implications for returns. If spreads at the time we make our forecast are unusually high (near the end of the credit cycle or when markets are nervous) this results in a high return forecast because we assume that over time spreads will return to normal levels, significantly boosting long-term returns. Over the medium term (a three-year horizon) our view of the stage the credit cycle has reached is central to our approach to forecasting credit returns. This view is heavily influenced by our view of recession risk. Typically a recession coincides with the end of the credit cycle, as the two cycles are causally linked in various ways. For example, a recession results in a collapse in corporate earnings, which often forces weaker companies into default. It is highly unlikely that a recession would occur without triggering a wave of defaults, and vice versa.
“Our medium-term recession risk indicator gives us a rough guide to the position in the credit cycle.” Our medium-term recession risk indicator gives us a rough guide to the position in the credit cycle. When it suggests recession risks are very low and the business cycle has several more years to run, we assume credit spreads will be below their long-term average. At mid cycle, when recession risks are at their average level, we assume spreads will be around their long-term average. Then as we get to late cycle we assume a rising path for spreads and defaults. Recession risk is not the whole story. For example, in 2016 the US energy sector underwent a major default cycle following the collapse of oil prices over 2014 and 2015. This resulted in very wide spreads on bonds from oil-related companies and a high default rate. Credit spreads increased elsewhere over 2015 and early 20163 due to fears that credit problems might emerge in other sectors. In the end, this turned out not to be the case and credit spreads fell sharply during 2016 back towards average levels (and providing investors with a bumper capital return). We consider a range of factors in coming to our credit views: average leverage levels, interest servicing cover, credit quality; as well as market factors such as the ‘hunt for yield’. US credit outlook Our recession risk indicator suggests that the US credit cycle has several more years to run (see discussion in Chapter 4). This simple perspective suggests a fairly neutral view on credit spreads.
Source: Bloomberg, Bank of America Merrill Lynch.
3
aberdeen-asset.co.uk
71
Credit continued However our view is complicated by the arrival of the Trump administration and the likelihood of major changes to US policy. On the plus side, deregulation and corporate tax cuts should boost after-tax corporate cash flows, thereby increasing interest cover and increasing corporate credit quality. On the negative side, congressional Republicans have been discussing the possibility of a major change to the tax system, removing the tax deductibility of interest payments. There is an argument that the tax system currently discriminates against equity in favour of debt. Dividend payments to equity investors are not tax-deductible, unlike interest. If this change were introduced it would significantly increase the post-tax interest costs for highly indebted companies in the near term, although it is not clear that overall credit quality would be impaired over the long run as companies alter the balance between equity and debt funding. Another possible headwind to returns is a rise in underlying government bond yields as the Federal Reserve raises policy interest rates. This is a particular issue for investment-grade credit, which has a higher duration (and so a higher capital loss from rising yields), although for some investors it is possible to hedge this duration risk. Higher rates also create a refinancing risk: as bonds that were issued when rates were low expire, new ones must be issued at higher rates. Defaults are elevated but likely to fall back in the near term While the credit outlook is improving thanks to proposed policy changes, the credit cycle is fairly mature and defaults are running above average levels. There are a substantial number of distressed companies, as measured by the number of CCC issuers, without access to credit markets. Default rates across several sectors, especially telecoms, consumer products, metals and chemicals are likely to remain elevated. It is not just energy companies facing difficulties, so we don’t expect a return to cycle lows for defaults or spreads.
“Low spreads provide a relatively low reward for bearing credit risk.” On the positive side the extent of ratings downgrades of high quality companies is modest. The rate at which investmentgrade companies are being downgraded to junk status is around 1.8% per year, below the long-run average of 2.1%. This could suggest that the end of the credit cycle is still some way off. Nevertheless, low spreads provide a relatively low reward for bearing credit risk. European credit outlook The credit cycle in Europe is in some ways at an earlier stage than in the US. European economies still have significant output gaps and interest rates are likely to remain below zero for some years. Many companies, especially financials, are still deleveraging their balance sheets, and at an aggregate level credit is yet to start rising faster than nominal GDP. This suggests that the European credit cycle has longer to run.
72
Credit
Fig. 7.2: Credit growth to private non-financial sector provided by domestic banks (%) 10 8 6 4 2 0 -2 -4 -6
2011 UK
2012 US
2013 Euro Area
2014
2015
2016
Japan
Source: Aberdeen, Thomson Reuters, Datastream, April 2017. Note: Credit to private non-financial sector, provided by domestic banks, unadjusted for breaks, market value, in local currency. Calculated as YoY % change.
However, even if the stage of the credit cycle in Europe would suggest a more positive view than for the US, there is usually contagion in credit markets. If the US credit cycle were to come to an end, we would likely see European spreads widen in sympathy.
Corporate loans In addition to issuing bonds, companies also borrow money from the capital markets via senior secured loans. These credit instruments have somewhat different characteristics from typical corporate bonds so they must be considered separately. The biggest market for loans is the US, but there are growing numbers of issuers in Europe and emerging markets. Loans tend to be issued by companies with lower credit ratings, more similar to the type of companies that issue high-yield bonds.
“Loan holders have a claim on corporate assets. This means we would expect the risk of loss to be lower.” However, these loans are ‘senior’ in the sense that they are first in the queue for repayment in the event a company enters bankruptcy. Bonds are further down the pecking order, while equity is at the bottom. They are also ‘secured’ in the sense that loan holders have a claim on corporate assets. This means that the risk of loss is lower than for comparable unsecured bonds, given a higher recovery rate.
Another key distinguishing feature of loans is that the coupon they pay is ‘floating-rate’. Coupons are typically linked to LIBOR, the standard money-market measure of short-term rates. As interest rates rise the coupon rises, meaning that loans offer protection against rising rates. By contrast, ordinary bonds experience a capital loss in proportion to their duration as interest rates rise. So in a rising US interest rate environment, loans are particularly attractive for investors. To forecast returns on loans we take a similar approach to highyield bonds, forecasting capital and income returns. Our income forecast is similar in that it is based on our assumptions about underlying yields. Changes in yield are a function of our forecast for the path of short-term interest rates to which loan coupons are linked. If a loan pays 4% over short term interest rates and we think these will be 2% in two years’ time, then the yield will be roughly 6%. Calculating the capital return is more complicated. The fact that the coupons on loans float with LIBOR rates means that they have little interest rate sensitivity – their ‘duration’ is near zero. However, in a typical loan index the size of the gap between coupons and LIBOR (the ‘spread’) does change from time to time as concerns about credit risk change. Capital returns are sensitive to this, resulting in capital gains and losses. Outlook for loans We expect LIBOR rates to rise steadily in the next three years as continued growth with low unemployment creates inflation pressure and a response from the Fed. This will increase the yield on loan portfolios.
“Spreads are nearing their tightest levels, so only a little more capital return can be expected.” However, spreads are nearing their tightest levels, so only a little more capital return can be expected. The return is therefore primarily a function of our income yield expectations, adjusted downward by 80-90bps to reflect the modest level of default losses we expect. This leaves us with a fairly positive view on loans. In a time of rising policy interest rates, floating rate credit is particularly attractive. Fig. 7.3: Global credit returns
3Y
5Y
10Y
UK IG Bonds
2.8
2.4
2.8
US IG Bonds
3.4
3.4
3.8
Euro IG Bonds
1.7
1.8
2.3
US High-Yield Bonds
4.6
4.8
5.3
Europe High-Yield Bonds
2.5
2.4
3.0
EMD (Hard)
3.5
4.9
5.5
EMD (Local)
5.6
6.3
6.5
Senior Secured Loans
5.8
5.4
5.5
Source: Aberdeen, March 2017. Note: Returns are in local currency and in percentage, per annum. Projections are estimates and provide no guarantee of future returns.
In 2015 and 2016, loan prices fell sharply (matching falls in prices of HY bonds). This reflected concerns about defaults in the energy sector as well as worries about a deteriorating credit environment more generally. However, with a recovery in the oil price, improved credit quality in the energy sector, and an improved credit outlook more generally, spreads tightened sharply, resulting in strong returns for loans in the last 12 months.
aberdeen-asset.co.uk
73
Chapter
08 74 document_name
Emerging-market sovereign debt
• Emerging-market government debt is less risky than in the past • Yields are high relative to developed-market bonds, offering higher returns • EM currencies are no longer expensive, so there is less structural currency risk for long-term investors.
aberdeen-asset.co.uk75
Emerging-market sovereign debt Emerging-market (EM) sovereign debt has grown rapidly as an asset class in the last decade. With developedmarket government bond yields persistently low, the much higher yields on offer from emerging-market governments make this asset class very attractive. Approach There are two categories of EM sovereign debt: ‘hard currency’ debt issued in foreign currency (dollars, or less frequently euros); and ‘local currency’ debt issued in the currency of the issuer (e.g. Korean won). This distinction means somewhat different approaches are required for these two categories of debt. Hard currency Hard-currency EM debt is modelled in a very similar way to corporate credit. We start with our forecasts for yields on government bonds in the currency concerned. For EM bonds issued in US dollars this would be US Treasury yields. Then, as with corporate credit, we form a view of the evolution of the spread of EM debt yields over the government yield. On a 10 year horizon we typically assume spreads will revert to their long-term median. However, in the short term we consider whether spreads are likely to deviate temporarily from this reversion trend. For example, if we see significant probability of downside risks to the market, we may assume spreads will widen before reverting.
“Hard-currency EM debt is issued in dollars but the sovereign issuers must repay the debt from tax revenues raised in local currencies.” One of the main downside risks comes from possible currency depreciations. Hard-currency EM debt is issued in dollars but the sovereign issuers must repay the debt from tax revenues raised in local currencies. Strong appreciation of the dollar against the local currency increases the risk of default. If we foresee upward movement in the dollar against emerging-market currencies, we may adjust the spread upward in the short term. Returns are calculated in the same way as for other bonds, based on income from the yield and capital returns from yield movements (multiplied by duration). We also make a small adjustment for default risk – though the risk is low. Our modelling is based on the most popular index for this asset class, which incorporates a wide range of emerging-market sovereign issuers. Local currency Local-currency debt is not modelled as a spread to US government bonds, given the bonds are issued in multiple currencies and prices are not closely linked to movements in the US government yield curve. In other respects our approach is similar. We assume the yield of the local-currency EM government bond index will revert to its long-term median. As with hard-currency bonds, we consider the probability that in the shorter term the path of yields will deviate upwards or downwards in response to risks. Local-currency debt may be attractive because yields tend to be higher than on hard-currency debt. The principal reasons for this are higher real equilibrium interest rates and higher inflation expectations in EM countries than in the US. These largely reflect faster (catch-up) growth and less credible inflation targets for emerging-market central banks, and in combination more than offset the spread on hard-currency EM bonds above US yields. 76
Emerging-market sovereign debt
“While yields are high, local-currency bonds also expose investors directly to currency risk.” However, while yields are high, local-currency bonds also expose investors directly to currency risk. This has a major impact on returns. The chart below (Fig. 8.1) shows returns of the localcurrency EM index in dollars and in local currency (equivalent to assuming no changes in spot exchange rates). As you can see, the local-currency return is fairly stable, but the dollar return is volatile. This difference reflects the volatility of EM currencies against the dollar. Fig. 8.1: EM debt total returns (%) 50
40 30 20 10 0 -10 -20 -30 2003
2005
2007
2009
Hard-currency debt in USD Local-currency debt in USD
2011
2013
2015
2017
Local-currency debt in local currency
Source: Aberdeen, Bloomberg, J.P. Morgan indices, as at 8 March 2017. Note: Calculated as 12 month rolling total return. Past performance provide no guarantee of future results.
When returns from foreign assets can be dominated by currency risk, investors may consider hedging. For hard-currency EM debt issued in dollars it is straightforward to hedge the dollar exposure. However, local-currency EM debt is issued in a dozen or more individual currencies, such as the Brazilian real, Mexican peso and Indian rupee. It is practically difficult to hedge some of these currencies, and in many cases also expensive. In hedging the currency exposure, investors must give up much of the high yield on offer from local-currency debt, so investors tend accept the risk. The outlook for EM currencies is therefore very important for our overall view on EM local-currency bond returns. It is worth noting that it might not generally be fair to expect returns as high as the yields on local-currency bonds would suggest. Over long time horizons, high EM inflation would be expected to result in nominal currency depreciation, directly reducing returns. Investors should only hope to capture the differential in real (inflation-adjusted) interest rates, and must face the currency risk.
“Investors should only hope to capture the differential in real (inflation-adjusted) interest rates, and must face the currency risk.”
Outlook for EM sovereign debt returns Investors with long memories will remember the Latin American sovereign debt crisis of the 1980s or the Asian crisis of the late 1990s. During these periods investors faced the risk of sovereign debt default, and dramatic currency depreciations. However, since then many lessons have been learned. In most emerging-market countries, government debt levels today are much lower than they were in the past. This is a very positive development for EM debt as an asset class. Indeed, in many cases they are lower as a share of GDP than debt levels in many developed economies, although this comparison is not always instructive.
“A greater proportion of EM debt is now issued in local currencies and central banks have built up much larger foreign currency reserves.”
That said, while some EM governments now have relatively low debt levels, some large EM companies have built up high debt burdens. Where these companies are state owned, or considered ‘too big to fail’ and likely to be bailed out by their governments, they represent a contingent liability for sovereign issuers. The vulnerability to ‘sudden stops’ – where capital inflows dry up rapidly – as a result of currency mismatches – used to be a major problem, but is less of an issue today. During the Asian crisis, countries had issued large amounts of debt in dollars but found that their central banks did not have sufficient foreign exchange reserves to repay them in the event of local currency depreciation. The risk of this is now much lower. A greater proportion of EM debt is now issued in local currencies and central banks have built up much larger foreign currency reserves. So the risk of default on hard currency debt is lower.
Fig. 8.3: Forecast yield, EM debt (local currency) (%)
Fig. 8.2: Forecast spread over US Treasuries, EM debt (hard currency) (%)
900
1,300 1,100
800
900 700
700 500
600 300 100 1996
2000
2004
2008
2012
2016
2020
2024
500 2002
2005
2008
2011
2014
2017
2020
2023
2026
Source: Bloomberg, Aberdeen, as at March 2017. Past performance is not necessarily a reliable indicator of future results. Note: Black dashed lines indicate historic median. Projections are estimates and provide no guarantee of future results.
Risks EM economies are growing at a solid but unspectacular pace, boosted recently by stronger performance in China, Brazil and Russia. This boosts investor optimism about the asset class. However, there is a risk that China’s slowdown will accelerate in the next few years, creating stress for some of the more Chinaexposed EM economies and governments. This risk is reflected in our forecasts in a slightly higher path for spreads (EM hard currency debt) and yields (EM local-currency debt). Trump administration economic policy is also a potential source of risk, both because of possible protectionist measures and because his trade, immigration and fiscal policies may result in a stronger dollar and corresponding EM currency depreciation. Returns To reflect this somewhat negative balance of risks, we assume EM sovereign debt yields will rise a little in the next few years, before falling toward the long-term median. This modest capital loss will be largely offset by the relatively high income that this asset class provides. Yields are around 3% for hard currency EM debt and 6% for local-currency EM debt. If the China and Trump risks we describe do fully materialise, returns will be lower than we forecast. But on our central case they will be a little higher. In the meantime, the high income yield provides compensation for bearing the risk. Overview on EM debt over the long term remains fairly positive.
Currency risk Investors in EM local currency sovereign debt must also take a view on currency risk, given the difficulties in hedging. At the moment, long-term equilibrium exchange rate models tend to show that on average emerging-market currencies are currently trading near fair value on a trade weighted basis. This means that over the long term there is no strong reason to expect either a negative or a positive currency return for EM local bonds. However, in the short term the risk of a stronger dollar or of a sharp China slowdown and possible renminbi devaluation mean there is a significant risk of volatility. But even taking currency risk into account, the volatility of returns for EM local debt is only around 12% on annualised basis. This is rather lower than developed equities (at 18%), and yet the returns on offer are around the same, or perhaps slightly higher. So on a risk-adjusted basis this asset class is attractive. Fig. 8.4: EM sovereign debt returns 3Y 5Y
10Y
3.5 4.9
5.5
EM Sovereign Debt (Local) EM currency basket 5.6 6.3
6.5
Forecast currency
EM Sovereign Debt (Hard) USD
Source: Aberdeen, March 2017. Note: Returns are in local currency and in percentage, per annum. Projections are estimates and provide no guarantee of future returns.
aberdeen-asset.co.uk
77
Chapter
09 78 document_name
Listed alternatives and property
• Investors are increasingly looking to alternative sources of diversification from equities • Liquid alternatives include infrastructure, insurance linked securities, and litigation finance • Correlations with equities are low, but returns may be much higher than bonds
aberdeen-asset.co.uk79
Listed alternatives and property A central goal of asset allocation is to build portfolios with diversified sources of return. Bonds have historically provided the primary source of diversification to equities. But, as we explain in Chapter 1, their diversification benefit is offset by their extremely low expected returns.
“Alternatives, such as infrastructure, property and insurance-linked securities can also provide diversification from equities.” This makes alternative assets much more important than they have been in the past. Alternatives, such as infrastructure, property and insurance-linked securities can also provide diversification from equities, but with returns that may be significantly higher than those we currently expect for government bonds. The fact that, after an extended bull market, most equity markets are now on the expensive side, makes the need for diversified returns even more acute. In addition, new alternatives strategies, and different ways of accessing more mature ones, have led investors to adjust their asset allocation. Alternatives is a very big category – including everything from private equity and private debt; infrastructure, property and other real assets; litigation finance, aircraft leasing and trade finance; to the many hedge fund strategies. Our clients invest in all of these asset classes and more besides. This chapter does not aim to be comprehensive, but instead will focus on the small selection from this sector that has, in the last two years, secured the biggest allocation in our main multi-asset portfolios. Our selection is defined by two main features: liquidity and diversification from equities. Liquidity can be a challenge for alternatives. Most alternative assets tend to be relatively illiquid, for example, infrastructure funds invest in physical social or environmental infrastructure assets – schools, roads, wind farms – that are not easily traded. However, investors can now access many alternative assets via permanent capital structures such as investment trusts. The underlying assets may remain illiquid, but the trusts are themselves traded on a daily basis.
“The listed-alternatives sector has expanded rapidly.” In the past the relatively small size of the listed alternatives sector has limited its capacity to absorb investor capital. However, as the chart (Fig. 9.1) shows, the listed alternatives sector has expanded rapidly. This means that it is now an option for all but the largest investors.
Fig. 9.1: Rapid growth in the UK listed alternatives market (market cap £m) 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Infrastructure Hedge Funds
Debt
Specialist
Private Equity
Property
Source: Numis, March 2017.
We also include property in this chapter. This asset class can be accessed by real estate investment trusts but also via open-ended property vehicles which square the liquidity circle by other means. While our current focus is on liquid alternatives, it is worth mentioning that attitudes towards illiquid assets are changing. With lower expected returns from conventional asset classes, clients are increasingly interested in the potential of private, unlisted assets to boost returns via an illiquidity premium. To meet this need, Aberdeen has been rapidly expanding its capability in this area. Future versions of this report will include our views on these assets.
“Our portfolios use alternatives because of their combination of low correlation to equities and higher returns than government bonds.” In addition to liquidity, we narrow our focus in this chapter on alternatives that provide diversification from equities. Many of our portfolios use alternatives because of their combination of low correlation to equities and higher returns than government bonds.
“For many alternative assets, cash flows have other sources that are not sensitive to economic shocks.” The reason why many of these alternative asset classes have low correlations to equities is related to the source of their cash flows. Equity returns are sensitive to market expectations about future economic growth because this is where corporate dividend growth ultimately comes from. But for many alternative assets, cash flows have other sources that are not sensitive to economic shocks. For example, social infrastructure investment in schools, hospitals and roads has cash flows that are normally related to long-term, government-backed contracts, which typically do not change materially with the economic growth outlook. Insurance-linked securities pay investors an insurance premium in return for investors taking some exposure to the risk of natural catastrophes, such as the risk of hurricane. Economic recessions are not correlated with severe hurricanes, so the returns of these assets are uncorrelated.
80
Listed alternatives and property
Not all listed alternatives offer this kind of diversification benefit. For example, the cash flows for liquid private equity have the same source as public equities and so the correlation with equity returns is high. We therefore exclude this asset class from the current chapter.
Our assumption about changing yields has two purposes in our model: it affects our assumption about market prices and so our capital return assumptions, while also providing us with our income return assumption.
Commercial Property
Outlook
Fig. 9.2: Global commercial property returns
The rental growth picture is fairly positive in most markets at present. The exception is the UK, where the picture is clouded by fears about the impact of Brexit on the UK economy and on the London office market in particular.
3Y
5Y
10Y
UK
3.0
3.8
5.9
US
5.1
5.3
5.8
Europe ex UK
2.7
4.3
5.9
Japan
4.4
4.1
3.2
Australia
7.9
8.3
7.0
N/A
N/A
N/A
Global
Source: Aberdeen, March 2017. Note: Returns are expressed over 3, 5, and 10-year time horizons in local currency and in percentage, per annum. Projections are estimates and provide no guarantee of future returns.
Approach Commercial property returns have two main components, income and capital growth. Income comes from the rents received, net of costs. Capital growth arises from growth in rental incomes and re-pricing following periods of over- or under-valuation. This in turn is a function of the supply and demand for buildings and investor sentiment. Our rental income growth expectations are developed by Aberdeen’s global property investment team. Their views are based on analysis of long-term historical trends for each property sector in each regional economy we cover (US, UK, Europe ex UK, Japan and Australia). This is based on assumptions about how trends might change in the future. For example, we look at projections on the extent to which internet retailers may displace physical retailers, resulting in increasing demand for distribution warehouses and lower demand for retail stores. The team also considers the evolution of supply and demand conditions in each market and the state of the business cycle.
The supply/demand picture suggests a low risk of oversupply in global property markets. In the US and UK a substantial amount of construction is underway, but less so than the pre-crisis property boom. In the US, a fair amount of speculative construction is concentrated in logistics, but this is still falling behind demand.
“The spread between property yields and underlying government bond yields remains well above average, except for the US market. On this basis, property is not particularly expensive.” In an environment of low government bond yields, there is strong demand from investors for alternative asset classes with higher yields. Strong demand has driven property yields lower than in the past (see regional yield charts Fig. 9.3 and Fig 9.4). However, the spread between property yields and underlying government bond yields remains well above average, except for the US market. On this basis, property is not particularly expensive. Government bond yields around the world have started to rise from their historic lows, so we expect some upward pressure on yields. However, we do not expect bond yields to return to the much higher levels seen in the decade before the financial crisis, so this factor is not as strong a concern as it might be. Fig. 9.3: UK commercial property yields and spreads (%) 8
Our long-term view on the direction of market valuations is based primarily on assumptions about the property yield. Our forecasts are based on yields for regional variants of the IPD index. Our forecast for yields has two components: assumptions about the future path of government bond yields in each market and assumptions about the spread between the property yield and the government bond yield. This is a measure of the risk premium that property investors receive.
7
We assume that this property risk premium spread will revert to its 20-year mean over the next 10 years. We do not always assume that the mean reversion in property risk premia is linear. In the short-term there may be times when the risk premia might be higher or lower than straightforward mean-reversion suggests. For example, in the UK the risk of a hard Brexit resulting in a weaker economy and partial loss of access to European markets for financial and office-based service sectors means that we assume the risk premium for UK commercial property will remain elevated for the next three years.
0
6 5 4 3 2 1 -1
2001
2005
IPD yields
2009
2013
10Y Gilt yields
2017
2021
2025
Spread
Source: Aberdeen, MSCI Real Estate Analytics, Bloomberg, March 2017. Note: Yield and spread projections are estimates and provide no guarantee of future results.
aberdeen-asset.co.uk
81
Listed alternatives and property continued The fairly large spread between property yields and government bond yields provides a cushion, allowing property yields to remain stable. The one obvious exception is the US, where we expect Treasury yields to rise fastest and where spreads are low (Fig. 9.4). Fig. 9.4: US commercial property yields and spreads (%) 9 8
The UK listed-infrastructure market is relatively concentrated with 13 listed investment companies split across three sub-asset classes: social infrastructure, renewable infrastructure and infrastructure debt. Social-infrastructure investment companies were the first to come to market just over a decade ago, while renewable investment companies are on average under four years old.
“Social-infrastructure returns are insensitive to the changing fortunes of the global economy.”
7 6 5 4 3 2 1 0 2001
2005 IPD yields
2009
2013
10Y Treasury yields
2017
2021
2025
Spread
Source: Aberdeen, MSCI Real Estate Analytics, Bloomberg, March 2017. Note: Yield and spread projections are estimates and provide no guarantee of future results.
The UK market also faces risks. The property risk premium is already high, but if Brexit fears mount, we expect it to move higher, resulting in capital loss (as indicated by our total return forecast below, Fig. 9.5).
As with direct infrastructure, social-infrastructure funds generate returns from long-term contracts to build and operate schools, hospitals, roads and other public infrastructure assets. Their returns come from long-term, government guaranteed public private partnership (PPP) and public finance initiative (PFI) contracts. This means they are insensitive to the changing fortunes of the global economy and so demonstrate low correlations to equity markets. Infrastructure funds also offer a degree of inflation protection given that contracts are typically inflation linked. There is little growth in the cash flows from PPP/PFI contracts, so the vast majority of return is in the form of income yield rather than capital gain. Fig 9.6: Listed social infrastructure during the financial crisis 155 140
Fig. 9.5: UK IPD total return index
125
2,400
110
2,000
95 1,600
80
1,200
65
800
1995
2000
2005
2010
2015
2020
2025
Source: Aberdeen, MSCI Real Estate Analytics, March 2017. Note: Index uses December 1986 as 100 base. Returns are estimates and provide no guarantee of future results.
Elsewhere, the combination of rental income growth and fairly stable yields makes for potentially higher property returns in most markets, particularly in Europe ex UK, which is our preferred market.
Listed infrastructure Traditionally, infrastructure has been the preserve of institutional investors prepared to lock away their capital for several years at a time. But over the last decade, a new breed of listed infrastructure funds has emerged, particularly in the UK, where these funds are now worth £13 billion1. The UK is our preferred source of listed infrastructure assets, given the stable regulatory environment, the strong governance of funds and the liquid nature of the market. Numis, as at 17 February 2017.
1
82
2007
2008
HICL Infrastructure co. ltd.
400 0
2006
Listed alternatives and property
2009
2010
2011
FTSE 250 Index
Source: Bloomberg, as at 21 February 2017. Notes: Total return (using gross dividends), rebased to 100 at 31 March 2006. The chart shows that while equity markets experienced a significant draw down during the financial crisis and again in the Euro-crisis in 2011, social infrastructure was largely unaffected. Past performance provides no guarantee of future results.
Renewable infrastructure funds hold portfolios of wind and solar energy generation facilities. As with social infrastructure, the cash flows from these assets relate to long-term, governmentbacked cash flows, this time in the form of renewable energy subsidies. As a result they offer a similar return profile to social infrastructure. Again, the vast bulk of the return is in the form of income – and income yields tend to be a little higher than for social infrastructure. This reflects the fact that the nature of renewable energy subsidy policies, electricity markets and even weather patterns, means that there are different risks. These funds are structured as ‘closed-ended’ investment trusts. This offers investors liquidity (they are traded on the London Stock Exchange) but the underlying assets are illiquid. To make this possible, the fund’s price floats independently of the fundamental value of its assets; its ‘net asset value’ (NAV). When investors are particularly enthusiastic about the return prospects of a given trust, its price will trade at a premium to NAV. When they are unenthusiastic, it will trade at a discount. This factor has an important part to play in the returns provided by these funds.
Fig. 9.7: Evolution of listed social infrastructure share premium (%)
Fig. 9.9: Catastrophe bond returns since 2001
350
30
Hurricane Irene, Tornadoes in the US, Tohoku Earthquake
300
20
250
10
200 0
Hurricane Katrina, Rita, Wilma
Hurricane Ike
150 -10 -20 2008
100 2009
2010
Premium to NAV
2011
2012
2013
Discount to NAV
2014
2015
2016
2017
Average
50 2002
Financial crisis 2004
2006
2008
Swiss Re Cat Bond Index
Source: Aberdeen, Numis, as at 21 April 2017. Notes: Difference between closing share price and NAV (Net Asset Value) for HICL Infrastructure Co. Ltd.
Our returns forecasts for social and renewable infrastructure are built from capital growth and income components. Reported weighted average portfolio discount rates provide an anchor for our target returns over the short and medium term. We make adjustments for fund management expenses and material sensitivities. This includes harnessing our projections for inflation and government bond yields over the period. We also account for reversion to nil premium/discount to NAV over a 10-year period. For both social and renewable energy infrastructure, income is by far the biggest component of return – 4.5% for social infrastructure and around 5.5% for renewables. There is also a modest capital return available on this asset class, but as (Fig. 9.7) indicates, premiums to NAV are somewhat elevated relative to their history. We assume premiums revert to zero over 10 years, which reduces the capital return we assume to a little above zero. Fig. 9.8: Listed alternative returns 3Y
5Y
10Y
UK listed social infrastructure
5.3
5.2
5.2
UK listed renewables infrastructure
5.9
5.8
5.8
Insurance linked securities
3.9
4.3
5.1
Source: Aberdeen, March 2017. Note: Returns are in local currency and in percentage, per annum. Returns are estimates and provide no guarantee of future results
Insurance-linked securities The insurance-linked securities market is a way of transferring the risk of insurance losses due to extreme natural catastrophes from insurers to capital market investors. In return, investors receive a stream of insurance risk premium. This market has grown markedly over the past 10 years, from $20bn in 2008 to $75bn in 2016. The market includes a diverse range of underlying risks including Florida hurricanes, Californian earthquakes, European windstorms, Australian floods, and Japanese typhoons. The most accessible and liquid route for accessing this market is through catastrophe bonds. Catastrophe bonds pay investors a regular premium for bearing a share of the risk of loss associated with a specific event – such as US property damage associated with a severe hurricane. The value of a catastrophe bond falls if risk event occurs and the insurer’s losses exceed a certain threshold trigger. Triggers can be structured to adjust the magnitude of actual losses borne by the issuer, thereby finding the most appropriate balance between reinsurance appeal and investor return.
2010
2012
2014
2016
MSCI World Index
Source: Bloomberg, Swiss Re as at 31 December, 2016. Note: Rebased to 100 at 31 January, 2002. Past performance provides no guarantee of future results.
The chart (Fig. 9.9) shows that negative returns for catastrophe bonds are correlated with natural catastrophes, not equity market events. A major attraction of catastrophe bonds and other insurance-linked securities (ILS) is that their returns are uncorrelated with equities. This is because equity returns are fundamentally driven by the behaviour of the economy’s growth. The worst equity losses occur during downturns in the business cycle. The risks for insurancelinked assets, on the other hand, are associated with natural disasters and other events that are unrelated to the business cycle. Hurricanes, for example, are not correlated to recessions.
“Hurricanes, for example, are not correlated to recessions.” Typically investors hold a diversified portfolio of different kinds of insured risks to ensure that any losses are spread. Our projections for ILS are informed through Aon Benfield Analytics’s reinsurance pricing, which provides market metrics for risk-linked securities. We aggregate the data in order to obtain a realistic average yield estimate, while accounting for estimates of expected losses. Catastrophe bonds tend to offer investors the highest returns in the period immediately following a major loss: premiums are usually highest at this point. At this point investors are risk averse and yields are high. The lowest returns are experienced during the loss event. Catastrophe-bond yields are relatively low at present, this is partly as a result of substantial investor interest in this asset category and the fact that insured losses have not been particularly bad in recent years. In the light of these low returns we prefer more specialist insurance-linked funds which write direct collateralised reinsurance contracts with insurers. These can offer significantly higher levels of expected return than the catastrophe bond market. These higher returns result from a combination of exposure to high insurance loss risks, added value from manager skill, and an illiquidity premium. Expected net returns on diversified collateralised reinsurance funds can be 4-5% or more higher than those of the catastrophe bond market for similar levels of portfolio risk.
aberdeen-asset.co.uk
83
Appendix: Expected returns by base currency Risk and return for UK investors Local
GBP Hedged
Local Currency
3Y
5Y
10Y
3Y
5Y
10Y
3Y
5Y
10Y
UK Equities
GBP
5.5
5.5
6.1
5.5
5.5
6.1
5.5
5.5
6.1
17.2
0.28
US Equities
USD
7.2
5.5
5.3
8.7
4.6
3.5
5.9
4.2
4.1
16.8
0.21
Europe ex UK Equities
EUR
6.2
5.5
5.2
6.9
5.7
5.0
7.0
6.2
6.0
20.4
0.27
Japan Equities
JPY
4.6
4.7
4.6
5.0
5.1
5.1
5.1
5.4
5.9
22.1
0.22
Pacific ex Japan Equities
AUD
5.6
6.3
6.6
4.4
4.7
5.0
3.8
4.5
5.1
17.7
0.22
Emerging Markets Equities
Various
5.9
6.1
6.3
4.8
4.8
5.0
n/a
n/a
n/a
n/a
n/a
3Y 5Y Volatility Sharpe Ratio
Global Equities
Various
n/a
n/a
n/a
7.4
4.9
4.3
5.8
4.7
4.8
16.7
0.25
UK Gilts (All Maturity)
GBP
1.5
0.9
1.1
1.5
0.9
1.1
1.5
0.9
1.1
5.9
0.05
UK Inflation-Linked Gilts
GBP
0.2
0.5
1.3
0.2
0.5
1.3
0.2
0.5
1.3
10.8
-0.01
US Treasuries (All Maturity)
USD
1.8
1.9
2.3
3.2
1.0
0.6
0.6
0.6
1.2
4.3
-0.01
Euro Govt Bonds (All Maturity)
EUR
0.8
0.6
1.1
1.4
0.8
0.9
1.5
1.3
1.9
4.0
0.17
Japanese Govt Bonds (All Maturity)
JPY
0.1
0.0
-0.1
0.5
0.3
0.4
0.6
0.7
1.2
3.7
0.01
Global DM Govt Bonds
Various
n/a
n/a
n/a
1.8
0.8
0.7
0.9
0.8
1.4
3.3
0.06
UK IG Bonds
GBP
2.8
2.4
2.8
2.8
2.4
2.8
2.8
2.4
2.8
7.5
0.23
US IG Bonds
USD
3.4
3.4
3.8
4.8
2.5
2.0
2.2
2.1
2.6
6.3
0.23
Euro IG Bonds
EUR
1.7
1.8
2.3
2.3
2.0
2.1
2.4
2.5
3.1
4.7
0.40
Global IG Bonds
Various
n/a
n/a
n/a
4.1
2.4
2.1
2.3
2.2
2.7
5.8
0.27
US High Yield Bonds
USD
4.6
4.8
5.3
6.0
3.8
3.5
3.4
3.4
4.0
9.8
0.29
Europe High Yield Bonds
EUR
2.5
2.4
3.0
3.2
2.6
2.8
3.3
3.1
3.8
12.3
0.20
EM Debt (Hard)
USD
3.5
4.9
5.5
4.9
4.0
3.7
2.3
3.5
4.2
9.0
0.32
EM Debt (Local)
Various
5.6
6.3
6.5
4.4
5.5
5.8
n/a
n/a
n/a
n/a
n/a
Senior Secured Loans
USD
5.8
5.4
5.5
7.2
4.4
3.7
4.6
4.0
4.2
8.7
0.39
Insurance Linked Securities
USD
3.9
4.3
5.1
5.3
3.4
3.3
2.6
3.0
3.8
8.3
0.28
UK Commercial Property
GBP
3.0
3.8
5.9
3.0
3.8
5.9
3.0
3.8
5.9
13.8
0.23
US Commercial Property
USD
5.1
5.3
5.8
6.6
4.4
3.9
3.9
4.0
4.5
13.4
0.25
Europe Commercial Property
EUR
2.7
4.3
5.9
3.4
4.4
5.7
3.5
5.0
6.7
13.7
0.32
Global Commercial Property
Various
n/a
n/a
n/a
5.1
4.5
4.8
3.9
4.5
5.4
10.0
0.39
UK REITs
GBP
4.2
4.6
6.0
4.2
4.6
6.0
4.2
4.6
6.0
23.5
0.17
Global Private Equity
USD
11.3
10.8
10.7
12.8
9.8
8.7
10.0
9.4
9.3
14.2
0.62
Global Private Infrastructure
USD
10.2
10.2
10.2
11.7
9.2
8.3
9.0
8.8
8.9
10.1
0.81
Global Private Real Assets
USD
5.9
6.0
6.4
7.3
5.0
4.5
4.6
4.6
5.1
19.0
0.21
UK Infrastructure Social
GBP
5.3
5.2
5.2
5.3
5.2
5.2
5.3
5.2
5.2
11.3
0.40
UK Infrastructure Renewables
GBP
5.9
5.8
5.8
5.9
5.8
5.8
5.9
5.8
5.8
11.3
0.46
Alternative Risk Premia
USD
5.5
6.0
6.8
7.0
5.0
5.0
4.3
4.6
5.5
10.6
0.37
Hedge Funds
USD
3.5
4.0
4.8
4.9
3.0
3.0
2.3
2.6
3.5
7.1
0.28
ABS - Mezzanine
USD
5.0
5.5
6.3
6.4
4.5
4.4
3.8
4.1
5.0
9.0
0.38
Commodity Futures
USD
1.5
2.0
2.7
2.8
1.0
0.9
0.3
0.6
1.5
21.6
0.00
UK Cash 3M LIBOR
GBP
0.5
0.8
1.7
0.5
0.8
1.7
0.5
0.8
1.7
1.1
0.14
USD Cash 3M LIBOR
USD
1.9
2.3
3.1
3.3
1.4
1.3
0.7
1.0
1.9
1.1
0.34
EUR Cash 3M LIBOR
EUR
0.0
0.3
1.1
0.7
0.5
0.9
0.7
1.0
1.8
1.1
0.35
Source: Aberdeen, February, 2017. Note: Volatility and Sharpe ratio refers to GBP Hedged. Return projections are estimates and provide no guarantee of future results. Returns and volatilities are in percent per annum.
84
GBP
Asset
Appendix
Risk and return for USD investors Local
USD
USD Hedged
Asset
Local Currency
3Y
5Y
10Y
3Y
5Y
10Y
3Y
5Y
10Y
3Y 5Y Volatility Sharpe Ratio
UK Equities
GBP
5.5
5.5
6.1
4.0
6.3
7.7
6.7
6.9
7.4
17.4
0.28
US Equities
USD
7.2
5.5
5.3
7.2
5.5
5.3
7.2
5.5
5.3
17.1
0.21
Europe ex UK Equities
EUR
6.2
5.5
5.2
5.5
6.4
6.6
8.2
7.6
7.3
20.7
0.27
Japan Equities
JPY
4.6
4.7
4.6
3.5
5.8
6.6
6.3
6.8
7.2
22.4
0.21
Pacific ex Japan Equities
AUD
5.6
6.3
6.6
3.0
5.6
6.8
5.1
5.9
6.4
18.0
0.22
Emerging Markets Equities
Various
5.9
6.1
6.3
3.3
5.6
6.7
n/a
n/a
n/a
n/a
n/a
Global Equities
Various
n/a
n/a
n/a
5.9
5.8
6.0
7.0
6.1
6.1
16.9
0.25
UK Gilts (All Maturity)
GBP
1.5
0.9
1.1
0.1
1.7
2.6
2.7
2.3
2.5
6.0
0.06
UK Inflation-Linked Gilts
GBP
0.2
0.5
1.3
-1.1
1.2
2.6
1.4
1.9
2.6
11.0
-0.01
US Treasuries (All Maturity)
USD
1.8
1.9
2.3
1.8
1.9
2.3
1.8
1.9
2.3
4.5
-0.02
Euro Govt Bonds (All Maturity)
EUR
0.8
0.6
1.1
0.1
1.5
2.4
2.7
2.7
3.1
4.2
0.16
Japanese Govt Bonds (All Maturity)
JPY
0.1
0.0
-0.1
-1.0
1.1
1.8
1.8
2.0
2.4
3.9
0.01
Global DM Govt Bonds
Various
n/a
n/a
n/a
0.5
1.6
2.3
2.1
2.1
2.6
3.5
0.05
UK IG Bonds
GBP
2.8
2.4
2.8
1.4
3.2
4.3
4.0
3.8
4.1
7.8
0.23
US IG Bonds
USD
3.4
3.4
3.8
3.4
3.4
3.8
3.4
3.4
3.8
6.4
0.22
Euro IG Bonds
EUR
1.7
1.8
2.3
1.0
2.7
3.7
3.6
3.9
4.4
4.9
0.40
Global IG Bonds
Various
n/a
n/a
n/a
2.7
3.3
3.8
3.5
3.5
3.9
5.9
0.26
US High Yield Bonds
USD
4.6
4.8
5.3
4.6
4.8
5.3
4.6
4.8
5.3
10.0
0.28
Europe High Yield Bonds
EUR
2.5
2.4
3.0
1.9
3.4
4.3
4.5
4.5
5.1
12.5
0.20
EM Debt (Hard)
USD
3.5
4.9
5.5
3.5
4.9
5.5
3.5
4.9
5.5
9.1
0.32
EM Debt (Local)
Various
5.6
6.3
6.5
3.0
6.3
7.4
n/a
n/a
n/a
n/a
n/a
Senior Secured Loans
USD
5.8
5.4
5.5
5.8
5.4
5.5
5.8
5.4
5.5
8.8
0.39
Insurance Linked Securities
USD
3.9
4.3
5.1
3.9
4.3
5.1
3.9
4.3
5.1
8.5
0.28
UK Commercial Property
GBP
3.0
3.8
5.9
1.6
4.6
7.4
4.2
5.2
7.2
14.0
0.23
US Commercial Property
USD
5.1
5.3
5.8
5.1
5.3
5.8
5.1
5.3
5.8
13.7
0.25
Europe Commercial Property
EUR
2.7
4.3
5.9
2.0
5.2
7.3
4.7
6.4
8.0
14.0
0.32
Global Commercial Property
Various
n/a
n/a
n/a
3.7
5.4
6.5
5.1
5.8
6.6
10.1
0.38
UK REITs
GBP
4.2
4.6
6.0
2.8
5.4
7.5
5.4
6.0
7.3
23.8
0.17
Global Private Equity
USD
11.3
10.8
10.7
11.3
10.8
10.7
11.3
10.8
10.7
14.4
0.61
Global Private Infrastructure
USD
10.2
10.2
10.2
10.2
10.2
10.2
10.2
10.2
10.2
10.2
0.81
Global Private Real Assets
USD
5.9
6.0
6.4
5.9
6.0
6.4
5.9
6.0
6.4
19.3
0.21
UK Infrastructure Social
GBP
5.3
5.2
5.2
3.8
5.9
6.7
6.5
6.5
6.5
11.6
0.40
UK Infrastructure Renewables
GBP
5.9
5.8
5.8
4.5
6.5
7.3
7.1
7.2
7.1
11.5
0.45
Alternative Risk Premia
USD
5.5
6.0
6.8
5.5
6.0
6.8
5.5
6.0
6.8
10.8
0.37
Hedge Funds
USD
3.5
4.0
4.8
3.5
4.0
4.8
3.5
4.0
4.8
7.3
0.28
ABS - Mezzanine
USD
5.0
5.5
6.3
5.0
5.5
6.3
5.0
5.5
6.3
9.2
0.38
Commodity Futures
USD
1.5
2.0
2.7
1.5
2.0
2.7
1.5
2.0
2.7
21.9
0.00
UK Cash 3M LIBOR
GBP
0.5
0.8
1.7
-0.9
1.5
3.1
1.7
2.1
2.9
1.4
0.11
USD Cash 3M LIBOR
USD
1.9
2.3
3.1
1.9
2.3
3.1
1.9
2.3
3.1
1.4
0.25
EUR Cash 3M LIBOR
EUR
0.0
0.3
1.1
-0.7
1.2
2.4
1.9
2.3
3.1
1.4
0.26
Source: Aberdeen, February, 2017. Note: Volatility and Sharpe ratio refers to USD Hedged. Return projections are estimates and provide no guarantee of future results. Returns and volatilities are in percent per annum.
aberdeen-asset.co.uk
85
Appendix: Expected returns by base currency continued Risk and return for EUR investors Local
EUR Hedged
Local Currency
3Y
5Y
10Y
3Y
5Y
10Y
3Y
5Y
10Y
UK Equities
GBP
5.5
5.5
6.1
4.8
5.3
6.3
4.7
4.8
5.3
17.0
0.29
US Equities
USD
7.2
5.5
5.3
8.0
4.5
3.7
5.2
3.5
3.3
16.7
0.21
Europe ex UK Equities
EUR
6.2
5.5
5.2
6.2
5.5
5.2
6.2
5.5
5.2
20.2
0.27
Japan Equities
JPY
4.6
4.7
4.6
4.2
5.0
5.3
4.4
4.7
5.1
21.9
0.22
Pacific ex Japan Equities
AUD
5.6
6.3
6.6
3.7
4.5
5.2
3.1
3.8
4.3
17.6
0.22
Emerging Markets Equities
Various
5.9
6.1
6.3
4.1
4.6
5.2
n/a
n/a
n/a
n/a
n/a
3Y 5Y Volatility Sharpe Ratio
Global Equities
Various
n/a
n/a
n/a
6.7
4.8
4.5
5.0
4.0
4.0
16.5
0.25
UK Gilts (All Maturity)
GBP
1.5
0.9
1.1
0.9
0.8
1.3
0.8
0.3
0.5
5.8
0.06
UK Inflation-Linked Gilts
GBP
0.2
0.5
1.3
-0.4
0.3
1.3
-0.5
-0.1
0.6
10.7
-0.01
US Treasuries (All Maturity)
USD
1.8
1.9
2.3
2.5
0.9
0.8
-0.1
-0.1
0.4
4.2
-0.01
Euro Govt Bonds (All Maturity)
EUR
0.8
0.6
1.1
0.8
0.6
1.1
0.8
0.6
1.1
3.9
0.17
Japanese Govt Bonds (All Maturity)
JPY
0.1
0.0
-0.1
-0.3
0.2
0.6
-0.1
0.0
0.4
3.6
0.02
Global DM Govt Bonds
Various
n/a
n/a
n/a
1.2
0.7
0.9
0.2
0.1
0.6
3.2
0.06
UK IG Bonds
GBP
2.8
2.4
2.8
2.2
2.3
2.9
2.0
1.7
2.0
7.5
0.24
US IG Bonds
USD
3.4
3.4
3.8
4.2
2.4
2.2
1.5
1.4
1.8
6.2
0.23
Euro IG Bonds
EUR
1.7
1.8
2.3
1.7
1.8
2.3
1.7
1.8
2.3
4.5
0.41
Global IG Bonds
Various
n/a
n/a
n/a
3.5
2.3
2.3
1.5
1.5
2.0
5.7
0.28
US High Yield Bonds
USD
4.6
4.8
5.3
5.4
3.7
3.7
2.6
2.7
3.3
9.7
0.29
Europe High Yield Bonds
EUR
2.5
2.4
3.0
2.5
2.4
3.0
2.5
2.4
3.0
12.2
0.20
EM Debt (Hard)
USD
3.5
4.9
5.5
4.3
3.8
3.8
1.6
2.8
3.4
8.9
0.33
EM Debt (Local)
Various
5.6
6.3
6.5
3.7
5.3
6.0
n/a
n/a
n/a
n/a
n/a
Senior Secured Loans
USD
5.8
5.4
5.5
6.6
4.3
3.9
3.8
3.3
3.5
8.5
0.40
Insurance Linked Securities
USD
3.9
4.3
5.1
4.6
3.2
3.4
1.9
2.3
3.0
8.2
0.28
UK Commercial Property
GBP
3.0
3.8
5.9
2.4
3.6
6.0
2.3
3.1
5.1
13.6
0.23
US Commercial Property
USD
5.1
5.3
5.8
5.9
4.3
4.1
3.2
3.3
3.7
13.3
0.25
Europe Commercial Property
EUR
2.7
4.3
5.9
2.7
4.3
5.9
2.7
4.3
5.9
13.6
0.32
Global Commercial Property
Various
n/a
n/a
n/a
4.4
4.4
4.9
3.2
3.8
4.6
9.9
0.39
UK REITs
GBP
4.2
4.6
6.0
3.6
4.5
6.1
3.5
3.9
5.2
23.2
0.17
Global Private Equity
USD
11.3
10.8
10.7
12.1
9.7
8.9
9.2
8.7
8.5
14.0
0.62
Global Private Infrastructure
USD
10.2
10.2
10.2
11.0
9.1
8.5
8.2
8.0
8.1
10.0
0.81
Global Private Real Assets
USD
5.9
6.0
6.4
6.7
4.9
4.7
3.9
3.9
4.3
18.8
0.21
UK Infrastructure Social
GBP
5.3
5.2
5.2
4.6
5.0
5.3
4.5
4.5
4.4
11.2
0.40
UK Infrastructure Renewables
GBP
5.9
5.8
5.8
5.3
5.6
5.9
5.1
5.1
5.0
11.2
0.46
Alternative Risk Premia
USD
5.5
6.0
6.8
6.3
4.9
5.1
3.5
3.9
4.7
10.5
0.38
Hedge Funds
USD
3.5
4.0
4.8
4.3
2.9
3.1
1.6
1.9
2.7
7.0
0.28
ABS - Mezzanine
USD
5.0
5.5
6.3
5.8
4.4
4.6
3.1
3.4
4.2
8.9
0.39
Commodity Futures
USD
1.5
2.0
2.7
2.1
0.9
1.1
-0.4
0.0
0.7
21.4
0.00
UK Cash 3M LIBOR
GBP
0.5
0.8
1.7
-0.1
0.6
1.8
-0.2
0.1
0.9
0.8
0.18
USD Cash 3M LIBOR
USD
1.9
2.3
3.1
2.6
1.3
1.5
0.0
0.3
1.1
0.8
0.43
EUR Cash 3M LIBOR
EUR
0.0
0.3
1.1
0.0
0.3
1.1
0.0
0.3
1.1
0.8
0.45
Source: Aberdeen, February, 2017. Note: Volatility and Sharpe ratio refers to EUR Hedged. Return projections are estimates and provide no guarantee of future results. Returns and volatilities are in percent per annum.
86
EUR
Asset
Appendix
121028114
Aberdeen Asset Management 40 Princes Street Edinburgh EH2 2BY Tel: +44 (0)131 504 4060 Fax: +44 (0)131 528 4400