The Economic Consequences of President Mitterrand [PDF]

policies emphasized aggregate demand management during a period in which .... Reaganomics supply-siders focus on the eff

27 downloads 22 Views 4MB Size

Recommend Stories


Economic Consequences
At the end of your life, you will never regret not having passed one more test, not winning one more

PDF Download Economic Report of the President PDF Online
Knock, And He'll open the door. Vanish, And He'll make you shine like the sun. Fall, And He'll raise

The economic consequences of leaving the EU
I want to sing like the birds sing, not worrying about who hears or what they think. Rumi

The Economic and Fiscal Consequences of Immigration
In every community, there is work to be done. In every nation, there are wounds to heal. In every heart,

Economic Consequences of the Ukraine Conflict
Almost everything will work again if you unplug it for a few minutes, including you. Anne Lamott

Economic consequences of the new rice technology
If your life's work can be accomplished in your lifetime, you're not thinking big enough. Wes Jacks

The Economic Consequences of Oil Shocks
You can never cross the ocean unless you have the courage to lose sight of the shore. Andrè Gide

The Economic Consequences of Hospital Admissions
Don't count the days, make the days count. Muhammad Ali

The economic consequences of ageing populations
Don't watch the clock, do what it does. Keep Going. Sam Levenson

Early Evidence on the Economic Consequences
And you? When will you begin that long journey into yourself? Rumi

Idea Transcript


The economic consequences of President Mitterrand Sachs Jeffrey

HarvardUniversity and CEPR

CharlesWyplosz

INSEAD, Fontainebleauand CEPR

1. Introduction The economic managementof France under the Socialistsin the 1980s is widely regarded as a major failure. Relative to the enthusiastic promisesof PresidentMitterrand'sgovernmentfollowingthe election victoryof May 1981, thatharsh assessmentis warranted.The governmentcame to officepromisingrapid growthand a reductionin unemployment.Instead, it has presided over fouryearsof slow growth,and a risein unemploymentfrom7% in 1981 to 10% in October 1985. The Socialistattemptto revivethe French economy sank on the shoals of risinginflationand a foreignexchange crisisby early 1983, so thatthe governmentof the left has in the end introduced a tougher, more marketorientedprogrammethananythingconsideredby the previous administrationof Giscard d'Estaing. centre-right In politics,consistencycan be more importantthan correctness,so that the Socialist U-turn and subsequent policies have garnered less applause thanmightbe expected.The austeritysince 1983 has reversed a trend of high and increasinginflationthat gripped France in the 1970s and early 1980s. The policies have also stabilized the fiscal accountsand greatlyimprovedFrance's externalbalance. Nonetheless, the Socialistsremain condemned by the failureof theirinitialexperiment. According to opinion polls withinFrance, they have alienated the leftwithoutattractingnew electoralsupport fromthe centreand the right. I

I

* Usefulcommentshave been receivedfromseveral participantsat the Panel and our discussants, fromparticipantsat the Paris Seminaired'Economie MonetaireInternationaleand the Business Economics Workshopat INSEAD. We also thankH. Delestre,P. Dubois, J. Melitz,G. Laroque and G. Oudiz for assistance withdata collection,and Fergal Byrne for useful research help. Financial support fromINSEAD is gratefullyacknowledged.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

EconomicPolicy April 1986 Printedin GreatBritain

SUMMARY

FranceunderMitterrand Sachs and CharlesWyplosz Jeffrey

While French economic performancehas not conformedto the enthusiasticpromisesof early 1981, thispaper argues that the widelyheld view thatit has been an unmitigatedfailure is unwarranted.First,the general deteriorationcan largely be dated back to the early seventies.Second, the early and unfortunateattemptat a demand-led expansion was both moderate in size and quicklyreversed.Finally,the post-1983 anti-inflationary policieshave been successfulso that,overall, the disappointingFrench performancehas been on a par with,and in some waysbetterthan, the restof Europe. But the main theme of the paper is the emphasis on the supply side. The analysisis organized around the NAIRU, the thresholdrate of unemploymentbelow which inflation rises. In an attemptto explain the factorswhich have led to a continuousriseof the NAIRU since 1973, the paper focuses on labour costs,particularlyon the 'wedge' betweenthe costs borne by employersand net take-homepay. To achieve and sustain a significantly lower level of unemployment,labour taxes mustbe cut,thus reducingthe wedge and the NAIRU, and there mustbe a correspondingdemand stimulusso that actual unemploymentcan fallto thislevel.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

263

Our paper aimsto providea broader perspectiveon Socialistmanagementof the Frencheconomy,by placingrecenteconomicperformance in the context of longer-termdevelopments,and by comparing the macroeconomicoutcomes in France with the resultsachieved in the other major industrialeconomies. We argue that the economic crisis in France did notbegin withthe Socialistsin the 1980s,butin the 1970s, under Presidents Pompidou and Giscard d'Estaing. Macroeconomic policies emphasized aggregate demand managementduring a period in which aggregate supply conditionsseriouslydeteriorated.Because of thisdeterioration,both inflationand unemploymentrose sharplyin the seventiesin France as elsewherein WesternEurope. Certainly,the early Mitterrandperiod did nothingto reverse this trend; the policies pursued were more naivelydemand orientedthan in anyothermajor industrialcountry.Startingfroma positionof double digitinflation,the Mitterrandgovernmentblithelystepped on the fiscal accelerator,hoping that the stimuluswould raise output,spur investment, and thereby lead to anti-inflationary productivityincreases. Labour costs were substantiallyraised by higher minimum wages, increasedpaid holidays,and a shorterworkingweek. In the shortterm, France avoided the deep 1982 recession sufferedby its neighbours. However, French inflationremained veryhigh in 1981 and began to risein 1982, whileinflationabroad fellsignificantly. Also,withdomestic demand expansion and a worldwide recession, the trade balance deterioratedsharply.Withthe Franc under attack,the fiscalexpansion was quicklyabandoned. Since mid-1983,the fiscalexpansion has been nearlyreversed,and wage policieshave been successfulin holdingdown the growthof real labour costs. Table 1 shows the time paths of inflation,unemployment,and the external balance, in the past fifteenyears. Clearly,PresidentGiscard bequeathed a poorlyperformingeconomyto the Socialistsin May 1981. The unemploymentrate had risen in everyyear since 1974 to a rate of 6.3% in 1980, and the inflationrateincreasedto 12.2% in 1980. Both measures were far higher than in Germany,France's major political and economic partner.The currentaccount was in substantialdeficit. In 1982, the situationdeteriorated,with higher inflation,unemployment,and currentaccount deficits.By 1985, aftermore than twoyears of austerity,substantialprogresswas finallymade on two of the three and major targetsof the government:inflationhad fallensubstantially, the currentaccount was back in balance. However,the unemployment rate,once the prioritytargetof the Mitterrandgovernment,had risen sharplyto over 10%. One popular wayto compare the relativesuccessof the fightagainst inflationin different countriesis to calculatesacrificeratios.For a given

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

264 Table 1. Macroeconomic indicators 1974-80

1980

1981

1982

1983

1984

10.5 4.7 16.1

12.2 4.5 19.8

11.8 4.2 11.8

12.6 4.4 7.4

9.5 3.3 5.0

7.1

1.9 4.4

6.0 2.3 5.0

Unemploymentrate(%) France 4.8 3.2 Germany UK 5.3

6.3 3.0 6.6

7.4 4.4 10.0

8.1 6.1 11.4

8.3 8.0 12.6

9.7 8.5 13.0

10.1 8.8 13.2

3.1 2.4 1.5

1.0 1.9 -2.2

0.5 -0.2 -1.3

1.8 -0.1 1.8

0.7 1.3 3.2

1.6 2.6 1.6

1.0 2.3 3.3

Currentaccount (% of GDP) France -0.3 0.2 Germany UK 0.5

-1.4 -1.8 1.8

-0.8 -0.8 2.9

-2.2 0.5 1.9

-0.9 0.6 0.8

-0.2 1.0 0.0

0.1 2.1 0.3

Structuralbudget deficit(% of GDIP) -0.7 0.6 France 3.2 3.2 Germany -4.0 UK 1.9

0.3 3.0 -4.4

0.7 1.6 -4.5

0.8 0.4 -1.5

0.2 -0.8 -1.8

Net public debt (% of GDP) France 11.5 -8.2 Germany UK 75.3

10.6 17.5 48.2

12.5 19.8 47.2

15.0 21.5 49.0

17.6 21.7 49.8

Inflation(% p.a.) France Germany UK

GDP growth(% p.a.) France Germany UK

9.1 14.3 48.9

1985

Sources:Price and Muller (1984), DRI data bank. 1985 data are forecastsfromOECD EconomicOutlook(December 1985). Notes:CurrentAccount data in column 1 referto 1975-80 for France and Germany, 1977-80 for the UK. StructuralDeficitis inflation-adjustedat mid-cycle.Net Public Debt Data in column 1 referto 1970.

period, the sacrificeratio is defined as the cumulative increase in unemploymentdivided by the reductionin the annual inflationrate. It is a crude measure of the additional short-termunemployment burden required to get inflationdown 1 percentagepoint.From Table 1 we see thatFrenchinflationfellfrom11.8% in 1981 to 16.0% in 1985, a reductionof 5.8%. Over the same period the cumulativeincrease in unemploymentwas 12.1%. Hence during 1981-85 the sacrificeratio was 2.1. For Germanyand the UK, over the same period the sacrifice In thissense,Table 1A showsthat ratioswere 10.9 and 4.0 respectively. than trade-off France faced a more favourableinflation-unemployment other European countries. Whilstthe Socialist experimenthas been no great success, French performancein some ways has been better than that in the rest of

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

265

Table 1A. Sacrificeratios, 1981-85

Cumulativeextra unemployment(%) Reduction in inflation(%) Sacrificeratio

France

Germany

UK

12.1 5.8 2.1

20.8 1.9 10.9

27.2 6.8 4.0

Source:as Table 1.

Europe. During the 1970s and 1980s, for most European countries, ever-higherunemploymentrateshave proved necessarymerelyto hold inflationin check,muchlessto reduce it.The so-called'non-accelerating inflationrate of unemployment'(NAIRU), whichmeasures the lowest unemploymentrate that can be sustained withoutfear of increasing inflation,appears to have risensharply.Note, forexample, thatduring the Giscard d'Estaing administration,the historicallyhigh and rising unemploymentrates during 1977-79, between the two oil shocks,did littleto reduce French inflation(see Table 1). In our view, the anti-supplyside policies of both Giscard d'Estaing and Mitterrand,like the policiespursued in othercountriesof Europe, contributedto this distressingdevelopment. The post-1983 period under Mitterrandhas begun to alleviate the problem,but in spite of thewillingnessofthegovernmentto undertakepolitically toughpolicies, it has stillnot shown a deep appreciationof the need forcoordinated supplyand demand measures.Some measurementsofaggregatesupply conditionsthatwe develop latershow onlyslightglimmersof improvement. Our policy recommendationsat the end of this paper focus squarely on the need for furtherimprovementsin aggregate supply conditions. We must spell out at the beginninghow our approach differsfrom The 'supply-side'economicsas preachedbythe Reagan Administration. on on the effects of income tax cuts focus Reaganomics supply-siders the incentivesto save and work. Individuals, it is argued, will work withlowermarharder,and undertakemore entrepreneurialactivities, of household behaviourto ginal tax rates. In our view,high sensitivity to be demonstrated: we do not depart from rates has tax yet marginal thetraditionalmacroeconomicassumptionofa fairlyinelasticlong-term labour supply of primarywage-earners. In our interpretation,sensitivity of workersto tax rates occurs not at the household level, but at the level of wage bargaining between firmsand workers,and in political lobbying for governmentwage policies (see Bruno and Sachs, 1985, and Knoesterand van der Windt, 1985). As we showbelow, unions willtryto pass on labour tax increases

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

266

Sachs and CharlesWyplosz Jeffrey

by raisingwages, even if such wage increaseswillengender unemployment for other workers,and even if unions faithfullyrepresentthe interestsof memberswhoseindividuallabour supplyis completelywage inelastic!Because unions have monopolypower,theychoose willingly to acceptmore unemploymentin returnforhigherwages.When labour taxes or oil price increases adverselyshiftthe demand for workers, unions are likelyto accept higherunemploymentin order to preserve real incomes.The same argumentcould be applied to politicalpressures forincreasesin governmentsalaries,or in government-regulated wage levels (as with the minimum wage). Even those voters who would willinglywork at much lower wages will often support government policies that preserve higher wages at the cost of higher aggregate unemployment(of others?). To quantifythe factorsthathave contributedto a risingNAIRU we constructa variablethatmeasuresthe 'wedge' betweencoststo the firm and net-of-taxtake home pay of workers.The theorysuggeststhat a rise in the wedge willtend to raise the NAIRU, since unions willresist the real wage reductionsthatwould otherwisebe necessaryin order to preventhigherlabour costs fromreducing the demand for labour by firms.We findthatthe 'wedge' has increasedmarkedlyin France,under Giscard d'Estaing as well as Mitterrand,mainlybecause of the increase in social securitytaxes.The risein thewedge seemsto be wellcorrelated with the rise in the non-inflationary thresholdof unemploymentin France. For this reason, we share the Reagonomics emphasis on the efficacyof tax reductions as a way of reducing unemployment, especiallysince France has had one of the fastestincreasesin the Social Securitytax burden in the OECD in the past fifteenyears,as shownby the data in Table 2. The plan of the paper is to review the Mitterrandexperience, emphasising both aggregate supply and aggregate demand factors. Section 2 describes the Socialist outlook at the time of the election victoryin May 1981, and its economic program,which combined a Keynesiandemand expansion,a Blum-stylesetofsupply-sidemeasures, and nationalizationof major industrialgroups. Why this initialprogrammefailedis the subjectof some dispute.In Section3, we consider threepossible reasons: the worldrecession,adverse supplyconditions, and a crisisof confidence.While all of these explanationshave some merit,we make one centralpoint:the NAIRU, or thresholdunemployment rate, was simplytoo high to allow for a sustaineddemand-led recovery,and thiswas compounded by adverse supply-sidemeasures. While the debacle of this phase is well known, the brevityand moderation of the demand expansion that actuallytook place is not widelyappreciated. As we show, the expansion was already winding

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

267

Table 2. Social securityexpenditures (Structurallevel as % of potential GDP)

1970 1975 1980 1981 1982 1983

France

Germany

UK

US

16.8 19.1 22.6 23.4 24.2 24.1

13.3 16.2 16.3 16.3 15.8 14.9

8.5 9.2 10.0 10.0 10.5 10.5

7.5 9.7 10.0 10.0 9.7 9.8

Source:Mullerand Price (1984). Note: Data measure social securityexpendituresat full GDP. The strucemployment,as % of full-employment tural level of social securityexpendituresis constructed analogouslyto the structuralbudget deficit;it is the level of spendingwhichthe systemwould produce at a normal level of economic activity.

down by the second devaluationcrisis,in June 1982, a mere 13 months afterthe Socialistscame to power. However, March 1983 broughtthe decisivebattleforcontroloftheFrencheconomyand theSocialistparty. Since March 1983, demand managementhas been geared, successfully, towardsa reductionof inflation.Althoughthesupplyside has continued to receiveonly fragmentary attention,the wage policyhas been one of moderation (at significantpoliticalcost). Even so, unemploymenthas risensharply.The outcome to date of 'socialistausterity'is reviewedin Section 4. In Section 5, the currentmacroeconomicenvironmentis surveyed, and some policyrecommendationsare offered.Accordingto our estimates,a reductionof the unemploymentrateto 5% is achievablein the next few years,througha combinationof demand expansion coupled witha significantreductionin labour costs.A cut in labour taxes would be useful in reducing labour costs to firms,therebyspurringemploymentand investment. 2. The firstthousand days 2.1. The political background

The election of Francois Mitterrandas Presidentin May 1981 was a major politicalchange. The lefthad not been in power since 1958, and Mitterrand'sbrand of socialism was perceived as more ferventand the dogmaticthan the socialismof the countlesscoalitionsin the fifties, shiftbeing clearlyunderlinedbythe explicitagreementto rule together withthe Communistparty.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachsand CharlesWyplosz Jeffrey

268

The changes were to cover the whole spectrumof economic policy: nationalizationof severalof the largestindustrialcorporationsremaining in the privatesector,new taxes (especiallya wealthtax), new labour market regulations (the Auroux Laws), and a new orientationfor macroeconomic policies. Many of the changes were based on a specificallysocialistperspectiveon the economy,but there was also a general perceptionin France thatthe previous economic policies had failed and thata change was needed. The left wing of the Socialistparty,nurturinga strongscepticism towardmarketforces,was attachedto the ideological principlesincorporated in the joint programmeadopted withthe CommunistPartyin 1972,and enshrinedin the 'ProgrammeCommun'. They wereprepared to go ahead with a large expansion of the public sector,quickly to redistributeincome and strengthenthe hand of the governmentin runningthe economy.It seems fairto say thatthiswas the dominating view in the Spring of 1981. At the otherend of the spectrum,the approach was more pragmatic and cautious. The examples of the Front Populaire in 1936 and of Chile's Allende governmentwere cited.' Several adviserswere aware thathighunemploymentand highinflation, a delicateexternalsituation and verynervousemployersand investorsrequireda carefulapproach. 2.2. The French economy in 1981

A carefulstudyof the economic policies of the Giscard era is beyond the scope of this paper. As later under Mitterrand,there were two periodsassociatedwiththetwoPrimeMinisters.JacquesChiracpresided over an expansion in the aftermathof the firstoil shock. It led to a quicklyrisinginflationrate,a stubbornlyupward creeping unemploymentrate,and severeexternalimbalanceswhichtwicepromptedFrance to leave the European currencyarrangementknown as the Snake. Raymond Barre replaced Chirac in late 1976 and quicklyestablished his mark as a 'dismal' economistbent on the hard work of deflation. His main themeswere balancingthe budget, strengthening the Franc and wage moderation.In a laterstage he engineereda historicalbreak fromthe entrenchedpracticeof pervasivepricecontrols.He succeeded in balancingthebudget,in leading France intothe European Monetary Systemand in never havingto devalue the Franc. Over the period 1973-81, in France as elsewherein Europe, the rise in unemploymenthad been met by a series of measures which conC. Kom,La 1S. S. C. Kolm, 'La

TransitionSocialisteFrancaise,' Le Monde 17-18 June 1981. TransitionSocialisteFrancaise,' Le Monde 17-18 June 1981.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

269

tributedto a furtherrise of unemployment.Unemploymentbenefits, mainlyfinancedbytaxes on wages,increasedfasterthanactual salaries, along with other social transfersalso financed by labour overheads (healthand retirementbenefits,as wellas the elaboratesystemof family allowances). By 1981, 25% of GDP was being thus redistributed, financedby wage taxes,the singlelargesttax in France. In thiscontext, stabilizinglabour costswould have requiredeitheran explicitagreement withthe trade unions on wage moderationor relyingon the pressure of high unemployment.Giscard and Barre's hostilerelationshipwith the trade unions meant thatthe formersolutionwas ruled out. Rising unemploymentin 1980 allowed for a stabilizationof real take-home wages, but increasingoverhead charges kept real labour costs rising, althoughat a slower pace. The combined rise of inflationand unemployment,along withan emergingcurrentaccount deficitin 1980, was of course related to the two oil shocks.But, as evidenced by Germany'ssuperiorperformance, thiscould not be the only explanation (for some detailed comparisons of Franceand Germany,see De Meniland Sastre,1985). Indeed, France had largelyfollowedthe general European trend of supply-sidemismanagement,withexcessivelabour costs,a fastgrowthof the welfare system,and continuouslyspreading regulation of labour markets. Anotherliabilityinheritedby Mitterrandwas an overvaluedexchange rate.Since thecreationoftheEMS and despitean unfavourableinflation differentialvis-a-visGermany,the parityof the Franc had remained unchanged. In theevent,thereal appreciation(about 10 to 12%) proved to be the firsthurdle on whichthe Socialistsstumbled. The brightside of Giscard'sand Barre's legacyresided in the public finances.With the exception of the early Chirac expansion and the temporary'locomotive'agreementof the 1978 summit,the budget had remained almostbalanced. As a result,France's public debt (relativeto GDP) was in 1980 one of the smallestamong OECD countries(Table 1). Mitterrandenjoyed fromthe beginningmore budgetaryfreedom of manoeuvre than was actually available to most other European countries,and it is perhaps not surprisingthatthis is where he chose to move forcefully.

2.3. Policy measures

Fairlyquickly,the new governmentenacted a seriesof policymeasures: a fiscal expansion with monetaryaccommodation, minimum wage policies,a reductionof the workweekand new laws concerninglabour relations.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

270

Sachs and CharlesWyplosz Jeffrey

% 26 24- -

2

-

'

'

i

22 20 -

I

10

;I 8

I

/May81

'IV 8 1I 10 80

II

i

I, 1I I ,I I

I

LondonEuro Market l ,i (Frenchfrancrate)

\'-

Tv..

I I ' I March83 ~-, Paris I I I I I| I I I Ii MoneyMarket I

-

81

1I

-

82

83

84

85

Figure 1. Short-terminterestrates Source:IMF

2.3.1. Fiscal policy.The supplementary1981 budget increased spending

withtaxes almostunchanged. In 1981, mostof the increasedspending was in transfersto households,and subsidiesto both state-ownedand privatecorporations,while investmentspending by public enterprises was boosted in 1982. Also, the governmentdirectlyhired 200,000 civil servantsin 1981 and 1982. As a measure of the expansionaryimpulse,we considerthe OECD's structuralbudgetbalance (fora criticism of anysingle inflation-adjusted measureof fiscalpolicystancesee Buiter,1985), shownin Table 1. The fiscalexpansion contributed1% of GDP in 1981, risingto a cumulative impactof 1.5% of GDP in 1983. As a matterof comparisonthe Chirac expansion contributed1.1% in 1975 over 1974, while the post-Bonn summitboost amounted to 1.5% in 1978 (followedby a reductionof 2.8% between 1978 and 1980!). 2.3.2. Monetarypolicy. Monetary policy was mainly accommodating. The

sharp increase in interestrates shown in Figure 1 may be misleading in this respect,because of the French practiceof creditceilings:each yeareach bank is givena maximumallowed growthrate foritsvolume of creditoutstanding.These ceilingswere moderatelyraised in June 1981 and graduallybroughtback down in 1982. Interestrates do not clear the marketfor credit,but ratherare directedby the Banque de France withan eye set towardeitherinvestmentand public borrowing

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences ofPresidentMitterrand

271

conditions,or the foreignexchange market,depending on the prevailing circumstances.In 1981-83 themainconsiderationwas the perceived need to protectthe Franc. Consequently,the 24 hour interbankrate jumped from 10.9% in February 1981 to 18.9% in June, to decline thereaftertoward 15% by end 1982. In the event, capital controls preventedan even steeper rise in interestrates in face of an intense speculation.This is shown in Figure 1 as the London Eurofrancrate whichis an indicationofthe levelcompatiblewiththe expected depreciation (and temporarilyfallsaftereach EMS realignment).Yet all this as the M2 target(10%) was overrun does not signalmonetarystringency to an in 1981 (11.4%), leading upward adjustmentof the reference trendfor1982. This overrunin factdid notcorrespondto anyexpansion of credit to the private sector,but rather to direct financingof the governmentborrowingrequirement(one thirdof the deficitfor 1981 was money financed). 2.3.3. Labour market.Mitterrand'selection was perceived by the trade unions as a momentouschange. The pledge to fightunemployment, togetherwith a host of specificpromises to redistributeincome and strengthenthe hands of trade unions enshrined in the 'Programme Commun', were perceived as the beginningof a new era, akin to the historictransformationsachieved during the short tenure of Leon Blum at the head of the Front Populaire in 1936-37. Nationalization and the new tax on wealth were preciselydesigned to symbolizethe shiftof power.While itis hard to assess the macroeconomicsignificance of some of these changes, we shall emphasise the effectsof three particularmeasures concerningthe labour market. The minimumwage: In 1981 a view commonlyheld in France, and incorporated in some macroeconometric models, was that an autonomous increase in real wages has an expansionaryeffect.The usual mechanismis a transferfromfirmsto groupswitha highmarginal propensityto consume, increasingprivateconsumptionwhich,via the familiarmultiplier-accelerator mechanism,generatesmore investment in of the spending spite profitsqueeze. The adverse effectof the profit squeeze on outputand investment spendingis delayedand is dominated in the interimby a reductionof unused capacity.2The sharp increases in the minimumwage (the SMIC) were thus expected to fulfilltwo and higheremployment.The major objectives;income redistribution SMIC was raised by 10% in nominal termsin June 1981, followedby

I 2

1 This resultwas commonin 1981 models likeMETRIC but seemsto have been modifiedrecently. Sterdyniaketal. (1984) reportthatin the OFCE model investmentspending fallsimmediately, but total spending increasesfortwo years,followedby a severe recession.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

272

Sachsand CharlesWyplosz Jeffrey

severalsmallerincrements.It rose bya cumulated38.9% over 1981-82, and in real termsby 11.4% between April 1981 and July1982. The number of workersreceivingthe SMIC doubled to about 8% of the labour force (Bourit, Hernu and Perrot, 1982). Overall, the average real hourlywage rate increased by 5.2% during the same period. The Auroux Laws: Several measures were aimed at changing the relationshipbetweenemployersand trade unions. The best knownof themconstitutedthe 'Lois Auroux', a seriesof new laws whichcodified and often enlarged the statutoryrole of trade unions. These laws, named after the Minister of Labour, Jean Auroux, provide union leaders with increased protectionagainst disciplinarymeasures and establishthe rightof workersat the shop floorand factorylevels to be involved in productionorganizationchoices. They also require wage eliminating negotiationswithtradeunionsat leastonce a year,effectively the scope for non-union wage settlements.Despite the fact that less than 20% of the labour forceformallybelong to a union, most of it is de factounionized, forexample because in manylow-paid,non-union sectorswages are set close to the minimumwage level, which in turn is set by the governmentafterconsultationwiththe unions. Other labour measurescontinuedthe trendof the seventiesin reducing the abilityof employersto dismissredundantworkersor to resort to part-timeand temporaryemployment.Since 1975 collectivedismissals must be submittedto administrativeapproval. Such approval, althoughusually granted,may be, and was, routinelydelayed. Shorterhours and longer holidays: Two decisionswere designed to reduce the lengthof workingtime. The firstextended the lengthof mandatorypaid vacations from 4 to 5 weeks per year. The second reduced the 'legal' workweekfrom40 to 39 hours as of January1982, and was meantto be a firststeptowardsa 35 hour weekto be established by 1985.3 The legal workweekis the number of weeklyhours beyond whichworkersmustbe offeredovertimerates,whichincrease steeply. The actual systemis quite complicatedand need not be explained in detail. It is enough to note that average hours per year declined by more than 3% between 1981 and 1983. Aftera major debate, Mitterrandexpressed publiclyhis desire to see thatmonthlypay remainunaffectedby the reductionof the workweek. Estimatesreported by Marchand, Rault and Turpin, (1983), indicate that such was indeed the outcome for the overwhelmingmajorityof cases. Consequently,real hourlysalariesrose quite sharplyin 1982: the 3 These decisionswere also a powerfulsymbolicgesture,echoingthe introductionof the 40 hour workweekand paid vacations in 1936 by Leon Blum, considered by many people to be his major achievement.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

273

additional vacationweek raised them by about 2%, withanother 2.5% correspondingto the reduced workweek. 2.4. The nationalizationprogramme

Since the signingof the 'ProgrammeCommun' in 1972, nationalization had been a major commitmentforMitterrand.The argumentsbehind this programme reached deep into the social historyof France, and there was littleroom in 1981 for questioningthe principlesinvolved: the governmentfelt the need to control directlysensitiveparts of the economy (bankingsector,high-technology, naturalmonopolies) in order to guaranteemore growthand greaterequity.Debate concerned only the size of the programme. The final decision was half-waybetween the maximalistproject of the CommunistPartyand the shorterlistof the Socialists'rightwing. It involved 12 industrialfirms,36 banks and 2 financialcorporations. As of May 1981, industrialstate-ownedcompanies representedabout 5% of GDP, withthe share risingto about 8% afterthe new nationalization measures. The newly nationalized industrialgroups employed some 550,000 workers,2.6% of the work force. By 1981, the publicly owned banks already received60% of all deposits.New nationalization took this share to almost 90%. When it is recognized that the largest insurance companies and a host of financialinstitutionswere already in the public sector,virtuallythe whole financialsectoris now directly controlledby the government. The question of interesthere is the macroeconomic impactof nationalization. This would require a difficult analysis,beyond the scope of this article.We limitourselvesto remarksabout the conditionsunder whichthe macroeconomiceffectsare likelyto be significant. First,the governmentmay intend to use the nationalized sectorto achieve macroeconomictargets.While jobs were created in 1981-82 in the governmentadministration, nothingmuchwas evidentin the public industrialsector.At most,lay-offs were postponed.But since 1983 there have been deep cuts in the steel industry,the automobile sector (Renault), and several other newlynationalizedcorporations. Second, the governmentmay change the allocation of resources, drawingfundstowardstate-ownedcorporations.It is truethatbetween 1981 and 1983 subsidies to the public sector rose markedly,but this trendhas now been reversed.By 1984, mostof the newlynationalized firmshave seen theirsubsidiesfallor disappear and have restoredtheir financialequilibrium. Third, the governmentmaynow be temptedto alteritstrade policies to protectthe public corporationssince, while before 1981 the output

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

274

Sachs and CharlesWyplosz Jeffrey

of the public sectorwas largelyinternationally nontradable(electricity, railroad,telephoneand postalservices),all the newlynationalizedfirms tradable sector.Casual evidence suggests belong to the internationally that France has always protectedits corporations,public or private, whenitwas deemed necessary.The governmenthas mainlyencouraged a wave of mergers (the so-called restructuring), all motivatedby the of international As noted pressure competition. by Stoffaes(1985), this is likelyto be the major impactof nationalization,and we tentatively conclude that nationalization has probably not had major macroeconomiceffects. 2.5. The outcome

The evolutionof the French economyduring 1981-83 certainlyfailed to conformto the Government'sexpectations.The officialforecastsfor 1982 and 1983 announced a rate of growthof real GDP of 3.1 and 2% whiletheactual resultswere 1.8 and 0.7%. For the inflation respectively, the forecasts of 13.4 and 8.2% mustbe compared withthe realized rate, values of 12.6 and 9.5%. Unemploymentdid not fall,and the Franc was buffetedby several crises. This sequence of setbackshas deeply impressedobserversof the Frencheconomy,to the pointof frequently overshadowingthe subsequent economic accomplishments.Yet the episode of expansionarypolicieswas relativelyshortlived,and a complete explanation of what went wrong does not seem to have been offered.Before turningin the nextsectionto an analysisof thisperiod, we firstpresentthe main factsand show how a successionof foreign exchange crisesbroughtabout a complete policyreversal. Unlike other large OECD countrieswhichsufferedtwoyearsof real GDP reductionin theearlyeighties(Table 1), FrenchGDP neveractually fell. On the other hand, inflation,which had remained stubbornly around 10%, startedto rise in 1980 afterthe oil shock and increased again in 1982, while most other industrialcountrieswere achieving significantreductions(see Table 1). To characterizethe employmentsituation,Table 3 separates the competitivesectorfromthe shelteredsector.The latterincludes those industrieswhich are not subjected to marketforcesbecause theyare dominated by state-ownedmonopolies. We furtherseparate out the competitivesectorinto its tradable and non-tradablecomponents(for details see the notes on Table 3). Total employment,measured by the number of employed workers, had been practicallystagnantsince the early seventiesand declined marginallyafter1980, hence the steadyincreasein the unemployment rate given a trend growthof 0.7% in the labour force.The stabilityof

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

275

Table 3. Indices of employment(E) and manhours (MH) (1980 = 100) 1973

1979

1981

1982

1983

1984

Competitivesector: E MH

100.1 108.5

100.2 100.2

98.8 96.8

98.1 93.7

96.7 91.4

95.0 88.9

tradable

E MH

111.0 125.0

101.6 103.2

98.8 95.7

95.0 91.6

92.9 87.8

90.4 84.7

nontradable

E MH

90.7 95.6

99.0 97.8

100.5 97.6

100.8 95.5

100.0 94.3

99.0 92.5

Shelteredsector:

E

92.3

99.1

101.1

103.3

105.1

106.1

Total

E MH

97.9 101.5

99.9 98.5

99.4 100.8

99.6 98.3

99.1 97.7

98.1 98.0

Source:INSEE: Rapportsur les Comptesde La Nation. Notes:tradables (those sectorswhere exportsplus importsat least 5% of value added; agriculture,manufacturing);nontradables:construction,retailand wholesale services; financial. sheltered:energy,transport,telecommunications,

total employment,in France as in most other European countries, conceals the divergentevolution of a declining traded sector and a growingnon-tradedsector.Overall employmentwas maintainedclose to its 1980 level largelyas the resultof an increase in the number of workersin the shelteredand governmentsectors.In conclusion,while the expansionarymeasuresmayhave insulatedFrance fromthe worldwide recession in 1982, they certainlyfailed to deliver the expected increase in employment. The new governmentwas also confrontedwith several speculative attackson the Franc. The firstone actually started in the last days beforethe electionand swelledimmediatelythereafter,promptingthe centralbank to interveneheavilyand to raise interestrates. Minister of Economic AffairsJacques Delors wanted an immediatedevaluation whilethe President'sotheradvisersdid not wantto give an impression of weaknessin the face of speculativeattacks.The latterviewprevailed, leavingFrance withan overvaluedexchange rate,thatbecame increasinglyso as inflationedged upward while abating abroad. The current account deficitworsened,lending strengthto the beliefthata devaluationwithintheEMS could notbe avoided. And theFrancwas devalued, not once but three times. The firstdevaluationoccurredin October 1981. Aftersix monthsof continuingpressure,the CentralBank had exhaustedmostofitsoptions of capital controls,and heavyuse of (high interestrates,strengthening the foreignexchange reserves).The officialviewwas thatthe exchange rate devaluation representeda technicaladjustmentpromptedby the

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

276

Sachs and CharlesWyplosz Jeffrey

legacy of a high inflationrate and a wave of unfriendlyscepticismon the financialmarkets. Paradoxically,the devaluationstrengthenedthe hand of the leftwing withinthe Government.It was, they said, a proof that 'new' policies could not be successfulwithinan economydominatedby the iron law of markets.More control, not less, was required. The Government pressed ahead withthe implementationof the ProgrammeCommun. To manyobservers,France had embarkedon a path thatwould eventually lead it towards more controls,more dirigismand a break-away fromthe EMS and EEC. Givensuchexpectations,theexchangemarkets remained sceptical.By March 1982, the pressurebuiltup once again, interestrates were raised and capital outflowsswelled to new heights. In June 1982, Mitterrandwas hostingthe Chiefsof Stateforthe annual summitmeetingin the magnificent'grandeur' of Versailles,while the Banque de France was depletingits reserves.Immediatelythereafter, the Franc was devalued by 10% vis-a-visthe DM. This time,the devaluationwas accompaniedbystabilizationmeasures.First,pricesand wages were frozenuntilOctober 1982. Second, new taxeswere raised through higher social securitycontributions.Third, the budget deficitwas targettedat a level not to exceed 3% of GDP. Fourth, some public expenditureswere delayed untilfurthernotice. For the firsttime,the word 'austerity'surfaced,suggestingan end to the expansionaryphase of the previous year, and a returnto the rulesof the game of the EMS. Thejoint freezeof bothpricesand wages, for the firsttime since 1958, signalled a decisive shifttowards more controls.In viewof thiscombinationof policies,it was unclear whether the Socialistswould abandon theirmacroeconomicpoliciesand return to Barre's medicine, or would go furtherin restrictingthe market economy,eventuallyleavingthe EMS. The answercame in March 1983 when, aftera furtherexchange rate crisis,the Franc was devalued by 8% vis-a-visthe DM. The external constrainthad forced Mitterrandto make a political decision he would rather have avoided. For Industry Minister Chevenementand his left-wingbackers,the only solutionwas to leave the EMS, abandon austerityand expand byincreasingpublicspending, especiallythroughthe enlarged public sector.The externalconstraint would be taken care of by protectionistmeasures, and if need be, a temporarysuspensionof EEC agreements.For Finance and Economics MinisterDelors and the more market-oriented wing,the externalconstraintrequired a strengtheningof the austeritymeasures phased in since the previous devaluation, while leaving the EMS would only worsen the situation. The President, after wavering, finallychose to shifttowards more restraintat home. This was the end of the

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

277

short-livedexpansionaryprogramme,alreadyshakenafterthe previous devaluationofJune1982. Hereafter,theSocialistswould pursuepolicies which looked a lot more like those of Barre than those they had advocated during the campaign. Chevenement left the Ministryof Industryand the Cabinet, and was replaced by Fabius, while Delors' responsibilitieswere enlarged. 3. Anatomyof a policy failure It took the Mitterrandgovernmentno more than two yearsto reverse its expansionarypolicy stance,only about a year if one notes that by June 1982 fiscalexpansion was already being cut back markedly.It is now officialdoctrinein the Socialistpartythat'we have learned a lesson'. Yet itis not clear whathas been learned. One interpretation, articulated Fonteneau and Muet is that the main was the delay (1983) by problem in the world recovery.In this view France was sufferingonly from insufficient aggregate demand. Accordingly,an expansion was warrantedin 1981, and could have been sustainedwithoutmajor external deficitshad the worldrecoverymaterializedthatyear,as mostforecasts then predicted. Another possibility,which we stronglyendorse below, is that the various early measures affectingthe labour marketsand the taxes on corporations furtherworsened supply-side conditions which were alreadyunfavourableby 1981. Several studies(Bruno and Sachs, 1985, Bruno, 1986,Artus,1984), have previouslysuggestedthatan expansionary policywould largelyfail to generate more employmentand more output,or would do so onlyat the costof higherinflation.But we would stillhave to account for the sharp slowdown in investmentspending and the sheer size of capital outflows.Could it be, then,thatthe mere electionof a Socialistgovernmentcommittedto sweepingsocial changes and income redistribution,prompted a crisis of confidence among investors,forcingthe Franc down and provokinga deep slump on the stockmarket? This sectionpresentsan attemptat disentanglingthe threeexplanations.Theyare not mutuallyexclusive,and all carrysome explanatory power. Each of them,though,has different implicationsregardingthe evaluationof the post-1983policiesand, more importantly, the outlook for France as of 1986. 3.1. Can demand factorsexplain everything?

One importantinterpretationof the failureof the 1981-82 expansion is thatthe world recessionof 1982 overwhelmedthe domesticdemand

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

278

stimulusin France. Such a view has received widespread expression both in France (Fonteneau and Muet, 1983, Muet, 1985, Beudaert, 1983) and abroad (WorldFinancial Markets,1983). There is certainly muchmeritto thisview.As we showbelow,thesize ofthecontractionary impulsecoming fromabroad in 1981 and 1982 was a large proportion of the size of the direct demand stimulusfromfiscalpolicy. If world demand had recoveredin 1982, ratherthancontractingsharply,French growthwould probablyhave been significantly higher,and the external deficitwould have been reduced. But this would not have been the end of the story.Higher growth would almost surelyhave produced an even fasterupturnin inflation than in fact occurred, and the real appreciation of the Franc would have been even more severe. Accordingto our estimatesFrance's 7% thresholdat that unemploymentin 1981 representedthe inflationary time.A lower unemploymentrate could certainlyhave been sustained, accelerationof inflation. but only at the cost of a significant The Mitterrandexpansion did have a temporaryeffectin boosting growthin the French economyin 1981. A comparisonof France and Germanyprovidesa good illustrationof thispoint.Based on the structuralbudget deficitfiguresof Table 1, the shiftof fiscalpolicybetween 1980 and 1983 amountsto a contractionof 2.8% of GDP in Germany, and an expansion of 1.5% of GDP in France. Over 1981-83, the rise in unemploymentin France is a moderate 1.9%, as compared with5% in Germany,and 4.2% in the EEC as a whole. By buckingthe trendof fiscalcontractionin the rest of the OECD, the Socialist government clearlyboughtsome moderationin theupwardtrendin unemployment, at least temporarily,but by less than had been expected in 1981. To gauge the effectsof the world recessionon France, we consider the change in the size of the export market in which France was competing. Export volume growthfor the industrialcountries as a whole slowed down quicklyafter1980, actuallydecreasingby 1.7% in 1982. Given the previoustrend,we estimatea shortfallof exportsdue to the world recession of 4.5% in 1982, and another 0.4% in 1983.4 The elasticityof demand for French goods withrespectto the overall marketforindustrialcountryexportsis approximately1.0. Thus, as a rough approximation,we can argue thatFrenchexportswere reduced by approximately4.5% in 1982 because of the world recession.Since French exports of goods and non-factorserviceswere about 20% of GDP in the early 1980s, the first-roundcontractionaryeffectof the i 4

1 Estimatebased on authors' regressions,and is similarto the 5.2% reductionin foreigndemand estimatedby Fonteneau and Muet (1983).

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

279

reduced exports(i.e. before any demand multiplieris considered) was approximately0.9% of GDP in 1982. This is more than half of the increase in the structuralbudget deficitin 1981 and 1982. In this sense, the assertionthat the demand stimuluswas sharply curtailed by the world recession is probablycorrect.This conclusion underscores the relativelysmall amount of fiscal stimulusthat was impartedin 1981 and 1982, as well as the severityof the global downturn.However, the world recessionwas not the only limitingfactorin successfullyreducing unemployment.For we should ask what would have happened to French inflation,and to the pressureon the Franc, in the event of a larger net expansion domestically.The inflation equations of the next sectionare emphaticon a key point: France was already at the inflationarythresholdlevel of unemploymentin 1981. Withmore expansion, inflationwould have had no place to go but up! It is importantto stressthat the rebound in inflationin France in 1982 was home made, rather than imported by the devaluations of 1981 and 1982. Import prices rose at the same rate as domesticprices in 1982 (12.2% forimportpricesand 12.6% forthe GDP deflator).This is not surprisingin thatthe devaluationsdid littlemore than keep the Franc depreciating at a rate equal to the inflationdifferentialwith France's EMS tradingpartners.Thus, withno world recession,there is littlecase for arguing that importprice inflationwould have been sharplylower. It is sometimesargued that even withhigher,and rising,inflation, the policy would have been sustainable in the absence of a world recession, since the external balance would have been substantially improved.On the firstround, exportswould have risenby about 0.9% of GDP, with the ultimatecurrentaccount improved by a somewhat smaller amount (after second-round induced importsare accounted for). Thus, the currentaccount deficitmighthave deterioratedfrom about 0.8% of GDP in 1981 to perhaps 1.7% of GDP in 1982, rather than the 2.2% level thatwas actuallyreached. Nonetheless,the current account positionwould stillhave been adverse (as should be expected afterany fiscalexpansion). Furthermore,the currentaccount is not the only channel through whichthe externalconstraintis felt.All threerealignmentswhich,taken together,led to a 32.2% devaluationof the Franc vis-a-visthe Deutschmark,followedintensespeculativerunswhichall but depleted France's non-goldforeignexchange reserves.As can be seen fromTable 4 much of the action in 1981 took place throughthe capital account: between the firstand the second subperiod,the overallbalance worsensby 2.1% of GDP, of whichonly 0.4% is attributableto the currentaccount. As for 1982, the dramaticimprovementin the capital account is largely

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

280

Table 4. Selected items in the balance of payments:quarterlyaverages

(% of GDP)

CurrentAccount Long-Term Capital Short-TermPrivateCapital Errorsand Omissions Overall Balance Short-termPublic Capital

1980Q1-1981Q1

1981Q2-1982Q2

1982Q3-1983Q1

1983Q2-1985Q1

-0.8 -0.3 0.6 0.2 -0.3 -1.1

-1.2 -0.5 -0.2 -0.4 -2.4 2.9

-2.6 1.5 -0.0 -0.1 -1.3 0.4

-0.3 0.5 -0.1 0.2 0.3 -1.0

Source: Banque de France, BulletinTrimestriel. Note: The ShortTerm Capital Accountexcludes the BankingSectorwhichis included in the Overall Balance. All data are on a quarterlybasis and the ratiosare averaged over each subperiod.

explainedbytheincreasinguse ofthetimehonouredpracticeofinstructing publicly owned corporations to borrow abroad ('emprunts encourages') as an indirectmeans of interveningon exchange markets. The currentaccount,therefore,is onlya partialmeasureofthe external constraint.Even withoutthe 1982 world recession,the Franc would almostsurelyhave been under stress. 3.2. The supply side

While mostEuropean countrieswentthrougha period of slow growth and regularlydeclining inflationin 1982 and 1983, consumer price inflationbegan to risein mid-1981and remainedstronguntilmid-1982 when most prices were frozen. We have already noted that foreign pricesand the exchange rate depreciationscannot be held responsible forthisdevelopment.Rather,7% unemploymentrepresenteda 'tight' labour market in 1981, offeringlittle scope for noninflationary expansion. A NAIRU of 7% was of course extraordinary,given the much lower average employmentrates of the 1960s and 1970s, and few analystsshowed any awarenessof how tightthe situationwas. The evidence rejects the key element of simple demand-centred policyanalyses:theassumptionofa constantNAIRU. WhentheNAIRU can change overtime,movementsofactualunemploymentratesrelative to theirhistoricalaverages providelittledirectevidence as to whether a demand expansion is warrantedor is likelyto be inflationary.As voluminousrecentresearchhas shown,the scope fordemand expansion can be determinedonlyaftera carefulanalysisof supplyconditionsin the economy.This analysisis unfortunately hampered by the absence of an agreed framework.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident Real takehome pay (W)

281

S

--

--

-\

L

I

I

Lu

Lu

F F

L

Employment(L)

Figure 2. The labour market

Our approach makes the NAIRU the cornerstoneof the analysisof supplybehaviour. In doing so we seek to answer the question: what is the lowest unemploymentrate that can be sustained in the medium and long-termwithoutproducinga continuouslyrisinginflationrate? Thus the NAIRU is not a barrierabove whicha fiscalexpansion will not raise output or lower unemployment;it is only a barriervis-a-vis the long-run sustainabilityof a demand expansion. In the short-run unemploymentmay fall below the NAIRU, but ultimatelyonly at the cost of risinginflation.Our model of supply behaviour is compatible witha varietyof storiesabout short-runbehaviour. In particularit is quite consistentwithan approach that has been especiallypopular in France which treatswages and prices as fixed in the short run and admits rationingin goods and/orlabour markets(see Artus,Laroque and Michel, 1984; Lambert,Lubrano and Sneessens, 1984; Bouissou, Laffontand Vuong, 1984; Malinvaud, 1986). Such models address the different question of whethera demand expansion would be successful at reducing unemploymentin the shortrun irrespectiveof any effect on inflation. Figure 2 depicts a simple diagram of the labour market. Firms' demand for labour will,for a given capital stock,be lower the higher is the real cost of labour. However the real cost of labour differsfrom real take-homepay of workers(W) for three reasons. First,whereas workersdeflatenominalwages by consumer prices,firmscare about the

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

282

Sachs and CharlesWyplosz Jeffrey

level of wages relativeto the productprice. Second technicalprogress which increases the productivityof labour for a given capital stock reducesthe effective costof labourand increasesthe demand forlabour even whentake-homepayis unchanged.Finallyemployerand employee taxes drive an additional wedge between take home pay and labour costs. Figure 2 shows a labour demand curve DD for given levels of capital,technicalknowledge,taxes, etc. An increase in taxes, or lower level of capital or technicalproductivity would reduce the demand for labour at any given level of take-homepay, shiftingthe curve to D'D'. Econometricevidence suggests,at least for primarywage-earners, that real take-homepay has littleeffecton labour-forceparticipation. Hence we draw a verticallabour supply schedule at L in Figure 2. A major theme of our discussion,however,is that if wages are set by unions, ratherthan by competitivelabour markets,the union 'supply' schedule mightin factbe quite elastic5.In fact,one importantspecial case derived in Appendix B showsthata rationalunion mightchoose to fixthe level of real take-homepay at a level such as Wu,no matter what the implicationsare for the employmentlevel. In other words, the union 'supply' schedule would be perfectlyelasticat the wage level Wu,as shown in Figure 2. Now letus considersome alternativescenarios.Beginningwithlabour demand DD, fullemploymentin a competitivelabour marketis at point F, withreal take-homepay W and employmentL. Now suppose instead thatthe labour marketis fullyunionized,with unions demanding real take-home pay level Wu. The union-based equilibriumwillbe at the point U. Unemploymentis given as L- L", and the (proportional)deviation of the actual union wage Wu fromthe level warrantedby fullemploymentis (Wu - W)/W. This deviationof the actual fromthe warranted wage is called the wagegap. Suppose now thatpayrolltaxes are raised, so thatthe demand curve shiftsto D'D'. The full-employment wage fallsto W'. The new equilibrium is given by U' in Figure 2, with employmentLu and real take-homepay fixedat W". Note thatthe verticalshiftof the demand curve, and hence the increase in the wage gap equals the amount of the tax increase. In this union-based theoryof the NAIRU, the 'equilibrium'rate of unemploymentdepends on the size of the wage gap and the slope of the labour demand curve. (Obviously,this formulationabstractsfrom 5Although,as a monopolisticsupplierof labour,the union does not have a supplyschedule that is independent of demand, it is expositionallyconvenientto maintainthis fiction.A rigorous analysisis given in Appendix B.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

283

purely frictionalcomponentsof the NAIRU, say due to job turnover or search; we assume thatsuch frictionalcomponentsare constant,for reasons that we outline below.) When the demand curve is DD the NAIRU is then determinedas U* = L-L". Factorswhichshiftthe labour forceL, the demand curve DD, or the targetwage Wu shiftthe NAIRU. Primecandidatesforsuchshiftfactors include the wedge, the capital stock,and the stateof technicalknowor a ledge. A fall in the capital stock,a fall in technicalproductivity, rise in the wedge would both raise the NAIRU and raise the wage gap. Our empiricaldiscussionthatfollowsexamines these presumptions. In our interpretationof the Phillips curve, we assume that union wage-settersgradually reach their target employmentlevel LU and targetreal wage level WU. In particular,we assume thatunions increase theirreal wage demands wheneveractual employmentL exceeds their targetlevel Lu, and lower theirdemands when L is less than LU. We thereforespecifythe nominalwage change 7rwas a functionof expected or lagged inflation r-1,and the deviationof the unemploymentrate fromthe NAIRU, U-U*: w = a(U-

U*)+

r_-

(1)

Price inflation r7 is closely related to wage inflation mwand the precedingdiscussionof whatdeterminesthe NAIRU, U*, coupled with our hypothesisabout how unions gradually adjust actual wages and employmentto targetlevels, allows us to derive (in Appendix B) the empirical Phillips curve which forms the basis of our econometric estimatesin the next section. The key noveltyin thisapproach is thatthe model underscoresthe factorsthat shiftthe NAIRU, and in turn cause shiftsin the Phillips curve. Variables that shiftthe NAIRU should also shiftthe wage gap. Thus, in examiningthe evidence for the 1970s and 1980s, we expect to findthatboth the wage gaps and the NAIRU have risen in France. Indeed we findstrongevidence forboth of these trendsafter 1973. 3.2.1. Wage gaps. Wage gaps have been calculated forFrance's manufacturing sector by Bruno and Sachs (1985), Artus (1984) and Bruno to obtainas theyrequire,in principle, (1986). Such measuresare difficult theestimationof productionfunctions,fromwhichto deduce the slopes of labour demand schedules. Our intentionhere is to update previous estimatesand to measure wage gaps forsectorsotherthan manufacturing, as the manufacturingsector represents a small, and secularly declining,portion of France's output (22% in value added terms in 1984, down from26% in 1963).

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

284

Table 5. Wage gaps (% of labour costs, base years 1976-78)

Manufacturing 1976-78 1980 1981 1982 1983 1984

0 1-7 4-9 3-9 4-12 6-17

Building and Construction

Retail and Wholesale Trade

0 1-5 3-7 2-8 2-9 2-10

0 1-5 0-6 0-10 2-12 4-14

Notes: For details of constructionsee Appendix A. Range of estimates assume an elasticityof substitutionof between 1/2 and 1 and allow for eitherHarrod or Hicks-neutraltechnicalprogress.The base years correspond to an average unemploymentrate of 5%.

Contraryto the other worksmentioned,we do not use econometric estimatesto measure the wage gap. Ratherwe specifyalternativeparametersof an assumed productionfunction,and finda resultingrange of estimatesforthe gap. The procedureis fullypresentedin Appendix A and only brieflyexplained here. First,by observingvalue added, factorinputsand factorshares,we directlymeasure the rateof technoto be labouraugmenting, logicalchange (whichis assumed,alternatively, or both capital and labour augmenting).The production functionis then fully characterizedby assumingthe elasticityof substitutionand measuringthe labour share in the base period. The derived labour demand is then computed (this is the DD curve in Figure 2) which allows us to findthe warrantedreal labour cost,i.e. the level at which firmswould hire the amount of labour correspondingto fullemployment.A baselineof sectoralfullemployment,in turn,is estimatedgiven an aggregate full employmentlevel and the trend in inter-sectoral employmentshifts. We presentin Table 5 a range of estimatesforthe wage gaps in three sectors: the manufacturingsector, the building and construction industry,and retailand wholesaletrade. The range of estimatescorresponds to the extremevalues obtainedallowingthe slope of the labour demand curve to varywithinreasonable bounds and forboth general and labour-augmentingtechnologicalprogress.In all cases, the chosen base period (definedas 'fullemployment')is 1976-78 when unemploymentaveraged 5%. These numbersare meantto answerthe followingquestion: by how much should labour costsbe reduced in a given sectorand in a given year to prompta demand for labour correspondingto an overall 5% unemploymentrate? The table shows large numbers (around 10%), with some differencesacross sectors. These numbers do not differ

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

285

markedlyfromthose computedforthe manufacturingsectorby Artus, Bruno and Sachs. 3.2.2.The NAIRU. In view of the traditionallink of the NAIRU with 'frictional'unemployment,it mightseem fruitfulto tryto explain its rise withvariablesthat mightshiftthe frictionalrate. Such candidates include: demographicshiftsin the labour force,such as a risingproportion of young workers,who have historicallyhad high rates of unemploymenteven when the aggregateunemploymentrate has been low; job mismatch,as evidencedbya shiftin therelationshipbetweenvacancy and unemploymentrates; regional and industrialmismatch,as evidenced byan unusual rateof structuralchange or compositionalchange in the labour force.A large numberof studieshave triedto trackdown these possible culprits,and the resultshave been disappointing(see Layard etal., 1984, and the recentcountrystudiespresentedat an LSE conferenceon the Rise in Unemploymentincludinga studyon France by Malinvaud). The inabilityof standard frictionalvariablesto account formuch of the increase in the NAIRU in Europe has led us to emphasize explanationsdrawingdirectlyon themodel ofthefirm-union bargaindeveloped in the previous section,and formallyset out in Appendix B. Thus we expect the NAIRU to increasewhen the unions' targetwage increases, when productivitygrowthslows,and when the wedge increases. In this section,we verifythat such a mechanismhas been at work. Recent studies have shown thatFrance, like mostEuropean countries, has suffereda rise in the NAIRU in the 1970s and 1980s (See Grubb, Layard and Symons 1984, Grubb, Jackmanand Layard, 1982, 1983, Layard etal., 1984, Coe and Gagliardi,1985). The significant European exceptionsappear to be Italy and Austria,and the remainingnotable exception in the OECD is the United States.6 To estimate these effectswe use an equation which generalizes Equation (1). The NAIRU depends on the wedge x, technicalproductivity(/, and the targetwage W'. We cannot gather data directlyon the targetwage, and we adopt the simple device of proxyingit witha time trend t. The wedge (the differencebetween real take-homepay and real labour cost to the firm)depends on labour taxes and on any I I 6

The followingestimatesof the NAIRU are taken fromCoe and Gagliardi:

1967-70 1981-83

Germany

France

UK

Italy

Austria

Netherlands

US

0.7 3.6

2.2 7.7

2.1 9.4

7.5 5.4

1.1 2.4

3.0 8.7

5.4 6.1

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

286

Sachs and CharlesWyplosz Jeffrey

Table 6. The Phillips curve (Dependent variable: change in inflationrate, AwT, annual data, 1963-84)

Variable constant unemployment(u) lagged change in relative importprices (A(pm-p)_1) minus wedge productivity (4 - x) lagged change in inflation rate (Air-) lagged error (e1) timetrend (t) dummyvariables:DV1 DV2 Standard error (percentagepoints)

Regression1 estimated t coefficient statistic

Regression2 t estimated coefficient statistic

205.64 -4.16

(3.77) (4.86)

251.51 -4.75

(6.69) (8.41)

0.07

(2.33)

0.06

(2.81)

-42.57

(3.76)

-52.05

(6.67)

-0.31

(4.03)

3.23

(7.47)

2.01

(8.61)

-

-

-0.58 2.76 1.16

(3.29) (4.27) (2.48) 0.98

0.60

Notes: (a) p and pm denote GDP deflatorat factorprices and importprice deflator respectively.Lower case lettersdenote logarithmsand r= 100 (p-p-_). $/ is the log of productivity(adjusted for capital deepening) and x the log of the wedge (see Appendix B). (b) Dummy variables are defined thus: DV1 takes the value minus unityin 1964, 1965, 1974, 1976, and 1983, the years in which severe wage or price controlswere in force,and otherwisetakes a zero value. DV2 additionallytakes the value plus unityduringthe years 1963, 1973, 1975, and 1982 to capture the effectof anticipatedtighteningof price controls.

discrepancybetweenthe pricesrelevantto firmsand to workers.Since income tax rates change littleover our sample period, we simplifyby omittingincome tax ratesincludingonly employerand employee payroll taxes and the ratio of importto domesticprices. Ratherthan assume thatunions continuouslyattaintheirtargetreal wage, we postulatesluggishadjustmentof actual to targetreal wages. In termsof our estimatedequation, this introduceslagged values of some variables to capture this sluggishadjustmentof actual to target real wages. Finally,we includetwoalternativespecifications ofa dummy variableDV to captureperiodsin whichunusuallyseverepricecontrols temporarilydistortedthe wage bargainingprocess. Having examined several regression equations for 1963-84 with annual data, we show our preferredequations in Table 6. They differ primarilyin the definitionof the price control dummyvariable (see Footnote (b) to Table 6). Regression(2) uses a dummyvariable which allows anticipatedas well as contemporaneouseffectsof severe wage and price controls and produces a better fit,though we also show

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

287

Table 7. Actual unemploymentand the estimated NAIRU (%)

Actual NAIRU:

Equation (1) NAIRU: Equation (2)

1963

1973

1978

1980

1981

1982

1983

1984

1.2

2.7

5.3

6.4

7.3

8.1

8.3

9.7

1.1

2.9

5.4

6.9

7.4

7.7

8.3

8.9

1.2

2.9

5.4

6.8

7.4

7.7

8.3

9.0

estimatesfor an equation in whichthe dummyvariable only includes contemporaneouseffects(Regression(1)). The choice of dummyvariable leads us to choose a slightlydifferent dynamicspecificationin each the but two case, reportedequations are verysimilarin theireconomic implications. Both equations show the crucial significanceof the level of productivity4fand the wedge x. In fact,it is an implicationof our theorythat the two variablesshould have equal coefficients but opposite signs,an empiricalpredictionwe testedbeforecombiningthese two variablesin Table 6. The time trend and eitherspecificationof the price controls dummyhave the expected effect,and changes in importpricesrelative to domestic prices also affectthe wage bargain. Other results are we could not detecta role, jointly reportedin Appendix C. Specifically, or severally,forthe followingvariables:changes in the unemployment rate,the ratio of unemploymentbenefitsto wages, the level of unemploymentbenefitsor of the minimum real wage, and the ratio of vacancies to unemployment(a crude indicatorof mismatch). As we explain in Appendix B, the equations of Table 6 can be solved to derive an estimatefor the NAIRU. Essentiallythis involvesusing theequation to determinewhatvalue of unemploymentwould generate no change in inflation,given other variablesin the equation. Table 7 shows the values of the NAIRU thus derived. Both regressionsgive verysimilarestimates.The NAIRU has increased steadilyover the last 15 years,withmarked jumps in 1974 and 1980-81. The moststriking implicationof Table 7 is thatactual unemploymenthas been veryclose to the NAIRU, at least until 1984. Our procedure cannot distinguish betweentwo competingexplanationsof thisfinding:first,thatmarkets adjust veryquickly,and, second, thatpolicymakers,being sensitiveto acceleratinginflation,have chosen or been forced to adjust demand and employmentto levels correspondingto the NAIRU. The role of the differentfactorsin the evolutionof the NAIRU can be considered using Regression (1). During 1964-73, productivity growthwas strongenough to absorb both a growinglabour forceand

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

288

Table 8. Changes in unemploymentand NAIRU (percentage points)

Change in actual unemployment Change in NAIRU contributionof: slowdown productivity wedge increase others

1973-80

1981-84

+3.7 +4.0

+3.3 +2.0

+1.6 +1.3 +1.1

+0.9 +0.5 +0.6

Note: NAIRU decomposition based on Regression (1) of Table 6.

expectationsof risinglivingstandards.Had productivity growthduring 1974-84 remainedat the 1963-73 rate (in otherwords,5.1% per year, ratherthan the actual 2.7%) the NAIRU would have been reduced by 2.5% in 1984 (Table 8). During the same period 1974-84, the wedge rose from41 to 69% of net take-homepay, contributinganother 1.8% to the overall 6.0% increasingin the NAIRU. The restis eitherunexplained or captured by the time trend. The time trend may be interpretedin severalways.Firstit picksup risingexpectationsof increased standardsof living,so thatthetargetwage setbyunions growsirrespective of actual productivityadvances. Second, it may also representa gradual increase in labour marketregulation.Third, there have been severaldemographicfactorsincreasingthe labour force,includingthe postwarbaby boom and the entryof women. Finally,there is a trend towardsreduced workingtimethatwe discussbelow. 3.2.3. The effectsof reduced workingtime. In January 1982, two changes

were introduced:the legal workweekwas reduced from40 to 39 hours, withan objectiveof 35 hours by 1985, and a compulsoryfifthweek of paid vacations.As alreadynoted,hourlywage rateswere raised to leave workers'yearlyearningsunchanged. While some reductionsin labour taxeswere grantedon a selectivebasis (smallfirms,or large firmswhich hired new workers),labour costs neverthelessincreased. It was the official viewthatreductionofworkingtimeshould lead to moreemploymentthroughwork sharing. The argumentthat employmentmightrise has been articulatedin the French case by Oudiz, Raoul and Sterdyniak(1979). It explicitly recognizesthatthe desired job creationrequiresthathourlywages do not increase and argues that the actual measues and the foreseen reductionof the legal workweekto 35 hours would promptplant-level reorganizationswhichshould in factreduce labour costsand raise the overall demand for labour. This would be the case, for instance,if a

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterranid The economicconsequences ofPresident

289

plantwere to use two shiftswhereit formerlyused one. But apart from exceptional cases, the reductionof the workweekis likelyto produce rathermarginalchanges in the plant level organizationof labour. It is more plausibleto assume that,giventhe existingequipmentand lay-out of most plants, operating less hours with more people will be more costly,if only because if it had been in the firms'best interestto do so, theywould not have waited for the mandatorychanges to introduce the new system. A formalanalysisof the issue (see Calmfors,1985; Wyplosz, 1985) in the frameworkof the union wage model showsthatif workerscare a lot for leisure,theywill accept a reductionin theirearned incomes in exchange fora shorterworkweek.Their unions would thengo along withno, or little,hourlywage increaseas itwould givefirmsan incentive to maintaintheirdemand formanhoursand increase the chances that work sharing occurs. If, in addition, there are no significantcosts of producingwithmoreworkersworkingshorterhours,thenemployment mustrise. The evidence reported so far is that work sharingdid not occur in France in 1982. We know thathourlywage rateswere not reduced so thatlabour costsincreased.And thereis no evidenceof large scale plant re-organizationwhichwould have broughtabout benefitsfromoperating with more workers.Thus the two necessaryconditionswere not satisfied.Attemptsat measuringthe effectdirectlyconclude thatit was, at best,verylimited.For example,Colin, Elbaum and Fonteneau (1984) conclude that some 25,000 jobs were created in 1982, withno further effectsthereafter.This representsabout 1.25% of the unemployed population and 0.12% of total employment.If we consider that the weekof paid vacationsreduced reductionin workinghoursand thefifth the numberof hours per workerby about 4.5% (2.5 and 2.0%, respectively),we conclude that these measures reduced the total volume of manhoursbyabout 4%. As actual manhoursfellbyabout 2%, itappears that these 'work sharing' measures in and by themselvesmore than offsetthe effectsof all the other expansionarymeasures enacted in 1981-82. 3.2.4. Appraisal. This section has produced evidence that labour costs

have increased steadilyover the past 10 years,resultingin 'wage gaps' of about 3% in 1973 to 7% in 1981. The 1981-82 expansion, coupled with measures which furtherincreased labour costs, could not have succeeded in creatingjobs in thecompetitivesectorin a non-inflationary way.Withthe actual rate also around 7%, Mitterrand'selectoralpledge to reduce unemploymentquicklycould not, in any case, be fulfilled withoutan accelerationin prices.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

290

Sachs and CharlesWyplosz Jeffrey

Our estimationof the NAIRU (Table 7) showsthata 10% reduction in labour costs (via increased productivity or a reductionin the wedge) would lower the NAIRU by a littlemore than 1%. This would require labour costs to fallby about 50% in order to reduce the NAIRU to its 1976-8 level! There is a large discrepancybetweenthe implicationsof thewage gaps and thoseoftheNAIRU. Clearly,measurementproblems may be part of the explanation. Yet there is another one. Given the manylabour regulationsand the costsof alteringthe size of the labour of demand force,itmaybe thatin theshortrunthe firms'wage elasticity for labour is quite small7. In this case, the Phillips curve estimates whilethe wage gap estimatescapturethe capturea short-runelasticity, of demand to wages. Thus, we suspect that labour long-runresponse the long-runfall in the NAIRU followinga rise in productivityor a fallin the wedge is probablygreaterthan is suggestedby the estimates of the Phillipscurve in Table 6. However,we have not yetinvestigated the natureof union behaviourunder the assumptionthatthe long-run labour demand elasticityexceeds its short-runvalue. 3.3. The confidence factor

No matterhow misplaced,thepolicymeasuresof 1981 wereofmoderate magnitude.This standsin contrastwiththe verystrongreactionof the exchange marketsto the change in governmentin 1981. We therefore offera third hypothesisto explain whythe Socialistsfailed to sustain theirpolicies,namelythattheyneverinspiredmuch confidenceamong the investorsand privatecorporations,and thatwhatevergoodwillthey may initiallyhave relied on was quicklyshattered. To evaluate the role of thesefearsin the deteriorationon the French situationbetweenMay 1981 and March 1983, we focuson two markets where these expectationsare likelyto have been most powerful;the stockmarketand the exchangemarket.Figure 3 showsa monthlyindex of industrialshare prices deflatedby an industrialgoods price index and theone-monthforwardexchangeratediscountoftheFrancvis-a-vis the Deutschmark (the percentage forward discount is stated at an annual rate). In principle,the firstvariablereflectsthe marketexpectation of futurediscountedprofits,while the second is a crude measure I

I

7 The estimatesin Tables 6-8 suggestan elasticityof 0.1 of the NAIRU withrespectto (41-x). Productionfunctionestimates(see Appendix B) suggesta value r/(1- SL) where r is the elasticity of substitutionand SL (=0.66 in 1984) the share of labour in value added. Even an elasticityof as low as 1/6would suggesta long run NAIRU elasticityof 0.5, considerablyabove substitution the shortrun estimategiven above.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

291

Mitterrand The economicconsequences ofPresident

I_ 1

discount ^QTForward

/

\

d 0

0.8 -

Source:

I

.

Data

:

0.6-

Ban&

:

.

May81

i

10

E

i March83

June82 0.4

Ill

74

llllllt

75

lll

76

77

78

79

u30

80

81

82

83

84

85

Figure 3. The real share price and the forward discount Source: DRI Data Bank & IMF

of expected futureexchange rate depreciation(a negativevalue indicates an expected depreciation). Three subperiods appear. In May 1981, there is a clear and sharp fallin bothmeasuresfollowedbya hecticevolutionaround a downward trenduntilJune 1982. Thereafterthe real share price remainsflatup until early 1983 where it entersa thirdphase withan upward trend. By mid 1985 it is back at its 1980 level,thoughsome 50% below itspre oil-shocklevel of 1973. The patternof the forwarddiscountis similar, in the intermediateperiod. The three subexcept for more voltatility 1981-85 withto the electionof Mitterrand commence periods during in May 1981, the second devaluationaccompaniedbythe firstrestrictive measures in June 1982, and the definiteturn to austerityat the time of the thirddevaluationof March 1983. In principle,we would liketo be able to saywhetherthe sharpchange observed in May 1981 and untilMarch 1983 is a rationalanticipation of whatwas to follow(indeed corporateprofitsfell,severalcorporations were nationalized8and the Franc was devalued) or whetherit reflected exaggeratedfears.In the firstcase, bybringingforwardfutureadverse developments correctlyanticipated,the financialmarketsfulfillthe paid (see Langohr and iallet, 1986). 8 Though compensationwas Though compensationwas paid (see Langohr and Viallet, 1986).

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

292

Sachs and CharlesWyplosz Jeffrey

useful role of reducing possible misallocation of resources, both nationallyand internationally.In the lattercase, a panic undermines whatcould have been sensiblepolicies.Unfortunately we are unable to distinguishthese two views so that the objective of this section is the more limitedone of documentingthe crisisof confidence. 3.3.1. The stockmarket.Share pricesare verymuchexpectationalvariables,

incorporatingall available informationabout future movementsin profitsand interestrates. Changes occur mainlybecause of the arrival of new information.In order to organize the evidence, we therefore examine whetherquarterlychanges in this variable are explained by changes in the factorspresumed to have an influence:monetaryand fiscalpolicy,real interestrates,real wages (which affectprofits),and importprices. We obtain the followingregression: Aq = 0.03-7.21 Aw- 0.04 AB + 1.70 Am- 0.08D1 + 0.06D2 (1.43) (3.23) (1.74) (1.24) (2.07) (1.45)

(2)

Sample period, quarterly,1974Q1-1984Q1; where t statisticsin brackets;standarderror= 0.08; DW = 1.92 Aq = quarterlygrowthin real sharepriceindex (nominalindex deflated by index of industrialgoods' prices), Aw =quarterly growthin real labour costs,AB = quarterlychange in public deficitas a proportionof GDP, Am= quarterlygrowthin thereal moneystock(M2). Thus higher real labour costs,a larger public deficit,and a lower real money stock tend to reduce real share prices.We also include two dummyvariables in Equation (2): D1 takes the value unityfromMitterrand'selectionto the major policy reversalof March 1983, zero elsewhere; D2 is unity thereafter,zero elsewhere.D1 documentsthe initialconfidencecrisis, whilst D2 provides some evidence for a subsequent restorationof confidence. Thus, Mitterrand'selection had two major effectson share prices. Firstthe directconfidenceeffectwhichalone accountsfora fall of 8% during the initialperiod when share prices actuallyfell by some 20% in May-June1981. Second, Equation (2) impliesthateach 1% increase in real labour costsreduces equitypricesby 7%. Since we have argued to real that Mitterrand'slabour market policies added significantly labour costs,thisreinforcedthe fall in share prices. Nor are these effectsunimportant.Reductionsin real share prices, by hittinginvestmentspending,mayhave weakened the initialstimulus to aggregate demand. To investigatethis possibility,we estimatethe

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

293

followingregressionon quarterlydata from 1976Q1-1984Q4.9 i =0.24+0.02q -0.17r+0.18A4y+0.87i1 (2.72) (2.00) (1.67) (2.21) (17.31)

(3)

t statisticsin brackets;standarderror= 0.02; where: q = (logarithmof) the real share price, i = (logarithmof) the share of investmentin output, r=(ex post) real interestrate, 4y= annual growthrate of corporateoutput. This regressionconfirmsthe usual investmentequations of French macro-modelswhich stressthe acceleratoreffectof growthand of the real interestrate. What we add is the role of stockprices,and through The them, the impact of market expectations and of profitability. the the effect of to measure when simulated 40% drop in regression, a real share prices during the early Mitterrandperiod predicts fall of 0.7% in the ratio of investmentto output as compared withan actual fallof 1.7%. These crude computationssimplyillustratethe plausibility of the storywhichlinksadverse expectationsin financialcirclesto a fall of stockprices,and thento a depressingeffecton investmentspending. Ignoringthese links,as do the usual French investmentequations,led to unpleasant surprises,as the governmentwas obviouslybettingon theresumptionofgrowthto boostinvestmentspending.The investment recoverynever occurred. 3.3.2. Capital flows and the EMS constraint. The other gauge of confidence

is the internationalmonetarysituation.We already noted the high correlation between share prices and the discount on the forward exchange rate. Capital outflowsare anotherchannelof interestbecause theymay exacerbatethe pressurefaced by the monetaryauthoritiesas theyattemptto maintaina given paritywithinthe EMS. In order to assess the role of capital flight,a certainnumber of precautionsmust be taken. First,because exchange controlsare widespread in France (see Claassen and Wyplosz,1985) and have been strengthenedin June 1981 and furtherconsiderablytightenedfromMarch 1983 to January 1984, capital flightis observed both in the shortterm capital account and in errorsand omissions.Second, both directlyand throughstateowned corporations,the Governmenthas been able to borrowabroad when private outflowsswelled. This is why we show in Figure 4 the sum of the short-termprivate capital account and the errors and omissions,labelled the speculativecapital account. 9 Recognisingthat q is endogenous, we have re-estimatedEquation (2) using 2SLS, but obtained practicallyidenticalestimates.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

294

Sachs and CharlesWyplosz Jeffrey

2

Iq!,l, !,

'

,!

SI,

-2 -

~/ V| 81:2 -4

i I I I I 80

E"s

82:2 I I I I II 81 82

I I 83

-~

Current account Speculativeaccount

83:1 I i I 84

II I 85

I

Figure 4. The currentaccount and the speculative capital account Source: Banque de France

The quarterlyfigures,the only ones available, probablyunderstate the size of outflowsimmediatelypreceding a devaluation, since the outflowsbecome inflowsin the aftermathof the devaluation,or when the expectationsturnout to be unwarranted.Thus, the quarterlydata tend to smoothout larger short-term swings.Yet, we observe massive outflowsin 1981 untilthe firstdevaluationin October,a refluxdelayed untilthe second devaluationofJune 1982, quicklyreversedbythe crisis in the firstquarter of 1983 whichled to the thirdand last devaluation in March. As in thecase ofstockprices,capitaloutflowsand theforwarddiscount may well have been rational anticipationsof futurecurrentaccount or else theymighthave reflected deficitsand devaluations,respectively, a componentof irrationalfear.What is importantis thattheyall point towarda significantworseningin confidence. The role of externalpressurewas extremelyimportant,since there is good reason to believe thatFrench commitmentsto the EMS tipped the balance towards austerity.Unlike the much looser commitments under the European Snake in the 1970s, whichFrance abandoned on twooccasions,membershipoftheEMS has been investedwithenormous politicalimportanceat the veryhighestlevels of the government.That is whythe debate over leavingthe EMS was treatedas synonymouswith the debate over abandoning other spheres of cooperation in Europe,

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

295

including participationin the Common Market. Apparentlyfew participantsin the debate opted foran intermediatestrategyof continued free trade, but witha floatingFranc outside the EMS. Thus, the episode is an importantcase study of the possibilityof applying international monetary agreements to restrain domestic policies.Economistshave long debated whetherinternationalexchange rate agreementscan reallybind national policies, since the sanctions forbreakingagreementsare so small and diffuse.The example of the Socialistturnaroundin March 1983 suggeststhatan internationalagreementcan help to tip the balance towardsdomesticrestraint. Perhaps a more interestingaspect of the EMS constraintis that it representeda way for one politicalpartyin France to bind the future actionsof another.Afterall, Giscardtook France intothe arrangement in 1979, but the constraintunder the arrangementreallyonlybecame exerciseda strong bindingunder Mitterrand.Thus, Giscard effectively vote on French policy long afterleaving office.This form of future control might be an argument for the Conservative Party under Thatcher to lead the UK into EMS membership:not to bind its own austere by itself,but ratherto bind the behaviour,whichis sufficiently actions of a successor government,which is likely to be far more expansionary. Note thatthe EMS could one day servethe interestsof the Socialists by preventingoverlycontractionarypolicies of a right-winggovernment,since excessiveappreciationcarriesresponsibilitiesforintervention and adjustment as does excessive depreciation (of course, the principalburden of adjustmentfallson the countrywiththe depreciating currency). 3.4. Summing-up

The Socialist spending expansion added a verymodest 1.5% of GDP to totaldomesticdemand, and in theend, employmentdeclined outside the non-competitive and public sectors.We agree thatthe unexpected world recessionin 1982 played an importantrole in underminingthe Frenchattemptat unilateralexpansion,yetthisis onlya partialexplanation of the policy failure.Confidencewas shaken, and we found that thisalone accountsfora fallof privateinvestmentspendingrepresenting 0.7% of GDP. Coming on top of the 0.9% of GDP lost in exports because of the world recession,we see thattherewas never much of a net demand stimulus.At best therefore,the monetaryand fiscalpolicy mix of 1981-82 was enough to keep France out of the world recession. The resultingcurrentaccount deficit,togetherwithconfidence-related capitaloutflows,createdthe conditionsfora balance of paymentscrisis.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

296

Sachs and CharlesWyplosz Jeffrey

In the end, Mitterrand'schoice to remainwithinthe EMS lefthimwith no other choice than to reverse his policy stand and join France's partnersin austerity. The strikingelementof 1981-83, datingback to the earlyseventies, was the simultaneousproblem of high unemploymentand continued stronginflationary pressures.The NAIRU had risen sharplyover the 1970s, and continuedto rise in 1981 and 1982. The high level of the NAIRU was accompanied by large and growingwage gaps. The proximate cause of the risingNAIRU was the combinationof slow productivitygrowthin France and a continuingincrease of real labour costs, due to higherlabour taxes, the secular growthof real take-homepay, and cost-raising measuressuchas theshortenedworkweek.Hence, even withoutthe problemsinduced by a world recessionand a collapse of confidence,Mitterrandhad littlescope in 1981 to embarkon a sustained demand expansion withoutquicklygeneratingacceleratinginflation. 4. Socialist austerity:a success story Hesitantlybetween June 1982 and March 1983, and more boldly thereafter,the Socialistgovernmenthas completeda turnaroundin its economic policies.The fightagainstunemploymentis no longerat the forefrontof the government'sobjectives. The key words are now: external balance, modernizationand rationalizationof the economy, and wage moderation.By thetimeofthe March greatercompetitiveness 1986 election,therefore,Mitterrandwillhave presided over two years of expansionarypoliciesand threeyearsof austerity.In thissectionwe present an overviewof these new policies and an early assessment, concluding that the governmenthas successfullyfought inflation, restoredthe currentaccountbalance, made some progresson the wage front,but has not yet grasped the full importanceof the supply side. 4.1. The external balance target

The decisionto stayin the EMS meantthat,unless the currentaccount was quicklybroughtback into balance, a new crisiswould erupt. Such an objective was stated explicitlyby Mitterrand,and was part of the agreementreached withGermanyat the timeof the currencyrealignment.Assigningtasksin 1983 for Mauroy's new government,Mitterrand asked him to 'bringdown inflationto a rate compatiblewithour competitors... restore within two years our external trade balance... and respectthe financialequilibriumof Social Securitywhile maintainingthe publicbudgetdeficitwithinitscurrentlevels' (Le Point, March 28, 1983).

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

297

This is exactly what happened. The budget deficitwas stabilized around 3% of GDP. This was achieved by freezingbudgeted expenses and by raising taxes: a 1% additional income tax to finance Social Securityand a compulsorysubscriptionto publicborrowingrepresenting 10% of 1982 income taxes. The OECD measuresofstructuraldeficitconfirmtherestrictive stance of fiscalpolicy after 1983 (Table 1). Indeed, while the actual deficit rose slightlyfrom 1983 to 1984, the structuraldeficitwas reduced by 0.5% of GDP. Monetarypolicyalso remained contractionary, although the squeeze has come from a differentchannel. While previous monetarycontractionswere achieved throughbinding credit ceilings and (relatively)low interestrates,since 1983, and more so since 1984, interestrates have been kept so high thatthe creditceilingshave not been binding. The role of the exchange rate is more difficultto assess. With an withGermanyrunningat about 4%, and assuming inflationdifferential that the Franc-Deutschmarkparitywas satisfactoryafter the March 1983 devaluation,this would implya gradual overvaluationreaching about 8% by mid-1985. But the Franc-Deutschmarkparityis not the only indicatorof France's competitiveness.An effectiveexchange rate index also includes the dollar and reflectsits appreciation. From the end of 1980 to the end of 1984, the dollar appreciated in real terms vis-a-visthe ECU by about 50%. Againstthe dollar the ECU was undervalued, while within the ECU, the Franc was overvalued and the Deutschmarkundervalued. Viewed this way,the improvementof the currentaccountin 1983 and 1984 was helped bythedollarappreciation. Thus, 1983-84 in many respects looks like the mirror image of 1981-82. A standard Keynesianexpansion was replaced by a standard contraction,while the rest of the world pulled itselfout of recession. In 1982, final domestic demand in France grew by 3.9% while it remained flatin the OECD zone. For 1984, the numbersare 0.5% for France and 5.1% forthe OECD. This reversalmusthave been a major factorbehind the balancing of the currentaccount in 1984 and 1985. in themeantime the situationis notperfectly symmetric: Unfortunately, both the public debt (see Table 1) and the externaldebt have grown, on thebudgetand thecurrent therebycomplicatingthebalancingefforts account. Also, as noted by Fonteneau and Gubian (1985) the 1981-82 fiscalexpansion was based on permanentfeatures(new hiringsof civil servants)whichare not as easilyreversableas increasesin spending on goods. Debt service in 1984 represented4% of public spending, i.e. about 2% of GDP, a figurewhichcan be compared to the budget deficit of 3%. A returnto budgetbalance would now requiremore contraction of spending or fastergrowthgeneratinghigherrevenues. This leaves

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

298

the French government,presentand future,witha limitedmarginof freedomto use fiscalpolicy. The external debt situationis oftendescribed withalarmistundertones. Cumulated currentaccount deficitshave raised France's gross liabilitiesby some 8% of GDP between 1980 and 1984. But as of June 1985, France's external assets (export credits, foreign exchange the externaldebt liabilities.'0It is truethatmany reserves,etc.) exceeded of the French assets are knownto be held in LDCs and are therefore at risk. Yet, by all measures, the external debt situationcannot be considered as a major source of concern,as France's ratingson world financialmarketsamplytestify. 4.2. Inflation

The inflationdifferentialwithGermany has been more than halved between 1981 and December 1985. Two broad policies have been put to work:a setof directwage and pricecontrolsand the moretraditional strategyof demand restraint. While price controlsare partof France's tradition,the impositionof a wage freezeas earlyas July1982 was a greatsurpriseto publicopinion. This freeze was set for4 months.The policyadopted afterthe freeze is interestingas it illustratesone of the many effortsby European countries to achieve effectivede-indexation. There has never been formal cost-of-living adjustment in France but in practice,this has been the rule for a very long time. The existence of a de facto full indexationmechanismis confirmedbyall theeconometricstudies(Artus etal., 1981; Oudiz and Sterdyniak,1985; Feroldi and Meunier, 1984; and themean lag is foundto be betweenone and twoquarters).Similarly there is a strong,fast,and well-documentedeffectof wage increases on inflation. The procedure since 1982 is as follows.At the beginningof each year, the governmentissues inflationrate targetsand asks the firms and the trade unions to base theirwage increaseson thesetargets.This is simplya wish,backed by a carrotand a stick.The carrotis the fact that the governmentpledges to followthis targetfor its own labour force and the prices of public servicesand utilities.As for the stick, prices are not free but resultfromcontractsat the branch level with theauthorities(in theabsence ofagreementthegovernmentsetsprices). I

10Data

I

computed by the Ministerede l'Economie,des Finances et du Budget (Sept. 1985). Gross liabilitiesare definedas borrowingforoverone year; shorterliabilitiesare omitted.Grossassets exclude exchange reservesand are mainlyexport creditsor public loans. At June 1985, net externaldebt was estimatedas 5.2% of GDP.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

299

Table 9. Wage guidelines(%p.a.)

Official Target ActualIncreases: GrossHourlyWage GrossAverageSalary SMIC CPI

1985*

1982

1983

1984

10.0

8.0

5.0

4.5

15.4 12.6 17.6 11.7

11.2 10.1 12.2 9.7

7.6 7.6 9.4 7.5

5.6 5.6 6.9 5.6

Source: Les NotesBleues No. 251.

*Note:official forecast.

The interestingfeatureof this scheme is that it does not state that wages must be de-indexed. On the contrary,it specificallyendorses indexationbut ties it to futureinflation,not to past observations.The advantagesofa forward-looking indexingmechanismwhen disinflation is under wayare obvious.Withtargetssetat the beginningof each year, the 1 or 2 quartertraditionallag is now replaced (at least in principle) byan average lead of six months.Whiletherehave been some inevitable slippages above the targets,Table 9 shows thatwage growthhas been graduallyconvergingtowardprice inflationsince 1983. Because social bornebyemployeeshave risen,especiallyin 1984, securitycontributions thishas meant a reductionin net real wages (-1.8% in 1984). At the same time,unemploymenthas been risingquickly,ahead of our estimatesof a steadilyincreasingNAIRU. The interestingquestion is whetherthe systemof price and wage restrainthas added much, if anything,to the normal Phillipscurve mechanism.In its 1985 annual surveyof France, the OECD reportsan INSEE studywhichhas simulated itsestimatedwage-priceequationsfor1984 findinga 2% difference betweenpredictedand observedinflation,suggestingthatindeed wage and price controlsdid work.Our own estimatessupportthese calculations.When we re-estimateour Phillipscurve (Regression(1) of Table 6) over the period 1963-80 and generateforecastsfor1981-84, we find significantresiduals of, respectively,2.9 and negativeand statistically 3.6% inflationfor 1983 and 1984. 4.3. Labourcostsand employment policies

Withreal gross wages roughlyconstantin 1984-85, real hourlylabour costshave stillincreased by about 4% in 1984 because of a furtherrise in the wedge. As a result,our measures of wage gaps show a further worseningof about 3% in the manufacturingsector.Not surprisingly therefore,while employmentmeasured by the number of registered

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

300

Sachsand CharlesWyplosz Jeffrey

workersdid not fallin 1984, afterdecliningin the years 1981 to 1983, the reductionof effectiveman-hoursworked in the competitivesector has not yetcome to a halt. Thus it appears thatthe apparentlysuccessfuleffortat curbingwages has not been mainlymotivatedbythe desire to reduce real labour costs and promote employment,but by a concern to break the wage-price spiralin order to fightinflation.Here again we findin 1983-84 a mirror image of the policies followedin 1981-82: reducingreal wages is now meant to reduce demand, verymuch as the earlier wage hikes were stillthe dominatsupposed to raiseit.The demand side is unfortunately ing concern. A similarinterpretationapplies to employmentpolicies. In the face ofrisingunemploymentnumerouspolicieshavebeen designedtoreduce the number of registeredunemployed,withoutreally increasingthe numberof employedworkers;earlyretirementbenefits,trainingprogrammes,specialcontractswithfirmsfortemporaryon-the-jobtraining have been continuouslyexpanding in size, scope and cost. By 1983 thesemeasurestogetherwithunemploymentbenefitsrepresented3.4% of GDP. These numbers underestimatethe true cost as they do not include subsidized loans to corporations,but theyare large enough to prompta change of heart. In 1984, for example, the unemployment benefits system has been restructured,reducing the number of beneficiariesand the volume of benefits.Several trainingprogrammes are being scaled back. The civilserviceand state-ownedcorporations are no longer called upon to hire. There have been spectacularlargescale firingsin industriessuch as steel, coal and shipbuildingwhich have long been recipientsof public subsidies,as well as at state-owned Renault, long a symbolof generous social policies. The government itselfis now engaged in a processof trimmingits ranks. 4.4. Conclusion

Withobvious successesregardinginflation(the officialtargetof 5% for 1985 has been surpassed) and the external balance (with a current account surplus in both the second and thirdquarters of 1985), the major casualtyof the Socialists'policies is paradoxicallyemployment. The combination of demand expansion and supply side errors in in 1983-84, has resultedin a 1981-82, followedby demand restriction continuousrise in the unemploymentrate, in line withgrowingwage gaps and a growing wedge. Only in 1984 did it emerge that the Government had realized the importance of the supply side and profitability. At the time of writing,there is littlefactualevidence on which to base a theorythattheSocialistshave now movedthewholecircletowards

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

301

defendingthe virtuesof competitionin a freemarketeconomy.Prices are being gradually freed, but controls remain extensive. Foreign exchange regulationshave been eased but remainas extensiveas before May 1981. The long standing traditionof credit ceilings has been replaced by a market-basedmonetarypolicy relyingon management of bank reserves,but this change was probablymotivatedby the very large amount of unused creditbelow the ceiling (due to the weakness of demand and highreal interestrates)whichcould be absorbed should the economypickup. Financialmarketshave been allowed to growand withthe activesupportof tax incentives,and develop new instruments, stockpriceshave soared since 1983,butthebanksare stillalmostentirely state-ownedand 'selectivefinancing'(i.e. credit allocation by government) does not appear to have fallen.

5. Currentsituationand policy options In the thirdquarterof 1985, withunemploymentat 10.2%, the budget for 1986 was being finalizedaround the general objective of further disinflationand no relapse fromausterity.The argumentsbehind a strategyof continuingausterityare thatthereis no room forexpansion because: the external constraintwould again surely undermine any such attempt;the budget is stillin deficit,thus ruling out any fiscal policy-led recovery; monetary policy must remain tight in France because it is tightabroad, expeciallyamong EMS countries;the process of rationalizationand trimmingof French companies is not yetcompleted, and any expansion would relax the pressure to pursue this fundamentalstructuraleffort. We do not findthese argumentsconvincing,since we believe thatif demand stimulusis combinedwithappropriatesupplymeasures,many of the adverse consequences of a demand expansion could be avoided. We now sketchsome policyoptionswhichcan be enacted fairlyquickly and which,according to our analysis,could deliver sizeable gains in employmentwithonlymodestcostsin inflationor externaldeficits.The cornerstoneof the proposal is a reductionin labour costs designed to reduce the wage gaps and bringthe NAIRU down, achieved through a decrease of labour taxes and a social contract.This policy must be accompanied by suitablestrategiesforthe budget and forthe external constraint. 5.1. A marginal wedge reduction

A central conclusion of Section 3 is that any attempt at reducing unemploymentthrougha demand expansion alone would quicklyhit

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

302

Table 10. The wedge. (Ratio of employers' real hourly labour costs to employees' real take-homepay) 1963

1968

1973

1976

1980

1982

1983

1984

1.34

1.41

1.41

1.50

1.60

1.62

1.65

1.69

Source: INSEE, ComptesTrimestriels. Note: The wedge is (T1/T2)(PCIPV)where T, is the ratioof totallabour coststo the wage bill, T2 is the ratio of net to gross take-homepay when employees' social securitycontributionsare deducted, and Pc and Pv respectivelyare price indices for consumptionand value added. As in Table 6, we ignore income tax rates,whichchanged littleover the period.

the inflationthreshold,withthe attendantexternalcrisis.The immediate implicationis thatthe firstpriorityshould be to reduce the NAIRU and the wage gap. One way of doing thatwould be to reduce hourly eitherthroughnegotiatedconcessions, wages relativeto productivity, or through'surpriseinflation'.The required reduction,10 to 15% is such that the inflationsolution is untenable, while the reduction in hourlywages via concessionsis likely,by itself,to lead only to small progress achieved over a protractedperiod. A differentapproach is required. Our analysisof the Phillipscurve directsattentionto the wedge, the ratio of total real hourly labour costs to net real hourly take-home salaries.The evolutionof the wedge is shownin Table 10. Cuttingthe wedge at constantreal take-homesalaries will reduce real labour costs in the same proportion.Thus a 15% reductionin the wedge, which bringsitback slightlybelow its 1976-78 level,would probablyeliminate mostof the prevailingwage gap. The strategyof achievingwage concessions should be combined witha strategythatreduces the taxation of labour income. A 15% cut in the wedge would not immediatelyreturnthe NAIRU to its 1976-78 level of around 5%, since in the meantimeseveralother factors(rising social protection,an increase in the targetwage, new labour regulations)have taken theirtoll. Our estimatesshow that,in the shortrun,the NAIRU would onlyfallby a bitmore than 2%, which would leave it in the range of 6 to 7%. However, as we have stressed earlier, we believe that the Phillips curve estimatesof the NAIRU probablyunderstatethe long-termbenefitsof a reduced wedge. Most of the wedge correspondsto labour taxes used to financesocial securityexpenditures,whichat presentamount to almost25% of GDP (Table 2). A 15% reductionin the wedge, if achieved by an across-theboard reductionin payrolltaxes,would then contributeto a deficitin the social securitybudget of 3.75% of GDP, a very large amount to recoup through spending cuts. However, if the cuts in payrolltaxes

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

303

can be placed more on the margin,to apply mainlyto new workers, and not to existingworkers,the budgetarycosts of wedge reduction would be very substantially reduced." As a very rough example, if a 15% marginal wedge reduction succeeds in bringing employmentup by 5%, the shortfallin social securityreceiptswould amount to about 0.5% of GDP. It must be stressedthat in order to achieve the long-termbenefits from payroll tax reduction, and to get those benefitsas rapidly as possible,firmsmustbe convincedthatthe lowertaxes willbe sustained witha marginaltax reductionis that for several years. One difficulty firmsmay be able to manipulatethe systemto theiradvantage without creatingmore jobs in total. For example, faced witha subsidyon new workers,firmsmightbe induced to make dismissalsforthe purpose of subsequent hirings,or to create new businesses to replace, for legal purposes,existingbusinessactivities.One possibilitywould be to relate the reductionin payrolltaxes onlyto new workersthatraise the firm's total labour force above the size at the startof the new policy. Thus dismissalsby a given firmfor the sake of re-hiringthe same number of workerswillbe pointless.It mayalso be advisable to plan to make a gradual reductionin the level of labour taxes on existingemployees, so thataverage tax ratesare ultimatelyaligned withthe lower marginal taxes,thusalleviatingthe incentiveto close down one businessand start up another to take advantage of the marginalwedge reduction.

5.2. A Social Contract

A reduction in labour taxes will lead to a distributionalfightover sharing the benefitsbetween employersand employees. The union wage model suggeststhatunions willtryto capturea part,thoughnot the entire benefit,of tax reduction. To make certain that the tax intoreduced labour costsforthe firm, reductiontranslatessignificantly it is worthwhileattemptinga social compact betweenthe government, trade unions and employers'associations.We envisage an agreement to splitthe reductionin labour taxesaccordingto a formulaof one-third to the employees(in the formof highertake-homepay),and two-thirds to the employers,(as a reductionin labour costs). Note that initially, the employees'average benefitswould be verysmall,to the extentthat the tax reductionapplies on the marginonly. If the average tax rate is I

I Our proposal for a marginal employmentincentivediffersfromthat of Steinherrand Van Halperen (1983) in three respects. Our proposal is designed to be permanent rather than temporary;it acts via a reductionin the wedge ratherthan an explicitemploymentsubsidy; and we do not insistthatitsbudgetaryincidenceis offsetby changes in unemploymentbenefit.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachsand CharlesWyplosz Jeffrey

304

reduced over time, the average rise in wages would become larger. Also, negotiationsshould be opened to deal withissueslikeproductivity incentivesand labour marketregulations(e.g. hiring-firing restrictions, overtimework or part-timejob regulations,various unemployment programmes),reductionsin subsidiesto ailing corporations,and price and wage moderation. 5.3. The need for a demand expansion

Fixing the supply side creates the necessary conditions for a nonincrease in employment.But unless demand is presentto inflationary absorb the increased supply,a fallin the NAIRU will,in the shortrun withoutdentingunemployment.The government onlyreduce inflation, could instead aim for a gradual reduction of actual unemployment, trailingbehind the reductionof the NAIRU (see e.g. Blanchard etal., 1985). Whatis theamountof demand expansionthatwould be required to achieve a gradual reductionin unemployment(assuming a falling NAIRU), say at a rate of 1% a year? For this purpose, it is useful to turn to Okun's Law, which relates GDP growthto reductionsin theunemploymentrate.Using simulations of demand expansion performedwith METRIC model (Artus et al., 1981) we find that real GDP must grow about 2% faster than its underlyingtrend (a proxy for supply expansion) if unemploymentis to be reduced by 1% a year.12 to say how much demand stimulusshould accompany It is difficult the supply side measures,in order to achieve the targetgrowthrates. In essence, both monetaryand fiscalpolicies will have to operate on a feedback basis, by settingnominal GDP targetsthat are sufficiently expansionaryto allow forthe desired real growthof the economy.The supply measures will of course themselvesraise demand, via higher take-home pay, higher investmentdemand as labour costs fall, and higher internationalcompetitiveness.However, the precise details of such an induced demand expansion will require separate study. It should also be pointedout thatany precisionin the size of the necessary demand expansionis vitiatedbytheuncertaintiesin theworldeconomy. 5.4 The external constraint

As has been amplydemonstratedbythe 1981-82 experiment,no policy measure is likelyto succeed unless it pays due attentionto the external I 12

l As a rough estimateto fixan order of magnitude,if trend annual growthis estimatedat 3% the growthrate the governmentshould targetis about 5% a year over the fiveyearsfollowing the wedge reduction.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

305

constraint.This constraint,in turn has two components: the current account and speculativecapital flows.It should be clear that supplyfriendlymeasuresof the typeenvisionedshould boostthe stockmarket, build up confidenceand therefore,at the veryleast, leave the capital account neutral. As for the currentaccount, we already noted that a reduction in labour costs has the importantadvantage of creatinga stronggain in competitiveness.Yet simulationsperformedwithFrenchmacromodels (METRIC or OFCE) concerningreductionsin employers'socialsecurity contributions,find that the currentaccount slightlyworsens because This result such policiesgeneratefastergrowthand a rise in imports.13 (of which we are a bit sceptical)should at least be taken as a warning thatany expansion has a potentialforcreatingcurrentaccount difficulties. The best wayto resolvethe externalpressuresfroman expansion would of course be forEurope as a whole to undertakemeasuressimilar to the ones that we are outlining.Afterall, France's difficulties are shared in various degrees by all EEC countries,where one also finds severelabour marketrigidities,large wage gaps, and low ratesof profit (see Bruno and Sachs, 1985; Bruno, 1986; Dornbusch et al., 1983; Layard etal., 1984). If theotherEEC countriesrefusesucha coordinated expansion, the second best would be a devaluation withinthe EMS. Our estimatesindicatethatthe Franc is currentlyovervaluedby about 6% vis-a-visthe Deutschmark.Of course with a reduction in labour costs,the extentof the overvaluationis considerablyreduced. Yet, with an anticipatedgrowthdifferential, a devaluationis inevitable. 5.5. Fiscal and monetarypolicy

The proposed wedge reductioncould result in significantbudgetary costs,especially if the reductionis generalized to all workers,rather than applyingonlyto marginalworkers.A general cut in payrolltaxes of 15% would cost the governmentrevenues of about 3.75% of GDP, as we noted earlier. Of course therewould be large associated savings in unemploymentsupport programmes,that could be as much as 1.5-2% of GDP if the unemploymentrate drops from10 to 5%.M Assumingthattherewould remaina significant budgetarycostto the tax reductions,the question is whetherthat cost should be made up I I ' Artus et al. (1981) and Sterdyniaket al. (1984). In the OFCE model a 10% cut in employers'

social securitycontributionsleads afterone year to a 0.8% increase in outputand a worsening of the currentaccount by 0.1% of GDP. 14 We assume thatthe cost of unemploymentprogrammes(4.5% of GDP in 1985) conservatively would be cut by onlyone third,giventhe fixedcost of bureaucracyneeded to administerthese programmes.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

306

throughcutsin spendingof varioussorts,or byincreasesin othertaxes, or by somewhatlarger deficits. We have repeatedly stressedthat the supply-sideconsequences of any other tax increases should be scrupulouslyexamined, and not underestimated.If thereis a major conclusionof our analysis,it is that cutting deficitsvia tax increases and spending cuts have different macroeconomicimplicationsforthe French economy.Aftera decade and a half of tax increases,withthe attendantsupply-sidedifficulties, it is probably time to stress spending cuts as a way to resolve the budgetaryproblem.

Discussion WillemBuiter

Yale University

The authors'thesisis thatFrance,under PresidentMitterrand,undertook a modest fiscalexpansion in May 1981 whichhad adverse supplyside effects.There was some increase in the growthof output,but the economy soon ran into both internaland external constraints.First, and more importantly, the non-acceleratinginflationrateof unemployment (NAIRU) was too high to permitcontinued expansion without an increase in inflation.The authors argue that Mitterrand'spolicies raised the NAIRU and theycitea numberof contributory significantly in the minimumwage; employers'payrolltaxes the increase factors; were also increased; there was an extension of employee protection legislation;workinghours were reduced increasingmarginal labour costs; and finallythere was nationalization(although its effecton the NAIRU is ambiguous). Second, the economyalso ran intoa twofoldexternalconstraint.The currentaccount deterioratedowingto the world recession.There was also a lack of confidencein the Socialistadministrationon the part of the financial markets. This produced increased capital outflows. Together these placed the Franc under pressure and raised doubts about the continued participationof France in the EMS. The authorsargue the originalMitterrandprogramlastedlittlemore than six months.The firstcrackswere showingby the timeof the first devaluation in October 1981. Furtherdevaluationsin June 1982 and March 1983 put the finalnails into the coffin.The fiscalstimulus(as measured by the cyclicallycorrectedbudget deficitwhich is not the correctmeasure) was reversedas earlyas June 1982. I will firstexamine their claims about the increase in the NAIRU. such as increasedjob or skill The authors'dismissalof frictionaleffects, are difficult mismatch,is perhapsa littlehastysincemanyoftheseeffects

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

307

to measure adequately. They are probably neverthelesscorrectthat their impact was small. Instead they focus on what has been dubbed 'household-involuntary,union-voluntary'unemployment.Oil price increases, reduced growthin technicalproductivity(relativeto wage aspirations)and increases in the tax wedge have all raised unemployment.However, I do not findtheirparticularmodel, withcompetitive firmsfacinga real wage set by a monopolisticunion, eithernecessary or useful.A more plausiblestoryis to assume unions setthe moneywage and imperfectlycompetitivefirmsset prices and hence also the real wage. The optimizingfoundationsof the model disappear completely when their purely staticmodel is embedded in an 'error-correction' mechanismto obtain the Phillipscurve. The finalmodel looks like a standard expectations-augmented Phillips curve where the difference between the target and warrantedwage drives inflationand unemployment. A crucial missing link in the storyis why the extra involuntarily unemployed union memberscannot find,or choose not to find,work in the non-union sectorof the economy. Is thisbecause wages in this sectorare too low relativeto unemploymentbenefitsas has been suggested by PatrickMinfordforthe UK? Or is it because relativewages are rigid across the two sectors?The mechanismat work here needs to be spelt out. The estimatesof the Phillipscurvein Table 6 providethe centrepiece oftheanalysis.However,theseare equationswithrelatively low explanatorypower estimatedon only twenty-oneobservations.While there is some attemptin Appendix C to examine the robustnessof the results withrespectto dynamicspecification and theinclusionofotherplausible determinantsof the NAIRU, theystrikeme as a veryslender basis on which to make the very importantpolicy conclusion that France was supply constrained at the beginning of the Mitterrandexperiment. of the equationsover 1981-84 would be one useful Testingthe stability testthe authors could carryout. Turning now to the wage gap calculations,theseare potentiallyuseful to assesstheirstatistical althoughI finditdifficult significance.However, the authors make a crucial non-sequiturin assuming that a positive wage gap impliesthe absence of a significant marginof unemployment due to deficientdemand. The real wage can be too high withoutthis being the bindingconstraint.The findingof a positivewage gap merely impliesthatone cannot get back to fullemploymentwithoutan accompanyingreductionin real wages. More generallyI do notthinktheauthorsexamine sufficiently closely the possibilitythat producers in France were constrainedby a lack of demand in the early years of the Mitterrandexperiment. Data on

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

308

Sachs and CharlesWyplosz Jeffrey

capacityutilizationand vacanciescould give some clues here. Is it the case that if France was in a demand constrained regime any fiscal expansionwould neverthelesshave failedbecause ofa lackofconfidence by the financialmarkets?In other words did the external constraint bind even if the internalone did not? Finallytheauthorssuggestthata crucialrolewas playedbytheFrench commitmentto the EMS. This is interestingbut not altogetherconvincing. Why did the authoritiesfeel constrainedto staywithinthe EMS? They also argue thatMrs Thatchercould tie the hands of a futureUK Labour governmentbyjoiningtheEMS. I findthisimplausible:a future Labour governmentwould just leave the EMS in order to be free to pursue expansionistpolicies. GerhardFels

Institut der DeutschenWirtschaft, Cologne

From the French Socialists'experimentof the early 1980s one can learn an importantlesson on economic policy.This episode is highly instructivefor other countries as well. At least in Germany it has disillusioned many advocates of a 'soft approach' which unions and Social Democrats stillfavouras an alternativeto the presentconduct of economic policy.To be sure,Germany'sperformancewas not much better,and was actuallyworseon employment.But theexternaldifficulties faced by France demonstratethat Mitterrand-style policies cannot solve internaleconomic problems.Sachs and Wyploszprovidean interestingevaluation of these policies,though I should have preferreda greateremphasis on comparisonwithother countries. I feel much in sympathywiththe basic analyticalapproach of the paper. I do, however,have some reservationswithrespectto the policy recommendations.In substance,I do not thinkthatthe recommended policy option currentlyexists in France. In particular,the external constraint,whichwas responsiblefor France's U-turnin March 1983, mustbe given more attention.France's experience highlightsthe fact that in an open economy,it is impossibleto conduct an autonomous economic policy independent of conditionsin the rest of the world, unless one revertsto trade restrictions and capitalcontrols.I thinkthis also has implicationsforthe therapyproposed by the authors. I think the authors may overstatethe degree of success on the employmentfront.One reason for France's moderate employment decline in the period 1980-84 may be a substantialamount of hidden unemploymentor unemploymenton the job. Table 3 lends supportto thishypothesis:there was a substantialincrease in employmentin the

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

309

sheltered sector. But this was not merely a feature of the Socialist administration;it had already occurred during the seventies. France's successagainstinflationalso seems overstated.It is truethat the inflationdifferential between France and Germanyis substantially smaller today than it was in 1982, but in both countriesthe inflation rate in 1985 is only about one thirdof what is was in 1982. There are differingviews about what caused inflationto subside. Sachs and Wyploszmaintainthatthe main factorwas wage and price controls.If that is true, the success may only be temporary.But this explanation does not seem to be very convincing since all major countries experienced a sharp drop in the rate of inflation.This phenomenon has to be explained by monetaryrestraintabroad. Paul Volker successfullyreduced monetarygrowthin the US, and Germany followedsuit. Hence France, too, was forcedto adopt a policyof tight money. Otherwise,she would have been forced to leave the EMS. So again it is the externalconstraintwhich has to be given creditfor the improvementafter 1983. The paper suggeststhat a macroeconomicstrategyfor overcoming the problemsstillexists.I agree withthe argumentthata reductionin labour costsof about 10% would be warrantedand useful.But I doubt whetherthis can be achieved by a significantcut in labour taxes, i.e., in compulsorysocial securitycontributions.Can we reallyexpect a shift of the wedge burden - of whatevermagnitude- fromemployersand employeesto the public sectorto overcomethe rigiditiesof the French economy? It seems to be more importantto reduce the burden than to shiftit to the public budget. Although the rationale of the US tax cut policy is differentfrom that of the Sachs-Wyplosz analysis,the structureof theargumentis quite similar:more incentivesto the private sectorat the expense of an increased public debt. If any countryat all can affordto have a significant increase in public debt relativeto GDP it is the United States,due to favourableconditionswithregard to its labour market,profitrates,equitycapital and the like. By contrast,the French economysuffersfromso manyrigiditiesthatthe fiscalstrategy recommendedby the authorswould be highlyriskygiventhe external constraintand the budget situation.I thinkthe same is, in principle, also true in the case of Germany. To reduce the wedge burden, the authors propose a social contract and a gradual reductionin public spending, especiallyin subsidiesto corporationsand individuals.But has a social contractever been more than a dream in France? Even the Socialistgovernmenthas not been able to establishit. And the level of subsidies is as much a subject of the debate over income distribution as wages and salaries.The financial

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

310

Sachs and CharlesWyplosz Jeffrey

supportwhich is given to decliningindustrieshas more the character of social policy than of industrial policy. And what about a large devaluationof the Franc as a complementarymeasure?My guess is that accelerated inflationwould be the most likely result. If the unions respond to increased inflationthe burden will be shiftedback to the businesssector. What is the alternative?Austerityis no solution.But in the environmentofa scleroticEuropean economy,privateconsumersand investors have lost confidencein macroeconomicpolicieswhichlead to a higher budget deficit.Under European conditions, supply policy must be geared towards improvingtotal factor productivity,which will also bridge the wage gap. In the French case I would propose to continue and reinforcethe course of economicpolicywhichthe governmenthas been pursuingsince 1983. This would mean; furthereffortsto reduce the budget deficitby maintainingonly a modest growth of public spending; continuationof monetarydisciplineto ensure low inflation rates; combatingmajor legislativerigiditieswhich constrainand discourage employersfromhiringadditionalworkers;deregulationof the servicesector to increase competition,and the privatizationof public service activities; easing of the constraintsimposed on small and medium-sizedfirmsin the formof workerparticipationand additional social changes; and finallythe introductionof tax sheltersfor newly establishedenterprises. Of course, such a strategypromises only a gradual improvement. However,ifsupplyconditionsare improvedthefreedomformanoeuvre in fiscalpolicy will become greater. The greater freedom should be used to lower taxes on investmentand high-riskactivities.My beliefis thatsuch a strategywould enable the Frencheconomyto growat a rate of about 3%. The German economy may serve as an example. It is about to recover and will accelerate its growthmeasures to improve supplyconditionswhichallowed the economyto overcome stagnation withoutan increase in the budget deficit.

Georges de Menil

Ecole des Hautes Etudes en Sciences Sociales, Paris

Afterthe legislativeelectionsof March 1986, the new Frenchgovernmentwilltryto learn fromthe errorsof the precedingfouryears.Why did thisSocialistexperimentfail?My commentswillconcentrateon the authors' three principal and most provocative conclusions: budget deficitswere not the most important problem; reductions in the

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

311

and competivenessof Frenchenterprise- in short,supplyprofitability side disturbances- were the most seriouslydisruptiveaspects of the Socialist programme; and Mitterrand'spolicies exacerbated these supply-sideproblemsbut did not constitutea fundamentaldeparture fromthe previoustrends.They were in essence a ripple on the rising tide of Eurosclerosis.My discussionis organized under threeheadings: the demand side aspectsof the programme,itssupply-sideconsequences, and the implicationsfor presentand futurepolicy. The importance of relativedemandstimulus. In theirclaim for the relative unimportanceoftheKeynesianstimulusin 1981-82, Sachs and Wyplosz make a myopicerror.They pointout thatthe increaseof the structural deficitof the French governmentbudget during 1980-82 (1.4% of average GDP) was smaller than in the United States during 1980-83 (1.8% of GDP) and about the same as in Germany(1.2% of GDP during 1977-80) followingthe Bonn summit(see Muller and Price, 1984). It is not budgetaryexpansion perse whichdisruptsexternalbalance, but expansion relativeto the policies being followed by major trading partners.In 1981 these partnerswere simultaneouslymoving in the direction of restraint.The substantial divergence between French stimulusand foreignrestraintwas neitheran accident nor a surprise: the French governmentconsciouslydecided to expand when others were contracting. Supply-side consequences.The mostappealing aspectofthe Sachs-Wyplosz

studyis the crucial emphasistheyplace on the effectof the Mitterrand programme on supply. The heart of their studyis an application to France of the wage-gapanalysiswhichhas gained prominencein recent years.A numberofeconomistsfollowingGiersch(1981), includingmost recentlyBruno and Sachs (1985), have argued thatexcessivereal costs of labour are the major reason manyWesterneconomies,particularly in Europe, have experienced steadilyrisingunemployment.'Natural' unemploymentis largely determined in this view by the difference betweenthe effective cost of labour and itsmarginalproductivity when all resourcesare fullyemployed.This discrepancyis the wage-gap.The authorsthen estimatea Phillipscurve whichrelatesshort-runchanges in inflationto the differencebetween actual and 'natural' unemployment. Whilst inflationand actual unemploymentare observable the unobservable natural rate of unemployment must be modelled Whereas forthisI believethatwage-gapanalysisis required, implicitly. Sachs and Wyploszstressonly one aspect of thisanalysis,the so-called

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

312

Sachsand CharlesWyplosz Jeffrey

'wedge' between gross labour coststo the firmand net take-homepay is not warranted.'5 receivedby workers.In myview thissimplification the of in the authors' role Moreover, key supply analysisis deficient fortwo reasons. First,inadequate attentionis paid to considerationsof competitivenessand external balance. It was not inflationor rising unemploymentbut the deteriorationof the currentaccount, and the consequent speculativeattackon the Franc, whichforceda reversalof budgetarypolicyin the Springof 1983. It is regrettablethattheauthors, in contrastto their careful studyof inflationand unemployment,do not bring this same detailed analysis of supply and demand to the externalposition. The second deficiencyis more fundamental.In theirintroduction, the authors recognize the long-runstructuralreformswhich,in the eyes of both the Socialistsand the opposition,constitutedthe heart of the programme. In many areas these were major departures. The nationalized sector increased by one quarter, to 18% of total nonagriculturalvalue-added. Marginal tax rates were increased,reaching up to 75% in 1982, and a new annual wealth tax of up to 2.5% was introduced. New labour laws enhanced trade union power, private rented housing was subjected to rent controls,and privatemedicine was curtailed. Whilstbudgetarypolicyhad been reversedby 1983, and some incentives for enterprisehad been restored,by 1985 the full array of the structuralreformswere essentiallyleft untouched. What were their effectson productivityand competitiveness?Sachs and Wyplosz are silenton the question.Having emphasizedthe importance surprisingly of supply-sideconsiderations,theauthorshave givenus Hamlet without the Prince of Denmark. Of these omissions,the most glaring is the neglect of nationalization,especially in the banking sector. Given exchange controlsand the underdeveloped nature of French capital markets,bank loans constitutethe most importantsource of external funds for privateinvestmentin France. The argumentthat the 1981 l

I

15The wage-gap, in Sachs-Wyplosznotationis

(h + t,- pv)-_ whichmay be rewrittenas (h - t2-pc)+(tl

+ t2+ P -pV)-

q

The firstterm is the log of real take-homepay, the second is the log of the wedge. Having made this decomposition,the authors implicitlyassume in their estimationthat changes in NAIRU are caused exclusivelyby changes in the last two terms,and since the data will not supporttheirimplicitomissionof the firstterm,take-homepay, theyare obliged to introduce an ad hoc time trend whichonly partiallycapturesthisomission.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

313

banking nationalizationwas of secondary importance because twothirdsof the industryhad previouslybeen nationalizedin 1946, misses the potentiallydecisive influenceon the entire industryof the threat of competitionfromthe privatesector.Similarcommentsapply in the manufacturingsector. Not onlyis the reductionin competitionlikelyto reduce productivity and efficiency,each nationalizationcreates a precedent for future nationalizationsand may significantly reduce incentivesforinnovation and risk-taking in the privatesector.If StephenJobsor KennethOlson, the foundersof Apple and Digital Equipment,had been French,they mightwell have emigrated. Of course these are long-runeffectsand difficultto measure. But they should not be neglected on that basis. By concentratingtheir analysison short-runphenomena which can be discerned quicklyin nationalstatistics, Sachs and Wyploszgivean unbalanced accountof the Socialistprogramme,and may miss more than half the story. Implicationsforpolicy.The messageoftheSachs-Wyploszstudyforfuture

Frenchpolicymakersis clear: reversesome of the measuresof 1981-82, but do not merelyrevertto the policies of the previous decade. The relentlessincrease of structuralunemploymentover that period is a tellingtestimonyto the cost of social rigiditiesand misguided policies whichpush labour costsever higher.This trendmustbe reversed.Sachs and Wyploszpropose a 15% cut in labour coststo reduce the NAIRU to 5%. This radicalsurgery,whichwould implythatthesumofemployee and employertaxes on labour fell from60% to 35% of wages, could notbe implementedovernight.And anyattemptto achievecoordinated reductions throughout Europe could only take longer. Given the authors' premise that there is little current slack in French labour markets,neitherrapid demand expansion nor a quick solutionto the unemploymentproblem is feasible. An alternative,and in my view more promising,way to reduce the wage-gap is to erode rigiditiesand distortions,improve competition, and thusmake marketsworkbetter.Examples of such measureswould include privatizationand the abolitionof controlson prices,rents,and and competitiveness, this foreignexchange. By increasingproductivity offersa more durable solution. It would be a radical departure from trendsin France over the last twentyyears. General Discussion Jacques Melitz believed the comparison with the experience of the Giscard administrationwas instructive.The Chirac expansion was also

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

314

Sachsand CharlesWyplosz Jeffrey

short-lived,but the Barre plan stabilizedthe currentaccount and the Franc relativelyquickly.Inflationwas allowed to rise over the period 1977-81. By contrastthere had been three successive devaluations under the Socialistsand even in 1984 the currentaccount was stillin deficit.This forcedMitterrandto maintaina contractionary fiscalstance feltthe continued downward on inflation. He also exerting pressure model of union behaviourwas inappropriateforFrance. Union membershiphad been fallingrapidlyunder the Socialistsand the preservation of union membershipratherthan the pursuitof higherwages was probablymore importantto union leaders. Patrick Minford supported Buiter's criticismof the lack of an adequate storyabout behaviourin the non-unionsectorand whyunemployed union memberscould or would not findjobs there.Even if the whole economy were treatedas unionized the storycould not be sustained because the shadow wage of displaced workerswould fall to zero. He thoughtthat the unemploymentbenefitsystem,the 'Aide Sociale', early retirementschemesand the like played a centralrole in underpinningthe model of labour marketbehaviour. Torsten Persson suggested the distinctionbetween internationally between tradable and non-tradablegoods could help in distinguishing and 'confidence'explanations for the fall in investthe 'profitability' ment. If firmsin the tradable sectorfound it less easy to pass on cost increases on account of foreigncompetitionthan firmsin the nontradablesectorone would expectto findthe share priceof the twosorts of firmsreactingdifferently to increasesin labour costs.On the other hand fear of nationalizationwould affectboth sectorsequally. He also thoughtthe distinctionwas relevantto the specificationof the Phillips curve which implicitlyassumed prices were a mark-upover costs: this would not be valid forthe tradable sector. MervynKing found the estimatedinvestmentequation a peculiar hybridbetween 'accelerator'models which stressedthe role of output and 'Tobin's q' models whichstressedthe role of the stockmarket. Paul Krugman pointed out a peculiar propertyof the estimated Phillipscurve, which implied thatunions were perpetuallyseeking to returnto thetrendvalue ofthetargetwage irrespectiveofdevelopments in productivityor the wedge in the interim.This did not seem very plausible over time spans of a decade or more. Giampaolo Galli noted thatneitherunemploymentnor the NAIRU was higherin France than in mostother European countriesand that the somewhatlaterdecelerationin inflationcould be attributedentirely

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

315

to the later increase in-interestrates. Monetarypolicyseemed capable of explaining most of the differences. Appendix A. Computation of wage gaps For each sectorwe assume a CES productionfunction Y = A[SK-P + (1 - )L-P]-'P

(Al)

where Y is real value added, K the stockof capital and L is labour in units efficiency t

L = L (1+i)

(A2)

i=l

withyi the rate of labour augmentingtechnologicalprogressin year i and L, is labour (in man-hours). If SL is the share of labour in value added and H represents(hourly) labour costs,SL = HL/ Y. Let small lettersdenote logarithms.Then an approximationto (Al) and (A2) is Ay,= (1 - SLt)Akt+ SLt(Alt +

(A3)

yt)

From (A3) it is easy to obtain estimatesof Yt since all of the other variablesare observed. In practice,we smoothSL by takinga two-year movingaverage of the originaldata. Next we normalize Y, K and L to 1.0 in the base period so thatA = 1. Maximizingprofitgiven K yieldsthe usual relationship SL=(1 - 8)

t

I ( +7)

_i=1

-P

(Yt/Lt)P

(A4)

Withthe above normalization,we findthatin the base period S? = 1- 8 so that parameter 8 is directlyobtained as the share of capital in the base period. We can now express (A4) as a demand for labour aftersubstitution of (Al) (up to a constantterm) i=l

P

Thus the real wage rate ensuringfullemploymentLf giventhe stock of capital is (up to a constantterm) h*y=

i=l

.

t 1+p In [8(K,/f)P P

+(1 -8)]

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

(A6)

316

Sachs and CharlesWyplosz Jeffrey

The wage gap is then simplyh,-h* where ht is the (log of the) observedreal labour cost. It is computed foralternativevalues of p = 0 and p = 1. A similarprocedure is followedforthe case of Hicks-neutral technologicalprogresswhere Equation (A2) is replaced by an equation in whichA growsover time. Appendix B. The union wage model 1. Firms' demand for labour

We assume that the technologyis representedby a CES production functionwithHicks-neutraltechnologicalprogress Y = A[SK-P + (1 - S)L-P]-P

(B1)

whereA = A(t) is the productivity factor.The unitcostof labour is the gross nominal wage H, divided by the value added deflatorPV, witha mark-up T1 foremployers'taxes. C=v

H

(B2)

TI

The firms'soptimizingconditionis C = a Y/aL = A-P(1 - 8)(Y/L)'+P

(B3)

With lower-case letters denoting logarithmsand omittingconstant terms,(B2) and (B3) imply c = h - p+ t = -pa+(1 +p)(y-l) (B4) (B1) can be log-linearizedas y = a +(1-

SL)k+SLl

(B5)

where SL is the share of labour in value added. of substitutiono = 1/(1+ p), (B4) and Denotingthe constantelasticity (B5) yield the demand forlabour: 1= k-p(h -pv + t- a) with /3= /(1 - SL).

(B6)

2. Trade union's wage setting

FollowingMacDonald and Solow (1981) we assume thata singleunion maximizes the expected utilityof the representativeworker (or, equivalently,the weighted average of employed and unemployed

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

317

workers): E (u) = (L/L)u( W) + [(L- L)/L]u(l)

(B7)

where L is the labour force,L the number of employed workers; W the real take-homewage, f real unemploymentbenefits.Real takehome wage is; W=

(B8)

PcT2

withPC the consumptiondeflatorand T2 the employee tax mark-up. The trade union maximizes(B7) withrespectto the gross salarylevel H, given the firm'sdemand for labour in (B6). If the union is able to impose its optimal salary level, it acts as a monopolist,leaving to the firmthe decision to hire labour accordingto (B6). The union's optimizingconditionis; L au aW -+ u-u? _ aL aC =O L aC aH LaW +H

(B9)

where u = u(W) and u?= u(fl). This can be shown to become W au

u uaW.+

u-u0C

u

dL

LaC

0

(B 10)

If the utilityfunctionhas constantrelative risk aversion, u= W'-Y/(l- y), and with(B6) we obtain 1-, =3P

U-U?

(Bl 1)

U

or

W" =

l withO =

y

(B1l')

Iftheunionisabletoenforcethisrealtake-home wage,theemployment levelis givenby lu= k+3(a-x-wu) where w" = 0+

(B12) from (Bll') and x is the logarithmof the wedge

separatinglabourcostsand take-home wages: pCT T, C PT, PCT PV W

=

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

(B 13)

318

Sachs and CharlesWyplosz Jeffrey

The union sets the salary level as a mark-up over unemployment benefits,withthe rate of mark-updepending inverselyupon the elas- SL) = as in a standard ticityof the firm'sdemand for labour r/(1 monopolypricingmodel. In the Cobb-Douglascase 3 is exactlyconstant and the targetreal wage is constant. Quite logically,(B12) shows thatthe equilibriumemploymentlevel and labour costs depends on the differencebetweenfactorproductivity as measured by the sum of the wedge x, the unemploymentbenefitsw and the mark-upimposed by the trade union 0. Importantly,it also depends on the stock of capital k. While k is taken as exogenous, it should properlybe related to labour costsvia the factorprice frontier, and is another channel throughwhich high labour costs may reduce employment. 3. The Phillips curve

Whilethe previoussectionpresentsthe labour marketequilibrium,the Phillips curve is intended to describe how this equilibrium is approached. It requires some lags in the adjustmentmechanism.We adopt two standard assumptions:first,unions are not able to achieve their target real take-homewage but work towards it; second, gross nominalwages are set at the beginningof the currentperiod when the CPI inflationrate wrc is not knownand approximatedby last period's rate rc_-(eitherbecause of the indexation mechanismor because of myopicexpectations).Using (B8), the resultingrate of increase of the gross hourlynominal wage is; Ah = Aw + r_In+At2+ a(l- IU)

(B14)

where Az = z - z- and wu is the (log of) real wage targetof the unions given in (B 11') as w = 0 + o. From (B6) we derive the rate of change of the value added deflator; Tv = Ah+ At - Aa - (Ak-Al)/3P

(B15)

If pm is the price index for imported goods, Pc =(Pm)A(PV)l-A plugging(B14) into (B15) gives (assuming 0 constant): Xrv=

rl+A(7rl-7t1)-

a(U-

U*)+A(w+tl+t2

+

)- AU/1

and (B16)

of labour where: i = a + (k- 1)/,1is the log of the marginalproductivity evaluated at fullemployment(see (B3) and (B5)) U = I- I is the rate of unemployment. U* = l- u is the union's targetunemploymentrate.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Mitterrand The economicconsequences ofPresident

319

This definitionof unemploymentassumes that hours worked have remained unchanged, which is grosslyinaccurate. We returnto this issue in the followingsection. (B16) is the Phillips curve referred to in the text. The NAIRU ANh4, i.e. corresponds to 7rv= 7r1, 7rm= 7Tr,AU = 0 and A(w + tl+ t2)= when inflationand unemploymentstabilize and labour productivity gains are exactlyabsorbed by labour cost increases.Then the NAIRU is the union's targetunemploymentrate and is given by U* =p(w

+ x- /)

(B17)

4. Hours worked

In the above formulationthe number of hours worked is taken as exogenous. A proper treatmentof thisissue would requireintroducing three changes in the model: leisure should enter the representative worker'sutilityfunction;the productionfunctionshould recognizethat there may be costs in operating given equipment withmore workers workingless hours; the numberof hours worked per workeris in the end a decision taken by the firmgiventhe extra cost of overtimework beyond the normal workweek(39 hours in France since 1982). The resultingproblemhas been studied in Wyplosz(1985) and in a similar model by Calmfors(1985). Its implicationsare discussed in the text. Appendix C. Alternativespecificationsof the Phillips curve Table Cl augmentsthe tworegressionsreportedin Table 6. Regression (3) augments Regression (2) by adding furtherlags in 7r,U, (i-x), contemporaneousA(pm-p) and separate currentand lagged values of x. Individually, their t statisticsare very small. An F test on the hypothesisthat the additional variables are jointlysignificantis also comfortablyrejected. (F(6, 8) = 0.33). to improve Having checked that it is not possible straightforwardly on the dynamic specificationof the equations shown in Table 6, we then investigatethe consequence of omittingcertaincontemporaneous variablesfromtheseequations.(Ideally,ofcourse,one would investigate both extensions simultaneously,but with annual data we exhaust degrees of freedom veryquickly.)Regressions(4) and (5) investigate the consequences of augmentingRegressions (1) and (2) of Table 6 withthree variables (all in logarithms): B: real unemploymentbenefits,the ratioof totalunemploymentoutlay to the product of the numberof unemployedand the consumerprice index.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Sachs and CharlesWyplosz Jeffrey

320

Table C1. Augmented Phillips curves (Dependent variable AlT, annual data, 1963-1984) Variable

Regression(3)

Regression(4)

Regression(5)

Constant U A(pm-p)(i+-x) Ar-1 e_l t DV1 DV2

201.38 -4.34 0.07 -40.28 -0.33

200.65 -3.83 0.06 -49.00

242.14 -4.46 0.06 -53.90 -0.32

IT-2

U_1 A(pm p) ($1- x)-_ x x-1 B RR UV Standard error

(0.90) (1.42) (1.62) (1.93) (2.15)

2.69 (0.98) 1.92 (3.69) -0.01 (0.02) 0.07 (0.04) 0.02 (0.39) -2.01 (0.05) 22.82 (0.77) -12.25 (0.29)

(3.64) (4.28) (1.79) (3.54)

-0.65 (3.89) 2.87 (4.32) 1.98 (2.80)

-5.39 (0.81) 6.90 (1.13) -1.51 (1.54)

0.72

(5.88) (6.82) (2.57) (6.22) (3.97)

3.22 (7.07) 1.95 (7.69)

-3.17 (1.05) 3.15 (1.24) -0.66 (1.21)

0.96

0.62

(t statisticsin parentheses)

RR: the replacementratio,definedby (1- U (total unemploymentoutlays\

U

/k

totalgrosswages

J

ratio. UV: the unemployment-vacancy t statistics and the F testson the 3 variables individual both Again, 1.15 for jointly, (F(3, 12)= Regression (4), and F(3, 12)=0.62 for regression (5)) indicate that their omission from Table 6 is not the source of any serious specificationerror. References IMF Staff An Empirical RealWageHypothesis: Evaluation', Artus,J.(1984).'The Disequilibrium Papers.

Model Macroeconomic of a Quarterly Artus,P., G. Laroqueand G. Michel(1984).'Estimation Econometrica. withQuantity Rationing', and R. Teyssier,(1981), Artus,P., J.Bournay,P. Morin,A. Pacaud,C. Peyroux,H. Sterdyniak de l'EconomieFrancaise',INSEE, Paris. 'Metric.Une Modelisation etStatistique. les ComptesNationaux',Economie Beudaert,M. (1983).'L'Annee1982a travers R. Layardand M. Monti(1985).'Employment Blanchard, O., R. Dornbusch, J.Dreze,H. Giersch, in Europe:a Two HandedApproach',CentreforEuropeanPolicyStudies,Paper and Growth No. 21.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

The economicconsequences Mitterrand ofPresident

321

Bouissou, M. B., J.J.Laffontand Q. H. Vuong (1984). 'Econometriedu Desequilibresur Donnees Microeconomiques',Annalesde 1'INSEE. Bourit,F., P. Hernu and M. Perrot(1982). 'Les Salaires en 1981', EconomieetStatistique. Bruno, M. (1986). 'AggregateSupply and Demand Factorsin OECD Unemployment',Economica. Basil Blackwell. Bruno, M. and J. Sachs (1985). Economicsof Worldwide Stagflation, Buiter,W. H. (1985). 'A Guide to Public Sector Debt and Deficits',EconomicPolicy. Calmfors,L. (1985). 'Trade Unions, Wage Formationand MacroeconomicStability',Scandinavian Journalof Economics,forthcoming. Claassen,E. and C. Wyplosz(1985). 'Capital Controls:Some Principlesand theFrenchExperience', in J. Melitz and C. Wyplosz (ed.) The FrenchEconomy, WestviewPress. Coe, D. and F. Gagliardi(1985). 'Nominal Wage Determinationin Ten OECD Economies', OECD WorkingPaper No. 19. Colin, J.F., M. Elbaurmand A. Fonteneau (1984). 'Chomage et Politique de 1'Emploi1981-1983', etDiagnosticsEconomiques, Observations Revue de I'OFCE No. 7. De Menil, G. and J. Sastre (1985). 'TransferPolicies,Income and Employmentin France', in G. De Menil and U. Westphal(eds.) Stabilization Policyin Franceand theFederalRepublicofGermany, North-Holland. Dornbusch, R., G. Basevi, O. Blanchard, W. Buiter, and R. Layard (1983). 'Macroeconomic Prospectsand Policiesforthe European Community',CentreforEuropean PolicyStudiesPaper, No 1. Feroldi, M. and F. Meunier (1984). 'La Boucle Prix-Salaireset l'Inflationdepuis 1970', Economie etStatistique. Fonteneau,A. and A. Gubian (1985). 'Comparaisondes RelancesFrancaisesde 1975 et 1981-1982', etDiagnosticsEconomiques, Observations Revue de I'OFCE No. 12. Fonteneau, A. and P. A. Muet (1983). 'La Politique Economique depuis Mai 1981: un Premier etDiagnosticsEconomiques, Revue de I'OFCE No. 4. Bilan', Observations Giersch,H. (1981). 'Aspectsof Growth,StructuralChange, and Employment- A Schumpeterian PoliciesforGrowthand Stability, Weltwirtschaft Perspective.'in H. Giersch (ed.) Macroeconomic an der UniversitatKiel. Grubb,D., R. Layard and J.Symons(1984). 'Wages, Unemploymentand Incomes Policy',Centre for Labour Economics Discussion Paper No. 168. Grubb, D., R. A. Jackmanand R. Layard (1982). 'Causes of the CurrentStagflation',Reviewof EconomicStudies. Grubb, D., R. A. Jackmanand R. Layard (1983). 'Wage Rigidityand Unemploymentin OECD Countries',EuropeanEconomicReview. Knoester,A. and N. van der Windt (1985). 'Real Wages and Taxation in Ten OECD Countries', Institutefor Economic Research, Erasmus University,Rotterdam,Discussion Paper No. 8501. Kolm, S. C., 'La TransitionSocialisteFrancaise', Le Monde 17-18 June 1981. Lambert,J. P., M. Lubrano and R. Sneessens (1984). 'Emploi et Chomage en France de 1955 a 1982: un Modele Macroeconomique Annuel avec Rationnement',Annalesde l'INSEE. Langohr,H. and C. Viallet(1986). 'Compensationand WealthTransfersin theFrenchNationalizations' JournalofFinancial Economics,forthcoming. Layard, R., G. Basevi, O. Blanchard,W. Buiterand R. Dornbusch (1984). 'Europe: The Case for UnsustainableGrowth',Centre for European PolicyStudies, Paper No. 8/9. MacDonald, I. M. and R. M. Solow (1981). 'Wage Bargainingand Employment',AmericanEconomic Review. Malinvaud, E. (1986). 'The Rise of Unemploymentin France', Economica. Marchand,O., D. Raultand E. Turpin (1983). 'Des 40 heuresaux 39 heures: Processuset Reactions des Entreprises',EconomieetStatistique. Muet, P. A. (1985). 'Economic Managementand the InternationalEnvironment,1981-1983', in H. Machin and V. Wright(eds) EconomicPolicyand PolicyMakingundertheMitterrand Presidency 1981-1984, Frances Pinter,London. Muller,P. and R. W. R. Price (1984) 'StructuralBudget Deficitsand FiscalStance',OECD Working Paper No 15. Oudiz, G., E. Raoul and H. Sterdyniak(1979). 'Reduire la duree du travail,quelles consequences?', EconomieetStatistique. Oudiz, G. and H. Sterdyniak(1985). 'Inflation,Employmentand External Constraints:An Overviewof the French Economy duringthe Seventies'. in J. Melitzand C Wyplosz(eds.) The FrenchEconomy,Theoryand Policy,WestviewPress. Price,R. W. R. and P. Muller (1984) 'StructuralBudget Indicatorsand the Interpretationof Fiscal PolicyStance in OECD Economies' in OECD EconomicStudies,No. 3.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

322

Sachs and CharlesWyplosz Jeffrey

Steinherr,A. and B. Van Halperen (1983). 'Approche pragmatique pour une Politique de Plein-Emploi:les Subventionsa la Creation d'Emplois', EconomicPaper No 22, EEC. Sterdyniak,H., M. A. Boudier, M. Boutillier,F. Charpin and B. Durand, (1984). 'Le Modele etDiagnosticEconomiques. Trimestrielde l'O.F.C.E.', Observations Stoffaes,C. (1985), 'the Nationalizations1981-1984: an Initial Assessment',in H. Machin and V. 1981-1984, Frances Presidency Wright(eds.) EconomicPolicyandPolicymakingUndertheMitterrand Pinter,London. World Financial Market,(1983), 'The FrenchAusterityProgram'. Wyplosz,C. (1985). 'A Note on the Reductionof the Workweek'.unpublished,INSEAD.

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

ERRATA Readers attentionis drawnto thefollowingerrata: p. 264. Nine linesfrombase lineshouldread "to 6% in 1985"

C (T P'_\ p. 317. Equation on lastlineshouldread: X = W= T2 (p p. 379. Two linesfrombase lineshouldread "under2% per month"

This content downloaded from 129.194.8.73 on Thu, 26 Dec 2013 12:38:48 PM All use subject to JSTOR Terms and Conditions

Smile Life

When life gives you a hundred reasons to cry, show life that you have a thousand reasons to smile

Get in touch

© Copyright 2015 - 2024 PDFFOX.COM - All rights reserved.