Europe’s Productivity Europe’s Productivity Growth Slumps But Growth Slumps But Employment Surges Employment Surges

Ian Dew-Becker, NBERIan Dew-Becker, NBERRobert J. Gordon, Northwestern Robert J. Gordon, Northwestern

and NBERand NBERPresented at Sciences-Po and Presented at Sciences-Po and

OFCEOFCEParis, 24 April 2007Paris, 24 April 2007

Ian in SF, you can’t see Ian in SF, you can’t see “MV=PY”“MV=PY”

This is a Work in This is a Work in ProgressProgress

At the end I’ll tell you some of our plans for At the end I’ll tell you some of our plans for further researchfurther research

Today’s Presentation Combines a Joint Today’s Presentation Combines a Joint Version from Last September with a Solo Version from Last September with a Solo Version of Ian’s from FebruaryVersion of Ian’s from February

One Thing we have in Common: Loud One Thing we have in Common: Loud ColorsColors

Since his first day as my RA 3.5 years ago, Since his first day as my RA 3.5 years ago, he has come up with inspired color schemes, he has come up with inspired color schemes, like everything involving EU must be yellow-like everything involving EU must be yellow-blue and US must be red-white-blueblue and US must be red-white-blue

Occasional lapses here toward black and Occasional lapses here toward black and whitewhite

The US Accelerates,The US Accelerates,Europe DeceleratesEurope Decelerates

From 1950 to 1995 EU productivity growth was From 1950 to 1995 EU productivity growth was faster than in the USfaster than in the US

But in the past decade since 1995 we have But in the past decade since 1995 we have witnessedwitnessed An explosion in US productivity growthAn explosion in US productivity growth A slowdown in EU productivity growth roughly equal A slowdown in EU productivity growth roughly equal

in sizein size An explosion in research on the US takeoff and but An explosion in research on the US takeoff and but

much less research on Europe’s slowdownmuch less research on Europe’s slowdown The magnitude of the shift (average EKS&GK The magnitude of the shift (average EKS&GK

Groningen)Groningen) EU/US level of labor productivity (ALP)EU/US level of labor productivity (ALP) 1979 1979 1995 1995 20042004

80%80% 97%97% 89%89%

Point of Departure: Post-95 Point of Departure: Post-95

Turnaround Plus New Turnaround Plus New HeterogeneityHeterogeneity This paper begins with two simple This paper begins with two simple

observations:observations:1. While European productivity (Y/H) has fallen 1. While European productivity (Y/H) has fallen

back since 1995 relative to the US, output per back since 1995 relative to the US, output per capita (Y/N) has not fared nearly as badlycapita (Y/N) has not fared nearly as badly

►►Y/H growth gap: .9%Y/H growth gap: .9%►►Y/N growth gap: .2%Y/N growth gap: .2%

2. After 1995, we see divergence across the 2. After 1995, we see divergence across the EU-15 in Y/H growthEU-15 in Y/H growth

► ► St. Dev. 1970-1995: 0.62St. Dev. 1970-1995: 0.62► ► St. Dev. 1995-2005: 1.01St. Dev. 1995-2005: 1.01

The Key Identity SuggestsThe Key Identity Suggeststhe Tradeoffthe Tradeoff

An identity links Y/N and Y/H to H/N:An identity links Y/N and Y/H to H/N:Y/N = Y/H * H/NY/N = Y/H * H/N

Thus the paradox of high European Y/H and low Y/N Thus the paradox of high European Y/H and low Y/N must be resolved by lower H/Nmust be resolved by lower H/N

Also, Y/H and H/N are jointly determinedAlso, Y/H and H/N are jointly determined The task of this paper is going to be figure The task of this paper is going to be figure

out which direction the causation runsout which direction the causation runs We will argue that a good deal of the decline in We will argue that a good deal of the decline in

ALP growth is due to exogenous employment ALP growth is due to exogenous employment shocksshocks

Also we will highlight the reversal of almost Also we will highlight the reversal of almost everything at 1995, comparing 1970-95 vs. 1995-everything at 1995, comparing 1970-95 vs. 1995-20052005

Bringing Together the Bringing Together the Disparate LiteraturesDisparate Literatures

Literature #1, why did Europe’s hours per capita Literature #1, why did Europe’s hours per capita (hereafter H/N) decline before 1995? Prescott, (hereafter H/N) decline before 1995? Prescott, Rogerson, Sargent-Lundqvist, Alesina, BlanchardRogerson, Sargent-Lundqvist, Alesina, Blanchard High taxes, regulations, unions, high minimum wagesHigh taxes, regulations, unions, high minimum wages Europe made labor expensiveEurope made labor expensive Movement up Labor Demand curve => low Movement up Labor Demand curve => low

employment + high ALPemployment + high ALP Literature #1 has missed the turnaroundLiterature #1 has missed the turnaround

Since 1995 there has been a decline in tax rates and Since 1995 there has been a decline in tax rates and employment protection measures; unionization earlieremployment protection measures; unionization earlier

Big increase in hours per capita, turnaround in both Big increase in hours per capita, turnaround in both absolute terms and relative to the US Move back absolute terms and relative to the US Move back down Ldown LDD curve curve

Textbook Labor Textbook Labor EconomicsEconomics

-2

-1

0

1

2

3

4

5

6

7

1 2 3 4 5 6 7 8 9 10 11

Labor Input

Re

al W

ag

e

Labor Demand Curve

High-Cost LaborSupply Curve

Low-Cost LaborSupply Curve

(W/P)0

(W/P)1

N0 N1

Downward shift in labor supply curve reduces real wage and productivity

A

B

Pre-1995: Moving Pre-1995: Moving NorthwestNorthwest

1970-95 EU climbs to the northwest1970-95 EU climbs to the northwest Hours per capita decline, average labor Hours per capita decline, average labor

productivity increasesproductivity increases In this sense much of Europe’s 1970-95 In this sense much of Europe’s 1970-95

productivity catchup was “artificial,” productivity catchup was “artificial,” propelled by policies making labor propelled by policies making labor expensiveexpensive No busboys, grocery baggers, valet parkersNo busboys, grocery baggers, valet parkers Product regulations kept stores shut tight Product regulations kept stores shut tight

many hours of the day/nightmany hours of the day/night All this reduced Europe’s employment share All this reduced Europe’s employment share

in retail/services in retail/services

Post-1995: Moving Post-1995: Moving SoutheastSoutheast

1995-2004 EU slides southeast1995-2004 EU slides southeast Hours per capita start increasing while they Hours per capita start increasing while they

decline in the USdecline in the US Effects are magnified by slow reaction of Effects are magnified by slow reaction of

capital, eventually capital should grow faster capital, eventually capital should grow faster offsetting much or all of productivity slowdownoffsetting much or all of productivity slowdown

Literature #1 misses the turnaroundLiterature #1 misses the turnaround Since 1995 decline in tax rates and Since 1995 decline in tax rates and

employment protection measuresemployment protection measures We are unaware of much macro-level research We are unaware of much macro-level research

on the turnaround in hourson the turnaround in hours Allard and Lindert (2006) do not really Allard and Lindert (2006) do not really

mention it – data only goes to 2001mention it – data only goes to 2001

Literature #2: The EU-US Literature #2: The EU-US ALP gapALP gap

Central Focus of Lit #2 on post-1995 Central Focus of Lit #2 on post-1995 turnaround in US Productivity Growthturnaround in US Productivity Growth Jorgenson, Ho and Stiroh (2006): ’95-’00 Jorgenson, Ho and Stiroh (2006): ’95-’00

due to ICT, ’00-’05 something elsedue to ICT, ’00-’05 something else Retail is often notedRetail is often noted

Van Ark, Inklaar and McGuckin (2003)Van Ark, Inklaar and McGuckin (2003) Foster, Haltiwanger and Krizan (2002) on Foster, Haltiwanger and Krizan (2002) on

new establishmentsnew establishments Baily and Kirkegaard (2004) on Baily and Kirkegaard (2004) on

regulationsregulations Need to free land use restrictionsNeed to free land use restrictions

Fully 85% of EU productivity slowdown Fully 85% of EU productivity slowdown has its counterpart in a speed-up of EU has its counterpart in a speed-up of EU H/NH/N Europe paid for lower ALP mainly with Europe paid for lower ALP mainly with

higher hours rather than less consumptionhigher hours rather than less consumption Saltari and Travaglini have made a similar Saltari and Travaglini have made a similar

point with respect to Italypoint with respect to Italy This runs counter to the Blanchard This runs counter to the Blanchard

story about preferences for leisurestory about preferences for leisure Now we hear that they’re not lazy, just Now we hear that they’re not lazy, just

unproductiveunproductive Huge literature on different structural Huge literature on different structural

reasons for EU sclerosisreasons for EU sclerosis

Literature #3: relationship Literature #3: relationship between Y/H and H/Nbetween Y/H and H/N

There is a long line of research examining There is a long line of research examining the relationship between hours and the relationship between hours and productivityproductivity

Even using an IV approach, increases in Even using an IV approach, increases in H/N drive down Y/HH/N drive down Y/H This makes sense in a single factor model or This makes sense in a single factor model or

with any slow adjustment of capitalwith any slow adjustment of capital Measuring the speed of adjustment of Measuring the speed of adjustment of

investment is difficult – future research for usinvestment is difficult – future research for us View today’s talk as a report on research in View today’s talk as a report on research in

progress, not the final polished wordprogress, not the final polished word

Figure 1. Trends in Output per Hour, Output, and Hours,

U.S. and EU, Anual Growth Rates, 1970-2005

0

1

2

3

4

5

6

1970 1975 1980 1985 1990 1995 2000 2005

Per

cen

t

E.U. Output per Hour

U.S. Output per Hour

Interpreting the Post-1995 Interpreting the Post-1995 TurnaroundTurnaround

Simple HP trendsSimple HP trends Europe is continuing its long slow declineEurope is continuing its long slow decline Turnaround is generally pegged at 1995Turnaround is generally pegged at 1995

The EU-15 stops catching up, and the US The EU-15 stops catching up, and the US takes offtakes off

We are mainly going to examine the We are mainly going to examine the determinants of the turnaround – i.e. determinants of the turnaround – i.e. changes in Y/H growth post-1995changes in Y/H growth post-1995

Qualification: US trend peaks in 2002-03 Qualification: US trend peaks in 2002-03 and is now decliningand is now declining

New US Productivity Trends New US Productivity Trends BasedBased

on March 2007 Quarterly on March 2007 Quarterly DataData

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

NFPB LP

Total economy LP

Difference

-2

-1

0

1

2

3

4

1970 1975 1980 1985 1990 1995 2000 2005

Pe

rce

nt

US Output per CapitaEU-15 Output per Capita

EU-15 Hours per Capita

US Hours per Capita

We Need to Look at We Need to Look at EverythingEverythingPer CapitaPer Capita

Population growth in EU 0.7 percent per Population growth in EU 0.7 percent per year slower than US over the past year slower than US over the past decadedecade

Output per capita in the EU doesn’t look Output per capita in the EU doesn’t look bad at allbad at all

Post-1995 hours turnaround is a Post-1995 hours turnaround is a counterpart to the Y/H turnaroundcounterpart to the Y/H turnaround

We will see that there is a similar pattern We will see that there is a similar pattern withinwithin the EU – strong negative the EU – strong negative correlation between the hours and ALP correlation between the hours and ALP turnaroundsturnarounds

-2

-1

0

1

2

3

4

5

6

7

1984 1988 1992 1996 2000 2004

US Hours

US CapitalEU Capital

EU Hours

The US has experienced an enormous The US has experienced an enormous decline in hours growth when capital decline in hours growth when capital growth fellgrowth fell Thus “capital-deepening” numbers for US Thus “capital-deepening” numbers for US

are misleading as they reflect as much are misleading as they reflect as much movements in the denominator as in the movements in the denominator as in the numerator.numerator.

Cumulative hours growth zero 2000-06, Cumulative hours growth zero 2000-06, growth in hours per capita negativegrowth in hours per capita negative

The EU had strong hours growth while The EU had strong hours growth while the US went through its recession and the US went through its recession and recoveryrecovery

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

1984 1988 1992 1996 2000 2004

US TFP

EU TFP

Defining Tigers and Defining Tigers and Tortoises, Tortoises,

Pop Shares and Private ALP Pop Shares and Private ALP GrowthGrowth

Tigers: Ireland, Finland, GreeceTigers: Ireland, Finland, Greece Pop Share: 5%Pop Share: 5% ALP 4.79%ALP 4.79%

Middle: Sweden, Austria, UK, Middle: Sweden, Austria, UK, Germany, Portugal, FranceGermany, Portugal, France Pop Share: 61%Pop Share: 61% ALP: 2.45%ALP: 2.45%

Tortoises: Belgium, Netherlands, Tortoises: Belgium, Netherlands, Denmark, Luxembourg, Spain, ItalyDenmark, Luxembourg, Spain, Italy Pop Share: 34%Pop Share: 34% ALP: 0.72% ALP: 0.72%

1970-1995 1995-2005 Difference 1970-1995 1995-2005 Difference 1970-1995 1995-2005 Difference

US 1.42 2.30 0.88 0.55 -0.14 -0.69 1.97 2.15 0.18EU 2.89 1.40 -1.49 -0.80 0.55 1.35 2.09 1.95 -0.14Tigers 2.93 2.95 0.02 -0.67 1.22 1.89 2.26 4.17 1.91Middle 2.80 1.86 -0.94 -0.84 -0.08 0.76 1.96 1.78 -0.19Tortoises 3.05 0.39 -2.66 -0.75 1.59 2.34 2.30 1.98 -0.32

Growth RatesGrowth RatesGrowth RatesProductivity Hours per Capita Output per Capita

We break the EU-15 into three groups based on post-’95 Y/H growth: Tigers: Ireland, Finland and Greece Middle Countries: Sweden, Austria, UK, Germany,

Portugal and France Tortoises: BeNeLux, Denmark, Spain and Italy

A closer look at the TortoisesA closer look at the Tortoises

Mainly driven by Spain and ItalyMainly driven by Spain and Italy

Spain:Spain:

►►-4.44% turnaround in Y/H-4.44% turnaround in Y/H

►►+5.01%+5.01% turnaround in H/N turnaround in H/N

Italy:Italy:

►►-2.25%-2.25% turnaround in Y/Hturnaround in Y/H

►►+1.08% turnaround in H/N+1.08% turnaround in H/N Had we ranked the countries Had we ranked the countries

according to output per capita, Spain according to output per capita, Spain would be a Tiger would be a Tiger

Figure 2. Private Economy Labor Productivity Growth by Country: 1979-1995, 1995-2003

1995-20031979-1995

0 1 2 3 4 5 6 7 8 9

Italy

Spain

Luxembourg

Denmark

Netherlands

Belgium

France

Portugal

Germany

United Kingdom

Austria

Sweden

Greece

Finland

IrelandTigers

Middle

TortoisesPre-1995

Post-1995

Making Sense of Cross-EUMaking Sense of Cross-EUHeterogeneityHeterogeneity

Notice the homogeneity pre-1995 and Notice the homogeneity pre-1995 and heterogeneity post-’95heterogeneity post-’95

The only two countries with a noticeable The only two countries with a noticeable acceleration are Greece and Irelandacceleration are Greece and Ireland

Sweden a bit up and UK a bit downSweden a bit up and UK a bit down Sharp declines for France, Portugal, and Sharp declines for France, Portugal, and

all the Tortoisesall the Tortoises For most of the remainder of the paper, we For most of the remainder of the paper, we

focus only on the middle countries and focus only on the middle countries and tortoisestortoises The tigers are special cases – they do not The tigers are special cases – they do not

provide any policy lessons for the rest of the EUprovide any policy lessons for the rest of the EU

The New Results in thisThe New Results in thisPaper at the Industry LevelPaper at the Industry Level

We aggregate productivity growth by We aggregate productivity growth by industry in a way that allows us to determine industry in a way that allows us to determine the relative role of productivity and sharesthe relative role of productivity and shares

The “productivity” effect is just the difference The “productivity” effect is just the difference in productivity growth in a given industryin productivity growth in a given industry

The “share” effect is the addition or The “share” effect is the addition or subtraction from growth as shares shift subtraction from growth as shares shift within industries.within industries. Example: Ireland shifts to high tech Example: Ireland shifts to high tech

manufacturing, this comes out as a “share” effect manufacturing, this comes out as a “share” effect within manufacturingwithin manufacturing

Contributions, Productivity vs. Contributions, Productivity vs. Share Effects, in EU-US, 1995-Share Effects, in EU-US, 1995-

20032003

-0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2

Farms/mining

Const./utilities

Manufacturing

Retail/wholesale

Trans.

Finance

Serv.

Comm.

Real estate

ProdShare

Non-ICT share

Non-durables share

Non-ICT prodICT prod

Non-durables prod

ICT share

Manufacturing is nearly as importantas retail

But ICT is tinyOnly ~2% hours share

ALP growth multiplied by nominal sharesALP growth multiplied by nominal shares

-0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

Real Estate

Communications

Services

Finance

Transportation

Retail/Wholesale

Manufacturing

Construction Utilities

Farms/Mining

U.S.

E.U.

US acceleration is widespread, not just in retailand manufacturing.

EU weakness is also widespread

Tortoises vs. MiddleTortoises vs. Middle

-0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1

Farms/mining

Const./utilities

Manufacturing

Retail/wholesale

Trans.

Finance

Serv.

Comm.

Real estate

Share

Prod

Failure is more widespread.Totally unrelated industries account for the declineNote that this is largely driven by productivity, not share effects

Interpreting the TortoiseInterpreting the TortoiseProblem after 1995Problem after 1995

Failure is across the boardFailure is across the board Consistent with basic theme of paper, that there Consistent with basic theme of paper, that there

is a macro causeis a macro cause How much due to a reduction in taxes and in How much due to a reduction in taxes and in

regulations?regulations? How much remains for an exogenous decline in TFP How much remains for an exogenous decline in TFP

growth?growth? Understanding Share EffectsUnderstanding Share Effects

ICT Share higher in US vs EU and also middle vs ICT Share higher in US vs EU and also middle vs tortoisestortoises

Big EU share deficit in retail/wholesale and services, Big EU share deficit in retail/wholesale and services, consistent with high tax storyconsistent with high tax story

Part of Tiger success is moving resources, out of Part of Tiger success is moving resources, out of agriculture for Greece and Ireland, into ICT mfg agriculture for Greece and Ireland, into ICT mfg for Ireland and Finlandfor Ireland and Finland

Research StrategyResearch Strategy

Divergence across the EU has Divergence across the EU has increasedincreased

The Y/H slowdown in the tortoises in The Y/H slowdown in the tortoises in most countries is balanced by most countries is balanced by healthy H/N growthhealthy H/N growth

We are going to then try to break We are going to then try to break down the determinants of the down the determinants of the middle-tortoise gap in Y/H growth middle-tortoise gap in Y/H growth and relate it to H/N growthand relate it to H/N growth

Qualification: We’re NotQualification: We’re NotDealing with Capital Dealing with Capital

AdjustmentAdjustment ALP Growth = ALP Growth =

ΔΔlabor qualitylabor quality+ + ΔΔcapacity utilizationcapacity utilization+ capital deepening+ capital deepening+ TFP+ TFP

We focus for now on capital deepeningWe focus for now on capital deepening Simple one-factor framework based on the textbook Simple one-factor framework based on the textbook

labor demand curve with fixed capitallabor demand curve with fixed capital Making capital adjustment endogenous next on our Making capital adjustment endogenous next on our

agendaagenda Also next on agenda is tracing link from policy changes Also next on agenda is tracing link from policy changes

to labor quality (e.g., changes in Female LFPR decreases to labor quality (e.g., changes in Female LFPR decreases average labor force experience and perhaps average average labor force experience and perhaps average education)education)

Figure 4. Employment per Capita

90

95

100

105

110

115

120

125

1983 1988 1993 1998 2003

Tortoises

Middle Countries

EU

Hours per Employee

84

86

88

90

92

94

96

98

100

102

1983 1988 1993 1998 2003

Tortoises

Middle Countries

EU

Interpreting the Graphs ofInterpreting the Graphs ofE/N and H/EE/N and H/E

(H/N) = (E/N) * (H/E)(H/N) = (E/N) * (H/E) ’’79-’95 US minus EU H/N growth: 1.01%79-’95 US minus EU H/N growth: 1.01%

Half from employment per capita (E/N), half Half from employment per capita (E/N), half from hours per employee (H/E) from hours per employee (H/E)

US had rising E/N, EU had falling H/EUS had rising E/N, EU had falling H/E ’’95-’04, gap was -.76% (EU had higher 95-’04, gap was -.76% (EU had higher

growth)growth) E/N gap was -.85%, H/E .09%E/N gap was -.85%, H/E .09% Almost entirely explained by a shift up in EU E/NAlmost entirely explained by a shift up in EU E/N H/E seems to have stabilizedH/E seems to have stabilized

So when comparing employment to ALP, So when comparing employment to ALP, E/N is the margin we are going to focus onE/N is the margin we are going to focus on

0

0.2

0.4

0.6

0.8

1

1.2

15 to 19 20 to 24 25 to 34 35 to 44 45 to 54 55 to 64 65+

EU-15Tortoises

Middle Countries

E/N Ratio to the US

-- A lot is explained around 45-54 and 15-19

-- All are very similar for 35-44

Figure 7. Difference in Growth Rates of Employment per Capita by Sex-Age Group, Tortoises minus Middle Countries, 1995-

2005 minus 1985-1995, Employment and Share Effects

Employment Share

-1 -0.5 0 0.5 1 1.5 2

Women 65+

Women 55-64

Women 45-54

Women 35-44

Women 25-34

Women 15-24

Men 65+

Men 55-64

Men 45-54

Men 35-44

Men 25-34

Men 15-24

Contributions to the difference in the turnaround in the Middle countries versus the Tortoises

This is the standard shift-share analysis This is the standard shift-share analysis from industry-level productivity studies from industry-level productivity studies (see Stiroh and van Ark and Inklaar)(see Stiroh and van Ark and Inklaar)

Note that the Tortoises have a big Note that the Tortoises have a big passive advantage – share effects for 25-passive advantage – share effects for 25-3434

Large employment effects for prime age Large employment effects for prime age womenwomen Slightly smaller for prime age menSlightly smaller for prime age men

Teens and retirement aged contribute Teens and retirement aged contribute littlelittle

Male and Female employment rates

Notice the enormous growth in female E/N It even manages to have the biggest

acceleration following 1995 Men in the Tortoises have caught up,

women still have a long way to go

Average Growth RatesMiddle 1985 1995 2005 85-95 95-05 turnaroundMale 65.85 62.30 60.79 -0.55 -0.25 0.31

Female 41.46 44.81 48.09 0.78 0.71 -0.07

Tortoises 1985 1995 2005 85-95 95-05 turnaroundMale 57.72 57.93 60.94 0.04 0.51 0.47Female 26.02 30.97 39.88 1.74 2.53 0.79

Variables to explain E/NVariables to explain E/N Tax wedgeTax wedge EPL – measures of bargaining EPL – measures of bargaining

coordination, firing restrictions, etc.coordination, firing restrictions, etc. Percentage of employees part timePercentage of employees part time

Actually see little evidence of the business cycleActually see little evidence of the business cycle We can see whether part time employees are We can see whether part time employees are

new entrants to the labor forcenew entrants to the labor force Union densityUnion density

Union density and union power aren’t the sameUnion density and union power aren’t the same France has always had lower union density France has always had lower union density

than the US than the US

Explanatory variables are the tax Explanatory variables are the tax wedge, EPL, union density and net wedge, EPL, union density and net reservation wagereservation wage Net reservation wage measures Net reservation wage measures

generosity of unemployment benefitsgenerosity of unemployment benefits

We don’t worry about factors We don’t worry about factors affecting teens or those near affecting teens or those near retirement because those age retirement because those age groups don’t drive much of the groups don’t drive much of the divergence within the EUdivergence within the EU

Figure 5. Tax Wedge

20

25

30

35

40

45

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Tortoises

EU

Middle Countires

Recall Prescott’s claim that the entire Recall Prescott’s claim that the entire gap between EU and US employment gap between EU and US employment can be explained by tax wedgescan be explained by tax wedges

If tax wedges are the main drivers of If tax wedges are the main drivers of employment variation, the employment variation, the compression in EU taxes is interestingcompression in EU taxes is interesting

►►Policy and E/N are converging but Y/H Policy and E/N are converging but Y/H is divergingis diverging

Employment Protection Legislation (EPL)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

EU

Middle Countries

Tortoises

Figure 6. Union Density

20

25

30

35

40

45

1960 1965 1970 1975 1980 1985 1990 1995 2000

Middle Countries

Tortoises

EU

Net Reservation Wage

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

1960 1965 1970 1975 1980 1985 1990 1995 2000

Middle Countries

Tortoises

EU

Interpreting the GraphsInterpreting the Graphsof the Explanatory of the Explanatory

VariablesVariables EPL shows the same convergenceEPL shows the same convergence Union density shows the familiar Union density shows the familiar

declinedecline This is a messy variable because union This is a messy variable because union

power is criticalpower is critical The US has more unions than FranceThe US has more unions than France

The net reservation wage has risen, The net reservation wage has risen, with the Tortoises converging up with the Tortoises converging up rather than downrather than down

Regressions of employment per capita Population weighted, US and Lux. excluded Notice the importance of fixed effects Net reservation wage and EPL have positive

coefficients

VariableTax Wedge -0.51 *** 0.01 -0.68 *** -0.30 ***EPL -0.01 0.10 ***Union Density -0.23 *** 0.15 ***Output Gap 1.12 * 1.88 ** 0.79 1.42 *Net Reservation Wage 0.10 *** 0.06 ***

R2 0.59 0.01 0.66 0.23RMSE 0.135 0.205 0.122 0.181Number of Observations 352 352 352 352Fixed Effects? yes no yes no

E/N regressions by age, FE included Note the effect of the output gap declines with age (see

Jaimovich) Tax wedge has smaller effect on men and prime age

workers Union density almost always has negative effects

Gender Age R2Men 15-24 -1.02 *** 0 -0.05 *** 2.95 *** 0.1 ** 0.81Women 15-24 -1.03 *** 0.02 -0.04 * 2.5 *** 0.14 *** 0.88Men 25-34 -0.23 *** 0.01 -0.02 *** 1.26 *** -0.02 * 0.66Women 25-34 -0.43 *** 0.13 *** 0.08 *** 1.14 ** -0.07 * 0.74Men 35-44 -0.26 *** 0 0.01 0.73 *** -0.04 *** 0.53Women 35-44 -0.8 *** 0.13 *** 0.19 *** 0.56 -0.28 *** 0.82Men 45-54 -0.5 *** -0.03 0.09 *** 0.25 -0.21 *** 0.49Women 45-54 -0.93 *** 0.08 0.23 *** 0.2 -0.54 *** 0.8Men 55-64 -0.43 *** -0.07 ** -0.11 *** 0.77 * 0.19 *** 0.82Women 55-64 -0.67 *** 0 0.01 0.81 * -0.15 *** 0.95Men 65+ -1.26 *** 0.08 -0.47 *** -2.1 0.15 0.78Women 65+ -1.34 *** 0.07 -0.42 *** -1.83 0.18 0.75

Gap DensityTax Wedge EPL WageNet. Res. Output Union

Qualifications for the Next Qualifications for the Next PhasePhase

of the Researchof the Research One problem with all of these regressions is One problem with all of these regressions is

that they have no place for a trendthat they have no place for a trend Any exogenous trends are forced to show up in Any exogenous trends are forced to show up in

the coefficients of trending RHS variablesthe coefficients of trending RHS variables In future work, we need to explore either In future work, we need to explore either

adding a linear trend or some sort of adding a linear trend or some sort of kalman filtered trendkalman filtered trend

We also need to check for coefficient We also need to check for coefficient instabilityinstability Marginal effects may be different at different Marginal effects may be different at different

levels of employmentlevels of employment

Next We Turn to the Next We Turn to the PossiblePossible

Tradeoff of Y/H vs. E/NTradeoff of Y/H vs. E/N We next run regressions of productivity We next run regressions of productivity

growth on employmentgrowth on employment See Gordon(1997), Beaudry and Collard See Gordon(1997), Beaudry and Collard

(2001), McGuckin and van Ark (2005), (2001), McGuckin and van Ark (2005), basically any 1-factor modelbasically any 1-factor model

Even with instruments, the relationship Even with instruments, the relationship is robust across countries and time is robust across countries and time periodsperiods Beaudry and Collard provide evidence that Beaudry and Collard provide evidence that

the coefficient has shifted over timethe coefficient has shifted over time

Regressions of Productivity on Employment

Instruments are explanatory variables from prior regressions

Variable LagsEmployment Rate 0 -0.59 *** -0.52 ***

1 -0.09 -0.1

5 -0.05 -0.02

10 -0.07 -0.05Sum of all Lags -0.81 *** -0.68 *** Standard Error [0.13] [0.13]Sum of Lags 1 and 2 -0.69 *** -0.62 *** Standard Error [0.1] [0.09]

Change in Output Gap 0.82 *** 0.78 ***Ratio to US LP -0.022 *** -0.041 ***

Fixed Effects? no yes

Comments on the Comments on the ProductivityProductivityRegressionsRegressions

Coefficient on employment is -.7 to -.8Coefficient on employment is -.7 to -.8 No bounce back with later lagsNo bounce back with later lags Significant catch-up effectSignificant catch-up effect

Being 10% behind the US adds .2-.4% to Being 10% behind the US adds .2-.4% to ALP growth each yearALP growth each year

Country fixed effects do not affect Country fixed effects do not affect results much, as opposed to results much, as opposed to employment regressionsemployment regressions

We can now ask how policy shifts We can now ask how policy shifts affected productivity growthaffected productivity growth This is very much back of the envelope This is very much back of the envelope

– we need to be more careful in the – we need to be more careful in the futurefuture

Two basic effectsTwo basic effects Policy effectPolicy effect Female cultural effectFemale cultural effect

We can’t identify the total cultural effect We can’t identify the total cultural effect on women; we just get the gap the on women; we just get the gap the middle countries and tortoises:middle countries and tortoises: Take residual male employment growthTake residual male employment growth

Call Middle-Tortoise gap the endogenous partCall Middle-Tortoise gap the endogenous part To get exogenous female growth, take the To get exogenous female growth, take the

Middle-Tortoise gap for female residuals, Middle-Tortoise gap for female residuals, and subtract the endogenous effectand subtract the endogenous effect

Basically, female residual growth minus Basically, female residual growth minus male residual growth equals cultural effectsmale residual growth equals cultural effects

We can consider alternative identifying We can consider alternative identifying assumptions: get the B functions from assumptions: get the B functions from regressionsregressions

EEs,gs,g = A = Ass(POLICY(POLICYgg)+B)+Bss(ALP(ALPgg)+C)+Cs,gs,g S indexes genders {M,F}, G indexes country groups S indexes genders {M,F}, G indexes country groups

{I,T}; C represents cultural forces{I,T}; C represents cultural forces POLICY is the vector of policy variablesPOLICY is the vector of policy variables ALP is labor productivity growthALP is labor productivity growth Lower case letters represent first differencesLower case letters represent first differences

The residuals from the earlier regressions include The residuals from the earlier regressions include the B termsthe B terms

ResidResids,g s,g = e= es,g s,g – A– Ass(policy(policygg)=B)=Bss(alp(alpgg)+c)+cs,gs,g ResidResidM,IM,I-Resid-ResidM,TM,T=B=BM M (ALP(ALPII)- B)- BM M (ALP(ALPMM)) ccF,IF,I-c-cF,TF,T=(Resid=(ResidF,IF,I-Resid-ResidF,TF,T))

-(B-(BFF(alp(alpII)-B)-BFF(alp(alpTT)))) Two identifying assumptions:Two identifying assumptions:

BBMM= B= BFF

ccM,IM,I= c= cM,TM,T=0=0 ccF,IF,I-c-cF,TF,T=(Resid=(ResidF,IF,I-Resid-ResidF,TF,T)-(Resid)-(ResidM,IM,I-Resid-ResidM,TM,T))

Excess employment growth in the Tortoises

Using the above methodology, we get excess female growth of .63% per year

Excess policy driven employment growth of .13% Note the massive overprediction for US

employment growth Short digression on US trends and forecasts

Actual Predicted Actual Predicted Actual Predicted ResidualUS 62.89 61.23 62.34 70.74 -0.10 1.61 -1.70Middle 53.22 53.37 53.96 55.16 0.15 0.37 -0.21Tortoises 44.02 43.33 49.58 45.31 1.32 0.50 0.82

1995 2004 Avg. Growth Rate

Breaking Down the Middle-Breaking Down the Middle-Tortoise GapTortoise Gap

►►.13% gap in predicted .13% gap in predicted ΔΔE/NE/N

→→.1% gap in Y/H.1% gap in Y/H

►►.63% excess female E/N growth.63% excess female E/N growth

→→.48% gap in Y/H.48% gap in Y/H Adding the two exogenous employment Adding the two exogenous employment

shocks and multiplying by .75 gives a shocks and multiplying by .75 gives a predicted shortfall of .58%predicted shortfall of .58%

Of the 1.47 percentage point gap, we Of the 1.47 percentage point gap, we can explain 38% with employment can explain 38% with employment effectseffects

Should we expect this to continue?Should we expect this to continue? Women in the Tortoises still need to Women in the Tortoises still need to

raise employment by 8% to catch up raise employment by 8% to catch up to the middle countriesto the middle countries Translates to a 7.7% total gapTranslates to a 7.7% total gap Implies a further 5.75% shortfallImplies a further 5.75% shortfall Over ten years would imply a shortfall Over ten years would imply a shortfall

of .58% per yearof .58% per year Increased investment would offset Increased investment would offset

some of thissome of this

ConclusionsConclusions Across Europe we find a negative correlation Across Europe we find a negative correlation

between employment and productivity growthbetween employment and productivity growth As labor markets have been liberalized, some As labor markets have been liberalized, some

countries have experienced huge rises in countries have experienced huge rises in employmentemployment

Exogenous shocks can explain about 40% of Exogenous shocks can explain about 40% of the shortfall in ALP in the tortoisesthe shortfall in ALP in the tortoises

Future research needs to identify the sources Future research needs to identify the sources of the other 60%, starting with of the other 60%, starting with a return to the industry-by-industry analysisa return to the industry-by-industry analysis A dynamic analysis of capital adjustmentA dynamic analysis of capital adjustment