europe’s productivity growth slumps but employment surges ian dew-becker, nber robert j. gordon,...
TRANSCRIPT
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