Towards completing EMU: The case for wage coordination ... · Motivation Theoretical Background Data & Empirical Approach Discussion & Conclusion Motivation European debt crisis brought

Motivation Theoretical Background Data & Empirical Approach Discussion & Conclusion

Towards completing EMU: The case for wagecoordination among member states

Martin Gachter1, Alexander Gruber2 and Aleksandra Riedl1

1Oesterreichische Nationalbank

2University of St. Gallen / Stanford University

Empirical Economics and Econometrics SeminarUniversity of Innsbruck, March 2014

The views expressed are strictly those of the authors and do in no way commit the

Oesterreichische Nationalbank (OeNB) or the Eurosystem.


Motivation

European debt crisis brought optimum currency area (OCA) theoryback to the center of economic policy debates

• What policy measures are needed to stabilize the euro area?

• Reform of EMU governance structures

→ European Semester→ Macroeconomic Imbalances Procedure (MIP)→ European Systemic Risk Board (ESRB)→ Reform of the Stability & Growth Pact, Fiscal Compact→ Banking Union→ EFSF, ESM→ Europe 2020 etc.

Martin Gachter (OeNB) The case for wage coordination among EMU member states 1 / 23


Macroeconomic Scoreboard - MIP

Source: Alert Mechanism Report 2014, European Commissionhttp: // ec. europa. eu/ europe2020/ pdf/ 2014/ amr2014_ en. pdf


http://ec.europa.eu/europe2020/pdf/2014/amr2014_en.pdf


Macroeconomic Scoreboard - MIP (ctd.)

Source: Alert Mechanism Report 2014, European Commissionhttp: // ec. europa. eu/ europe2020/ pdf/ 2014/ amr2014_ en. pdf


http://ec.europa.eu/europe2020/pdf/2014/amr2014_en.pdf


Motivation (ctd.)

Empirical literature on OCAs

• Business cycle (BC) synchronization as a ’meta-criterion’ for enteringa currency union (see e.g. de Haan, 2008)

Are EMU governance reforms in line with OCA theory?

• Determinants of business cycle synchronization

→ Bilateral trade relations (Frankel & Rose, 1998)→ Fiscal policy (Darvas et al., 2005)→ Financial integration (Imbs, 2004)→ Industrial specialization (Krugman, 1991)→ EMU membership (Gachter & Riedl, 2013)→ etc.



Motivation (ctd.)

Empirical literature on OCAs

• Labor market developments and competitiveness - while at the centerof policy discussions - have not been examined so far as a majordeterminant of BC synchronization.

Research question: The impact of diverging labor costs on (bilateral) BCsynchronization

• First empirical study that examines this empirical relationship both forEMU and non-EMU countries

• New BC synchronization measures allow panel estimation methods

• Empirical results show that labor cost developments are crucial forEMU member countries, while being insignificant for other countries



Theoretical Background I: OCA

Early contributions: Several prerequisites for successful monetaryintegration

• price and wage flexibility (Friedman, 1953)

• mobility of factors of production including labor (Mundell, 1961)

• financial market integration (Ingram, 1962)

• similarities of inflation rates (Fleming, 1971)

• fiscal integration (Kenen, 1969)

• high degree of economic openness (McKinnon, 1963)

• criteria are endogenous (Frankel & Rose, 1998)

• etc.

→ All criteria intended towards facilitating cyclical adjustment→ High level of BC synchronization lowers costs of common currency



Theoretical Background II: OCA and Labor Markets

Early OCA theory emphasized crucial role of labor markets

• Crucial role of labor mobility (Mundell, 1961)

• “The principal danger ... in participating in a fixed exchange rate area arises fromthe certainty, in the absence of perfect competition in product and factor markets,that developments would occur from time to time that pushed the relative costlevels of the participating countries out of line.” (Fleming, 1971)

• Similar inflation rates more likely with increasing similarity of

- national employment goals,- rates of productivity growth, and- the degree of trade union aggressiveness.

→ Divergencies in relative nominal unit labor costs (NULC) more likely if trade unionsand employers’ organizations are organized on a national basis. (Fleming, 1971)



The Role of Divergent Labor Cost Dynamics in EMU

• Labor cost divergence would require corresponding exchangerate adjustments

• Members of a currency union lose their control over monetary policyand the possibility of exchange rate movements to adjust to internal(inflation) and external (trade deficits) imbalances.

• Recent studies show that BC correlation increases with increasingexchange rate volatility (de Haan et al, 2002; Bergman, 2006)

→ Crucial role of labor market institutions and wage setting mechanisms



Labor Cost Developments in EMU



NULC and BC Synchronization I

• Pre-crisis boom phase: The internal perspective

- Peripheral countries recorded above-average increases in NULC- Persistent inflation differentials led to differences in real interest rates- Stimulation of domestic demand, credit growth and housing markets- Increasing demand further drives up NULC and causes wage-driven

inflation→ Self-reinforcing vicious circle→ Further facilitated by capital market failure of converging nominal

interest rates across countries (Keuschnigg, 2012)

→ NULC divergence decreases BC sychronization

• Inflation differential within EMU were commonly explained by theBalassa-Samuelson effect

- But: Inflation differentials originated from the non-tradeable sector,indicating domestic demand booms (reversed sequence)



NULC and BC Synchronization II

• Post-crisis bust phase: The external perspective

- Inflation differentials in a CU stemming from wage developmentsdirectly translate into changes in real exchange rates

- Decrease of export market shares, deterioration of trade and currentaccount balances

- Once again, the adjustment mechanism of nominal exchange rates isnot available

- Over time, decrease in competitiveness dampens domestic booms, thefollowing bust is considerably larger

→ Effect further accelerated by global financial crisis, diverging nominalinterest rates and sharp decline in external demand

→ Rigid labor market policies and nominal wage stickiness furtherimpeded necessary adjustments

→ NULC divergence decreases BC sychronization



Labor Cost Developments and BC Synchronization

→ How do NULC developments affect (bilateral) BC co-movement?



Measuring BC Synchronization on an Annual Basis

Adapted version of a synchronization index developed by Cerqueira &Martins (2009):

Correlij ,t =1

2log

(1 +

ρij,t2T−3

1− ρij ,t

)

ρij,t = 1− 1

2

cj,t − c j√1T

∑Tt=1(cj,t − c j)2

− ci,t − c i√1T

∑Tt=1(ci,t − c i )2

2

cj,t ... cyclical component of real GDP (using HP- & BK-filter) / real GDP growth rates

It can be shown that 1T

∑t ρij,t = ρij with ρij =

Cov(ci ,cj )

σ(ci )σ(cj )



Measuring BC Synchronization on an Annual Basis

(a) AT&DE: 1T

∑t ρ

Cycleij,t = 0.93 (b) PT&SK: 1

T

∑t ρ

Cycleij,t = 0.34



Panel Data Sample

• 27 EU countries (excl. Croatia), i.e. 27×(27−1)2 = 351 country pairs

• Annual observations, period 1993− 2011

• Overall observations: 6669



Main Explanatory Variables

• Nominal Unit Labor Costs

Labor costsij,t = | ˙nulc i,t − ˙nulc j,t |

where ˙nulc i(j),t =nulci(j),t−nulci(j),t−1

nulci(j),t−1∗ 100

• EMU membership

EMUij,t =

{1, if i and j are EMU members in year t,

0, otherwise.

• Interaction term

Labor costsij,t × EMUij,t



Main Explanatory Variables (ctd.)

Sample averages for country pairs by EMU membership, 1993-2011

Variable Full sample Non-EMU EMU

Correlij,t 1.26 1.21 1.50Labor costsij,t 7.70 8.84 2.13

Compensation per employeeij,t 10.58 12.41 2.00

No. of observations 6,048 4,983 1,065

Partial Correlation Coefficients for Business Cycle Synchronization

nominal unit labor costs compensation per employee

Labor costsij,t -0.04*** -0.10***Labor costsij,t × EMUij,t -0.05*** -0.07***

EMUij,t 0.09*** 0.10***



Control Variables

• Bilateral trade

Bilateral tradeij,t =Exportsij,t + Exportsji,t

GDPi,t + GDPj,t

• Fiscal policy differences

Fiscal policyij,t = |fbcai,t − fbca

j,t |where fbca

i(j),trepresents the cyclically adjusted fiscal balance (net lending/net borrowing in

percent of GDP) of country i(j) at time t

• Financial integration

Financial integrationij,t =Ai,t + Li,t + Aj,t + Lj,t

GDPi,t + GDPj,t

where Ai(j),t and Li(j),t represent a country i(j)’s total external assets and liabilities(portfolio equity, foreign direct investment, debt and financial derivatives) in year t (seeLane & Milesi-Ferretti, 2007)



Empirical Model

Correlij ,t = α + β1Correlij ,t−1 + β2Labor costsij ,t + β3EMUij ,t

+β4(Labor costs× EMU)ij ,t + β5Zij ,t + µij + λt + νij ,t

Zij,t ... ijtth observation on three control variablesµij ... country-pair specific effectsλt ... time effectsνij,t ... error term

Feasible system GMM estimator (Blundell & Bond, 1998)

• endogenous variables: EMU, labor costs, interaction term, bilateral trade

• arbitrary pattern of heteroskedasticity

• external instruments for trade (distance, common border, population size)

• number of instruments is restricted (up to five time lags)



Empirical Results

No Baseline Altern. ULC Altern. ULC BK-filter Unfiltered Excluding RE-modelcontrols measure 1 measure 2 GDP growth crisis

(1) (2) (3) (4) (5) (6) (7) (8)

Labor costs -0.002 -0.007* -0.001 -0.001 -0.001 0.007 -0.005 -0.004*(-1.08) (-1.81) (-0.28) (-0.46) (-0.13) (1.58) (-1.39) (-1.74)

EMU member 0.499*** 0.289*** 0.340*** 0.378*** 0.663*** 0.303*** 0.644*** 0.212***(6.83) (3.41) (3.66) (4.59) (6.11) (3.22) (6.78) (3.36)

Labor costs × EMU -0.133*** -0.108*** -0.092*** -0.145*** -0.212*** -0.143*** -0.236*** -0.052***(-4.90) (-3.55) (-2.96) (-4.50) (-6.02) (-4.09) (-7.41) (-2.63)

Bilateral Trade 0.074*** 0.069*** 0.072*** 0.053*** 0.051*** 0.058*** 0.064***(5.12) (4.47) (5.03) (3.45) (3.96) (3.60) (5.94)

Fiscal Policy -0.014** -0.022*** -0.022*** 0.007 -0.003 -0.006 -0.042***(-2.35) (-3.92) (-3.87) (0.88) (-0.53) (-0.90) (-8.20)

Financial integration 0.189*** 0.065 0.161*** 0.083* 0.106*** 0.081* 0.056***(4.69) (1.63) (3.88) (1.73) (2.62) (1.94) (2.66)

Correlt−1 0.158*** 0.073*** 0.083*** 0.063*** 0.039** 0.138*** 0.093***(4.64) (4.62) (4.89) (3.96) (2.06) (4.51) (5.51)

Obs 6196 5701 4974 5537 4606 5703 4299 5446Hansen 337.3212 341.9314 344.3582 348.3925 343.5011 343.6728 340.7910

Hansen p 0.1053 0.8119 0.9812 0.7379 0.7952 0.7107 0.4625AR(1) p 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000AR(2) p 0.8056 0.2333 0.0930 0.2370 0.8300 0.0999 0.1147



Robustness Check: Country Exclusion

Country Labor costs Labor costs×EMU Country Labor costs Labor costs×EMU

Austria -0.007* -0.110*** Ireland -0.007* -0.130***(0.004) (0.035) (0.004) (0.035)

Belgium -0.009** -0.105*** Italy -0.006* -0.092***(0.004) (0.032) (0.004) (0.034)

Bulgaria -0.006 -0.109*** Lithuania -0.011*** -0.107***(0.004) (0.031) (0.004) (0.031)

Cyprus -0.006 -0.123*** Latvia -0.010*** -0.112***(0.004) (0.031) (0.004) (0.031)

Czech Rep. -0.007* -0.109*** Luxembourg -0.008** -0.102***(0.004) (0.030) (0.004) (0.030) )

Germany -0.006 -0.083** Malta -0.008** -0.098***(0.004) (0.034) (0.004) (0.033)

Denmark -0.006* -0.106*** Netherlands -0.007* -0.089***(0.004) (0.031) (0.004) (0.032)

Estonia -0.009** -0.113*** Poland -0.004 -0.114***(0.004) (0.031) (0.004) (0.032)

Spain -0.007* -0.110*** Portugal -0.007* -0.119***(0.004) (0.035) (0.004) (0.031)

Finland -0.006 -0.095*** Romania -0.002 -0.103***(0.004) (0.034) (0.005) (0.032)

France -0.006* -0.107*** Sweden -0.007* -0.101***(0.004) (0.033) (0.004) (0.031)

UK -0.005* -0.101*** Slovakia -0.005 -0.122***(0.004) (0.031) (0.004) (0.031)

Greece -0.007* -0.089** Slovenia -0.006 -0.125***(0.039) (0.040) (0.004) (0.029)

Hungary -0.004 -0.114***(0.004) (0.032)

Sensitivity of Labor Cost Coefficients to Country Exclusion



Discussion & Conclusion

• Early literature highlighted several prerequisites for a country group toconstitute an OCA

- BC synchronization as a “meta-criterion” to assess the appropriatenessof a CU

- Remaining prerequisites are still important to facilitate economicadjustment in case of asymmetric shocks

• Some countries experienced unsustainable pre-crisis booms

- Policy-makers partly reckoned those developments as being part of awelcomed catching-up process

- Policy institutions in the EA were also powerless to correct imbalances



Discussion & Conclusion (ctd.)

• Impact of ULC dynamics on BC synchronization has so far beenneglected by empirical literature

- Empirical results show a strong effect of ULC dynamics within EMU- ULC developments hardly matter for countries outside a CU where

exchange rates can adjust accordingly

• Results add to current debate on structural reforms in EMU

- Crucial role of (national) wage-setting mechanisms and institutions- Alert Mechanism Report (AMR) / Macroeconomic Imbalances

Procedure as an important step in the right direction- Nominal thresholds for each variable are however insufficient to monitor

relative wage developments- Further efforts needed to establish effective monitoring systems


Towards completing EMU: The case for wage coordination ... · Motivation Theoretical Background Data & Empirical Approach Discussion & Conclusion Motivation European debt crisis brought

Documents