Speed of adjustment to selected labour market and tax reforms.

Mourougane, Annabelle ; Vogel, Lukas

INTRODUCTION

The long-term impact of structural reforms on economic performance has been studied in depth, for example, in the context of the revisited OECD Jobs Strategy. The impact of institutions on economic resilience to temporary shocks is also well documented (Duval et al., 2007; Ernst et al., 2006; Inklaar and Timmer, 2006). By contrast, only few studies have focused on transitional dynamics in the aftermath of structural reforms and the adjustment from one steady state to another (Roeger et al., 2008).

Still, analysing the adjustment path is interesting from a political economy perspective. It gives an idea of the impact lag of reforms on economic performance, putting the long-term benefits into perspective. It is also useful for macroeconomic stabilisation policies and helps gauging complementarities between structural adjustment and monetary or fiscal accommodation.

A number of studies use single-equation cross-country panel or industry-level difference-in-differences estimation (Belot and van Ours, 2001; Bassanini and Dural, 2006; OECD, 2007). They consider transmission channels from structural policies to macroeconomic performance in isolation and link a performance indicator (eg the unemployment rate or productivity) to institutional variables. The results are useful to assess the long-term impact of structural reforms. But they predominantly rely on static equations and they provide little information about the adjustment dynamics. Indeed, institutional variables in panel estimation usually display limited time variation, so that only a small set of variables can be tested at a time. Moreover, these approaches do not properly account for the fact that the impact of a particular institutional feature may differ across countries due to cross-country differences in the remaining institutional structure (Duval and Vogel, 2008; Freeman, 1998). Product terms may capture such interaction between institutions, but their ability to analyse a sequence of reforms is limited (Dreger et al., 2007).

Papers examining the short-term effect of structural reforms on economic performance have generally relied on dynamic neo-Keynesian or micro-founded dynamic general equilibrium (DGE) models (eg Bean, 1998; Coenen et al., 2007; Everaert and Schule, 2006). Neo-Keynesian models incorporate the traditional properties of large macroeconometric models. As the lag structure of these models is determined by the empirical fit, that is, by country specificities, they are generally well-suited to analyse the short-term impact of policy shocks. However, many institutional variables cannot be directly integrated in the model, because they lack sufficient variation over time (tax and expenditure data are notable exceptions). Consequently, simulations are often limited to illustrative Non Accelerating Inflation Rate of Unemployment (NAIRU) or mark-up shocks and aim at describing the adjustment mechanisms at play (Bean, 1998; Dural and Elmeskov, 2005; Hunt and Laxton, 2004). (1)

DGE models are explicitly derived from (assumptions about) firm and household behaviour and allow examining a wide range of structural reforms, while also accounting for potential spillovers between the included variables. (2) Based on structural equations with deep parameters, such models are less subject to the Lucas critique. Given the micro-foundations, it is also possible to assess the welfare consequences of alternative policies. However, the lag structure of DGE model equations derives from optimising behaviour and is thus generally limited and similar across countries and regions. Consequently, the DGE models tend to overemphasise similarities and to attribute differences to shocks rather than to economic structure. Furthermore, the characteristics of the model dynamics and the steady state may be quite sensitive to the choice of particular functional forms and parameters.

Against this background, the paper examines the length of economic adjustment to selected structural reforms, drawing on two macroeconomic models: a micro-founded DGE model and an estimated neo-Keynesian macro model. The paper does not seek to establish a ranking of different approaches, but rather takes an eclectic stance. Each framework presents advantages and limitations to examine the short-term impact of structural reforms. Putting them together brings complementary information, emphasises the different aspects of the topic and ideally can provide some indication about the robustness of the findings. Simulations with these models illustrate how labour and product market rigidities affect the pace of adjustment to structural reforms. Within the estimated neo-Keynesian framework the paper also examines how the conduct of monetary policy affects the speed of adjustment to structural reforms. A final section concludes.

MOST OECD COUNTRIES HAVE EXPERIENCED A DECLINE IN THE STRUCTURAL UNEMPLOYMENT RATE

Institutional data and how they relate to economic performance in OECD countries over the last two decades provide some insights into the ex-post effects of structural reforms. A time-series indicator of the structural unemployment rate can be constructed using the empirical approach undertaken in Bassanini and Duval (2006), which expresses the actual unemployment rate as a function of the output gap and a number of institutional variables, including the tax wedge between labour costs and take-home pay, employment protection legislation (EPL) and product market regulation in non-manufacturing sectors. The resulting structural unemployment rates, which are the unemployment rates net of the cyclical component associated with the output gap, tend to be more volatile than the OECD Economic Outlook NAIRU estimates, which are derived from a Kalman-filter estimation and core price Phillips curves. However, both measures broadly display common patterns. Notwithstanding significant differences for some countries at some points in time (especially for Germany) the associated unemployment gaps generally evolve in line. (3)

[FIGURE 1 OMITTED]

The estimated structural unemployment rate has declined markedly since 1995 in the OECD area, the G7 average and the European Union (Figure 1), in line with a general trend towards product market liberalisation, a gradual decline in the tax wedge and--since the beginning of the decade--a reduction in the average replacement ratio. Arpaia et al. (2007) suggest that these trends have continued in recent years. Contrary to the clear trend towards product market liberalisation, labour market reform across OECD countries display a more heterogeneous pattern. Moreover, some elements of labour market reform may not be sustained but reversed over time (OECD, 2008). This complicates the identification of the short-term impact of structural reforms, as it becomes difficult, for example, to distinguish the weak effect of a particular reform on economic performance from the offsetting impact of subsequent backtracking.

Three institutions appear to have played a major role in the evolution of structural unemployment in OECD countries over the period 1983-2003:

* Among all the institutional variables, the tax wedge has contributed the most to explain the level of structural unemployment in many OECD countries. (4) The evolution of the tax wedge over time and across OECD economies lacks a uniform pattern. The wedge has steadily declined in the United States, Italy and the United Kingdom, but risen in Japan, France and Canada. It has experienced ups and downs in Germany.

* In selected countries the average replacement rate also explains a significant portion of structural unemployment. Again, this variable displays no clear international trend.

* In addition to labour market institutions, product market regulation is estimated to have had a sizable impact on the structural rate of unemployment. A consistent and marked decline in the indicator, signalling the move towards more competitive markets, appears in most OECD economies, especially in Anglo-Saxon countries, where product market reforms have started earlier than in continental Europe.

* By contrast, changes in EPL have played only a minor role, except for a few European countries such as Germany, Italy and the United Kingdom. Indeed, there has been little change to employment protection measure in most countries and the impact of employment protection on the structural unemployment is estimated to be very small.

Correlations between the cumulative increase in the NAIRU between time t-i and t and the change in the institution at time t-i have been computed to provide further insights about the dynamic impact of structural reforms (Figure 2). (5) The correlations have been computed using a small number of observations and may be distorted by the presence of other omitted determinants of structural unemployment. Moreover no causality can be inferred from this calculation. Nonetheless a number of observations can be made:

* Changes in institutions, in particular the tax wedge, the replacement ratio and product market regulation, are associated with gradual changes in the NAIRU. The correlation peaks after 5-10 years, depending on the specific measure.

* In the short-term the strongest correlations are obtained for the tax wedge and for product market regulation. Correlations rise over the first 4 years and then gradually diminish, ending up close to zero after 7 years. By contrast, correlations between the NAIRU and anti-competitive product market regulation, albeit small, continue being significant over 10 years.

* The correlation between changes in the average replacement ratio and changes in the NAIRU is negligible in the short term, but gradually increases over time, implying long lags in the process of adjustment. The correlation becomes significant only after about a decade. (6)

* The correlation between changes in EPL and changes in the NAIRU remains insignificant over a 10-year period. This result holds for EPL on both temporary and regular contracts alike.

[FIGURE 2 OMITTED]

Overall, these patterns suggest the impact of reforms on structural unemployment to be gradual and to spread over a number of years. Structural reforms may require a costly reallocation of resources, so that efficiency gains will take time to fully materialise. The subsequent sections will focus on a number of market rigidities and macro policies and on the degree to which they account for the delay and for country differences in the pass-through of selected labour market and tax reforms.

ADJUSTMENT COSTS IN LABOUR AND PRODUCT MARKETS AND THE SPEED OF ADJUSTMENT

Previous studies have suggested that interactions between different areas of structural reforms are crucial for the long-run aggregate impact. Long-term gains from one reform may increase following the implementation of complementary measures (Bassanini and Duval, 2006). Increasing competition in product markets may also moderate political opposition and resistance to labour market reforms as the former tends to reduce the rents to be redistributed between unions and firms (Blanchard and Giavazzi, 2003). This section focuses on the short run and investigates to what extend policy complementarities may affect the short-term economic impact of structural reforms. In particular it examines the impact of product and labour market flexibility on the pass-through of tax, benefit and product market reforms. (7) The analysis relies on a stylised closed-economy DGE model, calibrated with parameter estimates for the euro area. The framework includes monopolistic competition in product and labour markets, which provides firms and unions with price and wage setting power. Firms use differentiated labour services to produce differentiated goods. Labour is the only production factor and yields constant returns to scale. The household sector consists of Ricardian and of liquidity-constrained consumers.

The basic model includes three mechanisms that delay the adjustment to external shocks: employment adjustment costs, price adjustment costs and habit persistence in private consumption. The introduction of adjustment costs is standard in the DGE literature to proxy for nominal and real rigidities (eg Coenen et al., 2007; Grenouilleau et al., 2007; Campolmi and Faia, 2006; Moyen and Sahuc, 2005). Employment adjustment costs [PHI] seek to capture existing hiring and firing costs for firm j, while price adjustment costs [THETA] account for menu or re-optimisation costs of firm j and generate a forward-looking inflation dynamics. Different specifications for adjustment costs (eg linear versus quadratic, symmetric versus asymmetric) coexist. (8) In line with many other studies, this paper adopts the following quadratic and symmetric costs specification:

[[PHI].sup.j.sub.t] = [phi]/2 [([N.sup.j.sub.t]/[N.sup.j.sub.t-1] - 1).sup.2]

[[THETA].sup.j.sub.t] = [theta]/2 [([N.sup.j.sub.t]/[N.sup.j.sub.t-1] - 1).sup.2]

Quadratic adjustment costs provide an incentive to smooth quantity and price adjustment over time, which delays the transition of employment, production and consumption towards a new steady state in the aftermath of structural reforms, without having an impact on the long-term or steady-state effects of the reform itself. (9) Other functional forms, such as linear or even marginally decreasing adjustment costs, could generate different transition dynamics and generally imply faster and more abrupt adjustments (Cahuc and Zylberberg, 2004). From this perspective, the quadratic adjustment costs specification in this paper may constitute some upper bound for the duration of adjustment and the impact of adjustment costs on the pass-through of structural reforms. The simulations focus on reforms increasing the level of employment rather than on adjustment behaviour over the business cycle, which makes the choice between symmetric or asymmetric specifications less decisive. One major difference between our specification of employment adjustment costs and parts of the literature is that adjustment costs in our model are a function of firm specific output, rather than being a fixed costs or a function of aggregate production. Consequently, adjustment costs are an endogenous variable in the optimisation problem. The costs parameters are set to [[phi].sup.EA] = 63, [[phi].sup.US] = 30, [[theta].sup.EA] = 21 and [[theta].sup.US] = 3.4, using estimates of Grenouilleau et al. (2007). They imply adjustment costs of 0.32% and 0.15 % of GDP per percentage point change in euro area and US employment, respectively. Raising or lowering prices by one percentage point implies costs of 0.11% and 0.02 % of euro area and US output, respectively. This calibration of price adjustment costs is compatible with the degree of euro area and US price rigidity documented in Altissimo et al. (2006) and Bils and Klenow (2004). Habit persistence in consumption has been set to the value of 0.85 from Grenouilleau et al. (2007). Habit persistence delays the adjustment of consumption and aggregate demand to structural shifts. However, as it is a behavioural feature rather than a rigidity the analysis of transitional dynamics will focus on the impact of employment and price adjustment costs.

Mourougane and Vogel (2008) provide a detailed presentation of the model and its microeconomic foundations. The key equations are summarised here:

Consumption:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Wages:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Production:

[Y.sub.t] = [N.sub.t]

Demand:

[Y.sub.t] = [C.sub.t] + [phi]/2 [([N.sub.t]/[N.sub.t-1] - 1).sup.2] [N.sub.t-1] + [theta]/2 [([P.sub.t]/[P.sub.t-1] - 1).sup.2] [Y.sub.t]

Prices:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Government budget:

(1 - [[tau].sup.w.sub.t])(1 - [N.sub.t])[W.sub.t][R.sub.t] + (1 + [i.sub.t])[B.sub.t] = ([[tau].sup.w.sub.t] + [[tau].sup.e.sub.t]) [W.sub.t][N.sub.t] + [[tau].sup.c.sub.t][P.sub.t][C.sub.t] + [T.sub.t] + [B.sub.t|1]

Monetary policy rule:

[i.sub.t] = -1n[beta] + [[phi].sub.[pi]][[pi].sub.t]

with:

               C   Aggregate consumption
       [C.sup.o]   Consumption of Ricardian households
[[lambda].sup.o]   Marginal utility of consumption of Ricardian
                     households
       [C.sup.k]   Liquidity-constraint consumption
[[lambda].sup.k]   Marginal utility of consumption of
                     liquidity-constraint households
               N   Employment
               W   Nominal wage
               P   Price level
   [[tau].sup.c]   Tax on consumption
   [[tau].sup.w]   Tax on labour income
   [[tau].sup.e]   Employer social security contribution
               R   Gross benefit replacement rate
               T   Lump-sum taxes
               Y   Output
               B   Outstanding government bonds
               i   Nominal interest rate
            [pi]   Inflation

[FIGURE 3 OMITTED]

The impact of employment adjustment costs on real adjustment turns out to be moderate. Reducing euro area adjustment costs to the lower US level accelerates the adjustment towards the new equilibrium production and consumption levels (Figure 3). (10) The overall gain from lower employment adjustment costs in terms of faster output and consumption adjustment is relatively modest, however, amounting to only two or three quarters. Differences are more pronounced for inflation, as reductions in income tax rates, benefit replacement rates and employer social security contributions all directly or indirectly reduce production costs. Such measures initially have some deflationary impact, which leads to an accommodating monetary expansion. The simulations suggest that lowering employment adjustment costs in the euro area to their US levels would lead to more pronounced price cuts and a more moderate decline of gross real wages in the transition process. Changing the specification of employment adjustment cost and introducing linear rather than quadratic adjustment costs does not modify this finding in a significant way.

Lowering euro area price adjustment costs to the del lower US level, while leaving employment adjustment costs constant, does not visibly affect the adjustment speed of real variables either. The differences between higher and lower price adjustment costs in the transition dynamics of output, consumption and real wages are negligible. The levels of price adjustment costs almost exclusively affect the adjustment of nominal variables. Lower costs lead to a steeper initial decline in prices and, consequently, to a slightly stronger monetary expansion. Price adjustment costs tend to be more important in the context of temporary shocks, dampening the initial amplitude of the impulse responses and delaying the return to the steady state (Duval and Vogel, 2008).

The previous results also hold in an extended model with nominal wage rigidity, where--in analogy to the Calvo price-setting model--sluggish adjustment takes the form of staggered wage setting (eg Canzoneri et al., 2007; Erceg et al., 2000). Nominal wages W follow:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

[W.sup.*.sub.t] is the optimal wage chosen by a household that resets wages in period t and 1 - [xi] is the probability for such re-setting to occur in a given period. The optimal wage itself follows from a dynamic optimisation problem taking into account that a wage set in period t may remain fixed for a number of subsequent periods (see Mourougane and Vogel, 2008). The impact of lower employment and price adjustment costs on real production and consumption is very small in this scenario (Figure 4). With nominal wage rigidity the transitory decline in real wages is more contained than under fully flexible wages. In contrast to Figure 3, the delay in nominal wage adjustment initially leads to positive inflation and to some monetary tightening.

Both the baseline model and the extension with nominal wage rigidity assume a neoclassical labour market without frictional unemployment. The search-and-matching model provides an alternative specification. It incorporates frictional unemployment and implies a steady turnover of jobs even under constant total employment. Each period a certain share of workers loses or quits a job for new positions or unemployment. Consequently, gross flows are larger than the net flows in and out of employment (see Mourougane and Vogel, 2008, for more details on the modelling). (11) Employment adjustment costs are larger than in the previous specifications if they apply to gross instead of net flows into employment:

[[PHI].sup.g,j.sub.t] = [phi]/2 [([N.sup.j.sub.t]/[N.sup.j.sub.t-1] + [rho] - 1).sup.2]

where [rho] is the job separation rate, set to the value 0.08 taken from Christoffel and Linzert (2005).

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

Within the search-and-matching framework, the role of employment and price adjustment costs in delaying the economic adjustment to structural reforms is also limited. As before, reduced employment adjustment costs have fairly small effects on the adjustment speed of production and consumption, and the contribution of lower price adjustment costs tends to be negligible (Figure 5). However, the differences are more pronounced for price, real wage and interest rate dynamics. Higher employment adjustment costs amplify the downward pressure on real wages. Given the strength of these effects, monetary policy will be more accommodative than in the baseline DGE model with neoclassical labour market.

INTERACTION WITH MONETARY POLICY

The implementation of labour market reforms affects potential output and inflation dynamics and is likely to trigger a macroeconomic policy reaction. The conduct of monetary policy may in turn affect the speed of adjustment to structural reforms, though to a different degree across OECD countries, depending on the strength of the transmission channels and on the sensitivity of policy rates to output and inflation. The expansion and smoothing of aggregate demand may reduce the transitory costs of reforms and the unemployment stemming from the necessary restructuring of particular industries. From a political economy perspective, the ability and willingness of monetary policy to accommodate structural reforms may weaken political opposition and facilitate their implementation.

Simulations undertaken with estimated neo-Keynesian macro models for the United States and the euro area (see Mourougane and Vogel, 2008, for details) illustrate the interaction between structural reforms and monetary policy. The models include equations for all demand components and associated price indicators, employment, labour supply and wages. The behavioural equations are backward-looking. Expectations are treated implicitly by the inclusion of lags of the dependent variables, and most behavioural equations are estimated in an error-correction form. Real short-term rates are determined endogenously following a Taylor rule, with equal weights on inflation and the output gap.

The short-term behaviour of the model is influenced by standard Keynesian features such as wage and price stickiness, liquidity constraints, capital adjustment costs and labour hoarding. Unemployment and the output gap are important determinants of wage and price adjustment. In the short run output is determined by demand. In the medium to long run, the supply side of the economy, which is modelled through a neo-classical production function, plays a prominent role. Prices and wages adjust as well as relative factor prices and incomes. Output and unemployment move back to their long-term equilibrium levels. A decline in the NAIRU increases potential output, leading to a negative output gap (Figure 6). Gaps exert downward pressure on prices and wages. Labour demand rises following the decline in real wages, and--as labour supply increases slowly--the unemployment rate declines. Gaps and their disinflationary impact trigger a monetary easing.

[FIGURE 6 OMITTED]

Simulations with these neo-Keynesian models confirm that different levels of price and employment adjustment costs, which here are captured by different degrees of employment and price inertia in the respective equations, have only a moderate impact on the adjustment speed. Without an endogenous monetary policy reaction, both the euro area and the United States models adjust at a similar pace to a NAIRU shock, notwithstanding the higher price and nominal wage inertia in the euro area.

Expected gains from monetary accommodation are estimated to be negligible for individual euro area economies. As the European Central Bank (ECB) focuses on aggregate euro area output and inflation, any monetary reaction to a reform introduced in an individual European country is improbable unless there was a coordinated effort to implement structural reforms in a sufficient number of euro area economies. This holds for small but also large euro area countries. For instance, a domestic reform lowering the NAIRU by one percentage point in France, which accounts for about 20 % of euro area GDP, would elicit almost no monetary policy response.

The contrast between the two sides of the Atlantic in adjustment speed under monetary accommodation reflects differences in the monetary transmission channels as modelled in the neo-Keynesian models. In line with previous research (eg Angeloni et al., 2003a, b), demand components, especially business investment, appear more sensitive to real interest rates in the United States than in the euro area. Consequently, with supportive monetary easing the United States adjust faster than the euro area to its new equilibrium production level. (12)

Modifying the monetary policy reaction function can alter the pace of adjustment for the euro area. This is illustrated by simulating a NAIRU decline in the euro area under different policy reactions: a Taylor rule with equal weights on current inflation and the output gap, a Taylor rule with stronger weight on inflation, and pure inflation targeting without reaction to output gaps. Increasing the weight of inflation in the policy rule appears to slow the adjustment in the very short term, but to accelerate it thereafter. As a result, the economy reaches its long-term equilibrium substantially faster, but with some overshooting (Figure 7). In the case of pure inflation targeting, the adjustment speed would nevertheless remain slower in the euro area than in the United States.

Interest rate persistence, which implies that central banks are reluctant to move the policy rate too rapidly to stabilise the economy, can decelerate the adjustment to structural reforms. However, both the DGE and the neoKeynesian model simulations suggest that interest rate persistence has to be very high--close to unity--to have a visible effect on the transition speed of the real model variables.

CONCLUSIONS

Economic adjustment to structural reforms is a gradual process. Drawing on two different modelling approaches--micro-founded DGE models and estimated macro-economic neo-Keynesian models--this paper has investigated the extent to which labour and product market rigidities as well as monetary policy affect the adjustment to structural reforms.

Employment adjustment costs are found to have only a limited effect on the pace of adjustment to labour market and tax reforms, while the impact of price adjustment costs on output dynamics appears to be marginal. This result is robust to the choice of the policy variables (income tax, benefit replacement rate or employer social security contributions) and holds in both the DGE and the neo-Keynesian model.

[FIGURE 7 OMITTED]

Macroeconomic policy can have a sizable impact on the magnitude of short-term transition costs. In particular, monetary accommodation can accelerate and smooth the adjustment to a new equilibrium though to a varying degree in different OECD countries and regions. Reforms in individual euro area countries are unlikely to trigger a sizable policy reaction, unless there is an area-wide effort to implement structural reforms. The adjustment is found to be faster in the United States than in the euro area, reflecting mostly a higher sensitivity of domestic demand to real interest rates. The latter result appears to be robust to the choice of the monetary policy rule.

The two approaches can provide complementary insights. Micro-founded DGE models illustrate the transmission of structural reforms. The variation of employment and price adjustment costs seem insufficient to generate sizable heterogeneity in adjustment speed across countries. The neo-Keynesian models have the advantage of using estimated macroeconomic equations for the euro area and the United States to yield a more accurate representation of the time series dynamics. However, the effect of product and labour market institutions on economic dynamics is sometimes captured only very indirectly, using off-model information, and subject to the Lucas critique. Methodological eclecticism is also a way to check the robustness of findings. In this regard, it is reassuring that main findings of the paper, i.e. the limited role of employment and price adjustment costs for the adjustment of real variables, holds independently of the modelling approach.

Acknowledgements

The authors especially thank Jorgen Elmeskov, who had the initial idea for this paper. They are grateful to Jonathan Coppel, Davide Furceri, Claude Giorno, Peter Hoeller, Vincent Koen and Jean-Luc Schneider for helpful discussions and suggestions. We also thank the participants of the 14th Dubrovnik Conference, in particular Maroje Lang.

REFERENCES

Acemoglu, D. 2001: Credit market imperfections and persistent unemployment. European Economic Review 45 (4-6): 665-679.

Altissimo, F, Ehrmann, M and Smets, K 2006: Inflation persistence and price-setting behaviour in the euro area: A summary of the IPN evidence. ECB Occasional Papers, No. 46.

Angeloni, I, Kashyap, A and Mojon, B (eds). 2003a: Monetary policy transmission in the euro area. Cambridge: Cambridge University Press.

Angeloni, I, Kashyap, A, Mojon, B and Terlizzese, D. 2003b: The output composition puzzle: A difference in the monetary transmission mechanism in the euro area and United States. Journal of Money, Credit, and Banking 35(6): 1265-1306.

Arpaia, A, Roeger, W, Varga, J, in't Veld, J, Grolo, I and Wobst, P. 2007: Quantitative assessment of structural reforms: Modelling the Lisbon strategy. European Economy Economic Papers, No. 282.

Bassanini, A and Dural, R. 2006: Employment patterns in OECD countries: Reassessing the role of policies and institutions. OECD Economics Department Working Papers, No. 486.

Bean, C. 1998: The interaction of aggregate demand policies and labour market reforms. Swedish Economic Policy Review 5(2): 353-382.

Belot, M and van Ours, J. 2001: Unemployment and labour market institutions: An empirical analysis. Journal of the Japanese and International Economies 15(4): 403-418.

Bils, M and Klenow, P. 2004: Some evidence on the importance of sticky prices. Journal of Political Economy 112(5): 947-984.

Blanchard, O and Giavazzi, F. 2003: Macroeconomic effects of regulation and deregulation on goods and labour markets. Quarterly Journal of Economics 118(3): 879-907.

Cahuc, P and Zylberberg, A. 2004: Labour economics. MIT Press: Cambridge, MA. Campolmi, A and Faia, E. 2006: Cyclical inflation divergence and different labor market institutions in the EMU. ECB Working Papers, No. 619.

Canzoneri, M, Cumby, R and Diba, B. 2007: The cost of nominal rigidity in NNS models. Journal of Money, Credit and Banking 39(7): 1563-1586.

Christoffel, K and Linzert, T. 2005: The role of real wage dynamics and labor market frictions for unemployment and inflation dynamics. ECB Working Papers, No. 556.

Coenen, G, McAdam, P and Straub, R. 2007: Tax reform and labour-market performance in the euro area: A simulation-based analysis using the new area-wide model. ECB Working Papers, No. 747.

Dreger, C, Artis, M, Moreno, R, Ramos, R and Surifiach, J. 2007: Study on the feasibility of a tool to measure the macroeconomic impact of structural reforms. European Economy Economic Papers, No. 272.

Duval, R and Elmeskov, J. 2005: The effects of EMU on structural reforms in labour and product markets. OECD Economics Department Working Papers, No. 438.

Duval, R, Elmeskov, J and Vogel, L. 2007: Structural policies and economic resilience to shocks. OECD Economics Department Working Papers, No. 567.

Duval, R and Vogel, L. 2008: Oil price shocks, rigidities and the conduct of monetary policy: Some lessons from a new Keynesian perspective. OECD Economics Department Working Papers, No. 603.

Erceg, Ch, Henderson, D and Levin, A. 2000: Optimal monetary policy with staggered wage and price contracts. Journal of Monetary Economics 46(2): 281-313.

Ernst, E, Gong, G, Semmler, W and Bukeviciute, L. 2006: Quantifying the impact of structural reforms. ECB Working Papers, No. 666.

Everaert, L and Schule, W. 2006: Structural reforms in the euro area: Economic impact and role of synchronisation across markets and countries. IMF Working Papers, No. 137.

Freeman, R. 1998: War of models: Which labour market institutions for the 21st century? Labour Economics 5(1): 1-24.

Grenouilleau, D, Ratto, M and Roeger, W. 2007: Adjustment to shocks: A comparison between the euro area and the US using estimated DSGE models. Paper presented at the Workshop on Structural Reforms and Economic Resilience, OECD, Paris, 14 June.

Hunt, B and Laxton, D. 2004: Some simulation properties of the major euro area economies in MULTIMOD. Economic Modelling 21(5): 759-783.

Inklaar, R and Timmer, M. 2006: Resurgence of employment growth in the European Union: The role of cycles and labour market reforms. Economics Letters 91(1): 61-66.

Mourougane, A and Vogel, L. 2008: Speed of adjustment to selected labour market and tax reforms. OECD Economics Department Working Papers, No. 647.

Moyen, S and Sahuc, J-G. 2005: Incorporating labour market frictions into an optimising-based monetary policy model. Economic Modelling 22(1): 159-186.

OECD. 2007: Employment outlook. OECD: Paris.

OECD. 2008: Going for growth. OECD: Paris.

Roeger, W, Varga, J and i'nt Veld, J. 2008: A simulation-based analysis using the Quest model with endogenous growth. European Economy Economic Paper, No. 351.

Disclaimer The views expressed are those of the authors and do not necessarily reflect those of their institutions.

(1) A partial remedy is to use a two-step procedure, where the Non Accelerating Inflation Rate of Unemployment (NAIRU) or mark-up shock is first calibrated off model using external information, for example, on the impact of institutions in a reduced-form unemployment equation, and then simulated within the macroeconomic model. However, this strategy leads to unbiased estimates only if other direct or indirect effects of institutions on economic performance can be neglected.

(2) For instance, Coenen et al. (2007) examine the effects of temporary fiscal measures. Everaert and Schule (2006) use the IMF's global economy model to explore transitory costs of reforms. Imperfect competition in labour and product markets is modelled in a stylised manner through the existence of mark-ups. Similarly, Kilponen and Ripatti (2005) have investigated the quantitative effects of an increase in competition in both product and labour markets. Batini et al. (2005) examine the impact of combined fiscal adjustment and structural reforms for Japan.

(3) Removing the country fixed effects and institutional variables that do not come out significant in the Bassanini and Duval equation would increase the level estimate of structural unemployment, without significantly modifying its profile.

(4) It should be noted, however, that a national-accounts based measure of the tax wedge would lead to a lower contribution.

(5) For this exercise OECD Economic Outlook NAIRUs rather than structural unemployment rates have been used, as the latter are by construction correlated with institutions. NAIRUs and structural unemployment rates usually display similar trends in OECD countries, but levels differences can be observed for some countries.

(6) A simple rule of thumb derived from regression analysis is that the correlation is significant when it exceeds 0.1 in absolute value.

(7) As the transition dynamics are very similar for different reform measures, the discussion of impulse response patterns focuses on labour tax cuts, however. Impulse responses for benefit and product market reforms are displayed in Mourongane and Vogel (2008).

(8) See Cahuc and Zylberberg (2004) for an excellent overview on labour adjustment costs.

(9) An implication is that no long-term complementarities between reforms increasing flexibility and reforms increasing labour supply can be found in our simulations.

(10) Similar results are found for a reduction of the benefit replacement ratio and a cut in employer social security contributions. See Mourougane and Vogel (2008).

(11) Contrary to the baseline model, labour adjustment costs also affect the steady state production and consumption level in the search-and-matching framework. As there are separations in each period, positive adjustment costs will even accrue in the steady state, reducing the level of consumption and equilibrium employment.

(12) Because consumption equations have not been estimated over the same period in the United States and the euro area, the traditional result (eg Angeloni et al., 2003b) that consumption is more sensitive to interest rates in the United States than in the euro area does not apply in this simulation. Hence, differences in the adjustment process between the United States and the euro area in the presence of monetary reaction may even be understated.

ANNABELLE MOUROUGANE (1) & LUKAS VOGEL (2)

(1) OECD Economics Department, (2)0 rue Andre Pascal, 75775, Paris CEDEX 16, France.

(2) DG Economic and Financial Affairs, European Commission, 1049 Brussels, Belgium.