Job Protection Deregulation in Good and Bad Times
  • 1 https://isni.org/isni/0000000404811396, International Monetary Fund

Contributor Notes

Authors’ E-Mail Addresses: rduval@imf.org, dfurceri@imf.org, jjalles@imf.org.

This paper explores the short-term employment effect of deregulating job protection for regular workers and how it varies with prevailing business cycle conditions. We apply a local projection method to a newly constructed “narrative” dataset of major regular job protection reforms covering 26 advanced economies over the past four decades. The analysis relies on country-sector-level data, using as an identifying assumption the fact that stringent dismissal regulations are more binding in sectors that are characterized by a higher “natural” propensity to regularly adjust their workforce. We find that the responses of sectoral employment to large job protection deregulation shocks depend crucially on the state of the economy at the time of reform——they are positive in an expansion, but become negative in a recession. These findings are consistent with theory, and are robust to a broad range of robustness checks including an Instrumental Variable approach using political economy drivers of reforms as instruments. Our results provide a case for undertaking job protection reform in good times, or for designing it in ways that enhance its short-term impact.

Abstract

This paper explores the short-term employment effect of deregulating job protection for regular workers and how it varies with prevailing business cycle conditions. We apply a local projection method to a newly constructed “narrative” dataset of major regular job protection reforms covering 26 advanced economies over the past four decades. The analysis relies on country-sector-level data, using as an identifying assumption the fact that stringent dismissal regulations are more binding in sectors that are characterized by a higher “natural” propensity to regularly adjust their workforce. We find that the responses of sectoral employment to large job protection deregulation shocks depend crucially on the state of the economy at the time of reform——they are positive in an expansion, but become negative in a recession. These findings are consistent with theory, and are robust to a broad range of robustness checks including an Instrumental Variable approach using political economy drivers of reforms as instruments. Our results provide a case for undertaking job protection reform in good times, or for designing it in ways that enhance its short-term impact.

I. Introduction

Job protection regulation matters for productivity, employment, and other important economic and social outcomes such as informality or inequality. However, virtually nothing is known about the short-term impact of deregulation, and even less so about how this impact may vary depending on prevailing business conditions at the time of a given reform. Some have argued that labor market and other reforms could raise output immediately by boosting aggregate demand through expectation effects (Draghi 2015), while others have stressed that deregulation may pay off only very slowly (Rodrik 2015) or even entail short-term costs in the presence of economic slack and constrained macroeconomic policies (Eggertsson, Ferrero, and Raffo 2014; Krugman 2014). In particular, the latter concerns have been expressed against the background of weak labor markets and broader (subdued) economic performance in the aftermath of job protection reforms in Southern European countries.

Theory also points to potential short-term macroeconomic costs from job protection deregulation. In a partial equilibrium framework, Bentolila and Bertola (1990) showed that high firing costs reduce the responsiveness of firms’ lay-offs and employment to business conditions. When facing high (low) costs of dismissing workers, firms will optimally choose to lay off less (more) workers, and adjust their workforce more (less) gradually through attrition, in bad times. More recently, and more closely related to the topic of our paper, in a dynamic stochastic general equilibrium model featuring sticky prices, endogenous firm entry in product markets and search-and-matching frictions in labor markets, Cacciatore and Fiori (2016) and Cacciatore et al. (2016a) show that lowering firing costs triggers immediate layoffs of less productive workers, while their re-employment takes time, resulting in transitory declines in employment and output. Cacciatore et al. (2016b) find that such transitory costs are larger when reform is carried out in a recession, as the adverse impact of relaxing a constraint on layoffs is larger in a depressed economy—in line with Bertola and Bentolila (1990), strict job protection raises the productivity level cut-off below which it becomes profitable for firms to lay off workers, which dampens job losses and has a stabilizing impact on employment and output when the economy is temporarily depressed by an adverse shock.

This paper addresses this gap in the empirical literature. We build a new “narrative” crosscountry dataset of major reforms of job protection legislation for permanent workers covering 26 advanced economies over the period 1970-2013, and estimate the dynamic response of sectoral employment to these shocks using a local projection method (Jordà, 2005)—which has been used recently to study the dynamic impact of macroeconomic shocks such as financial crises (Romer and Romer, 2015) or fiscal shocks (Jordà and Taylor, 2016). The role of macroeconomic conditions is explored using the smooth transition function proposed by Auerbach and Gorodnichenko (2012) to estimate fiscal multipliers in expansions and recessions; here we use this approach to estimate the response to job deregulation shocks instead of fiscal shocks. In this respect, an important novelty and strength of our reforms database is to identify the precise timing and nature of major legislative actions (reform “shocks”), and to so over a much longer period—starting from the 1970s—than had been possible thus far using other sources.

We estimate the response of sectoral employment to large job deregulation shocks using country-sector-level data and applying a differences-in-differences identification strategy à la Rajan and Zingales (1998). Following Basannini et al. (2009), our identifying assumption is that stringent dismissal regulations are more binding in sectors that are characterized by a higher “natural” propensity to regularly adjust their workforce, that is a higher “natural” layoff rate. The advantages of having a three-dimensional (j industries, i countries and t time periods) dataset are twofold:

  • First, it allows us to control for aggregate and country-sector shocks by including country-time (i, t), country-industry (i, j) and industry-time (j, t) fixed effects. The inclusion of the country-time (i, t) fixed effect is particularly important as it controls for any unobserved cross-country heterogeneity in the macroeconomic shocks that affect countries’ employment growth. In a pure cross-country analysis, this would not be possible, leaving open the possibility that the impact attributed to employment protection legislation (EPL) reforms would be due to other unobserved macroeconomic shocks.

  • Second, it mitigates concerns about reverse causality. While it is typically difficult to identify causal effects using cross-country time-series data, it is much more likely that EPL reforms affect cross-industry differences in employment than the other way around; since we control for country-time fixed effects—and therefore for aggregate employment—reverse causality in our set-up would imply that differences in employment across sectors influence the probability of reforms at the aggregate level. Moreover, our main independent variable is the interaction between EPL reforms and industry-specific natural layoff rates; this makes it even less plausible that causality runs from industry-level employment growth to this composite variable.

To further strengthen the causal interpretation of our results, we perform two final exercises. First, we estimate Instrumental Variable (IV) regressions, in which EPL reforms are instrumented by a small set of external variables drawn from the literature on the political economy of reforms, namely demographics, the timing of elections and the political orientation of the government. Second, we check for robustness to several additional controls—including interactions between other reforms and sector-specific natural layoff rates, whose omission could bias our estimates.

Consistent with recent theory, we find that the short-term effects of job protection deregulation vary depending on prevailing macroeconomic conditions at the time of reform—they are positive in an expansion, but become negative in a recession. These findings are robust across different estimation strategies and robustness checks. Our results provide a case for undertaking job protection reform in good times, or for designing it in ways that enhance its short-term impact—for example, by passing reforms that come into force only later when economic conditions improve.

Our paper relates to the extensive literature on the macroeconomic effects of job protection legislation. While strict job protection legislation has been found to reduce productivity (Eslava et al., 2004; Autor et al., 2007; Bassanini et al., 2009; Van Schaik and Van de Klundert, 2013), notably by reducing job turnover (Micco and Pages, 2006; Haltiwanger et al., 2010; Cingano et al., 2010), its impact on aggregate employment remains empirically unclear (for a review, see for example OECD, 2013), in line with the ambiguous effect underscored by theory (for example, Pissarides, 2000). With some exceptions—not least the seminal paper by Lazear (1990), cross-country time-series studies have typically found mixed results (see e.g. Nickell et al., 2005; Bassanini and Duval, 2009; for a review, Boeri et al., 2011). They focus on the long-term impact of (de)regulation and ignore their short-term effect and/or how the latter may vary depending on prevailing macroeconomic conditions—the two issues of focus of the present paper. Focusing more on the short-to-medium term, Blanchard and Wolfers (2000) find some evidence that pre-existing labor market institutions, including job protection, shape the employment effect of macroeconomic shocks in a panel of OECD countries covering four decades. Although quite different in focus and approach, our paper can be seen as addressing the symmetric issue, namely whether pre-existing macroeconomic conditions shape the employment effect of labor market regulation shocks. Closer to our empirical approach, more recent studies have strengthened identification strategies by exploiting the differential impact of legislative changes on different groups of workers, firms, regions and/or industries. Although results vary, they tend to point overall to small adverse negative employment effects of stringent job protection (see Hernanz et al., 2003; Autor et al., 2004, 2006; Kugler and Saint-Paul, 2004; Behaghel et al., 2008). One common issue across those studies that focus on differential effects across groups of workers is the difficulty to account for substitution effects, and therefore to infer aggregate effects. Importantly in our context, our focus on differences across country-industry units mitigates this concern. Most crucially, unlike these papers, ours addresses the question of whether macroeconomic conditions matter for the short-term impact of changes in job protection that has been studied in the most recent literature—see for example Bassanini and Cingano (2016) and Duval and Furceri (2017).

Also related to this paper is the small but growing theoretical literature on the dynamic effects of labor market—including job protection legislation—reforms. A number of small DSGE models, as well as large-scale models built at central banks and international institutions, have been used to explore the dynamic impact of labor market reforms in normal times (for example, Arpaia et al. 2007; Everaert and Schule 2008; Gomes et al. 2013), and most recently at the zero lower bound (Eggertsson et al.2014). However, these studies model job protection deregulation as a compression in wage mark-ups, which rules out any effects stemming from the dynamic impact of deregulation on layoffs vs. hires. Cacciatore and Fiori (2016) address this issue by incorporating explicit search-and-matching frictions in the labor market and, using such a model, Cacciatore et al. (2016b) show that the more depressed the economy is relative to its steady state at the time of reform, the more job protection deregulation entails short-term employment and output losses. The latter conclusion also follows naturally from older partial equilibrium models of labor demand in the presence of shocks and firing, starting with Bentolila and Bertola (1990). To our knowledge, our paper is the first to test empirically this prediction, for which we find strong support.

The remainder of this paper is organized as follows. Section II presents our new dataset of major employment legislation reforms as well as other data used in the empirical analysis. Section III sets up our econometric framework. Section IV provides the main regression results and performs several robustness checks. Section V concludes and elaborates on policy implications.

II. Data

A. Employment Protection Legislation Reforms

Major reforms of employment protection legislation (EPL) for permanent workers are identified by examining documented legislative and regulatory actions reported in all available OECD Economic Surveys for 26 individual advanced economies since 1970, as well as additional country-specific sources.2 In this respect, the methodology is related to the “narrative approach” used by Romer and Romer (1989, 2004, 2010, and 2015) and Devries et al. (2011) to identify, respectively, monetary and fiscal shocks and periods of high financial distress.

In a first step, we identify all legislative and regulatory actions related to regular EPL mentioned in any OECD Economic Survey for any of the 26 countries over the entire sample. Over 100 such actions are identified overall. In a second step, for any of these actions to qualify as a major reform or “counter-reform”—namely a major policy change in the opposite direction—one of the following three alternative criteria has to be met: (1) the OECD Economic Survey uses strong normative language to define the action, suggestive of an important measure (for example, “major reform”); (2) the policy action is mentioned repeatedly across different editions of the OECD Economic Survey for the country considered, and/or in the retrospective summaries of key past reforms that are featured in some editions, which is also indicative of a major action; or (3) the existing OECD regular EPL indicator of the regulatory stance is in the 5th percentile of the distribution of the change in the indicator—or would be in the top 5th percentile if the OECD’s scoring system were applied, but no OECD regular EPL indicator score is available for the country and year considered.

When only the third condition is met, an extensive search through other available domestic and national sources, including through the internet, is performed to identify the precise policy action underpinning the change in the indicator. Following this process, we end up with a variable that, for each country, takes value 0 in non-reform years, 1 in reform years, and −1 in “counter-reform” years. In all cases, implementation years are considered.

An important advantage of this database in our context is to identify the precise timing and nature of major legislative actions taken by advanced economies since the early 1970s. Specifically, compared with existing databases on policy actions in the area of labor market institutions (such as the European commission Labref, the Fondazione Rodolfo de Benedetti-IZA, and the ILO- EPLex database), the approach allows identifying major legislative reforms as opposed to just a long list of actions that in some cases would be expected to have little or no bearing on macroeconomic outcomes. Likewise, compared with an alternative approach that would infer major reforms from large changes in existing EPL indicators produced by the OECD, we are able to identify the exact timing of major legislative actions—including for a few actions that would not be identified by an indicator-based approach altogether—and also have longer time-series coverage, starting in the early 1970s rather than in the mid-1980s. These features are particularly useful for empirical analysis that seeks to identify the short-term effects of reform shocks.

The major strengths of this narrative reform database come with one limitation; because two large EPL reforms can involve different specific actions (for example, a major simplification of the procedures for individual and collective dismissals, respectively), only the average impact across major historical reforms can be estimated. It should also be highlighted that the reform database provides no information regarding the stance of current (or past) EPL, which is not the purpose of this paper.

B. Stylized Facts

Table 1 provides the detailed list of EPL reforms (and counter-reforms) identified using this procedure, and Figure 1 sums up their cross-country and time-series patterns for the entire sample. Looking at Panel A, while some core European countries (“Other Europe”) consistently reformed throughout the four decades under scrutiny, most of the action took place in the 2000s, particularly in Southern European countries after the Global Financial Crisis. In some cases, this was the result of the conditionality embedded in the financial assistance programs provided by international organizations, while in others it partly reflected financial market pressure. In contrast, most counter-reforms took place in the 1970s, as part of a broad tightening of labor market regulations amid rising job losses in the wake of the oil shocks (Panel B).

Table 1.

Major Reforms of Job Protection Legislation for Permanent Workers in 26 Advanced Economies over 1970-2013

article image
article image
article image
article image
article image
article image
Figure 1.
Figure 1.

EPL Reforms and Counter-Reforms by Groups and Periods

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: “Southern Europe” includes Greece, Portugal, Italy and Spain. “Other Europe” includes Austria, Belgium, France, Germany, Ireland, Luxembourg, Netherlands and Switzerland. “English Speaking” includes U.S., U.K., New Zealand and Australia. “Nordic” includes Denmark, Norway, Sweden, Finland and Iceland. “Other” includes Korea, Japan, Czech Republic and Slovak Republic.

Finally, and importantly in the context of our empirical analysis, the implementation of major EPL reforms does not appear to have depended significantly on economic conditions (Table 2). In fact, while the share of EPL reforms is almost identical between periods of high and low economic growth, when it comes to counter-reforms, these tended to be implemented mostly during expansionary times (such as the boom years immediately before the crisis).

Table 2.

Reforms and Counter-reforms Over the Business Cycle (%)

article image
Note: lower (higher) economic growth = real GDP growth below (above) the reforming country’s sampleverage.

C. Other Data

Our sector-level series of output and employment come from the EU KLEMS and World KLEMS databases. Data on layoff rates are taken from Basannini et al.(2009) and are computed based on industry-level U.S. layoff rates reported in the 2004 CPS Displaced Workers Supplement. While relying on U.S. layoff rates can be considered a good proxy for underlying layoff propensity in the absence of dismissal regulations, one potential problem with this approach is that they may not be representative for the whole sample—that is, U.S. layoff rates may be affected by U.S.-specific regulations or sectoral patterns. To check for the sensitivity of the results to this assumption, the analysis is replicated using U.K. layoff rates computed from U.K. Labor Force Surveys in the robustness checks section further below.

III. Econometric Framework

Two econometric specifications are used to assess the short-term effects of EPL reforms. The first establishes whether these reforms have significant effects on sectoral employment. The second assesses whether these effects vary with overall business cycle conditions prevailing at the time of the reform. In both cases, following Bassanini et al.(2009), our identification assumption is that stringent dismissal regulations are likely to be more binding in sectors that are characterized by a higher “natural” propensity to adjust their workforce to idiosyncratic shocks, that is a higher “natural” layoff rate.3 This differences-in-differences identification strategy is in the spirit of Rajan and Zingales (1998).

The statistical method follows the approach proposed by Jordà (2005) to estimate impulse-response functions. This approach has been advocated by Auerbach and Gorodnichencko (2013) or Romer and Romer (2015), among others, as a flexible alternative to vector autoregression (autoregressive distributed lag) specifications since it does not impose dynamic restrictions. It is also particularly suited to estimating nonlinearities (including interactions between shocks and other variables of interest) in the dynamic response. The first regression specification is estimated as follows:

yi,j,t+kyi,j,t1=αi,t+γi,j+μj,t+βkνjRi,t+θXi,j,t+εi,j,t(1)

in which yi,j,t+k is log employment in sector j of country i in period t+k; αi,t are country-time fixed effects, which control for any variation that is common to all sectors of a country’s economy, such as country-wide macroeconomic shocks and reforms in other areas, including other types of labor market reforms; γi,j are country-sector fixed effects, included to take account of cross-country differences in the average employment growth of certain sectors; μj,t are sector-specific time dummies to control for different employment growth rates of different industries; Ri, t is our EPL reform variable, which takes value 0 in non-reform years, 1 in reform years and −1 in counter-reform years; νj is the “natural” layoff rate, which is taken to be the U.S layoff rate reported in the 2004 CPS Displaced Workers Supplement—the analysis also presents robustness checks using a measure of natural layoff rates for the UK, computed from waves of the Quarterly UK Labor Force Survey (for details see Bassanini et al., 2009); Xi,j,t is a set of controls including two lags of sectoral employment growth and two lags of EPL reforms interacted with sector-specific “natural” layoff rates.

In the second specification, the dynamic response is allowed to vary with the state of the economy:

yi,j,t+kyi,j,t1=αit+γij+μj,t+βkLF(zi,t)νjRi,t+βkH(1F(zi,t))νjRi,t+θMi,j,t+εi,j,t(2)

with

F(zit)=exp(γzit)1+exp(γzit),γ>0

in which zit is an indicator of the state of the economy normalized to have zero mean and unit variance. Following Auerbach and Gorodichenko (2012), the indicator of the state of the economy is GDP growth (or the unemployment rate in some robustness checks), and Fit is a smooth transition function used to estimate the macroeconomic impact of major reform shocks in expansions versus recessions. They further argue for setting γ = 1.5, which we also use. In the robustness checks, we present the results based on alternative measures of economic slack (unemployment) and using a dummy to identify the effects across business cycle regimes, rather than a smooth transition function. M is the same set of control variables used in the baseline specification, but now including also the interaction term between Fit and the sectoral-specific layoff rate.

This approach is equivalent to the smooth transition autoregressive model developed by Granger and Terävistra (1993). The advantage of this approach is twofold. First, compared with a model in which each dependent variable would be interacted with a measure of the business cycle position, it permits a direct test of whether the effect of reforms varies across different regimes such as recessions and expansions. Second, compared with estimating structural vector autoregressions for each regime it allows the effect of reforms to change smoothly between recessions and expansions by considering a continuum of states to compute the impulse response functions, thus making the response more stable and precise.

Equations (1) and (2) are estimated for each k=0,..,4. Impulse response functions are computed using the estimated coefficients βk, and the confidence bands associated with the estimated impulse-response functions are obtained using the estimated standard errors of the coefficients βk, based on clustered standard errors at the country-sector level. The equations are estimated using OLS as the inclusion of the rich set of fixed effects is likely to largely address the endogeneity concerns related to omitted variable bias. In addition, reverse causality is unlikely to be a concern in our set-up. First, the natural propensity to layoff in the U.S. is arguably orthogonal to sectoral employment growth in other countries. Second, it is very unlikely that sectoral employment patterns can influence EPL reform. Aggregate employment may well do so, and employment growth may co-move across all sectors, but this potential source of reverse causality is addressed through the inclusion of country-time fixed effects. In other words, claiming reverse causality would mean arguing that differences in employment growth across sectors lead to EPL reforms; this, we argue, is unlikely. Nonetheless, one possible remaining issue in estimating equations (1) and (2) with OLS is that other macroeconomic variables might affect sectoral employment growth when interacted with industries’ natural layoff rates. This, in particular, may apply to reforms of EPL for temporary workers—these may correlate with regular EPL reforms, while affecting employment growth differently across sectors depending on their natural layoff rate. In addition, in order to further mitigate any endogeneity concerns, we also re-estimate our specifications using an IV approach, drawing our instruments from the political economy literature on the drivers of reforms. The next section starts with baseline OLS regression results, followed by robustness checks and IV estimates.

IV. Main results and robustness checks

A. Baseline

Figure 2 presents the results obtained by estimating equation (1). It shows that over the medium term—that is, four years after the reform takes place—EPL reforms for regular contracts tend to increase employment in industries with a higher propensity to layoff relative to those with a low layoff rate. The results suggest that the differential medium-term employment gain between an industry with a relatively high natural layoff rate (at the 75th percentile of the cross-sector distribution of layoff rates in the U.S) and one with a relatively low natural layoff rate (at the 25th percentile of the distribution) is about 1 percent. However, these unconditional effects mask considerable variation depending on business cycle conditions, as shown by the OLS estimation of equation (2) reported in Figure 3. Under strong conditions, reforms have a sizable positive impact on employment, whereas their impact becomes contractionary when undertaken during periods of slack. This heterogeneity in the response is quantitatively large. As an illustration, when EPL is deregulated during a strong expansion—in which the smooth transition function F takes value 0, the differential medium-term effect on employment between a sector that has a relatively high propensity to lay off (at the 75th percentile of the distribution of natural layoff rates) and one that has relatively low propensity to lay off (at the 25th percentile) is about 4 percent (Figure 3, Panel B). By contrast, this differential effect of EPL reforms on sectoral employment becomes negative, reaching −2 percent after two years, during a major recession—in which F takes value 1. This negative effect is statistically significant in the first two years following the reform. The difference in the response of sectoral employment between strong and weak business cycle conditions is statistically significant (at 5 percent) at each horizon k=0,..4.

Figure 2.
Figure 2.

The Effect of EPL Reforms on Sectoral Employment (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (1). Solid line denotes the differential employment effect of reform between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level.
Figure 3.
Figure 3.

The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2). Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using F=0 (F=1). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the unconditional response presented in Figure 2.

As noted above, the theoretical rationale for this asymmetric effect of deregulation across different economic regimes stems is that reform affects differently firms’ hiring versus firing incentives in good and bad times. In a recession, firms seek to dismiss more and hire less than in a boom, but stringent job protection partly discourages them from laying off (Bentolila and Bertola, 1990). Relaxing that constraint triggers a wave of layoffs, increasing unemployment, weakening aggregate demand and delaying the recovery (Cacciatore et al., 2016b).

B. Robustness Checks and Instrumental Variables Estimation

Using U.K. natural layoff rates

As a first robustness check, we replicate the analysis using sector-specific layoff rates for the U.K. According to the EPL indicators published by the OECD, dismissal procedures in the UK are the second least regulated in the OECD after the U.S, thus making U.K sectoral layoff rates a valid alternative measure. The results obtained estimating equations (1) and (2) are presented in Figures 4 and 5. They are largely similar to, and not statistically different from, those reported in the baseline.

Figure 4.
Figure 4.

The Effect of EPL Reforms on Sectoral Employment: Using U.K Layoff Rates (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (1). Solid blue line denotes the differential employment effect of reform between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black line indicates baseline response presented in Figure 2.
Figure 5.
Figure 5.

The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Using U.K Layoff Rates (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2). Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using F=0 (F=1). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the corresponding responses (using U.S layoff rates) presented in Figure 3.

Omitted variable bias

As discussed above, a possible concern in estimating equations (1) and (2) is that the results are biased due to the omission of other reforms affecting sectoral employment through the sector-specific natural layoff rate that may at the same time be correlated with EPL reforms. A prime candidate is EPL reform for temporary workers. In order to enhance the labor market prospects of disadvantaged groups such as youth, governments have often chosen to deregulate temporary contracts either as an alternative to, or—of potential concern in our context—in combination with EPL reform for regular contracts, particularly during the 1990s and the first half of the 2000s (see e.g. OECD, 2006). Deregulating temporary contracts may have a positive short-term effect even in recessions, since it increases incentives to hire (temporary workers) without affecting incentives to lay off—giving rise to a transitory, positive “honeymoon effect” on employment that eventually vanishes in the steady state (Boeri and Garibaldi, 2007).

To test whether the results are robust to the inclusion of EPL reforms for temporary workers, we re-estimate equations (1) and (2) adding the interaction effect between these reforms and the sectoral-specific layoff rate (and the state of the business cycle in equation (2)). Data on major reforms of EPL for temporary contracts, as well as major reforms in other areas considered further below, are taken from Duval and others (forthcoming). The results presented Figure 6 and 7 show that the effect of regular EPL reforms on sectoral employment are very close to, and not statistically different from, those reported in Figure 2 and 3. Interestingly, and somewhat surprisingly, we do not find any significant effect of employment protection reforms for temporary workers on sectoral employment, even during expansions/recessions.

Figure 6.
Figure 6.

The Effect of EPL Reforms on Sectoral Employment: Controlling for Reforms of EPL for Temporary Contracts (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (1). Solid blue line denotes the differential employment effect of reform between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black line indicates baseline response presented in Figure2.
Figure 7.
Figure 7.

The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Controlling for Reforms of EPL for Temporary Contracts (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2). Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using F=0 (F=1). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the corresponding responses presented in Figure 3.

Another potential candidate is unemployment benefit reform. Reducing the generosity of the unemployment benefit system can increase employment by strengthening job search incentives (e.g. Pissarides 2000), and some recent evidence suggests this effect may be weaker or even negative in the short-term if benefits are cut during a recession (Duval and Furceri 2017). However, the omission of these reforms from our baseline specification could bias our estimates only insofar as they are correlated with EPL reforms and also tend to have larger effects in industries with higher natural layoff rates. While this is far from obvious a priori, we still check the robustness of our results by adding in equations (1) and (2) the interaction of these reforms with the sector-specific natural layoff rate as well as with the state of the business cycle.4 The results reported in Figure 8 and 9 show that the inclusion of these controls does not significantly affect our results.

Figure 8.
Figure 8.

The Effect of EPL Reforms on Sectoral Employment: Controlling for Unemployment Benefit Reforms (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (1). Solid blue line denotes the differential employment effect of reform between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black line indicates baseline response presented in Figure2.
Figure 9.
Figure 9.

The effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Controlling for Unemployment Benefit Reforms (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2). Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using F=0 (F=1). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the corresponding responses presented in Figure 3.

By the same token, the omission of product market reforms could potentially be a source of omitted variable bias. We check the robustness of our results by adding in equations (1) and (2) the interaction of these reforms with the sector-specific natural layoff rate as well as with the state of the business cycle. Data on major product market reforms are drawn from Duval and others (forthcoming), as used in Duval and Furceri (2017). The results reported in Figure 10 and 11 are very similar to, and not statistically different from, those reported in the baseline.

Figure 10.
Figure 10.

The Effect of EPL Reforms on Sectoral Employment: Controlling for Product Market Reforms (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (1). Solid blue line denotes the differential employment effect of reform between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black line indicates baseline response presented in Figure2.
Figure 11.
Figure 11.

The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Controlling for Product Market Reforms (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2). Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using F=0 (F=1). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the corresponding responses presented in Figure 3.

Identification of slack

We check the sensitivity of our results to alternative measures of economic slack. Based on data availability considerations and the fact that it is generally accepted as a measure of underutilization, we use the unemployment rate, instead of the GDP growth rate, to compute the smooth transition function across business cycle regimes (for a similar approach, see Ramey and Zubairy, forthcoming). The results presented in Figure 12 show that the responses of sectoral employment to EPL reforms across business cycle regimes are rather similar to those obtained in the baseline. They are somewhat larger than, but not statistically different from, the baseline responses.

Figure 12.
Figure 12.

The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Using Unemployment Rather than GDP Growth to Measure Business Cycle Conditions in the Smooth Transition Function (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2). Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using F=0 (F=1). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the corresponding responses presented in Figure 3.

We also re-estimate equation (3) without measuring business cycle conditions through a smooth transition function, but instead more simply through a dummy variable that takes value 1 when the GDP growth rate of the country considered is below its sample average and zero otherwise. Again, the estimated responses are somewhat larger than, but not statistically different from the baseline responses (Figure 13).

Figure 13.
Figure 13.

The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Using a Dummy Variable Rather Than a Smooth Transition Function to Measure Business Cycle Conditions (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2). Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using a dummy variable than takes value 1 when a country’s GDP growth is below its sample average, and zero otherwise. Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the corresponding responses presented in Figure 3.

Instrumental Variables Estimation

Even though, as discussed earlier, we believe that our empirical set-up addresses endogeneity concerns, we still check the robustness of our results to an IV approach that instruments EPL reforms. Specifically, we use the following set of variables as external instruments: (i) the share of old-age population (Age), which is the percentage of the population aged 65 and over—taken from the World Bank’s World Development Indicators; (ii) the number of months to the next legislative elections (Months)—taken from the World Bank’s Database on Political Institutions; and a dummy for the political orientation of the government that takes value 1 for left-wing parties and 0 for center-right parties (Parties) —also taken from the Database on Political Institutions.

We proceed in two steps. In the first step, we regress EPL reforms on these indicators, controlling for time and country fixed effects. The results of the first stage suggest that these instruments can be considered as sufficiently “strong instruments”—that is, the joint F-test is above the indicative value (10) identified by Staiger and Stock (1997).5 In the second step, we re-estimate equations (1) and (2) using the exogenous component of EPL reforms driven by these instruments—that is, the fitted value of the first step.6 The results reported in Figure 14 and 15 are similar to, and not statistically different from those obtained using OLS, confirming that endogeneity is not a serious concern in our set-up.

Figure 14.
Figure 14.

The Effect of EPL Reforms on Sectoral Employment: IV Estimates (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (1) with IV. Solid blue line denotes the differential employment effect of reform between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black line indicates baseline response presented in Figure2.
Figure 15.
Figure 15.

The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle: IV Estimates (%)

Citation: IMF Working Papers 2017, 277; 10.5089/9781484333013.001.A001

Note: estimates based on equation (2) with IV. Solid blue lines denote the differential employment effect of reform under weak (Panel A) and strong (Panel B) business cycle conditions between a sector with a high natural layoff rate (at the 75th percentile of the distribution) and a sector with a low natural layoff rate (at the 25th percentile of the distribution). Effects under strong (weak) business cycle conditions are shown here using F=0 (F=1). Dotted lines indicate 90 percent confidence interval based on standard errors clustered at country-sector level. Solid black lines denote the corresponding responses presented in Figure 3.

V. Conclusion

This paper estimated the short-term impact of deregulating regular job protection and how it varies with prevailing business conditions, applying a local projection method to a new dataset of major reforms in advanced economies spanning over four decades, and using as an identifying assumption the fact that stringent dismissal regulations are more binding in sectors that are characterized by a higher “natural” layoff rate. Our main finding is that employment rises when reform is undertaken during economic expansions, but declines when reform happens during recessions. This result is robust to a battery of sensitivity checks, including to IV estimation using political economy drivers of reforms as instruments. It is also consistent with economic theory that, to our knowledge, had been untested thus far.

To the extent that streamlining job protection for regular workers yields long-term economic gains and that reform is therefore worth pursuing, what do our results imply for its timing and design? Under adverse macroeconomic conditions, other than postponing job protection deregulation until better times, one way to address short-term costs may include passing—or credibly announcing—today a reform that will come into force only the future. A more extreme option is to grandfather the new legislation altogether, that is, to apply it only to new contracts—a design feature of many of the post-Global Financial Crisis reforms of employment protection legislation in Europe, notably in Italy, Portugal and Spain where some of the provisions of the reforms of the 2010s were grandfathered. Such strategies have the advantage of frontloading the positive hiring effect of reform while delaying its adverse impact on layoffs. However, these short-term gains should be weighed against the possible efficiency losses from a more gradual phasing-in of the reform. We leave a cost-benefit analysis of these reform strategies for future theoretical and empirical research.

Job Protection Deregulation in Good and Bad Times
Author: Mr. Romain A Duval, Davide Furceri, and João Tovar Jalles
  • View in gallery

    EPL Reforms and Counter-Reforms by Groups and Periods

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment: Using U.K Layoff Rates (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Using U.K Layoff Rates (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment: Controlling for Reforms of EPL for Temporary Contracts (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Controlling for Reforms of EPL for Temporary Contracts (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment: Controlling for Unemployment Benefit Reforms (%)

  • View in gallery

    The effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Controlling for Unemployment Benefit Reforms (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment: Controlling for Product Market Reforms (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Controlling for Product Market Reforms (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Using Unemployment Rather than GDP Growth to Measure Business Cycle Conditions in the Smooth Transition Function (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle Conditions: Using a Dummy Variable Rather Than a Smooth Transition Function to Measure Business Cycle Conditions (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment: IV Estimates (%)

  • View in gallery

    The Effect of EPL Reforms on Sectoral Employment Depending on Business Cycle: IV Estimates (%)