## 1. Germany: Employment Rate

## Source details

**Canonical URL:** [1. Germany: Employment Rate](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1662.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1662.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1662.pdf.json)

---

### Introduction and identification challenges
- Structural reform effects are difficult to pin down empirically because of endogeneity between reform timing and the economic environment.
- Endogeneity can create:
  - upward bias if reforms are implemented just before a cyclical upswing; and
  - downward bias if reforms are implemented early during a downturn (reforms observed when employment or growth is low).
- Business-cycle interactions and macro policy responses can alter estimated impacts: reforms during low aggregate demand can produce small or negative short-run effects; supportive macro policy can sustain aggregate demand and increase positive effects.

### Methodology used in this study
- Local projection (LP) techniques estimate dynamic effects by computing cumulative impact on the change of employment over a 5-year horizon (h = 0, 1, ... , 5).
- Endogeneity addressed using augmented inverse probability weighting (AIPW) to estimate treatment effects while controlling for selection bias.
- Controls include lagged change in the employment rate (3 lags), lagged output gap, output loss during crisis (Laeven and Valencia, 2013), country and year fixed effects; Driscoll and Kraay standard errors used.
- Reform shocks defined from OECD indices (range 0 to 6); reform dummy = 1 when change exceeds two standard deviations (labor market reforms floor reduced to one standard deviation).

### Baseline empirical findings (LP baseline)
- Key headline: "By year 5, the reform increases the employment rate by 1.5 percentage points."
- Labor market reforms (EPLR) — cumulative impact on employment rate (coefficients; t-statistics):
  - Year 0: -0.185  (-1.15)
  - Year 1: -0.311  (-0.97)
  - Year 2: 0.103   (0.19)
  - Year 3: 0.705   (1.55)
  - Year 4: 1.233** (2.59)
  - Year 5: 1.468** (2.53)
- Product market reforms (PMR) — cumulative impact on employment rate (coefficients; t-statistics):
  - Year 0: 0.134   (1.37)
  - Year 1: 0.261*  (1.70)
  - Year 2: 0.444*  (1.93)
  - Year 3: 0.645*** (2.75)
  - Year 4: 0.781*** (3.14)
  - Year 5: 0.964*** (3.60)
- Observations and sample:
  - Labor market reforms: Observations by year — Year 0: 555; Year 1: 555; Year 2: 555; Year 3: 526; Year 4: 497; Year 5: 468; Number of countries: Year 0–4: 29; Year 5: 24.
  - Product market reforms: Observations by year — Year 0: 709; Year 1: 709; Year 2: 709; Year 3: 683; Year 4: 657; Year 5: 631; Number of countries: 26 (all years).
- Interpretation:
  - Near-term impacts generally not significant; positive impacts emerge in the medium term (years 3–5).

### AIPW (doubly robust) estimates addressing endogeneity
- First-stage probit predictors: lagged GDP growth, legislative election dummy, forward EU accession dummy, age dependency ratio (share aged 65 and older), political leader’s education dummy.
- AIPW LP results — cumulative impact on employment rate (coefficients; t-statistics):
  - Labor market reform:
    - Year 0: 0.014    (0.09)
    - Year 1: -0.162  (-0.77)
    - Year 2: -0.149  (-0.36)
    - Year 3: 0.296   (0.53)
    - Year 4: 1.123** (2.20)
    - Year 5: 1.276** (2.15)
    - Observations: Year 0: 555; Year 1: 555; Year 2: 555; Year 3: 526; Year 4: 497; Year 5: 468
  - Product market reform:
    - Year 0: -0.019  (-0.20)
    - Year 1: 0.115   (0.66)
    - Year 2: 0.457*  (1.85)
    - Year 3: 0.750** (2.42)
    - Year 4: 0.866*** (2.66)
    - Year 5: 1.188*** (3.27)
    - Observations: Year 0: 709; Year 1: 709; Year 2: 709; Year 3: 683; Year 4: 657; Year 5: 631
- AIPW summary interpretation:
  - By year 5 AIPW estimates imply increases in the employment rate of 1.276 percentage points for LMR and 1.188 percentage points for PMR.
  - Confirms lagged but positive impacts and largely corroborates LP baseline.

### State dependence: interaction with the business cycle
- Cyclical dummy I = 1 when output gap < –1 percent ("bad times"), 0 otherwise ("good times").
- LP with interaction — select results (coefficients; t-statistics):

  - Labor market reform, no slack:
    - Year 0: -0.0373  (-0.133)
    - Year 1: 0.529**  (2.194)
    - Year 2: 1.069**  (2.148)
    - Year 3: 1.243**  (2.276)
    - Year 4: 1.344**  (2.095)
    - Year 5: 1.886*** (3.287)

  - Labor market reform, slack:
    - Year 0: -0.436**  (-2.23)
    - Year 1: -1.563*** (-3.92)
    - Year 2: -1.853*** (-3.19)
    - Year 3: -1.027**  (-2.33)
    - Year 4: -0.024    (-0.03)
    - Year 5: 0.193     (0.24)

  - Product market reform, no slack:
    - Year 0: 0.180*   (1.780)
    - Year 1: 0.245*   (1.826)
    - Year 2: 0.365**  (2.166)
    - Year 3: 0.696**  (2.744)
    - Year 4: 0.755**  (2.526)
    - Year 5: 0.953**  (2.702)

  - Product market reform, slack:
    - Year 0: 0.143   (0.80)
    - Year 1: 0.051   (0.15)
    - Year 2: -0.143  (-0.27)
    - Year 3: -0.145  (-0.23)
    - Year 4: -0.034  (-0.07)
    - Year 5: -0.032  (-0.12)

- Interpretation:
  - Labor market reforms in bad times (slack) produce immediate and statistically significant negative effects on employment (years 0–3).
  - Product market reforms in no-slack periods show positive medium-term effects; PMR in slack periods do not show significant negative employment losses.
  - Timing correlates with reform intensity: average declines in OECD indicators are larger when reforms are initiated during good times.

### State-dependent AIPW results (business cycle)
- AIPW estimates by business cycle state — select results (coefficients; t-statistics):

  - Labor market reform, slack:
    - Year 0: -0.831    (-0.640)
    - Year 1: -2.196    (-0.968)
    - Year 2: -4.674*   (-1.838)
    - Year 3: -4.951    (-1.406)
    - Year 4: -7.957*   (-1.939)
    - Year 5: -11.94*   (-1.907)

  - Labor market reform, no slack:
    - Year 0: -0.00324  (-0.0224)
    - Year 1: 0.00188   (0.00635)
    - Year 2: 0.584     (0.848)
    - Year 3: 1.637*    (1.686)
    - Year 4: 2.391**   (2.112)
    - Year 5: 2.778**   (1.961)

  - Product market reform, slack:
    - Year 0: -0.448    (-1.467)
    - Year 1: -0.374    (-0.742)
    - Year 2: 0.208     (0.312)
    - Year 3: 0.594     (0.736)
    - Year 4: 0.730     (0.876)
    - Year 5: 1.330     (1.535)

  - Product market reform, no slack:
    - Year 0: 0.105     (0.794)
    - Year 1: 0.306     (1.096)
    - Year 2: 0.590     (1.426)
    - Year 3: 0.826*    (1.678)
    - Year 4: 0.758*    (1.660)
    - Year 5: 1.020**   (2.014)

- Caveat:
  - Limited number of LMR-slack episodes (12 LMR-slack vs 16 LMR-non-slack) may affect first-stage probit stability; PMR counts: maximum of 34 PMR-slack vs 63 PMR-non-slack occurrences.
- Interpretation:
  - AIPW corroborates that reforms are most effective when implemented in periods of limited slack (no slack); LMR during slack can show large negative AIPW coefficients (interpret cautiously).

### Role of macroeconomic policy stance (fiscal and monetary) — AIPW interactions
- General finding:
  - Structural reforms produce stronger positive employment effects when implemented with non-restrictive (supportive) fiscal or monetary policy; effects weaken or reverse under restrictive stances.

- PMR × Fiscal policy (AIPW) — select results (coefficients; t-statistics):
  - PMR, non-restrictive fiscal policy:
    - Year 0: -0.0962  (-0.800)
    - Year 1: -0.108   (-0.483)
    - Year 2: 0.258    (0.919)
    - Year 3: 0.633**  (1.981)
    - Year 4: 0.668**  (1.995)
    - Year 5: 1.113*** (2.764)
  - PMR, restrictive fiscal policy:
    - Year 0: 0.325   (0.791)
    - Year 1: -0.104  (-0.206)
    - Year 2: -0.524  (-0.761)
    - Year 3: -1.097  (-1.300)
    - Year 4: -1.667**(-1.970)
    - Year 5: -2.271***(-2.818)
  - Interpretation: PMR raises employment in the medium term under non-restrictive fiscal policy; under restrictive fiscal policy PMR yields negative effects 4–5 years after reform.

- LMR × Monetary policy (AIPW) — select results (coefficients; t-statistics):
  - LMR, non-restrictive monetary policy:
    - Year 0: -0.471** (-2.214)
    - Year 1: -0.723***(-3.735)
    - Year 2: -0.294  (-1.233)
    - Year 3: 0.513*  (1.770)
    - Year 4: 1.540*** (3.682)
    - Year 5: 2.897*** (6.026)
  - LMR, restrictive monetary policy:
    - Year 0: 0.163   (0.672)
    - Year 1: 0.204   (0.475)
    - Year 2: -0.0234 (-0.0299)
    - Year 3: -0.887**(-2.013)
    - Year 4: -0.543  (-1.089)
    - Year 5: -0.666  (-0.977)
  - Interpretation: Under non-restrictive monetary policy, LMR yields strong positive medium-term employment gains; under restrictive monetary policy, these gains disappear and can be negative.

- PMR × Monetary policy (AIPW) — select results (coefficients; t-statistics):
  - PMR, non-restrictive monetary policy:
    - Year 0: -0.017  (-0.16)
    - Year 1: 0.199   (1.05)
    - Year 2: 0.563** (2.25)
    - Year 3: 0.956*** (2.96)
    - Year 4: 0.999*** (2.88)
    - Year 5: 1.191*** (2.81)
  - PMR, restrictive monetary policy:
    - Year 0: -0.084  (-0.59)
    - Year 1: -0.049  (-0.26)
    - Year 2: -0.006  (-0.03)
    - Year 3: -0.132  (-0.53)
    - Year 4: 0.012   (0.04)
    - Year 5: 0.390   (1.06)
  - Interpretation: PMR yields positive medium-term employment effects when monetary policy is non-restrictive; positive effects disappear under restrictive monetary policy.

- Macro-policy synthesis:
  - Results align with simulation/theoretical work: restrictive monetary policy (e.g., binding ZLB absent non-conventional measures) attenuates or reverses positive output/employment effects of reforms; non-restrictive policy (including active asset purchase programs at ZLB) enables positive effects.

### Policy implications and conclusions
- Robust empirical evidence (LP and AIPW) indicates:
  - Structural reforms (LMR and PMR) tend to increase employment rates by about one percentage point over a five-year horizon in baseline LP estimates, and by about 1.276 (LMR) and 1.188 (PMR) percentage points by year 5 in AIPW estimates.
  - Positive effects are lagged, materializing mainly in the medium term (years 2–5).
- State dependence:
  - Reforms are more effective when implemented in periods of limited slack (no slack) and when supported by non-restrictive fiscal and monetary policies.
  - Labor market reforms implemented during slack can produce immediate and sizable negative employment effects.
- Policy recommendation:
  - Consider initiating structural reforms in conjunction with supportive fiscal or monetary policy if policy space is available to enhance the probability of realizing medium-term employment gains.
- Limitations and open questions:
  - Data limitations constrain assessment of reform sequencing, sectoral heterogeneity, and interactions among different structural reforms; richer sectoral-level reform data are needed.

*Source: IMF working paper chapter "1. Germany: Employment Rate" (excerpt) from _wp1662.*

### 1. Germany: Employment Rate.............................................................................................

### 1. Germany: Employment Rate

### Introduction and identification challenges
- Structural reform effects are difficult to pin down empirically because of endogeneity between reform timing and the economic environment.
- Example: the Hartz reforms in Germany were implemented during a downturn in the early 2000s; employment appears to have risen afterwards, but the effect of the reform must be disentangled from the subsequent cyclical recovery.
- Endogeneity can create:
  - upward bias if reforms are implemented just before a cyclical upswing; and
  - downward bias if reforms are implemented early during a downturn (reforms observed when employment or growth is low).
- The business cycle can affect both the magnitude and direction of reform impacts:
  - Reforms may free up resources that cannot be absorbed when aggregate demand is low, producing small or negative short-run effects.
  - Reforms launched during recessions can add uncertainty (for example, due to required legal clarification).
- Macroeconomic policy responses (monetary or fiscal) during reform episodes can alter estimated impacts; supportive macro policy can reduce uncertainty and sustain aggregate demand, potentially increasing positive effects.

### Methodology used in this study
- Local projection techniques are used to estimate dynamic effects of structural reform by computing cumulative impact on the change of employment over a 5-year horizon.
- Endogeneity is addressed using the augmented inverse probability weighting (AIPW) method to estimate treatment effects while controlling for potential selection bias.
- The role of the business cycle and macroeconomic policies is examined by explicitly controlling for these variables.

### Empirical findings
- Structural reforms have a lagged but positive impact on employment.
- The positive effect remains after accounting for endogeneity of the decision to reform.
- Both labor and product market reforms increase employment rates by about a little over one percentage point over 5 years.
- Supportive macroeconomic policy plays an important role in reaping the medium-term benefit of labor and product market reforms by enhancing their impact on employment.
- The short-run effects of structural reforms are mixed and limited in the literature; some studies find business-cycle-dependent effects that vary in size and direction across studies.

### Context from related model-based studies (as referenced)
- Model-based analyses (e.g., using IMF’s GIMF or EC’s QUEST) can quantify impacts under different scenarios.
- Example cited: Hobza and Mourre (2010) find that structural reforms could boost real GDP growth from 1.7 percent to 2.2 percent between 2010 and 2020 (employment gains range from 1 percent to [text truncated in source]).

### Policy implications
- Account for endogeneity and business-cycle timing when evaluating and designing structural reforms.
- Consider initiating some structural reforms in conjunction with supportive fiscal or monetary policy if policy space is available, to enhance the probability of realizing medium-term employment gains.

*Source: IMF working paper chapter "1. Germany: Employment Rate" (excerpt).*

### 4.5 percent. Anderson and others (2014), using the GIMF model, find that structural reforms

### _wp1662 - 4.5 percent. Anderson and others (2014), using the GIMF model, find that structural reforms

### Model-based evidence on structural reforms
- Anderson and others (2014), using the GIMF model, find that structural reforms in the euro area can increase real GDP, but it will take time for the benefits to fully materialize.
- Gomes and others (2013) use a multi-country general equilibrium model of the euro area and find that increasing competition in the labor and services markets in Germany and the rest of the euro area would increase long-run output.
- Eggertsson and others (2014), using a standard dynamic stochastic general equilibrium model, show that structural reform do not increase output during a crisis; their simulations show that structural reforms may have negative impact when monetary policy is constrained by the zero lower bound (ZLB).
- Decressin and others (forthcoming) find that structural reforms have a positive impact under quantitative easening.

### Empirical findings on long-run effects
- Bouis and Duval (2011) (drawing on Bassanini and Duval (2006), Basannini and others (2009), Bourles and others (2010b)) report:
  - A gradual alignment of product market regulations with best practice in a broad range of non-manufacturing sectors could boost aggregate labor productivity by several percent over 10 years in many OECD countries and by more than five percent across most of continental Europe.
  - Labor market reforms in unemployment benefit systems, activation policies, labor taxes and pension systems could raise employment rates by several percentage points in many OECD countries over a 10 year horizon if these reforms are phased in faster.
  - The impact of structural reforms is larger in the long run (10 years) than in the medium run (5 years).

### Short-run effects and heterogeneity
- OECD (2012b) notes that benefits from structural reforms usually take time to fully materialize, but seldom involve significant losses and often deliver gains already in the short run.
- Some argue short-run effects of structural reforms could be negative especially when slack in the economy is large; structural reforms may free up resources that cannot be absorbed in more efficient sectors.
- Theory on policy types:
  - Unemployment benefits and activation policies: likely to boost employment rates relatively quickly.
  - Job protection reforms: ambiguous near-term effects; layoffs likely to rise in the short-run if legal or regulatory constraints are relaxed.
  - Product market reforms: ambiguous short-run effects—inefficient firms may exit, but new entrants may invest and create jobs (OECD (2012a)).
- Model-based studies and empirical evidence:
  - Anderson et al. (2014) find weak demand conditions could dampen the short-run impact of structural reform; reforms initiated in weaker initial demand conditions can have very little positive and possibly negative impact on growth and employment even in the medium run.
  - Bouis and others (2012a) find structural reforms deliver short-run benefits for certain reforms: increased spending on ALMP employment incentives raises employment even in the short run; reductions in unemployment benefit duration also boost employment relatively quickly.
  - Bouis and others (2012a, 2012b) report tentative evidence that reductions in initial unemployment benefit replacement rates and job protection reforms pay off more in good times than in bad times and can entail short-term losses in severely depressed economies.

### Measuring structural reform shocks (OECD indicators and identification)
- OECD indices:
  - Range from 0 to 6 to capture the restrictiveness of regulation in labor and product markets.
  - Computed as a weighted sum of scores assigned to several underlying criteria.
  - A higher value indicates more restrictive regulation; the introduction of a reform is represented by a fall of the index.
- Reform shock definition and properties used in this paper:
  - Reform shock identified as a drop in the OECD index; reform variable is a dummy equal to one when a reform shock is observed.
  - Large: a change is considered a reform shock if it exceeds two standard deviations of the change in the indicator over all observations. Exception: the floor for labor market reforms was reduced to one standard deviation because there are far fewer labor market reforms than product market reforms in the sample.
  - Discrete: reform shock represented by a dummy variable (intensity of reform not captured).
  - Unsequenced: does not address reform sequencing or capture reform reversals; focus on drops in OECD indicators ignores episodes where the indicator later increases.
  - Non-sectoral: deals with aggregate macroeconomic indicators and outcomes, not sectoral effects.
- Data coverage and counts:
  - Data for 36 countries from 1960 to 2013.
  - 28 reforms in employment protection legislation for regular workers (EPLR).
  - 102 reforms in product market reforms (PMR).
- Examples illustrated:
  - Spain EPLR indicator captured labor market reforms applying to regular workers in 1995 and 2011.
  - UK PMR indicator shows gradual declines from early 1980s to early 2000s; only 3 reform episodes are picked up because the declines were in small steps.

### Caveats on OECD indicators
- Limitations:
  - Complexity of employment protection legislation can be difficult to summarize in an index.
  - Interactions across reforms may not be well captured.
  - Product market indicator summarizes reforms in seven industries in energy, communication, and transportation sectors; it does not capture reforms outside these industries.
- Implication: analysis focuses on aggregate macro level and may miss sectoral heterogeneity or interaction effects.

### Methodology for estimating reform impacts
- Outcome variable: change in the employment rate (employment rate defined as the ratio of total employment to the labor force).
- Data sources:
  - Employment and real GDP growth from the IMF World Economic Outlook (WEO) database.
  - Output loss during financial crises from Laeven and Valencia (2013).
- Econometric approach:
  - Local projection (LP) estimates (Jorda, 2005) used to estimate the impact of reform shocks over horizons h = 0, 1, ... , 5.
  - Advantages of LP: flexibility to accommodate nonlinear or state-dependent impacts; useful for investigating booms vs slumps and interactions with fiscal and monetary policies.
  - LP framework allows for identification strategies using treatment effect methods (as in Jorda and Taylor (2013)) to reduce endogeneity bias.
- Baseline specification and controls:
  - Control variables include lagged change in the employment rate, the output gap, a banking crisis dummy variable, and country and year fixed effects.
  - Models estimated using a sample of about 30 OECD countries observed over the period 1980–2013.
  - Driscoll and Kraay (1998) standard errors are computed to account for correlations in the error terms.
- Interpretation of coefficients:
  - Coefficients ߠ௛ measure the impact of reforms on the cumulative change in the employment rate at each horizon starting in year h = 0 up to 5 years after the reform is identified.
- Reported baseline result preview:
  - Table 1 (described) reflects cumulative impact of the reform on the employment rate on years 0 to 5.
  - For labor market reforms proxied by changes in the EPLR indicator, the impact is not significantly different from zero in years 0 to 3; it becomes significantly positive in year 4.

*Source: _wp1662 - 4.5 percent. Anderson and others (2014), using the GIMF model, find that structural reforms (PDF chapter/section).*

### 5. By year 5, the reform increases the employment rate by 1.5 percentage points. This is the

### _wp1662 - 5. By year 5, the reform increases the employment rate by 1.5 percentage points. This is the

### Baseline empirical findings (LP baseline)
- Key statement:
  - "By year 5, the reform increases the employment rate by 1.5 percentage points."
  - For product market reforms, "the impact is positive and statistially significant in years 1 to 5, with the employment rate rising by about 1 percentage point by year 5."
- Table 1: Effect of Reform on the Employment Rate (Baseline) — Dependent variable: Deviation in employment rate relative to pre-reform year
  - Labor market reforms (coefficients; t-statistics in parentheses):
    - Year 0: -0.185  (-1.15)
    - Year 1: -0.311  (-0.97)
    - Year 2: 0.103   (0.19)
    - Year 3: 0.705   (1.55)
    - Year 4: 1.233** (2.59)
    - Year 5: 1.468** (2.53)
  - Observations by year: Year 0: 555; Year 1: 555; Year 2: 555; Year 3: 526; Year 4: 497; Year 5: 468
  - Number of countries: Year 0–4: 29; Year 5: 24
  - Product market reforms (coefficients; t-statistics in parentheses):
    - Year 0: 0.134   (1.37)
    - Year 1: 0.261*  (1.70)
    - Year 2: 0.444*  (1.93)
    - Year 3: 0.645***(2.75)
    - Year 4: 0.781***(3.14)
    - Year 5: 0.964***(3.60)
  - Observations by year: Year 0: 709; Year 1: 709; Year 2: 709; Year 3: 683; Year 4: 657; Year 5: 631
  - Number of countries: 26 (all years)
- Notes and controls:
  - t-statistics from Driscoll-Kraay standard errors in parentheses.
  - Additional controls: Lagged annual change in employment rates (3 lags), lagged output gap, output loss during crisis (Laeven and Valencia, 2013).
  - Significance legend: *** p<0.01, ** p<0.05, * p<0.1.
- Interpretation:
  - Near-term impacts are generally not significant; positive impacts emerge in the medium term (years 3–5).

### State-dependent effects: interaction with business cycle (LP with interaction)
- Approach:
  - Interaction of reform variable with a cyclical dummy I that equals 1 when the output gap < –1 percent ("bad times"), and 0 otherwise ("good times").
  - Controls include fiscal and monetary stance dummies (P) to avoid confounding short-term demand-supporting policy responses.
- Table 2: Effect of Reform on the Employment Rate, Accounting for the Economic Cycle — Dependent variable: Deviation in employment rate relative to pre-reform year
  - Labor market reform, no slack:
    - Year 0: -0.0373  (-0.133)
    - Year 1: 0.529**  (2.194)
    - Year 2: 1.069**  (2.148)
    - Year 3: 1.243**  (2.276)
    - Year 4: 1.344**  (2.095)
    - Year 5: 1.886*** (3.287)
    - Observations: Year 0: 442; Year 1: 442; Year 2: 442; Year 3: 421; Year 4: 400; Year 5: 379
    - Number of countries: Year 0–4: 21; Year 5: 20
  - Labor market reform, slack:
    - Year 0: -0.436**  (-2.23)
    - Year 1: -1.563*** (-3.92)
    - Year 2: -1.853*** (-3.19)
    - Year 3: -1.027**  (-2.33)
    - Year 4: -0.024    (-0.03)
    - Year 5: 0.193     (0.24)
  - Product market reform, no slack:
    - Year 0: 0.180*   (1.780)
    - Year 1: 0.245*   (1.826)
    - Year 2: 0.365**  (2.166)
    - Year 3: 0.696**  (2.744)
    - Year 4: 0.755**  (2.526)
    - Year 5: 0.953**  (2.702)
    - Observations: Year 0: 430; Year 1: 430; Year 2: 404; Year 3: 379; Year 4: 354; Year 5: 329
    - Number of countries: Year 0–5: 26 (except Year 2–5: 25 for some years)
  - Product market reform, slack:
    - Year 0: 0.143   (0.80)
    - Year 1: 0.051   (0.15)
    - Year 2: -0.143  (-0.27)
    - Year 3: -0.145  (-0.23)
    - Year 4: -0.034  (-0.07)
    - Year 5: -0.032  (-0.12)
- Interpretation:
  - Labor market reforms implemented during bad times (slack) produce immediate and statistically significant negative effects on employment (years 0–3).
  - Product market reforms implemented during bad times do not show negative and significant employment losses; product market reforms in no-slack periods show positive medium-term effects.
  - Suggests differing mechanisms by reform type and possible role for reform intensity.

### Reform intensity and timing
- Empirical observation:
  - The average decline in the OECD indicator (capturing reform size) is larger when reforms are initiated during good times for both labor market reforms and product market reforms.
  - Suggests reforms implemented during non-slack periods tend to be more ambitious/serious.
- Implication:
  - Timing (good vs. bad times) may affect both the size of reforms and their employment outcomes.

### Robust approach: AIPW treatment-effect estimates (addressing endogeneity)
- Methodology:
  - Use Augmented Inverse Propensity-Score Weighting (AIPW) to obtain doubly robust matching estimates.
  - First-stage probit to estimate propensity to reform (predictors include lagged GDP growth, legislative election dummy, forward EU accession dummy, age dependency ratio, political leader’s education background).
  - Second-stage uses LP model with inverse-propensity weights to estimate average treatment effects (ATEs).
- Table 3: Effect of Reform on the Employment Rate (AIPW) — Dependent variable: Deviation in employment rate relative to pre-reform year
  - Labor market reform (coefficients; t-statistics in parentheses):
    - Year 0: 0.014    (0.09)
    - Year 1: -0.162  (-0.77)
    - Year 2: -0.149  (-0.36)
    - Year 3: 0.296   (0.53)
    - Year 4: 1.123** (2.20)
    - Year 5: 1.276** (2.15)
    - Observations: Year 0: 555; Year 1: 555; Year 2: 555; Year 3: 526; Year 4: 497; Year 5: 468
  - Product market reform:
    - Year 0: -0.019  (-0.20)
    - Year 1: 0.115   (0.66)
    - Year 2: 0.457*  (1.85)
    - Year 3: 0.750** (2.42)
    - Year 4: 0.866*** (2.66)
    - Year 5: 1.188*** (3.27)
    - Observations: Year 0: 709; Year 1: 709; Year 2: 709; Year 3: 683; Year 4: 657; Year 5: 631
- AIPW summary interpretation:
  - Labor and product market reforms increase the employment rate by 1.3 and 1.2 percentage points, respectively, by the fifth year (AIPW estimates).
  - Confirms lagged but positive impact and suggests larger PMR effect under AIPW than baseline LP.

### State-dependent AIPW results (business cycle)
- Table 4: Effect of Reforms on the Employment Rate, Accounting for the Business Cycle (AIPW)
  - Labor market reform, slack:
    - Year 0: -0.831    (-0.640)
    - Year 1: -2.196    (-0.968)
    - Year 2: -4.674*   (-1.838)
    - Year 3: -4.951    (-1.406)
    - Year 4: -7.957*   (-1.939)
    - Year 5: -11.94*   (-1.907)
    - Observations by year: Year 0: 557; Year 1: 557; Year 2: 557; Year 3: 528; Year 4: 499; Year 5: 470
  - Labor market reform, no slack:
    - Year 0: -0.00324  (-0.0224)
    - Year 1: 0.00188   (0.00635)
    - Year 2: 0.584     (0.848)
    - Year 3: 1.637*    (1.686)
    - Year 4: 2.391**   (2.112)
    - Year 5: 2.778**   (1.961)
    - Observations: same counts as above
  - Product market reform, slack:
    - Year 0: -0.448    (-1.467)
    - Year 1: -0.374    (-0.742)
    - Year 2: 0.208     (0.312)
    - Year 3: 0.594     (0.736)
    - Year 4: 0.730     (0.876)
    - Year 5: 1.330     (1.535)
    - Observations by year: Year 0: 709; Year 1: 709; Year 2: 709; Year 3: 683; Year 4: 657; Year 5: 631
  - Product market reform, no slack:
    - Year 0: 0.105     (0.794)
    - Year 1: 0.306     (1.096)
    - Year 2: 0.590     (1.426)
    - Year 3: 0.826*    (1.678)
    - Year 4: 0.758*    (1.660)
    - Year 5: 1.020**   (2.014)
- Caveats:
  - Number of reform-slack episodes is limited for LMR (12 labor market reforms-slack vs 16 LMR-non-slack) which may affect first-stage probit stability; product market reform counts: maximum of 34 PMR-slack vs 63 PMR-non-slack occurrences.
- Interpretation:
  - AIPW results corroborate that structural reforms (LMR and PMR) are most effective when implemented in periods of limited slack (no slack).
  - Labor market reforms during slack show large negative coefficients in AIPW (caution due to limited sample).

### Role of macroeconomic policies (fiscal and monetary) — AIPW interactions
- General finding:
  - Structural reforms produce stronger positive employment effects when implemented with non-restrictive (supportive) fiscal or monetary policy; effects weaken or reverse under restrictive stances.
- Table 5: Product Market Reform (PMR) × Fiscal Policy Stance (AIPW)
  - PMR, non-restrictive fiscal policy:
    - Year 0: -0.0962  (-0.800)
    - Year 1: -0.108   (-0.483)
    - Year 2: 0.258    (0.919)
    - Year 3: 0.633**  (1.981)
    - Year 4: 0.668**  (1.995)
    - Year 5: 1.113*** (2.764)
    - Observations by year: Year 0: 429; Year 1: 429; Year 2: 429; Year 3: 404; Year 4: 379; Year 5: 354
  - PMR, restrictive fiscal policy:
    - Year 0: 0.325   (0.791)
    - Year 1: -0.104  (-0.206)
    - Year 2: -0.524  (-0.761)
    - Year 3: -1.097  (-1.300)
    - Year 4: -1.667**(-1.970)
    - Year 5: -2.271***(-2.818)
    - Observations: same counts as above
  - Interpretation: PMR raises employment in medium term under non-restrictive fiscal policy; under restrictive fiscal policy PMR yields negative effects 4–5 years after reform.
- Table 6: Labor Market Reform (LMR) × Monetary Policy Stance (AIPW)
  - LMR, non-restrictive monetary policy:
    - Year 0: -0.471** (-2.214)
    - Year 1: -0.723***(-3.735)
    - Year 2: -0.294  (-1.233)
    - Year 3: 0.513*  (1.770)
    - Year 4: 1.540*** (3.682)
    - Year 5: 2.897*** (6.026)
    - Observations by year: Year 0: 551; Year 1: 551; Year 2: 551; Year 3: 522; Year 4: 493; Year 5: 464
  - LMR, restrictive monetary policy:
    - Year 0: 0.163   (0.672)
    - Year 1: 0.204   (0.475)
    - Year 2: -0.0234 (-0.0299)
    - Year 3: -0.887**(-2.013)
    - Year 4: -0.543  (-1.089)
    - Year 5: -0.666  (-0.977)
    - Observations: same counts as above
  - Interpretation: Under non-restrictive monetary policy, LMR yields strong positive medium-term employment gains; under restrictive monetary policy, these gains disappear and can be negative.
- Table 7: Product Market Reform (PMR) × Monetary Policy Stance (AIPW)
  - PMR, non-restrictive monetary policy:
    - Year 0: -0.017  (-0.16)
    - Year 1: 0.199   (1.05)
    - Year 2: 0.563** (2.25)
    - Year 3: 0.956*** (2.96)
    - Year 4: 0.999*** (2.88)
    - Year 5: 1.191*** (2.81)
    - Observations by year: Year 0: 689; Year 1: 689; Year 2: 689; Year 3: 663; Year 4: 637; Year 5: 611
  - PMR, restrictive monetary policy:
    - Year 0: -0.084  (-0.59)
    - Year 1: -0.049  (-0.26)
    - Year 2: -0.006  (-0.03)
    - Year 3: -0.132  (-0.53)
    - Year 4: 0.012   (0.04)
    - Year 5: 0.390   (1.06)
    - Observations: same counts as above
  - Interpretation: PMR yields positive medium-term employment effects when monetary policy is non-restrictive; the positive effect disappears under restrictive monetary policy.
- Macro-policy interpretation:
  - Results align with theoretical/simulation work: restrictive monetary policy (e.g., binding ZLB absent non-conventional measures) attenuates or reverses positive output/employment effects of reforms; non-restrictive policy (including active asset purchase programs at ZLB) enables positive effects.

### Conclusion — synthesized implications
- Robust empirical evidence (LP with AIPW) indicates:
  - Structural reforms (labor market reforms and product market reforms) tend to increase employment rates by about one percentage point over a five-year horizon in baseline LP estimates, and by about 1.3 (LMR) and 1.2 (PMR) percentage points by year 5 in AIPW estimates.
  - Positive effects are lagged, materializing mainly in the medium term (years 2–5).
- State dependence:
  - Reforms are more effective when implemented in periods of limited slack (no slack) and when supported by non-restrictive fiscal and monetary policies.
  - Labor market reforms implemented during slack can produce immediate and sizable negative employment effects.
- Policy implication:
  - Structural reforms are best initiated in conjunction with supportive fiscal or monetary policy if policy space is available.
- Limitations and open questions noted by the authors:
  - Data limitations constrain full assessment of interactions and sequencing of reforms; need richer, sectoral-level reform data.
  - The paper does not address which specific reforms to implement in what order, sectoral impacts, or interactions among different structural reforms.

*Source: _wp1662 - 5. By year 5, the reform increases the employment rate by 1.5 percentage points. This is the (IMF working paper content provided).*

### REFERENCES

### _wp1662 - REFERENCES

### Key bibliographic scope
- The REFERENCES section lists academic and policy works on structural reforms, labor and product market regulations, estimation methods for treatment effects, and databases on systemic banking crises. (Full citations appear in the source document.)

### Appendix I — Country sample (data and sample restrictions)
- Sample coverage: OECD countries with population of more than 5 million, with large episodes of labor market reforms for regular workers (EPLR) and product market reforms (PMR).
- Additional restrictions: non-missing output gap and employment rate variables with at least 3 lags; availability of political variables required for the first stage AIPW regression.
- Baseline EPLR regressions (reflected in Tables 1 and 3): 555 observations from 29 countries.
- Baseline PMR regression (reflected in Tables 1 and 3): 709 observations from 26 countries.
- Regressions in Tables 2, 4-6 (that include policy variables) further restrict the sample.
- Country list included in the sample:
  - Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Finland, France, Germany, Hungary, Indonesia*, Israel, Italy, Japan, Korea, Mexico, Netherlands, Poland, Portugal, Russia*, Slovak Republic, South Africa, Spain, Sweden, Switzerland, Turkey, United Kingdom, United States*
  - Note: * Dropped in the PMR regression sample owing to the absence of any large reform episode.

### Appendix II — Determinants of labor and product market reforms (probit analysis)
- First-stage design: propensity scores estimated via a pooled panel or random effect probit model for the probability of implementing a labor or product market reform.
- Regressors used:
  - Lagged GDP growth (to capture cyclical conditions; included because countries in recession may be more likely to implement reforms).
  - Legislative election dummy (1 when there is a legislative election in that year; drawn from the database on political institutions (DPI); captures timing within the political cycle).
  - EU accession dummy (1 if the country is in the EU; captures external pressures from the EU).
  - Age dependency ratio (share aged 65 and older; captures demographic pressure on pensions and social security).
  - Political leader’s educational background dummy (1 if Prime Minister or Minister of Finance has a degree in economics; drawn from Hallerberg and Wehner (2014)).
- Summary interpretation provided in the source:
  - Labor market reforms are more likely to be implemented following periods of lower growth, off election cycles, and in countries with EU accession commitments.
  - Product market reforms tend to occur in good times, in countries with a high dependency ratio, and where key political leaders have an economics background.

- Table A.1. Determinants of Structural Reforms — Pooled Probit Regression (Dependent variable: Probability of adopting a reform)
  - Labor market reform / Product market reform: coefficient estimates (with reported t-statistics in parentheses)
    - Lagged GDP growth: -0.0493*  ( -1.660 ) ; 0.0653***  ( 4.053 )
    - Legislative election dummy: -0.450*  ( -1.738 ) ; -0.0226  ( -0.189 )
    - EU accession dummy, 1 year ahead: 0.644**  ( 2.155 ) ; 0.00966  ( 0.0780 )
    - Aged 65 up: -0.0382  ( -0.947 ) ; 0.0396**  ( 2.539 )
    - Technocratic leader dummy: -0.0309  ( -0.168 ) ; 0.254**  ( 2.366 )
  - Observations: 688 (labor market reform) ; 1105 (product market reform)

### Empirical implications and methodological notes (as stated in source)
- The probit results support the inclusion of cyclical, political, integration, demographic, and leader-education covariates in the first-stage propensity score estimation for reforms.
- The source notes related work: Duval and others (forthcoming) examine drivers of structural reforms using similar variables and alternate definitions of large changes in OECD indicators.

*Source: _wp1662 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1662.pdf_