Heterogeneous Effects of the German Minimum Wage on Working Hours and Minijobs

In 2015, Germany introduced a national minimum wage. While the literature agrees on at most limited negative effects on the overall employment level, we go into detail and analyze the impact on the working hours dimension and on the subset of minijobs. Using data from the German Structure of Earnings Survey in 2010, 2014, and 2018, we find empirical evidence that the minimum wage significantly reduces inequality in hourly and monthly wages. While various theoretical mechanisms suggest a reduction in working hours, these remain unchanged on average. However, minijobbers experience a notable reduction in working hours which can be linked to the specific institutional framework. Regarding employment, the results show no effects for regular jobs, but there is a noteworthy decline in minijobs, driven by transitions to regular employment and non-employment. The transitions in non-employment imply a wage elasticity of employment of −0.1 for minijobs.

The introduction of a national minimum wage in Germany on 1 January 2015 — set initially at €8.50 per hour — is one of the largest labour-market interventions in the post-reunification era. Existing studies have found at most modest aggregate employment effects, but the policy introduced an institutional setting that creates heterogeneous incentives across job types: in particular, the German minijob category (jobs paying up to €450/month, exempt from social security contributions on the worker side) creates a kink in the effective wage that interacts with a binding hourly minimum.

Workers below the threshold prior to the reform were paid an average of 26 percent below the minimum wage threshold in 2014. Minijobs made up as much as 21 percent of all jobs in 2014. In absolute terms, 104,000 minijobs vanished due to the minimum wage; we find that 54,000 minijobbers entered non-employment, while the other half was upgraded to regular employment.

We use three waves of the German Structure of Earnings Survey (SES) — 2010, 2014, and 2018 — supplemented with the Establishment History Panel (BHP) and Integrated Employment Biographies (IEB). The SES covers establishments with at least 10 employees subject to social security (regular employees) and excludes agriculture, the public sector, apprentices, and a small set of further restricted categories. Sample sizes are large (typical analytic N in the millions of worker-year observations); the sample size of the SES varies materially between 2010 and 2014/2018 due to wave-to-wave panel-design changes.

Table 1 reports descriptive statistics. In our main SES sample of establishments with at least 10 regular employees, the minijob share is 16.4% — distinct from the 21% headline figure in the introduction, which is taken over the full German employment population. Mean nominal hourly wages are reported in current euros; for the cross-wave wage-distribution comparisons in Section 4 we deflate to 2014 €.

We exploit county-level variation in the bite of the minimum wage. The county bite ${\text{Bite}}_r$ is the share of workers in county $r$ earning below €8.50 in 2014. The reduced-form specification regresses an outcome $Y_{irt}$ on a treatment-period interaction:

$$Y_{irt}=\alpha +\delta {\text{Bite}}_r\cdot 1[t=2018]+\beta X_{irt}+{\gamma }_r+{\gamma }_t+{\epsilon }_{irt}.$$

We use one pre-period (2010 → 2014) for placebo testing and inspect the implied parallel-trends pattern visually. Distributional effects on the wage are estimated via Recentered Influence Function (RIF) regressions following Firpo–Fortin–Lemieux. For the small-establishment subsample (Sec. 5.5) we use the IEB-augmented sample and identical bite-interaction specification.

Table 2 reports a county-bite coefficient of 0.456 on the log hourly wage for the affected segment. If we extrapolate and compare the average hourly wage effect of 45.6 percent (for affected workers) with the average initial wage gap of affected workers of 26 percent, the implied increase exceeds the gap — consistent with spillovers above the minimum threshold.

The RIF estimates show that the lower tail of the hourly wage distribution compresses sharply between 2014 and 2018; the upper tail moves little. Monthly wage inequality (measured by the Gini and the 50/10 ratio) falls correspondingly. Figure 1 shows the real hourly wage distributions for 2014 and 2018, deflated to 2014 €; the distributional shift is visible at the new statutory minimum.

Average working hours show no statistically significant response. Conditional on the minijob status, however, hours decline materially: minijobbers above the implicit hourly cap (working more than 50 hours per month at the new floor would breach the €450 monthly threshold) bunch downward.

Section 5.3 reports the institutional adjustment. About 600,000 minijobbers with hours beyond 50 per month are no longer observed in 2018. Of these, 200,000 jobs may have been downgraded to 45–49 hours. Consequently, the other 400,000 jobs must have been either terminated or upgraded to regular jobs. The implied causal minijob loss is on the order of 104,000 net positions, with 54,000 transitions into non-employment (implying a wage elasticity of employment of −0.1 for minijobs).

For regular jobs, the bite-interaction coefficient is statistically indistinguishable from zero across both panels of Table 5 (large vs small plants). Section 5.4 presents the small-plant analysis using the IEB sample and a bite-specific trend ${\text{Bite}}_r\cdot {\text{Year}}_t$ to control for differential pre-2015 employment trends, following Ahlfeldt, Roth, and Seidel (2018) and Bossler and