Effects of the German Minimum Wage on Wages and Household Income

1 Introduction

In recent years, there has been increasing interest in the distributional effects of minimum wages. Mostly positive effects of minimum wages on the incomes of dependent employees in a number of countries have been shown in the international literature (to name just a few: Aeberhardt, Givord, and Marbot 2012 for France; Dube 2019 for the USA; Vandekerckhove, van Gyes, and Goos 2018 for Belgium). Studies on the distributional effects of the statutory minimum wage in Germany carried out to date have mainly focused on the effects on individual income from wages of dependent employees. Recent studies in this line of research were written by Bossler and Schank (2023) who identified considerable effects in the lower half of the wage income distribution (on a wage per day basis) and Dustmann et al. (2022) who found both statistically and economically significant and positive minimum wage effects on daily wages of workers. Research into the effects of minimum wages on household incomes in Germany has been more limited. There are some studies that have used survey data on net household income or receipt of social transfers. Even though those surveys are far smaller than the administrative datasets used in the studies mentioned above, significant effects on net household income could be shown (e.g. Pusch et al. 2021; Schröder et al. 2020).

In this study we use Unconditional Quantile Regressions (UQR), which have previously been applied to study the distributional impact of the minimum wages in a number of countries (France: Aeberhardt, Givord, and Marbot 2012; US: Dube 2019; Germany: Bossler and Schank 2023). The minimum wage effect is measured by elasticities of the considered income measures to the regional minimum wage bite in difference-in-difference terms embedded in the UQR estimations. The data stem from a large household survey in Germany, the Survey of Income and Consumption (German: Einkommens- und Verbrauchsstichprobe, EVS). The EVS comprises comprehensive household income information, it allows us to consider minimum wage effects on a number of income measures. As the EVS is conducted only every five years, we consider the year 2018 for measuring the combined effect of the minimum wage introduction and its first uprating (in 2017) on individual wage income, needs-weighted gross and net household income.

So far, none of the above-mentioned studies have estimated the minimum wage effects for individual wage income, needs-weighted gross and net household income. This is the major novum of this study as it allows a comparison of the relative size of income effects at the individual level considering only individual wage income and after consideration of other income sources at the household level (not all minimum wage earners live in poor households). First, using individual information, we combine all income from wages at the individual level, also including secondary jobs. There might be deviations from other studies considering time spent in jobs (which can be also secondary jobs), as some of the low paying secondary mini-jobs^[1] are combined with wage incomes of holders higher up in the income scale, in our study. However, the effect of this combination of different wage incomes does not seem to be very large. Our measured minimum wage effects on individual income from wages are broadly in line with those of Bossler and Schank (2023). We can confirm their results with a different data set.

Second, as we are working with a household survey, we can also do the same analysis as with wage income for the gross and net household income of workers. The largest effects are to be expected for individual incomes from dependent main employment, since the minimum wage has a direct effect on individual employees. If gross household incomes from wages are considered, effects are expected to be more stretched out along the line of distribution, since minimum wage recipients can also live with better-off partners and minimum wage effects thus spread more widely in the income distribution. As regards gross household income, since these also contain other income components, such as income from public transfers, which partly decline with increases in earnings from wages, weaker effects are to be expected especially at the lower end of the line of distribution. In addition, capital income and income from self-employment may decline in some households due to the minimum wage. The last estimations were carried out for net disposable household income. We expect minimum wage effects on net household income to be lower in the middle of the income distribution as compared to effects on gross household income due to the functioning of the tax progression.

Third, we also account for possible endogenous effects of the minimum wage, such as employment flows of people of working age (not just job losses, also possible inflows into the labour market, as have been argued by Krebs and Drechsel-Grau 2021 or changes from dependent work into self-employment or vice versa). We expect the minimum wage effects to be less pronounced here, mostly due to the fact that only about 68 % of the German population of working age (19–65 years^[2]) are workers, another 6 % are self-employed, about 4 % are apprentices. According to the EVS, about 22 % of the working age population are not involved in paid work. This can be for a number of reasons: participation in higher education, disability, reduction in earning capacity pensions, care responsibilities, early pensions etc. In many of these cases it can be assumed that the wage level is not the reason for being out of work.

The article is structured as follows. In Section 2 we give an overview of the background to the introduction of the statutory minimum wage in Germany. Section 3 presents the state of research on the effects of the statutory minimum wage on incomes in Germany. Section 4 describes the data used. Section 5 provides an overview of the development of individual income from wages, hourly wages and household income (gross and net) during the period under study. Subsequently, in Section 6 Unconditional Quantile Regressions are applied for the minimum wage effects on the distribution of individual as well as household income from wages, gross household income and net household income of employees and all persons of working age. Section 7 summarises the results.

2 The Statutory Minimum Wage in Germany: Background and Institutional Framework

The autonomy of collective bargaining enshrined in the German constitution (Grundgesetz) ensures that the parties to collective agreements in Germany have the right to shape wage determination (with exceptions, such as the churches and charitable institutions). In principle, this policy has been applied broadly for the vast majority of employees for decades, as shown by the very high collective bargaining coverage well into the 1980s (Bosch, Schulten, and Weinkopf 2021). However, in the 1990s, after reunification, this system began to show cracks. In east Germany, which did not have an established structure of collective bargaining parties, trade union membership and collective bargaining coverage fell sharply shortly after reunification. In addition, there was a deep crisis in the construction sector after the end of the reunification boom. At the same time, there was significant competitive pressure from companies and posted workers from other EU countries (Apel et al. 2012). This formed the background for the introduction of the first sectoral minimum wages declared by law to be generally binding, which from 1997 also extended to companies not bound by collective agreements and their employees.

However, the actual background for the introduction of the statutory minimum wage was the expansion of the low-wage sector (low wages are defined as wages below 2/3 of the median hourly wage), which since the end of the 1990s had also increasingly affected the west German labour market. While the low-wage share was still around 12 % of the workforce in 1998, it rose to 18 % in only 5 years (Kalina and Weinkopf 2017: 3). From 2001 onwards, Germany found itself in an economic crisis that lasted several years, and from 2004 onwards a series of labour market reforms (Hartz reforms) were passed in this context, in which working conditions were deregulated. Among other things, there was an expansion of marginal employment in the form of mini-jobs, which were exempt from income tax and for which the maximum number of hours was abolished. The regulations for the use of temporary work were also deregulated. For the unemployed, the reasonableness conditions for accepting a job offer were tightened, which subsequently also put pressure on wages (Deutscher Bundestag 2006). The binding force of collective agreements continued to decline during this period.

Although employment began to grow again from around 2006, the share of low wages continued to stagnate at a high level of around a quarter of the workforce in subsequent years (Kalina and Weinkopf 2017: 3). Trade unions and employers were apparently increasingly less able to shape the low-wage sector by means of collective agreements (Dütsch et al. 2023; Schulten and Pusch 2019). Generally binding sectoral minimum wages (as a possible alternative) did come about in some sectors (e.g. temporary work – for a current overview see Bispinck 2023), but their wider implementation failed. The debate on the introduction of a statutory minimum wage as a wage anchor to prevent a further reduction of the lowest wages, therefore gained increasing weight – even among trade unions, some of which (especially the industrial trade unions) had previously viewed the minimum wage with scepticism. In 2006, there was a majority decision at a congress of the DGB (German Federation of Trade Unions) to support the introduction of a statutory minimum wage.

The statutory minimum wage was introduced on 1 January 2015 by the then Grand Coalition at an initial level of 8.50 euros per hour. Exceptions to the statutory minimum wage exist for workers under 18 years of age who have not completed vocational training, apprentices and the long-term unemployed. The Minimum Wage Commission, composed equally of representatives of employers and trade unions and academics (without voting rights), draws up a proposal for the adjustment of the minimum wage every 2 years, which is then implemented by the legislator. In the first years, the Minimum Wage Commission followed the collectively agreed wage index with its proposals. Nevertheless, the minimum wage was still increased at a slower pace and over time the Kaitz index fell from 48.6 % (2015) to 46.7 % in 2021 (Herzog-Stein et al. 2023: 18).

Against the backdrop of a renewed public debate on the appropriate level of the minimum wage, the minimum wage was increased to 12 euros on 1 October 2022, deviating from the usual adjustment path, which at that point in time corresponded to approximately 63 % of the median wage (Mindestlohnkommission 2023: 31). As a result of this large increase, the statutory minimum wage in Germany is currently close to the target values of 60 % of the gross median wage or 50 % of the gross average wage, which are anchored in the recently adopted EU Minimum Wage Directive (EU 2022, Article 5 (4)).

3 Distribution Effects of the Minimum Wage in Germany

Research on the relationship between the minimum wage and various income measures at the household and personal level has produced a large number of empirical contributions in recent years. An overview of international studies would go too far for this article, the focus of which is the statutory minimum wage in Germany, for which a number of studies on income effects have already been published. An overview of these studies is given in Table 1. The overview shows that the majority of studies so far have been conducted on employee’s income from wages. We will discuss some of these studies (the most pertinent) in the following.

Bossler and Schank (2023, Table 2) identified significant minimum wage effects in the year 2017 which reach up to the 5th decile of the wage income distribution. For example, for the 2nd decile (where the relative effect is highest) a minimum wage effect of approx. 24 percent in employee’s wage income can be calculated in regions with an average minimum wage bite (own calculation based on the point estimate of the elasticity). The absolute employee wages in this area of the distribution are low at about 650 euros. However, the increases are comparable in magnitude to the results established by Himmelreicher (2020), who documented a similar rate of increase (27 percent) in wage income for jobs at the 20th percentile of the hourly wage distribution in the period from 2014 to 2018. Somewhat lower results were estimated by Dustmann et al. (2022) who found a minimum wage-related increase in daily wages of about 10.7 per cent for the lowest wage group. In contrast, Caliendo et al. (2018) did not find any increases in employee’s wage income as a result of the minimum wage. They attribute this result to the reductions in working hours. However, the effects documented by the authors only include the year 2015. In the study by Bossler and Schank (2023), the effects become larger over time.

According to a study by Schröder et al. (2020), the effects of the minimum wage on net household incomes in Germany are significantly smaller than in some of the studies mentioned above for employee’s wage income. In households with at least one minimum-wage employee (in 2014), disposable income increased by 4.3 percent between 2014 and 2016 compared to the control group. This finding is probably mainly due to the fact that minimum wage recipients are also to be found partly in higher areas of the household income distribution and income increases triggered by the minimum wage are therefore not only concentrated at the lower end of the distribution.

Another study on the income effects of the minimum wage was presented by Pusch et al. (2021), who were able to show effects up to the 30th percentile of the distribution of the means-weighted household net incomes of employees. They also find a decrease in the probability of receiving means-tested social benefits after the introduction of the statutory minimum wage. However, this cannot be attributed to the regional depth of intervention of the minimum wage. With a different specification of treatment and control group, Bruckmeier and Schwarz (2022), on the other hand, were able to show effects of the minimum wage on social benefit receipt, which decreased for recipients of unemployment benefit II when income from wages increased due to the minimum wage.

Last but not least, job losses caused by the minimum wage can also influence the distribution of income. Neumark, Schweitzer, and Wascher (2005), for example, come to the conclusion that an increase in the risk of poverty in the USA is due to the minimum wage and attribute this to job losses. If job losses were a dominant effect of the minimum wage, an increased receipt of social benefits and thus, as a rule, lower incomes could be expected. However, the literature on the employment effects of the statutory minimum wage in Germany has so far documented only minor effects (neutral: Garloff 2016; Herr et al. 2018, negative: Bonin et al. 2018; Bossler and Gerner 2020; Caliendo et al. 2017a).

4 Data

The Survey of Income and Consumption (EVS, Einkommens- und Verbrauchsstichprobe) is a household survey conducted every five years by the German Federal Statistical Office, the aim of which is a detailed recording of income and consumption expenditure in households. The survey is organised as a cross-sectional survey. Individual households cannot be observed over time. In addition to income and consumption data, the survey also collects some socio-demographic information. The EVS survey is implemented as a deliberate (non-probabilistic) selection of survey units (households) within the framework of a quota procedure (Statistisches Bundesamt 2017: 18). The reference for quota and extrapolation is the statistics of the micro census (MZ) which also serves as a reference for a number of other surveys in Germany. The net household income itself is one of the quota characteristics and is intended to ensure a representative income stratification. However, households with incomes above 18,000 euros are not surveyed in the EVS. The EVS therefore does not cover the upper end of the income distribution (Statistisches Bundesamt 2017: 19).

The Scientific Use Files of the EVS used here (Grundfile 3) comprise about 42.000 observations in each wave and represent an 80 % sample of the EVS (FDZ Bund Länder 2020). For the households surveyed, there is information for up to six employees, including wage components, employment status and working hours. Much of this information is only available at the quarterly level, so there are some uncertainties to be considered in the following analyses. Income components from employment and working hours are explicitly asked for in reference to the main job.7F^[3] In addition, the income from a dependent secondary gainful activity is also queried.8F.^[4]

One advantage of the EVS over the SOEP dataset, which is otherwise frequently used in minimum wage research for Germany, is the relatively high number of observations (see above) and the detailed recording of income and wage components. This leads, on average, to higher recorded wages and incomes (see Appendix A1 for a comparison with SOEP). It can therefore be assumed that incomes are measured more comprehensively in the EVS. Disadvantages are that the survey is only conducted every 5 years, there is a smaller scope of socio-demographic information and the nature of the survey is cross sectional.

For the following evaluations, only dependent employees (excluding apprentices) between 19 and 65 years of age or their households were taken into account. In the case of marginal employment (often called mini-jobs), there was a change in the questionnaire during the transition from EVS 2008 to EVS 2013, which is why the weighting factors for 2008 were adjusted with the help of data from the administrative statistics (BA 2023).^[5] The adjustment was made in such a way that essential characteristics of the distribution of marginally employed persons after reweighting corresponded to those of the administrative statistics (distribution by gender, age, west and east Germany) and the relative coverage corresponded to that of the subsequent 2013 and 2018 waves (about 2/3 of the number of observations in the administrative statistics, see Annex A1).

In 2008, the EVS survey was, in some respects, carried out differently. In particular this is the case with the status of apprentices, which was not queried – in contrast to the subsequent waves. For this reason, the apprentices were approximated in a comparable way for the sample delimitation in all three waves via age (up to 22 years) and the absence of a vocational training diploma.16F^[6] The apprentices thus delimited were not included in the employee sample for the descriptive overview, as apprentices are exempt from the statutory minimum wage. However, it has also been argued in the literature that there is a correlation between minimum wage and training investment (Acemoglu and Pischke 2003). For this reason, the regressions in Section 6 are also conducted with apprentices (and all other persons of working age, to allow for endogenous movements).

For the regression analyses in Section 6, a regional bite of the minimum wage, measured as the regional share of jobs below the minimum wage before its introduction, based on the Structure of Earnings Survey (VSE, Verdienststrukturerhebung) is used (Figure 1). This minimum wage bite measure was provided by the Federal Statistical Office (in Annex A2 we include also estimation results based on the EVS-based measure of the regional minimum wage bite). While the EVS-based measure of the regional minimum wage bite can only be calculated for main jobs due to the availability of working time information, the VSE figure also takes into account wages from secondary jobs. Both data sources nonetheless show a high correlation; the minimum wage thus had a greater prevalence in the east German states in particular (to be found in the top of Figure 1: Brandenburg, Mecklenburg-Vorpommern, Sachsen, Sachsen-Anhalt, Thüringen), a finding also known from previous research (Mindestlohnkommission 2016: 39). For this reason, the greatest effects of the statutory minimum wage are also to be expected for east Germany and were already documented early on in research (Amlinger et al. 2016). In the following presentation of the development of incomes, special attention is therefore also paid to the development in east Germany.

Figure 1:

EVS- and VSE-based measures, showing the depth of intervention of the minimum wage in Germany as a whole and by federal state (in %). Source: EVS 2013 (Grundfile 3), own calculations, VSE 2014: Statistisches Bundesamt (2016).

5 Descriptive Overview of Individual and Household Income Developments

6 Income Effects of the Statutory Minimum Wage: Results of Unconditional Quantile Regressions

7 Summary

In this study, data from the Survey of Income and Consumption (EVS) was used for the first time to examine the effects of the introduction of the statutory minimum wage in Germany. The EVS is a very detailed and, compared to other surveys, comprehensive survey of the income situation of households in Germany, with data from around 42,000 households in each of the waves considered: 2008, 2013 and 2018. Descriptive evaluations show substantial increases in wages in the lower and middle segments of the gross income distribution in east Germany, especially for the period after the introduction of the minimum wage. Since the minimum wage has a high prevalence in east Germany, this finding suggests a positive income effect of the statutory minimum wage. The development of gross and net household incomes overall was also very dynamic in the lower range of the income distribution after the introduction of the minimum wage.

These results are confirmed by means of Unconditional Quantile Regressions for employees’ incomes. The results of the regressions for individual wages show – similar also in magnitude to the results from the study of Bossler and Schank (2023) – clear effects that range roughly into the middle of the wage income distribution. As we used a measure of regional bite of the minimum wage (share of minimum wage jobs among all jobs), our results indicate that the minimum wage had a strong income-increasing effect especially in east Germany where there is a high prevalence of low wages as compared to west Germany. The estimated effects on total gross and net household incomes of workers are somewhat smaller, which is also not surprising, as the minimum wage addresses workers’ wages directly and not household incomes (which also derive income from other sources). In addition to the wider dispersion of minimum wage recipients across the entire household income distribution, a relative reduction in social benefit recipients at the lower end of the income distribution (as has been shown by Bruckmeier and Schwarz 2022) may have also contributed to this, which is also visible in the descriptive evaluations. Nevertheless, it is still interesting to see that household incomes are still positively affected, which is in line with the U.S. study carried out by Dube (2019).

Estimates for the minimum wage effect on net household income are broadly in line with the study carried out by Pusch et al. (2021) who based their results on the PASS survey, but indicate smaller increases at the 1st and 2nd decile in 2018. With respect to this difference at the lower end of the distribution, a possible limitation for difference-in-difference analysis may be seen in the increase of the marginal employment earnings threshold in 2013 (from 400 to 450 euros) – 2 years prior to the minimum wage introduction. This change is seldom mentioned in the German minimum wage literature, but it may complicate placebo tests or the modelling of trends in difference-in-difference estimations, possibly leading to biased estimates especially at the lower end of the income distribution where mini-jobbers tend to be concentrated (many of them being welfare recipients, students or even pensioners (see Pusch et al. 2021, p. 199).^[14] Since the increase of the minimum wage to 12 euros per hour in October 2022 the earnings threshold for mini-jobs has been linked to the development of the minimum wage, so that this problem for evaluation could be less important in the future.

The measured income effects for employees may be influenced by other endogenous minimum wage effects such as job losses or entries, transitions into/out of education or self-employment. For this reason, additional regressions were carried out for all persons of typical working age (between 19 and 65 years). The measured effects on household income are smaller here and more concentrated in the middle band of the considered income distributions; significant effects are found at the 40th and 50th percentiles of the distribution of gross and net household income. On the other hand, smaller minimum wage effects in the whole working age population are not surprising given the much higher number of people considered (about 50 % more as compared to workers alone, including many inactive), many of whom simply do not intend to work for a number of reasons. Since the minimum wage (by design of the policy) only intends to affect the wages of workers, it is therefore intuitive that the overall population (including unemployed and inactive individuals) experiences less of an income effect. Against this backdrop, it could be still considered an interesting result that even when all persons of working age are considered, the minimum wage has lifted incomes in the middle band of the income distribution in low wage regions.