Tag: work incentives

Discrimination in Work Conditions: The Case of Sexual Harassment

Posted on March 9, 2026 | Policy Brief

Blurred silhouettes of office workers behind glass symbolizing discrimination work conditions in a modern workplace environment.

The #MeToo movement put a spotlight on the severe and highly prevalent workplace problem of sexual harassment. New research argues that economists should treat sexual harassment as gender discrimination in work conditions. Both men and women are subject to this discrimination when their gender is in the minority in the workplace. These patterns reinforce segregation in the labor market and, by extension, economic gender inequality. By reducing the prevalence of sexual harassment, we not only reduce individual suffering but also have positive impacts at a societal level.

Introduction

Throughout the world, the sorting of women into lower-paying occupations and workplaces fundamentally determines economic gender inequality (see Penner et al. 2023 for an overview). The academic discussion about causes of this gender segregation typically centers on gender differences in preferences for work conditions. Women who have more responsibilities for children and the household prefer occupations and workplaces with more flexible schedules, work-from-home opportunities, and shorter commutes. To get these good work conditions (so-called work amenities), they accept jobs in occupations and workplaces with lower wages (Goldin 2014, Wiswall and Zafar 2016, Mas and Pallais 2017, Le Barbachon et al. 2019).

There is mounting evidence that the interpersonal work environment also matters greatly for job choices. Workers seem to put a large negative value on negative interpersonal work conditions such as hostility, bullying, and sexual harassment (see, for example, Folke and Ricke 2022, Collis and Van Effenterrre 2025, and Le Page et al. 2025). Unlike traditional amenities related to aspects such as schedule flexibility, training, or bonuses, the social work environment does not form part of the employment contract and is not under direct control of the employer. This implies that even among the most well-intentioned employers, the social work environment could differ across individuals – and, in particular, between men and women.

Gender differences in exposure to negative social behaviors may meet the standard definition of discrimination in empirical economics research. The mistreatment may imply that women and men with the same qualifications doing the same job receive different pay. Women and men may have the exact same job contracts and receive the exact same compensation on paper, but one gender may be exposed to negative treatment that dramatically reduces their total payoff from the job.

Sexual Harassment and Gender Inequality in the Labor Market

Folke and Rickne (2022) study how sexual harassment by colleagues and managers affects gender segregation across workplaces and, by extension, gender inequality in the labor market. The starting point is a general equilibrium model where the total pay of a job is a function of the wage and the gender-specific sexual harassment risk.

The model shows that sexual harassment leads to larger gender inequality in the labor market under three conditions. Sexual harassment risks need to increase in the share of opposite sex co-workers, wages should increase in the share of men in the workplace, and sexual harassment should affect labor market choices. The model explains that sexual harassment creates gender segregation by operating as a wedge in the payoff from jobs in gender-imbalanced workplaces. All else equal, women get a lower total compensation in male-dominated workplaces, and vice versa for men in female-dominated ones. This will create gender segregation as both women and men have smaller incentives to become a workplace gender minority. It will also create a larger gender wage gap by channeling women into lower-paying workplaces and men toward higher-paying ones.

Harassment Risks and Pay Across Workplaces

To empirically assess how harassment risks vary across workplaces, Folke and Rickne (2022) use survey data on self-reported sexual harassment from the Swedish government’s biannual survey on work conditions (N=40,000). This nationally representative survey contains questions on unwanted sexual advances, sexist hostility, and gender harassment from colleagues or managers in the last 12 months. The survey data can be linked to administrative data on the full Swedish workforce, enabling the computation of the share of men in each survey respondent’s occupation and workplace.

The relationship between self-reported harassment and sex ratios is shown in Figure 1. Clearly, both women and men self-report more harassment when they are the gender-minority in their workplace. The higher self-reported rate of harassment among gender minorities is not caused by systematically different demographic traits. Nor is it caused by gender minorities being more likely to have opposite-sex supervisors, or to themselves hold subordinate or supervisory positions, or by them having opposite-sex managers. Folke and Rickne (2025) show that these patterns also hold at the occupation level.

Figure 1. Sexual Harassment Incidence across Workplace Sex Ratios.

Source: Replication of the left-hand side of Figure II in Folke and Rickne (2022). Note: The figure shows binned averages of a binary variable for self-reports of sexual harassment in the last 12 months from colleagues or managers. Each sub-sample of men and women is split into 100 equally sized bins of the X-variable. N=19,975 for women, and 17,482 for men.

To examine how wages relate to sex composition, Folke and Rickne (2022) rely on the empirical framework developed by Abowd et al. (1999). This framework estimates workplace fixed effects in a wage regression that also includes individual fixed effects and a host of occupational and demographic controls. The workplace fixed effects (i.e., the wage premiums) capture how much a workplace pays in wages compared to other workplaces with the same occupation structure and workers’ socio-economic traits. The analysis shows that a 10-percentage-point larger share of men is, on average, associated with a 1-percentage-point higher wage premium.

To summarize the first set of results, male-dominated workplaces pay more. At the same time, both men and women face a higher risk of sexual harassment when they work in an occupation or workplace with more men. The combination of these results suggests that women have an incentive to work in lower-paying jobs, while men have an additional incentive to work in high-paying jobs.

Sexual Harassment and Job Choice

Sexual harassment can affect job choice in two ways: it can deter an individual from taking a job, or make a person leave a job that they have previously chosen. Folke and Rickne (2022) examine both these channels.

To examine if sexual harassment risks deter individuals from taking a job, Folke and Rickne (2022) use a survey experiment sent to ~4,000 employed Swedish citizens. The survey experiment follows the standard economic approach of exposing respondents to a hypothetical job choice experiment where they choose between fictional job offers with randomized wages and work conditions (for prominent examples of this approach, see, for example, Wiswall and Zafar 2017 and Mas and Pallais 2017).

Sexual harassment was incorporated into the experiment by showing respondents vignettes of sexual harassment incidents that took place in fictional workplaces (as in Hulin, Fitzgerald, and Drasgow 1996). These vignettes mimic the types of anecdotes or rumors that a prospective employee might hear about a potential employer. Importantly, the vignettes make it possible to vary the victim’s gender, which allows comparison of job choices among respondents who are exposed to a harassment victim of their own gender and respondents exposed to a victim of the opposite gender.

The experiment showed a large negative valuation of sexual harassment—the equivalent of a 10% lower wage in the full sample. This large valuation makes sexual harassment a relevant work condition for shaping people’s total remuneration from work and is quantitatively similar to the valuations of time/space flexibility in previous research (Wiswall and Zafar 2017; Mas and Pallais 2017; Maestas et al. 2018). While men and women had similar valuations, there was a substantial difference between those who see a victim of their own sex compared to the opposite sex: the negative valuation is equivalent to a 17% lower wage for same-sex victims but just 6% for opposite-sex ones.

Figure 2. Event Study of Workplace Transitions.

Source: Replication of Figure V in Folke and Rickne (2022). Note: The figure shows estimated differences in the proportion of employer-to-employer transitions out of the workplace in the Work Environment Survey between people who self-report sexual harassment in that survey or not. The X-axis denotes the number of years since the survey. Demographics controls from administrative records are four dummies for marital and parental status, four dummies for age categories, two dummies for having secondary or tertiary education, and two dummies for being born in a different European country or outside Europe.

Folke and Rickne (2022) rely on the work-environment survey matched to the administrative data to show that sexual harassment also affects the probability of leaving a workplace. Conditional on a host of controls, women who report sexual harassment are about 5 percentage points more likely to have left their workplace 3 years after having answered the survey than women who did not report sexual harassment. The equivalent gap for men was about 3 percentage points.

Conclusions

The case study of sexual harassment in Sweden highlights this work condition as an important barrier to gender equality in the labor market. It shows a higher prevalence of sexual harassment for workplace gender minorities and how it imposes costs on these minorities relative to their gender majority colleagues. The disincentive created by sexual harassment to become—and remain—a workplace gender minority reinforces gender segregation across workplaces. The gender wage gap also grows as women prefer not enter male-dominated workplaces with higher pay, or leave these workplaces and head to ones with more women and lower monetary compensation.  These macroeconomic impacts add to the “business case” for governments to prevent sexual harassment.

Sexual harassment is just one of many forms of discrimination in work conditions that could reinforce inequalities in the labor market. If we want to reduce gender inequality, it is clearly not enough to focus on gender differences in preferences for work conditions. We also need to pay attention to factors, such as sexual harassment, that lead to men and women facing different work conditions in the same job. Addressing this form of discrimination could not only yield large payoffs for individual well-being but also reduce inequalities in the labor market.

References

  • Abowd, J.M., Kramarz, F. and Margolis, D.N., 1999. High wage workers and high wage firms. Econometrica, 67(2), pp.251-333.
  • Folke, O. and Rickne, J., 2022. Sexual harassment and gender inequality in the labor market. The Quarterly Journal of Economics, 137(4), pp.2163-2212.
  • Folke, O., & Rickne, J. (2025). Sexual harassment across occupations: new evidence from Swedish Nationally representative data. European Sociological Review, 41(6), 903-918.
  • Collis, M.R. and Van Effenterre, C., 2025. Workplace Hostility. IZA-Institute of Labor Economics.
  • Goldin, C. (2014). A grand gender convergence: Its last chapter. American Economic Review, 104(4): 1091-1119.
  • Hulin, C.L., Fitzgerald, L.F. and Drasgow, F., 1996. Organizational influences on sexual harassment. Sage Publications, Inc.
  • Le Barbanchon, T., Rathelot, R. and Roulet, A., 2021. Gender differences in job search: Trading off commute against wage. The Quarterly Journal of Economics, 136(1), pp.381-426.
  • Lepage, L.P., Li, X. and Zafar, B., 2025. Anticipated discrimination and major choice (No. w33680). National Bureau of Economic Research
  • Maestas, N., Mullen, K.J., Powell, D., Von Wachter, T. and Wenger, J.B., 2023. The value of working conditions in the United States and implications for the structure of wages. American Economic Review, 113(7), pp.2007-2047.
  • Mas, A, and A Pallais (2017), “Valuing Alternative Work Arrangements”, American Economic Review, 107(12): 3722–3759.
  • Penner, A.M., Petersen, T., Hermansen, A.S., Rainey, A., Boza, I., Elvira, M.M., Godechot, O., Hällsten, M., Henriksen, L.F., Hou, F. and Mrčela, A.K., 2023. Within-job gender pay inequality in 15 countries. Nature human behaviour, 7(2), pp.184-189.
  • Wiswall, M. & Zafar, B. (2017). Preference for the workplace, investment in human capital, and gender. The Quarterly Journal of Economics 133(1): 457-507.

Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.

Equity and Efficiency in the Latvian Tax-Benefit System

Posted on November 3, 2014 | Policy Brief

20141103 Equity and Efficiency in the Latvian Tax-Benefit System Image 01

There is a trade-off between two major objectives of a tax-benefit system: equity and efficiency. The tax-benefit systems that redistribute a lot of income tend to generate disincentives to work. The tax-benefit systems that create good incentives to work and earn, are less effective in mitigating poverty, social exclusion and deprivation. In this brief we argue that, when contrasted to other EU countries, the Latvian tax-benefit system is less effective in achieving either of the objectives.

Equity-Efficiency Trade-Off

There is a fundamental trade-off between the two principal objectives of a tax-benefit system – income redistribution and efficiency. On the one hand, income redistribution is desirable as it helps to mitigate socially undesirable market outcomes such as poverty and deprivation. On the other hand, more income redistribution is often associated with higher distortions to labour supply and work effort.

There is no universal prescription as to how much a government should redistribute. The answer to this question depends, among other factors, on the relative value that society (government) assigns to the welfare of different population groups, and on the individuals’ labour supply elasticity.

However, a given degree of income redistribution can be achieved at a different cost of efficiency. In this brief, we analyse the degree of income redistribution generated by the tax-benefit system and work incentives in Latvia in the context of other EU countries. In our analysis, we use the European microsimulation tax-benefit model EUROMOD (Sutherland and Figari, 2013) version G2.0, EU-SILC data, and the analysis framework developed by Jara and Tumino (2013).

Income Redistribution in the EU

EU countries differ substantially in terms of inequality of original income and in terms of the degree of redistribution generated by the tax-benefit system (see Figure 1, data on 2007 and 2013). The Gini coefficient of equivalised household original income (which consists of income from employment and self-employment, property income, private pensions, private transfers and other relatively minor components) ranges from around 0.4 (Cyprus, Netherlands) to almost 0.55 (Romania in 2007, Ireland in 2013).

Inequality of original income in Latvia in 2007 was at the EU average level (Gini coefficient of 0.47), but the degree of income redistribution generated by direct taxes, benefits and pensions was the lowest in the EU. As a result, the inequality of disposable income in Latvia in 2007 was the highest in the EU (Gini coefficient of 0.37). Part of the answer as to why the degree of income redistribution in Latvia is so low is a relatively small contribution of pensions to redistribution – it is almost half of that observed in the EU on average, despite the fact that the share of public pension recipients in the total Latvian population in 2007 was above the EU average. Another important factor was the very minor role of means-tested benefits: in the EU on average, means-tested benefits generate a reduction in Gini coefficient by about 0.02, while in Latvia the corresponding figure is just one tenth of this.

Figure 1. Gini coefficients of original equivalised household income and degree of redistribution generated by tax-benefit systems in the EU in 2007 and 2013

fig1afig1b

Source: EUROMOD statistics, authors’ calculations.

In the course of the crisis and the following recovery, the degree of redistribution in Latvia increased (see lower panel of Figure 1). An important factor behind the increase was growing number of pension recipients and an increase in the average size of pensions (both in absolute terms and relative to employment income). The increase in the number of pension recipients was not a result of changes in eligibility criteria, but was due to population ageing and the fact that more people applied for other types of pensions. The growth in the average size of pension was due to generous indexation of pensions in 2008 and compositional changes, as pensions of new pensioners until 2012 were larger than the average pension. Another reason for a growing degree of redistribution was an increase in the size and the number of recipients of means-tested benefits (mainly Guaranteed Minimum Income (GMI) benefit). This was a result of reforms in the provision of the means-tested benefits and of falling incomes from employment, which made more people eligible for the social assistance programmes. Nevertheless, despite the increase in recent years, the degree of income redistribution in Latvia remains one of the lowest in the EU.

Work Incentives

The existence of a trade-off between income redistribution and better work incentives suggests that tax-benefit systems that ensure less income redistribution are likely to generate better work incentives. Jara and Tumino (2013) have demonstrated the existence of this trade-off in the EU countries in 2007-2010 by identifying a negative and statistically significant correlation between Gini coefficients and Marginal Effective Tax Rates (METR). The METR is a measure that is commonly used to quantify work incentives at the intensive margin. It shows what proportion of a small increase in earnings (which results from e.g. an increase in the supplied hours of work) is lost as a result of extra tax payments or foregone benefits that the person is no longer eligible for after the increase in earnings. The negative correlation identified in Jara and Tumino (2013) suggests that countries with less income redistribution (i.e., higher Gini coefficients) tend to have better work incentives (lower METRs).

In Latvia, the mean METR in 2013 was 32.2%, only slightly below the EU average (34.5%), and much higher than the average in Estonia (22.8%) and Lithuania (27.4%), despite a lower degree of income redistribution (EUROMOD statistics). Another feature of the Latvian tax-benefit system is that it is characterised by especially high METRs for poor individuals. Thus, in 2013, 94% of individuals who faced METRs in excess of 50% belonged to the two bottom deciles of distribution of equivalised disposable income. This is different from many other European countries, where distribution of high METRs is either more even across deciles or rising towards the top end of income distribution (Jara and Tumino (2013), data for 2007).

The main reason for high METRs faced by the poorest population groups in Latvia is the design of means-tested benefits (GMI and housing benefits), which generates 100% METRs for the recipients of these benefits. Namely, for each additional euro earned, the amount of benefit is reduced by one euro, which leaves the net income unchanged. This adversely affects employment incentives for the poorest individuals and increases the poverty risk.

Figure 2 illustrates mean METRs by deciles of equivalised disposable income in Latvia and shows the contribution of taxes, benefits and social insurance contributions (SICs) to the mean METRs. It clearly demonstrates that high METRs in the bottom deciles result mainly from the contribution of benefits, which disappears in the fourth decile. The contribution of SICs is slightly smaller in the bottom decile, which is due to the fact that the proportion of employed individuals is smaller in the bottom decile. For the same reason, and also because of basic tax allowances, the contribution of direct taxes is smaller in the bottom deciles, but then the contribution of taxes levels off, reflecting the Latvian flat tax rate.

Figure 2. The contribution of direct taxes, benefits and social insurance contributions (SIC) to METRs in Latvia by deciles of equivalised disposable income in 2013

fig2

Source: authors’ calculations using EUROMOD-LV

In their study on the incentive structure created by the tax and benefit system in Latvia, the World Bank (2013) pointed out the problem of bad work incentives generated by Latvian means-tested benefits. Our results, which are based on a population-representative database of incomes, also identify means-tested benefits as the major contributor to high METRs in the lowest deciles of the income distribution. Another concern expressed by the World Bank (2013) was that the problem of informal employment (either in the form of undeclared wages or work without a contract) can be exacerbated by high participation tax rates and METRs.

Conclusion

The Latvian tax-benefit system is characterized both by a relatively low degree of income redistribution and relatively weak work incentives, as measured by METRs. Recipients of means-tested benefits (GMI and housing benefits) are faced with 100% METRs, as benefits are withdrawn at the same rate as household income rises. This creates disincentives to increase labour supply for low-paid/low-skilled individuals, and hence creates a risk of poverty traps. Evidence from the literature suggests that the labour supply of low paid workers is particularly sensitive to the incentives generated by the tax-benefit system, hence reforms that would bring down METRs in the bottom deciles could yield positive results in terms of employment of low paid/low skilled workers.

A potential reform is to introduce either a gradual phasing out of the means-tested benefits, or to exclude a certain amount of employment income from the income test for the means-tested benefits. Such reforms would be targeted at the bottom end of the income distribution, help combat poverty, improve the incentive structure of the Latvian tax-benefit system, and positively affect the labour supply of low-skilled/low-paid workers.

References

  • EUROMOD statistics on Distribution and Decomposition of Disposable Income, accessed at http://www.iser.essex.ac.uk/euromod/statistics/ using EUROMOD version no. G2.0, retrieved on October 14, 2014
  • Jara, H. Xavier & Alberto Tumino (2013). “Tax-benefit systems, income distribution and work incentives in the European Union,” International Journal of Microsimulation, Interational Microsimulation Association, vol. 1(6), pages 27-62.
  • Sutherland, Holly & Francesco Figari (2013). “EUROMOD: the European Union tax-benefit microsimulation model,” International Journal of Microsimulation, Interational Microsimulation Association, vol. 1(6), pages 4-26.
  • World Bank (2013). “Latvia: “Who is Unemployed, Inactive or Needy? Assessing Post-Crisis Policy Options”. Analysis of the Incentive Structure Created by the Tax and Benefit System. Financial Incentives of the Tax and Benefit System in Latvia,” European Social Fund Activity “Complex support measures” No. 1DP//1.4.1.1.1./09/IPIA/NVA/001