The gender pay gap: Discrimination or structural differences?

IZAWOL.16.gaBy Solomon W. Polachek

Women earn less than men. In the US the gap is approximately 22%. Among OECD countries the gap averages 15%. One might argue this indicates rampant discrimination. But the story is far more complicated.

Many women earn more than men, and for those that earn less, the gap is not uniform. For 55–64 year olds in the US the gap is close to 25%, yet for 16–24 year olds the gap is just 5%. For single-never-marrieds the gender gap is about 5%, but for married men and women the pay gap is almost 23%. Children exacerbate the gap between 2% and 10% per child, and the gap is even bigger when children are spaced widely apart. In contrast, the median salary for full-time young single women in a number of metropolitan areas exceeds men’s by over 8%.

These same patterns persist across most countries. The gender wage gap for marrieds is between 3 and 30 times greater compared to singles. Similarly children widen the gap.

If discrimination were the reason, then one would need a theory about why employers discriminate less against young childless women employees, but so much more against older and married women.

Perhaps the explanation isn’t gender discrimination, but lifetime work commitment. Work patterns also differ by gender, marital, and child status. In 1970 US married men’s labor force participation rate was 86%; married women’s was 40%. In 2010 it was 76% for married men and 61% for married women. For singles these gender differences are much smaller. In 1970 the labor force participation rate was 66% for single men and 57% for single women, and in 2010 they were 67% for single men versus 63% for single women. Even today, many women still drop out to raise children. Studies indicate many are hesitant to work long hours. In short, single men and women accumulate experience at roughly similar rates, but married women accumulate far less labor market experience than married men.

One cannot overemphasize the importance of work continuity. Dropping out is costly. Earnings power depreciates up to 5% for each non-work year. Those who expect to drop out tend to enroll in less job-related schooling and to take jobs with less training and lower earnings growth. The same is true for those who seek shorter work hours and smaller commutes.

Over the last century female lifetime work increased dramatically, while male lifetime work has declined moderately. Concomitant with this gender work convergence is a decline in the gender earnings gap from approximately 70% in the early 1800s to the current 22% in the US. Similar patterns hold in other countries.

Policies consistent with promoting greater female lifetime work effort have reduced the gender wage gap. One such policy is making day care more available. An analysis by the OECD finds smaller gender wage gaps in countries with greater day care enrollment. Lower marginal tax rates that encourage greater participation of women in the labor force would work in the same direction.

Related IZA World of Labor article:
Equal pay legislation and the gender wage gap, by Solomon W. Polachek


Women’s lower wage sensitivity

Another IZA World of Labor article by Boris Hirsch (Leuphana University of Lüneburg and IZA) points to recent studies finding that imperfect competition in the labor market can account for a large part of the unexplained wage gap. The gap thus reflects “monopsonistic” wage discrimination—that is, employers exploiting their wage-setting power over women—rather than any sort of prejudice.

Hirsch argues that many women, due to domestic responsibilities, tend to care more about commuting times or working hours than salary offered. While equal pay legislation may help prevent employers from exploiting their wage-setting power, Hirsch recommends addressing the sources of women’s limited wage sensitivity by investing in additional and better childcare or enforcing more flexible working time arrangements.

Gender gap in leadership positions

Gender wage gaps are most pronounced in leadership positions. As Mario Macis (Johns Hopkins University and IZA) explains in his article, this is due to a combination of economic forces, cultural and social norms, discrimination, and unequal legal rights. Apart from social justice concerns, these gender disparities are economically  inefficient because they imply a sub-optimal allocation of female talent.

Gender differences in competitiveness

A common explanation of the gender pay gap is that attitudes towards competition differ between genders. Mario Lackner‘s (Johannes Kepler University Linz) article references laboratory and field experiments finding that women are more reluctant and less aggressive when it comes to initiating negotiations or applying for jobs with negotiable salaries.

According to the article, such differences in competitiveness are formed at young ages and are relatively persistent, exerting a profound influence on an individual’s future career. Consequently, policy measures should be targeted at early childhood education and education systems in general. The driving forces of this ‘competitiveness gap’ are, however, still under debate. Moreover, closing the gender gap in competitiveness might not be desirable under all circumstances, as men are often found to be overconfident and over-competitive.

Read more on the IZA World of Labor “key topic” page:

Posted in Opinions, Research | Tagged , , , , , , , , , , , , | Leave a comment

How the origins of language structures may help explain current economic outcomes

A growing literature in economics has demonstrated that grammatical features of languages influence our economic decisions. For example, whether or not a language makes a clear-cut distinction between present and future (while an English speaker uses a future marker as in “it will rain tomorrow,” a German speaker can simply rely on present tense as in “it rains tomorrow”) may affect intertemporal choices with regard to savings or retirement.

But why do languages differ in grammatical features such as tense, gender or politeness distinctions in the first place? New research by Oded Galor (Brown University and IZA), Omer Ozak (Southern Methodist University), and Assaf Sarid (University of Haifa) published in a recent IZA Discussion Paper argues that these differences can to a significant degree be explained by differences in historical economic and geographic circumstances – crop return, variation in agricultural productivity across genders, and ecological diversity – several hundred years in the past.

Crop return predicts the existence of a future tense

Languages differ in the structure of when and how they mark future events by whether speakers have to adjust a verb when talking about the future. For example, French speakers are required to change the form of the verb when changing a statement from the present (“Il fait froid aujour d’hui” (It is cold today)) to the future (“Il fera froid demain.” (It will be cold tomorrow)). Languages like this, which grammatically distinguish between present and future, have what is known as an inflectional future tense. By comparison, in Finnish, the present tense is used in reference to both the present (“Tänään on kylmää” (Today is cold)) and the future (“Huomenna on kylmää” (Tomorrow is cold)).

future-tenseIn an economic sense, possessing an inflectional future tense is believed to indicate an on average lower long-term orientation. The researchers argue that a long-term orientation is a cultural trait that should be inversely correlated to crop returns. In regions with high historical crop returns, people were able to rely on contemporary food production and thus would have had to care less about the future.

Results relating pre-1500CE potential crop returns to the existence of a distinct future tense confirm this hypothesis:  The higher the historical crop return in a geographical area, the less likely languages in this area developed a future tense.

crop-returnHistoric language development affects current economic outcomes

But do language structures indeed transmit pre-industrial crop returns into economic behavior today? Results focusing on second-generation immigrants in the US suggest there is such a long-term effect. Immigrants speaking a mother tongue with inflectional future tense display lower probabilities of attending college by 4 percentage points, indicating indeed a lower long-term orientation. Thus, the study by Galor and his colleagues sheds light on very long-run interlinkages between pre-industrial variation in geographical features, language structures and (current) economic outcomes.

Download the paper (IZA DP No. 10379):

Posted in Research | Tagged , , , , , , | Leave a comment

Decoding attitudes towards migrants

The recent political success of right-wing populists in the US and in many European countries is often attributed to a surge in anti-immigrant sentiments among broad parts of voter populations. Researchers, however, have struggled to understand the development of these hostile attitudes. What are key drivers and determinants? Can misperceptions be tackled? These questions are addressed in several recently published IZA Discussion Papers.

muslim_populationEmpirical data suggest that voters are often blatantly misinformed about the facts of immigration. As an example, the chart above shows a high degree of misperception among people worldwide about the share of Muslims living in their country. Similar figures have been found for over-estimations of immigrant proportions in general.

The good news is, people are willing to update their perceptions in response to newly provided factual information. This is the finding of an IZA Discussion Paper by Oxford economists Alexis Grigorieff and Christopher Roth with Diego Ubfal (Bocconi University, IGIER and IZA), who have studied whether providing official statistics about immigrants affects people’s attitude towards them.

The researchers designed an online survey for the US in which all 800 respondents were asked to estimate a range of figures regarding immigrants, which they wildly overestimated: share of immigrants (estimated 22% vs. correct 13%), of illegal immigrants (14% vs. 3%), of unemployed immigrants (22% vs. 6%), of incarcerated immigrants (13% vs. 2%) and of non-English speaking immigrants (33% vs. 8%).

Information about immigration positively affects attitudes

Half the sample was then provided with the correct figures followed by a continuation of the survey and further questions about attitudes and policy preferences. The information turned out to have a strong effect on self-reported attitudes: Respondents provided with factual information were 30% more willing to donate to a pro-immigrant charity. However, the effect does not extend to political behavior, as no difference was reported with regard to signing a petition asking for more Green Cards (permanent residence permits).

Those most affected by the new information were those most worried about immigration. Providing information displays a larger effect on Republicans than on Democrats. A follow-up survey a month later showed that the information effect persists.

Public information campaigns have a powerful impact

A second recent IZA Discussion Paper similarly demonstrates the power of information to counter misperceptions about immigration. Giovanni Facchini (University of Nottingham and IZA), Yotam Margalit (Tel Aviv University) and Hiroyuki Nakata (University of Leicester and RIETI) look at how broad information campaigns can decrease public opposition to immigration. They analyzed results from a large-scale experiment conducted in Japan, a country with widespread anti-immigrant sentiment. The researchers randomly exposed a large national sample of citizens to information pertaining to potential social and economic benefits from immigration.

The findings reveal that the campaign led to a substantial increase in support for a more open immigration policy and motivated citizens to take political action in support of this cause. Notably, while smaller in magnitude, many effects also persisted 10-12 days after the information was distributed, which highlights the potential value of public information campaigns to combat negative attitudes towards foreigners.

Minority salience influences extremism

Attitudes towards immigrants, however, do not form on externally provided information alone. People also become more aware of immigrants by being exposed to their cultural and religious practices in public life. The IZA Discussion Paper by Tommaso Colussi, Ingo E. Isphording and Nico Pestel investigates how a change in “salience” of Muslim communities in German municipalities influences voter behavior, potentially increasing the level of political extremism. The researchers use unique data on the construction of mosques and election results in municipalities over the period of 1980-2013.

The results indicate that the presence of a mosque increases residents’ political divergence from the political center. The negative effect of the presence of a mosque increases strongly if an election is scheduled right after the holy month of Ramadan, a period in which mosques happen to be much more visible to the general public due to extensive festivities and openly displayed religious practices. The findings show that vote shares for both far right- and left-wing parties become larger when the election date is closer to Ramadan. In addition, the change in visibility of the minority population increases the likelihood of politically motivated crimes against Muslims.

Read the complete papers:

Posted in Research | Tagged , , , , , , , | Comments Off on Decoding attitudes towards migrants

IZA/IAB Linked Evaluation Dataset: New resource for the analysis of labor market policies

High-quality data are the key to evaluating the effectiveness of labor market policies. While administrative datasets provide detailed and reliable information on individual job histories, survey data give insights into individual behavior, attitudes and subjective experiences. A joint project by IZA and IAB (Institute of Employment Research) now combines the “best of both worlds” into the newly available IZA/IAB Linked Evaluation Dataset 1993-2010 (LED).

idscThe dataset contains information on over 15,000 entrants into unemployment (between June 2007 and May 2008) who were interviewed for IZA over the subsequent three years and agreed to have their data matched with their individual employment biographies recorded by the IAB. The dataset is anonymized and individuals cannot be identified.

fdz-iabWhile the IAB administrative data provide detailed individual-level information on employment, benefit receipt and participation in active labor market programs, the IZA survey data cover a number of individual characteristics that affect search behavior and labor market outcomes.

The LED thus offers new opportunities for empirical labor market research – not only for the evaluation of active labor market policy instruments, such as training programs or wage subsidies, but also for many other aspects of the transition process from unemployment to employment.

More information on data access is available on the IAB and IZA websites:


Posted in IZA News, Research | Tagged , , , , , , , , | Comments Off on IZA/IAB Linked Evaluation Dataset: New resource for the analysis of labor market policies

On World Day of Social Justice: Insights on economic inequality from IZA World of Labor

Great Gatsby CurveThe UN has proclaimed February 20 the “World Day of Social Justice” to support efforts of the international community in poverty eradication, the promotion of full employment and decent work, gender equity and access to social well-being and justice for all. Despite substantial progress in the fight against poverty across the globe, the unequal distribution of income and opportunity between different groups in society features prominently in current political debates.

IZA World of Labor provides evidence-based insights from international research into various aspects of economic inequality. Here is an overview of related articles.

If you are born poor you will always be poor?

Income inequality and social origins
Promoting intergenerational mobility may make societies both more egalitarian and more efficient. The expectation that people, whatever their social origin, can raise their standard of living is a powerful incentive to human capital accumulation and personal effort. Policies to counteract disparities in family background, such as education interventions for poor children, may foster intergenerational mobility (Lorenzo Cappellari).

Does inequality persist across generations?
A strong association between incomes across generations—with children from poor families likely to be poor as adults—is frequently considered an indicator of insufficient equality of opportunity. Studies of such “intergenerational persistence,” or lack of intergenerational mobility, are concerned with measuring the strength of the relationship between parents’ socio-economic status and that of their children as adults. However, reliable measurement requires overcoming important data and methodological difficulties (Jo Blanden).

Not all people who are poor are persistently poor
Evidence suggests that unemployment, retirement, and single parenthood are closely associated with persistent poverty and that higher education tends to protect against it. There is also evidence of a poverty trap, meaning that policy should aim to prevent people from falling into poverty because once poor, the probability of being poor in the future increases (Martin Biewen).

What can be done to counteract inequality?

Do skills matter for wage inequality?
Differences in wage inequality across countries are driven primarily by differences in the return to skills, which is determined in part by labor market institutions, but also by how well the supply of skills meets the demand. A comprehensive policy package to tackle wage inequality should include a focus on skills, reforms of labor market institutions that influence how skills are rewarded, and alignment of skill supply and demand (Stijn Broecke).

Can social security programs reduce wealth inequality?
How well social security programs reduce inequality depends on program design and implementation and labor market and population characteristics. To reduce inequality and avoid labor market disincentives, the best design appears to be a mostly proportional contributory program complemented by a well-designed non-contributory component (Alvaro Forteza).

The integration of productive inclusion programs into social assistance systems can reduce poverty
However, productive inclusion will not work for everybody, and even when it shows impacts some households—especially the poorest and most marginalized—will still need assistance. To maximize impacts, it is important to tailor productive inclusion programs to match the beneficiaries’ profiles. Furthermore, the right monetary and design incentives must be provided so that social and productive inclusion programs can effectively coordinate activities, exchange information, and refer beneficiaries (Jamele Rigolini).

Can education reduce inequality?

What role does preschool play in reducing inequality?
Good-quality preschool programs more than pay for themselves by boosting achievement and reducing inequality of achievement. That is good news, especially for countries with persistent and high levels of inequality—and a good reason to expand preschool programs in countries where enrollment is far from universal (Jane Waldfogel).

Can higher education reduce inequality in developing countries?
Projections indicate that the global labor market will face continued disequilibrium. Excess labor supply is expected from less developed regions, while excess demand is expected from developed and emerging economies. At the same time, the global economy is becoming increasingly knowledge-driven. Hence, investment in vocational and higher education is important for developing countries to remain competitive; further, expanding the skill-base of the labor force may lead to lower levels of wealth inequality (Abebe Shimeles).

Slavery, racial inequality, and education
Evidence suggests that in some countries historical slavery has influenced the racial distribution of human capital and income inequality. A regional comparison of the influence of slavery on education, racial education inequality, and income distribution shows that policies aimed at addressing inequality can account for differential effects of past slavery on current outcomes. Policies designed to remove racial education inequalities in schools can favor income equalization, though given the resilience of the effect of past slavery they are by no means an immediate solution. Nevertheless, education policy clearly has a strong influence over time (Graziella Bertocchi).

Posted in Research | Tagged , , , , , , , , , , | Comments Off on On World Day of Social Justice: Insights on economic inequality from IZA World of Labor

Does modern technology slow down employment growth after recessions?

robots at workIn the last 25 to 30 years, recoveries from recessions in the US have been plagued by weak employment growth. While before 1990 employment growth during the two years following a recession was a little over five percent, it has been just under one percent since then. One possible explanation for the slower recovery of jobs is related to technological change.

A 2012 landmark paper on job polarization in the US found that middle-skilled jobs, those usually involving routine tasks that are particularly susceptible to replacement by new technologies, may be destroyed permanently during recessions. The displaced workers from these jobs are then forced into time-consuming transitions to different occupations and sectors, resulting in slow job growth during the recovery.

This phenomenon of labor market polarization (or “hollowing out” of middle-skilled jobs) has attracted widespread attention and contributed to the ongoing debate on the impact of technological change on labor markets. While much of the focus has been on the United States, Georg Graetz (Uppsala University and IZA) and Guy Michaels (London School of Economics and IZA) investigate in their recent IZA Discussion Paper whether labor markets in other countries have also been slow to pick up after recessions, and if modern technology could be to blame.

“Jobless recoveries” a US phenomenon

The authors analyze data from 71 recessions across 28 industries in 17 countries from 1970-2011 to first examine whether recoveries from recessions after 1985 produced slower employment growth than earlier recoveries. They then test if industries that are more susceptible to technological change have had particularly slow employment growth during recoveries. Finally, the researchers investigate whether routine-intensive industries have seen more replacement of middle-skill jobs during recessions and recoveries.

The results suggest that technology has not impeded job growth in recoveries outside the US. While GDP recovered more slowly after recent recessions in the countries under study, employment did not. Neither employment in industries more exposed to technological change nor middle-skilled employment experienced slower recoveries.

These findings pose a puzzle as to the nature of poor employment trends seen during recent recoveries in the US. Graetz and Michaels point to two possible explanations. The first is related to the differences in technology adoption found in previous studies. The second possible explanation appeals to US-specific policy and institutional changes: Unemployment benefit extensions, which increase workers’ reservation wages, may slow down employment growth during recoveries. Moreover, the declining role of unions may have facilitated the substitution of workers during recessions and recoveries. The authors stress, however, that further research is needed to establish the relative merits of the technology- and policy-based explanations.

CaptureRead the complete paper (IZA DP No. 10470):

See also the media coverage at Bloomberg View.

photo credit: Maksim Dubinsky via Shutterstock

Posted in Research | Tagged , , , , , , , | Comments Off on Does modern technology slow down employment growth after recessions?

How to improve compliance with sectoral minimum wages

Research on minimum wages focuses mainly on their impact on employment. More recently some papers have looked also at prices, profits and productivity. But they disregard a potentially important issue: non-compliance.

Economists tend to assume that the minimum wage is respected by all firms, which is not the case, especially where authorities seem to turn a blind eye. There have been some attempts to estimate non-compliance with national minimum wages, especially in developing and emerging economies (see IZA World of Labor for an overview).

A new IZA Discussion Paper by Andrea Garnero (OECD, ULB and IZA) extends the research on this topic by looking at non-compliance in the case of wage floors set by collective bargaining at the sectoral level in Italy. The results show that non-compliance rates are sizeable (around 10%) and also the amount of underpayment is quite large (20%, which means e.g. a loss of 200 euros for the worker when the wage floor is 1,000 euros per month).

minimum wage violations by region

Violation of hourly sectoral minimum wages, by Italian region in 2015
Source: IZA Discussion Paper No. 10511.

“Not surprisingly, non-compliance is particularly high in the South and in micro and small firms, and it affects especially women and temporary workers. Overall, the Italian collective bargaining system seems unable to safeguard a level playing field for firms and ensure that minimum wage increases are effectively reflected into pay increases for workers at the bottom of the distribution,” says Garnero.

His paper therefore develops a series of relatively simple and almost free-of-cost policies to increase compliance beyond more effective inspections:

  • Streamline the number of collective agreements
  • Ensure that agreements are signed by representative unions and employers’ organizations
  • Make the information on negotiated wages publicly and easily available
  • Establish a helpline and/or an online form for employers and workers
  • Awareness campaigns
  • Name and shame

For a detailed description see the complete paper (IZA DP No. 10511):

Posted in Research | Tagged , , , , , , , | Comments Off on How to improve compliance with sectoral minimum wages