Article

# Statistical Theories of Discrimination in Labor Markets

Authors:
If you want to read the PDF, try requesting it from the authors.

## Abstract

Examines economic discrimination in labor markets using a stochastic model. Analysis of several types of economic discrimination within the context of competitive market assumptions; Empirical plausibility and implications of the alternative models of economic discrimination; Role of statistical theories in the explanation of labor market discrimination. (Abstract copyright EBSCO.)

## No full-text available

... However, evaluation should generally be performed on as many demographic groups as are known. 3 (2) Handling progressively smaller groups (Sec. 5): the more identities we consider, the smaller each group is likely to be. ...
... Thus, we will not perform what might be considered a more model-agnostic approach of unsupervised learning on the dataset itself. 10 Differing heterogeneity is related to "second moment" statistical discrimination in economics: marginalized groups, for structural reasons like not being given sufficient opportunity to demonstrate ability, have a higher perceived variance and are discriminated against by risk-averse employers [3,30,36]. ...
... We acknowledge that the group identity delineations themselves are unstable and fraught with problems of operationalization[61,92].3 There are concerns regarding noisy measurements of small groups that are out of scope for our work; we refer the reader to Foulds et al.[39]. ...
Preprint
Full-text available
Research in machine learning fairness has historically considered a single binary demographic attribute; however, the reality is of course far more complicated. In this work, we grapple with questions that arise along three stages of the machine learning pipeline when incorporating intersectionality as multiple demographic attributes: (1) which demographic attributes to include as dataset labels, (2) how to handle the progressively smaller size of subgroups during model training, and (3) how to move beyond existing evaluation metrics when benchmarking model fairness for more subgroups. For each question, we provide thorough empirical evaluation on tabular datasets derived from the US Census, and present constructive recommendations for the machine learning community. First, we advocate for supplementing domain knowledge with empirical validation when choosing which demographic attribute labels to train on, while always evaluating on the full set of demographic attributes. Second, we warn against using data imbalance techniques without considering their normative implications and suggest an alternative using the structure in the data. Third, we introduce new evaluation metrics which are more appropriate for the intersectional setting. Overall, we provide substantive suggestions on three necessary (albeit not sufficient!) considerations when incorporating intersectionality into machine learning.
... The analysis is divided into three-part. First, to provide explanations and insights Information Frictions and Early-Stage Investors: Evidence from a Crowd-Rating Platform for my analysis, I developed a simple model of signal extraction based on the statistical discrimination literature (Aigner and Cain, 1977;Phelps, 1972). The model proposes that potential early-stage investors observe a noisy signal from the aggregate score of the crowd about the project's quality, and update beliefs according to this signal and their own information. ...
... Early-stage investors, as well as evaluators, which hold appropriate beliefs about aggregate group differences, fall into a statistical discrimination framework (Phelps, 1972). Statistical discrimination can be explained by rational expectations, where evaluators in period 0 form prior beliefs about projects' quality that differ by aggregate group characteristics, inferring individual quality based on these group statistics (Aigner and Cain, 1977;Altonji and Pierret, 2001). For example, evaluators can perceive certain group observable characteristics as higher, using these characteristics as a proxy for venture quality to infer screening of NPV. ...
... In this section, I develop a signal extraction model following the setups in Phelps (1972) and Aigner and Cain (1977), in which potential investors observe aggregate evaluations from platform users, and use this information to form investment decision 19 . ...
Thesis
This thesis studies different aspects of the factors that directly or indirectly impact innovative activities both at the macroeconomic and microeconomic levels. In a context where policymakers and firms consider innovation as a strategic asset for productivity growth, this thesis aims at contributing to the literature on the determinants of innovation and related market failures relying on primarily empirical contributions. The first chapter considers the impact of competition and trade openness on innovation. Country innovation intensity positively responds to less stringent regulation, but only domestic product-market reform is directly related to innovation. The second chapter evaluates a European program that supports SME’s innovation. R&D grants positively impact patenting, but this effect is stronger for more financially constrained firms by a certification mechanism on the quality of firms. Finally, the third chapter considers the role of information frictions among a crowd-rating framework, on ventures’ subsequent success. This chapter uses a novel sample of French ventures at both the idea and seed stage. Taken together, this thesis explores three different instruments that aim to spur innovation intensity, either in terms of R&D, patents, financing, and venture success outcomes.
... The author also provides a set of theories, which focus on the reasons of discrimination against the immigrants in the labour market. For instance, monopsonistic discrimination emerges as a result of lack of competition for labour on demand side (Madden 1973, in Kogan 2007; while the error discrimination takes place because of lack of information and false beliefs about the true productivity of the workers (Cain 1976;England 1992, in Kogan 2007; and statistical discrimination is similar to error discrimination, but which is a false belief about the true productivity of potential workers, therefore using the information about the group (Arrow 1972;Phelps 1972;Aigner andCain 1977, in Kogan 2007); whereas taste discrimination is based on the personal preferences of employers, employees or the costumers (Becker 1971, in Kogan 2007. ...
... Some country-specific human capital aspects such as language skills and cultural knowledge (Borjas 1994;Esser 1999, in Kogan 2007 and the reluctance of the immigrants in investing in the human capital after immigration (Kogan 2007) may affect the level of differentiation. Besides, there are certain types of discrimination like monopsonistic discrimination that emerges as a result of lack of competition for labour on demand side (Madden, 1973in Kogan 2007; the error discrimination appearing because of lack of information and false beliefs about the true productivity of the workers (Cain 1976;England 1992;in Kogan 2007); statistical discrimination, which is similar to error discrimination, but which is a false belief about the true productivity of potential workers, therefore using the information about the group (Arrow 1972;Phelps 1972;Aigner andCain 1977, in Kogan 2007); and finally taste discrimination based on the personal preferences of employers, employees or the costumers (Becker 1971, in Kogan, 2007. These sorts of discrimination are other obstacles before the migrants in the labour market of the host country. ...
... Although the empirical analysis in the study particularly the disparities in the labour market outcomes between the natives and migrants on the supply side, the findings are likely to point out the clues for the reasons of the differentials related to the demand side, which are examined by the theoretical approaches for discrimination against the immigrants in the labour market. Discrimination based on migration status may occur because of lack of competition for labour on demand side (Madden 1973, in Kogan 2007; or lack of information and false beliefs about the true productivity of the workers (Cain 1976;England 1992, in Kogan 2007; or a false belief about the true productivity of potential workers, therefore using the information about the group (Arrow 1972;Phelps 1972;Aigner andCain 1977, in Kogan 2007); or because of personal preferences of employers, employees or the costumers (Becker 1971, in Kogan 2007. The persistence of substantial wage gaps between the migrants with non-EU origins and the native population in Belgium, France, the Netherlands and Sweden shown in Chapter 1 after controlling all available socio-demographic, occupational, sectoral and work experience related variables suggests the existence of at least one type of such demand-side discrimination against migrants from particular regions of the world. ...
Thesis
The main purpose of this PhD study, which is titled as “Socio-Economic Disparities in the Integration Process of Immigrants in Western Europe: A Comparative Study for Six EU Countries”, is to examine several aspects of socio-economic disparities between the native population and the migrants in Belgium, Germany, France, the Netherlands, Sweden and United Kingdom. These countries are selected because of their diverse labour market and welfare state characteristics as well as their different migration and integration policies. The study covers literature review and empirical analysis on the disparities in the labour market, access to social benefits and education-related outcomes. The analyses also aim at demonstrating the relationships between these three domains, where there are available data. In the analyses, the European Union Statistics on Income and Living Conditions (EU-SILC) and European Union Labour Force Survey (EU-LFS), OECD Programme for International Student Assessment (OECD-PISA) and European Social Survey (ESS) data sets are used in different sections (in relation with their contents) as these data sources have high level of comparability across countries and over time. The analysis results of the study demonstrate the existence of remarkable socio-economic disparities between the migrants and the native populations in the selected Western European countries; re-generation of the disparities between groups in different socio-economic domains in different phases of lifecycle as a result of the interaction of the disadvantages in more than one domain; and the emergence of similar disparities across groups in all host countries regardless of the differences in their labour market, welfare state and education system settings in the macro-level. Further research topics and policy recommendations based on the literature and analysis results are also presented in the study.
... Many statistical discrimination models (see e.g., [2,3,9] and a survey in [14]) assume that individuals of different groups have identical a priori characteristics (e.g., cost-of-effort), but that the decision-maker uses their group-dependent belief when there is imperfect information to assess the performance of individuals in a group. In some cases, the discriminating beliefs (stereotypes) of decision-makers lead to equilibria in which these stereotypes are fulfilled. ...
... where is a zero-mean centered random variable from the standard normal distribution N (0, 1); the noise varianceˆ2 is assumed group-dependent. 2 The quality estimateˆis a noisy but unbiased measurement of the quality of a candidate , e.g., an interview result. The group-dependent variance of noiseˆ2 models the information inequality: if interviewers are more familiar with candidates of some demographic groups, they are more confident in their evaluation compared to that of candidates of other groups. ...
... For a given population , we denote by¯u n the average effort that candidates from group exert at equilibrium, and we denote by¯u n the selection rate which is the probability for a randomly chosen candidate from group to be selected. Since ( ; ) is the probability for a -candidate to be selected when exerting an effort (see (2)), these quantities satisfy: We say that the group (or ) is underrepresented if¯<¯(or¯<¯). Note that when we say "underrepresented', it means that a demographic group has less representation in the selection than its share in the population of all candidates. ...
Preprint
To better understand discriminations and the effect of affirmative actions in selection problems (e.g., college admission or hiring), a recent line of research proposed a model based on differential variance. This model assumes that the decision-maker has a noisy estimate of each candidate's quality and puts forward the difference in the noise variances between different demographic groups as a key factor to explain discrimination. The literature on differential variance, however, does not consider the strategic behavior of candidates who can react to the selection procedure to improve their outcome, which is well-known to happen in many domains. In this paper, we study how the strategic aspect affects fairness in selection problems. We propose to model selection problems with strategic candidates as a contest game: A population of rational candidates compete by choosing an effort level to increase their quality. They incur a cost-of-effort but get a (random) quality whose expectation equals the chosen effort. A Bayesian decision-maker observes a noisy estimate of the quality of each candidate (with differential variance) and selects the fraction $\alpha$ of best candidates based on their posterior expected quality; each selected candidate receives a reward $S$. We characterize the (unique) equilibrium of this game in the different parameters' regimes, both when the decision-maker is unconstrained and when they are constrained to respect the fairness notion of demographic parity. Our results reveal important impacts of the strategic behavior on the discrimination observed at equilibrium and allow us to understand the effect of imposing demographic parity in this context. In particular, we find that, in many cases, the results contrast with the non-strategic setting.
... For example, let σ 2 ǫB > σ 2 ǫW , so that the signal is more noisy for group B individuals. The productivity inference made by the employer is therefore more responsive to the signal for group W than it is for group B. Specifically, a group-B person who draws an above average signal realization is interpreted as having a lower 6 See Aigner and Cain (1977) for a thorough discussion of alternative definitions of statistical discrimination. expected productivity than a W -group person with the same signal realization. ...
... Based on this observation, Aigner and Cain (1977) qualify that Phelps' (1972) model of discrimination "does not constitute economic discrimination, statistical or otherwise." Instead, they propose an alternative definition of statistical discrimination, referring to a group-level phenomenon, rather than Phelps' individual-level assessment: Group-level statistical discrimination arises when two groups with the same underlying productivity distribution, but different productivity-signaling technologies, receive different wages on average. ...
... This result follows directly from the martingale property of posteriors: the expected posterior mean must equal the prior mean. Aigner and Cain (1977) note, instead, that group-level discrimination arises if employers offer a non-linear wage schedule -for example, due to the employer being risk-averse. In the model of normally distributed signals in the previous section, if the wage offered by the employer is a convex function of the induced posterior mean, then the group with higher signal accuracy receives higher wages on average. ...
Preprint
Full-text available
This paper surveys the literature on theories of discrimination, focusing mainly on new contributions. Recent theories expand on the traditional taste-based and statistical discrimination frameworks by considering specific features of learning and signaling environments, often using novel information- and mechanism-design language; analyzing learning and decision making by algorithms; and introducing agents with behavioral biases and misspecified beliefs. This survey also attempts to narrow the gap between the economic perspective on "theories of discrimination" and the broader study of discrimination in the social science literature. In that respect, I first contribute by identifying a class of models of discriminatory institutions, made up of theories of discriminatory social norms and discriminatory institutional design. Second, I discuss the classification of discrimination as direct or systemic, and compare it to previous notions of discrimination in the economic literature.
... However, if the decision-maker is restricted to not using the group information for wage assignment, the effort is equal for both groups. In many statistical discrimination models, the authors assume that individuals of different groups have identical a priori characteristics, but the decision-maker uses their group-based beliefs when facing imperfect information to assess the performance of individuals of a particular group (Aigner and Cain, 1977;Arrow, 1973;Coate and Loury, 1993;Fang and Moro, 2011). In some cases, these beliefs (stereotypes) lead to equilibria in which these discriminating beliefs are fulfilled. ...
... Many statistical discrimination models (see e.g., (Aigner and Cain, 1977;Arrow, 1973;Coate and Loury, 1993) and a survey in (Fang and Moro, 2011)) assume that individuals of different groups have identical a priori characteristics (e.g., cost-of-effort), but that the decision-maker uses their group-dependent belief when there is imperfect information to assess the performance of individuals in a group. In some cases, the discriminating beliefs (stereotypes) of decision-makers lead to equilibria in which these stereotypes are fulfilled. ...
Thesis
Data-driven decision-making algorithms are increasingly applied in many domains with high social impact, such as hiring, lending, or criminal justice. Recently, it was shown that such algorithms could lead to discrimination against certain demographic groups (e.g., they can discriminate by race, gender, or age). This led to a recent active line of research---called algorithmic fairness---which studies how to develop efficient algorithms with fairness guarantees. Most of the decision problems with high social impact mentioned above are essentially selection problems. In selection problems, the decision-maker must select a fixed fraction of the best candidates given their characteristics. The notion of a selection budget contrasts selection problems with classification problems typically studied in the algorithmic fairness literature.In this thesis, we study the causes of discrimination in selection problems and the impact of fairness mechanisms on the utility of selection. Our first contribution considers a selection problem with candidates whose quality is measured with a group-dependent noise---a phenomenon called differential variance. We study the impact of differential variance on group representations and how standard group fairness mechanisms affect the selection utility in the presence of differential variance. Our second contribution proposes a game-theoretic model of a selection problem with differential variance. We assume strategic candidates who maximize the individual utility by making a costly effort. The effort induces their quality, measured by a (group-fair) decision-maker with group-dependent noise. We characterize the equilibrium of such a game. In our third contribution, we consider a multistage selection problem. We extend classical group fairness notions to a multistage setting and propose the notions of local (per stage) and global (final stage) fairness. We then introduce and study the price of local fairness which is the ratio of utilities induced by the globally fair algorithm to that of the locally fair algorithm.
... First, if wage determination and career advancements are affected by taste-discriminatory behavior of (mainly male) supervisors and managers, a larger representation of women at the top of the occupational hierarchy is expected to reduce the gender wage gap and provide more opportunities (for women) to be promoted (Albrecht et al., 2003;Becker, 1957). Second, it has been argued that under imperfect information female managers might be better at inferring other women's unobserved productivity, hence reducing statistical discrimination toward women (Aigner & Cain, 1977). In this respect, females are likely to receive higher wages when employed by a female manager rather than by a male, while lower wages are likely to be paid to males by female managers. ...
... In the first case, female bosses are found to have no (or at least lower) taste for gender discrimination, as opposed to male bosses, which could be rationalized in terms of prejudice, cultural factors, and social norms (Bertrand, 2011). In the second case, female bosses are deemed to be better at assessing the (unobserved) productivity of their female subordinates, thus improving the (gender) allocation of work as well as the (gender) gap in rewards, thereby reducing perceived discrimination (Aigner & Cain, 1977). ...
Article
Full-text available
In this paper, we exploit rich cross-country survey data covering 15 European countries over the period 2000–2015 to investigate the relationship between the gender of the immediate supervisor (i.e. having a male or a female “boss”) and perceived gender discrimination at the workplace. We show that a female boss is associated with reduced gender discrimination, with positive spillovers mainly on female subordinates, in jobs where female presence is also higher and where work organization is more complex. The presence of more flexible work schedules and a better balance between work and life, further contributes to reinforce the mitigating effect of female leadership on discrimination. Results are shown to be consistent with available evidence on gender differentials in pay and career advancement, as well as being robust to a number of sensitivity checks.
... In other words, this type of bias is a form of "informational bias" in that some type of information will largely explain the observed disadvantages (Aigner and Cain 1977;Correll and Benard 2006;Ewens, Tomlin, and Wang 2014;Neumark 2018;Pager and Karafin 2009;Phelps 1972). On the other hand, a competing perspective suggests that decision makers may remain biased in light of rich, positive, and counterstereotypical information, with stereotypes and cultural beliefs creating more rigid biases that are relatively persistent and difficult to change (Correll and Benard 2006). ...
... Although sociologists, economists, and social psychologists vary in their exact formulas for describing these theoretical typologies, they are consistent in the primary differentiator between each typology: whether biases are responsive or resistant to clear, detailed, and positive or counter-stereotypical information (Bertrand and Mullainathan 2004;Biernat and Fuegen 2001;Bills et al. 2017;Bosch, Carnero, and Farré 2010;Correll and Benard 2006;Ewens et al. 2014;González, Cortina, and Rodríguez 2019;Kunda and Sherman-Williams 1993;Neumark 2018;Pager and Karafin 2009;Rubinstein, Jussim, and Stevens 2018). Informational bias, related to statistical discrimination in economics, occurs when decision makers are faced with insufficient information during the decision-making process and use assumptions about group characteristics to make inferences about a specific candidate (Aigner and Cain 1977;Chambers and Echenique 2018;Correll and Benard 2006;Neumark 2018;Phelps 1972). In other words, under this framework, rational evaluators fill in informational shortages with their own knowledge or with group-level stereotypes. ...
Article
Full-text available
Employment interruption is a common experience in today’s labor market, most frequently due to unemployment from job loss and temporary lapses to care for family or children. Although existing research shows that employment lapses cause disadvantages at the hiring interface compared to individuals with no employment disruptions, competing theories predict different mechanisms explaining these hiring penalties. In this study, the author uses an original conjoint survey experiment to causally assess perceptions of fictitious job applicants, focusing on a comparison of unemployed applicants and nonemployed caregiver applicants, who left work to care for family, to currently employed applicants. The author examines whether disadvantages for job applicants with employment gaps are receptive to positive information (and therefore represent a form of “informational bias”) or are resistant to information (reflecting “cognitive bias”) and further assesses which types of information affect or do not affect levels of bias in fictitious hiring decisions. Results show that positive information on past job performance and social skills essentially eliminates disadvantages faced by unemployed job applicants, but nonemployed caregiver applicants remain disadvantaged even with multiple types of positive information. These findings suggest that unemployed applicants face informational biases but that nonemployed caregiver applicants face cognitive biases that are rigid even with rich forms of positive or counter-stereotypical information. This study has implications for understanding the career consequences of employment disruption, which is especially relevant to consider in light of labor market disruptions during the recent pandemic.
... Statistical discrimination theory pioneered by Phelps (1972) and Arrow (1973) posits that in the absence of direct information about quality or productivity, individuals would use beliefs about group identities and other signals of quality or productivity to make inferences about the unobserved quality or productivity. Following Aigner and Cain (1977) , we present a simple model of statistical discrimination to highlight how patients may discriminate doctors on the basis of gender and to draw predictions that we test in the field experiment. ...
... The denominator v ar( x s ) of γ s is the sum of σ 2 q,s and σ 2 ∈ ,s . To explain the discrimination from which highly skilled female workers suffer in the labor market, the labor economics literature typically assumes that σ 2 ∈ , f > σ 2 ∈ ,m , while maintaining that β m = β f = β and σ 2 q, f = σ 2 q,m (e.g., see Aigner and Cain 1977 ). The assumption that σ 2 ∈ , f > σ 2 ∈ ,m corresponds to the notion that signal is a less reliable indicator of quality for female workers than for male workers. ...
Article
Using a field experiment in India where patients are randomly assigned to rank among a set of physicians of the same gender but with different castes and years of experience, we show that the differences in patients’ physician choices are consistent with gender-based statistical discrimination. Labor market experience cannot easily overcome the discrimination that female doctors suffer. Further, we find that gender discrimination is greater for lower caste doctors, who typically suffer from caste discrimination. Given the increasing share of professionals from a lower caste background, our results suggest that the ‘intersectionality’ between gender and caste leads to increased gender inequality among professionals in India.
... Statistical discrimination theory pioneered by Phelps (1972) and Arrow (1973) posits that in the absence of direct information about quality or productivity, individuals would use beliefs about group identities and other signals of quality or productivity to make inferences about the unobserved quality or productivity. Following Aigner and Cain (1977) , we present a simple model of statistical discrimination to highlight how patients may discriminate doctors on the basis of gender and to draw predictions that we test in the field experiment. ...
... The denominator v ar( x s ) of γ s is the sum of σ 2 q,s and σ 2 ∈ ,s . To explain the discrimination from which highly skilled female workers suffer in the labor market, the labor economics literature typically assumes that σ 2 ∈ , f > σ 2 ∈ ,m , while maintaining that β m = β f = β and σ 2 q, f = σ 2 q,m (e.g., see Aigner and Cain 1977 ). The assumption that σ 2 ∈ , f > σ 2 ∈ ,m corresponds to the notion that signal is a less reliable indicator of quality for female workers than for male workers. ...
Article
Using a field experiment in India where patients are randomly assigned to rank among a set of physicians of the same gender but with different castes and years of experience, we show that the differences in patients’ physician choices are consistent with gender-based statistical discrimination. Labor market experience cannot easily overcome the discrimination that female doctors suffer. Further, we find that gender discrimination is greater for lower caste doctors, who typically suffer from caste discrimination. Given the increasing share of professionals from a lower caste background, our results suggest that the ‘intersectionality’ between gender and caste leads to increased gender inequality among professionals in India.
... Statistical discrimination theory pioneered by Phelps (1972) and Arrow (1973) posits that in the absence of direct information about quality or productivity, individuals would use beliefs about group identities and other signals of quality or productivity to make inferences about the unobserved quality or productivity. Following Aigner and Cain (1977) , we present a simple model of statistical discrimination to highlight how patients may discriminate doctors on the basis of gender and to draw predictions that we test in the field experiment. ...
... The denominator v ar( x s ) of γ s is the sum of σ 2 q,s and σ 2 ∈ ,s . To explain the discrimination from which highly skilled female workers suffer in the labor market, the labor economics literature typically assumes that σ 2 ∈ , f > σ 2 ∈ ,m , while maintaining that β m = β f = β and σ 2 q, f = σ 2 q,m (e.g., see Aigner and Cain 1977 ). The assumption that σ 2 ∈ , f > σ 2 ∈ ,m corresponds to the notion that signal is a less reliable indicator of quality for female workers than for male workers. ...
... Stark racial disparities emerge in studies of employment outside of prison. This literature tells us that Black applicants are viewed differently than White applicants-for example, due to perceptions of Black individuals as costlier or riskier to employ in conjunction with racial stereotypes, such as strong associations between Blackness and criminality (Aigner & Cain, 1977;Altonji & Blank, 1999;Altonji & Pierret, 2001;Doleac & Hansen, 2016;Lang & Lehmann, 2012;Moss & Tilly, 2001;Phelps, 1972). To the extent similar biases play out when correctional staff and job managers make their hiring decisions, we should anticipate disparities in OPI employment to remain even after accounting for structural factors, such as prior education and behavioral records. ...
... Prison staff are more familiar with IPs than employers are with job applicants in the free labor market (Liebling et al., 2010). This may limit the role of stereotypes (Aigner & Cain, 1977;Altonji & Blank, 1999;Altonji & Pierret, 2001). Another factor unique to the prison context is the certainty that all potential workers were once criminally involved. ...
Article
Full-text available
This study assesses whether racial and ethnic disparities exist in prison industry employment and whether seemingly race‐ and ethnicity‐neutral eligibility requirements contribute to any such disparities. We examine whether there are racial/ethnic disparities in industrial prison work, the extent to which disparities are explained by administrative policies, and the conditions under which disparities are most pronounced. Using 10 years of prison administrative data from Ohio, this study employs multilevel and mediation analyses to examine the effects of race and ethnicity on the odds of working an industrial prison job. Results suggest that Black and Hispanic incarcerated persons (IPs) are less likely to work industry jobs than White IPs. The majority of this disparity stems from program requirements; however, some disparities maintain even when accounting for requirements. Black IPs who do not meet program requirements are less likely to work than White IPs who do not meet program requirements. Racial disparities are smaller in facilities with greater racial heterogeneity among correctional staff. In our discussion, we underscore how prison policies can contribute to racially and ethnically disparate incarceration experiences. The results suggest the importance of evaluating prison and other correctional policies that utilize selection criteria that appear race neutral but are likely to be disparate in their consequences. Moreover, policies aimed at diversifying staff may contribute to more equitable prison experiences for non‐White incarcerated people, although doing so does not directly address underlying policy problems that lead to inequalities.
... Drawing on Philippon's theoretical model research [31], we consider that the ability to capture personal credit quality is also a manifestation of risk management capabilities. Following the classic paper of Aigner and Cain [2], we construct a basic model to discuss how Fintech might affect risk-taking and what impacts the synergy between Fintech and traditional finance will bring about. ...
... Column (1) shows that the Fintech index and Z-Score are significantly negative at the 5% level during 2013-2016, while Column (2) shows that there is no significant correlation between the two in 2016-2020. The results of the Chow test in Column (3) show that, compared with 2013-2016, the synergy between Fintech and traditional finance in 2016-2020 can help commercial banks reduce their risk-taking. ...
Article
Full-text available
The development of financial technology (Fintech henceforth) has brought both opportunities and challenges to commercial banks’ risk management. Here, we discuss the interactive relationship between Fintech and traditional finance. We first construct a mathematical model and deduce that the synergy between Fintech and traditional finance can more effectively help banks measure the creditworthiness of customers, thereby reducing risk-taking. In order to verify the deduced result, we then conduct an empirical analysis using panel data of Chinese commercial banks from 2011 to 2020. The empirical results support the theoretical outcome that the synergy between Fintech and traditional finance has indeed effectively reduced commercial banks’ risk-taking. Our research also shows that Fintech development significantly reduces the risk-taking of commercial banks, and this mitigation effect is more prominent for larger-scale banks and those with a higher-developed traditional finance basis. Moreover, this mitigation effect is more significant after 2016, a period marked by an increased level of Fintech development.
... Below I sketch a simple signaling model to theoretically motivate how homophilous pairings between the board and loan applicants influence lending decisions and identify factors that influence the context of this relation. 3 The model is an adaptation of a model used by Aigner and Cain (1977) to explain discrimination in employment decisions. The key feature of the model is lenders do not know an applicant's true level of risk (q) due to asymmetric information and instead observe only a noisy signal (y) of an applicant's risk based on the loan application data y = q + u. ...
... The theoretical model I developed is adapted from the employment model of Aigner and Cain (1977). Therefore, my model's implications may also be generalizable to employment decisions, ...
Article
Full-text available
In this study, I examine the role racial minorities in the boardroom can play in reducing social injustice by promoting more equal access to mortgage credit to minority households. I develop a simple theoretical model that posits directors who are racial minorities provide the credit unions they govern with a perspective that shapes lenders’ trust of minority applicants. This trust is shaped by homophily and the tendency of individuals to prefer interactions with similar individuals. Using mortgage loan data from a cross-section of credit unions in the United States from 185,446 applications, I find that credit unions where the majority of board members are minorities are less likely to reject a similarly qualified minority applicant than their counterparts. Governance by minority directors significantly reduces the effects of discrimination faced by minority applicants. The board’s effect is strongest in minority neighborhoods and where the homophily is stronger between directors and applicants.
... Statistical discrimination has two components: the first is based on beliefs (Arrow, 1973) and studies how such beliefs can influence the decisions of funding institutions. The second stream, that of Phelps (1972) and Aigner and Cain (1977), focuses on the consequences of discrimination based on specific measures of expected outcomes. ...
Article
Full-text available
This article analyses gender difference in access to microcredit in rural Burkina Faso. Using 2017 data from the National Land Management Programme, a probit model combined with Fairlie's (An extension of the Blinder-Oaxaca decomposition technique to logit and probit models. Economic Growth, Center (Yale University), discussion paper no 873., 2003) decomposition method is applied to a sample of 6646 observations. The results indicate that there is indeed gender discrimination in access to microcredit in rural areas in Burkina Faso. This average discrimination is of 0.08 against women. The decomposition indicates that wealth has the largest relative contribution estimated at 22.27% followed by income (21.03%). We suggest that in order to improve women's access to microcredit in rural areas, policies should further promote women's economic empowerment. Women's non-farm entrepreneurship (7.42%) may be a channel to be explored. The robustness of this recommendation has been tested by examining the population of rural people who undertake non-farm activities. We observed that there is no gender discrimination in access to microcredit among non-agricultural entrepreneurs in rural areas.
... However, as Oettinger (1996) points out, the economic literature is reluctant to explain the gender wage gap through differences in "tastes". For this purpose, the concept of so-called statistical discrimination and informational models of discrimination is used (Phelps, 1972;Arrow, 1974;Aigner & Cain, 1977;Milgrom & Oster, 1987). Statistical discrimination theory assumes that employers have limited (imperfect) information about the skills of potential employees. ...
Article
Full-text available
Research background: There is a broad discussion in the literature on the situation of men and women in the labour market, especially about the differences in their remuneration. Due to the fact that females constitute a slightly different group of employees, certain factors have different impacts on the level of their remuneration in comparison to male employees. Hence, the question arises which factors cause these differences and how large the dissimilarities are. Purpose of the article: The aim of the presented study is to diagnose and evaluate differences in the impact of designated determinants on the level of monthly wages of women and men in selected European Union member states. The novelty of our approach consists in both comparison of the intensity of influence examined factors to men?s and women?s earnings, and a global approach to the remuneration of male and female employees. Methods: Due to the nature of the dependent variable (remuneration decile, which is a variable measured on an ordinal scale), the ordered logit model is applied in the analysis. The data comes from the Eurostat?s Labour Force Survey. Findings & value added: Presented results indicate that many factors have significantly different intensity of impact on the level of men and women wages. However, significant differences between parameters estimated for both genders are visible for the group of family variables the most often, then for variables describing the condition of work, the human capital variables, and characteristics of the workplace. This paper adds to the empirical literature a new approach to measure the intensity of factors influencing men and women wages. In addition, our investigation is a cross-country analysis.
... Statistical discrimination (Phelps, 1972;Arrow, 1973;Aigner and Cain, 1977) posits that differential treatment by race or gender can emerge without a distaste by an individual. 18 ...
... Second, our study contributes to the growing literature exploring the impact of motherhood and family policies on gender inequality in the labor market (Ruhm, 1998;Maurer-Fazio et al., 2011;Olivetti and Petrongolo, 2017;Kleven et al., 2019aKleven et al., ,b, 2021Chen et al., 2021;Cortés and Pan, forthcoming). The impact of parenthood on the labor market outcomes of women relative to their male counterparts is known as child penalties, which contribute to the gen-2Statistical discrimination occurs when employers make hiring decisions based on a calculation of a group's average or other statistical characteristics, rather than the applicant's actual quality (Aigner and Cain, 1977). der gap at work in many countries (Kleven et al., 2019a). ...
Article
Full-text available
We conduct a large-scale field experiment in China to investigate the effect of being of childbearing age on gender discrimination in the labor market. We send 35,713 fictitious resumes to real job postings on a major Chinese online recruitment platform for jobs in four leading cities, Beĳing, Shanghai, Guangzhou, and Shenzhen, which vary in the length of maternity leave. We send applications for positions advertised in the male-dominated field of information technology (IT), the female-dominated field of accounting (ACC), and the mixed-gender field of human resources (HR). We systematically vary the age and gender of the job applicants and record callbacks for interviews. To accurately mimic the job application process in the Chinese labor market, we do not disclose the applicants’ family status. We find that women of childbearing age are subject to discrimination in the field of IT, a problem that also exists in HR and ACC, particularly in Beĳing and Shanghai. There is no obvious discrimination against women of childbearing age in Guangzhou or Shenzhen, where maternity leave is longer. In the aggregate, the evidence indicates that women of childbearing age face statistical discrimination that prevents them from obtaining equal employment opportunities.
... For example, an applicant might present as having high job commitment, but there is still a risk that her apparent commitment is inaccurate or unstable over time. Importantly, past bias research in psychology and management has overlooked the riskiness of an applicant's suitability for a job as a potential contributor to personnel selection, despite the widespread role of risk calculations in decision-making processes in economics (e.g., Aigner & Cain, 1977;Arrow, 1973;Phelps, 1972;Kahneman & Tversky, 1979). ...
Article
Full-text available
Research grounded in gender role theories has shown that women face numerous employment disadvantages relative to men, with mothers often facing the greatest obstacles. We extend this literature by proposing that motherhood is not a necessary condition for women to face motherhood penalties. Instead, managers’ expectations that an applicant will have a child in the near future (i.e., “maybe baby” expectations) increases their perceptions of risk associated with employing childfree, childbearing-aged women–but not men. Investigating the intersection of gender and age, and integrating economic theories of discrimination, we conceptualize hiring as a risk assessment process, proposing that managers’ risk perceptions drive more precarious employment conditions for this group of women. Results from a field study with early career employees (Study 1) and a randomized experiment with hiring managers (Study 2) support our predictions across attitudinal (e.g., desire to offer a temporary job contract; Study 2) and objective indicators (e.g., having a temporary job contract; Study 1); female applicants can also mitigate this “maybe baby” risk by signaling a lack of interest in having children or by emphasizing their commitment and work ethic (Study 2). Our findings suggest that the perceived risks of parenthood can be hazardous for child-bearing-aged, childfree working women who simply may become parents (vs. men and mothers; vs. childfree women who are significantly younger or older than the average age of first childbearing in the local context).
... Following Aigner and Cain (1977), our model of statistical discrimination describes how rational voters use perception of competence and incumbency status to form beliefs about the candidate's quality. Based on a political candidate's physical appearance, a voter perceives the competence of the candidate, given whether the candidate is the incumbent or a challenger (i.e., incumbency status). ...
Technical Report
Full-text available
Prior studies have documented incumbency advantage for electoral candidates. A seemingly unrelated strand of literature has shown that perceived competence of candidates, as inferred from their physical appearance, also predicts their electoral success. We show that the theory of statistical discrimination helps understand the relationship between a political candidate's vote share, physical appearance and incumbency status. Voters possess actual information about the performance of incumbents, but not the performance of challengers. Competence of candidates, as inferred from their physical appearance, is therefore expected to have a relatively stronger influence on voters' decisions to vote for challengers. Data from Australian state elections support this prediction. The findings highlight the usefulness of statistical discrimination theories in explaining voting decisions.
... There are several reasons why employers might use information about applicants' history of mental health problems to make hiring decisions. Theories of statistical discrimination (Aigner and Cain, 1977;Arrow, 1973;Phelps, 1972) assume that employers are risk-averse and attempt to hire the most 'productive' candidate. Since they have limited information about individual applicants, they make decisions based on perceived group characteristics. ...
Article
Full-text available
Mental health problems are associated with poor labour market outcomes. Based on data from a field experiment, this article investigates the extent to which hiring discrimination limits the job opportunities of young applicants who disclose a history of mental health problems. From September 2019 to December 2020, 1398 job applications were sent in pairs to 699 employers with job openings in a broad selection of occupations in the Norwegian labour market. The applicants were equally qualified except that, in each pair, one applicant informed about mental health problems as an explanation for a past employment break. The results show that applicants who disclose mental health problems are discriminated against in hiring processes. Applicants with mental health problems have about 27% lower probability of receiving an invitation to a job interview and about 22% lower probability of receiving any positive employer response. These results do not seem to have been driven by the COVID-19 crisis that unfolded during the course of the study. As such, the study provides suggestive evidence that uncertain economic times might not necessarily increase the level of discrimination against applicants with mental health problems.
... There is an increasing body of literature that examines the role of race and gender in gig economy transactions (e.g., De Stefano, 2016;Edelman et al., 2017;Rogers, 2015). Drawing on theories of discrimination in traditional hiring (e.g., Aigner & Cain, 1977), several of these studies have uncovered evidence of discrimination in online hiring, thus raising concerns about labor market segregation and inequities in career opportunities as the gig economy expands (Chan & Wan, 2018;Galperin, 2019;Leung, 2018). ...
... The discriminator is more willing to lose important sources of income rather than sharing its position with communities considered "hostile". On the other hand, the lack of information about the "others'"economic and social features -perhaps a potential purchaser or a tenantleads to another kind of discrimination (Phelps, 1972;Aigner and Cain, 1977). It is mainly about an informative deficit, therefore it is not generated by ideology. ...
Article
Full-text available
New ethnicities and new cultures are taking up residence in our cities: for years the flow of foreign people has been compensating the decreasing rates of natural growth and population ageing. This trend comes along with several consequences in the real estate market, and the scientific literature has shown up events in common with many countries. Spatial segregation, as a consequence of real estate dynamics, has to be critically evaluated: it is by no means a neutral mechanism, yet it contributes in no small measure to the spatial segregation of the smaller communities. Nuove etnie e nuove culture abitano le nostre città: da anni il flusso di persone straniere compensa il decrescente saldo naturale e l’invecchiamento della popolazione. Ciò non è senza conseguenze sul mercato immobiliare e la letteratura scientifica mette in luce fenomeni comuni a molti paesi. La segregazione spaziale effetto di dinamiche di mercato va considerata criticamente: quest’ultimo non è dispositivo neutrale, ma concorre attivamente alla segregazione spaziale delle comunità.
... While only interview callback rates are measured, callback rates are found to empirically map directly to job offer rates, at least for in-person audit studies in the US (Mincy 1993). Whereas employer perceptions may be mistaken, incorrect beliefs are unlikely to persist in labor markets over time in a competitive market (Aigner and Cain 2013). Especially if decisions made by employers are based on previous experience working with similar workers, employer perceptions can reflect true productivity differences between workers. ...
Thesis
International migration and return are important channels through which individuals from migrant-sending countries stand to benefit from the world economy. This dissertation examines questions concerning the economics of why people go abroad and why they might return, and then looks at the consequences of such decisions. The first chapter investigates whether domestic employers value the foreign work experience of migrants when they return home from abroad. I conduct an audit study in the Philippines, sending 8,000 fictitious resumes in response to job postings across multiple occupations. Resumes are randomly assigned varying lengths of foreign work experience, among other things. Workers with foreign experience receive 12 percent fewer callbacks than non-migrants, with callback rates even lower for those who have spent a longer time abroad. I find that location-specific human capital is important to employers, and the value of this human capital deteriorates as a worker spends time away from the domestic economy. The second chapter discusses why migrants might decide to return to their home countries in the first place. I utilize exogenous exchange rate shocks arising from the 1997 Asian Financial Crisis to distinguish between return motivations of Australian immigrants. A depreciation in a migrant’s home country currency leads to a reduced likelihood of return. The results favor a life-cycle explanation for migrant behavior and reject the theory that migrants are target earners who seek to invest upon return. The third chapter, joint with Caroline Theoharides, considers “brain drain”, which is a chief concern for many developing countries. The “brain drain” view holds that migration is usually responsible for the scarcity of healthcare workers in a country, contributing to poor health outcomes. This chapter challenges prevailing wisdom by demonstrating how international migration may, in fact, lead to human capital formation. We focus on nurse migration from the Philippines. Using microdata on migrant departures, we find that the large nurse outmigration during the 2000s led to large subsequent increases in enrollments for nursing degrees in many provinces. The increase in human capital was large enough that, there was no reduction but an increased supply of nurses remaining in the country.
... A higher share of females in the managerial ranks may challenge the association between leadership and masculinity (Koenig et al., 2011), potentially paving the way for career advancement for women. Alternatively, women may be better placed than men to judge accurately the work performance of female colleagues and reward them accordingly (Aigner and Cain, 1977). One might therefore expect the gender wage gap to diminish as the share female in managerial positions rises. ...
Article
Full-text available
Using linked employer-employee data for Britain, we find a robust association between the share of female managers in the workplace and the size of the gender wage gap. In workplace fixed-effects estimates, the gap is eradicated when more than 60% of workplace managers are women, a scenario that obtains in around one fifth of all workplaces. The association between the share of female managers and the gender wage gap is more pronounced when workplace managers set pay at the workplace, and where employees are paid for performance. These findings are consistent with the proposition that women are more likely to be paid equitably when managers have discretion in the way they set pay or reward performance and those managers are women. They suggest that a stronger presence of women in managerial positions can help tackle the gender wage gap.
... Based on the notion of statistical discrimination, the expectation is that employers favor or disfavor individual job applicants based on group characteristics, e.g., known or expected lower average productivity (Phelps 1972) among women as a result of their greater family responsibilities, and longer and repeated expected periods of part-time work and work absence. Another version of statistical discrimination theory refers to group differences in variance rather than mean differences (Aigner and Cain 1977), but it is the latter that are (tacitly) claimed to lie behind employer discrimination against mothers (to be). ...
Article
Full-text available
A common assumption in comparative family policy studies is that employers statistically discriminate against women in countries with dual-earner family policy models. The empirical evidence cited in support of this assumption has exclusively been observational data, which should not be relied on to identify employer discrimination. In contrast, we investigate whether employers discriminate against women in Sweden—frequently viewed as epitomizing the dual-earner family policy model—using field experiment data. We find no evidence supporting the notion that Swedish employers statistically discriminate against women.
... The first is racial bias, in which judges discriminate against Black defendants at the margin of pretrial release due to either racial preferences (Becker, 1957) or some form of inaccurate racial stereotyping (Bordalo et al., 2016). The second is statistical discrimination, in which judges act on accurate risk predictions but discriminate due to racial differences in average risk or the precision of received risk signals (Phelps, 1972;Arrow, 1973;Aigner and Cain, 1977)." ...
Preprint
Full-text available
The Benchmark Test identifies an aggregate notion of discrimination (disparate treatment conditional on observables), while the Outcome Test identifies an aggregate notion of prejudice (biased beliefs and taste-based discrimination). However, no test for an aggregate notion of statistical discrimination has been proposed in the literature. Since prejudice can be thought of as discrimination net of statistical discrimination, this note explores whether aggregate notions of statistical discrimination can be identified by combining both tests. We discuss the difficulties in doing so and argue, through the provision of an identification result that relies on highly restrictive (presumably implausible) assumptions, that the answer is most likely negative.
... Statistical discrimination (Phelps, 1972;Aigner and Cain, 1977) occurs when the employer, which is not able to assess the true productivity of the worker, uses the statistical information he has on the group the worker belongs to (for example gender) to estimate the workers' productivity. This last type of discrimination can lead to occupational segregation as employers of already over-represented occupations will have better information on the over-represented gender and therefore can show a preference for it. ...
Preprint
Full-text available
The systematic differences of gender representation across occupations, gender-based occupational segregation, has been suggested as one of the most important determinants of the still existing gender wage gap. Despite some signs of a decreasing trend, there is evidence that occupational gendered segregation is persistent even though gender differences in human capital variables have been disappearing. Using an agent-based model we provide a framework that introduces discriminatory behavior based on labour market theories of discrimination where workers and firms can exhibit gendered preferences. The introduction of discriminatory behavior transforms the otherwise random dynamics of occupational choice into a persistent gender-based occupational segregation consistent with empirical evidence.
... 3 Therefore, immigrants are only hired at a wage lower than the wages of equally productive natives. Subsequently, economists have suggested statistical discrimination models that are based on imperfect information; these models posit the abilities of individual immigrants as an alternative explanation for wage inequality (Aigner & Cain, 1977;Dickinson & Oaxaca, 2009;Phelps, 1972). This class of models posit that in the absence of reliable information on individual immigrants, native employers often rely on information regarding the average characteristics of the group to which immigrants belong. ...
Article
Full-text available
In this study on Mainland Chinese women in Taiwan, I examined the effect of immigration-related disruption on the assimilation of these women into the Taiwanese labour market during 2005–2015. Accordingly, I used a unique dataset obtained by linking three administrative registers to measure the assimilation process. In addition, I employed the nearest-neighbour matching estimator to assess heterogeneous effects on each Mainland Chinese woman. The results indicated narrowing immigrant–native gaps in labour supply (full-time employment rates) and real monthly insured wages for their first traceable job in favour of Mainland Chinese women. In general, I found that Mainland Chinese women assimilated into the Taiwanese workforce at levels comparable with those of Taiwanese women.
... Information about individuals is imperfect, and individuals are often judged as members of social groups rather than as individuals (Phelps, 1972;Arrow, 1973;Aigner and Cain, 1977). When it is the case, the image of the group is substituted for the less available individual information and determines the members' economic success. ...
Article
We study the production of social image in social groups. In our experiment, we recruit pairs of real-life friends and study whether rule breaking in the form of misreporting decreases when misreporting may have negative spillovers on the image of the friend. We find that participants hurt their friends’ social image by misreporting: external observers update their beliefs and rightfully expect that a participant whose friend misreported is likely to misreport himself. However, participants misreport as often when their behavior can hurt their friends’ image as when it cannot, even though hurting their friends’ image reduces their own monetary gains. Our interpretation is that they underestimate the impact of their behavior on external observers’ beliefs about their friends. Our results show that, even in our case where group membership is salient, groups might have difficulties building a good image.
... Taste-based discrimination (Becker, 1957) argues that employers avoid hiring immigrants due to personal preferences or social factors unrelated to productivity (Busetta, Campolo, & Panarello, 2018;Carlsson & Rooth, 2012;Guryan & Charles, 2013). However, when it comes to the signaling value of educational credentials, immigrants' disadvantaged economic outcomes are usually explained as a result of employers relying on statistical discrimination (Aigner & Cain, 1977;Phelps, 1972). From this perspective, employers are more uncertain about interpreting signals of an immigrant's potential productivity from observable characteristics such as credentials and are not willing to invest in acquiring additional information. ...
Article
We examine the reasons behind a wage penalty that is applied to foreign-educated immigrants—a phenomenon identified across countries and levels of educational attainment. The lower wages employers offer to high-skilled immigrants are usually attributed to the foreignness of their educational credentials. We argue instead that their disadvantaged economic outcomes are related to the educational prestige of their universities and the cultural distance of their places of education. Employers may interpret signals of ability from the educational prestige associated with credentials and may view immigrants as lacking relevant cultural competence if they perceive the place of education as culturally distant. Using data from the U.S. H-1B visa program, we show that employers do not discount wage offers to immigrants with foreign degrees per se. Instead, we find that they offer lower wages to immigrants with educational credentials from less prestigious universities and more culturally distant places of education.
... 6 Aigner and Cain (1977)). Likewise, there is extensive work in finance that studies changes in lenders' ability to use public credit report data, their own private information, or other lenders' assessment of default risk. ...
Preprint
Full-text available
We show that lenders face more uncertainty when assessing default risk of historically under-served groups in US credit markets and that this information disparity is a quantitatively important driver of inefficient and unequal credit market outcomes. We first document that widely used credit scores are statistically noisier indicators of default risk for historically under-served groups. This noise emerges primarily through the explanatory power of the underlying credit report data (e.g., thin credit files), not through issues with model fit (e.g., the inability to include protected class in the scoring model). Estimating a structural model of lending with heterogeneity in information, we quantify the gains from addressing these information disparities for the US mortgage market. We find that equalizing the precision of credit scores can reduce disparities in approval rates and in credit misallocation for disadvantaged groups by approximately half.
... First, judges may use I i to predict (or interpret signals of) variables they do not observe that affect Y * i if distributions vary with I i . This is defined as statistical discrimination (Aigner and Cain, 1977) and directly affects the conditional outcome probabilities, p(I i , Z i ), and the bias, b j(i) (I i , Z i ), if these parameters impact how priors are updated. Second, if statistical discrimination is inaccurate and the degree of inaccuracy varies with I i , this induces a differential prediction bias to the decision. ...
Preprint
Full-text available
This paper tests for discrimination against immigrant defendants in the criminal justice system in Chile using a decade of nationwide administrative records on pretrial detentions. Observational benchmark regressions show that immigrant defendants are 8.6 percentage points less likely to be released pretrial relative to Chilean defendants with similar proxies for pretrial misconduct potential. Diagnostics for omitted variable bias -- including a novel test to assess the quality of the proxy vector based on comparisons of pretrial misconduct rates among released defendants -- suggest that the discrimination estimates are not driven by omitted variable bias and that, if anything, failing to fully account for differences in misconduct potential leads to an underestimation of discrimination. Our estimates suggest that discrimination stems from an informational problem because judges do not observe criminal records in origin countries, with stereotypes and taste-based discrimination playing a role in the problem's resolution. We find that discrimination is especially large for drug offenses and that discrimination increased after a recent immigration wave.
... Both the behavior of the applicant and discrimination by the employer may contribute to health-related selection. Studies have demonstrated that employers view signs of poor health, such as visible symptoms or gaps in employment history, to imply risk (e.g., sickness absences, poor productivity) [21,22], or associate them with other unwanted applicant characteristics (e.g., poor emotional or social skills) [16,17,[23][24][25]. However, employers are unlikely to possess full knowledge of an applicant's health. ...
Article
Full-text available
Background Successful transitions from unemployment to employment are an important concern, yet little is known about health-related selection into employment. We assessed the association of various physical and psychiatric conditions with finding employment, and employment stability. Methods Using total population register data, we followed Finnish residents aged 30–60 with an unemployment spell during 2009–2018 ( n = 814,085) for two years from the onset of unemployment. We predicted any, stable, and unstable employment by health status using Cox proportional hazards models. The data on specialized health care and prescription reimbursement were used to identify any alcohol-related conditions and poisonings, psychiatric conditions and self-harm, injuries, and physical conditions. We further separated physical conditions into cancer, diabetes, heart disease, and neurological conditions, and psychiatric conditions into depression, anxiety disorders and substance use disorders. Results The likelihood of any employment was lower among those who had any of the assessed health conditions. It was lowest among those with alcohol-related or psychiatric conditions with an age-adjusted hazard ratio of 0.45 (95% confidence interval 0.44, 0.46) among men and 0.39 (0.38, 0.41) among women for alcohol-related and 0.64 (0.63, 0.65) and 0.66 (0.65, 0.67) for psychiatric conditions, respectively. These results were not driven by differences in socioeconomic characteristics or comorbidities. All the included conditions were detrimental to both stable and unstable employment, however alcohol-related and psychiatric conditions were more harmful for stable than for unstable employment. Conclusions The prospects of the unemployed finding employment are reduced by poor health, particularly alcohol-related and psychiatric conditions. These two conditions may also lead to unstable career trajectories. The selection process contributes to the health differentials between employed and unemployed people. Unemployed people with health problems may therefore need additional support to improve their chances of employment.
... Thus it is not clear cut on exactly how employers are reacting to the discrimination shock. One possible explanation for this heterogeneity in effects across contract types is that employers become more risk averse about hiring minorities (Aigner and Cain, 1977) due to the shock and thus propose short-term contracts instead. Yet this is only speculation and does not speak to the change in composition of the types of minorities employers search for, nor if these minorities accept the offer from these employers. ...
Thesis
In the first chapter I study how the job performance of minorities changes depending on whether they work with managers who are more or less biased against their type. I show that when minorities work with more biased managers they perform significantly worse compared to majority workers on a range of performance indicators. Yet minority performance is higher when working with non-biased managers. We argue that this is evidence of a self-fulfilling prophecy whereby biased managers make minorities less productive and this generates statistical discrimination in the firm’s hiring policy. The second chapter explores how shocking the value of a vacancy through offering free recruiting services to firms affects their demand for labor. Offering free recruitment services leads to large increases in vacancy postings. Furthermore, this translates into significant increases in hires in permanent contracts. These results suggest that active labor market policies directed at generating firm labor demand may have substantial added value in the labor market. In the final chapter it is shown that the the Charlie Hebdo attacks significantly reduced Muslim minority job search effort. Frims also reduced their search for minority candidates, but only for the permanent contracts. This drop is partially offset by an increase in counselor matching effort made for minorities after the shock, but only in areas with low latent levels of discrimination, as measured by the local vote share for the Front National.
... Classic economic approaches to explaining ethnic discrimination in hiring distinguish between statistical discrimination (e.g., Aigner & Cain, 1977;Phelps, 1972) and taste-based discrimination (Becker, 1957). Statistical discrimination theory assumes that profit-maximising recruiters discriminate against non-natives due to imperfect information about their worker qualities (i.e., productivity). ...
Article
A bulk of experimental research has pointed to ethnic hiring discrimination as a key driver of social inequalities in multi-ethnic labour markets. However, the role of recruiter nationality as possible moderators of ethnic hiring discrimination has been widely neglected. Against this background, this study examines the role of recruiter nationality in hiring discrimination involving foreign applicants. I used data from a recent factorial survey experiment conducted with real recruiters in Luxembourg ( N= 677 from 113 recruiters). Respondents rated six experimentally manipulated profiles of fictitious applicants for jobs in different occupational fields. Luxembourg is a relevant case due to its multi-ethnic workforce at various levels of the vertical strata of labour market positions, which allows to test the role of ethnic homophily in hiring decisions. The results suggested that foreign recruiters discriminate less against foreign applicants than native recruiters do. Furthermore, the effect of being a foreign applicant was more negative if the applicant’s nationality matched the recruiter’s nationality. However, none of the observed differences in effects were statistically significant. Elucidating how both native and non-native recruiters make hiring decisions based on applicants’ nationality contributes to a better understanding of how social inequalities emerge and intensify in multi-ethnic labour markets.
... Statistical discrimination (Phelps, 1972;Arrow, 1973;Aigner and Cain, 1977) posits that differential treatment by race or gender can emerge without a distaste by an individual. 18 ...
... Έπειτα από τις δημοσιεύσεις των (Phelps, 1972) και , αρκετές οικονομικές μελέτες που εξετάζουν τις ανισότητες έχουν βασιστεί στην στατιστική θεωρία διακρίσεων. Τις ακαδημαϊκές έρευνες αυτές συμπλήρωσαν με την έρευνά τους και οι (Aigner & Cain, 1977), πάνω στις οποίες βασίζεται η παρακάτω ανάλυση της στατιστικής θεωρίας διακρίσεων. Η βασική υπόθεση στην οποία βασίζεται η θεωρία αυτή, είναι πως οι επιχειρήσεις διαθέτουν λίγη πληροφόρηση σχετικά με τις δεξιότητες και την παραγωγικότητα των υποψηφίων για θέσεις εργασίας. ...
Thesis
Full-text available
The main purpose of this paper is to examine possible discrimination that individuals may suffer in the labor market because of their sexual orientation. This analysis aims to find factors that trigger the creation of the wage gap between people of the opposite sex and with different sexual preferences. The detailed observation of discrimination in the labor market requires the use of both theoretical and econometric theories and methods.
... Becker concludes that employers would employ individuals of the discriminated group with same level of productivity as others only if the employers can pay a lower wage to them compared to others. The other strand of theoretical literature looks at statistical discrimination as a result of signal extraction problem (Arrow and et al 1973;Phelps 1972;Aigner and Cain 1977). ...
Article
Full-text available
In this paper, we attempt to study the gender pay gap in the case of Indian software developers. This is the only study in our knowledge that analyses gender pay gap in this important and dynamic sector in India. Previous studies on wage gap suffer from the problem of bias due to differences in the supports of the empirical distributions of individual characteristics for females and males. Lack of common support also leads to considerable model dependence of results. We overcome the limitations of previous gender wage gap studies by using entropy balancing weights in our recentered influence function-based wage decompositions. We find the existence of a gender pay gap after controlling for other covariates. We also find the presence of glass ceiling phenomenon for software developers in India.
Article
Full-text available
The paper analyzes earnings differences between rural–urban migrants and urban workers in China to examine the proposition that discrimination against migrant workers is tending toward zero. Blinder‐Oaxaca decomposition is applied using the 2013 China Household Income Project (CHIP) database to find out the extent of discrimination against migrant workers. The results show the existence of wage difference between the two groups, however, most of the wage difference can be explained by observable characteristics implying that discrimination is on the wane. Assuming the urban worker's wage as a nondiscriminatory wage, the benchmark model shows that 82.9% of the wage difference can be explained by individual endowments. Furthermore, including occupation and industry variables the extended model can explain 91.8% leaving only 8.2% as the source of wage discrimination. Using the combined and Cotton (1988) method, the explanatory part is still very high (89.2% and 86.3%) although a bit lower than the previous method. To further support the proposition, we also do the same exercise using 2007 CHIP data that shows a higher percentage of discrimination (36.2%), implying that discrimination decreased in 2013 compared to 2007. A cautious review of similar literature also supports the view. In short, wage discrimination against migrant workers is tending toward zero in recent urban China.
Article
Article
We evaluate the causes of the wage gap at the intersection of race, ethnicity and gender over time in the United States. We analyse the wage gaps for women of colour along three dimensions: relative to White women, relative to men of their respective race/ethnicity, and relative to White men. Using the American Community Survey, we replicate earlier findings based on the Current Population Survey data which show that, on average, Black women face an unexplained wage gap relative to White men that goes beyond the simple addition of the separate unexplained gender and racial wage gaps. This can be seen persistently between 1980 and 2019, and we find it is true across the entire wage distribution but especially notable at higher centiles. From 1990 through 2019, Black and Hispanic women saw stalled progress, while White women continued to make steady progress closing the wage gap relative to White men.
Chapter
Der vorliegende Beitrag beschäftigt sich mit der aktuellen Forschung zu ethnischer Diskriminierung auf dem Arbeitsmarkt. Dabei liegt der Fokus auf sozialpsychologischen und sozialwissenschaftlichen Erklärungsansätzen und auf den Ergebnissen sogenannter Korrespondenztests, die sich zunehmender Beliebtheit erfreuen und robuste Hinweise auf ethnische Diskriminierung bei der Auswahl von BewerberInnen geben.
Article
Full-text available
SDGs 2030 have at least two goals that strongly converged to gender equality of pay, gender equality (Goal 5), and decent work and economic growth (Goal 8). Marital status argued to have significant contribution in gender inequality of pay and the root of female wage penalty. The argument associated with traditional family division of labour, where men responsible for family breadwinner and women for family caregiver. We examine this argument in Indonesia experiences with additional feature of multiple breadwinner models in the family. Utilizing Indonesian National Labor Force Survey of 2015 of more than 150.000 workers, our twofold regression compatible Oaxaca-Blinder decomposition analysis confirmed the existence of statistically significant gender wage gap even when controlling for the human capital investment, family division of labour, institutional instrument and discrimination. Our findings suggest that married women tend to take a double roles than married men, being family breadwinner as well as family caregiver. In that double roles model, wage rate are highest among working women and marriage wage penalty died out. Unfortunately gender wage gap and wage discrimination persisted. While the role of minimum wage as current nominal institutional labor market instrument still trivial, other factors must also be addressed toward gender equality of pay and more efficient labour market.
Article
Does cultural dissimilarity explain discrimination against immigrant-origin minorities in the labour market? I conducted a factorial field experiment (N = 1350) to explore how explicit group cues trigger differential treatment and whether individuating information that counters cultural-based stereotypical representations mitigate discrimination. Employers were randomly assigned a job application with a putative female ethnic majority or immigrant-origin minority alias and CV photographs portraying the minority candidate with or without a headscarf—perhaps the quintessential marker of Muslim identity. Moreover, half the job applications conveyed information intended to reduce cultural distance by indicating a liberal lifestyle and civic participation. The results demonstrate that immigrant-origin women are significantly less likely to receive an invitation to a job interview, especially if they also wear a headscarf. Contrary to expectations, the differential treatment is not moderated by the individuating information in the applications. This indicates that the differential treatment is persistent and also targets immigrant-origin minorities who have acquired soft skills and signals cultural proximity.
Article
Using linked employer‐employee data for Britain, we find a robust association between the share of female managers in the workplace and the size of the gender wage gap. In workplace fixed‐effects estimates, the gap is eradicated when more than 60% of workplace managers are women, a scenario that obtains in around one fifth of all workplaces. The association between the share of female managers and the gender wage gap is more pronounced when workplace managers set pay at the workplace, and where employees are paid for performance. These findings are consistent with the proposition that women are more likely to be paid equitably when managers have discretion in the way they set pay or reward performance and those managers are women. They suggest that a stronger presence of women in managerial positions can help tackle the gender wage gap.
Article
Catastrophic events affect many people simultaneously. We exploit earthquake claim characteristics to test for racial discrimination in the adjudication of insurance claims. Using data from the Oklahoma Department of Insurance, the US Geological Survey, and the US census, we study eight earthquakes between 2010 and 2016 that were linked to oil and gas drilling activities. We test whether claim resolutions differ among zip‐code areas with different racial compositions, all else equal. We find evidence that claims from areas with higher percentages of Black population were less likely to result in payment, and when those claims did get paid, payments were lower in those areas. We further investigate the mechanisms through which such discrimination may exist. We do not find evidence that the percentages of Black, Native, or Asian population in an area are associated with the filing of marginal claims. We do find that areas with higher percentages of Hispanic population file fewer marginal claims.
Article
SDGs 2030 have at least two goals that strongly converged to gender equality of pay, gender equality (Goal 5), and decent work and economic growth (Goal 8). Marital status argued to have significant contribution in gender inequality of pay and the root of female wage penalty. The argument associated with traditional family division of labour, where men responsible for family breadwinner and women for family caregiver. We examine this argument in Indonesia experiences with additional feature of multiple breadwinner models in the family. Utilizing Indonesian National Labour Force Survey of 2015 of more than 150.000 workers, our twofold regression compatible Oaxaca-Blinder decomposition analysis confirmed the existence of statistically significant gender wage gap even when controlling for the human capital investment, family division of labour, institutional instrument and discrimination. Our findings suggest that married women tend to take a double roles than married men, being family breadwinner as well as family caregiver. In that double roles model, wage rate are highest among working women and marriage wage penalty died out. Unfortunately gender wage gap and wage discrimination persisted. While the role of minimum wage as current nominal institutional labour market instrument still trivial, other factors must also be addressed toward gender equality of pay and more efficient labour market.
ResearchGate has not been able to resolve any references for this publication.