25% of racial wage gap is due to prejudice: University of Chicago Professors find

Kerwin Kofi Charles and Jonathan Guryan, "Prejudice and Wages: An Empirical Assessment of Becker's The Economics of Discrimination," published in the Journal of Political Economy, Vol. 116, No. 5, October, 2008, pp. 773-809.

Abstract: We test the predictions from Becker's (1957) seminal work on employer prejudice and find that relative black wages (vary negatively with the prejudice of the marginal "white" in a state, (b) vary negatively with the prejudice in the lower tail of the prejudice distribution but are unaffected by the prejudice of the most prejudiced persons in a state, and (c) very negatively with the fraction of a state that is black. Our estimates suggest that one-quarter of the racial wage gap is due to prejudice, with nontrivial consequences for black lifetime earnings.

Race is off-limits factor in wrongful death case

From: Federal Judge Blasts Use of Statistics on Race to Set Damages in Ferry Crash, Anthony Lin, New York Law Journal, October 15, 2008


A Brooklyn federal judge has slammed the use of statistics showing racial differences in life expectancy to determine damages for a catastrophically injured black man.James McMillan was rendered a quadriplegic in the 2003 crash of the New York City-operated Staten Island Ferry. Last month, Eastern District of New York Judge Jack B. Weinstein awarded McMillan damages of $18.3 million.

The city had sought to limit McMillan's damages on a number of grounds, arguing that his past criminal records as much as his race indicated a shorter life expectancy. But Weinstein indicated during trial he would issue a written decision further explaining his reasoning on the race issue.

Issuing that decision Tuesday, Weinstein said the consideration of statistical differences in life expectancy among races in determining damages would be discriminatory and unconstitutional. He noted that a wrongheaded insistence on immutable racial differences had been behind the U.S. Supreme Court's infamous decision in Plessy v. Ferguson, 163 U.S. 537 (1896), which upheld racial segregation under the doctrine of "separate but equal.""Statistical reliance on 'race' leads to such questions as whether Plessy would have been today categorized as 'African American' for life expectancy purposes,"

Weinstein wrote. "In a more recent example, 'racially' characterizing for statistical purposes in a negligence lawsuit the current Democrat Party presidential candidate, born of a 'White' American mother and an 'African' citizen of Kenya, would be considered absurd by most Americans."

The judge also said racial statistics should be rejected on scientific grounds, and he approvingly cited a number of well-known anthropologists who regard race as a social construct rather than a biological fact."Reliance on 'race'-based statistics in estimating life expectancy of individuals for purposes of calculating damages is not scientifically acceptable in our current heterogeneous population,"

Weinstein wrote in McMillan v. City of New York, 03 civ. 6049.Though the judge acknowledged a documented mortality gap between blacks and whites, he said the gap likely owed much to socioeconomic factors masked as "race."

He noted some studies indicating that blacks and whites of equivalent socioeconomic status enjoyed similar longevity.Weinstein said that courts had increasingly moved toward race- and gender-neutral calculations of damages, and observed that racial differences were ignored by Special Master Kenneth R. Feinberg in his administration of the federal September 11th Victim Compensation Fund.The Corporation Counsel's Office declined to comment on Weinstein's decision.

Relative v. Absolute Risk employment discrimination cases

bbThe recent foreclosures in the housing market provides a very interesting example of relative versus absolute risk. The AP recently reported that 'Foreclosure filings up 121 percent from last year'. Dig a little further and it is clear that the AP is discussing relative and not absolute risk.

The story goes on and says that this year 1 out of every 171 homes is being foreclosed. This foreclosure risk number equates to approximately 0.005 or 0.5% of homes. Last year approximately 1 in 342 homes were being foreclosed or about 0.0025 or 0.25%.

Relatively speaking there was a big spike but in absolute terms, not many homes were in foreclosure.


In employment cases where statistical analysis comes into play it is also important to keep it in mind the difference between relative and absolute risk.

For instance in a wrongful termination case where there are allegations that the defendant/employer discriminated against older workers, statistical experts will typical evaluate the chance probability that a given employer would have been terminated had the employer been using a age neutral employee selection process. If the chance probability is small then it is viewed as suggestive of a discriminatory selection process.

The chance probability in an employment case that is typically a relative risk.

In other words, the chance probability measures the likelihood that one group, i.e. older workers, would have been terminated versus the likelihood that another group of workers, i.e. younger workers, would have been terminated if all factors were equal.

Accordingly, just like the case of foreclosures above, the actual number of individuals 'at risk' of being selected at any given time may be quite small. For instance, the relative risk may say that the older workers are 75% less likely to be promoted as younger workers. However in practice this relative risk of promotion may translate into an actual (absolute) gap of only 1 or 2 persons (out of many hundreds) that arguably should have been promoted. In this type of situation, a claim of class wide discrimination against older workers would be a little dubiuos to say the least.

Tall people are smarter than short people

New economic study confirms what Randy Newman are ready knew:

Tall people have higher self-esteem, social dominance, and now a new economic study finds that tall people have higher cognitive ability. Interesting work. Full CIte:

Journal of Political Economy, 2008, vol. 116, no. 3]
© 2008 by The University of Chicago. All rights reserved.

Stature and Status: Height, Ability, and Labor Market Outcomes
Anne Case and
Christina Paxson
Princeton University

Summary: The well-known association between height and earnings is often thought to reflect factors such as self-esteem, social dominance, and discrimination. We offer a simpler explanation: height is positively associated with cognitive ability, which is rewarded in the labor market. Using data from the United States and the United Kingdom, we show that taller children have higher average cognitive test scores and that these test scores explain a large portion of the height premium in earnings. Children who have higher test scores also experience earlier adolescent growth spurts, so that height in adolescence serves as a marker of cognitive ability.

Women who wait earn more

Hot Research Friday!

Women who delay childbearing earn more money than women who have children at earlier ages. Kasey Buckles (2008), "Understanding the Returns to Delayed Childbearing for Workign Women" (AER 2008, 98:2, 403-407) finds that there is a raw return of approximately 3 percent per year of childbearing delay. The author finds that about 90% of the delay premium is explained by differences in observable characteristics such as education and occupation.

In short, women who delay are more educated, more skilled and more likely to be in professional careers.

Finance companies overcharge African-American automobile loan borrowers

Hot Research Friday!

Charles, Hurst, and Stephens (2008) article, "Rates for Vechicle Loans: Race and Loan Source" (AER, 2008 98:2,315-320), finds that finance companies, such as GMAC, charge African-American more than similarly situated White borrowers at finance companies. Some of their statistical models point to differences as large as 120 basis points (1.2%) higher rates for African-American borrowers.

The authors do not find a similar differential for African-American borrowers at banks. They also find that most of the difference in rates is between similarly situated African-American high interest rate borrowers and White hihg interest rate borrowers.

What do EEO-1 reports tell you about discrimination in a hiring case?

Probably not much.

EEO-1 reports from the EEOC provide a breakdown of how many people of different racial, gender, and age groupings are hired into different job positions within a company. The detailed reports, which most times the public will not have access to, provides a breakdown by employer. The more aggregated reports, which the public generally has access to, provides a breakdown at a higher level, such as at the zip code level or higher, by industry type. So ultimately the EEO-1 reports present the composition of the employer at a given point in time, usually a year.

In contrast, in a hiring case the plaintiffs are usually alleging that the defendant failed to hire
them because of their age, race, gender or other characteristics. In these types of cases, statistical analyses that compare the individual employer’s hiring numbers to EEO-1 data are common.

However, the EEO-1 report generally makes a poor comparison.

Instead of focusing only on the employer/defendant’s hiring decisions, a firm’s EEO-1 report reflects the composition of the workforce which is the result of hiring decisions over a much longer period than are issue in the typical hiring discrimination case. In short, the composition of an employer’s workforce is not necessarily representative of the composition of the hires in any given year, or the composition of the applicants available for hire for any given year.

For example, consider a hiring case where there are allegations of age discrimination. In this example, the employer starts with a workforce of 200 employees that are under the age of 40. and zero that are 40 years of age or older. For this firm, it’s workforce composition is 0% 40 or over.

If during the following year the firm hires 20 employees and they all 40 years of age or over, the firm’s composition of older worker hires is 100%. However, the overall composition of the workforce is now just 9.09% that is 40 years of age or older. That is 20 workers who are over the age of 40 out of a total of 220 employees.

The workforce composition of 9.09% of older workers does not adequately represent that the firm’s composition of hires of 100% older workers in that year. Therefore, in this setting, the composition of the workforce is not an appropriate measure to analyze the hiring decisions in any given year.

Bayesian v. Frequenist analysis of employment discrimination

In the article 'Hierarchical Models for Employment Decisions', Joseph Kadane and George Woodworth (Journal of Business and Economic Statistics, April 2004, Vol 22. No.2) advocate the use of Bayesian analysis to assist the trier of facts in the determining if the employer had discriminated against the protected groups of workers.

Conceptually, the Bayesian statistical analysis approach turns the traditional approach to analyzing statistical evidence of discrimination upside down. The traditional approach, or what statisticians refer to as the frequentist approach, begins by assuming that no discrimination has occurred and the employment decisions were made in a age, race, or gender neutral manner. From that point, the data is used to determine the likelihood that a age, race, or gender neutral employment process would have generated the employment outcomes that are at issue in the lawsuit.

In short, the frequenist approach ask:

'Assuming that the defendant is utilizing a age, gender, or race neutral employment process, what is the probability that the unbiased employment process could have generated the observed employer's employment decisions?'


If there is a very small probability that a neutral employment process would have generated the outcome then it may be inferred that discrimination has occurred.

The Bayesian approach turns the approach around and upside down The Bayesian approach ask the question:

'Given that we observe the defendant's employment decision outcomes, what is the likelihood that the employment decision was age, race or gender neutral?'

If there is a small probability that the data is consistent with a neutral employment process, then it may be inferred that some type of discrimination has occurred.

What are the pros and cons of each approach? To be discussed....!