Discrimination Notes

Labour Economics Discrimination: Tutorial notes, revision notes & lecture summary Contents????Overview Gender Discrimination Role of Legislature and Judiciary Race Discrimination Taste based discrimination model: BECKER 1957 o Overview o Criticism Statisical Discrimination Model Implicit Discrimination Model Monoposony (Manning 2003) Essay Plan: 'The failure of women to achieve equality of earnings with men has more to do with their lifetime choices than with discrimination.' Discuss. Overview BECKER (1971): discrimination as unequal treatment among equivalent people due to race/gender Discrimination in recruitment, occupation or advancement vs wage discrimination. Occupational segregation may not necessarily mean occupational discrimination. May be due to preferences. Key: Difference between disadvantage and discrimination. Blinder-Oaxaca decomposition:?? Can't just use difference in averages/regression to claim discrimination exists. Decomposes wage differentials into two components. o 1. Explained component: disadvantage. Characteristics. A portion that arises because two comparison groups have different qualifications/credentials (eg years of schooling/experience in labour market). Lower personal characteristics could be consequence of discrimination outside labour market (eg education) or previous acts of discrimination in the labour market (statistical discrimination). For race: occupational crowding is the issue. o 2. Unexplained component: discrimination. Coefficient effect is a measure of labour market discrimination. o Can be self-fulfilling - eg. Manager not promoting because thinks women will get pregnant, then women does because not promoted. Controversy: some believe discrimination does not exist in labour market or B-O decomposition overestimates it. They point out that the wage equation cannot include all relevant variables. Problems of omitted variables, multicollinearity and autocorrelation. ECKSTEIN ET AL (1999): Problem in observing statistics: can only observe actual wage not offered wage. Eg. For ethnic groups wages offered are 3x greater than differences in those accepted. Need some kind of IV/natural experiment to get counterfactual what women would earn if male. Labour Economics? NEAL ET AL (1996): The role of Premarket Factors in Black-White Wage Differences Aim of paper: o Standard finding: controlling for observables still leaves large black-white wage gap. o Follow group of young in US between 1957 and 1964. In 1980 they all sat the Armed Forces Qualifying Test (AFQT), which is claimed to be a good measure of IQ. o Looked at their wage gap in 1990/91 with and without AFQT score. o Results: v.dramatic differences in scores on AFQT test by race. Black males earn 24.4% less than white males. Females earn 18.5% less. Controlling for education reduces this a bit but not much. But, including AFQT score eliminates 75% of the black-white gap for men and all the gap for women. o Implies wage gap is EXPLAINED. o Issues with paper: ? Perhaps other measures of IQ better ? Why do blacks do badly on AFQT test - ie discrimination from employers pushed to school Female Discrimination???Different position for full time and part time women. 26% of women in OECD countries work part time. Full time women: median now earning 90-91% of full time men in UK. But varies internationally - 87% in Slovenia vs 42% in Japan. BLAU AND KAHN (2003): despite huge earnings change; position of women not v.different. GREGORY (2009): trends towards convergence - though that was based primarily on growth in southern europe, which has now declined. UK worst in EU for female judges. Improvement since early 50s (60%) primarily due to improvement in observed characteristics - eg. Education, experience, job type, unionisation etc. Soft skills becoming more important. Fallen due to: o Education: upsurge in human capital - accentuates incentive to employee participation, though women specialise in humanities (less well rewarded) o Technological change: reduced role of non-routine analytical and interactive tasks Depends on type. Part-time women suffer when compared to full time men. But, BO decomposition. Discrimination does appear to exist - eg, Blau and Kahn (2001) find that statistical adjustments for human capital and industry or occupational differences only accounts for a limited part of the wage differential, leaving a 9% difference unaccounted for. o Criticism: could be other incommensurable factors eg negotiation, commitment But, evidence does still seem reasonably convincing eg. Goldin et al (2000): orchestra selection. Gregory (2009): earn more after leaving university but less during 'motherhood penalty' years. o Interrupted employment reduces development of job specific human capital o Job separation reduces value of job specific human capital o May earn lower returns because of real/perceived mental attributes eg job effort/motivation. Recent years: equal pay legislated for. Motherhood penalty explains 40-50% of gender gap in UK and USA. Crucial problem: rational from employers perspective. Difficulty of legislation in circumventing it. Labour Economics?Not just pay but also initial employment/promotion. Only 13.7% of women on boards in large listed companies. Only 3% CEOs and 8% senior execs are women. Being aided by technical based changed. Helped women replace men with increasing ease. Key model: monopsony (see below) Role of Legislation and judiciary ? 1970 Equal Pay Act: prohibits less favourable treatment between men/women. Came into force 1975. By end of 70s 50% ratio had become high 60s, so clearly equal pay act had some impact. Not necessarily much direct court action but unions used it.Impact increased by fact 70s were decade of periodic income policies. ? 1980s: British government taken to European courts for inadequate gender legislation. New legislation as a consequence. EU comparable worth principle: Equal Pay for Work of Equal Value. ? 80s: EU commission took British government to court for inadequate protection of women. Equal pay act not sufficient protection. Consequence: Equal Pay for Work of Equal Value. Equal Pay for Work of Equal Value. For a while pretty effective - unions have used it effectively (Blau and Kahn, 2001) ? Other mechanisms: o Vivian Reding (ECJ): proposed mandatory quotas but dropped in Oct 2012. Controversial. 30% club: should come voluntarily. Where imposed - eg. Norway, problematic. Various countries - eg. Britain, are opposing introduction of quotas. ? Other mechanisms: Scandinavia: family only gets leave if father takes it ? EOC - equal opportunities commission enforces acts, now CSJ - commission for social justice. ? 2010: Equal Pay Act repealed, replaced with Equality Act Race stats??ECKSTEIN ET AL (1999): Problem in observing statistics: can only observe actual wage not offered wage. Eg. For ethnic groups wages offered are 3x greater than differences in those accepted. Need some kind of IV/natural experiment to get counterfactual what women would earn if male. Complex. Can't just talk about non-whites. Experience of different non-white groups differs massively from one group to another. Apparent evidence: B-O decomposition: British born ethnic minorities have slightly higher wages than white native born peers, but should be even higher if incorporated characteristics. Eg. Hispanics in states a disaster. Racial groups: can become arbitrary. Context important in terms of intergenerational transmission mechanism: o Until early 50s all white. China towns Liverpool/London; small numbers of blacks, Caribbeans, few west Africans. o Early 50s: significant waves of non-white immigrants. o First two groups - West Indians/West Africans. Pakistanis/Bangladeshis. Groups encouraged in - economists thought running out of labour. o Other groups came in later - Indians, East African Asians, Chinese. Earliest groups did worse because less stringent immigration criteria. Now criteria of education and wealth. May not be to do with wealth but intergenerational transmission of advantage and disadvantage, though lack of conclusive empirical research. Labour Economics??DUSTMANN ET AL (2010): supports need to distinguish. o Study: how British born ethnic minorities perform in terms of education, employment and wages in relation to parent generation o Results: first generation ethnic minority immigrants have lower employment propensities for same wages than native born whites. Unexpected since have 0.6 years more in full-time education than white natives. Overall difference in employment differences: 5.8%. Breakdown: difference mainly due to lower employment probabilities of Pakistanis/Bangladeshis. o No conclusive evidence on discrimination. o Further analysis: indicates most of difference due to different participation functions, Inactivity primarily driven by choice. o Occupational crowding: every one of 5 migrant racial minority groups had more crowding. Chinese had most (45%). Eg Indians - among 10 most common occupations doctor, lawyer, accountant etc. o Criticism: measure of occupational crowding crude. Only conclusive for Chinese in restaurant trade and Irish in building trade. Future: both US and Canadian evidence indicates considerable success within the labour market. Eg. CARD ET AL (2000): finds that children of immigrants tend to have higher education and wages than children of natives. Most convincing evidence from field experiments - suggest discrimination widespread eg. BERTRAND et al (2004): field experiment replying to job adverts in Boston and Chicago. White sounding applicants get callbacks 3.2% more. Higher-quality white-sounding applicants receive more callbacks than lowerquality sounding applicants but not true for African-American sounding applicants/ Limitations: only examines callback rates. Ultimate hiring rates more important. Also only replies to job adverts. Eg. What about social networks? More likely to harm blacks? 

Buy the full version of these notes or essay plans and more in our Labour Economics: Discrimination Notes.