Case Evidence: Blacks, Hispanics, and Immigrants

Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 1 / 48 Blacks CASE EVIDENCE: BLACKS Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 2 / 48

Blacks BLACKS Recall from previous lectures: Lower education attainment relative to whites on average More adversely affected than whites by declining relative wages for less-educated workers Hurt more by decline in blue-collar and manufacturing jobs Black unemployment rates twice as high as white unemployment rates at all ages More likely to be imprisoned Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 3 / 48 Blacks Couch and Daly (2002) Couch and Daly (2002) - Black-White Wage Inequality in the 1990s: A Decade of Progress GROUP 5 PRESENTATION Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 4 / 48

Blacks Couch and Daly (2002) Couch and Daly (2002) - NOTES Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 5 / 48 Blacks Couch and Daly (2002) Couch and Daly (2002) - NOTES Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 6 / 48

Blacks Couch and Daly (2002) Couch and Daly (2002) - NOTES Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 7 / 48 Blacks Couch and Daly (2002) Couch and Daly (2002) - NOTES Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 8 / 48

Blacks Heckman et al. (2000) Heckman et al. (2000)- Understanding Black-White Wage Differentials, 1960-1990 Previous literature focused on two factors contributing to the advance of black economic status: 1 Relative improvement in schooling (years and quality) Replacement of older cohorts by newer cohorts School quality improved more for blacks than for whites over first half of 20th century 2 Social activism and tight labor markets Evidence for this argument from improvements across all cohorts in certain periods Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 9 / 48 Heckman et al. (2000) - Goals Blacks Heckman et al. (2000) Examine robustness of previous literature on the source of black male economic progress to: Alternative data and data restrictions Who is included in the data set? What are the exclusion criteria? Alternative model/equation specifications How is quality measured and specified in the earnings equation? Controlling for selective withdrawal of low-wage workers from the labor force Low-wage dropouts are excluded from most datasets. Does this affect the estimated results? Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 10 / 48

Blacks Heckman et al. (2000) Heckman et al. (2000) - Three main conclusions 1 Choice of sample matters Different sampling rules affect the weight placed on different components (secular factors, social activism, and tight labor market) 2 Model specifications matters Estimated quality effects are sensitive to the measure used and specification in the wage equation (i.e. variables and functional forms) 3 Selection withdrawal matters Mean black economic progress is significantly lower when adjusted for labor-force selection (nonworker potential; low wage workers most likely to drop out) Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 11 / 48 Blacks Heckman et al. (2000) Heckman et al. (2000) - Other Findings Across all samples, find in 1980 to 1990: Negative within-cohort effect Poor labor market conditions for low-skill workers, where blacks are overrepresented Lower between-cohort improvements relative to previous 2 decades Slowing down of the quality improvements between cohorts made largest gains in previous two decades (ex: school quality) Evidence unclear whether changes in relative quality explain rising relative rate of return to education Considerable part of the measured contribution of education to eliminating the wage gap in this period is due to the selective withdrawal of low-wage, unskilled blacks from the workforce. Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 12 / 48

Hispanics CASE EVIDENCE: HISPANICS Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 13 / 48 HISPANICS Hispanics Recall: Educational gap between Hispanics and whites larger than black-white gap Negatively affected by increasing returns to skill Negatively affected by decrease in blue-collar and manufacturing jobs Large number of young immigrants with language barriers Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 14 / 48

Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - The Wage Structure of Latino-Origin Groups Across Generations Compare wage structure of Latino male worker to male non-hispanic black and male non-hispanic white worker wage structures Account for differences in Latino origin and degree of generation status in U.S. Consider three Latino origins ( 90% of Latinos): 1 Mexican 2 Central/South American 3 Puerto Ricans Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 15 / 48 Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Generations Foreign-born versus native-born Latinos Generational advancement important Immigrants tend to be educated abroad and natives are U.S. educated First generation may have advantage of optimism and work ethic Second generation - born in the U.S. with at least one parent born outside U.S. Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 16 / 48

Fry and Lowell (2006) - Origin Hispanics Fry and Lowell (2006) Mexicans are 58.5% of Hispanics Least educated and lowest average wages Origin varies by where live and by their legal status Previous research tends to lump Latinos into single group and doesn t account for origin Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 17 / 48 Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Background Latino men earn approximately 60% of white wages and 90% of black wages Over 55% of the Latino labor force is foreign born On average, Latino immigrants have 3 years less education than whites Native-born Latinos have: 1.3 years less education than whites 0.2 years less education than blacks Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 18 / 48

Fry and Lowell (2006) - Goals Hispanics Fry and Lowell (2006) Analyze the extent of labor market success across Latino generations while accounting for origin. Does the payoff to HK characteristics (experience, education) differ across Latino groups and generations based on whether their HK skills were acquired within or outside the United States? Investigate whether Latino immigrant labor market outcomes differ for those who arrived during childhood compared to those who arrive during adulthood. Near-native immigrants Generation 1.5: arrived in US between ages 5 and 13 Generation 1.75: arrived in US before age five Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 19 / 48 Fry and Lowell (2006) - Data Hispanics Fry and Lowell (2006) Current Population Survey (CPS) Male workers between the age of 18 and 61 Hourly earnings Omit Cuban and other Spanish origin workers Dropped self-employed respondents Evaluate with and without constrained coefficients Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 20 / 48

Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Unadjusted Data for Men 1999-2000 Mean wage High School Bachelor s + Experience Mexican 1.0 10.18 21.4 3.4 19.0 Mexican 1.5 10.87 27.4 3.7 12.4 Mexican 1.75 10.67 36.5 3.9 10.7 Mexican 2 13.74 35.1 9.9 14.7 Mexican 3+ 14.04 39.9 11.3 16.9 C/S 1.0 12.64 30.2 15.4 20.06 C/S 1.5 13.31 33.3 10.7 10.33 C/S 1.75 14.61 29.2 14.2 8.02 C/S 2 15.04 22.9 24.4 11.09 C/S 3+ 14.89 26.3 20.8 15.46 P.R. 1.0 12.88 30.1 11.9 24.4 P.R. 1.5 13.63 35.9 11.4 22.1 P.R. 1.75 13.93 28.6 7.0 17.8 P.R. 2 15.05 36.1 14.0 15.2 P.R. 3+ 14.53 38.4 11.4 12.6 White 1.0 20.52 23.2 49.4 19.8 White 3+ 18.49 33.1 30.0 18.6 Blacks 3+ 13.98 41.5 16.0 18.3 Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 21 / 48 Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Unadjusted Results Substantial intergenerational progress Yet, little growth between second and third-plus generations Large unadjusted gaps between male white workers and male native-born Latino workers Overall, whites in any generation do much better than Latinos in most categories (wage, education, experience) Variation in education and experience across generations and origins Question: Does the variation in education and experience explain the Latino/White wage gaps? Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 22 / 48

Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Wage Equation where log(w i ) = G i α + E i β + P i χ + R i δ + T i φ + ɛ i G i - vector of ethnic/generation dummy variables 3+ generation non-hispanic whites are the omitted group E i - vector of educational attainment categories P i - quadratic in labor market experience R i - vector of geographic (residence) dummy variables T i - vector of interview time variables Estimate the equation twice: 1 Constrain coefficients (β, χ, δ, φ) to be the same for all ethnic/generation groups 2 Allow coefficients to vary by ethnic/generation groups allow returns to HK to vary by group Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 23 / 48 Hispanics Fry and Lowell (2006) Wage Differential relative to White 3+ (Constrained) Basic Model Control for Experience Control for Education Mexican 1.0-0.607-0.623-0.280 Mexican 1.5-0.563-0.473-0.176 Mexican 1.75-0.550-0.421-0.177 Mexican 2-0.359-0.256-0.101 Mexican 3+ -0.286-0.247-0.111 C/S 1.0-0.432-0.470-0.263 C/S 1.5-0.404-0.263-0.093 C/S 1.75-0.325-0.116 0.008 C/S 2-0.303-0.157-0.088 C/S 3+ -0.186-0.145-0.064 P.R. 1.0-0.368-0.413-0.237 P.R. 1.5-0.367-0.325-0.165 P.R. 1.75-0.312-0.314-0.108 P.R. 2-0.225-0.198-0.069 P.R. 3+ -0.258-0.153-0.017 White 1.0-0.003-0.025-0.072 White 2 0.073 0.070 0.024 Blacks 3+ -0.248-0.260-0.180 Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 24 / 48

Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Constrained Results Third-plus generation black workers paid 25% less than third-plus generation white workers Second-generation Latino workers compared to third-plus generation white workers (raw gap): Mexican: 36% less Central and South American: 30% less Puerto Rican: 23% less Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 25 / 48 Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Constrained Results Simple proxies for labor market skill explain a substantial portion of the wage gaps between native-born Latinos and whites Youthfulness also substantially shrinks gap (i.e. accounting for age/experience of respondents) Gap shrinks further to a substantially smaller size when account for schooling differences Progress stalls for Mexican Americans after second generation but not true for other origins Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 26 / 48

Hispanics Fry and Lowell (2006) Estimated Returns to Education and Experience 7th-12th Bachelor s + 10 Years Exp 30 years Exp Mexican 1.0-0.13 0.32 0.23 0.42 Mexican 1.5-0.15 0.53 0.29 0.49 Mexican 1.75-0.19 0.52 0.33 0.56 Mexican 2-0.21 0.60 0.36 0.66 Mexican 3+ -0.18 0.48 0.38 0.65 C/S 1.0-0.14 0.45 0.21 0.36 C/S 1.5-0.10 0.45 0.43 0.98 C/S 1.75-0.22 0.61 0.15 0.83 C/S 2-0.33 0.35 0.25 0.46 C/S 3+ -0.46 0.38 0.26 0.48 P.R. 1.0-0.15 0.59 0.20 0.47 P.R. 1.5-0.29 0.55 0.38 0.66 P.R. 1.75-0.17 0.35 0.23 0.49 P.R. 2-0.18 0.56 0.36 0.57 P.R. 3+ -0.01 0.44 0.51 0.52 White 1.0-0.24 0.56 0.31 0.57 White 3+ -0.21 0.48 0.39 0.68 Blacks 3+ -0.21 0.50 0.27 0.52 Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 27 / 48 Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Estimated Returns Returns to education: Estimated returns to education for native-born Latinos are at least as high as returns for 3+ generation whites Only group facing significantly different returns is first generation Mexican workers Returns to experience: Wage growth from experience lower for native Central/South Americans and Puerto Ricans relative to 3+ generation whites Question: How much of the wage gap is due to differences in HK? Use a standard decomposition Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 28 / 48

Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Wage Decomposition Differences in mean characteristics explain about 6% points of raw 24% difference between white and black workers Differences in skills explains at most a quarter of the gap Differences in mean characteristics between native-born Latinos and 3 rd generation whites explains about half of the wage gap Wage structure of native-born Latino workers similar to that for whites (not the case for black workers) Latino-white wage gap is primarily due to skill differences (unlike black-white wage gap) Most of Latino-white wage gap due to differences in quantity of skills rather than relative difference in price of skills (returns) Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 29 / 48 Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Conclusions Latinos are projected to be fastest growing segment of the labor force over the next 50 years Results suggest that significant resources devoted to equalizing educational opportunities for Latinos will help equalize labor market outcomes All Latino-origin groups do NOT have the same labor market experience Little support for hypothesis that arrival in U.S. during childhood offers advantage Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 30 / 48

Hispanics Fry and Lowell (2006) Fry and Lowell (2006) - Model Limitations and Suggested Corrections Assignment 7 Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 31 / 48 Hispanics Mora and Davila (2006) Mora and Davila (2006) - A Note on the Changes in the Relative Wages of LEP Hispanic Men between 1980 and 2000 Evaluated wage differentials between Hispanic and non-hispanic white men in U.S. Include measures for English proficiency Limited-English-Proficient (LEP) Hispanic workers Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 32 / 48

Mora and Davila (2006) - Data Hispanics Mora and Davila (2006) U.S. and foreign-born Hispanic men age 25-64 Must have reported positive wage and salary income Worked at least 20 hours a week for 32 weeks or more Raw data: Earnings gap between non-hispanic White and Hispanic men has widened since 1980 Similar growth in Hispanic/non-Hispanic White earnings gap for LEP and English fluent between 1980 and 2000 (despite return to skill increases) Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 33 / 48 Hispanics Mora and Davila (2006) Mora and Davila (2006) - Wage Decomposition Returns to observable skills is the largest component of the growing differential Hispanics, especially LEP, were able to offset some of the rising observable-skill returns through improvements in unobservable characteristics Most pronounced for LEP Hispanics during 1990s LEP Hispanics gained in unobservable skills relative to non-hispanic Whites Overall, LEP Hispanic men moved up in the earnings distribution relative to English proficient Hispanics between 1990 and 2000 Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 34 / 48

Immigrants CASE EVIDENCE: IMMIGRANTS Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 35 / 48 Immigrants Lubotsky (2007) Lubotsky (2007) - Chutes or Ladders? A Longitudinal Analysis of Immigrant Earnings Immigrants that stay in the U.S. may differ from immigrants who return to their country or immigrants who make multiple transitions Cross-sectional data may not capture these immigrants Cross-sectional data - a one-dimensional data set where data is collected on several individuals at the same point in time Need to account for selection issues when estimating the earnings gap Immigrant-native earnings gap has important policy implications Disagreement over how to measure experience of immigrants Debate on how open U.S. should be to low-skilled individuals Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 36 / 48

Immigrants Lubotsky (2007) Lubotsky (2007) - Important Immigration Policies Immigration Act of 1990 Increased number of visas allocated on basis of occupational skills from 54,000 to 140,000 per year 1996 Welfare Reform Restricted ability of new immigrants to obtain cash transfers, food stamps, and Medicaid during first 5 years in U.S. Goal: Discourage immigration solely for benefits Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 37 / 48 Immigrants Lubotsky (2007) Lubotsky (2007) - Permanent Out-Migration About one-third of immigrants to U.S. eventually return to home country permanent out-migration Plan to return to home country after limited period of time (ex: target level of savings) Poor labor market outcomes in U.S. (unanticipated) People with lower cost of migration more likely to return home Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 38 / 48

Immigrants Lubotsky (2007) Lubotsky (2007) - Permanent Out-Migration If low-wage immigrants are more likely to return to their home country, the earnings gap in cross-sectional data will overestimate returns to immigration and speed of assimilation Only evaluating those who stay may differ significantly from those who leave Permanent out-migration is disproportionately more common among low-earning immigrants Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 39 / 48 Immigrants Lubotsky (2007) Lubotsky (2007) - Temporary Out-Migration Selective back-and-forth migration may also exist temporary out-migration Respondents tend to list most recent arrival rather than initial arrival in survey Such immigrants will show up in later cohorts If temporary out-migration workers tend to have lower earnings or slower growth, selective migration will lower earnings and growth of more recent cohorts Found 14% of immigrants have earnings in the U.S. prior to reported arrival year Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 40 / 48

Lubotsky (2007) - Data Immigrants Lubotsky (2007) Matched annual earnings records from Social Security with 1990 and 1991 Survey of Income and Program Participation (SIPP) 1994 March Supplement to the Current Population Survey (CPS) Longitudinal data helps overcome selection issues Longitudinal data - repeated observations of the same individuals over a long period of time Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 41 / 48 Lubotsky (2007) - Data Immigrants Lubotsky (2007) Immigrant: anyone born outside U.S. Dropped if born to American parents, born in Puerto Rico or arrived in U.S. prior to age 18 Men born between 1930-1969 Issue: Arrival response may not accurately reflect initial year of arrival Question asked for date respondent decided to stay Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 42 / 48

Immigrants Lubotsky (2007) Lubotsky (2007) - Arrival Date Considered three alternative measures: 1 Reported date of arrival Understates exposure to U.S. labor market if make multiple trips May not have earnings for significant period after arrival - problem of compositional changes 2 Earliest of either reported date of arrival or first year of earnings Adjusted date of arrival Problem of compositional changes 3 First year of covered earnings Understates length of time spent in U.S. Easiest interpretation Definition of arrival date matters Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 43 / 48 Lubotsky (2007) - Evidence Immigrants Lubotsky (2007) Data consistent with large number of low-earning immigrants making multiple trips to U.S. and reporting most recent date of arrival Immigrants earnings growth faster in repeated cross-sectional data Evidence of selective out-migration among lower-earning immigrants Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 44 / 48

Lubotsky (2007) - Results Immigrants Lubotsky (2007) Earnings growth of immigrants slower when use longitudinal data compared to cross-sectional Longitudinal data accounts for immigrants who return and migrate multiple times Immigrant earnings grow by 10-15 percentage points more over their first 20 years in U.S. relative to native-born workers Temporary out-migration by low-wage immigrants has a significant effect on earnings trends across cohorts - overestimate deterioration in earnings of newly arrived immigrants Similar to discontinuous worker model Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 45 / 48 Immigrants Lubotsky (2007) Lubotsky (2007) - Model Limitations Possible errors when trying to match data sets Not able to be match all respondents Used population weights to control for this problem Self-employed and uncovered sector wages not included Men only Didn t account for family ties - assistance with assimilation Selection into survey Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 46 / 48

Immigrants Lubotsky (2007) Lubotsky (2007) - Model Limitations and Suggested Corrections Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 47 / 48 Upcoming Readings Case Evidence: Homosexuality Antecol et al. (2008) Black et al. (2003) Immigrants Lubotsky (2007) Case Evidence: Competitiveness Niederle and Vesterlund (2007) Price (2008) Case Evidence: Personality Mueller and Plug (2006) - Group 6 Case Evidence: Looks Perisco and Postlewaite (2004) - Group 7 Rosburg (ISU) Case Evidence: Blacks, Hispanics, and Immigrants Spring 2010 48 / 48