Program Report: Labor Studies, 2013
The Labor Studies (LS) Program is one of the largest and most active in the NBER, with almost 150 members producing nearly 200 Working Papers each year. The range of topics and expertise is stunning: it ranges from cutting edge research on aggregate labor market issues like unemployment and productivity to the effects of government programs like Disability Insurance, the differences in labor market outcomes among different educational, gender, and racial groups, and to many other topics in social science.
Reflecting their diversity, two-thirds of program members are affiliated with at least one other NBER program, and in the past few years the Labor Studies program has convened joint sessions at the NBER's Summer Institute with Public Economics, Economics of Education, Economics of Children, and with Working Groups in Personnel Economics and the Economics of Crime. This summer we will add a new joint session with Development Economics.
In this report I briefly summarize some of the main themes emerging from recent work by LS affiliates in three areas: immigration, gender, and unemployment. These topics barely scratch the surface of the vast body of work by LS affiliates, but give a flavor of some of the emerging ideas and latest techniques in the field.
Immigration
Over the past three decades, Labor Studies researchers have produced a series of major NBER research volumes on the economics of immigration (Abowd and Freeman, 1991; Borjas and Freeman, 1992; Borjas, 2000; and Borjas 2007) as well as many influential articles. This work continues with a focus on several new issues related to immigration.
One such issue relates to the growing importance of immigrants in the science and engineering workforce of the United States, particularly at the doctoral level, where immigrants now make up about one-half of all newly awarded Ph.D.s. John Bound, Sarah Turner, and Patrick Walsh (14792) point out that this has been driven in part by the rapid rise in the production of bachelors' degrees outside the United States, and they document its impact on the demand for advanced training inside the United States. Jeffrey Grogger and Gordon Hanson (18780) use data from the Survey of Earned Doctorates to study the determinants of which foreign-born students intend to stay in the United States. They find that the United States attracts the most talented foreign students, and they also show how changing economic conditions in sending countries affect the decision to stay. Using a unique survey of authors of recent scientific publications, Paula Stephan, Chiara Franzoni, and Giuseppe Scellato (18809) find that high prestige of the program or job and strong career prospects are the major factors driving the decisions of Ph.D. students and post-doctoral candidates to choose the United States over other potential host countries.
The impact of these foreign born scientists, engineers, and other highly trained workers is the subject of several recent studies. William Kerr and William Lincoln (15768) use data from the H-1B visa program to study the city-level and firm-level impacts of foreign-born science and engineering workers. While a traditional concern is that the foreign-born tend to crowd out natives, their analysis suggests that the opposite may be true, in part because of the direct employment contribution of foreign born inventors. George Borjas and Kirk Doran , who study the inflow of Russian mathematicians to U.S. universities following the collapse of the Soviet Union (17800), reach the opposite conclusion. Their analysis of publications and career mobility suggests that, in the case of advanced-level mathematics, the "pie" is essentially fixed, with no positive spillovers for native scholars.
Another emerging strand of research by LS members focuses on the broader impacts of immigration flows on local and national economies. Frederic Docquier, Caglar Ozden, and Giovanni Peri (16646) use new data on labor force stocks and migration flows by education level for OECD countries, combined with detailed models of the labor markets in each country, to simulate impacts of population movements on a country-by-country basis. They conclude that immigration has been a net positive factor for workers in most countries. Francine Blau and Lawrence Kahn (18515) similarly review the effects of immigration on the overall distribution of incomes in the United States and other major countries. They conclude that the presence of immigrants has contributed to wage inequality, although the effect is small relative to other forces, such as technology and trade. Peri (17570) similarly concludes that immigration has had little effect on poverty rates in the United States.
Language skills have long been recognized as a major factor in understanding differences between natives and immigrants, and several recent NBER working papers explore the impacts of language ability among immigrants. Jennifer Hunt (18696) finds that the lack of English language skills accounts for most of the pay gap between natives and immigrants with an undergraduate degree in engineering. Ethan Lewis (17609) shows that differences in language skills lead to a segmentation of the occupations held by immigrants and natives. The segmentation is most apparent in cities with large fractions of Spanish speakers. In the extreme case of Puerto Rico, Lewis finds that immigrants and natives are in much more direct labor market competition.
Gender
Members of the LS program have played a preeminent role in research on gender-related issues for many decades. Though the wide disparities in the labor market status of men and women that characterized earlier generations have narrowed, many differences remain, and in some cases new gaps have opened up. Much recent attention has focused on the dramatic reversal of the gender gap in educational attainment. As documented by Claudia Goldin, Lawrence Katz, and Ilyana Kuziemko (12139), female high school graduates had narrowed the differences with the male peers in achievement test scores, and were more likely to attend and graduate from college, by the early 1990s. Blau, Peter Brummund, and Albert Yung-Hsu Liu (17993) show that this sharp rise in relative education of women correlates with a decline in occupational segregation between male and female workers, as more educated women have entered traditionally male occupations.
While women are more likely to attend and complete college than men, there are still large differences in fields of study. Joseph Altonji, Erica Blom, and Costas Meghir (17985) use data on field of degree in the American Community Survey to document that women are under-represented in engineering, computer science, physics, economics, and business, but over-represented in communications, psychology, education, and English. The latter fields are associated with lower earnings for both men and women. These researchers show that differences in college major choices are an important contributor to the earnings disparities between college-educated men and women.
One factor that may explain some of the gender gap in education choices, career progression, and pay is a difference in "competitiveness". Muriel Niederle and Lise Vesterlund (11474) conducted a series of laboratory experiments to gauge willingness to compete in tournament-like competitions, and found sharp differences between men and women that remain even after controlling for risk aversion and over-optimism. Thomas Buser, Niederle, and Hessel Oosternbeek (18576) correlate similar measures for students in the Netherlands, and show that differences in competitiveness help to explain the lower fraction of girls who choose the most prestigious (science-based) track at high school.
More recent cohorts of women also have narrowed the gap in cumulative labor market experience relative to men. Martha Bailey, Brad Hershbein, and Amalia Miller (17922) show that some of this increase in career attachment was due to easier access to birth control and lower early fertility. Raquel Fernandez and Joyce Cheng Wong (17508) highlight the effect of the rising risk of divorce on women's decisions to acquire more education and stay attached to the labor market.
The rise in the relative success of women also may have been helped along by a decline in the demand for the manual skills traditionally supplied by less-educated men. Paul Beaudry and Lewis (18159) show that in cities where the college-high school wage gap has risen more quickly, the male-female wage gap has narrowed more quickly. They then show that both trends were correlated with more rapid local adoption of computer technology in the 1980s and 1990s, underscoring the role of changing relative skill demand. Similarly, Chinhui Juhn, Gergely Ujhelyi, and Carolina Villegas-Sanchez (18106) show that passage of NAFTA led Mexican firms to adopt new technologies that reduced the demand for physical skills and ultimately led to increased hiring of women relative to men.
Despite the progress made by recent cohorts, women still fall behind men, particularly in certain fields like science and engineering, and in high-profile careers in management and business. Marianne Bertrand, Goldin, and Katz (14681) study MBA graduates from a top U.S. school, and show that the male-female earnings gap widens steadily after initial completion of the degree, reaching 80 percentage points over 16 years. They find that the presence of children is a powerful predictor of career interruptions and lower hours for females, but not for men, and that these factors are highly related to earnings. Ty Wilde, Lily Batchelder, and David Ellwood (16582) reach a similar conclusion for a broader sample of women in the 1979 National Longitudinal Survey of Youth: childbearing has a strong negative effect on wage growth, particularly for higher-skilled women.
Unemployment, Job Displacement, and the Great Recession
The Great Recession has brought renewed interest in the study of labor market fluctuations, unemployment, and job displacement. While research is still ongoing, prominent contributions by LS members already have shed light on the labor market impacts of the Great Recession and its likely consequences.
Michael Elsby, Bart Hobijn, and Ayesegul Sahin (15979) provide an early analysis of the labor market consequences of the downturn that began in 2007. They note that the impact of the recession was particularly severe for men, who were disproportionately affected by job losses in construction and manufacturing. They also highlight the remarkable growth in long-term unemployment which is one of the hallmarks of the Great Recession. Henry Farber (17040) used data from the Displaced Worker Survey of January 2010 to show that nearly one in six U.S. workers reported having lost a job during 2007-9. Comparing recent job losers to those in earlier surveys, he noted a sharp decline in re-employment rates and a rise in measured earnings losses.
Hilary Hoynes, Douglas Miller, and Jessamyn Schaller (17951) provide a systematic analysis of the relative impact of the Great Recession across various demographic groups. Their research, based on monthly data from the Current Population Survey, shows that the relative responses to the most recent downturn were quite consistent with patterns in earlier recessions but larger in magnitude, reflecting the severity of the downturn. One of the least affected groups was older workers -- a fact confirmed by Alan Gustmann, Thomas Steinmeier, and Nahid Tabatabai (17547) based on data from the Health and Retirement Survey.
Steven Davis and Till von Wachter (17638) evaluate the longer-run costs of the massive job losses during the Great Recession. Using data from Social Security earnings records for job losers in the 1970s, 1980s, and 1990s, they estimate that a "typical" displaced worker (a male with three or more years of job tenure, laid off from a firm experiencing a 30 percent or larger cut in employment) experienced about a 12 percent loss in the discounted present value of earnings over the next 20 or so years. The loss rises to 20 percent, however, when the overall unemployment rate is greater than 8 percent at displacement. These estimates suggest that the longer-run costs of the Great Recession will be very large, and that many job losers will never see their salaries rebound to their pre-job-loss levels.
One potential mechanism accounting for the high cost of long-term unemployment is that workers become less likely to find a new job, either because of real or perceived deterioration in their skills. Kory Kroft, Fabian Lange, and Mathew Notowidigdo (18387) conduct an ingenious field experiment to measure this effect. They submit fictitious job resumes to a large sample of job postings in different U.S. cities, randomly varying the length of time the (fictitious) applicant has been out of work. They find that applicants who have been out of work longer are less likely to be called for an interview, although the effects are moderated in cities with higher unemployment rates.
An important policy response to the Great Recession was a large increase in the potential duration of unemployment benefits -- from the standard 26 weeks to as long as 99 weeks for job losers in some states. Jesse Rothstein (17534) uses month-to-month labor force transition data from the Current Population Survey to evaluate the impact of these longer benefits on re-employment rates and overall unemployment. Using a variety of approaches to control for unobserved variation in local labor market conditions, he concludes that the package of benefit extensions raised the unemployment rate in December 2010 by at most 0.5 percentage points -- a smaller effect than would have been expected given existing estimates of the effect of longer benefits on the duration of unemployment claims in the literature.
* In this article, the numbers in parentheses refer to NBER Working Papers
