ASSA 2022 Round-up: Day 1
Yesterday was the first day of the 3-day annual meeting of the Allied Social Science Associations, which is organized by the American Economic Association. The conference, held virtually again this year due to the ongoing coronavirus pandemic, features hundreds of sessions covering a wide variety of economics and other social science research. This year, Equitable Growth’s grantee network, Steering Committee, and Research Advisory Board and their research are well-represented throughout the program, featured in more than 60 different sessions of the conference.
Below are abstracts from some of the papers and presentations that caught the attention of Equitable Growth staff during the first day of this year’s conference and which relate to the research interests laid out in our current Request for Proposals. We also include links to the sessions in which the papers were presented.
Come back tomorrow morning for more highlights from day two, and Monday morning for highlights from day three.
Manish Raghavan, Harvard University; Solon Barocas, Microsoft Research; Jon Kleinberg, Cornell University; Karen Levy, Cornell University
Abstract: There has been rapidly growing interest in the use of algorithms in hiring, especially as a means to address or mitigate bias. Yet, to date, little is known about how these methods are used in practice. How are algorithmic assessments built, validated, and examined for bias? In this work, we document and analyze the claims and practices of companies offering algorithms for employment assessment. In particular, we identify vendors of algorithmic pre-employment assessments (i.e., algorithms to screen candidates), document what they have disclosed about their development and validation procedures, and evaluate their practices, focusing particularly on efforts to detect and mitigate bias. Our analysis considers both technical and legal perspectives. Technically, we consider the various choices vendors make regarding data collection and prediction targets, and explore the risks and trade-offs that these choices pose. We also discuss how algorithmic de-biasing techniques interface with, and create challenges for, anti-discrimination law.
Anna Sinaiko, Harvard University; Meredith Rosenthal, Harvard University; Vilsa Curto, Harvard University
Abstract: Vertical integration in healthcare has recently garnered scrutiny by antitrust authorities and state regulators. We examined trends, geographic variation, and price effects of vertical integration and joint contracting between physicians and hospitals using physician affiliations and all-payer claims data from Massachusetts in 2013–2017. Vertical integration and joint contracting with small and medium systems rose from 19.5 percent to 32.8 percent for primary care physicians and from 26.1 percent to 37.8 percent for specialists; vertical integration and joint contracting with large systems slightly declined. Geographic variation in vertical integration and joint contracting with large systems increased. We found that vertical integration and joint contracting led to price increases of 2.1 percent to 12 percent for primary care physicians and 0.7 percent to 6 percent for specialists, with the greatest increases in large systems. These findings can inform policymakers seeking to limit growth in healthcare prices.
Mark Curtis, Wake Forest University; Daniel G. Garrett, University of Pennsylvania; Eric Ohrn, Grinnell College; Kevin A. Roberts, Duke University; Juan Carlos Suarez Serrato, Duke University
Abstract: We study how tax policies that lower the cost of capital impact investment and labor demand. Difference-in-differences estimates using confidential U.S. Census Data on manufacturing establishments show that tax policies increased both investment and employment, but did not lead to wage or productivity gains. Using a structural model, we show that the primary effect of the policy was to increase the use of all inputs by lowering overall costs of production. The policy further stimulated production employment due to the complementarity of production labor and capital. Supporting this conclusion, we find that investment is greater in plants with lower labor costs. Our results show that recent tax policies that incentivize capital investment do not lead manufacturing plants to replace workers with machines. Note: This research was funded in part by Equitable Growth.
Arindrajit Dube, University of Massachusetts Amherst; Suresh Naidu, Columbia University; Adam Dalton Reich, Columbia University
Abstract: What is the value of workplace dignity to low-wage workers, and is it supplied efficiently? We present a methodology to elicit valuations of difficult-to-measure job characteristics under imperfect competition and use it to measure the value of workplace dignity to workers and its relationship to wages. We draw on extensive ethnographic work conducted with 87 Walmart workers to design and implement a survey experiment with over 10,000 Walmart workers recruited online. We first validate our experimental design by showing that our hypothetical quit elasticities are close to other estimates in the literature, and that subjects’ behavior (their likelihood of clicking on an outside job link) is consistent with their stated preferences and together imply a nontrivial degree of monopsony power. Next, we estimate workers’ valuations of different workplace amenities, from commuting and scheduling to more subjective measures of workplace dignity, including supervisor respect and self-expression at work. We find that workers value workplace dignity at approximately 6 percent of their current wage, making it comparable to amenities like commute time and more valuable than widely discussed amenities like control over one’s schedule. We find that workers’ experience of dignity at work is higher for White, older, and Southern workers, and is correlated with a measure of job rents. High-dignity jobs are also associated with a lower quit elasticity, but a higher bargaining elasticity, with no such heterogeneity by the nondignity job amenity values. Third, we use geographic variation in the bite of Walmart’s 2014 voluntary minimum wage policy to estimate the causal impact of higher wages on changes in dignity. Contrary to the theory of compensating differentials, we find that workplace dignity is not a substitute for wages, as reported dignity values do not decrease among those workers likely to have experienced a raise due to the voluntary minimum wage. Note: This research was funded in part by Equitable Growth.
Ihsaan Bassier, University of Massachusetts Amherst
Abstract: How does centralized bargaining affect the broader wage structure? And what does this tell us about the mechanisms that govern the wages and flows of labor? Using bargaining council contracts matched with firm- and worker-level data in South Africa, I show wages correspond sharply to large contracted wage increases especially in mid-wage and mid-size firms. A simple model of monopsonistic competition and strategic interaction predicts sizable cross-firm wage spillovers from such partially treated labor markets. To account for the overlapping and localized nature of labor markets, I show that a model-based measure of wage spillovers is proportional to worker flows between firms. Using observable worker-level flows, I isolate the empirically relevant labor market segment, and find a cross-wage elasticity of about 0.8. Previous estimates are lower, which is consistent with being based on less precise measures of spillovers. A microdata simulation suggests that accounting for these spillovers doubles the effect of contracted wage increases on the full wage distribution. Note: This research was funded in part by Equitable Growth.
Enghin Atalay, Federal Reserve Bank of Philadelphia; Sebastian Sotelo, University of Michigan, Ann Arbor; Daniel Tannenbaum, University of Nebraska-Lincoln
Abstract: Working in urban commuting zones (CZs) commands a large earnings premium, and this premium differs significantly by worker skill level. In this paper, we produce new descriptive evidence and introduce new measurement tools to understand the mechanisms behind the urban premium and why it differs by worker skill level. We use the near-universe of job vacancies and develop granular measures of job tasks—based on the natural language employers use, rather than survey-based categories—that allow for differences within occupations and across CZs. We find evidence for three mechanisms behind the earnings premium. First, jobs are more interactive and analytic in urban CZs, even within narrow occupation categories. Second, the computer software requirements of jobs are more intensive in urban CZs. Third, urban workers are more specialized, with less overlap in the sets of tasks performed, within occupations. Furthermore, these differences across CZs are more pronounced for college-educated workers than for noncollege workers. We show that job tasks and technologies account for a substantial portion of the urban CZ premium—even within occupations—and this relationship is stronger for white-collar occupations. Note: This research was funded in part by Equitable Growth.
Marshall Drake, Boston University; Neil Thakral, Brown University; Linh Tô, Boston University
Abstract: This paper introduces the use of a Bayesian Adaptive Choice Experiment, or BACE, a dynamic preference elicitation method. The method makes it possible to obtain individual-level willingness-to-pay estimates taking into account heterogeneity in inattention by using a dynamically chosen sequence of discrete choice scenarios that yields the largest information gain about preference parameters. We apply the method to understand differences in the values of job amenities by gender. The new data also allow us to estimate a new model of compensating differentials extending the Rosen (1986) framework to understand how workers trade off workplace flexibility.
Natalie Cox, Princeton University; Ernest Liu, Princeton University; Daniel Morrison, Princeton University
Abstract: Do government-funded guarantees and interest rate caps primarily benefit borrowers or lenders under imperfect competition? We study how bank concentration impacts the effectiveness of these policy interventions in the small business loan market. Using data from the Small Business Administration’s Express Loan Program, we estimate a tractable model of bank competition with endogenous interest rates, loan size, and take-up. We introduce a novel methodology that exploits loan “bunching” in the two-dimensional contract space of loan size and interest rates, utilizing a discontinuity in the SBA’s interest rate cap. In concentrated markets, we find that a modest decrease in the cap would increase borrower surplus by up to 1.5 percent, despite the rationing of some loans. In concentrated markets with a 50 percent loan guarantee, each government dollar spent raises borrower surplus by $0.64, boosts lender surplus by $0.34, and generates $0.02 of deadweight loss.
Sarah AlHaif, Howard University
Abstract: A rising tide does not lift all boats equally. During the expansion from the Great Recession, Black wage growth lagged that of Whites. While lower unemployment rates and the increased transition of workers from unemployment to employment is correlated with rising wages, this effect is through the correlation of job-to-job transitions on wages that increase as the labor; the actual mechanism is the increase in job-to-job transitions. But the minimum wage is important as a reservation wage and because increases in the minimum wage cause wage compression within firms create wage pressures on job-to-job transitions. But these pressures have different results by race because of differences in job-to-job transitions and because differences between minimum wage increases occur by race; Black workers disproportionately live where minimum wages have not increased. This paper decomposes those different effects.
Rachel M.B. Atkins, New York University; Lisa Cook, Michigan State University and National Bureau of Economic Research; Robert Seamans, New York University
Abstract: We assess the role of FinTech firms in loans made through the Paycheck Protection Program. The PPP, created by the U.S. government as a response to the coronavirus pandemic, provides loans to small businesses so they can keep employees on their payroll. We argue that FinTech firms’ reliance on technology rather than relationship-banking approaches used by traditional banks helps to address discrimination in lending, at least in part. Using newly released data on the PPP, we find support for our arguments: While Black-owned businesses received loans that were approximately 50 percent lower than observationally similar White-owned businesses, the effect disappears when FinTechs are allowed to provide loans.
William E. Spriggs, Howard University and AFL-CIO
Abstract: The dominant model of discrimination is to assume either barriers to entry, based on pre-market factors like schooling or distance to job locations, or discrimination in the market is viewed as client-, owner-, or worker-based discrimination. But the case of locomotive firemen in the late 19th and early 20th centuries, as the importance of the railroad grew, presents a more complex model of race and labor market discrimination. In the U.S. South, Blacks played a dominant role because the job of fireman on a steam locomotive was dirty and dangerous and was a servant role to the locomotive’s engineer. Their numbers were too large for White workers, seeking to exclude Blacks, for White railroad owners to agree to their exclusion. However, outside the South, Blacks were effectively barred from the job. This paper explores this complex setting and shows its relevance to understanding discrimination effects.
Yi Geng, District of Columbia Government; Daniel Muhammad, District of Columbia Government; Bradley Hardy, American University
Abstract: On May 6, 2008, Washington, DC passed the Pre-K Enhancement and Expansion Act of 2008 to provide all 3- and 4-year-olds in Washington universal access to high-quality pre-Kindergarten education. By school year 2018–19, around 80 percent of eligible children in Washington were served in a public pre-K program. While the primary goal of universal pre-K program is to invest in the human capital of children that low-income parents are unable to provide, the program is also justified by increasing low-income family pay and maternal labor supply. Using administrative data from the IRS and the District of Columbia, we designed a study to analyze the impact of the DC universal pre-K program on the labor supply of unmarried working mothers using a different-in-differences framework. Our results show that after the establishment of universal pre-K in Washington, single parents tended to work less before the child was eligible for the universal pre-K program and recover to pre-policy when the child was eligible for the program, when compared with earnings before the implementation of the universal pre-K policy and controlling other factors. This seems to imply that the city’s universal pre-K program produced income effects that significantly affected the labor supply for single parents in Washington with younger children eligible for universal pre-K program.
Robert Fairlie, University of California, Santa Cruz; Frank M. Fossen, University of Nevada, Reno
Abstract: Was the $278 billion reboot of the $800 billion Paycheck Protection Program disbursed equitably to minority communities? This paper provides the first analysis of how PPP funds were disbursed to minority communities in the third and final round of the program, which was specifically targeted to underserved and disadvantaged communities. Using administrative microdata on the universe of PPP loans, we find a strong positive relationship between PPP flows, as measured by the number of loans per employer business or loan amounts per employee, and the minority share of the population or businesses in the third round. In contrast, the relationship was negative in the first round of 2020 and less positive in the second round of 2020. We find a stronger positive relationship between minority share and loan numbers or amounts to employer businesses for first draw loans than second draw loans in 2021. The patterns are similar for loan numbers and amounts to nonemployer businesses but with a similarly strong positive relationship with minority share for both first draw and second draw loans. In comparison, in 2020, there was a negative relationship with minority share in the first round and a much weaker positive relationship in the second round. The restarted PPP program that ran from January to May 2021 appears to have been disbursed to minority communities as intended.
Thomas H. Byrne, Boston University; Robynn Cox, University of Southern California; Jamein Cunningham, Cornell University; Benjamin F. Henwood, University of Southern California; Anthony W. Orlando, California State Polytechnic University, Pomona
Abstract: We explore the extent to which income inequality influences crime, police contact, and police-related fatalities. Recent research has shown that income inequality is associated with increases in police expenditures; however, this increase could be a desire to deal with the social ills typically correlated with rising inequality. We combine information on inequality and crime to examine the relationship between inequality and public safety outcomes. Using an instrumental variables approach, we find evidence that income inequality is associated with increases in police expenditures, reaffirming previous research from Boustan et. al (2013). However, counter to the prevailing literature, we find that inequality is negatively related to homicide victimization, crime, arrests, and police killings of civilians.
Harrison Hong, Columbia University; Neng Wang, Columbia University; Jiangmin Xu, Peking University; Jinqiang Yang, Shanghai University of Finance and Economics
Abstract: Heat waves—periods of extreme heat spanning several days—damage regional productivity and are becoming more frequent in the age of climate change. We model their implications for welfare using a continuous-time growth model. The discrete arrival of heat waves leads to downward jumps in regional productivity. Firms can install cooling capital (e.g., air conditioners, refrigeration) which mitigate the fat-tail damage to productivity conditioned on the arrival of a heatwave. We apply our model to U.S. counties from 1960 to 2020. Our model can match regional economic moments, frequency of heatwave arrivals, and conditional loss distributions. We then use our model to generate regional estimates of capital formation, economic growth, and household welfare based on projected heat wave trends across counties.
Karl David Boulware, Wesleyan University; Kenneth N. Kuttner, Williams College
Abstract: A well-established fact is that Blacks (and other historically disadvantaged groups) are disproportionately affected by recessions. However, the reasons for this are not well-understood. The goal of this paper is to measure the high sensitivity of Black unemployment to macro conditions using a unified time series (error correction) framework. We also explore the possible role of occupational mix in contributing to Black workers high employment “beta”. We find that Black (un)employment is much more cyclical than White, which explains why the racial “gap” widens during recessions and narrows during expansions. Moreover, the “catchup” is especially rapid during high-pressure labor markets, and the excess cyclicality of Black unemployment cannot be explained by occupational stratification.