Seminars

Most poor families don’t have a full-time worker, and anti-poverty policies often focus on incentivizing parents to work more in order to reduce poverty. While previous research has identified on a variety of reasons that parents are unable to get a full-time job--disability, care responsibilities, poor work prospects due to criminal justice involvement or a high local unemployment rate, etc.--we identify a heretofore undocumented reason: hourly service jobs (a dominant and growing sector for non-college workers) that provide unstable schedules. Using a representative sample of hourly service worker parents, we document that cancelling shifts, sending workers home early, and telling them to come in late are all commonplace practices in these jobs, causing workers to frequently lose expected pay. In addition, calling workers in early or keeping them late, calling workers in at the last minute, and changing workers' schedules from week to week are commonplace, meaning that workers must hold many more hours available for work than they get paid for and making it very difficult to combine multiple jobs. We find that workers have to hold on average 70% more hours available for work than they actually work in a given month, and that doing so drives the effective hourly wage for some subgroups of workers below the federal minimum. Holding more hours available lowers the probability of having a second job and lowers the likelihood that a worker's child is enrolled in formal child care or Head Start.

Relying on comprehensive employment records of applicants in the universe of public jobs in Brazil from 1986 until 2017, an institutional reform that reduced discretion in hiring, and novel performance metrics for public administration, we provide the first evidence of how personal connections influence recruitment processes. We find that a reduction in the ability to leverage personal connections affects both the job search behavior of connected individuals and employers’ hiring decisions. Moreover, hires after the reform are of higher quality, have lower separation rates and absenteeism, and are more likely to be promoted. We also show that the decrease in connected hires improves the operational and budgetary efficiency of public sector establishments. These results carry significance in enhancing our understanding of how to bolster state capacity.

We leverage a comprehensive dataset of firm-level AI investments to examine how firm systematic risk changes with the advent of commercial artificial intelligence (AI). Firms that invest more in AI see increases in their systematic risk, measured by equity market beta. A one-standard-deviation increase in firm-level AI investments translates into a 0.05 increase in market beta. This result is robust to the choice of asset pricing model and controlling for changes in leverage. Other technologies, such as IT and robotics, are not predictive of changes in beta. Furthermore, the results cannot be explained by either increased correlation with the tech sector or by simple within-industry concentration: AI-investing firms experience similar increases in comovement in equity returns with their own industry, with the tech sector, and with other industries. This increased systematic risk appears to be benign: AI-investing firms experience reductions in cash flow volatility, and the increases in their market betas concentrate on the upside (during good market conditions). Overall, our findings suggest a novel market implication of new technologies: they do not only affect firm growth but also change firms' risk profiles.

We study the effect of enclave residence on immigrant assimilation in the early twentieth century by comparing matched neighborhoods that were later designated (or not) by the Catholic diocese to receive an ethnic church. In Polish neighborhoods, which organized their cultural life around the church, enclaves grew after church construction, and residents slowed their cultural and economic assimilation, delaying their English ability, taking manual jobs, giving their children more Polish-sounding names and increasing the rate of in-group marriage. Children who grew up in enclave neighborhoods were more likely to stay in the area and hold manual jobs. These patterns are not observed in Italian neighborhoods, which organized cultural life around home-based religious practice rather than the church.

We examine longevity and its transmission across generations using a unique dataset containing about more than 26 million individuals born between 1880 and 1920. We first document new facts about the transmission of longevity using both absolute and relative mobility measures. Absolute mobility was high: between 45 and 55 percent of individuals lived longer than their parents, though it was much higher for women than men. Relative measures on the other hand show substantially less variation across time and subpopulations. The intergenerational correlation in longevity (a measure of persistence rather than mobility) is 0.09 for both sexes – this low correlation is observed across races, education groups, cohorts and birth states. Finally, we document that the intergenerational persistence of longevity is much smaller than the persistence in socio-economic status. Moreover, correlations in longevity and in education are largely independent of each other, suggesting that mobility in wellbeing was larger than measures of income suggest.

In 1900, East and South Asia were extremely patriarchal. Men were revered as high status, while female sacrifice was glorified. By socialising women to marry, obey their in-laws and stay put, Asian families consolidated trusted networks of social cooperation. Since chastity was crucial for family honour, women were also tightly restricted. But, over the 20th century, East Asian women increasingly undertook paid work in the public sphere, forged solidarity and gained status. Growth also catalysed a broader process of cultural liberalisation: autonomy, dating, and divorce. South Asian patriarchy is much more persistent. Intimate partner violence remains normalised. To explain this divergence, I suggest that every patrilocal family faces a trade-off between honour (achieved by social policing) and income (earned by exploiting female labour). East Asian female employment rose because rising wages compensated for honour. East Asian culture also differed: they were exogamous and less concerned about female seclusion.

Open banking (OB) empowers bank customers to share transaction data with fintechs and other banks. 49 countries have adopted OB policies. Consumer trust in fintechs predicts OB policy adoption and adoption spurs investment in fintechs. UK microdata shows that OB enables: i) consumers to access both financial advice and credit; ii) SMEs to establish new fintech lending relationships. In a calibrated model, OB universally improves welfare through entry and product improvements when used for advice. When used for credit, OB promotes entry and competition by reducing adverse selection, but higher prices for costlier or privacy-conscious consumers partially offset these benefits.

Africa has some of the highest rates of unemployment globally. While a growing literature has linked ethnicity and ethnic bias to inefficient outcomes across a range of contexts in Africa, there is limited understanding of their contribution to labor market frictions and unemployment. Using new administrative data on 194,190 applicants and over 1.3 million matches from the largest online job platform in Nigeria, we study the role of ethnicity in matching and firm hiring. We find significant differences in the matching outcomes of applicants by ethnicity. While on average, applicants that share the same ethnicity as hiring managers from ethnic majority groups are more likely to be hired, the effects differ significantly by the applicant's gender. Co-ethnic female applicants are much less likely to be hired by hiring managers, while co-ethnic men are more likely to be hired. Male and female hiring managers exhibit similar hiring behavior by ethnicity. Evidence from an information experiment suggests that providing information on these patterns to hiring managers can dampen these effects. The results provide evidence of the role of social and cultural institutions in maintaining persistent group based inequality within formal labor markets.

We analyze whether an increase in the female share of an occupation causally leads to lower relative wages in that occupation. The literature has documented a negative correlation between the female share and relative wages on the occupation level in different countries and time periods, but causal evidence has so far been missing. We exploit the natural experiment of German reunification as an exogenous shock to female occupational shares. Before reunification, East German women were not only more likely to participate in the labor market than their West German counterparts, but female occupational shares also differed substantially between East and West. Using the share of women among East German potential migrants in an occupation-age cell as an instrument for the actual female share in the West after reunification, we document that an increase in the female share in an occupation causally leads to lower wages in that occupation. This evidence is in line with the “devaluation hypothesis” developed in the sociological literature.