Using the National Compensation Survey between 2004-2017, we document substantial cyclical heterogeneity among performance pay and fixed wage jobs. Using within-establishment variation among jobs that are classified on the basis of duties and responsibilities, we find that performance pay jobs respond primarily by adjusting the intensive margin of compensation per worker over the business cycle, whereas fixed wage jobs respond by adjusting the extensive margin of employment. These results are driven by the fact that firms can more easily adjust the strength of incentives in performance pay jobs over the business cycle.
This paper studies the rise of performance pay contracts and their aggregate effects on the labor market. First, using the Panel Study of Income Dynamics and National Longitudinal Survey of Youth, I document three patterns: (i) the share of performance pay workers grew from 15% in 1970 to 50% by 2000, (ii) performance pay workers experience higher earnings levels and growth rates and work longer hours, and (iii) invest more in their on-the-job human capital. These differences persist even when comparing similar jobs in the same establishment using the National Compensation Survey. Second, I build a dynamic Roy model with heterogeneity in performance pay, time-varying probabilities of receiving performance pay, and human capital accumulation. The model is calibrated using simulated method of moments on the NLSY79. Third, I use my model to gauge the role of incentives, the contribution of performance pay to rising earnings inequality, and evaluate a recently proposed counterfactual 73% marginal tax rate.
We use daily survey data from Gallup to assess whether households' beliefs about economic conditions are influenced by surprises in monetary policy announcements. We first provide more general evidence that public confidence in the state of the economy reacts to certain types of macroeconomic news very quickly. Next, we show that surprises to the Federal Funds target rate are among the news that have statistically significant and instantaneous effects on economic confidence. In contrast, surprises about forward guidance and asset purchases do not have similar effects on household beliefs, perhaps because they are less well understood. We document heterogeneity in the responsiveness of sentiment across demographics.
Using newly licensed individual-level data from Gallup between 2008 and 2017, this paper provides microeconomic evidence that sentiments about economic activity played an important role in amplifying and propagating the Great Recession. First, after controlling for aggregate shocks, a 1pp rise in county employment and housing price growth is associated with a 0.30sd and 0.67sd rise in perceptions about the current state of the economy and a 0.12pp and 0.27pp rise in perceptions the economy is improving. Second, exploiting plausibly exogenous variation in the 2016 Presidential election, consumption of non-durable goods grew by 4.2% with a larger 10-12% increase among conservatives. The causal effect of sentiment on consumption is robust to three separate instrumental variables strategies: a state Bartik-like measure of gasoline price shocks, county fluctuations in daily temperature, and exposure to different housing price shocks through social networks. A back-of-the-envelope calculation suggests that the decline in sentiment can account for 34-68% of the decline in consumption during the Great Recession and an additional 14-43 months of delayed recovery.
Using daily consumer survey data, we analyze the transmission of gas prices to consumer beliefs and expectations about the economy. We exploit the high frequency and geographic disaggregation of our dataset to facilitate identification. Consumer sentiment becomes more pessimistic with rising gas prices. This effect is strongest for consumers who lived through the recessionary oil crises in the 1970s, consistent with models of learning from personal experience. For younger respondents, the sensitivity of sentiment to gas prices is stronger for college-educated respondents. We use web search data to study how consumers deliberately acquire information to supplement or substitute for personal experience to interpret economic conditions. Our results also provide an additional amplification mechanism for the macroeconomic effects of energy price changes.
Despite evidence that contract workers are often tightly integrated into organizations' work routines and processes, researchers still do not understand when contractors might be more or less culturally integrated with their organizations. How do the prevailing cultural norms in organizations relate to contract workers' cultural integration? It is unclear whether strong norms are conducive to culturally integrating both regular employees and contractors, workers who may have different cultural preferences, and who differ in status as full versus provisional organizational members. Using data from an employer review website, we examine how norm strength in organizations relates to both contractors' and regular employees' cultural integration. While regular employees exhibit more integration in organizations with stronger norms, we find that norm strength is associated with less integration among contractors, specifically those doing more independent jobs. We also examine whether strong collaborative and strong hierarchical norms appear conducive to integrating contractors versus regular employees. The results have implications for the returns to contracting in organizations, and suggest that organizational cultures face trade-offs in integrating workers across different employment arrangements.
Corporate culture is increasingly important for retention and employee motivation. First, using a new survey tool with PayScale.com, I show that culture is strongly correlated with employee engagement and both firm productivity and occupational skills. Moreover, a value-weighted portfolio of companies with high culture earns a monthly four-factor alpha that is 7.5% higher than their counterparts. Second, to understand the mechanism behind these correlations, I exploit plausibly exogenous variation in employees' outside options and find that employees are willing to give up 1.7% of their annual earnings ($1,159/year) for a standard deviation increase in culture.
The Digital Economy & Cybersecurity
idesharing has become ubiquitous throughout the United States. We exploit quasi-experimental variation in the staggered entry of Uber into different metropolitan areas between 2010 and 2016. Using a combination of housing price data from Zillow and social well-being (SWB) data from Gallup, we find that the entry of UberX into a metropolitan area leads to a 3% rise in median housing prices, a 1.5% rise in economic optimism, and a 0.8% rise in life satisfaction. Our estimates are identified based on within-metropolitan comparisons before versus after their entry into the market, conditional on semi-parametric location-specific controls. Using our hedonic pricing elasticity, a back-of-an-envelope calculation suggests that the entry of Uber is associated with a welfare gain of over $633.5 million.
Do Product Market Reforms Raise Innovation? Evidence from Micro-data Across 12 Countries (with Anastasia Litina and Georgios Tsiachtsiras)
How does policy affect innovation and the digital economy? We revisit a classical question as to how standard product market regulation affects innovation and we develop a novel framework for thinking about digital regulation. Using new establishment-level micro-data across 12 countries between 1998 and 2012, this paper first estimates the effect of competition policy on innovation. We find that a standard deviation rise in product market regulation is associated with a 1.029% decline in innovation activities. These declines are a result of product market regulation on the incentives to invest in in-house R&D and make the appropriate capital acquisitions, as well as of the effects of regulation on the cost of innovation activities. We then theorize on the effect of digital regulation on innovation and we empirically test our hypothesis using a sub-sample of the years in our analysis. We find that “protective regulation” confers a positive effect on innovation, while “restrictive regulation” confers a negative effect on innovation. Thus, contrary to our findings about standard PMR, the digital regulation results are more sensitive to the content of regulation. We attribute this ambiguity to the fact that digital markers require an enhanced level of trust to be operative.
Increasing evidence suggests that technological change will have significant effects on the tasks and interactions in the workplace. Although technological change may displace some jobs, it will also affect employees’ experiences of the jobs that remain and the new ones that are created. First, this paper introduces a new measure of technological change at the county-level by drawing upon measures of the growth in the stock of intellectual property (IP) across industries. Second, we use this new measure, together with proprietary data on millions of employees between 2008 and 2018, to investigate the quantitative effects of technological change on employee attitudes about work and well-being. Our results suggest that technological change is associated with robust positive effects on self-efficacy and well-being. While we find the effect is strongest in workplaces with trust, we find that managers who behave more as a boss, rather than a partner, help workers buffer against technological change.
The expansion of digital technologies has heightened exposure to cybersecurity vulnerabilities, culminating in a number of data breaches for large, publicly-traded companies. Using the universe of online job postings from 2010 to 2019, combined with all publicly reported data breaches for these companies, we investigate how firms change their demand for different skills, particularly cybersecurity related, following the public revelation of a data breach. First, we find an increase in demand for job postings that contain at least some mention of cybersecurity skills. Second, the increase in demand for these skills occurs a quarter before the data breach becomes public, suggesting that companies begin acting on private information over a breach before it becomes public. Third, we show that the increase in cybersecurity skills reflects a broader increase in demand for high-skilled labor that demands a heightened wage premium, potentially due to the negative effect that these breaches have on firm brand.
The Technology Content of Consumption (with Giovanni Gallipoli)
Household Finance (and Subjective Well-being)
Why do firms offer non-wage compensation instead of the equivalent amount in financial compensation? We argue that firms use non-wage benefits, specifically maternity leave, to efficiently target workers with desirable characteristics. Using Glassdoor data, we show that firms offer higher quality maternity benefits in industries and locations where female talent is relatively scarce -- a relationship robust to an instrumental variable analysis. Second, using plausibly exogenous variation in the timing of government policy, we show that these benefits can increase firm value. Third, we document novel stylized facts about non-wage benefits and how they are correlated with firm characteristics.
Rise of the 'Quants' in Financial Services: Regulation and Crowding Out of Routine Jobs (with Alberto Rossi)
This paper documents a rise in high-technology workers in the financial services sector and finds evidence that financial regulation has contributed to this change in labor force composition. First, we document three stylized facts: (a) the share of science, technology, engineering, and math (STEM) workers grew by 3 percentage points between 2005 and 2017, (b) while the earnings premium of working in finance have grown, the STEM premium has declined since 2005, and (c) regulatory restrictions in the financial services sector have grown faster than any other sector over the past two decades. Second, we subsequently show that increases in regulation are associated with increases in STEM jobs and declines in routine and manual jobs. These results are consistent with models with increases in compliance risk raise the returns to automation.
Industrial production in the United States declined by 47 percent between 1929 and 1933. This paper quantifies how the severity of the Great Depression within a location may affect contemporaneous entrepreneurship rates. On one hand, a more severe decline in productivity could have persistent effects that adversely affect entrepreneurship today. On the other hand, a more severe decline could have prompted individuals growing up during the Great Depression to become more entrepreneurial and frugal, thereby influencing the values they emphasized to their children. Consistent with the latter hypothesis, we find that a one percentage point increase in retail sales growth is associated with a 0.04 percentage point decline in entrepreneurship two generations later. Using the Panel Study of Income Dynamics, we explore the role of personal experience and upbringing as a moderating factor, finding that individuals in areas more affected by the Depression are more financially sophisticated.
We exploit the staggered and discontinuous changes in interest rates among adjustable rate mortgages to identify the effects of foreclosures independently of housing prices. First, interest rate resets predict foreclosure, accounting for up to 18% of the change in foreclosures. Second, a 10% rise in foreclosures is associated with a 1.14% and 2.57% decline in non-tradables employment and hiring, respectively, accounting for up to 10% of the decline in the hiring rate between 2006-2011. Third, we introduce a new mechanism independent of housing prices whereby foreclosures reduce local optimism and raise uncertainty, thereby leading to a contraction of credit that affects hiring, especially small businesses.
This paper explores the causal effect of foreclosure on individual well-being and social capital. Using plausibly exogenous variation in the timing of interest rate changes on different types of adjustable rate mortgages (ARMs), we find that a 10% rise in foreclosures is associated with a 0.58% and 0.28% decline in current and expected future life satisfaction. These effects are primarily driven by the effects on local optimism and social capital. The results are consistent with models of spatial externalities where large-scale shocks generate adverse effects on communities, not just individuals.
Over six million households experienced foreclosure during the financial crisis. Where did they move, how did they fare, and why? First, we create a new longitudinal dataset between 2006 and 2011 from households' date of foreclosure to their relocation. Despite significant heterogeneity in mobility outcomes, we find that individuals move to, on average, higher quality locations. However, these locations are sometimes worse than what a household would have chosen at random, on average, within the same state. Second, to investigate the source behind these plausibly suboptimal moves, we quantify the contributions of three different hypotheses: (i) local labor market conditions, (ii) local composition effects, and (iii) state foreclosure institutions towards mobility outcomes. Third, we find that the return to moving to another county following foreclosure, relative to census tracts within-county, is 2.3%. These results suggest that, while individuals who moved are better off for it, labor market frictions play an important role in moderating reallocation over the business cycle.
This paper studies the spatial and time series patterns of religious liberty across countries and estimates its effect on measures of human flourishing. First, while there are significant cross-country differences in religious liberty, it has declined in the past decade across countries, particularly among countries that rank higher in economic freedom. Second, countries with greater religious liberty nonetheless exhibit greater levels of economic freedom, particularly property rights. Third, using micro-data across over 150 countries in the world between 2006 and 2018, increases in religious freedom are associated with robust increases in measures of human flourishing even after controlling for time-invariant characteristics across space and time and a wide array of time-varying country-specific factors, such as economic activity and institutional quality. Fourth, these improvements in well-being are primarily driven by improvements in civil liberties, such as women empowerment and freedom of expression.