With Jiwon Choi and Leah Boustan, Review of Economics and Statistics, accepted.

Abstract: The diffusion of computerized machine tools in the mid-20th century was a pivotal step in the century-long process of factory automation. We build a novel measure of exposure to computer numerical control (CNC) using initial variation in tool types across industries and differential shifts toward CNC by type. Industries more exposed to CNC from 1970-2007 increased labor productivity and reduced production employment. Workers in more exposed labor markets adjusted by shifting from metal to non-metal manufacturing. Union members were shielded from this job loss, and some workers returned to school to retrain.

With Scott Shane. Management Science. 2018.

Abstract: Accredited investors finance more than 75,000 U.S. startups annually. We explain how training aspiring entrepreneurs to pitch their new business ideas to these investors affects their odds of continued funding discussions. We model accredited investors’ decision to continue investigation as a real option whose value is a function of their experience and the information contained in the entrepreneurs’ pitches. We derive four hypotheses from the model, which we test through a field experiment that randomly assigns pitch training at four elevator pitch competitions. The data support all four hypotheses and are inconsistent with alternative explanations.

Full paper

With Asim Ijaz Khwaja and Michael Kremer. Quarterly Journal of Economics. 2009.

Abstract: We estimate the impact on pilgrims of performing the Hajj pilgrimage to Mecca. Our method compares successful and unsuccessful applicants in a lottery used by Pakistan to allocate Hajj visas. Pilgrim accounts stress that the Hajj leads to a feeling of unity with fellow Muslims, but outsiders have sometimes feared that this could be accompanied by antipathy toward non-Muslims. We find that participation in the Hajj increases observance of global Islamic practices, such as prayer and fasting, while decreasing participation in localized practices and beliefs, such as the use of amulets and dowry. It increases belief in equality and harmony among ethnic groups and Islamic sects and leads to more favorable attitudes toward women, including greater acceptance of female education and employment. Increased unity within the Islamic world is not accompanied by antipathy toward non-Muslims. Instead, Hajjis show increased belief in peace, and in equality and harmony among adherents of different religions. The evidence suggests that these changes are likely due to exposure to and interaction with Hajjis from around the world, rather than to a changed social role of pilgrims upon return.

Full paper

The Economic Journal. 2017.

Abstract: Scholars have conjectured that the return to speaking a language increases with the number of speakers. Long‐run economic and political integration would accentuate this advantage, increasing the population share of the largest languages. I show that, to the contrary, language size and growth are uncorrelated except for very small languages (< 35,000 speakers). I develop a model of local language coordination over a network. The steady‐state distribution of language sizes follows a power law and precisely fits the empirical size distribution of languages with ≥ 35,000 speakers. Simulations suggest the extinction of 40% of languages with < 35,000 speakers within 100 years.

Full paper

With Roman Sheremeta. Experimental Economics, 2018

Abstract: Some economists argue that consumption of publicly visible goods is driven by social status. Making a causal inference about this claim is difficult with observational data. We conduct an experiment in which we vary both whether a purchase of a physical product is publicly visible or kept private and whether the income used for purchase is linked to social status or randomly assigned. Making consumption choices visible leads to a large increase in demand when income is linked to status, but not otherwise. We investigate the characteristics that mediate this effect and estimate its impact on welfare.

Full paper

Journal of Human Resources. 2014.

Abstract: Bilingualism is a distinct and important form of human capital in linguistically diverse countries. When communication among workers increases productivity, there can be economic incentives to learn a second language. I study how the growth of industrial employment increased bilingualism in India between 1931 and 1961. During that period, Indian factories were linguistically mixed. I exploit industrial clustering and sectoral demand growth for identification. The effect on bilingualism was strongest in import-competing districts and among local linguistic minorities. Bilingualism was mainly the result of learning, rather than than migration or assimilation, and was not a byproduct of becoming literate. My results shed new light on human capital investment in developing economies and on the long-run evolution of languages and cultures.

Full paper

With Jiwon Choi and Leah Boustan. In The Economic History of American Inequality, eds. Martha Bailey, Leah Boustan, and William Collins. University of Chicago Press, 2025

Summary: This paper examines the political consequences of computer numerical control (CNC) automation in metal manufacturing from 1970–1990. CNC significantly reduced manufacturing employment, particularly among less-educated and non-union workers. Workers in industries highly exposed to CNC reported lower institutional confidence, especially in the South, where unionization was weaker. CNC exposure also reduced Democratic Party identification among affected workers and shifted Southern votes toward third-party candidates, while depressing overall voter turnout. The results highlight technological displacement as an important driver of declining institutional trust and partisan realignment.

Full paper

With Moran Cerf, Will Drover, and Scott Shane. Journal of Business Venturing. 2020.

Abstract: We explore how variation in entrepreneurs' displayed passion affects informal investor interest in start-up ventures by examining neural responses to entrepreneurs' pitches using functional Magnetic Resonance Imaging (fMRI). We find that founders displaying high passion increase investor neural engagement by 39% and investor interest in the venture by 26% over those displaying low passion. A one standard deviation increase in neural engagement is associated with an 8% percent increase in investors' interest in investing in a start-up company relative to the mean. Moreover, our findings indicate that neural engagement may account for some of the effect of founder passion on investor interest. Our study has implications for both research on, and the practice of, entrepreneurship.

Full paper

With Jeffrey Williamson. Explorations in Economic History. 2008.

Abstract: India was a major player in the world export market for textiles in the early 18th century, but by the middle of the 19th century it had lost all of its export market and much of its domestic market, primarily to Britain. The ensuing deindustrialization was greatest c1750–c1860. We ask how much of India’s deindustrialization was due to local supply-side forces—such as political fragmentation and a rising incidence of drought, and how much to world price shocks. An open, three-sector neo-Ricardian model organizes our thinking and a new relative price database implements the empirical analysis. We find local supply side forces were important from as early as 1700. We then assess the size of Indian deindustrialization in comparison with other parts of the periphery.

Full paper

With Will Drover and Scott Shane. Small Business Economics. 2023.

Abstract: With the rise of accelerators, angel groups, and business plan competitions, pitching has become an important step for most entrepreneurs raising capital. In this exploratory study, we investigate the effects of pitch training, exploring a variety of outcomes over two time horizons. We conducted a field experiment that randomly assigned 271 would-be entrepreneurs at four elevator pitch competitions to receive one of four pitch training treatments or a null treatment. We observe that pitch training — when received the day of the competition — leads entrepreneurs to improve their pitches, although it causes short-term disruption to pitch delivery. Over the following 30 months, all varieties of pitch training cause entrepreneurs to work more on their pitches, to participate in more business plan competitions and accelerator programs, and to engage in entrepreneurial learning beyond the pitch itself. Entrepreneurs who receive pitch training also are less likely to have employees and are more likely to abandon their initial ventures and founder roles. We discuss the implications of these exploratory observations for the development of theory about pitch training.

Full paper

With Mark Conley and Scott Shane. Journal of Innovation Economics and Management. 2022.

Abstract: The rise of business accelerators, angel groups, and startup competitions has meant that founders increasingly pitch their businesses to investors in group settings, raising the question of whether the order in which ideas are pitched affects outcomes. We test in a field experiment whether range-frequency theory or the theory of bounded rationality better predicts the effect of serial position on pitch outcomes. We find that range frequency theory better predicts the empirical patterns than the theory of bounded rationality.

Full paper

Journal of the Economic Science Association. 2019

Abstract: Neoclassical theory holds that different sources of income are fungible at the margin. In contrast, mental accounting holds that appropriate uses for income vary by source, making them infungible. This study investigates which theory better describes giving at the margin when income may have multiple sources. Dictators accrue differing amounts of (1) earned income from a real-effort task, (2) windfall income, or (3) both. I find that dictators treat marginal earned and windfall income as partially infungible, supporting mental accounting. Dictators who had a single income source gave 14% of a marginal windfall token and 5% of a marginal earned token. Strikingly, dictators who had income from both sources were sharply less generous with each, giving only 2% and 1%, respectively. Multiple accounts enabled greater selfishness at the margin. A follow-up experiment shows that two accounts must be qualitatively different, not just multiple in number, to produce this effect.

Full paper

Journal of Economic Behavior and Organization. 2016.

Abstract: I use instrumental variables to estimate the causal effect of family income on the frequency with which individuals experience negative emotions. Doubling income reduces the experience of negative emotions by 0.26 SD on average. Percentage changes in income have a constant effect on negative emotion for family incomes below $80,000. Above $80,000, the effect of percentage changes declines, reaching zero at $200,000. A dollar change in family income has an eight times larger effect at the 20th percentile of income than the 80th percentile. Effects of income are similar on the high levels of negative emotion characteristic of mental illness, except there is no satiation.

Full paper

With Scott Shane. Fordham Law Review. 2016.

Abstract: This Article reviews the empirical literature on the effects of individual income tax policy on entrepreneurship. We find no evidence of consensus, even on relatively narrow questions such as whether individual income tax rates deter or encourage entrepreneurial entry. We believe the absence of consensus reflects both the complexity of mechanisms connecting tax policy to entrepreneurial decision making and the infeasibility of employing the most reliable empirical methods, such as experiments, in this domain.

Full paper