Working PapersView/Hide Abstract
I study the effect of noncompete agreements (noncompetes) on aggregate productivity growth. I build a dataset matching venture capital funded startups to the previous employers of their founders. It exhibits a statistically and economically significant relationship between incumbent R&D spending and employee startup formation in the same industry (within-industry spinouts). Motivated by this finding, I develop and quantify a general equilibrium model of endogenous growth with R&D-induced within-industry spinouts and noncompetes. In the quantified model, eliminating restrictions on noncompetes increases the annual growth rate by 0.21 percentage points and social welfare by 3.24% in consumption-equivalent terms. A subsidy to R&D spending reduces growth and welfare. Both mechanisms involve shifting the R&D labor allocation between incumbents and entrants. Sufficiently large R&D subsidies targeted at incumbent innovation can substitute entirely for noncompetes without stifling spinout innovation. Growth is maximized with targeted R&D subsidies of 88% and a ban on noncompetes.
I describe a novel mechanism through which an aging population can be a drag to productivity growth. I describe this mechanism using a general equilibrium model which adds a quality ladders endogenous growth model to Chari & Hopenhayn 1991, "Vintage Human Capital, Growth, and the Diffusion of New Technologies" (CH). In the model, young workers acquire technology-specific human capital on the job, eventually becoming old skilled workers. Old workers cannot accumulate skill in additional technologies. Hence, as technology improves, old skilled workers continue to use obsolete technologies. Moreover, when skilled and unskilled labor are complementary, young workers have an incentive to develop human capital in old technologies. The result of these two forces is slower adoption of new technologies and hence a lower value to an R&D firm of a successful invention. While in CH growth is exogenous, in my endogenous growth framework this reduces R&D spending in equilibrium and hence decreases the growth rate of productivity. Moreover, in contrast with CH, the model has a dynamic inefficiency: workers do not internalize the effect of their human capital accumulation decisions on the rate of technology adoption and hence the incentive to conduct R&D. In this context, a rightward shift in the age distribution of the workforce can either increase or decrease growth through its effect on the rate of technology adoption and hence R&D incentives. On the one hand, a longer life span in the old and skilled state incentivizes young workers to develop human capital in newer, and hence more valuable, technologies at the expense of lower wages when young, as they have a longer time during which to profit from being skilled in that technology. The distribution of unskilled employment shifts towards newer technologies. On the other hand, a longer life span means that the average time since an old worker became skilled increases, so more skilled workers use obsolete technologies. To understand the relative strength of these two mechanisms, I calibrate the model to employer-employee data from Germany, in particular disciplining the model with the joint distribution of employee age and the establishment's technological vintage. Using the quantified model, I consider the implications of an aging workforce for the growth rate of productivity. The relationship between population aging and labor productivity growth is an inverse U-shape: the human capital accumulation incentive effect dominates at first, but is eventually superseded by the effect of aging on the average time since an old person developed their human capital.
We analyze the per capita output growth performance since 1960 of countries in Latin America and the Caribbean (LAC). Using a worldwide panel of countries, we first find evidence of unconditional productivity divergence in that countries further from the frontier of output per capita have slower growth in output per capita and TFP. Second, we find that LAC countries have slower TFP growth, factor accumulation, and overall growth in output per capita than other countries with the same output per capita gap relative to the United States. This holds across countries and over time, although the shortfall is smaller during the post-1990 period. Further, TFP growth accounts for more than 3/4 of the overall shortfall in output per capita. We then ask to what extent pre- and post-tax and transfer income inequality, as measured by the market and disposable income Gini coefficients, can account for the LAC productivity growth and factor accumulation shortfalls. We find that the market Gini and the disposable income Gini (i.e., after taxes and transfers) have asymmetric effects on the measured LAC growth shortfall. Controlling for market Gini has little effect on the estimated LAC growth discrepancy in either productivity or factor accumulation. By contrast, controlling for disposable income Gini reduces by over 2/3 the LAC productivity growth shortfall while increasing the LAC shortfall in factor accumulation by a similar magnitude. These findings suggest are consistent with the hypothesis that a relative lack of redistribution in LAC countries can account for both low productivity growth and high factor accumulation compared to countries with similar output per capita.