Understanding the 30 year Decline in Business Dynamism: a General Equilibrium Approach

Recent empirical work has drawn attention to an unmistakable shift in U.S. firm dynamics since the late 1970s. Principally, the entry rate, measured as the share of new employer firms out of all employer firms, declined by nearly 40 percent from 1977 to 2007, even before the impact of the Great Recession. Remarkably, this steady decline occurred relatively uniformly within geographic areas and within relatively narrow industry aggregations. Taking this trend decline or startup deficit as given, Pugsley and Sahin (2014) show that both its direct effect on business formation and its indirect cumulative effect through a shift in the employer age distribution partly explain the emergence of slower employment recoveries with each business cycle. An important question is what explains this apparent decline in business dynamism? Understanding the source or sources of the decline are crucial to understand whether it is an efficient response to technological shifts or escalating misallocation from increases, for example, in the costs of starting or running a business. In this paper we provide the first, to our knowledge, quantitative analysis of the set of factors that may have affected business dynamism. Our results are surprising. Rather than signicant changes in fixed and entry costs on the firm side, we find that a shift in the growth rate of the working age population that begins in the late 1970s and its general equilibrium effect on labor supply drives the bulk of the declines along the entry margin. To reach this conclusion, we consider a variety of potential explanations in partial and general equilibrium. The first set of factors that might shift business formation are related to changes in barriers to entry and changes in fixed and variable operating costs. Changes in laws and regulations, market concentration, education and licensing requirements, and shifts in economies of scale might discourage firm entry by creating higher barriers to enter and/or a higher fixed cost of operating. The second set of factors we consider is related to demographic changes. There are various reasons through which the population growth rate can affect business formation. Most directly, an older population might be associated with a lower rate of business formation in the economy if younger workers are more likely to engage in entrepreneurial activity. Changes in population growth also affect the labor supply, which could have important effects on business formation through a general equilibrium channel. In all of these cases, but especially for the general equilibrium effects it is important to differentiate between the effects on incumbent firms and the effects on potential entrants. Shocks to the labor supply that put downward pressure on wages create incentives for incumbent firms to expand, but they also create opportunities for potential entrants. Any effect on the entrant share will depend on how the shocks to the labor supply are accommodated by expanding incumbents and new firms. We first evaluate the importance of these two sets of factors exploiting pooled cross-state variation in firm entry rates. In particular, we estimate a simple linear regression using OLS and document a strong correlation between the growth rate of working age population and changes in the firm entry rate at the state level, controlling for state and time fixed eects. We also instrument for the variation in the age composition with lagged birthrates and find a startup rate semi-elasticity of roughly 1 to 1.5. Given that the growth rate of working age population went down from around 2% in early 1980s to slightly above 1% by 1990s, this estimate suggests that more than half of the decline in the start-up rate can be explained by the decline in the growth rate of working age population. The second part of our empirical analysis focuses on various regulatory and policy changes, which is still in progress. While the regression based results are informative, they fail to provide us an internally consistent composition of the factors we considered. In addition, it is not possible to have aggregate time-consistent measures of the type of frictions that affect entry costs and fixed costs. To alleviate this problem, the second part of our analysis follows a more structural approach and estimate a variant of the Hopenhayn and Rogerson (1993) model with population growth to evaluate the quantitative importance of different sets of explanations. Since the model has implications for many other measures of firm dynamics, such as average firm size, survival rates, and employment growth rates, it allows for a more complete evaluation of alternative channels. Before we move on to estimating the model, we analyze the time variation along these additional dimensions. Interestingly, we find that despite the gradual decline in the firm entry rate, survival and employment growth rates by firm size remained stable since 1980s. These facts help us discipline the model and measure the contributions of various shifts in the economy to the decline in startup rates. The estimation of the model is still work in progress but we present comparative statics from the model. Our preliminary findings show that the measured declines in business formation are the optimal response to a shift in the growth rate of the population. Our paper is closely related to the emerging literature on the declining dynamism in the U.S. economy. Early work by Reedy and Strom (2012) first called attention to a decline in the aggregate entry rate of new employers. Using more disaggregated data, recent papers by Pugsley and Sahin (2014), Decker, Haltiwanger, Jarmin, and Miranda (2014a), Hathaway and Litan (2014), Gourio, Messer, and Siemer (2014) and Davis and Haltiwanger (2014) all document that these declines in the entry rate are pervasive within geographic areas and relatively narrow industry aggregations. All of these papers have also drawn attention to the relevance of this decline for the ongoing health in labor market. We are the first, to our knowledge, to provide cross-sectional and time-series evidence on the determinants of the declining entry rate. Specifically, we identify the principal role of shifting demographics on the equilibrium quanitity of startup activity, which is also consistent with the predictions of the workhorse model of industry dynamics.