Congress enacted The Americans with Disabilities Act of 1990 over the protests of small business advocates who claimed that the ADA would trigger a wave of bankruptcies. Although the profitability of firms may suffer from the costs of ADA compliance, no systematic evidence is available on the ADA’s effect on firms. This paper seeks to determine if the ADA had a measurable impact on the number of retail firms, the entry of new firms, and the exit of existing firms in each market. I first develop a theoretical model of the response of industry dynamics to increases in costs (an extension of Klepper’s (1996) model). I show that increases in marginal and fixed costs may have interesting and non-obvious effects on entry and exit. The restrictions the model places on observed changes in entry and exit allow inference about which components of the ADA raised marginal cost and which raised fixed costs. The same model could easily be adapted to examine the impacts of other forms of cost-increasing regulation or exogenous process innovation on industry dynamics. The data used in the study are counts of business establishments currently operating by county and type of business. While standard count models can be used to investigate changes in the number of firms after the ADA, investigating the ADA’s impacts on entry and exit is more challenging. Backing out the entry and exit rates from the establishment count data is a major econometric contribution of the paper. Applying techniques from queuing theory, we develop the maximum likelihood estimator for a generalized Poisson queuing system based on the available count data. The model incorporates unobserved heterogeneity in and correlation between the entry and exit rates. Identification of the entry and exit rates is secured through the assumption that entry and exit are Poisson stochastic processes, conditional on time-varying covariates and correlated, gamma-distributed mixing terms (i.e., random effects that relax the Markovian assumptions in the model). Although we use techniques drawn from the existing queuing theory literature, the likelihood for the count data is non-trivial to derive and we have not seen the likelihood for this model presented elsewhere. We develop this model here out of necessity, due to the particular limitations of the available data; however, there are many other potential applications for the econometric model. The empirical results imply that the ADA indeed decreased the number of retail firms. There were fewer retail firms after the ADA was passed, and the drop was larger in states in which the ADA was more of a legal innovation, and in states that had more disabled people, more ADA-related lawsuits, and more ADA-related labor complaints. The same conclusions hold when baseline trends for larger establishments (those least vulnerable to the costs imposed by the ADA) are differenced out. Regarding entry and exit, there is also evidence that employment and access discrimination suits imposed real costs on retail stores, encouraging exit. However, the exit of incumbents was partially offset by new entrants, which may imply that stores less able to adapt to the new requirements made room for the entry of stores better able to adapt. So, while the prediction by the pessimists that the ADA would cause firms to fail may be correct, the decline in the number of firms was partially offset by new entry.
