is just under 10%, while it is about 8% for large firms (100 or more employees). Numbers for other countries are similar. Hence, small firms are there to stay. They are not necessarily future large firms (although of course large firms tend to start small), nor are they doomed to disappear quickly. Their presence is not simply due to systematic size differences across industries either, as e.g. Foster et al. (2001) document that within-industry productivity dispersion dominates the productivity dispersion between industries.
In spite of the prevalence of small firms, most recent research attempting to match the firm size distribution has mainly focussed on the right tail of the distribution (see e.g. Luttmer, 2007 and Chatterjee and Rossi-Hansberg, 2012). Indeed, popular models have problems accounting for just how small and persistent small firms can be. For instance, in settings like that of Hopenhayn (1992) and the many models based on it, a fixed cost or a uniform outside option imply a strictly positive minimum firm size. Similarly, Lucas's (1978) seminal model of entrepreneurial choice predicts that only agents from the upper tail of the entrepreneurial ability distribution enter the market. In the data, however, minimum firm size (in terms of employees) is zero, and not all entrepreneurs have high ability. Hence, estimated versions of such models have trouble accounting for small firms and their persistence. While this may not be a problem for all applications, it should be taken into account in others, e.g. the analysis of entry. It may also affect quantification, as available data and model concepts may refer to different groups of firms. Heterogeneity in outside options could solve that problem, as shown below in the model.
Are all entrepreneurs out to pursue some golden opportunity? Despite the fact that many large firms started small, most firms stay small, and yet they persist. In fact,