What Determines Firm Size?

Motivated by theories of the rm, which we classify as \technological" or \organizational", we analyze the determinants of rm size across industries and across countries in a sample of 15 European countries. We nd that, on average, rms facing larger markets are larger. At the industry level, we nd rms in the utility sector are large, perhaps because they enjoy a natural, or o cially sanctioned, monopoly. Capital intensive industries, high wage industries, and industries that do a lot of R&D have larger rms, as do industries that require little external nancing. At the country level, the most salient ndings are that countries with e cient judicial systems have larger rms, and, correcting for institutional development, there is little evidence that richer countries have larger rms. Interestingly, institutional development, such as greater judicial e ciency, seems to be correlated with lower dispersion in rm size within an industry. The e ects of interactions (between an industry's characteristics and a country's environment) on size are perhaps the most novel results in the paper, and are best able to discriminate between theories. As the judicial system improves, the di erence in size between rms in capital intensive industries and rms in industries that use little physical capital diminishes, a nding consistent with \Critical Resource" theories of the rm. Finally, the average size of rms in industries dependent on external nance is larger in countries with better nancial markets, suggesting that nancial constraints limit average rm size. Preliminary version. We thank Eugene Fama, Doug Gollin, Ananth Madhavan, John Matsusaka, Andrei Shleifer, and workshop participants at Chicago, Copenhagen, Norwegian School of Management-BI, and USC for comments. Why is it that a small country like Finland has such large successful rms such as Nokia? This question is not of minor importance. In recent years, a great deal of attention has been paid to the process of economic growth. An interesting aspect of growth is that much of it takes place through the growth in the size of existing organizations. For instance, in the sample of 43 countries they study, Rajan and Zingales (1998a) nd that 2/3rd of the growth in industries over the 1980s comes from the growth in the size of existing establishments, and only 1/3rd from the creation of new ones. What determines the size of economic organizations? Are there any constraints to size and, hence, any potential constraints to growth? Organizational size seems important for various economic phenomena. For example, the work by Gertler and Gilchrist (1994) suggests that small rms account for a disproportionate share of the manufacturing decline that follows the tightening of monetary policy. Size has been found to be an important in uence on stock returns (see Banz (1981)). Similarly, various phenomena in corporate nance the extent to which a rm levers up, the quantity of trade credit it uses, the compensation its top managers get, all seem related to rm size. But what determines rm size? The data we have on the distribution of rm size across industries in 15 European countries are particularly useful for answering this question. These are all fairly well-developed countries, so the minimum conditions for the existence of rms such as a basic respect for property rights, the widespread rule of law, and the educational levels to manage complex hierarchies exist. A number of rst order factors such as war, economic system, or respect for basic property rights that would otherwise a ect rm size are held constant in this sample. This enables us to focus on more subtle economic and institutional factors for which there is some variation across this sample of countries. We also have a large number of industries, and the variation between industries in their use of di erent factors can give us some understanding of the e ects of production technology on rm size. Finally, the interactions between institutional and technological e ects give us perhaps the clearest insights into the determinants of rm size. We start by documenting broad patterns in rm size across industries and countries. We nd that, on average, rms facing larger markets are larger. At the industry level, we nd rms in the utility sector are large, perhaps because they enjoy a natural, or o cially sanctioned, monopoly. Physical capital intensive industries, high wage industries, and industries that do a lot of R&D have larger rms, as do industries that require little external nancing. While For the link between rm size and leverage see the extensive literature cited in Harris and Raviv (1990) or Rajan and Zingales (1995), for size and trade credit see Petersen and Rajan (1997), and for compensation, see Jensen and Murphy (1990), and the literature cited there.