Pierre-Cyrille Hautcoeur, Rutgers University and CNRS, Paris

This paper presents and tries to solve the paradox represented by the development of the French capital market in a context of high agency costs. The first part of the paper presents a new description of the development of the French capital market from 1890 to 1939 on the basis of new individual date on the issues realized by all the stock-market-listed firms (around 1000 firms). It shows that the stock market not only broadened by listing more firms, but also deepened since firms issued more stocks and bonds, and these issues represented a larger share of their total financial resources. We turn thereafter to a description of the agency costs resulting both from asymmetric information and a very insufficient legal and regulatory framework. Since the mere idea of agency costs is based on the expectations of the agents participating the market, we base our description of these agency costs on the legal flaws observed most frequently by contemporaries. This provides some indicators that may be consider (both for contemporaries and for us) as good proxies for agency costs. The third part of the paper consists in a quantitative estimate of the influence of agency costs on French firms finance both at an aggregate level and at the individual level of all listed firms. We try to measure for different periods the effect of the indicators of agency costs described previously on such variables as the relative shares of market vs. internal finance or the share of bonds relative to stocks in market finance. We show that the influence of agency costs is effective since firms presenting signs of a higher risk for savers are not able to issue securities as easily as the others. On the other hand, it is difficult to find any sign of an influence of agency costs at an aggregate level, whether comparing different industries for the same period or comparing different periods. This suggests that a firmÕs behavior vis-ˆ-vis its shareholders had no externalities on similar firms, and that the market was able to select ÒgoodÓ firm. A last part of the paper tries to explain why an important number of firms chose to adopt agency costs creating behavior even being listed on the stock-exchange, and why the effect of agency costs proxies is occasionally limited. We show that agency costs creating behaviors could result from the pursuit of family control preservation, and that outside shareholders didnÕt leave either because of the protection represented by the firmÕs commitment to distribute high dividends or because of the protection represented by high liquidity in the market for their securities. We conclude that the actual functioning of the stock market during this period was more complex than usually recognized but not less efficient. The most important consequence of these features is that the stock market functioning or the savers participating it cannot be made responsible for the low investment rate and the limited size of French firms of this period relatively to those of other rich countries.