21than the aggregate anti-director

21
than the aggregate anti-director variable) plus the five creditor rights (described in Table 3 and the next section) as dependent variables, and origin dummies and levels of GNP per capita in each country as independent variables. Because within a country specific laws are correlated with each other, the appropriate estimation method is Seemingly Unrelated Regression. We can then perform F-tests of the hypothesis that origin matters for shareholder rights. The results of the regressions are presented in Table 4, and the F-tests in Table 5,
1 The results in Table 4 confirm the findings of Table 2 that civil law, and especially French
civil law, countries have inferior protections of shareholders to those of the common law
countries. Controlling for per capita income does not change this result (we return to per capita income in section 7). The tests in Panel A of Table 5 furthermore show that, even controlling for per capita income levels, legal origin matters for laws protecting shareholders, With reasonable cofidence, we can reject the hypothesis that any given origin has the same laws as the rest of the world, Moreover, the table shows that the two origins that are most different from the others are French civil law and common law. We can also reject the hypotheses that the laws in the German family are the same as those in either the French or the common law fmilies.
2 These formal tests are consistent with the principal finding of this section, namely that
common law countries have the relatively better protections of shareholders, and the French civil law countries have the worst ones5. A minority shareholder in Australia or South Africa can vote
5 Some European legal scholars have objected to this conclusion on the grounds that our selection of variables was biased toward finding common law countries more protective of investors. One variable that these scholars objected to our not using is the mandatory preemptive right to buy new shares given by many European countries to the existing shareholders. Originally, this right did not appear to us to be critical, which is why we did not include it. Mer this criticism, we have analyzed the data on this right, although we did not include the results in the basic tables since this right was selected to favor civil law countries, As it happens, 56 22 by mail, can trade his shares during a shareholders’ meeting, is protected from certain expropriations by directors, and needs to have only 5°/0of share capital to call an extraordinary meeting. A minority shareholder in Italy or Belgium, in contrast, cannot vote by mail, has tis shares blocked during the shareholder meeting, is not protected from expropriation by directors and needs to have 20°/0of share capital to call for an extraordinary meeting. The differences between legal origins come out loud and clear from this analysis of shareholder rights.
4. Creditor Rights.
Conceptually, creditor rights are more complex than shareholder rights. The reason isthat
creditors exercise their power in several ways. Perhaps the most basic creditor right is the right to repossess -- and then liquidate or keep -- collateral when a loan is in default (see Hart 1995). For collateralized loans, the power of creditors against borrowers depends largely on the ease of repossessing collateral. In some countries, law makes it difficult for the lenders to repossess collateral, in part because such repossession leads to liquidation of firms that is viewed as socially undesirable. In these countries, lenders may still have some powers against borrowers, namely their votes in the decisions for how to reorganize the company and pay off the creditors. The creditor rights that experts consider to be essential for debt finance are the rights to repossess collateral and to have a say in reorganization (see Paul Vishny 1994, Aghion, Hart, and Moore percent of common law countries have this rule, compared to 76 percent of French civil law countries (t = 1.36), 50 percent of German civil law countries (t = .23 in the comparison to common law), and 100 percent of Scandinavian countries ( t = -3.69 in the comparison to common law). Thus the principal comparison of common law to French, and even German, civil law families is not materially affected by the inclusion of this variable. We have also checked all our firther results and none are affected by the inclusion of this measure of preemptive shareholder rights.