【正文】 he three country studies confirm this result. Guiso shows that the financial debt of small Italian firms in proportion to their total assets is substantially lower than for large Italian firms. Guiso carefully explains that this difference is because many small firms do not have any loans outstanding at financial institutions. Indeed, conditional on having financial debt, the financial debt ratio and the maturity structure of financial debt are broadly similar across size classes. In sharp contrast with the Italian case, Hommel and Schneider find that the Mittelstand (. German small and mediumsized enterprises) is much more indebted than large German firms. Two thirds of German firms operate with an equity ratio lower than 20 percent, and 41 percent of German firms report equity ratios below 10 percent. This pares to a European average equity ratio of around onethird (see Wagenvoort). Dietsch finds a similar equity ratio for French panies regardless of their size. Overall, while the average European, French, and Italian SME does not appear to be undercapitalised, German SMEs are. Wagenvoort also analyses how firms’ capital structure changes over time. He finds that the dynamics of the financial debt ratio are very different for the average firm in the small and medium size classes in parison to the average firm in the large and very large size classes. More specifically, SMEs appear to be less flexible than larger firms in adjusting the structure of their balance sheets to changing growth opportunities. In particular, the financial debt ratio increases (falls) at a slower rate in growing (shrinking) small firms than in growing (shrinking) large firms. Our interpretation of this result is that small firms have less flexibility in adjusting financial debt in response to changing growth conditions. 3. Finance constraints Is this lack of flexibility due to credit rationing? The three country case studies draw a firm conclusion: SME credit rationing is not a widespread phenomenon in Italy, France, and Germany. Guiso builds a model that can explain why some small firms carry financial debt whereas others do not. The empirical results show that those firms without bank loans are often the ones that finance a relatively high proportion of their assets with equity. Guiso argues that a negative relationship between the equity ratio and the probability of carrying financial debt stands in sharp conflict with the rationing hypothesis since a credit rationed firm is unlikely to substitute equity for financial debt. The absence of financial debt on the balance sheet of many Italian firms is thus mainly because they do not want to borrow, not because lenders do not want to lend. However, Guiso finds that when credit constraints are binding, size and lack of equity seem to play a key role. So, credit rationing happens more often with smaller firms than with larger firms. Dietsch observes that, except for very small French firms with an annual turnover of less than EUR 2 million, French SMEs do not increase bank borrowing when their credit status improves. In contrast with small and mediumsized firms, very small firms with a solid credit standing do raise more loans than their peers of equal size but lower credit standing. In light of this, Dietsch concludes that credit rationing is only relevant for very small firms with unfavourable credit ratings, and he shows that relatively few firms in France have these characteristics. Hommel and Schneider argue that the virtual standstill of credit growth in Germany in 2002 can mainly be attributed to the current cyclical downturn of the German economy. Whether, in addition, the Mittelstand suffers from structural adverse supplyside effects remains to be determined. However, given the large equity gap in German panies, lack of equity is the main finance constraint and additional debt does not seem to be the optimal way forward in Germany. A few qualifying remarks are worth making. One needs to bear in mind that the Stiglitz and Weiss definition of credit constraints implies that a firm is only considered to be rationed if lenders reject the demand for loans although the borrower is willing to pay the going interest rate (and to meet other conditions) on equivalent loans made to others borrowers of the same quality. In other words, according to this definition a firm is not considered credit rationed if it does not want to borrow at the requested interest rate even when the conditions imposed by the bank are too demanding relative to the true creditworthiness of the borrower. In this respect it is worthwhile observing that interest rates on bank loans are in general substantially higher for SMEs than for large Both the empirical findings of Dietsch and Wagenvoort suggest that from a portfolio credit risk viewpoint this may not be justified. It is true that on an individual basis smaller firms are riskier than larger firms because the expected default probability is negatively related to firm size. Banks in general use this argument to defend a higher risk premium on small business loans. But a portfolio of loans to small firms is not necessarily riskier than a portfolio of loans to large panies. Dietsch finds that default correlations are lower within the group of S