Competition in Credit Markets

Efficient financial markets have. to carry out several functions. Similar to other markets, efficiency requires that capital is provided at a price close to marginal costs. However, unlike in usual product markets, the allocation of capital is not only determined its its price. Banks choose her customers and play therefore an active role in allocating savings among alternative investment uses. Welfare is only maximized if they direct. funds to the firms with the highest leturns.

In this chapter, the incentives to acquire costly information in a sealed bid common value model of credit market competition will be analyzed. The existing work is generalized in two directions:

So far, all existing auction models of bank competition have used a sealed bid framework. However, the assumption that bank competition is a sealed bid auction lies that banks are only allowed to make one off, and do not receive any information about the rival’s bid. For large firms this is not vent realistic. In reality, the interest rate is determined o a process involving repeated bidding and information revelation. The precise features of this bidding process depend very much on the specific case. However, there are several reasons to assume that in general this process is rather an open than a sealed bidding. Banks certainly have incentives to learn the offer of another bank and seem sometimes to be able to do that. The firms are interested in open bidding as well. It seems to be common that firms use the credit offer of one bank to incite another bank to undercsa. They can credibly communicate the offer for example by showing the credit comma. In fact, they do not even need to (lo so. We also obtain an open bid setting when a hank, realizing that its credit offer will not be accepted, makes a second offer. Of course in order to obtain the perfect analogy of an open bid auction, firms would have to go back and forth from one hank to the other infinitely often, which is definitely not realistic. Is fact, firms also switch par of their bank loam quite frequently from one hank to another, depending on the offers and on the outcomes of the negotiations with the Hulks they do business with. In the case of as firms, this may be less frequent than N the case of larger firms. The case of switching infinitely often is thus a limiting case.

The participation of hanks in the equity capital of a firm is one of the most controversial issues in banking regIdation. The maximum amount of capital that a bank is allowed to hold in a non-financial firm riders widely between countries. Usually two types of restrictions can he found: limitations of the absolute level of equity investment in relation to a book’s capital and absolute limits on the participation in a single company. Whereas the first kind d restrictions normally only relevant for very big firms, direct limitations of link ownership aa, a severe limitation of the banks scope of action. For example in Denmark and Norway banks may own up to 50% of a firm’s voting stock, n Portugal 25%, Finland or Ireland 10%, and in Belgium, Japan and Sweden op to 5%. Since Italy has recently abandoned ownership Australia is now the only OECD which does not allow any investment in equity of non-financial firms. Similar to Germany, Austria, Greece, Spain and Turkey have no general restrictions on the percentage of a firm that a bank may own (OECD 1992).

The simple model of monitored finance which is constructed in the first part of this chapter helps to explain several features of a bank-firm relationship. However, the principal intention of this chapter is to give a new explanation for credit rationing by banks. The attitude of hanks with respect to credit risk has puzzled economists for a long time The interest rates banks quote do not differ very much among firms of different riskiness. For example Machauer and Weber (1998) analyze the dependence of interest rates on a bank’s own internal rating of borrowers. In their sample, the average interest rate difference between borrowers in the hest and the worst of five risk classes is only 1.2%. Given that t ire best class is defined as “good or very good creditors” whereas the worst. class consists of borrowers which are “very much in danger of default” these interest differences are much smaller than the interest rate spreads on corporate bonds of comparable r ratings. In addition, the results of Petersen and Rajah (1994) or Blackwell and Winter (1997) seem to indicate that the variations of the interest rate can rattier be explained with differences in bargaining power or credit market competition than with the riskiness of a loan.