When this seminar on modelling credit risk was being planned, we wondered just how interested people would be in what might be regarded as simply a narrow niche of the financial sector. So we are glad to see so many of you here today. We are presumably not alone in seeing this topic as important. I shall be opening the seminar by explaining why I consider the modelling of credit risk to be, not a narrow concern but, on the contrary, a major component of modern financial operations.
First let me take you back a number of years. In my early encounters with banking theory, one of the basic questions for me—as no doubt for many of you—was what makes banks so special. A simple answer is that banks borrow short and lend long, a strategy that is founded on what banks know about their borrowers’ future ability to pay. The banks’ long-term customer relationships, involving deposits, lending and payment transactions, gave them a unique advantage in this respect. Compared with other institutions, they were in a better position to conduct lending operations.
In recent years, however, there have been many developments which prompt the question whether that is still the case. Extensive deregulation has enabled banks to broaden their operations at the same time as other institutions have had opportunities of moving into traditional banking markets. Now that a Swedish banking permit is no longer tied to the criterion of need, new types of banks have been established with customer relationships that are not necessarily as clear as those of the traditional banks.
These developments, together with the accelerating pace of technological innovations, have fundamentally altered the financial markets. For a number of years now, currency and interest rate contracts are being actively traded over a range of maturities and the prices are updated in real time. The market for interest-bearing securities is attracting more and more issuers. Whereas previously this was primarily a market for government borrowing, it has been opened by degrees to financial enterprises and now also to a growing extent to non-financial companies.
One consequence of this is the emergence of rating companies, mainly engaged in assessing the creditworthiness of players wishing to borrow directly in the market. The credit assessments that banks used to make are now undertaken by independent raters. Another interesting point is that these ratings are paid for by the companies that apply for them. Instead of being absorbed by banks in the process of providing credit, the cost is carried directly by the borrower. The securities that are issued will be priced by the market in the light of the rating company’s assessment.
It is therefore no exaggeration to say that the banks have lost their earlier advantage as regards information about borrowers who are in a position to raise funds directly in the securities market.
A similar conclusion applies to the market for consumer borrowing, though developments here have been somewhat different. Consumer credits and house mortgage loans have become increasingly standardised. With a large stock of customers, the provision of credit can be based on much the same lines as insurance. The calculation of default risks in bank loan portfolios has much in common with how insurance companies manage the risk of accidents and damage.
Under these circumstances it is hardly surprising that credit risk assessments have become more sophisticated. The examples I mentioned make it natural to adapt the assessments so that they are appropriate for a particular category of borrower. It is then important that we and other authorities follow these developments and keep ourselves up to date.
Banks and authorities have been a little hesitant about using credit risk models in certain cases because this might be seen as an excessively mechanical approach to the provision of credit and risk assessments. One of the problems discussed in connection with the reform of the capital adequacy standards is that the models are only applicable to certain categories of borrower, usually those with a credit rating. In my opinion it should not be considered a drawback that banks use certain models for one type of credit and other models for other types, for instance credit scoring for standardised consumer credits. Moreover, it is perfectly possible to combine the models with more traditional credit assessments, using unique information, in the case of loans to small and medium-sized firms.
These developments in the financial sector have gone hand in hand with technical advances that permit a very high rate of innovations. One highly significant innovation in this field is credit derivatives, which make it possible to separate the provision of credit from the credit risk. A bank wishing to establish a long customer relationship can advance credit and then, by selling the associated risk to some other investor, still refrain from taking a risk it prefers to avoid. At the same time, someone with no experience of credit operations and the associated information about counterparties can still take credit risks as a means of altering the level of risk in their total portfolio.
There is a similar trend in securitisation, which in very simple terms can be described as selling credit items in the form of securities. In that credits can be securitised and marketed, while exposure to credit risk can be adjusted by trading in credit derivatives, it can be said that in certain respects credit risk resembles market risk. Exposures can be altered very quickly and market prices can be used to manage the aggregate risk with the aid of models. For those of us who represent public authorities, it is then more difficult to monitor the risk positions of financial institutions by analysing their statistical reports. Even the monthly data that are presented to the authorities can be difficult to penetrate.
In Sweden as well as internationally, these consequences have been recognised in the construction of the capital adequacy standards for market risks by admitting the use of Value-at-Risk models. But the approach to credit risk remains more traditional: each credit is considered separately and capital requirements are set for broad institutional categories. The accelerating pace of changes concerning credit risk makes it important to pay more attention to the management of these risks as well. Oversight and supervision of financial system stability can no longer be practised solely with analyses of a quantitative nature. That work must be augmented with more qualitative assessments of risk management in individual institutions. This has in fact been highlighted in a very satisfactory way in the recent report from the Bank Law Committee.
This, then, is the background to the decision by the Riksbank and the Financial Supervisory Authority to arrange a seminar on the modelling of credit risk. It is our hope that this will generate a broader discussion in Sweden’s financial market concerning these models and their purely practical applications for the banks. This is, of course, an important way for us to improve our knowledge in the field but it is no less important that we thereby demonstrate the Riksbank’s interest in following developments. Any shortcomings will need to be identified and overcome so that full use can be made of the positive effects. Another consequence is that in its own work on analysing systemic stability, the Riksbank will be paying increased attention to risk management in the major banks as a complement to the present analysis.
