In acquiring assets, banks should use the pricing mechanism in conjunction with product/ geography/ industry/ tenor limits. For example, if a bank believes that construction loans for commercial complexes are unattractive from a portfolio perspective, it can raise the price of these loans to a level that will act as a disincentive to borrowers. This is an instance of marginal cost pricing - the notion that the price of an asset should compensate the institution for its marginal cost as measured on a risk-adjusted basis. Marginal cost pricing may not always work. A bank may have idle capacity and capital that has not been deployed. While such an institution clearly would not want to make a loan at a negative spread, it would probably view even a small positive spread as worthwhile as long as the added risk was acceptable.

Institutions tend to book unattractively priced loans when they are unable to allocate their cost base with clarity or to make fine differentiations of their risks. If a bank cannot allocate its costs, then it will make no distinction between the cost of lending to borrowers that require little analysis and the cost of lending to borrowers that require a considerable amount of review and follow up. Similarly, if the spread is tied to a too coarsely graded risk rating system (one, for example, with just four grades) then it is more difficult to differentiate among risks when pricing than if the risk rating is graduated over a larger scale with, say, 15 grades.

A cost-plus-profit pricing strategy will work in the short run, but in the long run borrowers will balk and start looking for alternatives. Cost-plus-profit pricing will also work when a bank has some flexibility to compete on an array of services rather than exclusively on price. The difficulties with pricing are greater in markets where the lender is a price taker rather than a price leader.

The pricing is based on the borrower's risk rating, tenor, collateral, guarantees, historic loan loss rates, and covenants. A capital charge is applied based on a hurdle rate and a capital ratio1. Using these assumptions, the rate to be charged for a loan to a customer with a given rating could be calculated.

This relatively simple approach to credit pricing works well as long as the assumptions are correct - especially those about the borrower's credit quality. This method is used in many banks today. The main drawbacksof this method are:

Only 'expected losses' are linked to the borrower's credit quality. The capital charge based on the volatility of losses in the credit risk category may also be too small. If the loan were to default, the loss would have to be made up from income from non-defaulting loans.

It implicitly assumes only two possible states for a loan: default or no default. It does not model the credit risk premium or discount resulting from improvement or decline in the borrower's financial condition, which is meaningful only if the asset may be repriced or sold at par.

Banks have long struggled to find the best ways of allocating capital in a manner consistent with the risks taken. They have found it difficult to come up with a consistent and credible way of allocating capital for such varying sources of revenue as loan commitments, revolving lines of credit (which have no maturity), and secured versus unsecured lending. The different approaches for allocating capital are as under:

One approach is to allocate capital to business units based on their asset size. Although it is true that a larger portfolio will have larger losses, this approach also means that the business unit is forced to employ all the capital allocated to it. Moreover, this method treats all risks alike.

Another approach is to use the regulatory (risk-adjusted) capital as the allocated capital. The problem with this approach is that regulatory capital may or may not reflect the true risk of a business. For example, for regulatory purposes, a loan to a AAA rated customer requires the same amount of capital per Rupees lent as one to a small business.

Yet another approach is to use unexpected losses in a sub-portfolio (standard deviation of the annual losses taken over time) as a proxy for capital to be allocated. The problem with this approach is that it ignores default correlations across sub-portfolios. The volatility of a sub-portfolio may in fact dampen the volatility of the institution's portfolio, so pricing decisions based on the volatility of the sub-portfolio may not be optimal. In practical terms, this means that one line of business within a lending institution may sometimes subsidize another.

As it became clearer that banks needed to add an appropriate capital charge in the pricing process, the concept of risk adjusting the return or risk adjusting the capital arose. The value-producing capacity of an asset (or a business) is expressed as a ratio that allows comparisons to be made between assets (or businesses) of varying sizes and risk characteristics. The ratio is based either on the size of the asset or the size of the capital allocated to it. When an institution can observe asset prices directly (and/ or infer risk from observable asset prices) then it can determine how much capital to hold based on the volatility of the asset. This is the essence of the mark-to-market concept. If the capital to be held is excessive relative to the total return that would be earned from the asset, then the bank will not acquire it. If the asset is already in the bank's portfolio, it will be sold. The availability of a liquid market to buy and sell these assets is a precondition for this approach. When banks talk about asset concentration and correlation, the question of capital allocation is always in the background because it is allocated capital that absorbs the potential consequences (unexpected losses) resulting from such concentration and correlation causes.

RAROC allocates a capital charge to a transaction or a line of business at an amount equal to the maximum expected loss (at a 99% confidence level) over one year on an after-tax basis. As may be expected, the higher the volatility of the returns, the more capital is allocated. The higher capital allocation means that the transaction has to generate cash flows large enough to offset the volatility of returns, which results from the credit risk, market risk, and other risks taken. The RAROC process estimates the asset value that may prevail in the worst-case scenario and then equates the capital cushion to be provided for the potential loss.

RAROC is an improvement over the traditional approach in that it allows one to compare two businesses with different risk (volatility of returns) profiles. A transaction may give a higher return but at a higher risk. Using a hurdle rate (expected rate of return), a lender can also use the RAROC principle to set the target pricing on a relationship or a transaction. Although not all assets have market price distribution, RAROC is a first step toward examining an institution's entire balance sheet on a mark-to-market basis - if only to understand the risk-return trade-offs that have been made.

¹The hurdle rate is defined as the minimum acceptable return on a business activity.