Challenges in defining the cost of capital highlighted

The challenges in defining the cost of capital were discussed extensively at SCI's recent Capital Relief Trades Seminar. Panellists agreed that it is a nebulous concept.

"The most naïve view one usually starts with is that the cost of capital is effectively the coupon of the first year divided by the capital relief," said Jeremy Bradley, director at Lloyds. "Yet the problem with this is that it does not capture the full benefits of a transaction and ignores what happens in the years post-closing. The main challenge is to find a metric that captures the future view of portfolios and transaction benefits for a five- to seven-year period."

The next step in the pre-transaction capital analysis is to understand the likely cost of the transaction, which requires an estimate of how investors will look at the transaction, given the asymmetry of information. Gauging investors' risk assessment and value analysis ensures time is only spent on transactions with a realistic chance of achieving cost of capital hurdle rates. In fact, cost of capital should be considered across different scenarios and even as a distribution taking into account the appropriate risk factors.

"Transparent pricing/trading is also extremely helpful," Bradley added. "That allows us to work out what we think investors want to be paid on new trades. Using our optimisation tools, we can then identify before marketing transactions a fairly tight range of feasible outcomes in terms of level and tranching."

According to Mathias Korn, head of financial solutions at Caplantic, the cost of capital should not reflect market assumptions. It is "impossible to say to investors if the expected loss on average is tiny (10bp), but for accounting purposes it is important to know, since there is a difference between 10bp and 20bp per year from a cost of capital perspective. Yet from an accounting point of view, selection of portfolio is crucial to show investors that the transaction is stable and that the assumptions are consistent and possible."

Structural features are also important for RWA management. Korn stated that all transactions should have a replenishment feature, so the bank can avoid additional costs in issuing the next transaction and can keep the current transaction more efficient. Additionally, the thickness of the tranches is important, since if there is a mezzanine piece and it is possible to use certain excess spread, then the excess spread can lower the first loss retention.

However, transactions are not all about RWAs and effective risk transfer is critical, according to Bradley. "If the transaction is driven by a desire to release risk limits, some features would not be particularly helpful; for example, retained first loss or excess spread covering losses."

Nevertheless, everything seems to be about RWAs when it comes to how to price the equity factor into the model. On this issue, stated Bradley, Lloyds maintains a balance sheet model, with a targeted amount of gearing at different levels in the capital stack.

"What is important is the denominator (RWAs). The treasury, in turn, projects the balance sheet through time and discounts it back to today," he explained.

SP