With the impact of the credit crisis still being felt across the entire banking sector, there is some debate about potentially raising the cap on the amount of loan-loss reserves that can be counted as Tier 2 capital.
Those in favor of increasing the current cap of 1.25% of risk-weighted assets maintain that it would supply an incentive to build up reserves in response to portfolio or macroeconomic deterioration; those opposed contend that such an increase would dilute capital quality.
Though the views on either side are strongly held, the debate begs a fundamental question: What really is the difference between reserves and capital?
From the standpoint of a debtholder (or a depositor) — we intentionally use the term broadly, recognizing that there are different types of debtholders — whose key concern is the bank's solvency, the distinction between reserves and capital should be inconsequential. Such a statement may initially cause surprise, but one must ask why a debtholder evaluating two banks with similar risk profiles should distinguish between a position of 7% capital and 3% reserves versus a position of 9% capital and 1% reserves.
Ultimately, a debtholder should be assessing the likelihood and magnitude of a severe loss and comparing this against all resources available to absorb such a loss and, specifically, before its own position is affected. Such a debtholder should be indifferent to the differences in the two scenarios above because both imply that the bank can absorb losses of 10% before a debtholder loses anything.
To draw an analogy, the position of a debtholder is like that of an investor in the senior tranche of a securitization. Fundamentally, such an investor cares little how many subordinated tranches exist in the securitization, how they are rated or which tranche is subordinate to which. A senior tranche investor cares solely about the amount of loss that can be absorbed by all subordinated tranches and how likely such a loss is to occur, because this would affect the timely repayment of principal and interest.
Similarly, in the case of a bank's debtholders, the "subordinated tranches" are the various loss-absorbing resources that stand between them and the bank's insolvency. First, the components of Tier 2 capital are available, including reserves, then Tier 1 capital. The allocation of these resources to absorb expected loss and unexpected loss is merely a matter of allocation; it does not affect their overall level.
As a side point, many would argue that the bank with 3% reserves is signaling that it has a higher risk profile to begin with. Specifically, the argument is that reserves represent existing but not yet observed losses — that is, what is already "baked into the cake" — or future "expected losses" over some defined time horizon. Though a lot of this depends on the nuances of a bank's approach to reserving, we broadly agree with it.
This now brings up an important but separate (and subtle) point. Existing or expected losses are generally highly correlated with unexpected losses and "tail risk." This implies that reserves (existing or expected losses) should then be highly correlated with capital (unexpected losses) — but they are not.
In reality, as of Dec. 31, reserves and capital were negatively correlated, and in a nontrivial way. Most banks with high reserves had low capital.
This may not be a grave concern for some, but it highlights the fact that debtholders should be suspicious of a situation in which two banks are purported to have the same debt rating (that is, solvency standard) and loss absorbing resources, say 10%, but one has higher reserves and less capital. The bank with higher reserves is likely to be more risky overall, and its debt rating should be reevaluated relative to the bank with lower reserves.
Back to the original point, it appears that the debate over raising the cap on loan-loss reserves that can be counted as Tier 2 capital misses the crucial issue. The discussion's focus should not be balance sheet and income statement allocations but rather the aggregate risk of a bank and how that compares to its aggregate loss-absorbing resources. And these loss-absorbing resources must be limited to those that are tangible and truly available to an institution in a solvency-threatening scenario.
The value of goodwill and deferred tax assets, for example, has a nasty way of disappearing in a crisis. Similarly, the value of certain instruments that can be converted into equity is very small in a crisis. In this light, reserves are as valuable as equity or retained earnings because they are "banked money" on the balance sheet specifically set aside to absorb losses.
Also, more focus should be put on the countercyclical building of loss-absorbing resources based on early warning indicator models that let a bank preemptively detect market softening. For example, macroeconomic models were predicting commercial real estate deterioration as early as 2006 — implying that loss-absorbing resources, whether capital or reserves, should have been built up well before the recession began.
In a nutshell, instead of worrying how to divvy up the pool of loss-absorbing resources into different accounts, time should be spent to determine whether this pool is deep enough in the first place.