Allocation of Risk Capital in Financial Institutions

Financial Management (Financial Management Association), Autumn, 1999 by Francesco Saita

The easiest assumption would be that the amount of available CaR is fixed (i.e., CaR supply elasticity is zero). In this case, the amount of capital to be divided among units is not dependent on units' demand, and the only variable to be set is the price (i.e., a RAROC target) to be paid for the assignment of that capital. The higher the RAROC target a unit offers, the greater the amount of capital-at-risk it will receive. In this situation, the effects of excess or insufficient CaR at the corporate level are completely transferred to the individual units. For instance, undercapitalization at the bank-wide level will result in remarkably high RAROC targets (which could in turn make all units accept unrealistic objectives). In this regime, profitability objectives are not set by top management, since they derive from self-imposed targets submitted by risk-taking units. Consequently, each unit's appetite for risk becomes relevant. [11]

The intermediate and most common way to allocate CaR is through negotiation. CaR and RAP target levels are derived through an iterative process where senior management's judgments lead to initial proposals, which are then negotiated with the business units. In this case, RAP targets and capital-at-risk must be negotiated jointly, so that capital increments are provided only to units that accept a commitment to achieve more challenging RAP objectives. Planning and budgeting procedures should therefore be designed to support the key role of the CaR allocation process. At the same time, should the business units attempt to reduce the allocated capital after viewing their RAP targets, this would indicate that corporate headquarters had set an unrealistic initial objective (and should correct it). If properly managed, negotiation can substantially increase efficiency in capital usage and avoid internal market pitfalls. At the same time, a negotiation-based process is not exempt from shortcomings, since CaR alloca tion is based in this case not only on past and expected performance (as when top-down allocation prevails) and on promised targets (as in internal CaR markets), but also on each unit's organizational power (see Table 1). [12] This can slow the decision process or even alter CaR distribution between efficient and inefficient units, although, in the long run, persistent differences in risk-return performance must force reallocation to take place.

Three observations can be added at this point. First, drawing an analogy between the process of CaR allocation throughout the bank and asset allocation involving a diversified portfolio of securities, the following issue arises: what is the best estimate of the expected returns on each business unit in the bank's portfolio? Which matters most: the performance expected by top managers, the "promised" performance, or historical performance? Do divisional managers have better information than corporate headquarters? On the other hand, do they have a tendency to misrepresent potential future earnings, as seems to happen in the formulation of capital budgeting proposals? [13] How persistent is historical performance? At the same time, the analogy suggests that since correlations across businesses' returns matter, a certain level of top management intervention may be desirable.