Allocation of Risk Capital in Financial Institutions

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

The starting point in designing a process to allocate capital across units is to identify clearly the key elements that define the process, and precisely:

* the types of risk being measured;

* the architecture of the risk-taking centers and the degree of risk specialization;

* the type of capital allocation process and its degree of centralization-decentralization;

* the measures used to evaluate risk-adjusted performance; and

* the linkages between risk-adjusted performance measures and the reward/punishment system.

The first element--the range of risks for which capital-at-risk (CaR) [8] should be allocated--is fundamental, but it is determined by the development achieved by measurement models. This paper will therefore be devoted to analyzing the remaining issues in greater detail.

II. Identifying Risk-Taking Centers

The initial step in the design of any control system is to define profit centers and then assign revenues and costs to each profit center. When risk-adjusted performance measurement (RAPM) is implemented, the same process occurs, with the difference that risks also have to be allocated. This raises the issue of the optimal amount of risk specialization desirable in a financial institution: it is therefore necessary to analyze the elements in favor of and against risk specialization.

A. Single Risk vs. Multi-Risk Units: The Rationale for Risk Specialization

There are at least three reasons why some degree of specialization is desirable. First, specialization can be expected to produce higher returns, since it implies assigning responsibility for each risk to those people who have proved to be the best in handling it. Moreover, each operator can focus attention on the core risk of his or her activity, without wasting time and resources managing second-level risks. Second, the auditing process is easier and more effective, since there will be no way to cover losses coming from the core business when operators bet on risks that are different from those they are supposed to manage.

Finally, by clearly dividing risks among organizational units, the possibility of inadvertent internal hedges is reduced. This enables the bank to exploit all available capital-at-risk. Consider, for example, the simplified example of a currency trader who has entered an outright trade selling British pounds forward against dollars, and a money market trader investing on the British money market. The position of the currency trader can be split, for risk-measurement purposes, into an investment in dollars financed by a debt in British pounds. Therefore, the currency trader and the money market trader are taking opposite bets on British money market rates: the former will make money if they increase, the latter if they decrease.

Suppose now that each trader has been given a limit based on capital-at-risk: specifically, the bank is able to allocate to them a total amount of $200,000, which is equally divided between the two desks. Even if each trader fully exploits all his risk-taking capacity, the total risk for the bank will be less than the sum of the risks of the individual positions, since at least part of British money market risk is inadvertently offset internally. This is not a problem as far as risk control is concerned, but from a profit-and-loss viewpoint it means that earnings from one position will be offset by losses on the other. As a consequence, capital is not being used efficiently: a part of shareholders' capital that has been allocated to the two desks is providing a zero return. [9]