Chapter 14

Incorporating Subjective Views

Not everything that can be counted counts and not everything that counts can be counted.

—Albert Einstein

The fictional trustees from Chapter 10 established a strategic allocation to stocks and bonds (70/30) and then actively managed around that with the assistance of two managers. By definition, active management produces active risk. Because active risk is viewed as a cost over the benchmark risk of a passive-only strategy, prudence suggests that the trustees establish an upper bound on the acceptable amount of active risk and allocate that risk across the active positions in the portfolio. This is the basis of risk budgeting.

A strategic asset allocation is a set of weights that target a portfolio's asset mix consistent with some predetermined policy goal. This portfolio generates variable risk and return through time, which require management if the strategy is to remain optimal. Risk management operates on two levels: an allocation across strategic asset classes (70/30) and an allocation within programs (the 80/20 mix as well as the risk at the manager level). Aggregating manager-level risk within and across programs produces an estimate of total risk. It is total risk that is the management objective.

The measure of risk we are interested in is the covariation in asset returns. Risk, in general, is estimated as the portfolio-weighted average of these covariances. With given covariances, risk management reduces to managing a set of portfolio weights. In a passively managed portfolio especially, there is little room for budgeting outside of finding a mean-variance efficient portfolio; all risk is benchmark risk and there is nothing to budget without fundamentally altering the risk-return profile.

A more realistic depiction has investment managers pursuing both passive and active strategies. Because active returns are the difference between the manager's strategy and the benchmark return, active risk is a combination of risk associated with movement in the market as well as the residual return associated with nonmarket moves (the strategy). As such, risk management is now the act of allocating passive and active risk and it is the presence of active risk that makes risk budgeting an interesting problem. Why? The mean of the residual return, after the active return is adjusted for the market exposure, measures the manager's contribution over the purely passive benchmark return, that is, alpha. That is, α = (r_i) – β(r_m) in our simple CAPM. As such, it is uncorrelated with market risk and, hence, there will always be a demand for active risk in the plan as long as the ratio of return to the marginal contribution to active risk across programs is not equal. Recall that the marginal contribution to active risk MCAR is generated by incrementally allocating (cash) to a program (manager). If the ratios of returns to MCAR vary across programs, then risk can be budgeted to exploit these differentials until equilibrium is restored. The allure of active risk suggests that it be allocated much like any other asset in the portfolio. The challenge is do it efficiently.

Active risk is therefore the raison d'être to risk budgeting. The process of risk budgeting can take various forms, such as porting alpha and devising active risk overlays. Portable alpha, for example, is an attempt to exploit differences in the ratios of active returns to active risk. To see this conceptually, consider a case in which we are very confident that a particular small cap manager will generate relatively high active returns. Hiring that manager exposes the fund to active risk, for which there is a positive demand, but the increased small cap market exposure will violate the plan's strategic allocation to size. Independently budgeting a short small cap index position simultaneously with a long large cap index position in futures ports the small cap manager's alpha (active return) to the large cap index. Although he still manages to the small cap benchmark, the long-short futures position effectively eliminates the increased small cap exposure with the large cap index futures position. The plan gets exposure to the desired active risk without violating its strategic allocation.