CAPITALMANAGEMENT4YOU.COM

capital management cash - www.capitalmanagement4you.com

Menu


264 RISK BUDGETING decomposition-which may alternatively be described as the portfolio's diversification footprint-is in


line with expectations. What is being measured here is the extent to which the manager is investing capital in accordance with stated policies. This report should be run at the manager level as well as at the consolidated portfolio level to ensure that no undue (i.e., unacceptably large vis a vis budget) concentrations of risk are present. Figure 17.2 shows the largest active exposures and marginal contributions at the industry level. The risk monitor should be able to opine on whether the levels of risk concentration observed are in accordance with manager philosophy. Once again, this report should be run at the manager level as well as at the consolidated portfolio level to ensure that no undue (i.e., unacceptably large vis a vis budget) concentrations of risk are present that might put either a strategy or the overall plan at risk. Is the Risk Forecasting Model Behaving as Predicted? As indicated earlier, the risk forecasting model uses statistical methods to produce a forward-looking estimate of tracking error. Accordingly, the risk monitor is charged with knowing whether the model is producing meaningful estimates of risk. For example, GSAM's PACE tabulates the number of times that a portfolio's actual return is materially different from its risk forecasts. As an example of this test, please refer to Figure 17.3. Note that if the model is behaving as expected, the portfolio's actual returns should exceed the tracking error forecast by approximately one day per month. Over the four months ended April 30, one therefore expects that there should be four occurrences where actual returns exceed forecast. In fact, there are three. The risk monitor can conclude that the model is behaving appropriately over the period. Had this result not been reached, some of the model's assumptions might have needed to be revisited. Note from Figure 17.3 that this technique gives no guidance as to how much the model might underestimate risk in the event that the actual result exceeds forecast. It only explores the frequency with which this result occurs. The risk monitor- PACE Risk Monitor Summary iidiiHl.itru p?) -t.i3-v-o.-uc ;*";,' The largest active exposure {overweight vs. benchmark) is a 4.87% exposure to modia. That 4 ^7% -verwiPight m rrpdia consumes at thy margin, 8 553<o of thtt portfolio => predtd-iti tracking eiroi FIGURE 17.2 Industry-Level Exposures and Marginal Contributions