Traditional Number of Number of Manager Manager Structured Traditional Structured Traditional Information Correlation Correlation Managers Managers Allocation Allocation Ratio 0.00 0.00 2 4 55% 45% 0.81 0.25 0.35 2 4 70 30 0.67 0.25 0.35 2 8 70 30 0.71 traditional managers in a portfolio.12 This higher correlation does not mean, however, that investors should allocate more assets to traditional managers. In fact, the opposite is true: When the correlations among traditional managers increase, investors should allocate more assets (i.e., more of the active risk budget) to the structured equity program. We can see the impact on the active risk budget as follows. Suppose an investor has a tracking error target for the overall active program of 200 basis points. When the correlation of excess returns is zero, we determined that a 55/45 blend of structured and traditional managers achieved the target tracking error. This blend has an expected information ratio of 0.81, as shown in Table 14.7. Now let's consider what happens when we assume higher correlations among excess returns. Table 14.7 shows the results. All else being equal, higher correlations mean higher tracking errors and lower information ratios for both active programs. Because the correlation increases more for the traditional program, however, its tracking error also increases more (and its information ratio falls more). Consequently, investors should allocate more assets to the structured program in order to neutralize the impact of higher active risk in the traditional program. In fact, it now takes a 70/30 mix to hit the risk target of 200 basis points. The information ratio for the combined program is now 0.67, which amounts to a decline in expected return of 28 basis points relative to the zero-correlation case. This example highlights the importance of finding managers with independent and uncorrected sources of excess return. Of course, the expected information ratio for the U.S. equity program will also vary with the investor's views about manager performance. Since we have used first-quartile information ratios for both structured and traditional managers, our examples have implicitly assumed skill in manager selection. Suppose that we are less confident in our ability to pick managers, and instead decide to use median information ratios in our analysis. What happens to the mix of passive, structured, and traditional managers? Clearly, the information ratio for the total U.S. equity portfolio will decline at all tracking error levels. Table 14.8 illustrates this point by showing the active re- 12Of course, the diversification benefit of adding more managers must be balanced against the real cost of potentially higher fees. Adding more managers at what are likely to be lower allocations per manager makes it likely that investors will be unable to achieve fee breaks. Selection and monitoring costs are also likely to rise as the investor adds more managers.