highest information ratio. This portfolio has a tracking error of 160 basis points. If we construct a portfolio that has 38 percent invested passively and 62 percent invested in the optimal blend portfolio, the total portfolio will hit the tracking error target of 100 basis points. Thus, the passive position effectively dilutes the active risk in the optimal blend portfolio without reducing the total portfolio's information ratio. The total portfolio now has an information ratio of 0.87 and an expected excess return of 87 basis points, with 38 percent invested passively, 43 percent invested with structured managers, and 19 percent invested in traditional strategies. Thus, this portfolio clearly takes risk across the spectrum. How does this optimal portfolio compare to the barbell strategy? To achieve a targeted tracking error of 100 basis points in the barbell strategy, the investor would need to allocate 25 percent to the traditional portfolio and 75 percent to the passive portfolio. This portfolio would have an information ratio of 0.60. Moreover, we can easily see that the structured equity allocation comes almost entirely from the passive position: By putting more of the passive assets to work in a structured equity program, the information ratio for the total U.S. equity portfolio increases from 0.60 to 0.87, or almost 45 percent! Table 14.6 summarizes these two examples and provides the strategy split and information ratios for other tracking error targets. This table contrasts these figures with the barbell strategy: The information ratio increases as risk is taken along the active risk spectrum. What is more striking, though, is that for the most part funding for the structured equity position comes out of the passive allocation. So far, our analysis has assumed that excess returns are uncorrected across managers within an active management type, and across active management types. This assumption has been roughly consistent with the observed median correlation, as shown in Table 14.6. What happens to the information ratio if we assume the correlations are higher? For example, suppose the pairwise correlations are close to the first quartile level in Table 14.6. That is, the average excess return correlation among structured managers is 0.25, and the average correlation among traditional managers is 0.35. We'll continue to assume that each prospective manager in each strategy can generate first quartile risk-adjusted performance. In the two-manager structured equity program, the tracking error increases by about 12 percent, going from 152 basis points to 170 basis points. This increase in tracking error reduces the information ratio for the structured portfolio from 0.64 to 0.57. For the traditional equity program, the higher correlations increase the overall tracking error by 44 percent, from 400 basis points (with four managers) to around 575 basis points. As with the structured program, the information ratio declines, going from 0.60 to 0.42. Thus, the larger increase in correlation among traditional managers produces more significant deterioration in their total tracking error and information ratio. Suppose an investor decides to improve the efficiency of the traditional program by doubling the number of managers. The tracking error for the traditional program would fall from 575 to 525 basis points. Correspondingly, the information ratio would increase from 0.42 to 0.46. Thus, the higher correlation of excess returns among traditional managers may produce an incentive to hold more