Level II Portfolio Construction item sets

Marcus Bell is the investment director of Thornbury College Endowment, which supports scholarships and research at a private university. The endowment invests most of its equity allocation in a passive fund that tracks a broad global equity index. The investment committee wants to move part of that allocation to an active manager and has asked Bell to evaluate Sable Ridge Capital, a quantitative equity manager that builds portfolios from a stock-ranking model.

Aisha Rahman, the head of research at Sable Ridge Capital, presents the firm's process to Bell. Each quarter, the model ranks the stocks in the benchmark, and the portfolio managers adjust the active weights in response. Rahman estimates that the model makes the equivalent of 64 independent active decisions a year. Based on more than ten years of live results, she estimates that the correlation between the model's forecast active returns and the realised active returns is 0.04. Bell considers this a credible estimate for a model of this kind, because a small but stable information coefficient applied across many decisions can add value.

Sable Ridge Capital's strategy is long-only and has limits on position size and sector weights. Rahman explains that, because of those constraints, the active weights in the portfolio do not fully reflect the model's forecasts. The firm's analysis shows that the correlation between the active weights the model would choose without constraints and the active weights actually held is 0.60. Rahman says that Sable Ridge Capital also offers a version of the strategy that allows short positions of up to 30% of the portfolio, which Rahman says would reduce the effect of the position limits.

Bell asks Rahman to estimate the strategy's expected information ratio using the fundamental law of active management. He wants to compare it with the ratios of other managers on the endowment's shortlist. The shortlist includes two fundamental managers and one other quantitative manager, all benchmarked to the same global index.

Bell then turns to how much active risk the endowment should take. He assumes that the benchmark has an expected return in excess of the risk-free rate of 8.0% a year and a standard deviation of 16.0% a year. He wants to choose the level of active risk that gives the combined portfolio the highest possible Sharpe ratio, and he understands that this optimal level depends on the manager's information ratio relative to the benchmark's Sharpe ratio. He assumes that active returns are uncorrelated with benchmark returns.

The investment committee also wants to know how much the endowment's Sharpe ratio could improve if the allocation to Sable Ridge Capital were sized correctly. One committee member suggests that the improvement is simply the manager's information ratio added to the benchmark's Sharpe ratio. Bell is not sure this is right and plans to check it.

Finally, Bell considers whether to recommend the long-only version of the strategy or the version that allows short positions. He believes the stock-ranking model and the number of decisions it makes would be the same in both versions, and that only the extent to which forecasts are reflected in portfolio weights would differ. The endowment's investment policy allows short positions in externally managed portfolios with the committee's approval.

Exhibit 1 summarises the inputs Bell will use. He treats the figures Rahman provided as ex ante estimates and plans to compare them with the strategy's realised, ex post information ratio once the endowment has at least three years of its own results.

Bell plans to present a recommendation to the investment committee at its next meeting, together with a summary of the risks of the strategy and an estimate of the fees the endowment would pay.