Asset-Based Benchmarks

The seven types of benchmarks introduced in this section are

• Absolute (including target) return benchmarks.
• Broad market indexes.
• Style indexes.
• Factor-model-based benchmarks.
• Returns-based (Sharpe style analysis) benchmarks.
• Manager universes (peer groups).
• Custom security-based (strategy) benchmarks.

Absolute Return Benchmark

An absolute return benchmark is a minimum target return or a ‘hurdle rate’ for the investment manager. The return may be a stated minimum (e.g., 10%), stated as a spread above a market index (e.g., the Fed Funds Rate + 5%), or determined from actuarial assumptions. An example of an absolute return benchmark is 15% return annually for a private equity investment. Market-neutral long–short equity funds often have absolute return benchmarks.

Broad Market Indices

Broad market indexes are measures of high-level asset class performance, such as the NASDAQ for the overall US Equity Markets, or the MSCI World Index for global developed market equities. Broad market indexes are well known, readily available, and easily understood. The performance of broad market indexes is widely reported in the popular media, and thus easily accessible.

Style Indices

Market indexes have also been more narrowly defined to represent investment styles within asset classes, resulting in style indexes. An investment style is a grouping of investment disciplines that has some predictive power in explaining the future dispersion of returns across portfolio, such as a long/short US Small-Cap Stock style.

Factor-Based Benchmarks

Factor-model-based benchmarks are similar to style indices, except that they can be broken down into smaller components, or drivers of return. Examples of factors include:

• Broad market index returns.
• Industry exposure.
• Financial leverage.

To determine the factor sensitivities, the portfolio's return is regressed against the factors believed to predict returns. The general form of a factor model is:

$$
R_p = a_p + b_1(F_1) + b_2(F_2) \dots b_kF_k + \varepsilon_p $$

Where:

\(R_p\) = The portfolio's periodic return.

\(a_p\) = The “zero-factor” term, which is the expected portfolio return if all factor sensitivities are zero.

\(b_k\) = The sensitivity of portfolio returns to the factor return.

\(F_k\) = Systematic factors responsible for asset returns.

\(\varepsilon_p\) = Residual return due to nonsystematic factors.

The sensitivities \((b_k)\) are then used to predict the return the portfolio should provide for given values of the systematic-risk factors. The four-factor Carhart model is an example of such a benchmark, however, it remains customizable. As an example, if the investment manager believes that interest rates are inversely related to security prices, then the model can incorporate an interest rate factor. If interest rates unexpectedly fall, then security returns can be expected to rise by an amount determined by the security's sensitivity \((b_k)\) to interest rate changes.

Returns-based Benchmarks

Factors for returns-based benchmarks are derived by analyzing returns for various styles of stocks (e.g., small-cap value, small-cap growth, large-cap value, and large-cap growth). The style analysis produces a benchmark of the weighted average of these asset class indexes that best explains or tracks the portfolio's returns.

Manager Universes

A manager universe, or manager peer group, is a broad group of managers with similar investment disciplines. Although not a benchmark, a manager universe allows investors to make comparisons with the performance of other managers following similar investment disciplines. Managers are typically expected to beat the universe's median return.

Peer groups as benchmarks suffer from some significant weaknesses. Although managers within a peer group may all nominally be classified as by a particular style, they may not truly be substitutable for one another. Some may have tilts or constraints that create an investment product very different from that of the median manager.

A manager's ranking within the peer group may change quickly with very small changes in performance, often in response to factors outside of the manager's control: A change in the ranking may be driven not by something he did but by the actions of others in the peer group (e.g., other managers in the peer group may have chosen to overweight a “hot” sector, whereas the target manager is constrained from making similar bets).  

Custom Security-based Benchmarks

Lastly, custom security-based benchmarks are built to better reflect the investment discipline of an investment manager. Such benchmarks are developed through discussions with the manager and analysis of past portfolio exposures. After analyzing the manager's investment process, the benchmark is created by selecting securities and weightings consistent with that process (and any restrictions).

The benchmark is rebalanced on a periodic basis to ensure that it stays consistent with the manager's investment system. The purpose of custom security-based benchmarks is to serve as a reflection of a manager's strategy. Custom security-based benchmarks are appropriate when the manager's strategy cannot be closely compared to a broad market index or style index. These benchmarks are costly to calculate and maintain.

Question

Which of the following is least likely a limitation of using Manager Universes as benchmarks?

1. Differing portfolio constraints.
2. Manager rankings can change quickly.
3. Regulatory limitations.

Solution

The correct answer is C.

The curriculum says nothing about special regulatory burdens related to using a manager universe.

A and B are incorrect. Both differing portfolio constraints between managers' IPS's, and the fact that rankings can change quickly (and often due to no fault of the managers themselves) are mitigating factors that limit the usefulness of manager universes.

Performance Measurement: Learning Module 1: Portfolio Performance Evaluation; Los 1(j) Describe types of asset-based benchmarks