Sunday, February 17, 2008

Managing Investment Uncertainty - A Fiduciary Requirement

Managing uncertainty is an investor’s primary occupation. In scientific models uncertainty is categorized as either aleatory or epistemic. Aleatory uncertainty refers to random errors or, for our purposes, the random variations inherent in capital markets. “Random walk” in the markets cannot be reduced by evaluating more data points or having better information. Its role can be managed, though not eliminated, by basic statistical concepts of variability, probability and expert judgment. On the other hand, epistemic uncertainty or systematic errors stem from a lack of knowledge, inaccurate or incomplete information or flawed models. Systematic bias, which we refer to as selection bias in the investing process, can be either informational or procedural in nature.

Selection bias can distort the conclusions drawn in investment evaluation based on the way investment data is calculated, presented, evaluated or processed by investors. Selection bias probably accounts for a significant amount of investment under-performance as well as fiduciary risk. The prudent expert standard of care required of investment fiduciaries accommodates uncontrollable randomness in markets. It does not tolerate preventable selection bias. A requisite level of knowledge, information and analytical process necessary to overcome selection bias is required of investment fiduciaries. A sample of selection biases include:

Branding Bias - investors often use brand as a heuristic in their investing decisions since the indicia of investment quality are not readily apparent. Marketing studies show that familiar brands have an advantage in competitive, commodity like markets. Branding reinforces recognition and can become a “proxy for quality” in complex products like investments. Icons for investment quality, such as fund ratings, conspire with product branding to bolster investor’s perception that investment quality & suitability can be evaluated at a glance. Minimal diversification of investment managers in products or investment programs can indicate branding or other selection bias.

Universe Bias – many investors are presented with investment fund universes that have been pre-screened to facilitate their investment selection. In general, the best attainable portfolios in terms of risk/return efficiency are diminished when investment universe constraints are imposed. Our experience suggests that constrained fund universes can reduce diversification benefits and potential alpha. Since optimal investment selection can be subverted by other interests in investment screening, fiduciary investors have an obligation to understand any economic or quantitative selection biases that have been applied to their investment selection universe. Open fund architectures provide a richer palette from which to build efficient portfolios.

Fee Bias - there is significant regulatory and legal activity around fee presentation & disclosure in the 401(k) domain. The prevailing bias has been to bundle investment, administration and intermediary fees into net returns and provide disclosures about the bundled fee components in collateral literature. The majority of investors don't/won't read collateral no matter how simple. What investors will react to are explicit detailed service charges on their quarterly statements. When it is clear to investors that they are purchasing services, they will be more motivated to ensure they and their plan sponsors are purchasing efficiently. In other sectors of the economy such as healthcare, where consumers have choice and can control cost, falling costs and increasing productivity follow.

Benchmarking Bias – benchmarks are an important investment policy tool used to determine relative performance of portfolios and investments. To provide relevant information, a benchmark should be a viable passive investment strategy and it should have similar asset class and styles proportions as the implemented fund/portfolio. Improperly constructed benchmarks lead investors to mistake beta for alpha and under-performance for out-performance. Our research suggests inappropriate benchmarking is a large source of phantom “alpha” and selection bias in investor portfolios. Use of customized style benchmarks, benchmark correlation testing and timing/selection performance attribution are indicative of processes designed to overcome this bias.

Performance Presentation Bias – What investors see depends on what they are shown. Cumulative returns (i.e.1, 3, 5 year) can suffer from return smoothing and “endpoint bias”. Cumulative compound annual growth calculations smooth out returns through the time period. Any significant excess performance creates the illusion that the fund has consistently outperformed and hides characteristics such as return consistency and volatility which should be important to investors in setting ex-ante return expectations. Similarly, the particular time periods and starting points for which investment performance data is shown can drastically alter investor’s perception of fund quality. The ubiquitous presentation of cumulative investment returns can disguise more than it can reveal. Use of rolling return periods, backward graphing, annual period return evaluations and a mosaic of other statistical performance information can help investors overcome data presentation biases.

Fiduciary investors can’t overcome random bias in the capital markets. However, a prudent expert standard of care compels them to overcome systematic uncertainty, such as selection bias, in their investment decision making. Selection bias is rooted in low levels of investor awareness and knowledge. It can be propagated through behavioral manipulation and statistical artifices which pander to investor’s desire for simplicity. Evidence of selection bias is obvious to a prudent expert and therefore represents a source of risk to investment fiduciaries. Mitigating this risk can be as simple as opting out of the selection process by utilizing a passive investment strategy or investing resources in improving the investment selection process.

Links to this post:

Create a Link

<< Home