Investment Approaches For 2005 And Beyond
By: Wendy Brodkin
With the introduction of prudent person investing and the associated requirement for an investment policy, the pension industry developed standard practices for investing that haven’t really evolved over the years. Although the industry has become more quantitative and its glossary of terms has expanded, the substance of best practices is the same as it was 20 years ago. For example, the investment policy is stated as an asset mix of weighted market exposures, recently renamed the ‘beta policy.’ The policy is then implemented in a variety of ways including the use of active and passive management, ‘style boxes,’ and managers with varied investment decision- making processes in terms of the nature, frequency, and concentration of the portfolio bets (recently renamed the ‘alpha implementation’). This investment program is then monitored for compliance with the policy and the managers are monitored relative to their mandates.
These practices are deeply engrained despite the fact they have been less than effective for the past 20 years. Consider the ’80s and ’90s when the markets were so strong you made double-digit returns without any effort at all. The impact of the policy and active managers was negligible at best. The practices were similarly ineffective in protecting the investments during the interest rate declines and equity bear market of the early part of this century. In fact, they provided a false sense of comfort based on the diversification by asset class and by manager controlled risk. This resulted in nothing less than a funding crisis for many pension plans.
The pension crisis did result in some changes to industry practices. It drew attention to the shorter term period covered by the actuarial valuations and accounting principles. Before, the focus was on the duration of the liabilities, which are typically more than 10 years, and the life span of the plan which is intended to last for generations. No one considered that shortterm volatility mattered at all. However, the long- versus short-term issue has yet to be reconciled in investment policies.
Risk budgeting was also recently incorporated to address the issue of proper allocation of capital. The industry recognized that diversification alone was not enough and more often than not the result of the effort was a ‘closet index fund’ because of over-diversification. Risk monitoring was also added to the performance monitoring exercise, using value-at-risk techniques, assessing holding level data (the underlying securities), and monitoring more frequently the financial situation of the assets relative to the liabilities.
However, these changes did not go far enough. Pension funds are still stuck with investment programs that generally underperform in strong markets and lose money in weak markets because the objective is to beat a long-term, market-related benchmark. There is limited protection against short-term volatility. Although there are a variety of investment solutions available, they are not widely used because they don’t fit neatly into the rigid two-stage model of a market benchmark policy and its implementation. These solutions include alternative investments such as real estate and private equity and absolute return strategies such as hedge funds.
Consider Hedge Funds
Consider hedge funds for a moment. Some pension funds do not include them at all because they can’t be modelled in the standard asset/liability framework. Hedge funds are not about static market exposure.
Other funds have adopted them on a limited basis using a variety of contrived ways to make them fit with the beta framework. For example, some incorporate them in the policy as an asset class, applying an arbitrary allocation of five to 10 per cent. Others include hedge funds in their alpha implementation either to hedge some of the beta or to add value as a ‘portable alpha’ over a beta portfolio.
I believe this is backwards. Hedge funds shouldn’t be fit around the market exposures. Rather, the market exposures should be fit around the hedge funds (and other alternatives) if, and only if, beta is needed. In other words, the investments should more likely be 100 per cent in alternatives instead of 100 per cent in equity markets!
You could argue that the use of active managers reduces the level of market exposures and the respective short-term volatility. But this reduction is insignificant. Comparing the Mercer passive policy asset mix benchmark returns from 1994 to June 2004 to the median manager returns shows the rolling four-year correlations ranged between 98 and 99 per cent! For the same period, the correlation between the returns of the median manager and the CSFB/ Tremont hedge fund index ranged from about zero to 59 per cent. The range of correlation is even more extreme for individual hedge fund of fund managers.
There are two important observations from the data for this period.
First, hedge funds were actively changing their positions in response to market volatility in order to manage short-term volatility.
Second, hedge fund managers were successful in capturing market returns in the strong markets and protecting capital in the weak markets. There is no mystery how they did this and will continue to do this.
Hedge funds are not benchmarked to the market and they take advantage of both buying stocks and selling short.
Why do we have a beta investment policy then, especially in a market environment of low interest rates and limited equity return expectations? Is there anything to stop us from restating the policy as 100 per cent alternative investments and strategies? Pension legislation requires a policy asset mix that reflects the needs of the plan, its risk tolerance, and the prevailing market conditions. Ironically, the ubiquitous asset mix of ‘60/40’ is not conveying very much about the needs of the plan. And the fact that it is fairly static does not say much about prevailing market conditions!
The needs of the plan are reflected in a minimum risk portfolio which matches investments to the liabilities. But there is a huge gap between the matched position and the ‘60/40’ policy that is impossible to reconcile. The average pension plan is adverse to the cost of investing 100 per cent in the minimum risk portfolio, but that doesn’t justify completely ignoring it.
Minimum Risk Portfolio
So maybe we don’t need a beta investment policy. Maybe what we need instead is a policy asset mix built around the minimum risk portfolio by combining valueadded portfolios to the extent that they are necessary and desirable. Some of these portfolios may be equity beta portfolios, but not necessarily. Equity market betas typically generate returns over longer time horizons. This time horizon mismatch creates short-term volatility that may not be tolerable. The better solution is to add alternative investments that are not subject to the short-term volatility of the markets and hedge fund managers who actively manage the short-term volatility.
The allocation to alternative investments – such as private equity and real estate – would be determined based on the liquidity needs of the fund. The allocation to hedge funds can be done using an efficient frontier model of risk/return optimization. The policy might be stated as something like 50 per cent (minimum risk portfolio) plus 50 per cent (value-added portfolio), with an expected return of eight per cent and expected volatility of four per cent over the next three years.
In order to move away from a policy that is almost 100 per cent beta, we have to start understanding the source of returns generated by managers, both active and hedged. In the past, this wasn’t critical given that the portfolios were highly benchmarked to the markets. There are two approaches to manager evaluation today. The first is the ‘trust me’ approach where the evaluation is based on historical returns and the manager’s own explanation for the performance. Especially in cases of good returns, this approach is intuitively appealing.
Different Sources Of Beta
The second approach is more quantitative. It uses style analysis and attribution of returns. There is still a high reliance on historical returns, but holdings level data is used to uncover different sources of beta such as value and growth or small cap and large cap. What’s still missing, however, is an analysis or ‘audit’ of alpha. For example, a recent explanation for exceptional Canadian equity performance by a manager given by a large consulting firm was “they buy value stocks that go up and sell them before they go down.” Surely there’s more to the story than this, otherwise everyone would be doing it.
Hedge funds – and, in fact, all active managers – generate three types of return:
- ‘exotic’ beta
Laurence Siegel describes ‘exotic’ beta as betas not yet identified or not yet available through low cost vehicles1. For example, 20 years ago style investing was exotic beta. Today shorting growth stocks and buying value is exotic beta. Alpha is defined as anything other than beta and exotic beta.
Beta, of course, is market exposure. Clifford Asness, a prolific author in this area, believes that alpha becomes beta over time once the source is recognized. The expected return of the manager will be a weighted average of beta, exotic beta, and alpha. The actual return will also include a component of luck – good and bad.
We need to get to this level of sophistication in order to assess whether the returns are sustainable and under what conditions they will perform well or poorly. In addition to monitoring for performance, compliance, and style as we do now, the investment program should be monitored for changes that would impact the expected risks and returns. These could include changes in interest rates, credit spreads, an alpha becoming a beta, or even a new manager offering a different source of returns. Martin Leibowitz makes a strong case for more adaptive investment policies. Currently, the main tactical moves around the policy asset mix are ‘buy and hold,’ ‘rebalancing,’ and ‘market timing.’ Leibowitz suggests a new approach of “integrating” obvious changes in the environment into the policy2.
To summarize, I recommend the following changes to the way we invest:
- Abandon the two-stage investment policy and implementation model
- Replace equity market exposures with stable, longer term alternative investments, and with hedge funds that manage short-term volatility by managing the mix of beta, exotic beta, and alpha
- Explicitly include the minimum risk portfolio in the investment policy (even if it is zero weighted)
- Analyze the source of investment returns from active management as a combination of beta, exotic beta, alpha, and luck
- Recognize the need for long-term returns but manage around the risk of short-term volatility with regular policy reviews
How do you get started? Put the ideas forward at the 2005 investment committee meeting and take it from there. The transition to the new approaches won’t happen right away, but at least it’s a beginning.
Until then, no more talk of beta policies, alpha transport, benchmarking, tracking error, or anything else predicated on an underlying beta policy. Let’s start working on a methodology to create a policy asset mix with the right time horizon, proper risk management, and a focus on the right financial obligations.
Wendy Brodkin is president of Goodman Institutional Investments Inc.
1. The Changing Role of the Policy Portfolio: CFA Institute Conference Proceedings, February 2004
2. Distinguishing True Alpha from Beta: CFA Institute Conference Proceedings, February 2004
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -