When choosing an active equity manager, institutional investors typically focus on the manager’s ability to ensure that its products consistently outperform both their equity benchmark and their peers. Investors looking for absolute returns tend to focus on equity long/short strategies to profit from a manager’s stock selection skills. However, many investors’ investment guidelines do not allow for short sales.
The only way out of this dilemma is an investment strategy that is totally different from traditional index oriented strategies and allows for larger deviations from a global equity benchmark. Since the major global equity indices are cap weighted, they do not provide a suitable representative benchmark for most investors. The total return equity portfolio (TREP) is therefore optimised against a cash benchmark and boasts several features that typically appeal to risk averse investors.
The idea is just as simple as it is good: Based on the return forecast and risk estimates for a stock that were originally derived for a conventional equity investment strategy, the equity portfolio is optimised against a cash benchmark. Failure to use return forecasts would automatically produce a so-called minimum risk portfolio, the portfolio with the minimum standard deviation from the mean. Adding return forecasts will produce a superior portfolio if those forecasts add value in the long run. Long-standing experience in the ‘relative world’ of investing therefore supports the concept of an actively managed portfolio with a low risk target.
The objective of the TREP is to maximise the Sharpe Ratio and to offer investors an attractive absolute risk-return ratio. Even though the portfolio is fully invested in equities, the return target – that is the benchmark – of the portfolio is an active annual return of 400bps over money market, taking into account the absolute performance target.
Aside from a lower absolute risk compared with conventional equity strategies, the total return equity strategy is based on the rationale of better alpha generation. This is the main reason why investors increasingly demand ‘unconstrained’ management.
However, investors should be aware of deviations from a global equity index like the MSCI World. Historically, the rolling 12-month tracking error of our total return equity product ranged between 2% and 14% and averaged 8.5% (see Figure 1, left hand scale).
Figure 1 also illustrates a strong correlation (~ 0.78) between the tracking error of the total return equity products and the volatility of the global equity market represented by the MSCI World Index (right hand scale). During periods of strong market volatility, the TERP showed a high tracking error compared to the MSCI World Index. This meant that the portfolio performance deviated from that of the equity market index. During the period 2000 to 2002, in particular, the portfolio profited from these deviations because its losses were not as high as the losses of the cap weighted global equity market indices.
In the absence of return forecasts, the solution is the so called minimum risk portfolio. Since stock specific risk can easily be diversified, the remaining risk is mostly systematic risk. Common risk factors can be limited by selecting stocks with as many independent sources of common factor risk as possible from a broad global investment universe. The addition of stock specific return forecasts produces a superior portfolio compared with the minimum risk portfolio.
While traditionally managed equity products tend to focus on relative risk, the TERP seeks to maximise the expected return relative to the absolute risk. As per June 30 2005, the fund’s ex-ante calculated total risk (annualised standard deviation) amounted to just 5.8%. The MSCI World Index, as a proxy for an equity market portfolio, has an ex-ante calculated annualised standard deviation of 15.0%.
Looking at the attributes that emerged over the 15 years of back testing the portfolio, the participation rate was substantially lower in times of negative market performance than during periods of positive market performance. An analysis of 183 monthly observations yields 109 months with a positive market performance and 74 months with a negative performance.
On average, the loss of the minimum risk portfolio amounted to just 42% of the average monthly market loss. At the same time, the portfolio gains reached about 80% of the average monthly gain in months with positive market performance. This means a certain level of underperformance in rising markets but strong outperformance in falling markets.
As in the traditional relative investment approach, the alpha estimates for each stock in the investment universe are drawn from a proprietary global stock selection model that forecasts stock specific returns, which are neutral with respect to regions and industries. The portfolio is then optimised versus a cash benchmark, focusing on an optimal relation between the (absolute) target return and a minimum risk level. Input variables are the BARRA Aegis risk model and the forecasts from the global stock selection model. The portfolio contains no cash and is fully invested in equities.
The maximum monthly turnover is 7% to avoid excessive trading costs. The average expected alpha of a portfolio optimised against cash exceeds the expected alpha of optimisations against benchmarks as the latter contain a relatively large proportion of stocks included solely for the purpose of relative risk reduction.
The maximum exposures are limited to prevent outsized and overly risky positions in regions, countries or industries. Nonetheless, they offer enough leeway to implement an investment strategy that meets the performance target with a considerably lower risk. Currency exposure is almost fully hedged into the base currency. On a stock level, the maximum size of a position is based on the security’s liquidity to avoid market impact when buying or selling and to ensure sufficient investment capacity in the product. The number of portfolio holdings is expected to range between 90 and 110 stocks.
The optimal portfolio is a blend of (a) the minimum factor risk portfolio and (b) a specific risk/specific return optimised stock selection portfolio, subject to real world constraints and cost considerations.
Since the focus is on minimising total risk, the Sharpe ratio is an appropriate measure of the risk-adjusted attractiveness of the portfolio.
The portfolio showed a positive Sharpe ratio in 125 out of 148 rolling three-year periods compared to 80 incidences of a positive Sharpe ratio for the MSCI World Index. The average Sharpe ratio of the TERP was 0.67 during those periods compared to 0.18 for the global equity market.
With regard to total annualised return over the back test period, TERP showed a superior risk-return trade-off. It achieved a higher return with a lower risk compared to the global equity market, again represented by the MSCI World Index (see chart below).
The portfolio has outperformed its total return target of 400 basis points over money market by about 63 basis points a year (gross of management fees, net of transaction costs).
The product is geared towards institutional investors with a focus on absolute risk, high net worth individuals and retail investors. The first mandate of this kind was initiated by a UK retail bank and funded in June 2005. We expect demand for these strategies to increase sharply as the strategic hedging of currency exposure is viewed as another attraction.
Optimisation against cash combined with a proven stock selection process produces attractive risk-return features that differ substantially from those of most conventional equity products. Thanks to the comparatively unconstrained approach coupled with a solid, risk controlled investment process, the asset manager can select individual risk-return trade-offs while avoiding many of the pitfalls of intuitive approaches. It is therefore safe to assume market dominance in both absolute and relative terms over the long run.
Michael Fraiken is head of product management and Alexander Tavernaro is senior portfolio manager with Invesco structured product group