Strategy selection is never easy. The disclosure “Past performance does not guarantee future results” is displayed on every fact sheet for good reason. By definition, a strategy will look its best immediately preceding periods of underperformance.

Figure 1 shows the rolling three-year excess returns for an active strategy vs. its benchmark. Typical manager searches require at least three years of history, and a strategy at peak performance will often float to the top. How does a practitioner select a strategy that offers the best chance at long-term performance without picking an intermediate top?

Mitigating random strategy performance

Unfortunately, the answer is disappointing: There is no magic bullet. The reason for this is the randomness of performance patterns; it is nearly impossible to tell which strategy will outperform over the next 12 to 24 months. A solid manager with a good entry point can be chosen, but fate could have other ideas. Only with hindsight is it obvious which strategy worked.

Despite the knowledge that strategy selection is impossible to time, there is a way to mitigate the effects of random performance. Selecting complementary strategies that provide diversification helps temper performance cycles.

Figure 1: Active manager vs. S&P 500^® Index
rolling three-year excess annualized return (9/28/12–5/31/22)

Performance data shown represents past performance and is no guarantee of future results.
Source: FactSet. For illustrative purposes only.

The classic example of this is the style pairing of growth and value. While one strategy is in favor, the other is out, mitigating the timing effect of strategy selection. Expand this to each asset class in the portfolio, and the timing aspect is largely mitigated. 

What about a 50/50 mix? 

When looking at a 50/50 portfolio of the large-cap growth and large-cap value strategy, combined, their performance pattern becomes more stable, and excess returns against the benchmark become more reliable.

This process can be applied across asset classes, sectors, and style factors. No matter the pairing, two important considerations are the stability of the relationship over time and the theoretical grounding behind the relationship. This helps answer the question: “Will this relationship persist in the future as expected?”

Focusing on each strategy’s fit within the total portfolio lessens the emphasis on individual performance and tempers emotional decisions when a strategy underperforms. However, this does not remove the need for individual evaluation. Track record is important and should be long enough to experience a full performance cycle. This means at least 5 to 10 years of returns.

Applying additional screening criteria

A manager search applied to an investible universe should rank and weight criteria the practitioner finds meaningful. This often includes risk-adjusted performance, risk measures such as maximum drawdown, and other rankable data points such as fees. What criteria to include in the ranking and their weights depends largely on investment philosophy. For example, more defensively oriented portfolio managers may weight downside risk statistics more heavily.

In addition to quantifiable data points, the strategy should exhibit return and risk characteristics that reflect its stated investment process. Performance cycles should align with this process as well. For example, a more aggressive and value-driven process should outperform in strong cyclical uptrends. Since each strategy is a piece within a total portfolio, style drift is an important aspect to monitor. Two strategies that drift together stylistically over time can double up exposures. Finally, if it is an active strategy, it should exhibit outperformance against its benchmark over a full market cycle.

Complementing existing portfolio strategies

The four short-listed funds in Figure 2 are intended to complement a defensive growth equity strategy. They rank at the top of the large-cap value universe and have strong risk-adjusted statistics. The table shows seven-year statistics for the short-listed value managers and the growth strategy against the S&P 500^® Index. We can see strong alpha and Sharpe ratio statistics for Value strategy B and Value strategy D. While both strategies have strong risk-adjusted returns, Value strategy B is more aggressive, with a higher standard deviation and tracking error.

Looking at excess return correlation in Figure 3, all four strategies are good complements to the Growth strategy with a negative correlation. Value strategy C shows the lowest excess return correlation with the Growth strategy, at -0.67.

Figure 2: Multi-statistic review of value funds as complement for defensive growth strategy (9/30/15–9/30/22)

Performance data shown represents past performance and is no guarantee of future results.
Source: FactSet. For illustrative purposes only.

Figure 3: Correlation of excess returns with defensive growth strategy (9/30/15–9/30/22)

Performance data shown represents past performance and is no guarantee of future results.
Source: FactSet. For illustrative purposes only.

Charting rolling three-year excess returns vs. the S&P 500^® Index, both strategies act as a complement to the Growth Strategy (Figure 4). Value strategy B has more extreme peaks and troughs than Value strategy D, but directionally both have similar performance profiles. In the recent time periods, Value strategy D has underperformed, while its longer-term history shows more stability. This type of quantitative analysis combined with qualitative assessments helps align past performance with future expectations.

Figure 4: Value strategy B and D returns vs. defensive growth strategy
Rolling three-year excess annualized return vs. S&P 500^® (10/31/12–6/30/22)

Performance data shown represents past performance and is no guarantee of future results.
Source: FactSet. For illustrative purposes only.

Making the final decision

In Figure 5, both a 50/50 combination of Growth strategy with Value strategy B and Value strategy D look strong: higher returns, lower volatility and positive alpha compared to the S&P 500^®. While Value strategy B has higher absolute returns, it is higher risk, sporting larger maximum drawdowns and a higher tracking error. While Value strategy D has struggled recently, it has shown more consistency in the past. Depending on objective and risk tolerance, both strategies are solid complements.

Figure 5: Pairing growth strategy with Value strategy B or D
Rolling three-year excess annualized refturn vs. S&P 500^® (10/31/12–10/31/22)

Performance data shown represents past performance and is no guarantee of future results.
Source: FactSet. For illustrative purposes only.

Multi-statistic analysis (10/31/12–10/31/22)

Performance data shown represents past performance and is no guarantee of future results.
For illustrative purposes only.

The final decision comes down to a matter of preference: A more defensive, lower tracking-error strategy vs. a more aggressive and higher tracking-error strategy. No matter the preference, instilling a process around strategy selection and sticking to it will ensure the best chances of building a resilient portfolio. Pairing complementary strategies together reduces timing risk and the probability of a bad entry point. While it is impossible to predict the future, we can reduce uncertainty through thoughtful investment strategy selection.

Explore our model portfolios

Our multi-asset hybrid models combine strategic investments and active mutual funds with tactical positions and passive ETFs.