Benchmark concentration and the effects on active portfolios

Diversification and risk management cannot fully eliminate the risk of investment loss.

Concentration risk in a new regime

Equity markets have become increasingly concentrated, with a small group of mega-cap stocks driving most of the returns in indices like the S&P 500. Explore recent concentration trends in the charts below.

Portfolio construction considerations

Active managers have tools to adapt. As benchmarks become more concentrated, the way portfolios are constructed becomes crucial for maintaining performance.

In this research, we investigate the relationship between active risk and benchmark concentration across three strategies: Long-only, partial long-short, and full long-short implementations.

In this analysis, we focus on the expected risk-adjusted return, specifically the forecast information ratio, to compare and draw conclusions from simulated data. ¹ To evaluate the effects of benchmark concentration, we create permutations of the current S&P 500 Index weightings by adjusting concentration levels. Starting with the S&P 500 benchmark weights as of April 2024, we compute new asset weights using the following function:

In the equation, a scaling factor m is applied to simulate various concentration levels (from 0.8 to 1.2), increasing or decreasing concentration while preserving rank order. Figure 4 (below) shows the resulting effective breadth for the S&P 500 at varying values of n.

We conduct single period mean-variance optimizations using simulated return forecasts across various implementations: Long only, partial long/short (130/30 and 175/75), and full long-short market neutral (100% long and 100% short). We measure forecast IRs across different risk and benchmark concentration levels.

The return forecasts are generated using a random distribution of 900 sets of stock-level alphas where each set is normally distributed with a mean of zero and a standard deviation which corresponds to an expected IR of 1.0 in the unconstrained long-short portfolio.

Figure 5a visualizes forecast IR vs. active risk across the different strategies and concentration levels. Long-only portfolios show greater dispersion in IR as benchmark breadth narrows and partial and full long-short portfolios show flatter, tighter lines, suggesting less sensitivity to concentration changes.

Figure 5b depicts the % change in forecast IR as benchmark effective breadth drops from 122 to 29. In this instance, long-only portfolios suffer most with a ~19% drop in IR at 1.0% active risk. In partial long-short IR drops range from -2% to -4.4%. In full long-short, there is 0% change in IR, demonstrating full insulation from concentration effects. This indicates that partial and full long-short portfolios are more resilient in concentrated benchmark environments.

In figure 6 we explore the impact on long-only portfolios. At higher benchmark concentration and constant active risk, IRs decline for long-only portfolios. However, if active risk is reduced, IR can be preserved.

For example, at a 3% active risk level with a benchmark effective breadth of 122, the expected IR is 0.47. This is roughly equivalent to the expected IR at a 2% risk level when benchmark breadth drops to 29. Conversely, if the risk level remains at 3%, the IR declines to 0.39 when the benchmark breadth drops to 29.

Similarly, if the S&P 500 becomes twice as concentrated compared to 2024 levels, the expected IR in long-only strategies may be maintained by halving active risk.

Final takeaways

As cap-weighted indices become more concentrated, investors may consider an evolved approach. In our view, portfolio flexibility is key to maintaining consistent alpha. Whether through active risk management, partial long-short design, or portable alpha overlays, active managers can adapt and thrive — even in a market dominated by a few giants.