Factoring Bear Markets

For financial market stakeholders, March 2020 will be remembered as a key market turning point alongside October 1987 (Black Monday), September 2001 (9/11 attacks), September 2008 (Global Financial Crisis), and August 2011 (European Sovereign Debt Crisis). In March 2020, the S&P 500 recorded its fourth-largest monthly percentage decline (-12.5%) since the second World War. The three largest declines occurred in October 1987 (-21.8%), October 2008 (-16.9%), and August 1998 (-14.6%). On the European side, using the STOXX Europe 600 as our benchmark, March 2020 saw the largest monthly percentage decline ever (-14.8%).

However, in terms of price volatility, March 2020 has been the month with the highest standard deviation for the S&P 500 (93.2%) since WWII. For the STOXX Europe 600, March was the second most volatile month (67.9%) after October 2008 (74.2%).

Does history provide any pattern that will allow us to identify the factors reacting to stressed markets? Is it feasible to detect factors responding better than others to the bear market in terms of risk-adjusted returns?

To answer these questions, I have constructed 10 factors: Value, Growth, Momentum, Quality, Risk, Coverage, Profitability, Leverage, Size, and Liquidity. All factors are sector neutral and equally weighted. I have back-tested monthly data starting from December 1999, isolating three periods affected by shocks (September 2001, September 2008, and August 2011) to detect how these factors behave in the 18 months following the shock. I adopted a long/short strategy in which I buy the first fractile (F1) and sell the last fractile (F5, in this research) of each factor. The sample is composed of the 1000 companies with the highest market capitalization on a rolling monthly basis across two different regions, Europe and the U.S. The factors’ returns are compared to the STOXX Europe 600 and S&P 500 for the European and the U.S. sample, respectively.

October 2001 – March 2003 (9/11 Attacks)

For our European sample in the 18 months following 9/11, we see the long/short strategy applied to the Value factor yielded the best results in terms of returns, risk, and statistical significance. From October 1, 2001, to March 31, 2003, the Value F1-F5 cumulative return was 54.7%, exceeding the STOXX Europe 600 by 63.9%, which declined 9.2% over the same period. The long/short strategy beat the benchmark across all factors. To measure risk-adjusted returns, I used the Sharpe ratio. Value was the factor with the highest Sharpe ratio (0.62) for the 18 months after the 9/11 attacks.

From a statistical perspective, across all factors, Value had the highest Information Coefficient (IC) (0.09) and the IC t-stat was 11.75. The Information Coefficient (IC) measures the correlation between forward returns data and current factor figures. A significant IC t-stat means that you can reject the null hypothesis that the IC equals 0.

We see similar results for the U.S. sample. Value was the best performing factor for the 18 months after 9/11 with our long/short strategy with a cumulative excess return over the S&P 500 of +48.6%. Only Liquidity and Momentum underperformed the benchmark (-7.7% and 11.7%, respectively). At risk-adjusted returns level, Coverage’s Sharpe ratio (0.32) was slightly higher than Value and Size (0.31).

As we saw with the European sample, for our universe of U.S. companies, the Value factor had the highest IC (0.05) and IC t-stat (7.39).

November 2008 – April 2010 (Global Financial Crisis)

For our European sample of companies, Value was the best-performing factor following the global financial crisis when applying a long/short strategy in terms of returns. Between November 1, 2008, and April 30, 2010, its cumulative excess returns over the STOXX Europe 600 was +33.5%. However, compared to the post-9/11 period, where all factors overperformed the STOXX Europe 600, here the long/short strategy beats the market only on four factors (Value, Coverage, Size, and Leverage). Coverage (0.53) is the factor with the biggest Sharpe ratio in the 18 months following the financial crisis.

At a statistical level, Value factor had the greatest IC (0.05) and the IC t-stat was 6.67.

During the 18 months after October 2008, the U.S. sample gives results significantly different from the European sample. The long/short strategy underperforms against the S&P 500 on all factors and only three factors (Size, Coverage, and Leverage) have positive Sharpe ratios.

Size is the factor with the highest IC (0.04) and IC t-stat (5.10).

September 2011 – February 2013 (European Sovereign Debt Crisis)

For our European sample, the period following the European sovereign debt crisis differs from the two periods we analyzed above, as the long/short strategy on the Value factor underperformed most other factors and the STOXX Europe 600. Momentum is the factor with the highest cumulative excess returns (+11.3%). However, in terms of risk-adjusted returns, Coverage has the highest Sharpe ratio, which is what we also saw following the 2008 financial crisis.

From a statistical perspective, Momentum has the highest IC (0.05) and the IC t-stat is 5.95.

For the U.S. sample, backtesting results for the period following the European debt crisis yields similar results to the post-9/11 period. The long/short strategy underperformed the S&P 500 across all factors and Value has the highest risk-adjusted return (Sharpe 0.22).

Value has also the greatest IC (0.03) and IC t-stat (4.37).

Conclusion

The bear markets we analyzed all had different causes and characteristics. Moreover, the government interventions through monetary and fiscal policies differed between Europe and the U.S. as well as from crisis to crisis. Although backtesting these historical periods did not give us the same results, there are some consistent signals that we can glean from the last twenty years:

• The Value factor can provide good defense, in terms of excess returns, from stressed markets. Statistical tests corroborate this thesis. This appears to apply more in Europe than in the U.S.
• At a risk-adjusted return level, Coverage is a defensive factor. Even though it was never the best factor for excess returns in any of the periods analyzed, it was among the top three Sharpe ratios in two-thirds of the periods we analyzed. This hints that Coverage can provide a valid equilibrium between risk and returns during periods following bear markets.
• A long/short strategy wins more in Europe than in the U.S. The reason may be the higher relative strength of the U.S. benchmark over the European one (the S&P 500 outperformed the STOXX Europe 600 by 87% in the period March 2000 – March 2020, with both indices measured in local currency).