Value and Momentum During Recessions Ada Investment Management Research Note #2009/Q1 The Value strategy buys stocks that are cheap and sells those that are expensive. The Momentum strategy buys stocks that have had high recent returns and sells those that have had low recent returns. Value and Momentum strategies have been widely studied in academia. Several of these studies have shown the importance of these exposures in understanding returns. This has in part contributed to its popularity. Value and Momentum, in various guises, are two strategies employed by many quantitative hedge funds. In this research note we highlight how they have performed during recessions. In order to precisely date recessions, we use the National Bureau of Economic Research (NBER) business cycle dates. 1 The NBER recession dating procedure incorporates many economic and financial indicators in their determination of recession dates, including the performance of the stock market. 2 Having all of the data available as well as then analyzing it means that the recession dates are often announced with a substantial lag ‐ indeed the NBER announced in December 2008 that the current recession had begun in December 2007. So while the dates are an important standard for understanding when recessions occurred, they provide little real‐time guidance as to when a recession begins or ends. In order to generate returns to value and momentum factors, we rely on the most widely used proxies, first generated by Ken French and Eugene Fama. The Fama‐French version of Momentum is known as Up Minus Down (UMD). The Fama‐French version of the Value strategy is called High Minus Low (HML). 3 For purposes of a benchmark, we also include the strategy of investing in the market and selling Treasury bills (MKT). Ken French maintains a data library with returns to their versions of each strategy from January 1927 till February 2009. 4 Information ratios in recessions versus expansions Thus, our analysis will use data going back to 1927. During this period from 1927 to February 2009, we witnessed 14 recessions spread over 207 months. We also had 781 expansionary months.
1
Also freely available at http://www.nber.org/cycles.html. In fact, in the fourteen recessions in our sample, the NBER has never dated the end of a recession in a month where the stock market went down 3 For their Value strategy, Fama‐French categorize stocks as cheap or expensive based upon the ratio of the book value of the stock to its market value, although it should be noted that there are many other variants based upon different measures of value. Since 1927 4 The data and more detailed descriptions of the strategy construction are freely available at http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html. 2
Average Return Volatility Information Ratio (IR)
Table 1. Strategy Returns Jan 1927 – Feb 2009 Recessions and Non‐Recessions Market (MKT) Value (HML) All Rec. Non‐Rec. All Rec. Non‐Rec. 6.81 ‐8.73 10.93 4.70 1.51 5.55 18.96 28.06 15.49 12.46 16.77 11.03 0.36 ‐0.31 0.71 0.38 0.09 0.50
Momentum (UMD) All Rec. Non‐Rec. 9.29 9.14 9.33 16.17 25.25 12.74 0.57 0.36 0.73
Table 1 shows that Value has returned 4.70% annually at an IR of 0.38 and Momentum has done even better, returning 9.29% at an IR of 0.57. For reference, an investment in the market has returned 6.81% at an IR of 0.36. Therefore, each of the strategies considered have been sound investments over the last eighty years. However, when we consider performance during recessions, our conclusion shifts. Not surprisingly, investing in the market is a painful experience during recessions with an average return of ‐8.73% and volatility of 28%. Value performs similarly ‐ with returns falling to 1.51% and markedly increased volatility, the IR becomes an unremarkable 0.09. Momentum, on the other hand, holds up rather well with similar returns in expansionary and recessionary periods. The strategy does become volatile, thus causing the IR to drop but remain at a healthy 0.36. Diversification benefit in recessions versus expansions To get a sense of the diversification that value and momentum provide to an investor, we look at how these strategies are correlated with the MKT. To further separate the impact of recessions we look at these correlations in both recessions and expansions. Since diversification is of particular value when the market is falling, it is also useful to see if these correlation patterns are concentrated during rising or falling markets. Since both momentum and value take long and short positions in stocks, market exposure should be relatively low. In Table 2, we see that this is the case, with the full sample beta for Value at 0 and Momentum at ‐0.06. Yet these full sample betas hide interesting patterns. Drilling down to analyze the exposures during recessionary and non‐recessionary periods; as well as to rising and falling markets reveals these patterns. 5 All of the negative market exposure of momentum is concentrated during recessions, with both the up and down market exposures becoming ‐0.32. Outside of recessions, the strategy has no real exposure to rising markets and a positive exposure of 0.26 to falling markets. This suggests that much of the diversification benefit of investing in Momentum occurs during recessions. In stark contrast, Value’s low beta behavior disappears during recessions. In recessions, Value’s beta to falling markets becomes 0.14, precisely when one would not want it. 5
We estimate all up market and down market exposures in a single robust regression for each strategy and then a separate robust regression to estimate the full‐sample overall beta. Statistical significance is indicated with an asterisk.
Table 2 – Strategy Market Exposures Value (HML) Momentum (UMD) UpBeta DownBeta Beta UpBeta DownBeta Beta Rec ‐0.016 0.140* ‐0.322* ‐0.324* NonRec ‐0.047 ‐0.064 ‐0.038 0.256* Full Sample ‐0.004 ‐0.061* The current recession has seen these strategies behave as expected. Figure 1 plots an investment made at the end of 11/2007, right before the NBER start date for the current recession. The overall market has been a terrible investment, particularly since the beginning of 10/2008. Value returns were relatively stable and did well prior to 10/2008, rising 9% in three months. However, as the market began substantial declines in October, the strategy began to move more closely with the market and lost 25% over the next five months. In contrast, Momentum has performed relatively well, with annualized returns of 22% at 20% volatility throughout the recession. However, we also haven’t seen significant bear market rallies or recoveries in this sample (the data from Ken French does not yet include March 2009). One would expect momentum to give back some gains during these rallies and during the period of recovery in equity markets. We hope the above patterns help you understand these two popular quantitative factors. Figure 1 – Strategy Performance During Current Recession (2007/11 = 100)