Equity styles and the Spanish flu

Since the beginning of 2020, equity styles (or ‘factors’) have performed in very disparate ways, as the table below shows. The relative returns of these four styles are shown versus the market index, in US dollars, as at Tuesday 31 March. We use publicly available MSCI and S&P indices.¹

Momentum and low volatility displayed positive performance relative to the market, whereas small caps and high dividend (value) stocks underperformed.² In hindsight, investors should have focused on low volatility and momentum only, in order to find some protection during the first quarter of 2020. March was particularly challenging for factors. Small tilts to small value stocks would have hurt performance significantly during the sell-off.

To understand better the performance of equity markets and factors during pandemics, we analyze the period of the Spanish flu outbreak of 1918-1919 in more detail.

To paraphrase the first sentence of Tolstoy’s Anna Karenina, all bull markets are alike (nothing bad happened), but every bear market is bearish in its own way. Each crisis is different, but perhaps history can offer some guidance. The world has experienced other pandemics in the past, most notably the ‘Spanish’ Flu of 1918-1919.

This influenza pandemic occurred in three waves throughout 1918 and 1919. The first wave began in March 1918 and lasted until the summer. The second wave was the most lethal and occurred during the fall of 1918 (October to December). It was followed by a third wave in the spring of 1919 (February). After that, the virus disappeared, probably due to improved treatments, or a mutation of the virus into a less lethal form.

During this pandemic, which spread across the globe via travelers and soldiers engaged in World War I, around 40 million people died worldwide from early spring 1918 to late spring 1919, of which 675,000 people in the United States (about 0.8% of the 1910 population).³

During this pandemic, many businesses faced economic disruption due to a lack of available workers, temporary closure of other businesses, and a drop in demand. Moreover, the pandemic occurred towards the end of World War I, a period of widespread economic stress and a severe market correction, as highlighted below.

How did the equity market and factors behave during this period? The years around 1918 are not covered by the widely-used CRSP database, which starts in 1926. As a result, most investors ignore how factors performed during the Spanish flu outbreak period. Perhaps lessons could be drawn from this period (and other market sell-offs prior to 1926).

To answer this question, we use the data on all listed US stocks and their characteristics from Baltussen, Van Vliet and Van Vliet (2020). When studying equity styles, it is very important to correct for any size effects, since small stocks tend to be more biased and volatile and can dominate results. Also, the pre-CRSP period was characterized by many listings of small, thinly traded stocks, making it important to value-weight stock returns in order to study economic significant effects for investors. Baltussen, Van Vliet, Van Vliet (2020) therefore supplement their stock database with hand-collected data of the market capitalization of US stocks, extending the CRSP period to 1866.⁴

In line with Fama-French, factors are constructed by splitting all stocks in two groups: small and big, and are subsequently sorted on value (dividend yield; ‘DIV’), volatility (36-months return volatility: ‘LowVol’) and momentum (12-1 month returns; ‘MOM’) as a 2x3 sort. We will focus on the largest 50% of stocks for the equity styles. In addition, we construct the value-weighted returns on all small stocks (‘SMALL’).

As the Spanish flu occurred around World War I stocks markets were especially occupied with worries about the war. The peak of the stock market was reached in November 1916, but then sold off to bottom a year later. With the relief about the end of the war a recovery started in which the Spanish flu occurred. The figure below shows the performance of the stock market and the equity styles for the period from the stock market peak during World War I and the Spanish flu period (November 1916), to bottom (November 1917) and subsequent recovery (February 1919).

Over this period, the market dropped by about -20% from peak to bottom due to negative developments around World War I. High-dividend stocks and low volatility stocks offered protection, although not as much as in other market corrections. All stock moved in tandem and correlations went up. There was nowhere to hide. This is a bit like the recent March 2020 sell-off. After that markets recovered, with a dip around the first outbreak of the Spanish Flu.

The markets fully recovered by the end of February 1919. During the correction, small caps underperformed, and winner stocks performed in line with the market. In the following recovery period, small caps showed the strongest performance.

Historically, epidemics and pandemics have often caused short-term equity market corrections (such was the case with SARS, MERS, the Hong Kong flu and others). As equity corrections are less rare than pandemics, it is especially interesting to take a closer look at the different equity styles during other market corrections in the pre-1926 period.

This selection of market crises include market episodes which are not often studied. This includes the 1907 Bankers’ Panic, also known as the Knickerbocker crisis⁵ , and also the panic of 1903, 1893, 1884, and 1873. In these six market corrections, the stock market fell by about -20% to -25%. During these periods, we see the pattern emerging of low volatility and momentum offering protection, similar to March 2020.

Intriguingly, during the recoveries that followed (the timing of which is by the way generally hard to predict) small caps and high dividend outperformed, as shown in the figure below. We also present a multi-style ‘conservative’ strategy following Blitz and Van Vliet (2018), in which low volatility, momentum and yield are combined in one strategy (‘LowVol+).⁶ Overall, this ‘multi-style’ strategy managed to limit losses and outperform over the correction-recovery cycle.

Footnotes

¹We use respectively the MSCI Minimum Volatility, value-weighted, momentum, high-dividend, momentum, small cap US indices and the S&P500 low-volatility, pure value, momentum, high dividend US indices and the Russell2000 (small cap) index.
²Significant differences exist within the same style in 2020. We therefore combine (equal weight) the different style indices of MSCI and S&P and focus on the US market only.
³Garrett, T.A., 2007, Economic Effects of the 1918 Influenza Pandemic, working paper Federal Reserve Bank of St. Louis.
⁴Study will be available later this year including all details on the investment universe and database construction. For the 1918-1919 case study, we cover 523 US stocks for which we have market capitalization data and at least one year of returns.
⁵https://en.wikipedia.org/wiki/Panic_of_1907
⁶https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3145152