Style factors: bank failures spur unexpected and extreme returns

Following the failure of Silicon Valley Bank, returns for many style factors have been extremely large and in the opposite direction to what could be expected given their long-term history. Such performances may be of particular interest as style factors tend to go haywire at market inflection points and often in advance of a crisis.

Exhibit 1 shows the returns for the style factors in the Axioma Medium-Horizon Fundamental US model for March 10-17 compared with the long-term average value. The numbers above or below the bars are the percentile ranking of the return compared with rolling six-day returns (the higher the percentile ranking, the higher the return).

Although none of the returns over this period are all-time low or high values, Dividend Yield, Earnings Yield, Leverage and Value all fall into the first (lowest) percentile. On average, Earnings Yield and Value have positive returns, so these results run counter to expectations. In contrast, the Size factor’s return fell into the 99^th percentile, meaning that investors were shunning smaller names and piling into large-cap stocks. This is also counter to the long-term average levels. While Momentum’s magnitude of return was not quite as close to the tail as the other factors, it was nonetheless negative. In the first few days of the crisis, Momentum actually had a positive return; it was only in the last few days that it dipped down.

Not all factors produced returns counter to expectations. Both Volatility and Market Sensitivity saw negative returns, meaning investors were seeking the relative safety of lower Volatility and lower beta stocks. These factors produce returns that are negative on average, although they had been positive more recently as stocks surged early in the year. Finally, Profitability has had a strong run, as investors rewarded high cash flows and profit margins in this uncertain environment.

And finally, Exchange Rate Sensitivity measures a stock’s sensitivity in the movement of its home currency. A positive exposure means that the stock should go up as the currency strengthens. Over time, the average return to this factor is zero, but over the recent period in question it has been unusually strong, despite the weakness in the dollar, as many of the big banks had negative exposures to the factor (so negative exposure times negative return = positive contribution).

We have written in the past about our observation that style factors tend to go haywire at market inflection points. This is based more on anecdotal evidence working with factor models for the past 40 years rather than a statistical study. This often happens in advance of a crisis. For example, the ‘Quant Crisis’ of 2007 predated the onset of the Global Financial Crisis by more than a year, although factors were ‘behaving badly’ for much of 2008. 1987 is another example of quant models breaking down before the market crash – in that case we can attribute the performance to trying to invest rationally in an irrational world, when market valuations were through the roof.

In Exhibit 2 we show the factor returns from the day of the SVB crisis outset through six days later, to compare the current performance to what happened immediately surrounding other periods of market stress. In this chart we can see that while the returns for most factors have been large in absolute terms, in most cases they are comparable to what we have observed in the past – in magnitude, if not in direction. In fact, only Value has shown consistently negative returns at the immediate outset of a crisis; it is also the only factor in which this recent period has seen the lowest return of any of these particular crisis periods.

Exhibit 1 — US factor returns, March 10-17, 2023, versus average six-day returns, 1982-March 2023

Exhibit 2 — Six-day factor returns during selected periods of market uncertainty

Overall, we expect factor returns to settle down as they usually do, but that may not happen until we see panic selling in the banking sector subside. In the meantime, model users should expect to see higher volatility in many style factors along with changing correlations.

Qontigo is a leading global provider of innovative index, analytics and risk solutions that optimize investment impact. As the shift toward sustainable investing accelerates, Qontigo enables its clients—financial-products issuers, asset owners and asset managers—to deliver sophisticated and targeted solutions at scale to meet the increasingly demanding and unique sustainability goals of investors worldwide.

Qontigo’s solutions are enhanced by both our collaborative, customer-centric culture, which allows us to create tailored solutions for our clients, and our open architecture and modern technology that efficiently integrate with our clients’ processes.

Part of the Deutsche Börse Group, Qontigo was created in 2019 through the combination of Axioma, DAX and STOXX. Headquartered in Eschborn, Germany, Qontigo’s global presence includes offices in New York, London, Zug and Hong Kong.