The authors test their theory that constituents of ETFs react very differently to negative news, with some more resistant to macro shocks than others. They create a simple trading strategy that buys oversold index constituents when an ETF’s price falls amid a market panic caused by negative news flow about a sector or a prominent stock within that sector.

The authors analyse two broad market ETFs (the S&P 500 Index and small cap) and nine sector ETFs. Across the 11 ETFs, a total of 240 volume spikes are noted over an eight-year period. After each volume spike that results in falls in an ETF’s price, the methodology is to buy the “outsiders”—the ETF constituents with a low beta to the overall ETF. That is, their prices move less in tandem with the ETF than its other constituents.

These outsiders are all held for 40 days, and the authors measure the alpha produced from the first to the 40th day after the volume spike, with the results shown both before and after trading costs.

During ETF volume spikes, the correlations between the prices of companies within each ETF increase, despite clear differences in the fundamentals of the companies and despite some companies having little exposure to the ETF theme. The authors refer to this as a “correlation bubble.”

For example, after a September 2015 tweet by then presidential candidate Hillary Clinton in which she announced that she intended to cap prices of some pharmaceuticals, there was a rapid sell-off in health care ETFs. Although the prices of health care ETFs plummeted after her announcement, the stocks of some non-pharmaceutical companies rebounded strongly in the 40 days thereafter.

If investors buy ETF “outsiders,” immediately after a spike in volume and lever them to the same beta as the ETF, they are likely to benefit from reversion of the prices for a full 40 days. Although the average alpha on the first day after a surge in ETF sales volumes is slightly negative, over the next 39 days the average daily alpha is generally positive.

This strategy is profitable in the S&P 500 ETF and in the small-cap ETF, as well as in eight of the nine sector strategies—consumer discretionary, consumer staples, energy, financials, health care, industrials, materials, and utilities. The technology sector is the exception.

The highest returns after 40 days are found in the financials ETF, which produced alpha of nearly 8%. Applying the strategy to the S&P 500 ETF creates returns of around 6%. The technology ETF produces losses of around 3% 40 days after the market panic and subsequent spike in ETF trading volumes.

There seems to be a clear return opportunity for stock pickers who buy “outsider” stocks that are swept along in the high-volume selling of sector or thematic ETFs.

The strategy requires only the ability to measure a stock’s beta relative to its ETF and lever a portfolio of outsider stocks to the same beta as the ETF. Stock pickers with greater insight into fundamental valuations may be able to obtain even higher returns than this simple strategy presents.

For ETF investors, the mispricing during volume spikes of the companies that comprise an ETF may represent a hitherto unknown cost of ETF investing. But the authors note that it takes many types of investors with different motivations to provide liquidity to capital markets, which in turn provides more appropriate pricing.

Exchange-traded funds (ETFs) account for nearly a third of trading volumes in the United States, up from less than 2% in 2000. This sharp rise in volume may leave ETF investors vulnerable to getting “picked off” by active investors when trading in ETFs is particularly heavy.

ETFs enable investors to express top-down views on markets and sectors without paying attention to the relative merits of the stock fundamentals within each market or sector. This may create opportunities for stock pickers when ETF constituents are rapidly oversold.

The authors investigate what happens to the prices of the constituent stocks of an ETF after a strong sell-off in the ETF.

They reason that ETF investors are agnostic to stock-specific information, so they “throw the baby out with the bathwater,” effectively handing returns on less correlated constituents to stock pickers.