Who Sells During a Crash? Evidence from Tax Return Data on Daily Sales of Stock

Abstract

Using U.S. tax-return data containing the universe of individual taxable stock sales from 2008 to 2009, we examine which individuals increased their sale of stocks following episodes of market tumult. We find that the increase was disproportionately concentrated among investors in the top 1 and top 0.1% of the overall income distribution, retired individuals, and individuals at the very top of the dividend income distribution. Our estimates suggest that, following the day when Lehman Brothers collapsed, taxpayers in the top 0.1% sold $1.7 billion more in stocks than individuals in the bottom 75%. This difference is equal to 89% of average daily sales by taxpayers in the top 0.1%.

Volatility Expectations and Returns

“…A shortcoming of our model is that it features a representative investor and hence does not speak directly to this evidence (as there is no trade in equilibrium), though modest extensions of the model that allow for differences in the amount of bias would naturally be consistent with the evidence on trading behavior in Hoopes et al. (2021).…”

“…Taken together, the firm-level results support our main hypothesis that agents initially underreact to changes in variance and that this is reflected in implied volatilities. Hoopes et al (2021) provide evidence that investors do react to changes in volatility, with more sophisticated investors and older investors responding more strongly. Specifically, they show that higher income and older investors sell more aggressively following increases in volatility.…”

“…Giglio et al (2019) also find that agents' views about risk are informative about their portfolio decisions, with higher expectations of stock market risk being associated with a lower allocation to stocks. A shortcoming of our model is that it features a representative investor and hence does not speak directly to this evidence (as there is no trade in equilibrium), though modest extensions of the model that allow for differences in the amount of bias would naturally be consistent with the evidence on trading behavior in Hoopes et al (2021).…”

“…Hoopes et al. (2021) provide evidence that investors do react to changes in volatility, with more sophisticated investors and older investors responding more strongly. Specifically, they show that higher income and older investors sell more aggressively following increases in volatility.…”

“…Further, Hoopes et al. (2021) show evidence that investors do react to changes in volatility with more sophisticated investors (e.g., those in highest income brackets) responding most quickly. That is, it does not appear that agents are not aware and do not act on changes in volatility.…”