(Bloomberg) -- When the going gets rough, the 1 percent start selling.
That’s the finding of a new paper that says people with the highest income bailed from stocks disproportionately on the worst days of the financial crisis. The share of selling by the biggest earners rose “sharply” in days following spikes in volatility, according to data on millions of sales reported to the government in 2008 and 2009.
Mapping selling patterns in periods of tumult is of interest to researchers trying to get at the psychological underpinnings of events such as the financial crisis, when more than $10 trillion was erased from U.S. share values. Their main conclusion, that different people react with varying urgency to signs of trouble, could help identify behavioral biases that feed market meltdowns.
“Very, very high income people are disproportionately likely to sell a bunch of stock during a financial crisis,” said Daniel Reck, a doctoral candidate in the economics department at the University of Michigan and one of the paper’s authors, said by phone. “It’s difficult to say exactly how much high-income people are responsible relative to everyone else, but they’re certainly contributing more to volatility.”
One explanation for the divergence is that rich people have more at stake per person and are more sensitive to shocks, though it’s only speculation, Reck said. Another is they believe they’re better market timers. A third possibility is that investors who earn less are reluctant to sell at a loss, a cognitive tendency known as the disposition effect.
For the study, researchers from Ohio State University and the University of Michigan examined about 273 million Internal Revenue Service 1099-B forms that record sales in stocks and mutual funds, comparing them with levels of gross income and plotting them against fluctuation in the CBOE Volatility Index. To control for the fact that rich people own more stock, the study focused on changes in the typical selling levels among income groups versus the years before the collapse of Lehman Brothers Holdings Inc.
“We find that, starting in September 2008, the share of sales volume attributed to the top 0.1 percent of income recipients and other top income groups rises sharply until the beginning of 2009, and in 2008 and 2009 the sales of these groups are relatively more associated with stock market tumult as measured by the VIX,” they wrote.
Specifically, according to the paper, a roughly 10 percent increase in volatility correlated with a 3.3 percent increase in sales volume for the top 0.1 percent of earners, relative to less-affluent investors. When the VIX went up 25 percent, proportionate selling by the ultra-rich rose almost 8 percent.
Using average selling levels by the various wealth classes, the authors estimated that the effect might have accounted for roughly $142 million of excess selling by the 0.1 percent group on the day of Lehman’s collapse alone, and $1.7 billion in the 10 days after what was then history’s biggest bankruptcy filing.
Other investor demographics, from gender to marital status to place of residence, showed no signs of being related to the volatility sensitivity of stock sales, the study showed.
The stock market meltdown provided a special opportunity for the study’s authors to explore which cross section of investors contributing most to selling pressure. The paper joins an expanding pool of research assessing the effect of wealth on investment decisions, and how different groups judge risk.
“The older view of the stock market is that prices move in response to news that arrives -- it’s not usually thought of as a large collection of individuals trading on slightly different information and having different reactions to new events,” Reck said. “Documenting this kind of heterogeneity will allow us to paint a fuller picture of how stock markets actually work.”
The paper, titled “Who Sold During the Crash of 2008-9? Evidence from Tax-Return Data on Daily Sales of Stock,” was co-authored by Reck, his fellow doctoral candidate Bryan Stuart, Michigan economics professors Stefan Nagel and Joel Slemrod, Ohio State accounting professor Jeffrey Hoopes and IRS research analyst Patrick Langtieg.
While the IRS data provided new means for assessing who unloaded equities during the crisis, it doesn’t give a complete picture, the authors note. Absent from the numbers are share purchases, which would provide net selling data. It’s also possible that people increase levels of selling from non-taxable retirement accounts during crises.
So were people who sold during the 2008 rout guilty of a behavioral bias involving market timing or just good traders? Both could be true. Someone selling early in the plunge avoided at least part of the biggest equity wipeout since 1937, which saw the S&P 500 plummet 57 percent.
At the same time, holding stocks since the 2007 peak hasn’t been a money-losing proposition. The S&P 500 has posted annualized return of almost 6 percent since the high point in October 2007.
“At this point it’s too early to tell whether market timing has been successful,” Reck said. “We have the data -- we know what they sold and we can tie it to historical stock prices. But right now, we’re not conclusive one way or the other.”