A publication of the School of Business at Loyola University Chicago
Opaque Financial Reports, R2, and Crash Risk
Amy P. Hutton, Alan J. Marcus, and Hassan Tehranian
Journal of Financial Economics, 2009, 94, 67-86.
Hutton, Marcus, and Tehranian (HMT) investigate the financial reports as the three-year moving sum of annual
relationship between the opacity of financial reports and the discretionary accruals. The idea is that firms with consis-
distribution of stock returns, most notably the propensity tently large accruals are more likely to be managing earn-
for the stocks of certain firms to experience sudden large ings and thus to reveal less firm-specific information. R2
negative returns, that is, to crash. The intuition underlying comes from a regression in the spirit of Fama and French.
HMT’s analysis is that, for some firms, managers “…are The authors define a crash as a weekly negative stock
able to stockpile negative information, hiding it from inves- return that is 3.09 standard deviations (or more) below the
tors’ view until its accumulation reaches a tipping point suf- mean value, corresponding to a frequency of 0.1 percent
ficient to result in a stock price crash.” (68) from the normal distribution. A jump is a positive stock re-
The “R2” of the title refers to the goodness-of-fit be- turn of the same magnitude. Thus, a crash or a jump equates
tween the stock returns of a particular firm and the market to a weekly abnormal stock return with a magnitude of at
as measured in a modified index-model regression. HMT least 18 percent. If stock returns are normally distributed,
expect that R2 will be higher for firms with less transparen- the probability of a single jump or crash in a year would be
cy: “When less firm-specific information is publicly avail- 5.07 percent.
able, fewer observable reasons exist for individual stock HMT develop several models to explore their first hy-
returns to depart from broad market indexes and stock mar- pothesis. Controlling for a variety of other factors (such as
ket synchronicity increases…” (68) The restriction of firm- firm size) HMT find that their measure of opacity is posi-
specific information by managers that increases crash risk tively related to R2, and conclude that the impact of their
also increases R2. opacity measure on R2 is “…as strong as the appearance of
These intuitions give rise to one ancillary and three explicit firm-specific news in the financial press.” (79)
major hypotheses. HMT first hypothesize that firms with With respect to the probability of a crash, HMT find
opaque financial reports have stock returns that are more that the greater the opacity of financial reports by their
synchronous with the market, that is, these stock return re- measure, the greater is the risk of a crash. The authors find
gressions exhibit a higher R2. Second, firms with opaque that crash risk is largely independent of accounting practice,
financial reports face greater crash risk. But greater opac- yet the probability of a crash event is economically and sta-
ity should not lead to greater jumps, because managers do tistically significantly related to their opacity measure. Also
not try to hide positive information. This leads to the third as hypothesized, greater opacity of financial statements is
hypothesis: firms with opaque financial reports experience not related to a higher probability of a stock price jump.
no greater likelihood of large stock price jumps. As an Finally, HMT find that the relationship between opac-
ancillary hypothesis, HMT suggest that these effects may ity and the probability of a crash has diminished in the
be lessened in the post-Sarbanes-Oxley period due to in- post-Sarbanes-Oxley period: “This is consistent with the
creased transparency and reduced earnings manipulation. hypothesis that, in a period with greater monitoring and
HMT draw on Compustat and CRSP for the 1991-2005 scrutiny of accounting practice, earnings management has
period and focus on weekly stock returns for a sample con- substantially subsided…” (84)
sisting of 40,882 firm-years. They measure the opacity of
http://www.RMRR.com Fall 2009 4