The Calm before the Storm

• Author: FERHAT AKBAS
• Date: 01 February 2016
• DOI: http://doi.org/10.1111/jofi.12377
• Published date: 01 February 2016

THE JOURNAL OF FINANCE •VOL. LXXI, NO. 1 •FEBRUARY 2016

The Calm before the Storm

FERHAT AKBAS∗

ABSTRACT

I provide evidence that stocks experiencing unusually low trading volume over the

week prior to earnings announcements have more unfavorable earnings surprises.

This effect is more pronounced among stocks with higher short-selling constraints.

These findings support the view that unusually low trading volume signals negative

information, since, under short-selling constraints, informed agents with bad news

stay by the sidelines. Changes in visibility or risk-based explanations are insufficient

to explain the results. This evidence provides insights into why unusually low trading

volume predicts price declines.

IN AN INFLUENTIAL STUDY, Gervais, Kaniel, and Mingelgrin (2001) document

that unusually high (low) trading volume, measured over a day or a week,

predicts higher (lower) future returns. This phenomenon is known as the high

volume return premium. Their study spurred a flurry of research proposing

explanations for this premium, such as changes in visibility or compensation for

risk.1These explanations, however, mainly relate to the return premium due to

unusually high volume—the relation between unusually low volume and future

returns has received less attention, although it is equally important.2Our

understanding of why unusually low volume predicts future returns therefore

remains limited.

In this paper, I show that unusually low trading volume signals forthcom-

ing negative information about changes in firm fundamentals, as captured

∗Ferhat Akbas is from the School of Business, University of Kansas. I thank William Armstrong;

George Bittlingmayer; Ekkehart Boehmer; Bilal Erturk; Egemen Genc; Chao Jiang; TedJuhl; Ron

Kaniel; Sami Keskek; Paul Koch; Felix Meschke; Ralitsa Petkova; Sorin Sorescu; Musa Subasi;

Julie Wu; and seminar participants at the University of Kansas, the 2013 Southern Finance Asso-

ciation Meetings, and the 2013 Financial Management Association Meetings for helpful comments.

I am particularly grateful to Kenneth Singleton (the Editor) and two anonymous referees for their

insightful comments and suggestions. Part of the analysis in this paper was conducted while I was

a PhD student at Texas A&M University. I have received no external financial support for this

research and am aware of no conflicts of interest.

1See, for example, Garfinkel and Sokobin (2006), Barber and Odean (2008), Lerman, Livnat,

and Mendenhall (2008), Gallmeyer, Hollifield, and Seppi (2009), Schneider (2009), Banerjee and

Kremer (2010), and Kaniel, Ozoguz, and Starks (2012) for alternative explanations of the high

volume return premium.

2Gervais, Kaniel, and Mingelgrin (2001) show that both high and low volume shocks signif-

icantly contribute to the high volume return premium. For example, among small firms where

the return predictability is most observed, stocks with high (low) unusual volume outperform

(underperform) stocks with normal volume by 50 (−50) basis points over a month.

DOI: 10.1111/jofi.12377

225

226 The Journal of Finance R

by earnings surprises.3I find that stocks with unusually low trading volume

prior to earnings announcements experience significantly lower unexpected

earnings compared to stocks with volume or unusually high volume. The re-

sults are stronger among stocks with binding short-selling constraints. In ad-

dition, I show that unusually low volume predicts abnormal returns around

earnings announcements in a way that is consistent with the unexpected

earnings prediction, with a strong relation between unexpected earnings pre-

dictability and return predictability. In contrast, while high unusual volume

predicts higher returns, it does not predict positive changes in future funda-

mentals. These findings support the view that unusually low trading volume

is a signal of bad news about firm fundamentals since, under short-selling con-

straints, informed agents with bad news (short sellers) stay by the sidelines.

These findings further suggest that the underlying driver for the volume-

prompted price movements is different for unusually high and low volume

shocks.

The idea that, under short-selling constraints, unusually low trading activity

is a signal of bad news is formalized in the seminal theoretical work of Diamond

and Verrecchia (1987). According to Diamond and Verrecchia (1987), “Periods

of the absence of trade are bad news because they indicate an increased chance

of informed traders with bad news who are constrained from selling short”

(p. 303). Their rationale is as follows.

Short sellers are informed traders whose trading activity signals negative

information about firm fundamentals. If the stock market were frictionless,

there would be no short-selling constraints and short sellers would freely trade

on their information. In this case, trading activity in a stock and a stock’s price

would reflect current conditional expectations about the asset’s value. How-

ever, in the presence of barriers to short selling, when unusually low trading

volume is observed, it is not clear that short sellers choose not to trade. If we

observe a sudden lack of trading in a stock under short-selling restrictions, the

conditional expectation of the price will be upward biased since the negative

information that short sellers possess is not incorporated into stock prices in

a timely manner. Thus, unusually low trading volume may indicate a higher

probability that, due to short-selling constraints, short sellers are forced to stay

by the sidelines and hence the negative information they possess is not incor-

porated into prices. As a result, an unusual decrease in volume may signal bad

news about the firm.

It should be noted that the Diamond and Verrecchia (1987) model assumes

a pure rational expectations economy where investors learn from low trading

volume instantaneously and incorporate its informational content into stock

prices immediately. However, if price adjustment to low trading volume is not

instantaneous and investors do not learn from trading volume immediately,

3By fundamental information, I mainly refer to information about changes in a firm’s cash flow

prospects.

The Calm before the Storm 227

then the model will arguably predict that unusually low trading volume con-

tains negative information about future prices.4

Two explanations have been proposed to explain the return predictability

due to unusual trading volume. The first argues that unusual volume predicts

future returns due to a sudden change in a stock’s visibility, as proposed by

Miller (1977), Mayshar (1983), and Merton (1987).5However, in these models,

the visibility of a stock changes mainly through unusually high volume: in-

vestor awareness of a stock might suddenly increase as a result of unusually

high volume but is unlikely to suddenly decrease following unusually low vol-

ume. Because of this asymmetry, the relation between unusually low volume

and future returns would be hard to explain with the visibility hypothesis.6

The second explanation argues that the high-volume return premium is com-

pensation for risk. For example, Banerjee and Kremer (2010) argue that unusu-

ally high volume reflects a high level of disagreement about the future, which

leads to increased uncertainty and risk today.7Alternatively, Schneider (2009)

suggests high trading volume implies low information quality and thus higher

uncertainty. Gallmeyer, Hollifield, and Seppi (2009) further propose that large

volume signals an unusual degree of uncertainty about investor demand for a

stock. In each of these models, the focus is on understanding why unusually

high volume predicts future returns and the return-predictive power associated

with unusually low volume receives little attention.

In this paper, I examine the nature of the information that is signaled by

unusually low volume. My primary contribution is to present evidence that

unusually low volume signals unfavorable changes in firm fundamentals. The

fundamental information channel provides an important and alternative ex-

planation for the association between low volume shocks and future returns.

To the best of my knowledge, this is the first paper to empirically document the

relation between unusually low volume and firm fundamentals.8

4Relaxing the pure rational expectations assumption is reasonable, since ample evidence sug-

gests that price adjustments to information are far from instantaneous. For example, the literature

on short selling (see Boehmer, Jones, and Zhang (2008)), price momentum (Hong, Lim, and Stein

(2000)), insider trading, (Seyhun (1986,1998), Lakonishok and Lee (2001)), and unusual volume

shocks (Gervais, Kaniel, and Mingelgrin (2001)) suggest that price adjustments to information are

far from instantaneous.

5Miller (1977) focuses on overvaluation of prices due to excess demand under short-sale restric-

tions, and Merton (1987) focus on the reduction in the estimation risk faced by traders due to an

increase in a stock’s investor base after experiencing excess buy-side demand. Barber and Odean

(2008) claim that the high volume premium can be explained by Miller (1977). Kaniel, Ozoguz, and

Starks (2012) provide evidence in favor of Merton’s (1987) explanation of the high-volume return

premium.

6Chen, Noronha, and Signal (2004) make a similar argument regarding the asymmetric price

reaction after additions to and deletions from the S&P 500 index. They argue that investor recog-

nition will increase after listing on the index, but will not decrease suddenly after delisting.

7Var ian (1985) also argues that high volume levels capture disagreement among investors from

a risk perspective.

8Gervais, Kaniel, and Mingelgrin (2001, p. 912) briefly discuss the implications of Diamond and

Verrecchia (1987) but do not discuss these implications in detail or present any formal test. Reed