Glued to the TV: Distracted Noise Traders and Stock Market Liquidity

• Author: DANIEL SCHMIDT,JOEL PERESS
• DOI: http://doi.org/10.1111/jofi.12863
• Date: 01 April 2020
• Published date: 01 April 2020

THE JOURNAL OF FINANCE •VOL. LXXV, NO. 2 •APRIL 2020

Glued to the TV: Distracted Noise Traders

and Stock Market Liquidity

JOEL PERESS and DANIEL SCHMIDT∗

ABSTRACT

In this paper, we study the impact of noise traders’ limited attention on financial

markets. Specifically, we exploit episodes of sensational news (exogenousto the mar-

ket) that distract noise traders. We find that on “distraction days,” trading activity,

liquidity, and volatility decrease, and prices reverse less among stocks owned pre-

dominantly by noise traders. These outcomes contrast sharply with those due to the

inattention of informed speculators and market makers, and are consistent with noise

traders mitigating adverse selection risk. We discuss the evolution of these outcomes

over time and the role of technological changes.

THE LITERATURE ON LIMITED ATTENTION in financial markets focuses on informed

agents—speculators and market makers—and on the speed with which they

impound public information into stock prices. Less is known about how atten-

tion might affect a third, equally important, group of agents: noise traders.

By trading stocks for reasons unrelated to stock fundamentals, noise traders

allow speculators to profit from their information and market makers to recoup

losses from trading with speculators. As summarized by Fisher Black in his

Presidential Address, “Noise Makes Financial Markets Possible” (Black (1986,

p. 530)). In this paper, we aim to improve our understanding of how the limited

attention of noise traders affects financial markets. Toward that end, we con-

sider events that distract noise traders and we trace out their consequences for

markets. We find that when noise traders are distracted from trading, liquid-

ity and volatility decrease and prices reverse less. These effects are consistent

with what theory predicts.

∗Joel Peress is at INSEAD. Daniel Schmidt is at HEC Paris. We are grateful to David Str¨

omberg

for sharing detailed news pressure data, including headline information, and to Terry Odean for

providing the discount brokerage data. We thank Thierry Foucault; Marcin Kacperczyk; Yigitcan

Karabulut, Maria Kasch, Asaf Manela, Øyvind Norli, Paolo Pasquariello; Joshua Pollet (the AFA

discussant); Jacob Sagi; Paolo Sodini; Avi Wohl;Bart Yueshen; and Roy Zuckerman for their valu-

able comments, as well as seminar/conference participants at the AFA,EFA, 7th Erasmus Liquidity

Conference, ESSFM Gerzensee, 2015 European Conference on Household Finance, UC Berkeley,

Coller School of Management (TelAviv University), Marshall School of Business, Frankfurt School

of Management and Finance, Hebrew University, IDC Arison School of Business, Imperial Col-

lege Business School, University of Lugano, University of Paris-Dauphine, and Warwick Business

School. Joel Peress would like to thank the AXA Research Fund and the Institut Europlace de Fi-

nance for their financial support. The authors have nothing to disclose with respect to The Journal

of Finance’s Submission Guidelines and Conflict of Interest Disclosure Policy.

DOI: 10.1111/jofi.12863

C2019 the American Finance Association

1083

1084 The Journal of Finance R

There are two main challenges to identifying the effect of noise traders’ at-

tention on markets. First, events that draw attention to stocks are typically

associated with material news about stocks’ fundamentals, which raises ques-

tions about whether any observed pattern is caused by attention itself or by

news. Second, many attention-grabbing events affect investors at large and not

just noise traders. We overcome these problems by exploiting events that (I)

divert, rather than attract, investors’ attention and (II) appeal to irrational,

“biased,” or sensation-seeking investors. Specifically, we identify episodes of

sensational news that are exogenous to the stock market and temporarily dis-

tract traders. A vivid example of such an event is the murder trial verdict

in the case of football and movie star O.J. Simpson on October 3, 1995. Mil-

lions interrupted what they were doing to hear the verdict announcement.

Long-distance telephone call volume declined, electricity consumption surged

as viewers turned on televisions set, and water usage plummeted as they post-

poned using bathrooms (Dershowitz (2004)). More relevant for our purpose, and

as Figure 1shows, trading volume on the New York Stock Exchange (NYSE)

plunged by 41% in the first five minutes after the announcement, and by an-

other 76% in the next five minutes, before abruptly recovering. Because the

O. J. Simpson trial was unrelated to the economy, such an episode speaks to a

causal effect of (in)attention on the stock market. Moreover, a detailed analysis

of who is distracted by sensational news demonstrates that it mainly affects

noise traders.

Tocreate a sample of such distraction events, we use a variable (introduced by

Eisensee and Str¨

omberg (2007)) called news pressure. This variable measures

the median number of minutes that U.S. news broadcasts devote to the first

three news segments. For example, the O.J. Simpson trial verdict received

16.5 minutes of air time, the highest value for that year. Each year, we sort days

into news pressure deciles and identify days belonging to the highest decile. We

then parse through the headlines of news segments covered in the broadcasts

and retain only those days for which the sensational news event is plausibly

exogenous to the economy. Examples of such distracting news include the O.J.

Simpson trial verdict, the Cessna plane crashing on the White House lawn, and

the Challenger space shuttle explosion. Data on TV viewership confirm that

these events drew the attention of U.S. households. Our final sample contains

551 days (distraction events) over the period from 1968 to 2014.

We begin by examining whether these news stories actually divert noise

traders’ attention away from the stock market. Detailed trading records of

households (from a large discount broker) and institutions (from Abel Noser

Solutions, or ANcerno) show that trading volume drops on distraction days

by 6.5% and 4%, respectively, before reverting to normal on the following day.

We next construct several proxies for investor “biasedness,” in particular, the

extent to which investors make losses from trading, churn their portfolios, un-

derdiversify, and trade stocks covered in the media, and, for households, their

gender and whether they were quick to adopt online trading. These factors

are all related to the tendency of traders to be sensation-seeking, overconfi-

dent, or susceptible to media coverage. We find that, among both households

Glued to the TV: Distracted Noise Traders and Stock Market Liquidity 1085

Figure 1. Trading activity during the O.J. Simpson trial verdict.

This figure shows the value of aggregate trading volume (in logs) on the New YorkStock Exchange

on October 3, 1995, the day the verdict of O.J. Simpson’s murder trial was announced. The top,

middle, and bottom panels display trading volume for, respectively, all, small, and large trades.

Trades are sorted into five size groups, where small (large) trades are those in the bottom (top)

quintile. The horizontal axis labels five-minute intervals starting from 9:30 a.m. EST. The vertical

line marks the announcement time (10 a.m. PST or 1 p.m. EST). The solid horizontal line indi-

cates the average (log) trading volume during that day (excluding the time period from 10:00 to

10:10 a.m.) for the trade size category displayed in the panel. The dashed horizontal line indicates

the 5% confidence bound (1.96 times the standard deviation of (log) trading volume during the

day). Data for this figure come from Trades and Quotes (TAQ). (Color figure can be viewed at

wileyonlinelibrary.com)

and institutions, more “biased” investors are more prone to distraction, which

confirms that our sensational news episodes primarily distract noise traders.

We therefore exploit these episodes to study how short-lived changes to noise

trading affect financial markets.

We next examine how the stock market behaves on sensational news days

(i.e., when noise traders are inattentive). Although we find weak results for the

overall market, we uncover pronounced effects when focusing on subgroups of

stocks with high retail ownership—stocks in which noise trading is expected

to be more pronounced (e.g., Lee, Shleifer, and Thaler (1991)). We confirm that

trading activity declines (by about 3%) in the bottom tercile of stocks in terms

of firm size, stock price, and institutional ownership—three variables that are

negatively correlated with retail ownership. Exploiting CRSP daily data as well