Glamour among value: P/E ratios and value investor attention

• DOI: http://doi.org/10.1111/fima.12281
• Date: 01 September 2020
• Published date: 01 September 2020
• Author: Jordan Moore

DOI: 10.1111/fima.12281

ORIGINAL ARTICLE

Glamour among value: P/E ratios and value

investor attention

Jordan Moore

Rohrer College of Business, Rowan University,

Glassboro, New Jersey

Correspondence

JordanMoore, Rohrer College of Business,

RowanUniversity, 201 Mullica Hill Road, Glass-

boro,NJ 08028.

Email:moorejs@rowan.edu

Abstract

The trailing-four-quarter price–earnings (P/E) ratio is the most pop-

ular fundamental value proxy.This article is the first to examine the

P/E ratio as the preeminent measure of value investor attention.

Trailing-four-quarter P/E ratios predict significantly greater cross-

sectional variation in stock returns than lagged P/E ratios or current

price-to-book (P/B) ratios. P/E strategy returns are robust to vari-

ables that proxyfor fundamental risk, variables mechanically related

to P/E ratios, relativetrading volume, and liquidity. The role of atten-

tion is evidentin return patterns across long and short portfolios, day

of the week, and time since formation. Stocks with low P/E ratios

exhibit an increase in total trading volume driven by small trades,

an improvement in liquidity,and lower idiosyncratic volatility. These

patterns are consistent with the typical trading activity of individual

investors, who havethe strongest attention constraints.

1INTRODUCTION

Investors payattention to price–earnings (P/E) ratios far more than any other measure of fundamental value. Because

it is well known that value stocks earn high average returns, investors with limited attention are likely to search for

value stocks using P/E ratios.1P/E ratios are widely quoted as the ratio of a firm’s stock price and its total earnings

per share (EPS) over the last four quarters. This article is the first to examine trailing-four-quarter P/E ratios as the

preeminent proxy for value investorattention.

Trailing-four-quarterP/E ratios predict significantly greater cross-sectional variation in stock returns than lagged

P/E ratiosor current price-to-book (P/B) ratios. I provide evidence to support a behavioral explanation for this superior

return predictability.The relation between P/E ratios and attention is evident in return patterns across long and short

portfolios, day of the week, and time since formation. P/E ratios predict changes in overalltrading volume, percentage

c

2019 Financial Management Association International

1Graham and Dodd (1934) encourageinvestors to buy assets with low prices relative to underlying fundamentals. Asness, Moskowitz, and Pedersen (2013)

documenta value premium in equities, equity indices, bonds, currencies, and commodities.

Financial Management. 2020;49:673–706. wileyonlinelibrary.com/journal/fima 673

674 MOORE

of small trades, liquidity,and idiosyncratic volatility, which are all consistent with individual investor trading behavior.

Individual investors havestronger attention constraints than institutional investors.2

Attention is a scarce cognitive resource. Kahneman (1973) shows that people have limited ability to pay attention

to multiple pieces of information at the same time. It is impossible to obtain aggregate data on ideal proxies for atten-

tion, such as the pupil dilation of all market participants. However,Da, Engelberg, and Gao (2011) and Ben-Rephael,

Da, and Israelsen (2017) show that search activity on Google and Bloomberg are robust proxies for individual and

institutional investor attention, respectively.Figure 1 illustrates the P/E ratio’s unrivaled popularity among individual

investors. Hou, Xue, and Zhang (2015) examinea taxonomy of 80 cross-sectional anomalies, 12 of which proxy for the

value-versus-growthpremium. Among topics related to the 12 value anomaly variables, P/E ratios generate the highest

Google search volume in every month of the data. Individuals search Google for P/E ratios 6.9 times more often than

P/B ratios, the ubiquitous value proxy in the finance literature. Likewise, Figure 2 shows that institutional investors

search Bloomberg for P/E ratios more than any other value metric.

Averagevalue strategy returns using trailing-four-quarter P/E ratios are nearly twice the average strategy returns

using either lagged P/E ratios or current P/B ratios. From 1973 to 2015, a value-weighted decile trailing-four-quarter

P/E strategy earns an average monthly long–short return of 106 basis points with an annual Sharpe ratio of 0.84.

This strategy earns a significant positive alpha in the Fama and French (2015) five-factor model, which also controls

for exposure to market beta, size, value, profitability, and investment. It is important to control for variables that

relate mechanically to P/E ratios or changes in P/E ratios. Therefore, I show that P/E ratiosstill predict returns within

portfolios already sorted on market leverage, price momentum, monthly reversals, share price,accruals, or earnings

momentum.

The value effect is persistent, whereas the attention effect is fleeting. Merton (1987) associates a positive atten-

tion shock with a higher equilibrium price, higher trading volume, and lower future returns. Because value investors

primarily pay attention to P/E ratios, it makes sense that P/E strategies outperform P/B strategies at first and then

underperform later.The superior return predictability is driven by the short side of the portfolio. This is not surprising,

as there are greater limits to arbitragewhen implementing the short sideof the strategy. The strong pattern in strategy

returns across days of the week is consistent with findings documented by Abrahamand Ikenberry (1994), Lakonishok

and Maberly (1990), and Birru (2018).

Patterns in trading volume suggest that P/E ratios predict value investorattention. Barber and Odean (2008) find

that individual investors are especially likely to purchase attention-grabbingstocks.3Conversely, individual investors

rarely hold short positions. Individual value investors are likely to buy stocks with low P/E ratios and ignore other

stocks. There is a significant increase in trading volume for stocks with low P/E ratios relative to other stocks. I

apply the Battalio and Mendenhall (2005) algorithm to identify individual investor trading activity. Stocks with low

P/E ratios generate 2–3% in additional trading volume, all of which is attributable to increases in individual investor

trading.

The relation among P/E ratios and liquidity are also consistent with individual investor behavior.Kaniel, Saar, and

Titman (2008) show that individual investors provide liquidity when they purchase stocks. Individual investors are

especially susceptible to the disposition effect, the tendency to sell stocks with positive returns to realize gains.4The

disposition effect implies that individual investors also provide liquidity when they sell stocks. If individual investors

2Althoughindividuals are more constrained by limited attention than institutions, many studies present evidence of limited attention in equity market pr ices.

Cohenand Frazzini (2008) show that the stock prices of supplier firms adjust slowly to changes in future earnings expectations of customer firms. DellaVigna

and Pollet (2009) show that the stock price reaction to earnings is delayedwhen companies announce on Fridays, whereas Hirshleifer, Lim, and Teoh(2009)

showthat the reaction is delayed when many other firms release earnings on the same day.

3Many studies relate attention proxiesto subsequent returns. These include trading volume (Gervais, Kaniel, & Mingelgrin, 2001; Kaniel, Ozoguz, & Starks,

2012),search volume (Da, Engelberg, & Gao, 2011), and proximity to the 52-week high (George & Huang, 2004; Li & Yu, 2012).

4Forinstance, Odean (1998) finds robust evidence of the disposition effect in the trading activity of individual investors at a large discount brokerage. Hartz-

mark(2015) finds that individual investors tend to sell the stocks in their portfolios with extreme returns, and are much more likely to sell the stockswiththe

largestgains than the largest losses.

MOORE 675

FIGURE 1 P/E ratios and individual investorattention

Note. These figures show the Google Search VolumeIndex (SVI) for topics related to fundamental value. The sample

covers January 2004–April 2017. Data are from Google Trends(https://www.google.com/trends). SVI is the number

of web searches that Google assigns to specific search terms or broad topics. The P/E Ratio series charts searches for

the price–earnings ratio topic. Common search terms that Google assigns to the price–earningsratio topic include p/e,

pe ratio, price earnings, and price earnings ratio.The P/B Ratio series charts searches for the P/B ratio topic. Common

search terms that Google attributes to the P/B ratio topic include p/b, pb ratio,price book, price book ratio, and

market to book. The MarketLeverage series charts searches for the debt-to-equity ratio topic. Common search terms

that Google attributes to the debt-to-equity ratio topic include debt ratio, equity ratio,debt to equity, and equity debt

ratio. The Dividend Yield series charts searches for the dividend yield search term. PanelA shows the monthly SVI

time series for all four topics. The values in Panel A are normalized so the topic-month observation with the most

searches has a value of 100. This observation corresponds to the price–earnings ratio topic in October 2008. PanelB

shows the averageSVI for all four topics for every month of the calendar year. The values in Panel B are normalized so

that the averageJanuary SVI for each topic is 100. Panel B shows that the SVI for all four topics is above average in the

spring and fall and low in the summer and winter in the Northern Hemisphere, where the largest equity markets are

located.

actively buy and sell stocks with low P/E ratios, these stocks should be especially liquid. I estimate Jacobs and

Hillert (2016) predictive liquidity regressions and show that stocks with low P/E ratios are at least 20% more liquid

than expected.

The fact that trailing-four-quarter P/E ratios proxy for value investor attention can help explain the earnings

momentum anomaly. Ball and Brown (1968) are the first to show that the magnitude of a firm’s quarterly earnings

relative to investor expectations predicts stock returns. In a seasonal random walk model, investors expect a firm