The affect heuristic and stock ownership: A theoretical perspective

• Date: 01 January 2019
• Published date: 01 January 2019
• DOI: http://doi.org/10.1002/rfe.1026

SPECIAL ISSUE ARTICLE

The affect heuristic and stock ownership: A theoretical

perspective*

Jiang Luo

1

|

Avanidhar Subrahmanyam

2

1

Nanyang Business School, Nanyang

Technological University, Singapore City,

Singapore

2

Anderson Graduate School of

Management, University of California at

Los Angeles, Los Angeles, CA, USA

Correspondence

Avanidhar Subrahmanyam, Anderson

Graduate School of Management,

University of California at Los Angeles,

Los Angeles, CA, USA.

Email: subra@anderson.ucla.edu

Abstract

We consider asset prices and informational efficiency in a setting where owning

stock confers direct utility due to an affect heuristic. Specifically, holding equity

in brand name companies or those indulging in “socially desirable”activities

(e.g., environmental consciousness) confers positive consumption benefits,

whereas investing in “sin stocks”yields the reverse. In contrast to settings based

on wealth considerations alone, expected stock prices deviate from expected fun-

damentals even when assets are in zero net supply. Stocks that yield high direct

utility are, on average, more informationally efficient as they stimulate more entry

into the market for these stocks and, consequently, more information collection.

The analysis also accords with a value effect, high valuations of brand-name

stocks, abnormally positive returns on “sin stocks,”volume premia in the cross-

section of returns, proliferation of mutual funds and ETFs, and yields untested

implications. If, as psychological literature suggests, agents derive greater utility

from successful companies by “basking in reflected glory,”then asset prices react

to public signals non-linearly, leading to booms and busts, as well as crashes and

recoveries.

KEYWORDS

asset prices, behavioral preferences, informational efficiency, stock ownership

1

|

INTRODUCTION

In recent years, a body of anecdotal and empirical evidence indicates that perceptions of firms’products influence market

valuations and investment choices. For example, the market valuation of Tesla recently exceeded that of Ford Motors, even

as Tesla’s sales volume was about 1% of that of Ford.

1

Billett, Jiang, and Rego (2014) show that stocks of companies with

prestigious brands have higher market/book ratios and earn lower average returns in the cross section. Keloharju, Kn€

upfer,

and Linnainmaa (2012) show that investors prefer to trade stocks of firms whose products they use. Hong and Kacperczyk

(2009) show that “sin”(tobacco, gambling) stocks earn positive abnormal (risk-adjusted) returns. Further, the number of

mutual funds in the U.S. rivals or exceeds the number of publicly traded individual firms, a proliferation that seem s to

*We thank Knut Aase, Syed Zamin Ali, Dominique Badoer, Warren Bailey, Jean Canil, Martin Dierker, Paskalis Glabadanidis, Rachel Gordon, Francisco

Guedes, Majid Hasan, David Hirshleifer, Griffin Jiang, Thore Johnsen, Byoung Kang, Jøril Mæland, Jane Luo, Sebastian Pouget, Hersh Shefrin, Xunhua

Su, Karin Thoburn, Wilson Tong, Kam-Ming Wan, Sterling Yan, Jeffrey Yu, and participants in seminars at the Norwegian School of Economics, Chap-

man University, University of Missouri, University of Adelaide, and the Hong Kong Polytechnic University, and in the 4th Indonesian Financial Manage-

ment Association Conference, the 2017 Asia Finance Association Conference, and the 2017 China International Conference in Finance, for valuable

comments.

Received: 20 March 2018

|

Accepted: 23 March 2018

DOI: 10.1002/rfe.1026

6

|

©2019 University of New Orleans wileyonlinelibrary.com/journal/rfe Rev Financ Econ. 2019;37:6–37.

emanate from investor tastes, since the funds’aggregate performance, net of fees, underwhelms.

2

In addition, several

exchange-traded funds (ETFs) in the U.S. cater to tastes.

3

To explain the above phenomena, we consider a setting where the affect associated with the product(s) of a firm (viz.

Chaudhuri & Holbrook, 2001) carries over to the decision to invest in a firm’s stock. Finucane, Alhakami, Slovic, and

Johnson (2000) characterize this aspect of choice as an affect heuristic, which is the notion that “...images, marked by pos-

itive and negative affective feelings, guide judgment and decision making”(p. 3). Thus, we propose, for example, that

some users of Apple products have positive affect toward Apple stock due to a favorable inclination toward its product s.

This reason for holding Apple complements traditional risk-reward considerations. Similarly, anothe r premise is that some

investors who care about the environment buy shares, for instance, in Tesla Motors, again for reasons that go beyond

wealth-related motives. Indeed, plenty of references in the popular press allude to the notion that investors often “fall in

love”with stocks.

4

Fama and French (2007) indicate that the standard assumption that “investment assets are not also con-

sumption goods”is “unrealistic.”Further, Nagy and Obenberger (1994) find in a survey of individual investors that a pri-

mary reason for stock investment is “feelings for a firm’s products and services,”which is an emotional reason for

attachment to a firm’s stock. Keloharju et al. (2012), in empirically linking investors’preference for firms’stock with

usage of the corresponding products, state that a “setup in which customer-investors regard stocks as consumption goods,

not just as investments, seems to best explain [their] results.”Beal, Goyen, and Philips (2005) argue that owning socially

responsible investments leads to an increase in psychic well-being.

5

Hong and Kacperczyk (2009) mention that “there is

clearly a societal norm against funding operations that promote human vice, and consequently many investors may not want

themselves or others to support these companies by investing in their stocks.”

What is the effect of non-wealth stock-holding motives on asset prices, incentives to acquire information, and informa-

tional efficiency? We address this question, and, in turn, derive results that accord with many stylized facts and yield sev-

eral other untested implications. Our analysis considers an equilibrium with asymmetric information where stock ownership

provides direct consumption benefits. In our model (based on Grossman & Stiglitz, 1980), there are four class of agents,

the standard informed and uninformed traders, noise traders, and a class of utilit y-maximizing traders that value stock for

intrinsic reasons, in addition to incorporating a traditional risk-reward tradeoff. We term the last class “Atraders,”for want

of a better term. Such traders can be proxied by retail investors who are less sophisticated than neoclassical utility-maximi-

zers, as well as professional money managers who cater to their investors’tastes. We consider multiple securities and also

allow these agents to acquire information about stocks.

We characterize equilibria with information acquisition and derive cross-sectional implications for expected stock

returns. We find that stocks that yield extreme utility (or disutility) tend to be, on average, the most informationally effi-

cient and least volatile since they yield high certainty equivalents for Atraders and thus stimulate entry and information

collection by these agents. Further, expected market-to-book ratios (proxied by the ratio of the expected market price to the

unconditional expected value of the stock) are higher for stocks that appeal to investors more strongly. This is because in

equilibrium, the expected prices of these stocks positively deviate from expected fundamentals. Since it is these stocks that

also generate lower expected returns, our model is consistent with the value effect documented by Fama and French

(1992), among others. Going beyond just explaining the value effect, our approach predicts that specific proxies for direct

utility from owning stock are related to market/book ratios and future returns. Thus, we argue that there will be a negative

relation between such proxies (such as brand visibility or environmental responsibility) and average returns. Similarly, there

will be a positive relation between a proxy for disutility (such as whether a stock is a “sin stock”) and average returns. We

predict that such relations will tend to be more pronounced for stocks which yield extreme direct utilit y and in which Atra-

ders face lower costs of information acquisition, for example, due to more stringent disclosure policies.

6

Keloharju et al. (2012) and Frieder and Subrahmanyam (2005) show that individual investors are respective ly attracted

to stocks whose products they use and whose brands are familiar; in our setting these findi ngs obtain because agents derive

direct utility from holding stocks of companies with visible products and brands.

7

Also, we demonstrate that if investors

derive disutility from holding “sin”stocks, such stocks earn positive abnormal returns, as in Hong and Kacperczyk (2009).

Further, we predict that trading volume should be highest for stocks that yield the greatest levels of utility or disutility. We

provide other untested implications that cross-sectionally relate proxies for direct utility in a stock to the type of clientele

(institutional vs. retail) that holds that stock. We also show that, under reasonable conditions, mutual funds or ETFs that

cater to investor tastes earn negative expected returns in equilibrium, thus rationalizing the subpar performance of mutual

funds extensively documented since Jensen (1968).

We extend our model to a dynamic setting, where Atraders make an entry decision based on past fundamentals. Psy-

chological literature (e.g., Cialdini et al., 1976), considers the penchant for individuals to associate with successful ventures

and people, i.e., to “bask in reflected glory.”We thus propose that Atraders derive greater utility (disutility) from holding

LUO AND SUBRAHMANYAM

|

7

stock with good (bad) past fundamentals. We find that if fundamental-related signals are above a certain threshold, Atra-

ders enter

8

and prices overreact to these positive signals (and subsequently correct). Similarly, there is a crash and recovery

following periods where fundamentals are below a lower threshold. We predict that such booms and busts will be higher in

stocks held by relatively unsophisticated retail investors, since those agents are more likely to derive direct utility from past

firm fundamentals.

Note that when agents receive direct utility from owing a stock, the expected return to these investors from holding

the stock can be less than zero. Our model accords with the notion that some financial market agents lose money on

average by trading stocks; see, for example, Barber and Odean (2001) or Odean (1999). This feature is common to

other models of financial markets, e.g., Kyle (1985) and its various extensions, where it is standard to assume that

some agents earn negative expected profits. We note, however, that our model predicts negative expected profits only

in stocks where agents derive extreme utility or disutility from stock ownership; in the intermediate cases, profits can

be positive. We also point out that valuing stock as a consumption good is not the same as trading on irrational (or

mistaken) beliefs. In our framework, agents who derive such direct utility calculate all exogenous parameters correctly

and have rational expectations, but just maximize a different objective relative to neoclassical agents who trade for tra-

ditional wealth-related reasons.

Our work is related to Barberis and Shleifer (2003) on style investing. In their sett ing, investors categorize assets into

different styles and move money across these styles depending on past performance.

9

We complement this approach via

modeling investors who value stock because of the attributes of the company or product, rather than a broad style. Dorn

and Sengmueller (2009) consider the notion that investors derive entertainment from trading equities. In contrast, we exam-

ine a framework where agents derive direct utility from holding specific stocks, as opposed to deriving utility from trading

stocks generically.

10

Fama and French (2007) suggest that variation in tastes for different stocks can cause agents to hold a

different portfolio than the standard tangency portfolio in a mean-variance setting. Since their focus is on portfolio choice,

they desist from endogenizing the effect of differing tastes on the equilibrium. We instead focus on a closed-form equilib-

rium to consider the informational efficiency of prices, the cross section of holdings, value effects, volume premia, and

equilibrium expected profits when Atraders are present in the market, and also provide untested implications.

11

A relevant issue in our model is whether our Atraders present an arbitrage opportunity. On this point, we note that our

model does allow for agents who are free to arbitrage pricing discrepancies. Nevertheless, such discrepancies remain

because arbitrage is risky, which bounds positions. Another issue is whether Atraders form a sufficiently significant mass

to affect prices in reality. In this regard, Barber, Odean, and Zhu (2009) and Kumar and Lee (2006) both provide evidence

that retail investors (who are more likely to be unsophisticated Atraders) do move markets. Further, there has been pressure

on several public pension funds to effectively become Atraders by pressures to avoid “sin”stock s and predilection toward

environmentally conscious stocks,

12

and it is well-known (see., e.g., Gompers & Metrick, 2001) that large institutions do

have an effect on prices.

This paper is organized as follows. Section 2 presents the basic one-security model with symmetric information. Sec-

tion 3 presents the multi-asset setup with information acquisition. Section 4 endogenizes participation by Atraders in the

stock market. Section 5 presents a dynamic extension. Section 6 concludes. All proofs of propositions and corollaries,

unless otherwise stated, appear in Appendix A.

2

|

A SIMPLE MODEL

We first consider a simple one-security model with symmetric information that conveys some basic intuition, before mov-

ing on to multiple securities and allowing for information acquisition. A single stock is traded at Date 0. At Date 1, the

stock pays off a liquidation dividend

V¼

Vþhþ:



Vis a positive constant, which represents the expected dividend. The variables hand represent exogenous technology

shocks; is not revealed until Date 1, but hcan be observed by investors at Date 0. These variables have zero mean and

are mutually independent and normally distributed. Throughout the paper, we denote the variance of any generic random

variable, g,bym

g

.

A mass mof agents, where m2[0, 1] is a positive constant, have a standard exponential utility function. Specifically,

for the i’th agent,

8

|

LUO AND SUBRAHMANYAM