Stock market anomalies and baseball cards

• Date: 01 August 2020
• Published date: 01 August 2020
• Author: Linh Thompson,Jared Williams,Joseph Engelberg
• DOI: http://doi.org/10.1111/fire.12223

DOI: 10.1111/fire.12223

ORIGINAL ARTICLE

Stock market anomalies and baseball cards

Joseph Engelberg1Linh Thompson2Jared Williams3

1Rady School of Management, University of

California-San Diego, La Jolla, California

2College of Business Administration,The

University of Texasat El Paso, El Paso, Texas

3Muma College of Business, University of South

Florida, Tampa,Florida

Correspondence

JaredWilliams, Muma College of Business,

Universityof South Florida, 4202 East Fowler

Ave,Tampa,FL 33620.

Email:jwilliams25@usf.edu

Abstract

Baseball cards exhibit anomalies that are analogous to those that

are documented in financial markets, namely,momentum, price drift

in the direction of past fundamental performance, and initial pub-

lic offering (IPO) underperformance. Momentum is higher among

active playersthan retired players, and among newer sets than older

sets. Regarding IPO underperformance, we find that newly issued

rookie cards underperform newly issued cards of veteran players,

and that newly issued sets underperform older sets. The results are

broadly consistent with models of slow information diffusion and

short-selling constraints.

KEYWORDS

baseball cards, IPO underperformance, momentum, post-earnings-

announcement drift

JEL CLASSIFICATIONS

G12, G14, G40

1INTRODUCTION

Basicfinancial theory predicts that stock returns should be unpredictable after adjusting for risks. That is, any abnormal

returns earned by tradingstrategies are either due to random chance or they are compensation for risks that investors

care about that are not captured by the asset pricing model. However,financial economists have documented the exis-

tence of simple strategies that earn unusually high or low returns even though the strategies do not load heavily on

common risk factors. For example, Jegadeesh and Titman (1993) document stock price “momentum,” that is, the ten-

dency for stocks that have performed well in the past 3–12 months to continue outperforming stocks that have per-

formed poorly in the past 3–12 months.

Concerning momentum, the literature provides a plethoraof possible explanations for the phenomenon. Several of

these theories are consistent with optimal investorbehavior. Such theories rely on features that are unique to financial

markets such as time-varying expected dividend growth rates(Johnson, 2002), firms’ growth options (Berk, Green, &

Naik, 1999; Sagi & Seasholes, 2007), and mutual fund flow inertia (Vayanos & Woolley, 2013). Additionally, there are

behavioral theories based on suboptimal investor behavior (Barberis, Shleifer,& Vishny, 1998; Daniel, Hirshleifer, &

Subrahmanyam, 1998; Hong & Stein, 1999). Restricting oneself to financial data, it is difficult to test whether any one

Financial Review.2020;55:461–479. wileyonlinelibrary.com/journal/fire c

2019 The Eastern Finance Association 461

462 ENGELBERG ETAL.

of these theories are a source of momentum, because all of the theories apply to financial data. It is thus difficult to say

that momentum is explained by one of the theories but not the others. The point of this study is to use a nonfinancial

laboratory in which some, but not all, of these theories should apply.If we find momentum, this would be evidence that

momentum can exist naturallyin markets without the bells and whistles of dynamic growth rates, dividends, or mutual

funds.

Our laboratory is the market for baseball cards. Baseball cards have a long history,dating all the way back to the

late 1860s (Jamieson, 2011). By 1991, sales of baseball cards reached $1.2 billion annually (Jamieson, 2011). Although

baseball cards produce no cash flows, their market values can be substantial. For example, the T206 Honus Wagner,

which was produced from 1909 to 1911, has been sold for as much as $2.8 million.1Because there havebeen long peri-

ods of time overwhich their values have appreciated, baseball cards have often been perceived as investment vehicles.

This perception has been fueled in part by the popular press: in 1988, the New YorkTimes published an article on base-

ball cards noting that over the previous decade, cards’ values had risen 32% per year.2

Many theories that financial economists havedeveloped to explain stock price momentum do not apply in this mar-

ket. There are no growth options, dividends, or mutual funds.3Nevertheless, we find that there is momentum in this

market,and that it is substantially stronger than it is in financial markets. Whereas momentum strategies earn less than

1% per month among stocks, momentum strategiesof baseball cards earn up to 5.6% per month. If momentum is driven

by behavioralbiases, it should not be surprising that it is significantly stronger in this market than it is in financial mar-

kets.In the stock market, there are many hedge funds that can arbitrage away inefficiencies and keep prices in line with

fundamentals. In the baseball card market, there are dealers who are relatively sophisticated, but much of the activity

in this market is driven by children. Moreover,while it is common to short stocks, there is little (if any) short selling of

baseball cards. Hence, the opportunity for arbitrage is severelylimited in the baseball card market.

Two behavioral theories of momentum generate testable predictions in this environment: Daniel et al. (1998)

(henceforth DHS) and Hong and Stein (1999).4DHS posit that investors are overconfidentbecause they overestimate

the precision of private signals they receive about an asset’s fundamental value. This causes prices to overreact to the

private information, which results in long-run price reversals. Momentum arises in their model when investors have

biased self-attribution. In this version of their model, investors become even more confident in the precision of their

private signals whenever public information is consistent with the private signals they observe. In the baseball card

market, collectors might obtain private signals about a player’sability by watching him play.5In addition, to the extent

that people cannot process all of the information contained in (public) box scores, a collector’s knowledge of a player’s

current batting average and other statistics can be considered “private”information. Whereas momentum only arises

in the biased self-attribution version of their model, long-run reversals are an unambiguous prediction of DHS. We

examine whether there is anyevidence of long-run reversals, but we find none. When the formation period is 2, 3, and

4 years long, momentum returns remain positive (though often insignificant).

We turn our attention to Hong and Stein’s (1999) model of momentum. In their model, momentum arises because

information gradually diffuses across the investor population.6That is, at each time t, there is information release,

and different portions of the population observe different pieces of the information at different times—it is only at a

1Source:http://www.latimes.com/sports/sportsnow/la-sp-sn-honus-wagner-card-20150427-story.html

2Source:http://www.nytimes.com/1988/11/13/business/potpourri-a-grand-slam-profit-may-be-in-the-cards.html

3Goetzmannand Huang (2018) document momentum in imperial Russia, which also lacked mutual funds.

4Barberiset al. (1998) also develope a model that can generate both short-run momentum and long-run reversals. However,for some parameter values, there

is momentum but no reversals, and for other parametervalues, there are reversals but no momentum. Hence, it is difficult to develop concrete predictions

fromtheir model in our setting.

5Forexample, some hits are the result of random chance and/orluck, and they are unlikely to be predictive of future success. Conversely, some at-bats result in

outsdue to bad luck (e.g., an extremely good play by the defense). This information does not appear in the game’s publicly observable box scores or the publicly

observablestatistics about the player’s performance.

6Like DHS, Hong and Stein (1999) developetwo versions of their model. In DHS, long-run reversals are predicted in both models, whereas momentum only

arises in the special case where investorshave biased self-attribution. In Hong and Stein (1999), momentum is predicted in both versions, whereas long-run

reversalsare only predicted in the special case where there are “momentum traders.” We restrict attention to the baseline version of their model.