Therefore, the total returns of art investments can be decomposed into two components: psychic (or nonfinancial) returns and financial returns. The psychic returns include, but are not limited to, aesthetic returns and any other prestige and complementarity effects. To some extent, psychic returns are popularity premiums—that is, the premiums paid for pleasure and enlightenment, being admired or sought after. Financial returns relate to the change in the price of the art objects. The price changes can be actual market prices or changes in expert opinions. Financial returns are easier to quantify, of course, than psychic returns.
The key question is how to quantify the psychic returns. Much debate is going on in cultural economics about how to measure the psychic returns of art investments. The literature provides three ways to estimate the psychic returns in the art market: Jensen’s alpha, rental charge, and opportunity cost.
Jensen’s Alpha. Stein (1977) proposed that Jensen’s alpha should be taken as a measure of the returns from the viewing of an artwork. Chanel, Gérard-Varet, and Ginsburgh (1994) and Hodgson and Vorkink (2004) also associated the alpha estimate in the single-factor market model with a measure of psychic returns in art market investments.
This framework is based on the market model. In the market model, the returns on all securities are from two sources—a single market factor and idiosyncratic return. The result is the following regression equation:
R a – R f = α + β(R m – R f ) + ϵ ,
where
R a = return series on art investments
R f = return series on a risk-free asset
R m = return series on the market portfolio
β = sensitivity of the excess returns on art investment to the excess returns on the market portfolio
α = the part of the excess returns on an art investment that cannot be explained by its risk–return relationship with the market portfolio
ϵ = residual unsystematic and diversifiable risk
The psychic return is defined as the negative of Jensen’s alpha (i.e., –α). The logic is that if the investor had chosen to invest in securities with no psychic return instead of art, he or she would have earned a financial return higher by the amount of the alpha.
Empirical analyses typically estimate α to be negative, so psychic returns are positive. For example, Stein’s (1977) point estimate of α is –1.6%. Pesando (1993) studied the returns in the market for prints for the period 1972–1992 and separately examined returns on Picasso prints. Pesando’s estimation results for the market model yielded an α of –1.5% for the overall market for prints and an α of –1.2% for Picasso prints. Chanel et al. (1994) estimated the values of α to be close to –1.0%. Hodgson and Vorkink (2004) reported an estimate for α of –0.8%. In summary, when the market model framework is used, the estimated psychic return is in the range of 1%–2%.
Rental Charge. Renting or leasing a piece of art provides possession of the object without having ownership. Thus, the renter is not concerned with any changes in its market price. The renter is solely paying for viewing the object and enjoying any other intangible returns it yields.
Atukeren and Seçkin (2007) argued that the psychic returns from investing in artwork are the changes in their rental prices. The authors made use of the prices charged by a Canadian fine art company for its art rental services and calculated the implied psychic returns to be about 28% of the sale price (hammer price) in international auctions. In an alternative way, they followed Hodgson and Vorkink’s (2004) suggestion that the Jensen’s alpha captures the extent of net psychic returns. The evidence for alpha from the art market applications of the market model coupled with the transaction cost data from international art auctions also suggest that the psychic returns to investing in artwork might be about 28% of the sale price. Because transaction costs are quite large in art auctions, this factor can make a substantial difference in the value of the psychic returns derived from the market model framework.
Opportunity Cost. Candela, Castellani, and Pattitoni (2013) argued that using Jensen’s alpha as a way to measure psychic returns may be problematic when the assumptions of the market model do not hold. Applying an opportunity cost framework and the analytical tools of portfolio theory, they proposed a new psychic return measure, one that is not affected by the same issues as Jensen’s alpha. They applied this measure of psychic return to art investments and estimated psychic returns to be in the 1%–2% range, as found when the market model has been used.
Conclusion
To some extent, psychic returns are popularity premiums—premiums paid for pleasure and enlightenment, being admired or sought after. The estimated psychic return for art investments has ranged from 1% to 2%. Transaction costs are quite large in art auctions, and this factor can make a substantial difference in the psychic returns derived from the market model framework.
3. Popularity and Asset Pricing
The risk–return paradigm continues to dominate the way in which both academicians and investment professionals think about modeling and forecasting asset prices. 22 The cause is largely the influence of the capital asset pricing model (CAPM).
The CAPM is a logical application of neoclassical economics, which also led to the efficient market hypothesis in which markets are assumed to be efficient and investors are assumed to act rationally. The key assumptions of the CAPM are that (1) market participants act rationally, (2) markets are informationally efficient, (3) investors are risk averse, and (4) investors can diversify costlessly so as to eliminate all diversifiable (nonmarket) risk. These and other assumptions lead to the conclusion that only undiversifiable risk is compensated with a premium. The CAPM produces the simple formula that the expected return is equal to the risk-free rate of return plus the security’s beta (beta relative to the market) multiplied by a single premium for market risk.
After the development of the CAPM and the efficient market hypothesis, psychologists Tversky and Kahneman (1974) began to question the basic assumption that investors behave as rational agents (see also Kahneman and Tversky 1979 ). Behavioral finance, which Amos Tversky and Daniel Kahneman pioneered, has offered up a plethora of behavioral biases that lead to irrational behavior. Many of the biases seem to provide explanations for some of the documented ways in which observed security prices depart systematically from those that would exist in efficient markets.
Although behavioral finance tells a rich story, it has thus far not provided a full framework or theory for understanding asset prices. The CAPM and multifactor arbitrage pricing theory (APT) remain the baseline asset pricing models with which all other asset pricing models are compared.
Therefore, the analysis of investments needs a simple, coherent, and intuitive asset pricing framework for understanding and forecasting asset prices. We believe that any successful theory of asset pricing will start with the concept of popularity as we have discussed it.
In this chapter, we (1) continue to refine the popularity framework, (2) further explain the link of popularity to classical and behavioral finance, and (3) put forth a popularity-based asset pricing formula.
Refining the Popularity Framework
In Chapter 1 , we introduced the concept of popularity, which was first presented by Ibbotson and Idzorek (2014) , as an asset pricing concept that provides a unifying approach to explaining return premiums that is consistent with the risk–return framework and anomalies that are not consistent with efficient market explanations and are thus best explained by concepts in behavioral finance.
Our work started with the observation that assets represent bundles of characteristics that investors like or dislike. Conceptually, each characteristic has a supply and demand; high demand relative to supply is associated with high price, and vice versa. The price of an asset is formed by the aggregation of investor preferences. Assets with popular characteristics are expensive, and assets with unpopular characteristics are inexpensive.
The characteristics of an asset can change over time, as can investors
’ relative preferences. A popularity return premium goes to those who are willing to hold assets with unpopular characteristics. The premium is supplied by those willing to pay for assets with the most popular characteristics. Although individual security characteristics may migrate over time, the premiums for the unpopular characteristics themselves are relatively stable over the long term. In contrast, short-term popularity fads and distortions are interpreted here as mispricing.
The constituents of a given universe can be ranked by any characteristic or dimension of popularity—the most popular students in high school, presidential candidates, television shows, asset classes in the universe of investments, or stocks in a given market. The people/shows/asset classes/stocks with the most desirable characteristics will rank at the top, and those with undesirable characteristics will rank at the bottom. Today, reality stars and some other celebrities are famous for being famous or, in our preferred parlance, popular for being popular.
Different factors or characteristics influence popularity, but the overall popularity of the item in question can be thought of as the amalgamation of those various characteristics. Different investors make heterogeneous assessments of the benefits and costs of these characteristics, and the collective assessments result in a market-clearing price. This process relates directly to the basic principles of supply and demand: The most popular items are in short supply and high demand and, therefore, are dear.
Long-term asset pricing premiums may always be positive but can change in value. Short-term fads can lead to temporary mispricing. Finally, just as popularity is a naturally occurring behavioral phenomenon in which greater intrinsic value is attributed to popular items, so also is it natural that, within any universe for any characteristic or along any dimension of popularity, some of the most popular items will decrease in relative popularity over time while some of the least popular items will increase in relative popularity.
Precursors to the Popularity Approach
Popularity relates to a number of different bodies of literature, including efficient market equilibrium asset pricing, behavioral finance, and return premiums/anomalies. The popularity approach is most closely related to the New Equilibrium Theory (NET) put forth by Ibbotson, Diermeier, and Siegel (1984 ; hereafter, IDS), which we review in Chapter 4 . It also relates to what is referred to as “affect” in the behavioral finance literature, with a number of behavioral biases contributing to what we call popularity premiums .
According to NET, assets represent bundles of characteristics in which the cost of capital (expected return) for a given asset is the aggregation of the costs of all its characteristics. NET recognizes that systematic risks (say, from the CAPM or APT) affect asset prices, as do unsystematic risks that are costly to diversify (e.g., an individual’s human capital or house). NET also recognizes specific nonrisk characteristics: taxation, marketability, and information costs.
All else being equal, investors are willing to pay a premium price for an equivalent investment with a more desirable characteristic (e.g., the higher liquidity of an on-the-run US T-bond) even though it will have a lower expected return. Conversely, to hold an equivalent investment with a less desirable characteristic (e.g., the lower liquidity of real estate), investors expect a discounted price, resulting in a higher expected return.
Investors’ complex assessments of the numerous characteristics of an asset or investment drive asset pricing. NET is consistent with the classical view that investors are rational, but it goes beyond both the CAPM and APT view that only systematic risks drive asset prices. Popularity goes beyond NET to include anomalies associated with both rationality and irrationality. Today, we believe that both systematic/nonsystematic and rational/nonrational factors form the various dimensions of popularity that drive asset pricing.
Others have approached asset pricing similarly. From the perspective of a behavioral asset pricing model, Shefrin and Statman (1994) and Anginer and Statman (2010) found that stocks with greater benefits—what we would label “popular characteristics”—have lower expected returns. Statman and Glushkov (2011) expressed a perspective similar to that of NET and popularity:
What stocks do investors want? Many investors like large-cap stocks, growth stocks and, perhaps, stocks of “socially responsible” companies, such as those with good employee relations. Stocks with greater benefits fetch higher prices, and higher prices correspond to lower expected returns. (p. 5)
Surprisingly, in a 2016 interview, Eugene Fama, who was awarded the Nobel Prize in economics for developing the efficient market hypothesis, said the following:
Value stocks tend to be companies that have few investment opportunities and aren’t very profitable. Maybe people just don’t like that type of company. That to me has more appeal than a mispricing story, because mispricing, at least in the standard economic framework, should eventually correct itself, whereas taste can go on forever. (Fama and Thaler 2016 )
Popularity is an intuitive and naturally occurring behavioral phenomenon associated with being admired, sought after, well-known, and/or accepted. It is observed in countless settings. From a behavioral literature perspective, the idea of popularity is closely linked to affect , which Statman, Fisher, and Anginer (2008) described as the specific quality of goodness or badness. Affect, or sentiment , is closely linked to the intuitive concept of popularity in that it describes emotional or automatic feelings regarding an asset, investment, or company and the way those emotions influence cognitive decision making. Zajonc (1980) concluded that affect may have a stronger influence on decision making than do cognitive processes, with affect and cognitive processes under the control of partially independent systems. 23
Popularity goes somewhat beyond behavioral finance in that it explicitly accounts for rational preferences that may or may not be influenced by emotions. Specifically, affect focuses on emotional reactions and seems to exclude observable anomalies that are the result of rational preferences, such as greater liquidity or preferential tax treatment. Affect seems like a good contributor to popularity. Statman, Fisher, and Anginer (2008) and Anginer and Statman (2010) incorporated affect into a behavioral asset pricing model. But even without the affect heuristic, rational popularity premiums that are consistent with NET would still exist.
Efficient Markets, Behavioral Finance, or Something Else?
In Chapter 1 , we explained how popularity straddles classical and behavioral finance. We believe that popularity can be consistent with both camps; thus, it can be thought of as a unifying asset pricing theory. The formalization of the theory of popularity moves beyond the paradigm that more return requires more risk (e.g., CAPM or APT) to an enriched framework in which relative popularity drives returns.
From an efficient market perspective, one can take the view not only that the market is efficient at pricing risk but also that a number of other characteristics are being priced by investors with their heterogeneous attitudes toward those characteristics, as expressed in NET. Since popularity is a social phenomenon, the popularity approach seems to emerge from the behavioral finance perspective. Investors fail to make nearly instantaneously rational decisions for a variety of reasons: affect, lack of attention, loss aversion, overconfidence, anchoring, mental accounting, and so on. For example, investors who are overly confident may go after the most popular stocks and end up driving the price way up. If these biases are only temporarily connected with a security, they result in mispricing; however, more-permanent biases related to groups of securities can result in long-term premiums (e.g., the value effect). In both the classical and behavioral interpretations of popularity, in the long run, the winners hold the unpopular stocks and those who hold popular stocks are willing or unknowing losers.
The main idea of the efficient market hypothesis is that all relevant information about the value of securities is reflected in the market prices of securities. Hence, the market price of each security is “fair” and reflects its “intrinsic value.” This makes active manag
ement futile. However, if the market is inefficient, not all relevant information is reflected in market prices. Hence, some securities are underpriced and some are overpriced relative to their unknown fair values. If prices tend toward fair values, active management can succeed if active managers can estimate fair values with some accuracy.
Popularity theory suggests that the efficient-versus-inefficient, dichotomous view of market efficiency is inadequate. The failure of all relevant information to be reflected in market prices is not the only manner in which prices may not be fair. Prices may also reflect irrelevant information, such as the behavioral preferences of investors. In such a case, rather than being inefficient in the usual sense, the market can be said to be “beyond efficient.” 24 In an efficient market, prices are fair, but in a market that is beyond efficient, prices are “biased” because of investor preferences. To be precise, a biased security price is a price that reflects both relevant and irrelevant information in a beyond-efficient market. The bias in the price of a security is the percentage difference between the price of the security in a beyond-efficient market and its price in an efficient market. The bias could be positive or negative. In Chapter 5 , we present a formal model of a market that is beyond efficient with such biased prices.
Popularity Page 6