Academic journal article
By Dwyer, Gerald; Barnhart, Cora
Federal Reserve Bank of Atlanta, Working Paper Series , Vol. 2002, No. 15
Abstract: We examine the distribution of returns in new industries to determine whether stocks in new industries are similar to lottery tickets. We focus on one characteristic of lottery tickets: negative expected returns. We examine data from the United States on sellers of own-brand personal computers, airlines and airplane manufacturers, automobile manufacturers, railroads, and telegraphs. A relatively small number of companies generate outstanding returns in some industries. We find no evidence of low expected returns. On the contrary, firms in new industries typically have high volatility of individual stocks' returns and high expected returns relative to indexes for the same periods. None of our evidence suggests that investors reasonably might expect to pay to play when investing in new industries.
JEL classification: G10, G12
Keywords: stock prices, returns, lottery
The prices of Internet stocks seemed extraordinary before the crash in prices in 2001, and many still seem extraordinary relative to likely or even plausible outcomes. One explanation is craziness or euphoria, an explanation given more plausibility by the crash in prices. On the other hand, the predictability of the valuation of individual Internet firms may make this explanation less attractive than it would otherwise [Hand 2000].
An alternative explanation is that Internet stocks are similar to lottery tickets. This possibility has occurred to many observers, one prominent person being Alan Greenspan who was quoted as saying "When you are dealing with stock the possibilities of which are either it's going to be valued at zero or some huge number--you get a premium in that stock price which is exactly the same sort of price evaluation process that goes on in the lottery." [Blumberg Capital 1999]. Perhaps Internet stocks have a distribution of returns that includes the possibility of large gains and small losses? If so, then expected utility-maximizing investors can be locally risk-averse and simultaneously willing to pay for a probability of a large gain for the same reasons that people are willing to pay for lottery tickets. This explanation could be consistent with high valuations and even negative expected returns because investors are willing to pay on average to have a small probability of a large gain. This argument is distinct from the point made by Schwartz and Moon [2000, 2001] that it is important to correctly allow for the dramatic importance of the volatility of returns, although their point is a crucial part of a correct analysis [Fisher 2002].
On one level, it is not possible to test the proposition that people perceive stocks in Internet firms to be similar to lottery tickets. If investors have an unfounded belief that Internet stocks are like lottery tickets, then even the subsequent evolution of the industry will not provide evidence on whether people thought the stocks were like lottery tickets. On the other hand, subsequent returns on average will be related to what people expect if people's expectations are rational in the sense of Muth. Ten or 20 years from now, Internet firms will provide one observation on whether Internet firms are like lottery tickets.
If investors' expectations are consistent with prior experience, the likely distribution of returns can be examined using the distribution of returns in earlier new industries. Many analogies have been drawn between Internet stocks and earlier industries as far back as railroads and telegraphs, e.g. in Business Week [Mandel 1999, Hof and Hamm 2002] and the Economist [Economist 2000]. (1) If the distribution of returns in prior new industries does have a few large positive-return firms and many firms with smaller, negative returns, this would provide support for the proposition that investors are basing their decisions on such a distribution for Internet firms. If the expected return in these industries also is negative, then the lottery explanation acquires some plausibility. …