ارزش دفتری: قیمت و کیفیت موقتی تبعیض در بازار ایالات متحده برای ارزش دفتری

Publishers produce books in hardcover and paperback versions with different prices and time of market introduction. Analysis of detailed book-level data reveals that (i) price-cost differentials cannot be explained by cost differences, making this an example of quality discrimination; (ii) market introduction time strongly affects sales, suggesting that time is the crucial dimension of discrimination; and (iii) there is substantial price rigidity across books and over time: prices depend on cost shifters but not on demand shifters. I discuss an explanation for this last finding in terms of the nature of demand in this market.

Books are a familiar example of intertemporal price discrimination and are frequently cited as such in the economic literature.1 Publishers use publication time and binding quality as ways of discriminating among consumers with heterogeneous valuations of those attributes. They typically publish a new title in hardcover format first, then follow it up with a paperback version several months (or even years) later. There is usually a large price differential between hardcover and paperback which, most people are convinced, does not reflect differences in production costs. Despite its broad acceptance, this conventional wisdom has never been corroborated with hard evidence. Doing so is the first step towards the broader goal of this paper, which is to analyze pricing strategies and the structure of demand in the book market. The analysis exploits a new and extensive dataset provided by Yale University Press. The dataset includes prices and sales for the hardcover and paperback versions of each book and identifies the pattern of introduction of the different versions into the market. Moreover, information on fixed and variable costs of production is available for a subset of the data. The econometric evidence from this study suggests that book prices exhibit two types of rigidity. One is intertemporal rigidity: prices are rarely changed over time regardless of demand fluctuations. The other is rigidity across books: once we control for observable physical characteristics such as size and binding quality, there is little price variation left. In particular, prices do not appear to depend on the pattern of market introduction or the author’s previous publishing record. Sales, on the other hand, are strongly affected by these factors. This leads to an apparent puzzle: why do prices depend on cost-related demand shifters but not on ‘pure’ (not cost-related) demand shifters? I propose an explanation for this result in terms of the nature of demand in this market. The introduction of an inferior version of an established product is common practice in many markets. New movies are usually released first in luxurious and expensive first-run theaters; several months later they hit the second-run theater circuit and are also made available on videotape. Patient consumers can therefore enjoy them at a significant discount. Other examples are bakers, who often sell their day-old bread at half-price, and fashion designers, who channel unsold items to stores that specialize in marked-down designer-wear. In none of these cases can the price drop be attributed to changing costs. The same is not true for, say, computer prices. Even though the price of a new computer falls substantially within months of its introduction, one can argue that it may be falling costs that drive much of this price decrease. This argument cannot be made for books, movies, bread, or designer clothing.2 The analysis in this paper provides insights into the operation of those markets and contributes to our understanding of product placement, marketing, and intertemporal pricing of new products. The rest of this paper is organized as follows. Section 2 provides an overview of the industry and introduces the dataset. Section 3 describes the cost data and tests for price discrimination. Section 4 presents an econometric analysis of demand and publisher behavior in this market. Section 5 discusses the findings and proposes an explanation for the observed patterns of behavior. Finally, Section 6 summarizes and concludes.