We demonstrate how to estimate a model of oligopoly pricing when products are multi-dimensionally differentiated. We provide an empirical counterpart to recent theoretical work on product differentiation. Using specifications informed by economic theory, we estimate price-cost margins for products differentiated in many dimensions.
Quarterly Journal of Economics1995110(3), 837-855open access
The determinants of bargaining power and price formation in a dynamic exchange market where new traders enter randomly over time are studied. When agents on the long side of the market possess the option to wait for the arrival of future partners, the terms of trade in the spot market must honor the value of this option. The equilibrium terms of trade are expressed in intuitive closed-form equations that highlight the distinct influences of short-run spot-market conditions and long-run market demographics.
[The relation between theorized components of the bid-ask spread and trade size for a sample of NYSE firms is examined. We find that the adverse selection component increases uniformly with trade size. Conversely, order processing costs decrease with increases in trade size for all but the largest trades. We find that order persistence decreases with trade size. The adverse selection component is highest at the beginning of the day and lowest at the end of the day for all but the largest trades. Trades of NYSE firms executed on regional exchanges or NASDAQ contain a large order processing cost component but no significant adverse information effect.]
[In this article, we suggest an efficient method of approximating a general, multivariate log-normal distribution by a multivariate binomial process. There are two important features of such multivariate distributions. First, the state variables may have volatilities that change over time. Second, the two or more relevant state variables involved may covary with each other in a specified manner, with a time-varying covariance structure. We discuss the asymptotic properties of the resulting processes and show how the methodology can be used to value a complex, multiple exerciseable option whose payoff depends on the prices of two assets.]
Using an econometric model of charitable giving and a ten-year panel of tax return data, the author finds that previous studies have underestimated the effects of permanent income and overestimated those of permanent changes in tax prices. The significant statutory tax changes that occurred during the 1980s, especially in 1986, serve to identify the key model parameters. The author's results imply that people smooth their giving when transitory income changes but also time their giving to exploit transitory changes in tax prices. The results also raise questions about how effectively the tax incentives permanently influence the level of charitable giving by individuals. Copyright 1995 by University of Chicago Press.
Using an econometric model of charitable giving and a 10-year panel of tax return data, I find that previous studies have underestimated the effects of permanent income and overestimated the effects of permanent changes in tax prices. The significant statutory tax changes that occurred during the 1980s, especially in 1986, serve to identify the key model parameters. My results imply that people smooth their giving when transitory income changes but also time their giving to exploit transitory changes in tax prices. The results also raise questions about how effectively the tax incentives permanently influence the level, rather than just the timing, of charitable giving by individuals.
Abstract. This study distinguishes between incremental and relative information content. Incremental comparisons ask whether one accounting measure provides information content beyond that provided by another, and apply when one measure is viewed as given and an assessment is desired regarding the incremental contribution of another (e.g., a supplemental disclosure). Relative comparisons ask which measure has greater information content, and apply when making mutually exclusive choices among alternatives, or when rankings by information content are desired (e.g., when comparing alternative disclosures). Questions of both incremental and relative information content arise frequently in accounting. However, few previous studies have examined questions of relative information content. Possible explanations include unfamiliarity with the relative versus incremental distinction, and the additional statistical complexity involved in testing for relative information content. First, we examine analytically the relation between incremental and relative information content, demonstrating that they address different research questions and require different tests for statistical significance. Second, we identify accounting research contexts in which questions of relative and incremental information content arise. Third, we propose a new regression‐based test for relative information content. This test applies to both returns and valuation studies, generalizes to any number of predictor variables, and can be used in conjunction with White's (1980) adjustment for heteroskedasticity. Fourth, we illustrate tests for relative and incremental information content in a familiar research setting that compares the information contents of net income, cash flows, and net sales in 40 industries. Résumé. Les auteurs établissent la distinction entre le contenu marginal et relatif en information. Les comparaisons du contenu marginal visent à déterminer si une mesure comptable livre un contenu en information supérieur à celui d'une autre, et elles s'appliquent dans les cas où une mesure est considérée comme étant donnée et où l'on souhaite évaluer la contribution marginale d'une autre mesure (par exemple, la présentation de renseignements complémentaires). Les comparaisons du contenu relatif visent à déterminer quelle mesure livre un contenu en information supérieur , et elles s'appliquent dans les cas où les choix entre différentes possibilités sont mutuellement exclusifs ou lorsqu'il s'agit d'établir des classifications selon le contenu en information (par exemple, dans la comparaison de différentes possibilités en ce qui a trait à la présentation d'information). La question du contenu marginal aussi bien que du contenu relatif en information se pose fréquemment en comptabilité. Dans les études réalisées jusqu'à maintenant, l'on s'est peu intéressé aux questions touchant le contenu relatif en information. Cette lacune pourrait s'expliquer, entre autres, par le caractère inusité de la distinction entre contenu relatif et contenu marginal en information et par la complexité statistique accrue que supposent des tests portant sur le contenu relatif en information. Selon un schéma analytique, les auteurs examinent en premier lieu la relation entre le contenu marginal et le contenu relatif en information, en démontrant que les mesures correspondantes se rapportent à des questions de recherche différentes et qu'elles exigent des tests de signification statistique différents. Ils déterminent en deuxième lieu quels sont les contextes de recherche comptable dans lesquels se posent les questions pertinentes au contenu relatif et marginal en information. Ils proposent, en troisième lieu, un test inédit fondé sur la régression et s'appliquant au contenu relatif en information. Ce test convient à la fois aux études de rendement et aux études d'évaluation, peut être généralisé à n'importe quel nombre de variables prédictives et peut être utilisé en association avec l'ajustement de White (1980) relatif à l'hétéroscédasticité. Ils illustrent, en quatrième lieu, l'application de tests du contenu relatif et marginal en information dans un contexte de recherche familier où le contenu informationnel du bénéfice net, des flux de trésorerie et du chiffre d'affaires net est comparé dans 40 entreprises.
Using the marginal value of characteristics, we show how to construct bounds on the exact hedonic price index. When prices are above marginal costs then our bounds still apply, but the value of characteristics cannot be measured so easily from a hedonic regression. Since the price--cost markups are an omitted variable, they will bias the coefficients obtained. For a special class of utility functions, we argue that a linear regression will still provide a measure of the marginal value of characteristics, but a log-linear regression will overstate these values.
The Review of Economics and Statistics199577(3), 486
Empirical estimation of input market power hindered by problems in measuring an input's value of marginal product (VMP). By estimating a variable profit function system, however, one can infer a factor's VMP through its shadow price. This technique is used here to specify a structural equation system, which is estimated using time series data for the U.S. sawmilling and paper industries, to empirically measure the degree of oligopsony power for sawlog and pulpwood inputs respectively. Results evaluated at sample means indicate that pulpwood markets are more oligopsonistic than sawlog markets, though both perform closer to perfect competition than monopsony. Time trends for market power differ for each product and perfect competition cannot be rejected for sawlogs in later years. Copyright 1995 by MIT Press.