The purpose of this paper is to emphasize the importance of selecting the efficiency evaluation model that best represents the known and postulated relation between inputs and outputs. By ignoring the input/output relationship, Mehrez, Brown, and Khouja (MBK) misspecify the efficiency evaluation model and misinterpret the distortions resulting from their misspecification as being paradoxical. This paper shows that the use of the appropriate data envelopment analysis (DEA) model reflecting the matched output/input (MOI) technology postulated by MBK will always provide efficiency estimates that satisfy their MOI axiom. Statistical consistency of the DEA estimator also implies that even with MBK's use of a misspecified model, the frequency of violation of their MOI axiom is likely to be small for large samples. Résumé. L'auteur insiste sur l'importance du choix du modèle de l'évaluation du rendement qui représente le mieux la relation connue et postulée entre les intrants et les extrants. En ignorant la relation intrant‐extrant, Mehrez, Brown et Khouja (MBK) définissent mal le modèle d'évaluation du rendement et jugent, à tort, les distorsions résultant de cette définition comme étant paradoxales. L'auteur démontre que le recours à l'analyse intégrale appropriée des données s'inspirant de la technologie de concordance extrant‐intrant postulée par MBK produit toujours des estimations du rendement qui respectent l'axiome de concordance. La cohérence statistique des estimations issues de l'analyse intégrale des données donne également à penser que même lorsqu'on utilise un modèle mal défini comme le font MBK, la fréquence des dérogations à l'axiome de concordance extrant‐intrant tend à être faible pour les gros échantillons.

[According to the December 1991 issue of the Survey of Current Business, expenditures of state and local governments account for more than 11 percent of the U.S. gross domestic product. Moody's 1991 Municipal Manual indicates that these governmental entities have an outstanding debt now approaching $800 billion, and a report by the Public Securities Association (1987) indicates that this debt grew at a compound annual rate of 12 percent from 1966 to 1986. State and local governmental activities continue to increase in magnitude, and evidently form an important part of the political and economic environment in which accounting operates. Important accountability issues distinctive to these organizations need accounting research attention. The articles by Feroz and Wilson and Deis and Giroux in this issue, which we have been invited to review, address some of these topics. The study by Feroz and Wilson can be regarded as an extension to the public sector of capital-market-based research that examines the effects of financial-accounting disclosures on security prices and returns. They hypothesize segmentation of the market for municipal obligations along national and regional lines and study the effects of differential information disclosure on borrowing costs. In the other study, Deis and Giroux utilize quality reviews that were conducted by the Texas Education Agency to evaluate and rate the audits (by public accountants) of public schools' financial reports. They test hypotheses about audit quality that were originally developed in the context of commercial firms. Both studies thus represent extensions of theories and methods used in research of private-sector accounting and auditing issues. The contributions of the two articles are discussed, and modifications that consider the unique aspects of governmental accounting are presented in sections I and II. Other possible avenues for research are discussed in section III.]

This paper analyzes a formal principal‐agents model of resource allocation and coordination in which demand for transfer pricing arises endogenously within a decentralized environment characterized by asymmetric information and divergence of preferences. It is shown that a modified Groves scheme achieves full information efficiency in a setting of the type considered by Harris, Kriebel, and Raviv (1982) and Cohen and Loeb (1984) only if the information asymmetry is postcontract and collusion is precluded. Conditions for the optimality of a coordination mechanism that is immune to collusion are also examined. It is shown that demand for a collusion‐free marginal cost‐type transfer pricing scheme arises if the agents are risk neutral, the cost function is separable but not necessarily linear, and the information asymmetry is postcontract. Résumé. Les auteurs analysent un modèle structuré d'affectation des ressources et de coordination mandant‐mandataire, dans lequel la demande de prix de cession interne est issue, de façon endogène, d'un contexte décentralisé caractérisé par une information asymétrique et une divergence des préférences. Les auteurs démontrent qu'un schéma Groves modifié permet d'atteindre l'efficacité maximum de l'information dans un contexte semblable à celui qu'utilisent Harris, Kriebel et Raviv (1982) et Cohen et Loeb (1984), uniquement si l'asymétrie de l'information est postérieure au contrat et si la collusion est rendue impossible. Ils examinent également les conditions d'optimalité d'un mécanisme de coordination qui est à l'abri de la collusion. Les auteurs démontrent qu'il y a demande de prix de cession interne, à l'abri de la collusion, du type coût marginal si les mandataires sont neutres à l'egard du risque, si la fonction de coûts peut être isolée sans être nécessairement linéaire, et si l'asymétrie de l'information est postérieure au contrat.

According to the December 1991 issue of the Survey of Current Business, expenditures of state and local governments account for more than 11 percent of the U.S. gross domestic product. Moody's 1991 Municipal Manual indicates that these govern- mental entities have an outstanding debt now approaching $800 billion, and a report by the Public Securities Association (1987) indicates that this debt grew at a compound annual rate of 12 percent from 1966 to 1986. State and local governmental activities continue to increase in magnitude, and evidently form an important part of the political and economic environment in which accounting operates. Important accountability issues distinctive to these organizations need accounting research attention. The articles by Feroz and Wilson and Deis and Giroux in this issue, which we have been invited to review, address some of these topics. The study by Feroz and Wilson can be regarded as an extension to the public sector of capital-market-based research that examines the effects of financial-accounting disclosures on security prices and returns. They hypothesize segmentation of the market for municipal obligations along national and regional lines and study the effects of differential information disclosure on borrowing costs. In the other study, Deis and Giroux utilize quality reviews that were conducted by the Texas Education Agency to evaluate and rate the audits (by public accountants) of public schools' financial reports. They test hypotheses about audit quality that were originally developed in the context of commercial firms. Both studies thus represent extensions of theories and methods used in research of private- sector accounting and auditing issues. The contributions of the two articles are discussed, and modifications that consider the unique aspects of governmental accounting are presented in sections I and II. Other possible avenues for research are discussed in section III.