More powerful tests of a theory of choice of accounting methods and the effect of changes in these choices on equity values are provided. The power increase comes from efficiently aggregating results across studies. One conclusion is that at least six variables common to more than one study have explanatory power. These variables are managerial compensation, leverage, size, risk, and constraints on interest coverage and dividends. Another conclusion is that the posterior probability that the theory taken as a whole has explanatory power is close to one. This conclusion includes the effect of variables that only appear in one study.

This paper examines cross-sectional analysis procedures common to many market-based accounting research papers. Both the economic and econometric properties of ‘levels’ and ‘returns’ studies are discussed. Topics covered include the relations between the accounting studies and cash flow valuation models, the role of expectations of accounting variables, deflators, spurious inference, risk adjustment and its relation to growth, size and leverage, residual dependence, dependence among explanatory variables, and the effect of scale differences across firms. Major conclusions are that market value is the correct deflator in returns studies, and that levels and returns studies are economically but not econometrically equivalent.

[Accounting numbers are an integral part of the firm's formal and informal contracts (Watts 1974; Holthausen and Leftwich 1983; Watts and Zimmerman 1986; Ball 1989; Christie 1990). This contracting-based theory of accounting is based on the premise that managers choose particular accounting procedures either efficiently to maximize the value of the firm, or opportunistically to make the manager better off at the expense of some other contracting party (Holthausen 1990). The relative amounts of efficiency and opportunism depend on controls on managers' accounting discretion. Such controls include monitoring by the board of directors, competition from the product markets and from within the firm by other managers, and the discipline of the market for corporate control. It is difficult to determine whether managers make accounting choices to maximize firm-value. Empirical tests often assume opportunism and usually reject the null hypothesis of no association between accounting choice and firm-specific variables such as leverage (Christie 1990). However, many of the empirical regularities interpreted as evidence of opportunism can also be interpreted as occurring for efficiency reasons, which serves to confound these findings (Watts and Zimmerman 1986; Watts and Zimmerman 1990; Sweeney 1994). The few tests based on efficiency rationales find an association between the contracting variables and accounting choice (Zimmer 1986; Whittred 1987; Malmquist 1990; Mian and Smith 1990a). This paper measures the relative influences of efficiency and opportunism in accounting choice by examining a non-random sample that maximizes the probability of finding opportunism. Economics and finance studies document that corporate control actions such as tender offers and proxy fights discipline opportunistic managers. Therefore, we select a sample of takeover targets and examine this sample for evidence of managerial opportunism in choice of accounting procedures. A key assumption of our tests is that, prior to the control action, corporate control targets contain more non-value-maximizing managers than surviving firms in the same industry that were not targets of corporate control actions. This assumption allows us to use surviving firms' accounting procedures as the benchmark for the efficient accounting choice. In this application of Alchian's (1950) "economic Darwinism," surviving firms on average are more efficient than non-surviving firms. An unbiased estimate of accounting opportunism in this sample is an estimate of the upper bound on the amount of accounting opportunism in a random sample of firms. If there is little or no opportunism in our non-random sample, then for firms generally we can discount opportunism as an explanation for choice of the three accounting methods we study. Efforts to explain choice of accounting methods could then be redirected towards efficiency explanations. We measure opportunism by comparing the frequency of choice of income-increasing procedures by takeover targets with the corresponding frequency of their surviving industry peers. We find that takeover targets select income-increasing depreciation, inventory, and investment tax credit (ITC) accounting methods more frequently than their surviving industry peers in up to 11 years preceding the control action. Our analysis includes a multiple regression that controls for efficient choice of accounting methods. Our estimate of the upper bound on the frequency of opportunism is small relative to the mean choices of surviving firms. We conclude that some accounting opportunism exists in the takeover targets, but that efficiency is the more important explanation of accounting choice. However, while the frequency of opportunism is relatively small, the dollar effect on retained earnings of selecting an income-increasing method is large. For the median target firm, the after-tax effect on retained earnings if the firm had been on the alternative method is about 26 percent for depreciation, 9 percent for inventory, and around 5 percent for the ITC.]

Value maximization requires either that knowledge is transferred to those with the right to make decisions, or that decision rights are transferred to those who have the knowledge. A tradeoff of knowledge transfer costs and control costs is required. Characteristics of firms' investment opportunity sets (IOSs) that affect knowledge transfer costs and control costs are identified. Testable predictions about the relations between these characteristics and firms' decentralization decisions are developed and tested. The evidence presented is consistent with our predictions and is robust to different ways of measuring variables.

A number of recent research papers use two-stage procedures in lieu of a single multiple regression, in some cases purportedly as a solution to colinearity among independent variables. We demonstrate that, since collinearity is inherently a data problem rather than a statistical problem, no partitions of dependent or independent variables, orthogonal or otherwise, can provide insights into the relative influence of collinear variables. For the class of linear unbiased estimators this follows directly from the Gauss-Markov Theorem, but we demonstrate some of the results in detail as an aid to interpreting particular papers.