Knowledge that Transforms

[Errors arise in measuring changes in prices of assets due to imperfection and incompleteness of asset markets. Furthermore, the rates of price-change, and the magnitudes of errors of measurement vary and are often correlated across assets. Suppose we characterize an economy by means and variances of price changes for individual goods and of measurement errors in these changes as well as by the degree of diversification in the asset portfolios held by individual firms. In such an economy, the linear valuation rule that yields the most efficient estimate of change in the economic value of these asset portfolios is the one that minimizes the mean squared error (MSE). This paper presents a linear aggregation model of valuation to help understand how the minimum MSE valuation rule is affected by various parameters that characterize the economy, and the circumstances under which historical-cost valuation rule yields a (statistically) more precise estimate of the unobserved economic value of firms' assets than the current valuation rule. The analytical findings of the paper are consistent with the reluctance of accountants to depart from historical cost in spite of the existence of low inflation, and in spite of scholarly critiques of this valuation rule by Chambers (1966), Edwards and Bell (1961), Sterling (1970) and others. They are also consistent with the use of specific price indexes by most firms to prepare SFAS 33 disclosures. Several testable implicatons of the results are provided. A direct comparison of the characteristics of valuation rules is complicated by the heterogeneity of the decision contexts in which accounting numbers are used. We use the mean squared error (MSE) between the principal value and its various estimators to rank the latter. Using this criterion, previous simpler models that ignore the presence of measurement errors in price changes have shown that the use of increasingly detailed price indexes yields more precise valuation; current valuation is the most precise valuation rule because it uses the most detailed set of indexes (Sunder 1978). We show that this basic result does not hold when the measurement of price changes is subject to errors. As the magnitude of these measurement errors increases relative to the magnitude of price changes, the most accurate valuation rule requires a less detailed set of price indexes. A key implication of this result is that the existence of inflation or deflation is not sufficient for general-price-level valuation, specific-price-index valuation, or current valuation to dominate historical-cost valuation as an estimator of the economic value of firms' assets. Historical-cost valuation is dominated by others only when the magnitude of price changes are large relative to the errors of measurement in price changes.]

[We examine risk, return, and abnormal return behavior in the days around quarterly earnings announcements, using a research design that allows risk to vary daily in event time. We test several hypotheses concerning the effect on security prices of earnings announcements per se (i.e., ignoring both the sign and the magnitude of earnings). The first hypothesis concerns the resolution of uncertainty over time. By conveying information about firms' activities, earnings announcements resolve some uncertainty about future cash flows, but the concurrent price reactions increase the variability and covariability of securities' returns during the announcements. Thus, it is hypothesized that return variances and betas, and therefore expected returns, increase during earnings announcement periods (Stapleton and Subrahmanyam 1979; Epstein and Turnbull 1980; Choi and Salamon 1989). Previous research has demonstrated anomalous positive abnormal returns during earnings announcements (Chambers and Penman 1984; Penman 1984, 1987; Chari et al. 1988). Because risk was not allowed to vary in event time in this research, it does not adequately distinguish between increased expected returns and true abnormal returns. We report that abnormal returns remain after controlling for risk increases at earnings announcements. The abnormal returns are not related to any over- or under-reaction by the market to earnings news (see, e.g., DeBondt and Thaler 1985, 1987; Bernard and Thomas 1989) because we do not condition on the earnings realization. The second hypothesis (the information hypothesis) is that the timing of an earnings announcement is informative because managers systematically announce good news early and bad news late (Givoly and Palmon 1982; Chambers and Penman 1984; Kross and Schroeder 1984). The hypothesis predicts that average abnornal returns: (1) are positive at the earnings announcement, (2) are negative prior to the announcement, and (3) cumulate to zero by the end of the announcement period. Our tests extend those of Chari et al. (1988), Kross and Schroeder (1984) and Chambers and Penman (1984) by examining the pattern of returns around earnings announcements for the population of stocks. The pattern we observe is not as predicted by the information hypothesis. Finally, we investigate whether cross-sectional variation in announcement-period risks and returns is a function of firm size, which is a proxy for the increase in information arrival during earnings announcement periods. The evidence reveals that, after controlling for risk increases, abnormal returns generally are positive and decreasing in firm size. For the smallest size decile, abnormal returns in the ten days up to and including the earnings announcement are approximately 1.75 percent in the average quarter, or approximately 7 percent over only 40 trading days per year. This adds to an impressive body of size-related anomalies. We use these results to reexamine Hand's (1990) reinterpretation of the functional fixation hypothesis. Hand investigated quarterly earnings that included previously announced book gains from debt-equity swaps. He distinguished between "sophisticated" and "unsophisticated" investors, hypothesizing that only the former correctly comprehend the different implications of swap gains and other components of earnings. He found that abnormal returns increase in a variable representing the interaction between the swap gain and a proxy for the probability that the marginal investor is unsophisticated. We are skeptical about both the hypothesis and whether it predicts the observed result. We interpret Hand's result as similar to the puzzling but typical size effect around earnings announcements. It seems unlikely to be due to swap gains, to the sign or magnitude of earnings information released at the time, to errors in measuring the earnings information released, or to functional fixation.]

[This study examines thrifts' supplemental disclosures with respect to default risk (scheduled items) and interest rate risk (repricing data). It represents an extension of a recent study (Beaver, Eger, Ryan, and Wolfson 1989; hereafter BERW), which also examines the relationship between banks' share prices and supplemental disclosures with respect to default risk (nonperforming loans) and interest rate risk (maturity data). However, there are differences between the research designs and the findings of the studies. The research design differences between the bank and thrift studies include: (1) the use of per share deflation instead of book value deflation; (2) the use of seemingly unrelated regressions (SUR), as well as a fixed-effects model; and (3) a partitioning of the sample according to availability of supplemental data to assess the impact of nondisclosure on the estimated coefficients. A sample of 165 publicly traded thrifts was used to regress market values on several variables, including supplemental disclosures. The findings differ from those of the bank study in several respects. First, the coefficient on the supplemental disclosure on interest rate risk is significant in the case of the banks but not in the case of thrifts. Second, the coefficient on the default risk variable is smaller for thrifts than for banks. Third, the following three additional findings extend the bank study. (1) Thrifts that do not disclose the default risk variable appear to be valued at a discount relative to disclosing thrifts. (2) The estimated effects of the incremental explanatory power of the default risk variable is robust with respect to two nonnested estimation methods, a fixed effects model, and a seemingly unrelated regression. (3) The finding regarding scheduled items is robust with respect to an alternative specification, which assesses differences in estimated coefficients for "good" loans versus "bad" loans.]

[Stratified sampling for inventories, as discussed in the literature, usually employs the recorded book value as the stratification variable (Neter and Loebbecke 1975; Roshwalb et al. 1987). Stratification by book value along with an appropriate estimator provides a cost effective and precise estimate of the total inventory cost. Other stratification variables, such as volume of activity, provide information not provided by the book value about the variability of the errors in inventories, but little is known about the effectiveness of designs based on them. Greater improvement in efficiency may be possible by simultaneously including book value and volume into one sample design. In this paper, the effectiveness of volume is investigated as a stratification variable. A bivariate stratification method based on a model-based sampling framework is developed which effectively folds both book value and volume into one collective and simple stratification scheme. Using a bivariate sample design is not without any cost, because it requires that both the book and the volume values for an item to be positive. However, another extension of the model is proposed to include all items with zero book value or zero volume into a single stratified design. To examine the effectiveness, sample designs based on book value alone, volume alone, and the bivariate method are constructed for three inventory populations. Since the errors are known for every item in these inventories, the exact efficiency of a sample design can be determined. By designing a sample to achieve a specified precision with known reliability, the resulting sample size is a measure of efficiency. The design for a population with the smallest sample size is the most efficient and cost effective. To test whether the sampling methodologies actually achieve their precision and reliability targets, the audit sampling process is repeated in a simulation. Random samples are repeatedly selected from the inventories using the alternative sample designs, estimates for the inventory total are calculated using each set of sample data, and the accuracy of the estimates are recorded. Designs based on the bivariate stratification method required up to 62.5 percent fewer items to achieve the same level of precision and reliability than designs based on book value alone or volume alone. The simulation results show that these bivariate stratified designs attain the same targer level of precision and reliability as designs based on book value alone or volume alone. In some cases, the bivariate designs improve the actual confidence interval coverage more than designs based on book value alone.]

[Using laboratory market data, this study demonstrates that consumers respond differently to a market event depending on the information reported about the event. Specifically, they respond more rapidly to an economically predicted price increase when they are informed that sellers' marginal costs have increased, but they resist price increases if they know that the sellers' profits have increased. These information effects are based on the principle of dual entitlements, which posits that purchase decisions are influenced not only by the direct economic utility of the purchase, but also by consumers' perceptions of the equity or fairness of a negotiated price. Survey evidence from prior studies indicates that consumers (buyers) justify price increases driven by increases in sellers' costs, but resist price increases that increase sellers' profits. This study goes beyond surveys to investigate these predictions in a market setting affected by an economic event that simultaneously increases both the marginal costs incurred and the profits earned by sellers. Specifically, we examine the combined effect of a change in the sellers' tax rate and tax base. Nine laboratory markets in three separate financial information structures were conducted to investigate the predicted information effects. Each market had ten traders (five buyers and five sellers), for a total of 90 subjects. In three markets, buyers were apprised of an increase in sellers' marginal tax costs. In three other markets, buyers were informed of an increase in sellers' after-tax profits. Finally, three control markets with no information disclosures served as a baseline. Subjects in each market were student volunteers who received their market profits in real cash, in conformance with the tenets of induced-value theory. The results have implications for the financial disclosures volunteered by firms or mandated by regulatory bodies. While the accounting literature has traditionally stressed information effects on investors (Lev 1989), the body of users affected by financial reporting is much larger and includes the consumers who purchase the goods and services of disclosing firms (Financial Accounting Standards Board 1978, par. 24). This study suggests that financial disclosures can influence consumer behavior in competitive markets for goods and services.]

[Firms sometimes file their 10-K or 10-Q with the Securities and Exchange Commission (SEC) several days before the corresponding earnings announcement appears in The Wall Street Journal (WSJ). In these cases, the 10-K or 10-Q filing constitutes the first public announcement of earnings. This study addresses the question of whether the price and volume reactions to these earnings announcements occur at the SEC filing date or at the subsequent WSJ announcement date. This research can be viewed as a limited test of whether the securities markets are efficient with respect to the data contained in 10-K and 10-Q filings. The generalizability of the test results is limited because of the nonrandom nature of the sample. Tests were performed on daily price and volume data for 342 firmquarters for which the SEC filing date preceded the WSJ earnings announcement by at least four trading days. The results suggest that there was no significant market reaction, on average, at the SEC filing date, even though the filing was the first public announcement of earnings for the quarter. However, there is evidence of the existence of a market reaction to the subsequent WSJ earnings announcement. The direction of the price reaction at the WSJ announcement date is consistent with the sign of unexpected earnings. Wright and Groff (1986) present evidence indicating that the existence of a market reaction to the public release of price-relevant information does not depend on exactly how that information is made public. In contrast, the results presented in this paper suggest that, in some limited cases, the method of disclosing accounting earnings is related to whether the information embodied in those earnings will be reflected in security prices in a timely fashion. The results also suggest that, at least in one specific set of circumstances, data disclosed as part of an SEC-mandated filing are not fully reflected in prices until a subsequent media disclosure is made.]

[The American Institute of Certified Public Accountants (AICPA) and its affiliated state societies promote restrictive accountancy laws that limit both the right to express opinions on financial statements and the use of certain occupational titles to licensed public accountants. Although occupational licensing, like other forms of government regulation, is justified as being in the "public interest," critics (e.g., Stigler 1971; Peltzman 1976) suggest that licensing arises because of the professional groups' interest in using the coercive power of government for their own economic advantage. Until 1979, CPAs were content to limit state regulation to the audit function, permitting unlicensed accountants to perform other accounting tasks. With the growing importance of review and compilation services, however, CPAs have sought to restrict the performance of these services too. In addition, the AICPA and state CPA societies have used their influence with state legislatures and licensing boards to impose limitations on the use of professional titles such as "public accountant," "accountant," and "auditor." Whereas some states have adopted relatively permissive licensing laws, others restrict all analytical work and professional titles to licensees. This study explains why some states have adopted more restrictive licensing regimes than others. Hypotheses are developed to test the power of interest groups, political systems, and socioeconomic variables in explaining such differences. The evidence, based on both univariate and multivariate techniques, supports the following general conclusions. Restrictive licensing regimes are more likely in states where the interest-group strength of CPAs is high, as measured by their numbers relative to public accountants who are not CPAs. Restrictiveness is inversely related to statewide competition between Republicans and Democrats and is slightly related to legislative turnover.]

[Firms' investment practices often seem to be suboptimal; for example, they may ration capital or underinvest in risky projects. Recent research (e.g., Antle and Eppen 1985, Antle and Fellingham 1990, Harris et al. 1982) argues that these practices might be optimal responses to contracting frictions that arise from information asymmetry among the members of the firm. Analyzing a two-person firm, these studies endow the manager with pre-decision information about the returns from a given project and examine the owner's optimal response to the information asymmetry. This article expands the problem in two ways. First, it considers the choice between two competing projects. Second, it explicates the source of the manager's information superiority: proximity to the firm's operations advantage the manager in both acquiring and knowing the cost-benefit of additional information pertinent to project choice. Because of the manager's advantage, the owner cannot infer why the project was selected. The owner's response to the information asymmetry must optimize over both the information-acquisition and resource-allocation decisions. This article considers a model in which the manager must choose between a risk-free and a risky project. The owner delegates project selection because of the manager's pre-contract information superiority and because only the manager can obtain post-contract additional, costly information. Delegating project choice, however, creates a coordination problem because of the owner's inability to observe whether the manager expended effort on information search and because the value of additional information depends on the manager's pre-contract information (i.e., on the manager's type). Using the observed variables (project choice and realized profit), the optimal contract must (1) motivate a subset of manager types to select the project after expending effort on information search, and (2) motivate the complementary set to condition project selection on their pre-contract information alone. The model considered here is related to that of Demski and Sappington (1987) which considers issues involved in contracting with an "expert" who is uniquely qualified to acquire pertinent information, and Lambert (1986) applying this concept to a project-selection setting. Pre-contract asymmetry about the value of search is the crucial difference between previous research and this study. Absent information asymmetry (in the first-best solution), the information-acquisition problem is separable from the resource-allocation problem. The owner asks for search only from those manager types for whom the value to search exceeds cost. (The set of such manager types is convex because the value to search is concave in manager type.) Also, the manager uses all available information to choose the more profitable project. However, the owner cannot attain this first-best solution without information about manager type and information search. The two problems cannot be decomposed. As a result, the optimal (second-best) contract always induces distortions in information acquisition and, for the case of the risk-averse manager, induces overinvestment in the risky project. Furthermore, communication of the search result is strictly valuable. The results suggest that optimal responses to information asymmetry among the members of the firm could result in practices that seem to disregard the value of information search and to use acquired information sub-optimally. The results also lend support to the use of accounting reports, such as budgets, in managerial performance evaluation.]