Knowledge that Transforms

[This study examines the effect of the uncertainty in analysts' earnings forecasts on the relation between unexpected returns and unexpected earnings. Numerous theorists have considered the effect of uncertainty on firm value, with particular interest in the uncertainty in a firm's future cash flows that underpin firm value. Since accrual accounting earnings represent a theoretical proxy for future cash flows, the effect of earnings uncertainty on firm value is also of considerable interest. However, observed uncertainty in accounting earnings may be attributable to noise (garbling) in the earnings signal or to the fundamental uncertainty of expected future cash flows, or both. Moreover, theory suggests these different forms of uncertainty may have differing effects on firm value. To date, there is little empirical evidence concerning the effect of uncertainty in earnings on firm value. We investigate the effect of ex ante earnings uncertainty by using the familiar linear relation between unexpected stock returns (UR) and unexpected earnings (UE), with 3,167 firm-year observations collected over the six-year period 1979-84. The variance in analysts' earnings forecasts just prior to a firm's annual earnings announcement is employed as our firm-specific proxy for ex ante uncertainty. Our results indicate a systematic relation between ex ante uncertainty and the information content of earnings. A given unit of earnings news has a greater effect on unexpected stock price change as the amount of pre-earnings-announcement uncertainty decreases. Firms with relatively high ex ante uncertainty exhibit little or no systematic price change at the time earnings are announced. Sensitivity tests reveal our results to be robust over numerous alternative specifications of the variables and models employed. Sensitivity tests also suggest that our results are not driven by either firm size or the amount of information available about the firm. In addition, we develop and report results of a model that controls for the effects of uncertainty. This results imply that the dispersion (disagreement) in analysts'earnings forecasts is more likely to be a proxy for noise in the financial reporting system than a proxy for fundamental uncertainty in a firm's future cash flows.]

[Defense firms typically produce a large number of products. The purpose of this article is to explain how two features of the current regulatory process create a significant incentive for these multiple-product firms to choose inefficient production methods. The first feature is that the marginal impact of accounting cost on price varies significantly among products. Prices for a defense firm's products are set according to a rather unique process that combines elements of both competition and cost-based regulation. Defense firms typically produce some purely commercial products and prices for these products are competitively determined. Aside from standard off-the-shelf items such as army boots, most defense products are purchased from a sole source and thus their prices are nominally cost-based. In reality, the negotiated price is likely to be affected by other factors as well. In particular, in cases where closer substitutes exist or where an alternative source might not be prohibitively expensive, the potential cost of these alternatives plays a role. The important consequence of this is that the negotiated price will not necessarily decline or rise by a full dollar when the projected cost of production declines or rises by a dollar. In more competitive procurements where the cost of alternatives plays a stronger role, changes in projected accounting cost are less important. The second feature of the regulatory process concerns the method that defense firms are allowed to use to calculate the cost of each product. Following traditional commercial accounting practices, only a relatively small fraction of costs are directly charged to products. The remaining costs are grouped together into overhead pools and allocated across products usually in proportion to directly charged labor use. These two features create the following incentive problem. Given the first feature, the firm would like to be able to assign more of its costs to well-funded sole source procurements instead of to more competitive procurements or commercial products. The second feature provides a method for accomplishing this task. Namely, the firm can increase (decrease) the amount of overhead allocated to a contract by increasing (decreasing) the amount of direct labor used on the contract. This means that the firm will have an incentive to engage in pure waste by padding direct labor usage on contracts with cost sensitive revenues. It will also have the incentive to distort its input substitution decisions between labor and other inputs by using too much (too little) direct labor on contracts with cost sensitive (cost insensitive) revenues. Two major types of input substitutes for labor exist. The first is capital. Thus, we would expect the firm to purposely under-capitalize production of products with cost sensitive revenues and over-capitalize production of products with cost insensitive revenues. The second possible input substitute is material. For many subcomponents of a weapon, a firm has the potential option of subcontracting production to another firm or making the component in-house. Subcontracting will result in higher direct material costs for the firm but lower direct labor costs. Thus, engaging in more in-house production is essentially a way of substituting towards direct labor and away from direct material. In particular, then, we would expect the firm to purposely engage in too much in-house production for its products with cost sensitive revenue and too much subcontracting for its products with cost insensitive revenue. An important point to note about this incentive effect is that it does not require the firm to report any cost projections untruthfully. That is, in the behavior predicted by this article, the firm does not make money by projecting that costs will be high (in order to get a high price) and then actually having low costs. The firm actually spends all of the money that is charged as a cost. The profit occurs through shifting the assignment of these costs. The importance of this point is that auditing is very poorly equipped to deal with this type of behavior. Auditing is relatively good at determining whether the firm actually spent as much as it projected. However, it is relatively poor at determining whether any expenditure that actually occurred was necessary. Braeutigam and Panzar (1989), Brennan (1990), and Sweeney (1982) have analyzed models of public utility regulation where the utility has commercial business segments. They make the general point that, depending upon how costs are allocated, the firm may have an incentive to distort its output and/or input decisions in order to shift overhead to the regulated sector. However, none of these articles analyzes allocation schemes based on direct labor or any other input base. Braeutigam and Panzar (1989) and Sweeney (1982) consider allocation schemes based on units of output under the assumption that comparable units of output exist across different products. Brennan (1990) considers allocation schemes where each product is allocated a fixed, invariant share of overhead. Thus all of the predictions of this article regarding the particular sorts of input distortions one would expect to see in defense procurement are new to this article. Furthermore, on a technical level, the model of this article is also somewhat different because it considers a multiple product case where products are not necessarily either perfectly competitive or perfectly regulated and the level of competitiveness varies from product to product.]

[The results of prior research on Statement of Financial Accounting Standards No. 33 (FASB 1979) and the effects of required disclosure of inflation have been mixed. Many studies report little or no information content for such disclosures (Beaver and Landsman 1983), although more recent studies (Bublitz et al. 1985; Lobo and Song 1989) report evidence of stock price reactions to releases of information on current cost. These studies were conducted at the aggregate level of the market and did not examine the trading behavior of particular classes of market agents. In contrast, this study focuses on the trading behavior of corporate insiders, and, instead of the commonly used security returns, a non-price variable is the dependent variable of interest. Managers (insiders) could have information about how disclosure might affect the value of a firm through political and contracting costs borne by the firm (Smith and Warner 1979; Holthausen 1981) well in advance of other traders (Jaffe 1974; Larcker et al. 1983). In conformance with their fixation on reported income and their general skepticism of market efficiency (Mayer-Sommer 1979), managers might perceive that investors would react negatively to the new information required by SFAS No. 33, especially when current-cost adjusted income falls below historical-cost income. As a consequence, they would be expected to sell their stocks in anticipation of a negative stock market reaction. For a sample of 441 firms, this study investigates the relation between the initial release of inflation-adjusted information and the trading behavior of insiders. Specifically, it is hypothesized that the expectation that disclosure income will be lower than historical-cost income leads to net selling by managers prior to the initial disclosures under SFAS No. 33. The results are generally consistent with the hypothesis. The analyses show results that are statistically significant (at conventional levels) for both the independent variables in this study-a proxy for income shrinkage and a proxy for earnings change in the expected direction. But one must keep in mind the limitations of inferences from analyses of insider-trading data (Larcker et al. 1983). Specifically, since there is no comprehensive theory of insider trading, it is difficult to interpret insidertrading activity, and the results of this study must be kept in perspective as additional evidence on the effect of SFAS No. 33 disclosures.]

[The 1987 Treadway Report recommended several ways the profession might enhance auditors' ability to detect fraud. The Auditing Standards Board implemented many of the recommendations in 1988 with nine new "expectation gap" pronouncements issued as SAS Nos. 53 through 61. These new standards clarify and extend an auditor's responsibility for (1) detecting and reporting errors, irregularities (including fraud), and illegal acts (SAS Nos. 53 and 54) and (2) assessing the internal control structure (SAS No. 55). The effect of recent and possible future increases in auditors' responsibility to deter and detect fraud warrants study. To provide insight into the strategic interaction between a manager and auditor, we develop a theoretical foundation through game-theoretic analysis and economic experimentation. The game involves one decision by the manager and two by the auditor. The manager moves first, choosing a probability of committing fraud. The auditor, without observing the manager's choice, decides whether to perform tests of controls and then decides the level of detailed tests of balances. This second stage of testing allows the auditor to detect fraud with a probability that increases with the level of testing (up to a probability of less than 1). We solve for Bayesian Nash equilibria in this setting. This study investigates the effects of four independent variables: (1) auditor's penalty, which reflects either a monetary loss (through lawsuits) or a loss of reputation from failure to detect fraud, (2) auditing standard requirements, which reflect the ability of auditing standard setters to increase auditors' responsibility for detecting fraud or other irregularities, (3) the quality of the internal control structure, which is represented by a distribution of unintentional clerical errors that the auditor must investigate to detect an irregularity, and (4) audit fee, which reflects increased competition in the market for auditing services. We examine the effect of these variables on: (1) tests of transactions and detailed tests of balances, (2) fraud detection, and (3) incidence of fraud. The results of the economic experiment indicate that increasing the auditor's penalty decreased fraud, increased detailed tests of balances, decreased tests of transactions, and increased fraud detection. Increasing testing requirements increased audit costs, decreased discretionary testing, increased fraud detection, and decreased fraud commission. With strong internal control, auditors increased tests of transactions and detected fraud more frequently; managers committed fraud less frequently. Increasing the audit fee resulted in less fraud. The empirical results support a direct relationship between testing, fraud detection, and fraud prevention. Two caveats are worth noting. Given our motivation for studying the interaction between manager and auditor, we frame the issue as fraud. However, the model actually encompasses all irregularities. Irregularities come in two varieties: fraud (misrepresentation of fact) and defalcations (misappropriation of assets). The situation modeled is sufficiently general to fit both types of irregularities, yet not so specific as to exclude either one. Management fraud possesses many unique features that are only implicit in the model (e.g., reporting). Hence, any generalizations should pertain to the detection of irregularities rather than to specific aspects of management fraud. A second caveat is that we focus solely on the economic implications of fraud in this study, and its ethical dimension is not explored.]

[An intriguing anomaly in recent market-based accounting research is that both the market and analysts appear not to recognize properly the time-series properties of quarterly earnings shocks. Bernard and Thomas (1990) and Freeman and Tse (1989) present evidence that the market underestimates the implications of previous period earnings for future earnings. Mendenhall (1991) and Abarbanell and Bernard (1991) find that analysts do not seem to utilize time-series information about earnings correctly when setting their forecasts. Specifically, these last two studies document a positive serial correlation in analysts' quarterly forecast errors and interpret this finding as analysts systematically underestimating the persistence of past earnings forecast errors in forecasting future earnings. The purpose of this article is to examine whether analysts properly recognize the time-series properties of annual earnings when setting their estimates of future earnings. Givoly (1985) investigates this issue and finds that analysts' forecasts of annual earnings per share (EPS) are unbiased and that prediction errors are not serially correlated. He concludes that forecasts are formed in an efficient manner. In contrast, this study finds that, on average, analysts set overly optimistic estimates of the next period's annual EPS and that forecast errors display significantly positive serial correlation. These results hold for short-term as well as for longer-term IBES consensus forecasts. Bias and positive serial correlation in forecast errors suggest that analysts do not properly recognize the time-series properties of earnings when setting expectations of future earnings. Studies show that, for any given year, earnings shocks have both permanent and transitory (i.e., mean-reverting) components (see, e.g., Brooks and Buckmaster 1976; Ou 1990; Ou and Penman 1989b) and that the level of earnings persistence varies across firm-years. We examine whether analysts are aware of the differences between permanent and temporary components in the previous year's earnings when predicting future earnings. The findings show that analysts are able to differentiate partially between permanent and temporary components in previous period earnings. We also find that the overestimation bias in forecasts is most pronounced for firms that recently experienced negative earnings. In contrast, the positive serial correlation is most evident for firms that had predominantly permanent earnings. These results suggest that the overestimation bias and serial correlation are not uniform across firms. To address the economic significance of our findings, estimates of bias and serial correlation in forecast errors are used to adjust existing forecasts. The accuracy of the adjusted forecasts are then compared to the unadjusted forecasts. A significant improvement in forecasting ability is found for the adjusted longer-term forecasts, as evidenced by a 12 percent improvement in the mean squared error. Further, the bias and serial correlation in the adjusted forecasts are significantly less than in the actual forecasts for both longer and short terms. These results further support the view that analysts do not utilize available information efficiently when setting forecasts.]

[This article reports evidence that the presentation order of audit working-paper documentation affects the outcome of audit partners' going-concern decisions. Although Frederick (1991) has found that memory organization and the organization of evidence interact to affect the retrieval of audit evidence from memory, this study establishes that memory organization and the organization of working-paper evidence interact to affect audit decisions. Order effects are hypothesized because the going-concern decision task is driven by voluminous working-paper evidence that is typically documented in an order convenient for compiling audited financial statements, not in an order optimal for minimizing the cognitive effort required to reconstruct the causal order of relevant events. The experiment reported in this article tests the proposition that evidence documented in a contextually meaningful causal order (rather than in the traditional working-paper order of the public accounting firm participating in the study) will diminish the cognitive effort required of highly experienced audit partners by facilitating their ability to evoke pre-existing going-concern schemata from long term memory. One hundred audit partners from a public accounting firm participated in the study. The experiment manipulated the presentation order of 60 sentences to compare the going-concern decisions of partners who read the audit evidence in variations of two distinctly different orders: causal order versus the participating firm's working-paper order. Evidence sentences were compiled from one of the firm's audit engagements by the researcher and by the engagement partner, manager, and reviewing partner assigned to the audit. Fifteen of the sentences referred to events consistent with a decision of substantial doubt about the entity's ability to continue as a going concern, 15 were consistent with a decision of no substantial doubt, 15 supported both decisions, and 15 supported neither decision. The sentences were arrayed in causal order by the researcher, engagement partner, and manager, using procedures outlined in Pennington and Hastie (1986, 1988). The working-paper order of presentation was prepared by the engagement partner, manager, and reviewing partner and represents the way an audit partner in the participating firm would likely encounter the evidence in a set of working papers assembled by the firm. The experiment was constructed as a 2 X 2 design: 25 subjects read the "yes" sentences (evidence for substantial doubt) in causal order and the "no" sentences in working-paper order, 25 read the "no" sentences in causal order and the "yes" sentences in working-paper order, 25 read both the "yes" and "no" sentences in causal order, and 25 read both the "yes" and "no" sentences in working-paper order. To control for within-cell order effects, randomly assigned subjects in each of the four treatment groups read the "yes" cues first, and the others read the "no" cues first. All subjects completed two tasks: a decision regarding substantial doubt based on the 60 evidence sentences and an assessment of their confidence in the decision. The results establish that the order manipulation affected the subjects' going-concern decisions but not their confidence in those decisions. Subjects decided there was substantial doubt about the client more often when the strongest evidence supporting this decision was read in causal order and, correspondingly, least often when the strongest evidence was read in the firm's working-paper order. The order effects documented in this study have important implications for audit working-paper design, particularly in complex unstructured decision tasks like the going-concern decision.]

[A number of formal models have been used to reflect the fact that people generally require less time to perform a complex task after they acquire some familiarity and experience with the task. These models are referred to as learning curves, and a sizable literature exists on their properties and uses (Yelle 1979). In contrast, little has been written on the nature of relearning curves, which are formal models used to estimate the reduction in task time that occurs while relearning skills that have been forgotton due to interruption. The relatively few studies addressing this aspect of interrupted production advocate "backing up" the original learning curve (typically a log-linear model) in some fashion and using it to model the relearning process (Adler and Nanda 1974; Carlson and Rowe 1976; Cherrington et al. 1987; Cochran 1968; Hoffmann 1968; Keachie and Fontana 1966; Lippert 1976). This article reports on a laboratory study designed to provide some evidence on the nature of relearning curves. Subjects were paid a realistic wage to assemble structures with Erector Set parts. They repeated the task for approximately four hours, and three forms of marginal-time learning curves were fit to their performance data. After breaks ranging from seven to 175 days, they repeated the construction projects. The two sets of times were used to compute a measure of skill decrement and to fit nine forms of learning curves. Issues of interest are: (1) similarity of the functional forms of the best-fitting relearning and learning curves, (2) whether backing up the learning curve is a reasonable method of modeling relearning, and (3) effects of skill decrements on the relative goodness-of-fit of relearning curves. The results show that the best-fitting relearning curves are not of the log-linear form commonly used to model learning, but are of a form that is nonlinear on both a log-log and an arithmetic scale. Further, this new form developed in our study also provided a better fit than the log-linear model when applied to learning (pre-interruption) data. Additionally, backing up the best-fitting learning curve or starting anew with this same curve form also provided a good-fitting relearning curve model; however, backing up the classical log-linear model did not fit the data as well as other models. Finally, differences between the goodness-of-fit of the various models became more pronounced with the increase in amount of skill decrement. These findings should be of interest to those who use learning curves to estimate labor time, especially under conditions of interrupted production.]

[In the last two decades of accounting research, many studies have investigated the relation between accounting variables and risk-adjusted security returns. The earliest studies (e.g., Ball and Brown 1968) used simple, nonparametric methods and focused mainly on the question of whether accounting earnings are associated with residual equity returns. Subsequent studies made methodological refinements in both the measurement and the statistical techniques. One important statistical refinement was the "unexpected earnings response regression model" (UERRM), a linear statistical model that uses the unexpected earnings variable as a regressor to explain risk-adjusted returns. The UERRM has become well-known, and many recent studies (e.g., Cornell and Landsman 1989, 686; Daley et al. 1988, 580; Doran et al. 1988, 392; Landsman and Damodaran 1989, 107; McNichols 1989, 15) have used some form of it as a "benchmark" model, against which to compare more complicated models. In one paradigm (so popular that it has become almost standard practice), various accounting variables are added to the UERRM, and their "incremental information content" is assessed by testing the statistical significance of their coefficients. The inferences derived from this procedure are, of course, conditional on the degree to which the UERRM is correctly specified. A critical problem caused by using a misspecified UERRM is that its least squares estimator can lead to erroneous inferences in the research design. Despite the UERRM's popularity, researchers have expressed concerns regarding its specification. For example, Lev (1989) recently surveyed a large number of research papers that used some form of an UERRM and found that for the most part R2 s were low and often bordered on "the negligible." His table 1, which includes statistics from 19 studies, indicates that most reported R2 s are less than 10 percent. Although low R2 s are not proof of major specification problems, Lev does suggest that they are a cause for concern and may be the result of specification problems. Investigation of multiple specification problems is an important aspect of the present study because the various specification issues are interrelated. For example, nonnormality can be associated with nonlinearity, which, in turn, can be associated with heteroscedasticity or variation in the coefficients (Judge et al. 1985, 455, 814, 839). Thus, ad hoc tests for a single specification problem can be misleading and can fail to identify the fundamental problems. To our knowledge, no studies have comprehensively and formally evaluated the specification of the cross-sectional, ordinary least squares model that relates unexpected earnings to risk-adjusted security returns. Therefore, the purpose of this study is to test such a model systematically and empirically for specification problems. Specifically, this study tests for nonlinearity, heteroscedasticity, residual nonnormality, omitted variables, and interfirm systematic and random coefficient variation. Also, when appropriate, adjusted R2 -statistics are included to indicate the degree of misspecification-information that is not directly observable from the tests themselves. A high degree of generality is obtained by using three samples of earnings forecasts as proxies for expected earnings. These were obtained from IBES financial analyst consensus forecasts, Value Line financial analyst forecasts, and COMPUSTAT-based time-series forecasts. Daily and monthly security returns are considered for short and long event-windows, respectively. In addition, one study recently published in The Accounting Review (Cornell and Landsman 1989) is replicated. The findings from all of the samples and the replication indicate that the specification error is large enough to affect conclusions regarding economic relationships. For example, in the replication, the specification problems are shown to lead to substantial instability in inferences from the model.]

[In this study we assess how different regimes of auditor liability affect the demand for and supply of auditing services. The assessment was made with 15 experimental markets, each of which involved two sellers of assets, two auditors (verifiers), and four buyers. The experimental markets paradigm allowed us to compare the negligence liability regime (six markets) that auditors currently face with two alternatives not currently in existence-a strict liability regime (six markets) and a no-liability regime (three markets). We focused on the extent to which the experimental results conformed to our predictions of (1) sellers' frequency of hiring verifiers and of selecting a costly investment that improved aggregate welfare, (2) verifiers' service fees and their frequency of testing the truthfulness of the sellers' disclosures, and (3) buyers' reliance on the sellers' disclosures and verifiers' reports when pricing the sellers' assets. The predictions varied across the liability regimes primarily because differences in the degree of the verifiers' liability changed their economic incentives to test the truthfulness of the sellers' disclosures. The results show that the no-liability and negligence markets operated in a manner consistent with the predictions, whereas the strict markets deviated from the predictions on several dimensions. Specifically, verifiers in the strict liability markets were hired less often than predicted because they submitted higher offers for their services than sellers were willing to pay. This in turn led to fewer than predicted costly investments by the sellers. Although our general conclusion recognizes that a legal system is an integral part of the auditing institutional infrastructure, we found no evidence of any systematic benefits from imposing a strict liability rule on the verification service. In fact, the results suggest that the negligence liability markets operated at a level of economic efficiency as high or higher than those in the other two regimes. This suggests that the current tendencies of courts and the auditing profession to expand the scope of auditors' liabilities may not achieve the net benefits expected from such expansions.]