Recent theory argues that concern about job security creates an incentive for managers to smooth earnings in consideration of both current and future relative performance. We find support for this theory. Our evidence suggests that when current earnings are ‘poor’ and expected future earnings are ‘good’, managers ‘borrow’ earnings from the future for use in the current period. Conversely, when current earnings are ‘good’ and expected future earnings are ‘poor’ managers ‘save’ current earnings for possible use in the future. However, sensitivity analysis indicates that we cannot rule out selection bias as a potential alternative explanation for our findings.

Abstract. We examine the empirical relationship between auditors' resource allocations and selected engagement characteristics. Our measure of resources is hours of grades of labor (partner, manager, etc.) “charged” to audit activities (planning, internal control evaluation, etc.). Engagement characteristics examined are client size, industry affiliation, client complexity, risk, auditor provision of management advisory services to the auditee, and degree of control reliance. The data were obtained from publicly available sources and a survey developed and administered by an international public accounting firm. We find the cross‐sectional variation in the labor charged to various audit activities can be explained by engagement characteristics found to be important in prior studies on audit fees, total labor inputs, and the mix of labor inputs. Measures of client size, industry, complexity, risk, and services provided are associated with changes in the allocation of labor among audit activities. We find no substitution of internal control review/testing for substantive testing on reliance audits. Task assignments vary by rank. Measures of client size, complexity, risk, and services provided are associated with activity‐specific changes in the labor mix.

This paper introduces a conditional Kolmogorov test of model specification for parametric models with covariates (regressors). The test is an extension of the Kolmogorov test of goodness-of-fit for distribution functions. The test is shown to have power against 1/√n local alternatives and all fixed alternatives to the null hypothesis. A parametric bootstrap procedure is used to obtain critical values for the test.

To obtain consistency and asymptotic normality, a generalized method of moments (GMM) estimator typically is defined to be an approximate global minimizer of a GMM criterion function. To compute such an estimator, however, can be problematic because of the difficulty of global optimization. In consequence, practitioners usually ignore the problem and take the GMM estimator to be the result of a local optimization algorithm. This yields an estimator that is not necessarily consistent and asymptotically normal. The use of a local optimization algorithm also can run into the problem of instability due to flats or ridges in the criterion function, which makes it difficult to know when to stop the algorithm. To alleviate these problems of global and local optimization, we propose a stopping-rule (SR) procedure for computing GMM estimators. The SR procedure eliminates the need for global search with high probability. And, it provides an explicit SR for problems of stability that may arise with local optimization problems.

In this paper I discuss alternatives to the GMM estimators proposed by Hansen (1982) and others. These estimators are shown to have a number of advantages. First of all, there is no need to estimate in an initial step a weight matrix as required in the conventional estimation procedure. Second, it is straightforward to derive the distribution of the estimator under general misspecification. Third, some of the alternative estimators have appealing information-theoretic interpretations. In particular, one of the estimators is an empirical likelihood estimator with an interpretation as a discrete support maximum likelihood estimator. Fourth, in an empirical example one of the new estimators is shown to perform better than the conventional estimators. Finally, the new estimators make it easier for the researcher to get better approximations to their distributions using saddlepoint approximations. The main cost is computational: the system of equations that has to be solved is of greater dimension than the number of parameters of interest. In practice this may or may not be a problem in particular applications.

[We explore the implications of financial system design for financial innovation. We begin with assumptions about the investment opportunities of firms, their observable attributes, and the roles of commercial banks, investment banks, and the financial market. We examine the borrower's choice between bank and financial market funding, the commercial bank's choice of monitoring capacity, and the investment bank's choice of whether to invest in financial innovation. Our main result is that financial innovation in a universal banking system is stochastically lower than innovation in a financial system in which commercial and investment banks are functionally separated.]

This article builds a theory of financial system architecture. We ask: what is a financial market, what is a bank, and what determines the economic role of each? Starting with basic assumptions about primitives–the types of agents and the nature of informational asymmetries–we provide a theory that explains which agents coalesce to form banks and which trade in the capital market. It is shown that borrowers of higher observable qualities access the financial market. Moreover, a financial system in its infancy will be bank-dominated, and increased financial market sophistication diminishes bank lending.