Andy’s Research Papers

These papers are very mathematical. In the first two, much of the math is presented in the appendices. In the next two, much of the math is presented in the body of paper. Most non-researchers should be able to follow the quality costs paper; however, the paper on conflicting incentives uses optimal control with lags, which requires—depending upon your perspective—either an advanced understanding of calculus or a thorough understanding of the basics. Your comments are welcome at andy@AESpero.com.
Completed Papers:
Deadlines as Management Control Devices. This paper is from my dissertation, which was the Best Management Accounting Dissertation in 1994 according to the American Accounting Association.

Within firms, workers often face deadlines to complete tasks. The environment may exogenously impose such deadlines, or the firm may endogenously impose them. To date, there has been no theory to explain the optimality of imposing endogenous deadlines. The model analyzed and the results presented in this paper fill that void by showing that in task-based settings—when there are economic implications of completing or not completing the task—deadlines are valuable managerial control devices only when incentive problems exist. Thus, a demand for deadlines is derived, and the optimal deadline-based contracts are characterized and contrasted.

This paper has been presented at: the University of British Columbia, Carnegie Mellon University, the University of Chicago, Harvard University, the University of Minnesota, Northwestern University, Stanford University, and Washington University.

Optimal Imprecision and Ignorance. With Chandra Kanodia and Raj Singh. This paper has been publishedby the Journal of Accounting Research. It was awarded the Best Corporate Finance Paper at the 2002 Western Finance Association Conference. (The final title is slightly different.)

Consider a setting in which an entrepreneur has an investment opportunity and desires to sell the investment’s long-term proceeds in the capital market. We assume the investment’s profitability potential is exogenous, while the investment level is endogenous. The entrepreneur knows both the investment's profitability and level (size).

When investors have full-information—perfectly know both profitability and level—the most efficient outcome for society is achieved. However, when investors have perfect information on one dimension and imperfect, or noisy, information on the other dimension, a dysfunctional equilibrium results. If investors know the investment’s profitability and receive an imperfect measure of the investment level—say through an accounting report—the entrepreneur myopically under-invests (relative to the first-best) as accounting reports cannot influence investors about the projects underlying profitability. If investors perfectly know the level of investment—from say a perfect accounting report—but are ignorant about it inherent profitability, then the entrepreneur over-invests (relative to the first-best) as a way to influence perceptions (about the investment’s profitability) in the capital market.

We show that given the market’s ignorance about investment type, there is an optimal degree of accounting imprecision. Conversely, given a level of accounting imprecision, there is an optimal level of ignorance (or information asymmetry between the entrepreneur and investors). In fact, in certain settings, choosing the optimal choice of imprecision (or ignorance) allows society to obtain the full-information outcome.

This paper has been presented at the American Accounting Association’s 2000 Doctoral Consortium, the American Accounting Association’s 2000 Annual Meeting, the University of Minnesota’s 2000 Theory Conference in Accounting, the Canadian Accounting Association’s 2000 Doctoral Consortium, the 2002 Western Finance Association Conference, and the University of Chicago.

A Quality Model & Field Study.

This paper provides a model a production problem with stochastic quality. I extend the basic cost minimization problem in economics by explicitly considering random quality in terms of failure-related activities and inputs. By making a series of reasonable assumptions, I explicitly model trade-offs amongst quality activities, and test this theory at a field site: a forge factory. I also model the design (problem-solving) and manufacturing phases as a Bayes’ decision problem.

Versions of this paper have been presented Carnegie Mellon University and Washington University.

On the Conflict Between Providing Short- and Long-Term Incentives in an Agency Setting.

This paper analyzes a multi-period investment problem where a manager must take both short-term operating actions and a long-term investment decision. When the effects of short-term actions and long-term investments cannot be disaggregated, under-investment may result ( valuable investments may be rejected). The incentive schemes used to motivate operating effort can destroy the manager’s desire for investmerns. I derive the optimal contracts for this problem, and provide and illustration to show how under-investment may be eliminated if the manager is allowed to manipulate his performance measure, income, subject to the constraint that cumulative income equals cumulative net cash flows.

Different versions of this paper have been presented at the American Accounting Association’s 1996 Management Accounting Section’s Annual Conference, Washington University, the University of Chicago, and the University of Minnesota.
Work-in-Progress:

On the Optimality of Seniority versus Merit-based Promotion Policies…. Or, does Early Entry in the NBA Draft Really Hurt the NCAA? Look for this working paper, finally, this year. This paper analyzes the efficiency of promotion policies within organizations. To narrow the focus of our study, we assume that those individuals who are promoted receive all the negotiating power once they are identified; so, the firm receives no further benefit once the individual is promoted. This fact eliminates any direct benefit (to the firm) of promoting "high types" and is similar to the NCAA losing the benefit of a player’s services after he is drafted into professional sports. In a model with two types of players, but no private information, we investigate whether an organization is—on average—better off losing the output of those perceived to be its best agents if such a policy induces that remaining agents to work harder. Compared to seniority-based (senior-draft only) policy, under a merit-based (early-draft) policy, the opportunity costs associated with lost effort from promoted individuals is higher, yet the benefits of increased effort and output from junior workers is also higher. (To increase his own probability of early promotion, each junior worker will take higher levels of effort. This increases total output.) So, in sports terms, we answer the question: for incentive reasons, do we have (on average) better play with (on average) worse players?

An Empirical Investigation of Draft Policies and Player Incentives. This research is an extension of the above work. Using an event study methodology, we will attempt to measure differences in underclassmen output in either or both of the major-college, draft-related sports, i.e., basketball and football. We will limit our attention to major NCAA conferences where draftable players tend to enroll. Unlike many firms, performance measures for such players are publicly available. In the past twenty years, there have been several events that should have changed underclassmen’s beliefs of being drafted. We will focus our analysis around these time periods.
A Note on the Flawed Logic of Activity-based Product Costing: Approximately fifteen years ago, several academics and practitioners began promoting the notion that activity-based costing (ABC) methods provided managers with “better information” about cost behavior, including accurate estimates of unit product cost. They observed that some costs may not vary with volume, but may vary along other non-volume-related dimensions: “activity-related” dimensions. As opposed to most traditional product costing systems, a benefit of ABC is that, in general, it does provide more accurate estimates of the costs of these non-volume-related activities. From this fact, proponents of ABC concluded that aggregating such costs along such dimensions and allocating them to products on a similar basis provides accurate unit product costs. Using geometry and linear algebra, I show that there is no justification to draw this conclusion. The logic and math are simple: iIf the reason to account for such activity costs is their low correlation with volume, then relating activity costs back to units will also have a low correlation, and, need not generate better information—just different product costs.