Catálogo de publicaciones - libros

Joel Demski has made significant contributions to accounting research and education for nearly forty years. This paper reviews and highlights many of his scholarly contributions. He has been innovative and thought provoking — always at the leading edge of our discipline

Accounting information is formed by an aggregation of the information available to the accounting system. Introduction of fair value accounting represents a new solution to the accounting aggregation problem as market information is merged into the accounting system. Multiple sources of information are available to market participants and accounting information is but one of these sources. Fair value information is available to the accounting system, to the public, and to individual market participants, hence, the aggregate information available in the economy — aggregate informativeness — depends on the confluence of accounting information and other sources of information. Particularly, the price process might well be informative but is influenced by the accounting policy chosen and, hence, it is not obvious the introduction of fair value accounting leads to an improvement in aggregate informativeness. Fair value accounting may destroy the aggregation mechanism of the market.

This paper presents an equilibrium model of voluntary disclosures for the seller of an asset who receives a random sample of information of random size about the asset’s value. Even though (a) antifraud rules prevent the seller from making false statements about the value of the items in his random sample, (b) all potential purchasers of the asset know that the seller’s random sample always contains at least one sample element, (c) all potential purchasers of the asset interpret the seller’s disclosure or nondisclosure in the same way, and (d) disclosure of any or all of the seller’s sample information generates no proprietary costs, we show that in equilibrium there is a positive probability that the seller will make no disclosure at all, and that, when the seller makes no disclosure, the nondisclosed information is not the worst possible sample information the seller could have had about the asset’s value. These results are contrasted with the “unravelling” result of [], [], and []. We show that, were potential purchasers of the asset to know the size of the seller’s random sample, “unravelling” (i.e., full disclosure) would occur. We conclude that the randomness of the seller’s sample size is key to determining the seller’s equilibrium voluntary disclosure strategy.

A standard control problem is analyzed using quantum probabilities. There are some advantages of conducting the analysis using the axiomatic structure of quantum probabilities: (1) there is synergy associated with bundling activities together, and, hence, a demand for the firm; (2) information occupies a central place in the analysis; (3) accounting information questions can be related to other information sciences. The main result is that control costs decline when aggregate performance measures are used; aggregation arises naturally. An implication is that the common practice of acquiring individual measures may be misguided in an environment where synergy is a first order effect. Also, double entry accounting appears well suited for processing information in a synergistic context.

Moral hazard with hidden information refers to a control problem where the agent’s actions are observable, but not the information on which they are based. This paper analyses the case in which an agent (for example a subcontractor or a dealer) obtains perfect information before deciding on his action. The close relationship to adverse selection allows easy derivation of the set of feasible sharing rules. The optimal action (production plan) is then derived, and it is shown that action efficiency (and incentive power) is uniformly lower than first-best (except in the best and the worst state), but greater than efficiency in the corresponding adverse selection problem. It is shown that efficiency and incentive strength is decreasing uniformly in the agent’s aversion to risk (properly defined). The level of risk may be endogenously as well as exogenously determined. Holding exogenous risk constant it is shown that risk averse agents tend to end up with more risky production plans. However, the effects of exogenous changes in risk are ambiguous. It is further demonstrated that the risk aversion of the principal will have the opposite effect as a more risk averse principal will tend to prefer more efficient (and less risky) production. Finally, it is argued that principals prefer agents to be informed before actions are taken, but after contracting. This information structure also represents a social optimum.

In this essay I focus on the equilibrium relation between the “risk” in a performance measure and the “strength” of the controlling agent’s “incentives.” The main motivation is that a large (mainly empirical) literature has developed postulating that the key implication of the principal-agent model is that this relation be negative. I first show that a standard principal-agent model, e.g., Holmström (1979), offers no equilibrium prediction about the relation between “risk” and “incentives.” Next, I show that except in the highly stylized limiting Brownian version of Holmström and Milgrom (1987), this model doesn’t yield a directional prediction for the equilibrium relation between “risk” and “incentives” either. This is due to the general property that risk arises endogenously in such principal-agent models. This, in turn, establishes that while the mixed empirical evidence on this relation may be useful from a descriptive vantage point, it does not shed any light on the validity of the principal-agent theory.

We characterize optimal investment and compensation strategies in a model of an investment opportunity with managerial incentive problems, caused by asymmetric information over investment costs and the manager’s desire to consume slack, and flexibility over the timing of its acceptance. The flexibility over timing consists of the opportunity to invest immediately, delay investment for one period, or not invest at all. The timing option provides an opportunity to invest when circumstances are most favorable. However, the timing option also gives the manager an incentive to influence the timing of the investment to circumstances in which he gets more slack.

Under the assumption that investment costs are distributed independently over time, the optimal investment policy consists of a sequence of target costs, below which investment takes place and above which it does not.

The timing option reduces optimal cost targets, relative to the case when no timing option is present. The first cost target is lowered because the compensation function calls for the payment of an amount equal to the manager’s option to generate future slack, should investment take place. This increases the cost of investing at the first opportunity, thus reducing its attractiveness. In order to ease the incentive problem at the initial investment opportunity, the second target cost is also lowered, even though no further timing options remain.

Making the additional assumption that costs are uniformly distributed, we generate additional insights. First, circumstances are identified in which not only does the cost target for immediate investment exceed that for delayed investment but also the probability of immediate investment exceeds the conditional probability of delayed investment, results impossible in the first-best context. Here, relatively speaking, incentive problems shift the probability of investment away from delayed investment towards immediate investment. Second, incentive problems are generally thought to reduce target costs, relative to opportunities with no incentive problems, in order to limit the manager’s slack on lower cost projects. Incentive problems, however, have more complex effects in the opportunity analyzed here. As a result, we are able to identify circumstances under which the target cost for immediate investment may be increased by incentive effects, relative to the target cost that exists in the absence of incentive problems.

A puzzling feature of many incentive compensation plans is the practice of capping bonuses above a certain threshold. While bonus caps are often justified on the grounds of keeping pay levels in check, it has also been argued that such caps can wreak havoc on a firm’s incentive problems. In this paper, we study a setting in which bonus caps can actually help align incentives. When a CEO is impatient, she may be tempted to take a hardline stance with a privately-informed manager in project selection: if she places little weight on future flows, she is fixated on cost-cutting and curtailing budget padding. A bonus cap can soften the CEO’s posture by inducing risk aversion and thus creating a preference for a middle ground. We show that this force can enable a judiciously chosen cap to achieve goal congruence between shareholders and a CEO.

In this paper, we illustrate some subtleties related to responsibility accounting by studying two settings in which there are interactions among multiple control problems. In the first setting, two agents are involved first in team production (e.g., coming up with ideas) and then in related individual production (e.g., implementing the ideas). We provide conditions under which the agents are not held responsible for the team performance measure, despite each agent conditionally controlling it. The conditions ensure the incentive problem related to individual production is so severe it drives out any demand for the team performance measure. The team incentive problem is not binding because of the large “spillback” from the individual problem to the team problem.

In the second setting, we provide conditions under which an agent is held responsible for a variable he does not conditionally control. Conditional controllability is a notion derived for one-sided moral hazard. Our model is instead one of two-sided moral hazard. Under two-sided moral hazard, it can be optimal for an agent’s pay to depend on variables conditionally controlled by the principal. This serves as a substitute for commitment by the principal.

We consider the consequences of public disclosure of insider trades on trading costs and price discovery in financial markets. Similar to Cournot competition in product markets, corporate insiders with common private information have incentive to trade more aggressively than a monopolist with the same information. Since, given periodic financial corporate reporting, insiders routinely have access to information in advance of the market, it is reasonable to expect them to seek ways to limit trades and, thereby, increase profits. Public reporting of insider trades may have the unintended effect of furthering tacit coordination by allowing insiders to monitor each others’ trades. Moreover, even without such reporting, we show how insiders may be able to sustain coordinated behavior depending on the distribution characterizing liquidity trading. Thus, competition among corporate insiders may be less influential in price discovery than previously thought.