Multitasking, Multidimensional Screening, and Moral Hazard with Risk Neutral Agents* SUREN BASOV School of Economics and Finance, La Trobe University, Bundoora, Vic., Australia SVETLANA DANILKINA Department of Economics, The University of Melbourne, Parkville, Vic., Australia In this paper we consider a model where a risk-neutral princi- pal devises a contract for a risk neutral agent who can exert effort along different dimensions and possesses private information about her cost of effort. We show that when the number of effort dimensions exceeds the number of performance measures observed by the principal hidden action leads to an additional welfare loss compared with pure adverse selection even if both parties are risk neutral and the production technology is independent of the agent’s type. The result implies that if effort has many dimensions it is beneficial to the principal to base employees’ compensation on many performance measures rather than on a single ‘bottom- line’ measure (e.g. their contribution to the company’s profits). I Introduction Incentive theory identifies two main difficul- ties in aligning incentives of parties in a stan- dard agency problem: adverse selection and moral hazard. Both of them are present in many interesting principal-agent relationships. Adverse selection is created by hidden informa- tion on the part of the agent, while moral hazard results from hidden action. Though both are caused by private information on the part of the agent, the welfare consequences of hidden action and hidden information are quite differ- ent. While hidden information alone almost always results in a welfare loss, 1 hidden action alone will not result in any welfare loss when both parties are risk neutral, because the basic trade-off between efficient risk sharing and pro- vision of optimal incentive disappears. 2 The models that considered coexisting hidden action and hidden information were developed by Laffont and Tirole (1986), Picard (1987), Rogerson (1988), Guesnerie et al. (1989), Melu- mad and Reichelstein (1989), and Caillaud et al. (1992). The main conclusion of these models was that under risk neutrality the pres- ence of hidden action does not create any addi- tional welfare loss. The logic behind this result is the following. Assuming the action of the agent is observable, solve the pure hidden infor- mation model to obtain a wage 3 that depends on the action of the agent. If the action is un- observable, one can find a wage that depends on * We are grateful to all participants of ESAM 09 and ACE 09 for helpful comments. Special thanks to Ronald Stauber, Ian King, and anonymous referees. JEL classifications: C6, D82 Correspondence: Suren Basov, School of Econom- ics and Finance, La Trobe University, Bundoora, Vic. 3086, Australia. Email: s.basov@latrobe.edu.au 1 An exception is the case when under the first best contract the agent’s utility does not depend on her pri- vate information. For an example of such a situation, see Basov and Bardsley (2005). 2 For a review of principal-agent problem in the case of pure hidden action, see Grossman and Hart (1983). 3 For concreteness, assume that the agent is hired by the principal to perform some tasks and the hidden action is the effort of the agent. THE ECONOMIC RECORD, VOL. 86, SPECIAL ISSUE, SEPTEMBER, 2010, 80–86 80 Ó 2010 The Economic Society of Australia doi: 10.1111/j.1475-4932.2010.00662.x