Working PapersIntra-Firm Bargaining and Learning in a Market Equilibrium February 2009 This paper introduces an agency relationship into a dynamic game with informational externalities. Two competing firms plan to launch a new product and in each firm headquarters bargain with its R&D division about the cost of developing this product. The development cost is the private information of R&D divisions and is correlated across the two firms. We find that the agency relationship creates an incentive for simultaneous launching of the product, even if this involves an inefficient delay. As the commitment power of the headquarters decreases, this incentive becomes stronger. The effect of competition is decomposed into two parts. Inter-period competition (from past and future sales) pushes firms towards launching the product simultaneously, while intra-period competition (from concurrent sales) does the opposite.
Competition in Bureaucracy and Corruption Journal of Development Economics, forthcoming January 2009 This paper studies the consequences of introducing competition between bureaucrats. Firms are supposed to invest into eliminating negative externalities of production, while bureaucrats administer the process by issuing licences. Some bureaucrats are corrupt, that is, they issue a licence to any firm in exchange for a bribe. The competition regime is found to create more ex ante incentives for firms to invest, while the monopoly regime is better at implementing ex post allocation, that is, distributing the licences given the firms' investment decisions. Additional results on the effect of punishments, intermediaries and staff rotation are provided. An older version (with endogenous pool of bureaucrats) is Department of Economics Working Paper #369
Slides Learning and Microlending (with Rocco Macchiavello) September 2008 (CEPR Discussion Paper 7011) For many self-employed poor in the developing world, entrepreneurship involves experimenting with new technologies and learning about oneself. This paper explores the (positive and normative) implications of learning for the practice of lending to the poor. The optimal lending contract rationalizes several common aspects of microlending schemes, such as “mandatory saving requirements”, “progressive lending” and “group funds”. Joint liability contracts are, however, not necessarily optimal. Among the poorest borrowers the model predicts excessively high retention rates, the contemporaneous holding of borrowing and savings at unfavourable interest rates as well as the failure to undertake profitable and easily available investment opportunities, such as accepting larger loans to scale-up business. Further testable predictions can be used to interpret and guide the design of controlled field experiments to evaluate microlending schemes. Slides
Financing Experimentation (with Rocco Macchiavello) August 2009 A shorter and newer version of "Learning and Microlending" but with less results for microfinance
When people start a new activity, they may not know how good they will be at it and, therefore, must experiment in order to learn. What happens when the experimentation is financed by a lender and there is an agency problem between the lender and the borrower? In a standard setting, experimentation is more profitable for a longer time horizon or a lower payoff of the known activity. In contrast, financing experimentation might be harder in these cases, that is, precisely when it is more valuable. The optimal contract resembles typical microfinance schemes observed in practice. Information and Delay in an Agency Model Revised & Resubmitted to The Rand Journal of Economics August 2009 Abstract This paper studies how delay in contracting depends on an exogenous signal. The agent whose cost is his private information may produce in the first period or be delayed until the second period. A signal about the cost of the agent is available between the two periods. The quality of the good can vary; in the benchmark case of no signal, the principal offers the standard Baron-Myerson contract and there is no delay, as is shown by Wang (1998). Delay is determined by the considerations at the margin and may increase or decrease with a better signal. The value of information can be negative as better information may aggravate the principal's commitment problem. We provide conditions for and examples of how more information affects the principal's profits and optimal delay. An older version (with some results for indivisible good) is Department of Economics Working Paper #298 Contracts with Informational Externalities May 2007 Abstract This paper studies informational externalities between contracts. Two principals (for instance the governments of two neighbouring countries) deal with two different agents (for instance a railway company in each country). If, in the first period, an agent refuses the contract offered by his principal, the principal can offer a new contract in the second period. If pair A has not signed a first period contract but pair B has, information about the contract signed by pair B (and also, maybe, some exogenous information) is transmitted to pair A in between the two periods. As a consequence, because the types of the agents are correlated, there are informational externalities between the two principal-agent pairs. Inefficient equilibria with delay, i.e., with agreement in the second period, are obtained. In any equilibrium with delay only the low cost agent produces in the first period. I also show that there is always some production in the first period. More results under specific assumptions on the information structure of the game are provided. In particular, I identify some cases when the delays in the two pairs are strategic substitutes/complements. Protecting from Employees' Hold Up: Overemployment, Over/Underproduction and Organizational Design July 2007 Abstract Overemployment is a strategy often used by firms as a response to employee hold up. This paper studies its consequences in a simple moral hazard model. First, its effect on agents' incentives to work is considered. As compared to the single agent case, duplication of agents may increase total effort, i.e., production, above the first best level (overproduction), increase it but to a level lower than the first-best one (partial restoration of the first-best) or decrease it (further distortion downwards). Second, possible consequences of overemployment on organizational design are studied, namely, distortions in the choice of (a) projects, (b) agents and (c) informational structure.
|