Choice of technology in outsourcing: an endogenous information structure

https://doi.org/10.1016/j.infoecopol.2003.06.001Get rights and content

Abstract

We study a monopsonist-supplier model in which each party has its own technology for production. The supplier may use its own or be required to adopt the buyer's technology. Because the efficiency of each technology is unknown to the other party, the choice of technology determines the informed party. We find that using the buyer's technology reduces not only the supplier's incentive to misrepresent, but sometimes the buyer's as well. As a result, when equally efficient, the buyer's technology is adopted. In cases where each technology has several states of efficiency, the less efficient technology may be assigned in the optimal contract.

Introduction

Outsourcing has become a common economic phenomenon due to diversified product lines and increasing opportunity costs of in-house production. As its popularity grows,1 firms have been exercising different ways of outsourcing. Baron and Kreps (1999) distinguish two different ways to outsource. One approach is to outsource completely. With complete outsourcing, the buyer is simply purchasing a final product from the supplier and the details of how that product is produced are left to the supplier. The other approach is to rely on some form of contingent labor, such as a leased employee. Under such an arrangement, the buyer is purchasing only the labor services from the supplier and the supplier uses the buyer's technology to produce the output.

Under complete outsourcing, the supplier may have an incentive to exaggerate the cost of production to command an information rent. A solution for this problem can be to transfer the buyer's technology to the supplier, and purchase only the labor service from the supplier. In such a case, the buyer may have an incentive to exaggerate the efficiency of her2 technology in an attempt to pay for a lower cost of production to the supplier. As Lovell (1968) reports, a common complaint among manufacturing firms is that the technology to be transferred is often overestimated. Although much of the literature in agency contracting deals with these incentive problems,3 they tend to be treated separately in each outsourcing approach, and merging them together has been largely overlooked.

In this paper, we analyze a situation where a monopsonistic firm (the buyer) is outsourcing its product to an outsider (the supplier) – the buyer can assign her own or the supplier's technology for production. Because the efficiency of one's technology is private information, the buyer's choice of technology determines the informed party in our model.

We first analyze the case with two states of nature (efficiency of the technology possessed by each party is either “efficient” or “inefficient”). In our model, the efficiency of a technology represents not only production cost, but quality as well4 – for instance, an efficient technology can produce a high-quality good at low cost. Both the buyer and the supplier announce the efficiencies of their production technologies and the buyer chooses a technology depending on the announcements. The supplier engages in production using the technology chosen by the buyer.

A main finding is that assigning the buyer's technology for production mitigates not only the supplier's misreporting incentive, but sometimes the buyer's as well. When both parties announce “inefficient”, assigning the buyer's technology removes the supplier's incentive, but when both announce “efficient”, by contrast, it reduces the buyer's misreporting incentive. The intuition to the former case is simple. In our context, the supplier with an efficient technology has an incentive to exaggerate the cost of production: he would misreport that his technology is an inefficient one.5 The buyer, however, can assign her own technology when the supplier announces that his technology is inefficient, thus removing the supplier's incentive to misreport.

Therefore, when the buyer's technology is announced to be efficient and the supplier's technology inefficient, the buyer assigns her technology not only for technological efficiency, but also for rent extraction. Assigning the buyer's technology after it is announced to be “efficient”, however, brings about the buyer's misrepresenting incentive. The buyer (with an inefficient technology) has an incentive to understate the cost of production in an attempt to pay less to the supplier: she would misreport that her technology is an efficient one.6 Because the supplier anticipates this when the contract is offered, the buyer must remove her incentive to misreport in equilibrium. This results in distortions in the output schedule. These distortions can be reduced using technology assignment for the case where both technologies are announced to be efficient.

When both announcements are “efficient”, allowing the supplier to use his technology may seem to relieve the buyer's misreporting incentive. Paradoxically, the result is contrary. The buyer can make misreporting less attractive to herself by assigning her own technology to the supplier. Suppose both parties announced “efficient”, but the buyer's technology is in fact inefficient. With the supplier's technology, the buyer can enjoy both the low cost of production and the high-quality products. With her own technology, on the other hand, the buyer gets the low-quality products although she pays the supplier for the low cost of production. Thus, by committing to assign her technology when both announcements are “efficient”, the buyer with an inefficient technology can discourage herself from misreporting since she fears losing the high-quality output in case the supplier has an efficient technology. This in turn reduces the distortions in the output schedule for all cases where the buyer announces that her technology is efficient.

Our discussion is extended to the cases where each technology has several states of efficiency. In the case mentioned above, dealing with the misreporting incentive of either party is accompanied only by inefficiency in the output schedule, and not by inefficiency in the choice of technology. With more than two states of efficiency, however, the less efficient technology may be chosen in the optimal contract. For example, consider a misreport by the supplier. Unlike in the binary setting, a misreported efficiency of the supplier's technology can be higher than the efficiency of the buyer's technology. Therefore, removing the misreporting incentive from the supplier (by assigning the buyer's technology) may cost the buyer technological efficiency. If the information rent of the supplier is potentially large enough, the buyer's less efficient technology may be assigned in the optimal contract.

Although there are numerous studies on outsourcing,7 only a few analyze buyer–supplier relationships under two-sided asymmetric information. For example, Demski and Sappington, 1991, Demski and Sappington, 1993 also model a situation in which the buyer has to cope with her own misreporting incentive as well as the supplier's. In their models, devices to resolve the incentive problems are limited to the output schedule and the transfers. We extend this line of study by endogenizing the information structure. In our paper, choosing the informed party is an additional device to ease the incentive problems in an outsourcing relationship.

This paper is also related to studies on organizational mode. Riordan and Sappington (1987) study a model similar to the one in our paper. In their model, the project has two tasks. The first task is always delegated to the supplier, but the second task can be delegated to the supplier (complete delegation) or done by the buyer (partial delegation). Each task has an efficiency parameter that is private information to the one who is in charge. The main difference from the model analyzed in this paper is that in their model, the buyer's private information under partial delegation does not play a key role. The reason is that under partial delegation, although the buyer has private information on the second task, the task is done by the buyer herself, not by the supplier. In our model, the supplier is the only producing party, and hence the buyer's private information becomes an issue when her technology is assigned.

Using an incomplete contract approach, Aghion and Tirole (1997) also study a principal–agent model in which the contract specifies which party has the authority to choose whose information is to be used for the project. The authors show that giving the authority to the agent may increase his effective control over decisions at the expense of the principal, but increases his incentive to be informed. In their paper, however, the only contractual choice is the assignment of the authority since the contract is assumed to be incomplete. Thus, they do not analyze the effect of choosing the effective informed party on distortions in the optimal outcome.

Finally, our study is also related to the principal–agent models with two-sided asymmetric information such as Myerson (1983), Riordan (1984), and Maskin and Tirole, 1990, Maskin and Tirole, 1992. In their papers, unlike ours, being the effective informed party is not a contractual choice.

The remaining sections of the paper are structured as follows. Section 2 presents the basic setup in the case with two states of nature. Section 3 discusses our results. An extension of our analysis to the cases with several states of nature is discussed in Section 4. Section 5 concludes.

Section snippets

Basic setup

We present a monopsonist–supplier model with adverse selection. A buyer (she) must hire a supplier (him) to produce a good and offers him a take-it-or-leave-it contract. We assume that all parties are risk neutral. In exchange for producing q units of output (q⩾0), the supplier receives a transfer t from the buyer. The buyer and the supplier each has her/his own production technology, and hence one of the specifications in the contract is whose technology is to be used for production,

Results

When both the buyer and the supplier have private information, the contract must be Bayesian incentive compatible for both parties because each has her/his own misreporting incentive. First, as in traditional buyer–supplier models, the supplier with an efficient technology (henceforth the θ1 supplier) has an incentive to misreport the efficiency, thus exaggerating the cost of production. In equilibrium, therefore, the following incentive constraint for the θ1 supplier must hold:B=12πB{tB1−[aB1θ

More than two states of nature

In the case with two states of nature, although there could be distortions in the output schedule, there was no distortion in the choice of technology. In other words, when the announcements of θB and θS are different, the efficient technology was chosen in equilibrium. This, however, is not always the case if θB,θS∈{θ1,…,θN} with θN>⋯>θ1, where N>2. A less efficient technology may be chosen in the optimal contract to reduce the misreporting incentives.

To make our point in the simplest set up,

Conclusion

In this paper, we constructed a model in which the supplier can be required to use the technology possessed by the buyer. Our analysis revealed that making the supplier use the buyer's technology mitigates not only the supplier's misreporting incentive, but sometimes the buyer's as well. Consequently, we found that when the technologies of the both parties are equally efficient the buyer's technology is adopted for the production, and the less efficient technology may be assigned in the optimal

Acknowledgements

We are most grateful to Fahad Khalil and Jacques Lawarrée for valuable comments. We also thank Gorkem Celik, Helen Popper, Bill Sundstrom, the seminar participants at the 67th annual meeting of the Midwest Economics Association, and anonymous referees for helpful suggestions. Financial support from Presidential Research Grants at Santa Clara University is acknowledged by the first author.

References (22)

  • F Khalil et al.

    Catching the agent on the wrong foot: ex post choice of monitoring

    Journal of Public Economics

    (2001)
  • T Lewis et al.

    Countervailing incentives in agency problems

    Journal of Economic Theory

    (1989)
  • O Shy et al.

    Strategic outsourcing

    Journal of Economic Behavior and Organization

    (2003)
  • P Aghion et al.

    Formal and real authority in organizations

    Journal of Political Economy

    (1997)
  • J Baron et al.

    Strategic Human Resources

    (1999)
  • J Demski et al.

    Resolving double moral hazard problems with buyout agreements

    RAND Journal of Economics

    (1991)
  • J Demski et al.

    Sourcing with unverifiable performance information

    Journal of Accounting Research

    (1993)
  • G Grossman et al.

    Integration versus outsourcing in industry equilibrium

    Quarterly Journal of Economics

    (2002)
  • Harrison, B., Kelley, M., 1993. Outsourcing and the search for flexibility, work. Employment and Society 7,...
  • F Khalil

    Auditing without commitment

    RAND Journal of Economics

    (1997)
  • Khalil, F., Kim, D., Shin, D., 2003. Optimal task design: to integrate or separate planning and implementation? Working...
  • Cited by (5)

    View full text