Article ID: | iaor199951 |
Country: | United States |
Volume: | 43 |
Issue: | 12 |
Start Page Number: | 1726 |
End Page Number: | 1744 |
Publication Date: | Dec 1997 |
Journal: | Management Science |
Authors: | Seidmann Abraham, Barron Terry, Wang Eric T.G. |
Keywords: | research |
Custom software development projects have special informational attributes that have challenged managers for many years: they are associated with information asymmetries regarding user valuation and developer costs, relationship-specific investments, and a resulting likelihood of positive externalities for the user or the developer from the other party's investment. Furthermore, in a custom project, market prices for software are not helpful in solving either the valuation or the cost problems. In this paper we analyze the unique nature of the software development agreements that can be reached between the user and the developer in such a setting. We compare the value of using internal and external developers, with the goal of better understanding the factors relevant to the outsourcing decision. For internal development, we derive a new mechanism that achieves the first-best system whenever the project has positive expected net value, while achieving ex ante budget balance. In contrast, the optimal mechanism for an external developer will not in general yield the first-best system. This implies that when internal and external developers have identical cost functions, internal development definitely yields the larger net value. More generally, this implies that an external developer must have considerable cost advantages over an internal developer in order to have the larger net value. Numerical results indicate that this difference in net values can be very large, as much as a 100% increase for internal development over external development. This is consistent with the strong bias in favor of internal custom development found in recent empirical studies. We also explain why the efficient levels of investment can be achieved only when there are no externalities, and we show that the presence of positive externalities results in underinvestment. Since using an external developer will typically yield a system that is not first-best, inefficient investments result with or without externalities. We present examples showing that uncertainty about system value is not a significant factor in choosing between internal and external development. However, uncertainty about the development costs is highly significant, with greater uncertainty making outsourcing less attractive.