Technology Adoption with Uncertain Future Costs and Quality

Abstract

In this paper we study the impact of uncertainty about future innovations in quality and costs on consumers' technology adoption decisions. We model the uncertainty in the technology's quality and costs as a Markov process and consider three models of the adoption decision. The first model assumes that consumers do a simple net present value (NPV) analysis that compares the NPV of adopting to that of not adopting, without considering the possibility of waiting. The second model is a stochastic dynamic program that considers the possibility of waiting and views the adoption decision as a one-time event, i.e., the consumer will only make a single purchase, the only question is when. The third model allows repeat purchases so the consumer may “upgrade” by purchasing new versions of the technology whenever it suits her. We study structural properties of these models, e.g., the following: What changes in qualities and costs will make the consumer better off? What changes will encourage adoption? We will see that the simple NPV and single-purchase model have many intuitive properties: with the right notion of improvements and reasonable assumptions about the technology changes, we find that improvements in the technology make the consumer better off and encourage adoption. Here improvements are defined using a partial order on quality and cost pairs. The results are more complicated in the repeat-purchase model. Under the same conditions on technology changes, technology improvements will make the consumer better off. However, except for special cases of transitions, these improvements may make the consumer better off and discourage adoption.