45 Pages Posted: 18 Jul 2006
Date Written: March 1, 2007
This paper estimates a structural model of demand for the personal computer (PC) by repeat purchasers. Taking advantage of a large dataset on household-level PC purchases, the econometric model uses variation in PC holdings among PC owners to identify households' marginal values of quality improvements. The analysis uses only cross-sectional data, and accounts for: stock effects, forward-looking behavior, and large amounts of household heterogeneity. The estimates allow us to measure sensitivity to long-term and short-term price and technology changes, as well as consumer welfare changes from technological improvements. The results show a large variation in marginal values for PC quality across households, and that failing to account for forward-looking behavior results in biased estimates and a poorer fit to the data. Incorporating stock effects proves especially important since, for the data used here, the model's parameters are not only biased, but virtually unidentified without them. The results also show that price elasticity is approximately 25% higher in the short-term compared to the long-term, and technology elasticity is approximately 35% higher in the short-term compared to the long-term. Further, welfare measurements are significantly underestimated when using a model that doesn't account for forward-looking behavior. Finally, the model is extended to include first-time purchasers. The results show similar patterns, but should be interpreted with much caution due to the likely presence of significant unobserved heterogeneity between new purchasers and repeat purchasers.
Keywords: Personal computers, demand, structural estimation, dynamics, technology elasticity, long-term elasticity, short-term elasticity, stock effects, durable good, replacement
JEL Classification: L0
Suggested Citation: Suggested Citation
Prince, Jeffrey, Repeat Purchase Amid Rapid Quality Improvement: Structural Estimation of Demand for Personal Computers (March 1, 2007). Available at SSRN: https://ssrn.com/abstract=917865 or http://dx.doi.org/10.2139/ssrn.917865