So Close and Yet So Far: Information Technology and the Spatial Distribution of Customer Service (Research Seminar, January 29th, 2004) David Fitoussi MIT Sloan School Abstract Economic research has only recently demonstrated the impact of information and communication technologies (IT) on economic growth and productivity. There is less systematic evidence though on the economic impact of IT on the work reorganization that accompanies the deployment of these technologies. On the one hand, the ability to gather, process and move information at very low cost is offering firms opportunities to structure work processes in innovative ways. This is giving rise to predictions that a new wave of globalization, of services this time, will fundamentally affect where and how work is performed. And indeed, faced with incessant pressure to control and reduce costs, American and European firms are urged to move operations off-shore where unit costs are much lower. On the other hand, recent research has emphasized the need to complement IT investments with substantial investments in intangible assets, some of which firms might have difficulty replicating in other locations. In this case, taking advantage of regional cost differentials may not only be more difficult, but might in fact increase the relative value of location-bound assets. In this research, I provide evidence on the impact of IT on localization patterns of firms between regions of the continental United States. Building on a detailed data-set of customer-care hiring decisions by about 100 Fortune 1000 manufacturing firms, I develop a model in which firms choose both the location of their customer-service representatives and how many to hire at each location. The framework is a discrete-choice model in which regions play the role of differentiated products. I allow for flexible substitution patterns between regions by using random coefficients. The latent variable of the model, the firm's profits from customer care, is derived from the premises of a queueing (stochastic) process. The estimated demand structure is used to assess the effects of information technology on customer volume, location choices and cost savings. The results confirm the higher price sensitivity of IT-intensive firms, but also suggest that the ability to exploit cost differentials is highly firm-specific and that the importance of local assets does not vanish