Timeliness and agglomeration

An important element of the cost of distance is time taken in delivering final and intermediate goods. We argue that time costs are qualitatively different from direct monetary costs such as freight charges. The difference arises because of uncertainty. Unsynchronised deliveries can disrupt production, and delivery time can force producers to order components before demand and cost uncertainties are resolved. Using several related models we show that this generates a hitherto unexplored mechanism for clustering. If final assembly takes place in two locations and component production has increasing returns to scale, then component production will tend to cluster around just one of the assembly plants.