ABSTRACT
Software poses an unprecedented challenge to economics because it is not subject to marginal cost in the usual sense. The cost of software does not depend on how many copies are made, how many copies are sold or how many are shipped from makers to customers. Anybody who has downloaded software knows this. A software application is a bundle of features of which some, perhaps all, are valuable to the users. The maker of a software application incurs the cost of inventing and developing its features. Therefore, the cost of software depends on its content, not on how many units are produced or sold. Notwithstanding the novelty of these issues, my analysis of the economics of software draws on concepts going back to Ricardo. The key assumption is this. Any software that can generate enough revenue to cover its total cost will be placed on the market. This is akin to Ricardo’s contention that any land that can yield rent will be drawn into cultivation. A formal model is indispensable for meeting software’s challenge to economics.
