Estimating economies of scale and scope with flexible technology

Economies of scope are typically modelled and estimated using a cost function that is common to all firms in an industry irrespective of their type, e.g. whether they specialize in a single output or produce multiple outputs. Instead, we estimate a flexible technology model that allows for type-specific technologies and show how it can be estimated using linear parametric forms including the translog. A common technology remains a special case of our model and is testable econometrically. Our sample, of publicly owned US electric utilities, does not support a common technology for integrated and specialized firms. Our empirical results therefore suggest that assuming a common technology might bias estimates of economies of scale and scope. Thus, how we model the production technology clearly influences the policy conclusions we draw from its characteristics.