Note also that this would seem to suggest what economists call a ‘production function’ of the form Yt = f(CIt)α, with the models distinguished by whether α is less than, equal to, or greater than 1. However, we do not seek to estimate a production function (and hence α) because of both insufficient consistent time series data for CI and the difficulties of interpreting a cross-country regression. Production functions for the cultural industries have been developed and estimated (e.g. Gapinski 1980, 1984), Throsby 2006), but what we refer to here is the notion of an entire industry as an input into a growth equation, which is less conventional.