If it were easy to classify most external productivity effects as either global
in scope or as so localized as to be internalizable at the level of the family or
the firm, then I think a model that incorporated internal human capital effects
only plus other effects treated as exogenous technical change would be
adequate. Such a model would fit time series from advanced countries about as
well as any I have advanced, being an intermediate model to those I discussed
in sections 2 and 4, which were in turn not distinguishable on such data alone.