Published online by Cambridge University Press: 14 March 2022
The relationship between a class of structural forms and a single reduced form econometric model is discussed. These are shown to be empirically equivalent. It is made clear that choice of a particular structural form for estimation, rather than another of the same class, rests on a priori heuristic and computational considerations, not on empirical or logical grounds.
Alternative scientific strategies are discussed briefly.
This paper was begun in 1959-60 while the author held a Ford Foundation Faculty Fellowship. Some additional work was done under support by the Office of Naval Research, Contract Nonr 233-02 under task NR 047-003, Management Sciences Research Project, University of California, Los Angeles.
I am indebted to my colleague, John Henning, for his thoughtful remarks. It is unnecessary to add that the responsibility for any errors rests with me.