money, they achieve this by using a number of different mone- tary totals: the monetary base, currency plus demand deposits, currency plus all commercial bank deposits. For the government budget they do so by using high employment expenditures and receipts. For both, autonomy is reinforced by using lagged values of monetary and fiscal policy variables with different methods of allowing for lags. All variants yield essentially the same results. Of course, no study can "prove” anything finally. Proof is re- served for logical, not empirical, propositions. What a study can do is to contradict or fail to contradict hypotheses, and even then, of course, any findings are always tentative because the evidence is necessarily incomplete and there is always the possi- bility that the hypothesis contradicted can be reformulated so as to be consistent with the initially contradictory evidence. However, as a veteran of many years’ standing of the kind of smoke-screen criticism levelled by Heller against the Ander- son-Jordan study, I would take it far more seriously if the as- sertion that something may be wrong were accompanied by some evidence, empirical or analytical, that something is wrong. 12. The references added by Professor Heller in footnote 1 of his paper do not meet this challenge. They are criticisms of my work of the same kind as, though naturally more elaborate and developed than, his brief criticisms of the Anderson-Jordan article discussed in the preceding footnote. None of the criticisms gives any systematic evidence supporting the hypothesis that fiscal policy by itself has a significant influence on nominal in- come. To complement the references in his footnote, I should note an article of mine, “Taxes, Money and Stabilization,” Wash- ington Post, November 5, 1967, because it is partly in answer to several critical pieces by Tobin in the Washington Post, including the one referred to by Heller. His footnote 8 gives two other references to OECD docu- ments. Since I have not yet seen these I cannot judge whether they provide relevant empirical evidence. Digitized for FRASER Federal Reserve Bank of St. Louis
88 Notes 1. Gramley, who may know more about Professor Friedman's work than anyone other than Friedman himself, has compared actual yearly changes in nominal Net National Product (NNP) with those predicted by the Friedman equation as reported in “Interest Rates and the Demand for Money,” Journal of Law and Economics , October, 1966. For the period 1948-60, the average yearly change in nominal NNP was $18.6 billion. Gram- ley found the average predicted changes to be $5.5 billion. It is quite true that for the period 1960-67, the two average changes are nearly identical ($37.2 billion versus $36.8 billion). But this does not contradict the point that a fixed rule ought to allow for trend changes in velocity. And if trend changes are allowed for, why not shorter run changes?
You've reached the end of your free preview.
Want to read all 95 pages?