The Use of Inputs by the Federal Reserve System: Comment
评论了Shughart和Tollison关于美联储扩张货币供给是为了增加自身官僚机构规模的观点，指出其分析存在概念、实证和方法上的问题。
In a recent article in this Review (1983), William Shughart and Robert Tollison (S-T) argue that an important motivation for expansion of the money supply is to finance the growth of the Fed's bureaucracy: specifically, that the level of Fed employment is a determinant of the nominal monetary base. Using a public choice framework, they present a money supply equation which yields a positive and statistically significant relationship between the monetary base and the total employment of the Federal Reserve System. While this conclusion is certainly interesting, the analysis suffers from a number of important conceptual, empirical, and methodological problems. Shughart-Tollison point out that the revenue side of the Federal Reserve's budget is closely linked to its monetary policy operations. Through open market purchases of securities, the Fed generates a stream of interest payments. Because monitoring of this income stream is difficult and costly, Congress has elected to impose the constraint that excess revenues be returned to the Treasury. As a result, the Fed pays its current operating expenses out of this current interest income and then turns over the remainder to the Treasury. Because the central bank cannot retain the profits it generates, S-T argue that the Fed will exhibit a form of expense-preference behavior in purchasing more amenities than are justified by costminimizing money production-in particular, by expanding Fed employment. In addition, since expansionary monetary policy increases opportunities for this profit/amenities tradeoff, Fed policies are likely to have a built-in inflationary bias. To test this hypothesis, S-T develop a regression model that estimates the monetary base as a function of Fed employment, real GNP, a number of financial variables, and linear and nonlinear time trend variables. Using annual time-series data from 1915 to 1981, they find that the employment variable is a significant determinant of the money supply, with their model explaining about 95 percent of the variation in the monetary base. To check that employment causes monetary expansion and not the other way around, S-T employ the test of causality developed by C. W. J. Granger (1969) and Christopher Sims (1972). This test estimates two-sided