Academic journal article Journal of Money, Credit & Banking

Further Evidence concerning Expense Preference and the Fed

Academic journal article Journal of Money, Credit & Banking

Further Evidence concerning Expense Preference and the Fed

Article excerpt

In an interesting paper, Boyes, Mounts, and Sowell (1988) (hereafter BMS) empirically studied the existence of expense-preference behavior on the part of the Federal Reserve. The paper hypothesized that the Fed, being a bureaucracy, might be more interested in maximizing its size or prestige rather than its profits, which are turned over to the Treasury Department and so the taxpayers. If so, then managers of the Fed might maximize their own utility that is a function of expenditures as well as profits, seeking to increase expenditures beyond their profit-maximizing levels and indulging in expense-preference behavior. Not being closely monitored by sponsors or owners would allow such managers to do this (Edwards 19877). BMS's empirical findings suggested that the Fed practiced expense preference toward labor over the sample period 1917-81. They posited that the Fed's expense-preference behavior might affect its conduct of monetary policy. Indeed, Shughart and Tollison (1983) posited that the Fed's preference for amenities gives its monetary policy an "inflationary bias."(1)

In this paper I seek further evidence on the existence of expense preference on the part of the Fed. I expand on the model of Fed behavior BMS used and test for expense-preference behavior in a more general specification. I find that the data for the sample period reject the joint hypothesis that the Fed uses a Cobb-Douglas production technology and exhibits expense-preference toward labor; the data also reject the hypothesis that the Fed uses a Cobb-Douglas technology and maximizes profits. Thus, the data imply that the Fed does not use a Cobb-Douglas technology, one of the assumptions of BMS's model. Using a more general test of expense preference, developed in Mester (1989), which permits a less restrictive production technology than Cobb-Douglas, I find evidence that the Fed did not indulge in expense-preference behavior over the sample period.

The rest of this paper is organized as follows. Section I derives an alternative joint test of expense preference and Cobb-Douglas technology to the one used by BMS. Section 2 presents the empirical results of this test. Section 3 discusses a more general test of expense preference that does not presume the Fed uses a Cobb-Douglas technology. Section 4 presents the empirical results of this test. Section 5 concludes.

1. TESTS FOR EXPENSE PREFERENCE ASSUMING A COBB-DOUGLAS TECHNOLOGY

I start with the model of the Federal Reserve System used in BMS. According to this model, the Federal Reserve System uses a Cobb-Douglas production technology of the following form:

(1) [Mathematical Expression Omitted] where [M.sub.t] = monetary base, [L.sub.1t] = labor at the Board of Governors, and [L.sub.2t] = labo at the Federal Reserve Banks. This model is consistent with Shughart and Tollison (1983) and Toma (1982). The demand function facing the Fed is of the following form:

(2) [Mathematical Expression Omitted] where [R.sub.t] = appropriate monetary interest rate and [Y.sub.1t] = real gross national product, and [Y.sub.2t] = the currency-to-deposits ratio.(2)

In Section 3, where I allow a more general production technology on the part of the Fed, I will expand the model to include outputs that might better reflect the Fed's output and that are more directly related to the Fed's use of inputs, such as number of pieces of currency and coin received and counted, and number of checks handled. For now, I analyze the model used by BMS in order to focus on its assumption of a Cobb-Douglas technology.

Letting [[omega].sub.1t] = wage of labor at the Board of Governors and [[omega].sub.2t] = wage of at the Federal Reserve Banks, then profits are

(3) [Mathematical Expression Omitted]

If the Fed managers act to maximize profits (that is, do not exhibit expense preference), then they choose inputs to maximize (3) subject to (1). This implies labor demand functions:3

(4) [Mathematical Expression Omitted]

(5) [Mathematical Expression Omitted]

On the other hand, the Fed might exhibit expense preference for labor by maximizing a utility function of the form U = U([[pi]. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.