Excess reserves during the 1930s: Empirical estimates of the costs of converting unintended cash inventory into income-producing assets

Journal of Economics and Finance, Summer 2001 by Lindley, James T, Sowell, Clifford B, Mounts, W M Stewart Jr

Abstract

It is often argued that the persistent amounts of excess reserves in the 1934-1941 period were sought either for protective liquidity or as a signal of bank safety to depositors. More recent explanations argue that these excess reserves were unintended inventory due to the high internal adjustment costs of converting reserves to income-producing assets. Our findings support the latter explanation and reveal high internal asset adjustment costs after 1933. Thus, a monetary policy focused on increasing reserves would have been ineffective. A successful monetary policy would be one that increased outside money. (JEL 6210, 6280, 0420)

Introduction

The modern theory of the banking firm posits that an increase in the demand for loans and securities (bank outputs) leads to an increase in the bank's demand for deposits (inputs).1 In this context, the acquisition of deposits is endogenous to bank managerial decisions. A departure from deposit endogeneity occurred during the decade of the 1930s. There were at least two contributing factors. First, banking legislation of the time reduced the ability of banks to price-compete for deposits by instituting Regulation Q and by placing a ban on paying interest on checking deposit accounts.2 Second, as shown in Table 1, there was a rapid return of deposits following the Bank Holiday (declared on March 6, 1933) but to a smaller number of banks. Between 1929 and 1933, the number of banks declined from 24,633 to 15,015.(3)

1 This view can be traced back to Sealey and Lindley (1977).

2 The legislation was contained in two acts: the Banking Act of 1933, which created deposit insurance, and the Banking Act of 1935.

3 Bank numbers are taken from the Board of Governors of the Federal Reserve, 1943, Banking and Monetary Statistics, Washington, D.C., p. 16.

4 Bankers were confronting a situation similar to that of a mutual fund that experiences large cash inflows when its investment opportunities are restricted to a narrow range of assets.

5 Excess reserves are reserves above required reserves. Warburton notes, "This episode is the only exception to the general tendency of Federal Reserve member banks to expand in line with changes in their reserve position" (Warburton, 1950, p. 540).

6 Alternative forms of protection would be a willing lender of last resort as suggested by Friedman and Schwartz (1963) and access to de novo capital as suggested by Calomiris and Wilson (1996) and Ramos (1996).

7 Indeed, most of the expansion from 1934 through 1941 has been attributed to a large increase in outside money, mostly gold inflows (Romer 1992).

s It should be noted that critics of the Federal Reserve's actions during this period, such as Friedman and Schwartz as well as Warburton, were critical of the Fed's feeble and often counterproductive attempts to conduct monetary policy of any kind. The Fed did not conduct policy that expanded either bank reserves or fiat money. (Friedman and Schwartz 1963, p. 511). Romer notes, "While they were clearly aware that other developments led to a rise in the money supply during the mid-1930s, Friedman and Schwartz appear to have been more interested in the role that Federal Reserve inaction played in causing and prolonging the Great Depression than they were in quantifying the importance of monetary expansion in generating recovery" (Romer 1992, p. 758).

It could be argued that the inclusion of R, and D, may be referring to those variables subsumed under an interest rate model (Frost 1971) and a protective liquidity model (Friedman and Schwartz 1963) of the accumulation.

'o The solution to expression (2) is given in Sargent (1987) and Hansen and Sargent (1981). Elsewhere, content-specific treatments appear in Cuthbertson and Taylor (1987) and Cuthbertson (1988).

" Data on excess reserves (XR,) and the yield on government securities (R,) were taken from Banking and Monetary Statistics (1943). XR, is the log of excess reserves, Table 101, pp. 369-72. R, is the yield on government securities, Tables 128, pp. 469-71. Data on deposits and the two monetary policy variables are given in Friedman and Schwartz (1963).

'2 The interest rate used in the estimation is consistent with those used by Frost. However, the results reported below are insensitive to the interest rate that is selected.

" Estimation was carried out using real variables for XR,, D,, and R, with no appreciable effect on the results. This is not surprising given the variables involved. Excess reserves and deposits are both in dollar terms where excess reserves are an amount over the percentage of deposits banks were required to hold. Deflation of deposits and excess reserves would be a scale shift in their values.

14 Successful joint estimation of (6) and (7) becomes unduly complex when dealing with monthly data if (7) has long lags and a VAR representation. The addition of surprise terms, so crucial to an explanation of the 1930s, imposes a formidable programming and computational burden. The problem is simplified by restricting (7) to have a univariate representation in each of the forcing variables.