Hostname: page-component-546b4f848f-w58md Total loading time: 0 Render date: 2023-06-02T09:48:41.927Z Has data issue: false Feature Flags: { "useRatesEcommerce": true } hasContentIssue false

The Antebellum Money Market and the Economic Impact of the Bank War

Published online by Cambridge University Press:  11 May 2010

Marie Elizabeth Sushka
University of Georgia


This paper examines the impact of the Bank War on the economic events of the 1830's. An economic model of the antebellum money market is developed and tested. Specifications for money demand and supply are drawn from contemporary monetary literature and empirically estimated. Next, the historical hypotheses are tested by exploring the structural stability of the model. The results clearly indicate that: the Bank War affected the economy because it altered the pattern of financial behavior; wildcat banking was not characteristic of the post-Bank period; and finally, the Panic of 1837 was the result of a severe monetary contraction.

Copyright © The Economic History Association 1976

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)


1 See, for example, Hammond, B., Banks and Politics in America from the Revolution to the Civil War (Princeton, 1957)Google Scholar; Schlesinger, A., The Age of Jackson (Boston, 1945)Google Scholar; and Meyers, M., The Jacksonian Persuasion (Stanford, 1960)Google Scholar.

2 For example, see Stevens, E., “Composition of the Money Stock Prior to the Civil War,” Journal of Money, Credit and Banking, 3 (02 1971), 84101CrossRefGoogle Scholar; Macesich, G., “Sources of Monetary Disturbances in the United States, 1834-1845,” The Journal of Economic History, 20 (09 1960), 407–34CrossRefGoogle Scholar; Rockoff, H., “Money, Prices and Banks in the Jacksonian Era,” in Fogel, R. and Engerman, S., eds., The Reinterpretation of American Economic History (New York, 1971)Google Scholar; and especially Temin, P., The Jacksonian Economy (New York, 1969).Google Scholar

3 For example, government statistics during the nineteenth century were collected on the basis of different fiscal years during different years. Prior to 1843, the government operated on a September-based year, but in 1844 it changed to a June basis (see United States Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1957 [Washington, 1960]).Google Scholar As a result, for one year, figures reflect only a nine month rather than a twelve month period. This problem is typically ignored in most of these researchers' work; Moreover, data are often compared which do not reflect the same reporting period so that a calendar year figure is inappropriately paired with a fiscal year figure. Temin's derivation of data series for total specie and currency in The Jacksonian Economy is an example of these problems. By the end of the sample period these differences amount to about $25 million.

4 See Temin, Jacksonian Economy.

5 The empirical literature on money demand is large, but among the more important contributions are Brunner, K., “Some Further Evidence on Supply and Demand Functions for Money,” Journal of Finance, 19 (May 1964), 240–83CrossRefGoogle Scholar; Meltzer, A., “The Demand for Money: The Evidence from the Time Series,” Journal of Political Economy, 71 (June 1963), 219–46CrossRefGoogle Scholar; and Latane, H., “Cash Balances and the Interest Rate—A Pragmatic Approach,” Review of Economics and Statistics, 36 (November 1954), 456–60.CrossRefGoogle Scholar See Laidler, D., The Demand for Money: Theories and Evidence (Scranton, 1969)Google Scholar for a comprehensive survey of this literature.

6 This model is drawn from the Latane “Cash Balances” formulation in which

where M* is desired money balances and it is hypothesized that actual money demand, M, adjusts only gradually to the desired amount:

Rewriting the above equation in logarithmic form and solving for M yields the specification (3) which is used for empirical testing.

7 This procedure is standard in such works as Slovin, M. and Sushka, M. E., Interest Rates on Savings Deposits (Lexington, 1975)Google Scholar and Modigliani, F., Rasche, R., and Cooper, J. P., “Central Bank Policy, the Money Supply and the Short Term Rate of Interest,” Journal of Money, Credit and Banking, 2 (May 1970), 166218CrossRefGoogle Scholar

8 The importance of price expectations in analyses of money demand has been a basic tenet in the views of the modern quantity theorists, especially Friedman, M., “The Demand for Money, Some Theoretical and Empirical Results,” Journal of Political Economy, 67 (June 1959), 327–51.CrossRefGoogle Scholar

9 Although the Second National Bank began operations in 1817, the initial years of its existence are omitted so that the observations form a sample period characterized by relatively consistent and straightforward financial policies. See Studenskii, P. and Krooss, H., Financial History of the United States (New York, 1963).Google Scholar

10 In the money demand literature which is formulated in aggregate terms, the dependent variable is usually taken to be Ml, defined as demand deposits plus currency, or M2, defined as commercial bank time deposits plus M1. For further information about money demand research which disaggregates money holdings, see Goldfeld, S., Commercial Bank Behavior and Economic Activity (Amsterdam, 1966)Google Scholar; Modigliani, Rasche, and Cooper, “Central Bank Policy”; and Slovin and Sushka, Interest Rates, ch. 8.

11 Each banking series is constructed from data in Fenstermaker's, J. Van, The Development of American Commercial Banking, 1782-1837 (Kent, 1965)Google Scholar for the years up to and including 1833, and from the Comptroller's data for the period 1834 to 1859 which is found in United States Treasury Department, Annual Report of the Comptroller of the Currency, 1876 and 1896 (Washington, D.C., 1876, 1896)Google Scholar, and the United States Treasury Department, The Executive Documents, 1892-1893, Vol. 5 (Washington, 1893).Google Scholar Empirical tests were also conducted on series obtained through private correspondence. See M. E. Sushka, “Estimating Income in the Andata for 1838 through 1859. The regression results were essentially identical so that the conclusions reported in this study are not sensitive to the choice between the two data sources.

12 The regressions in Table 1 are estimated using ordinary least squares. Whe n the presence of positive serial correlation in the residuals is indicated, the regression is re-estimated using the first-order autoregressive technique of Cochrane, D. and Orcutt, G., “Application of Least Squares Regression to Relationships Containing Autocorrelated Errors,” Journal of the American Statistical Association, 44 (March 1949), 3261.Google Scholar Throughout the paper, the t statistics are considered to be significantly different from zero if they lie outside a 95 percen t confidence interval about zero.

13 The income series is derived by midcasting a series developed by Gallman which was obtained through private correspondence. See M. E. Sushka, “Estimating Income in the Antebellum United States,” working paper, 1974.

14 All regressions were also estimated in real terms, but the results were very similar and thus are not reported. All reported equations are in logarithmic form so that the coefficients are the elasticities of the variables.

15 For an explanation of the midcasting techniques utilized in order to supply the missing observations in the interest rate series reported in Homer, S., A History of Interest Rates (New Brunswick, 1963)Google Scholar, see Sushka, M. E., “A Note on Interest Rates During the Early 19th Century,” working paper, 1974Google Scholar.

16 This short-term rate series is only available beginning in 1831; and therefore, inclusion of a dummy variable is required when RCP is specified in the regressions.

17 Only one interest rate variable is specified in each regression since severe multicollinearity in time series interest rate data prohibits the effective estimation of more than one interest rate parameter in a given equation.

18 The size of the income elasticity of the demand for money has been a topic of frequent discussions in contemporary literature. Overall, most economists believe that this elasticity is close to unity. See , Laidler, Demand for Money, p. 107Google Scholar, and Poole, W., “Whither Money Demand?,” Brookings Papers on Economic Activity, 1 (1970)Google Scholar.

19 Although each of these three interest rates were used throughout the empirical testing of the money demand models, in each case both RMUN and RCP uniformly fail to obtain a significant coefficient and, consequently, only the regression results for the government bond rate are analyzed further.

20 These speeds of adjustment are similar to the results implied by modern money demand equations. Usually the speed of adjustment, estimated for quarterly data, ranges between 10 and 20 percent per quarter, which is generally considered to be fairly slow—see, for example, Modigliani, Rasche, and Cooper, “Central Bank Policy.” However, faster speeds of adjustment have been reported by Slovin and Sushka, Interest Rates, ch. 8.

21 It should be noted that the statistical properties of equations with lagged dependent variables are sometimes suspect. In the absence of autocorrelated disturbances, however, least squares estimates remain consistent even if lagged values of the dependent variable are specified in the equation. As Griliches, Z. has demonstrated in “Distributed Lags: A Survey,” Econometrica, 35 (January 1967), 1649CrossRefGoogle Scholar, the problem is that lagged dependent variables can sometimes capture the serial correlation of the residuals rather than providing any information about the dynamic behavior of the dependent variable. In order to provide additional information about this question, a specification is estimated in which both current and lagged income enter as independent variables:

The regression results do not provide any indication of serious lags in money demand behavior and thus the results of equation (1.4) must be regarded with caution. It is reassuring that the steady state properties of equation (1.4) are almost identical to the results obtained by specifying only current income and interest rates, equation (1.1). In order to test further for the possibility that the results in equation (1.4) may reflect the impact of autocorrelation in the residuals rather than the presence of partial adjustment, a further test developed by Griliches was carried out in which a lagged independent variable is added to the specification. If the autoregressive model rather than the partial adjustment model is appropriate, such a variable should be negative and significant and obtain a coefficient equal to the coefficient of the lagged dependent variable multiplied by the coefficient of the unlagged independent variable. The re-estimated equation results are

The evidence clearly indicates rejection of the hypothesis of serial correlation and acceptance of the validity of a partial adjustment model since the coefficient of the lagged exogenous variable is neither significant nor equal to the product of the coefficients of the lagged dependent variable and the independent variable (.142 ≠ .582 × .522).

22 Since the ordinary least squares estimation was characterized by a high degree of serial correlation in the residuals, only the results obtained by applying the Cochrane-Orcutt, “Autocorrelated Errors,” autoregressive technique are reported. The variables are again in nominal terms and all equations are estimated in logarithmic form.

23 , Goldfeld, Commercial Bank Behavior, pp. 7991Google Scholar, hypothesizes such a result and finds empirical support in his currency and demand deposit equations when specified in non-logarithmic form. In his logarithmic equations, however, the income elasticity for currency exceeds that of demand deposits. For previous work on the interest elasticity of currency demand, see Goldfeld, pp. 80-81, and , Modigliani, , Rasche, and , Cooper, “Central Bank Policy,” pp. 177–79Google Scholar.

24 Once again, it is possible that the lagged dependent variable captures the effect of serial correlation in the residuals rather than any real dynamics in behavior so that these results should be treated with caution. See footnote 21 for further explanation.

25 See Fenstermaker, Van, Banking, 1782-1837, pp. 4353Google Scholar, for a description of bank asset portfolios during the early 1800's. For a contemporary analysis of bank portfolio behavior, see Goldfeld, Commercial Bank Behavior.

26 See , Modigliani, , Rasche, and , Cooper, “Central Bank Policy,” pp. 191–93.Google Scholar

27 This formulation is standard in the recent literature. See, for example, “Central Bank Policy,” p. 188. Th e original exposition of this new view is found in Tobin, J., “Commercial Banks as ‘Creators’ of Money,” in Carson, D., ed., Banking and Monetary Studies (Homewood, 1963).Google Scholar

28 An alternative specification for the money supply, based on an approach initially developed by Teigen, R., “The Demand for and Supply of Money,” in Teigen, R. and Smith, W., eds., Money, National Income, and Stabilization Policy (Homewood, 1965)Google Scholar, was also tested. Since the empirical results were quite similar, the development and testing of this model is not discussed in this paper.

29 The regressions are estimated using ordinary least squares but when the presence of serial correlation is indicated, the regressions are re-estimated using the autoregressive technique of Cochrane and Orcutt in “Autocorrelated Errors.” The t statistics appear in parentheses below the coefficients.

30 To provide a more efficient estimate of the remaining parameters, the equation is re-estimated omitting the interest rate variable. This procedure is applied throughout this section whenever an interest rate coefficient is not significant or has an incorrect sign.

31 Both the traditional historical view of the Bank War and the revisionist version assert conclusions which primarily deal with the effect of the Bank War on bank behavior. Thus, the stability of the money supply function is particularly important.

32 In order to test the sensitivity of these split sample results and to provide additional information about the exact pattern of the shifts, the money demand and supply equations were re-estimated to allow for a split in the sample period for each year from 1832 to 1841. This procedure is commonly used in modern econometric analyses in order to evaluate the stability of functional relationships. This technique is employed, for example, by Eckstein, O. and Brinner, R., “The Inflation Process in the United States,” in A Study, prepared for the use of the Joint Economic Committee, United States Congress, 92nd Congress, 2nd session (Washington, D.C., 1972), pp. 2425Google Scholar, in an analysis of the Phillips Curve; and by Goldfeld, S., “The Demand for Money Revisited,” Brookings Papers on Economic Activity, 3 (1973), 539–90Google Scholar, in his study of money demand. In particular, the empirical evidence obtained by applying this procedure clearly indicated that the shift in the pattern of behavior of both money demand and supply occurred and was completed prior to 1837, and thus these observed shifts could not have been induced by the Panic of 1837. This additional empirical evidence confirms the pattern of results reported in the text.

33 Temin, Jacksonian Economy.

34 The data indicate that total specie continued to rise throughout the period 1831-1839. See Bureau of the Census, Historical Statistics.

35 It is not clear that such policies would have bee n carried out by a central bank. In a similar situation during the 1930's, the Federal Reserve actually sought to reduce the quantity of high-powered money, a policy which intensified th e economic crisis. See Friedman, M. and Schwartz, A., A Monetary History of the United States, 1867-1960 (Princeton, 1963).Google Scholar