¹ For perhaps the most comprehensive formulation and extensive support of this hypothesis see North, D. C., The Economic Growth of the United States, 1790-1860 (Englewood Cliffs, N.J.: Prentice-Hall, 1961)Google Scholar.

² Each component of the balance of payments, of course, has direct effects upon the stability and growth of the economy apart from their role in determining compensatory flows. Furthermore, there are some who consider the link between monetary stability and general economic stability to be somewhat tenuous. While highly important, such considerations do not fall explicitly within the scope of this study and hence will be considered only peripherally.

³ Macesich, George, “Sources of Monetary Disturbances in the United States, 1834-1845,” Journal Of Economic History, XX (09 1960), pp. 407–34CrossRefGoogle Scholar;Williamson, Jeffrey, “International Trade and United, States Economic Development, 1827-1843,” Journal Of Economic History, XXI (09 1961), pp. 372–83CrossRefGoogle Scholar; and Macesich, George, “International Trade and United States Economic Development Revisited,” Journal Of Economic History, XXI (09 1961), pp. 384–85.CrossRefGoogle Scholar See also , William-son'sAmerican Growth and the Balance of Payments, 1820-1913 (Chapel Hill: The University of North Carolina Press, 1964)Google Scholar.

⁴ Richard Timberlake has postulated the existence of an ancillary fiscal mechanism whereby changes in the trade balance may reinforce the effects of specie flows through their effects on the size of the central government's surplus or deficit. Timberlake's mechanism does appear to have been in operation in the United States in the 1838 to 1843 period. Large capital movements and revenues from public land sales tended to mitigate the workings of this mechanism for the remainder of the period, however. See Timberlake, Richard H. Jr, “The Independent Treasury and Monetary Policy Before the Civil War,” Southern Economic Journal XXVII (10 1960), pp. 95–96Google Scholar.

⁵ , Macesich, “Sources,” p. 419Google Scholar.

⁶ In rejecting Macesich's specie-flow hypothesis, Williamson pointed out that in the period 1879-1904 there was an almost perfect correlation between gold flows and movements in the money supply and expressed doubt that the difference between these two periods could be explained by the increasing maturity of the American money market.

⁷ This point is considered further in section II, for it forms one of the major pillars of the alternative explanatory hypotheses presented in this article.

⁸ , Williamson, American Growth, p. 186Google Scholar.

⁹ Patinkin, Don, Money, Interest, and Prices (2d. ed.; New York: Harper and Row, 1965)Google Scholar. For a full treatment of this framework of analysis and explanation of terminology the reader is referred to this volume, especially chs. i-iii and ix-xi. For a more sophisticated application of this approach to external balance than is employed in this article see Mundell, Robert, “The Monetary Dynamics of International Adjustment Under Fixed and Flexible Exchange Rates,” Quarterly Journal of Economics LXXIII (05 1960), 227–57CrossRefGoogle Scholar.

¹⁰ He does, however, briefly make reference to this mechanism in his book (pp. 182-83) during his discussion of the 1879-1914 period.

¹¹ This is just the standard economic presumption that an increase in output with a given level of aggregate demand will decrease prices.

¹² Price and wage rigidities, if present, would divert downward price pressures into the labor market, causing unemployment and reducing real income. Through the Keynesian foreign trade multiplier this would discourage imports, improving the current account and attracting specie.

¹³ This situation would thus correspond to quadrant 1 of Figure 1. The “rest of world,” for which Great Britain would serve as a good proxy during our period of analysis, would display just the reverse tendencies, that is, its situation would be represented by quadrant 3.

¹⁴ Or to put it somewhat differently, “the cash-balance effect… rules out income equilibrium without balance-of-payments equilibrium.” Yeager, Leland B., International Monetary Relations (New York: Harper and Row, 1966), p. 77Google Scholar.

¹⁵ He generally referred to gold in parentheses when speaking of an excess demand for money.

¹⁶ The continuing controversy over the need or usefulness of “realistic” assumptions in economic theory is not really relevant here, for Williamson explicitly set for himself the task of presenting “an alternative explanatory model, more complex than Macesich's, but providing … a more realistic explanation of the observed interaction between external and internal conditions.” (“International Trade,” pp. 372-73).

¹⁷ Real growth would stimulate an increased demand for real balances which could be “manufactured” by a decline in the price level. Alternatively, equilibrium could be restored by increases in the interest rate which would decrease the demand for real balances.

¹⁸ Mundell, Robert, “A Fallacy in the Interpretation of Macroeconomic Equilibrium,” The Journal of Political Economy LXXIII (02 1965), 61–66CrossRefGoogle Scholar.

¹⁹ See the Federal Reserve Bank of New York's cost-of-living index, U.S. Bureau of the Census, Historical Statistics of the United States: Colonial Times to 1957 (Washington: U.S. Government Printing Office, 1960), E-157, p. 127Google Scholar.

²⁰ See , Williamson, “International Trade,” p. 377.Google Scholar Williamson dismissed the episode with the comment that “the flow of gold may have over-satisfied excess demands in the money market generated by real income growth” (p. 377).

²¹ , Williamson, American Growth, p. 187Google Scholar.

²² Ibid., p. 187.

²³ This would not, of course, be the case if the excess demand for money has been generated by a financial panic.

²⁴ In his book (American Growth, pp. 175-76), the argument is softened to read only that “the flow of gold … offsets variations in the ratio (of money supply to specie backing) as well as in domestic supplies and at least partially dominates the money supply.”

²⁵ This would imply that Williamson's XDM would not correspond to XDM in the conventional (Patinkin) sense. If, in fact, it does refer only to excess demands for specie, then the inclusion of the bond and commodity markets would not be sufficient to complete a general-equilibrium system such as Williamson posited he was using. On the other hand, it seems difficult to interpret consistently in any other manner, Williamson's statement that, “Rapid growth has a tendency to generate excess demands for money (gold) and goods (trade balance deficit), as well as excess supplies of securities (net capital inflow.)” (“International Trade,” p. 383).

²⁶ If banks actively attempted to import gold to replenish their holdings rather than passively responding to the demands of their debtors and creditors, then this implied correlation need not hold.

²⁷ Williamson himself confirms this relationship. See American Growth, p. 185.

²⁸ Williamson, ibid., p. 185.

²⁹ In fairness to both Macesich and Williamson it should be pointed out that for the purpose of contrast the main body of this article treats the more extreme forms of their models. Their actual differences would thus, I suspect, be overstated in this summary. Because of the almost exclusively critical tone of my analysis to this point I should also like to record my admiration for the many valuable contributions contained in their work.

³⁰ While the independent Treasury did undertake some of the functions of a central bank (see Timberlake, “The Independent Treasury”), enforcement of the “rules of the game” was not one of these.

³¹ The emphasis placed here on the manner in which specie flows affected the domestic monetary situation does not imply that they were the only cause of movements in the expansion ratio. As has been argued, die banks did have considerable leeway in the determination of the domestic money supply, and they did have economic incentives to vary the money supply in response to changes in the demand for money (often caused by growth as Williamson pointed out). The expansion ratio expresses the relationship between specie reserves and the money supply, and both of these variables could and did change considerably during the period. It makes little difference whether an expansion ratio increase was caused by a specie outflow while the money supply was held constant or was falling less rapidly, or by an increase in the money supply (such as in 1834-1836) which was more rapid than the rate of specie inflow. In either case, by initiating or allowing offsetting movements in the expansion ratio to take place, the banking system was able, over fairly long periods of time, to insulate, partially at least, the domestic monetary system from the constraints of external balance.

³² The mechanism by which variations in the expansion ratio could precipitate financial collapses will be discussed extensively in the latter portion of this section.

³³ The repeated overexpansion of the money supply during the period would at first glance seem to imply that in aggregate the banking community systematically underestimated the influence of specie flows and the expansion ratio on the demand for money; that is, that it never learned that rapid monetary expansion relative to its specie base was highly conducive to financial collapse. There may be an alternative explanation, however. Each banker might correctly feel that he individually would have little influence on the aggregate expansion ratio or on that of his region or town. Thus even if he knew that the banking system in aggregate were rapidly overexpanding, profit-maximizing behavior might call for him to expand his note issue rapidly also, hoping to make as much return as possible before the expected panic which it was impossible to avoid. The situation might thus be analogous to the famous prisoners' dilemma in which minimax behavior leads all players into suboprimal positions from their point of view. Each bank would act as if it had no influence on the demand for money or specie (through their influence on the expansion ratio), although all of the banks in aggregate would.

³⁴ The introduction of such a time lag in the statistical analysis was unavoidable, since North's specie-flow figures were estimated for the middle of each year and specie reserves and money-stock estimates are for the first of each year. The correlation of the specie level or money stock with the gold-flow figures of the preceding year thus introduces a real lag of approximately six months. The use of this lag, however, does not appear inappropriate for considering the impact of specie flows on the expansion ratio. The half-year lag allows us to observe the figures after the banks have had considerable time to adjust to the specie flows, while at the same time not being so long as to wash out the effects of the specie movement. The reference for North's estimates is North, Douglass C., “The United States Balance of Payments, 1790-1860,” Trends in the American Economy in the Nineteenth Century (“Studies in Income and Wealth,” Vol. XXIV) (Princeton: Princeton University Press, 1960), Table B-l, Column 6, p. 605Google Scholar.

³⁵ , Macesich, “Sources,” p. 418.Google Scholar These estimates do not sum to unity because of interaction and error in rounding.

³⁶ In his famous Studies in the Theory of International Trade (London: George Allen and Unwin, 1955)Google Scholar, Jacob Viner predicted that under such circumstances regularity could be found in the movement of the expansion ratio.

If there is no central bank or its equivalent, and if there are a large number of genuinely independent banks with power to issue bank money, whether in the form of demand deposits or of notes, and with their specie reserves left completely or substantially free from statutory regulation, the specie-movement phase of the inter-national mechanism can still be regarded as automatic if the average specie reserve ratio of the banking system as a whole is at any moment determined by the aggregate effect of the autonomous decisions of a large number of individuals or firms. The ratio under such circumstances of the total amount of means of payment to the amount of specie will be a variable one, but there will be some elements of regularity in this variability, discoverable by historical investigation if not by a priori cogitation alone, (p. 392)

As Viner went on to point out, such conditions are rarely met and Viner himself offered no examples to support his remarks. The analysis in this article does suggest, however, that the United States provides such an example during the period in question.

³⁷ Interestingly, after the United States returned to the gold standard in the second half of the century, a remarkably stable relationship existed between, international specie flows and die American money supply, rendering the particular specie-flow model developed in this article inapplicable to this latter period. See , Williamson, American Growth, pp. 175–83Google Scholar.

³⁸ All of the sign correlations presented in this article with the exception of the inverse relationship between the expansion ratio and the direction of the preceding years specie flow from 1835 to 1849 and the similar relationship between changes in the specie flows and the expansion ratio from 1850 to 1863 are significantly different from zero at the 0.01 level.

³⁹ A strong case can be made that in more than one of the panics during the period, sharp specie outflows induced by discount rate policies of the Bank of England were among the adverse events which may have helped initiate the collapse.

⁴⁰ For data on the wide dispersion of ratios of circulation to specie (which we may take for this purpose as proxies for the expansion ratios) between states and localities in the 1820's and 1830's, see Fenstermaker's, VanThe Development of American Commercial Banking: 1782-1837 (Bureau of Economics and Business Be-search, Kent State University, 1965).Google Scholar Evidence for the whole period, but particularly for the 1850's, is found in Walker's, AmasaThe Nature and Use of Money and Mixed Currency (Boston: Crosby, Nichols and Company, 1857).Google Scholar In 1855 the proportion of specie to circulation varied from 0.88 to 1 in Louisiana to 41.85 to 1 in Mississippi. For regions the ratio varied from 1.76 to 1 in the Southwest to 7.97 to 1 for the Eastern states. Evidence of considerable intrastate variation is also presented.

⁴¹ The “relatively high” postulate of the hypothesis may not have held for the 1837 panic, for according to Macesich's figures (which are based on different estimates of bank specie reserves) this rapid rise was merely a movement back toward the high level of 1834. See , Macesich, “Sources,” p. 430Google Scholar.

⁴² While the chief support offered in this article for my hypothesis has come from analysis in retrospect, the importance of international specie flows and fluctuations in the expansion ratio upon the domestic money situation did not pass unnoticed by contemporary observers of the period. Professor Amasa Walker, for instance, in his Science of Wealth stressed one of the main tenets of the hypothesis put forth in this article:

The cause which limits the expansion [of a mixed currency] and finally produces contraction, is the liability of the notes to be presented for payment in money [specie].

The occasion for this cause may be almost anything—a political convulsion, an adverse balance of trade, a failure of some large trading or banking company, or an unaccountable mood of the popular mind.

And to Walker the “most common and sensible” of these causes was “an adverse balance of trade.” (Walker, Amasa, The Science of Wealth: A Manual of Political Economy (Boston: Little, Brown and Co., 1866), pp. 157–58Google Scholar).

₄₃ For treatments of countries as financial intermediaries under the present inter-national monetary system see Salant, Walter S., “Capital Markets and the Balance of Payments of a Financial Center,” in Machlup, Fritz, ed., Maintaining and Restoring Balance in International Payments (Princeton: Princeton University Press, 1966), pp. 177–96Google Scholar, and the work by Kenen, Kindleberger, and Depres cited therein.

⁴⁴ For a theoretical analysis of the present international monetary system along somewhat similar lines see Kenen, Peter B., “International Liquidity and the Balance of Payments of a Reserve-Currency Country,” Quarterly Journal of Economics LXXFV (11 1960), 572–86CrossRefGoogle Scholar.