AN OLD DEBATE AND A NEW FRAMEWORK

After President Jackson's veto and his October 1833 executive order forbidding further deposits of Treasury receipts in the Second Bank, money supply growth accelerated. From 1834 to 1836 the money supply grew at an average annual rate of 30 percent, compared to 2.7 percent between 1831 and 1834. Historians traditionally attributed the acceleration in money growth to the removal of the federal bank, which had maintained “systematic pressure on the (state-chartered) banks to redeem their notes.” Now, instead, each state bank could “count the notes of other banks as reserves and expand its loans accordingly.”⁴

In addition, the removal of federal deposits from the Second Bank to state banks purportedly fuelled the boom in public land sales, as state-chartered banks lent federal deposits to land buyers, and as the land offices deposited receipts from land sales in the state banks.

This traditional view, resting largely on anecdotal evidence, was put to the empirical test by New economic historians and found wanting. If the earlier historians were correct, the state banks' specie reserve ratios would necessarily have fallen after 1834, and frontier banks would have had the lowest reserve ratios. In fact they did not.

Using a standard Friedman-Schwartz framework of money supply accounting, Temin constructed estimates of the national stock of specie, the money supply, the reserve ratio, and the proportion of the money supply held as specie for 1820–1858 and found that “the reserve ratio of the American banking system as a whole did not fall below its 1831 level throughout the 1830s.”⁵

Instead, the money supply grew because large influxes of foreign capital increased the nation's stock of specie. Although he did not fully extend his money supply accounting model down to the regional level, Temin presented evidence on the reserve ratios of state banks by region in 1834–1837, the years of most rapid money growth. He concluded that “the reserve ratios of the different regions stayed remarkably constant” in 1834–1837, and he noted that the lowest reserve ratios were not in the frontier areas, where the “freed” state banks were supposed to have been the most reckless in their expansion.⁶ The implication is that the growth of money and credit was not disproportionately high on the frontier.

But it was. Appendix 1 shows the annual balance sheet of the U.S. state-chartered commercial banking system in 1830–1837, and Tables 1–3 are based on this Appendix. (The assets and liabilities of the Bank of the United States of Pennsylvania, formerly the Second Bank of the United States, are not included in Appendix 1 or the tables derived from it.⁷

) Table 1 shows the regional pattern of growth in state-chartered commercial bank assets (promissory notes, discounted bills, and securities) and monetary liabilities in 1830–1836. Between 1830 and 1836 two regions grew disproportionately, as indicated by the fact that their share of the change in each of these aggregates was larger than their share of the total in 1830: the Southwest and the Northwest.⁸

Because the traditional hypothesis about disproportionate expansion in the West was concerned with the lending behavior of state banks, and because excessive lending was at the root of the financial crisis that followed the financial boom, this article focuses on the asset side of the banking system's balance sheet. In a Friedman-Schwartz framework, the determinants of bank credit growth are the same as the determinants of growth in bank money: the reserve ratio and bank reserves. This framework, however, was developed for quite a different world, one in which banks acquired most of their money through deposits. It needs to be modified for application to the 1830s United States in light of the “investment club” nature of commercial banks in this period.⁹

Antebellum banks relied relatively heavily on equity as a source of funds for asset acquisition. In 1830–1836 the capital-asset ratio of the state-chartered commercial banking system averaged 48 percent; the contemporary FDIC-insured U.S. commercial banking system's ratio is 10.4 percent.¹⁰

The use of equity as a source of funds for bank loans is illustrated in a simple model of the state-chartered commercial banking system, a model based on accounting identities, in order to understand what happened on these banks' balance sheets over time and to make some useful inferences. The balance sheet of the commercial banking system as a whole consisted of four primary elements, two types of assets: reserves (R) and loans (L); and two types of liabilities: monetary liabilities (D) and shareholders' equity (K).

Bank reserves, broadly defined here, include specie, specie funds, notes of other banks, and deposits held with other banks (“due from” other banks). The banks held these reserves as a proportion (r) of monetary liabilities. Loans include promissory notes, discounted bills, and securities. Monetary liabilities include deposits, notes (including notes held by other banks), and “due to” other banks. K is the paid-in capital stock.

Because total assets must equal total liabilities, it follows that

The sources of growth in bank credit can be separately identified by considering the differential of equation 2 with respect to r, R, and K

This expression identifies three sources of growth in bank credit: a reduction in the reserve ratio, which has a larger effect the larger the level of bank reserves and the smaller the value of the reserve ratio; an increase in bank reserves, which has a larger effect the smaller the value of the reserve ratio; and an increase in bank equity. The first two terms refer to loans created through fractional-reserve banking methods captured in a Friedman-Schwartz framework. The third captures equity-driven bank lending.

Applying the data on the state-chartered commercial banking system's balance sheet in Appendix 1 to equation 3, we find that this equation explains 88 percent of the observed $337 million change in aggregate state bank credit over 1830–1836, or $295 million (see Table 2). Of this amount, only $2 million, or under 1 percent, was due to the very small reduction in the state-chartered banking system's reserve ratio, and $129 million, or 44 percent, to the change in bank reserves (broadly defined). Fully $166 million, or 56 percent, was due to the change in bank equity. To be sure, not all of the equity was actually paid in; some portion of the loans “funded” by equity were loans to shareholders, who could pay for a fraction of their equity position in specie and repay the balance on an installment basis.¹¹

Nevertheless, even if half the loans funded by new equity were loans for buying bank stock itself, equity still funded $80 million of the growth in “true” financial intermediation performed by the state-chartered commercial banking system. In short, a model based on the Friedman-Schwartz framework fails to account for, conservatively, $80 million of equity-driven growth in bank credit to the nonfinancial sector.

In the 1830s, equity investments in state-chartered commercial banks increased dramatically, from $85 million in 1830 to $250 million in 1836 (not including the $35 million of equity converted from the Second Bank of the United States to the Pennsylvania-chartered Bank of the United States of Pennsylvania), or 32 percent per year on an average annual basis, as compared to 10 percent in the 1820s. Loan growth was even faster than equity growth, as state banks sought higher returns by increasing their leverage, thereby exposing note holders, depositors, and shareholders to greater risk.¹²

Each dollar of bank equity supported $2.32 of assets in 1836, compared to $1.85 in 1830. Although banks were unwilling to lower their reserve ratios and so increase their exposure to the risk of being unable to meet demands for deposit withdrawals and note redemptions, they were willing to increase their exposure to the risk that loan losses would render them insolvent (see Tables 3A and 3B).

The rising leverage of the banking system did not deter new equity investment in banks. Sixty million of the $166 million in new equity investment in state-chartered banks took place in 1835, at the height of the land boom and at the height of banking leverage. In that year alone, 90 new banks were organized, or one-fourth of all new banks organized between 1830 and 1836. Furthermore, almost half of new bank equity raised without state assistance was invested in the most highly levered regional banking systems, the Middle Atlantic and the Northwest.¹³

Instead of being deterred by the greater risk involved in greater leverage, private equity investors were drawn to the higher potential returns of these banks.¹⁴

Although the boom in state banking was not reflected in falling reserve ratios, it was still causally related to the Bank Veto, which brought about three destabilizing monetary effects—effects felt to varying degrees in different regions. First, the removal of the Second Bank opened a large gap between the supply of, and demand for, commercial banking services, creating conditions conducive to a large movement of financial capital into state banking—an entry effect. Second, the veto led to a qualitative change in the kinds of loans available from the commercial banking system, as the dual (federal-state) system gave way to an entirely state-chartered banking system—a credit policy effect. Finally, the removal of federal deposits from Second Bank branches to state-chartered banks eliminated a stabilizing effect that would have developed as a result of the buildup of Treasury deposits during the boom—a reserve retention effect.

THE ENTRY EFFECT

The state banking system grew because state governments decided to charter more banks, and because wealth-holders decided to invest more financial capital into commercial bank equity. Both responses were largest in regions where the Second Bank had dominated commercial banking and where bank equity gave investors access to capital gains on frontier lands.

At the time of the Bank Veto, the United States had a dual commercial banking system, with state-chartered banks and one large, federally chartered, nationally branching bank, the Second Bank. Nationwide in 1830, state-chartered banks provided 73 percent of the dollar value of all commercial bank loans; the Second Bank, 27 percent (Table 4). By that year, however, the mix of state and federal banking differed dramatically across regions; state-chartered banks provided 94 percent of all commercial bank loans in New England but only 33 percent in the Southwest and 22 percent in the Northwest, where the Second Bank was most dominant.¹⁵

By 1830 too, the quantity of banking services supplied by the dual banking system varied widely in different regions, in terms both of absolute levels (Table 5A) and of relative regional incomes, which indicate relative levels of demand for credit (Table 5B).¹⁶

In Table 5B, state and total bank loans per capita and real product per capita are indexed to the national average (columns 1–3). The ratio of the index of total loans per capita to the index of product per capita (column 5) shows whether a region enjoyed a larger (ratio greater than one) or smaller (ratio less than one) supply of banking services in relation to demand, using the national average as the standard. In 1830 the Northwest and the Southeast were underbanked, by this measure, and the Southwest slightly so. Column 4 similarly measures each region's relative level of banking services, but from state banks only.

A comparison of the last two columns shows the effect of the closing of Second Bank branches on the supply-demand balance in each region. By these measures, the Second Bank branch closings had no adverse impact on the relative standing of New England, the Middle Atlantic, or the Southeast, but the Southwest and the Northwest fell far behind the national average, even after adjusting for the lower incomes in these regions. In the 1830s the Southwest started out with a level of banking services not far from the national average, but the closing of the Second Bank left it deficient. In the same decade, the Northwest started out in an underbanked condition that became aggravated by the closing of the Second Bank branches there. The absolute level of loans per capita fell by two-thirds in the Southwest and by 78 percent in the Northwest, much more than in any other region.

Another indication of the impact of the closing of Second Bank branches on different regions is relative access to a banking office, measured by the number of residents per banking office (i.e., bank density). By 1830, state banks dominated the commercial banking system in terms of the population's access to banking offices; there were 43,000 residents per state banking office, 469,000 per Second Bank branch (Table 6A). Again, the relative density of state and federal banking offices varied across regions. In New England, access to a state banking office was 25 times as great as access to a Second Bank branch; in the Northwest only six times as great, and in the Southwest it was actually slightly lower than access to a Second Bank branch.

Overall access to a banking office was lower on the frontier than in coastal areas, but we expect regions with lower population density to have lower density of banking offices, especially when transportation and communications costs were high. To put these differences in perspective, we must compare bank density to population density. In Table 6B, state and total bank density by region are indexed to the national average, as is population density. The ratio of the total bank density index to the population density index (column 5) reflects whether a region enjoyed greater (ratio greater than one) or less (ratio less than one) relative access to banking facilities than would be expected based on its population density alone. All regions other than New England had lower relative bank density than their relative population densities would have predicted.

The difference between column 5 and column 4 indicates the effect of the removal of the federal bank on each region's relative access to banking facilities. New England's relative access improved, whereas the positions of the Middle Atlantic, Southeast, and Midwest deteriorated slightly or stayed about the same. The only region where the closing of the Second Bank had a significantly adverse impact on access to a banking office was the Southwest, where, consequently, the chartering rate grew the most.¹⁷

Between 1830 and 1836 all states granted new bank charters at increasing rates, as measured by the ratio of population increase to number of banks chartered (see Table 7), and by this measure the chartering rate at the national level more than doubled. In the 1830s, state governments had both offensive and defensive reasons to charter banks. Economic competition among states drove them to undertake internal improvement projects and to charter banks, both essential to an infrastructure for developing market economies. State legislatures were also afraid that “if they did not create new banks to replace branches of the Second Bank of the United States, other state banks would flood their particular state with notes over which they would have no control.”¹⁸

Bank chartering increased at varying rates across regions and states. The largest increases were in the Southwest, the Northwest, and the Southeast, regions where the closing of Second Bank branches did the most to restrict access to banking services.¹⁹

At the state level, chartering activity depended on both the willingness of state legislatures to provide new charters and the number of petitions from bank promoters seeking new charters. Without systematic data on the number of bank petitions submitted to state legislatures in the 1820s, we cannot know whether chartering increased in the 1830s because of increased demand for new charters, increased interest in granting new charters, or a combination of the two. In any event, either the supply of new bank charters or the demand for them—or both—had been relatively suppressed in these areas when the Second Bank operated. The closing of those branches brought forth more petitions for new banks and caused a shift in the political balance of power in favor of those advocating more liberal chartering policies at the state level.²⁰

The closing of Second Bank branches could have spurred state governments to charter more banks, but the banking system would grow only if owners of financial capital subscribed to the authorized banking capital. At the regional level, commercial bank equity grew disproportionately only in the two frontier public-land regions, whether or not the measurement includes equity raised with state assistance (Table 8). Equity investments in these western banks were an indirect way for individual investors and state governments to invest in western lands.²¹

Once in frontier banks, new equity investment helped significantly to drive the rapid credit expansion there, accounting for almost three fourths of the explained credit growth in the Southwest and about half in the Northwest (see Table 2).²²

In two states in slower-growth regions, New York and South Carolina, state bank equity also grew disproportionately, and each of these states had something in common with the two faster-growth banking regions. New York was not hurt by the closing of Second Bank branches (see Tables 5B and 6B), but it did find itself, by the late 1820s, in an underbanked condition, as measured both by bank loans per capita and by bank density. After a decade of reluctance to charter additional banks, the New York Assembly resumed bank chartering in 1829, when the shortage of banking capital had become a pressing political issue. Further, although New York was not a public-lands state, the value of large tracts of land owned by private land companies in “frontier” areas of central and western New York increased after the completion of the Erie Canal in 1825.²³

In South Carolina the level of bank loans per capita was disproportionately low in 1830, and it became significantly worse after the closing of the Second Bank branch in Charleston (see Table 5B). Bank density was also very low in proportion to the state's population density, both before and after the exit of the Second Bank. In this southeastern seaboard center of cotton marketing and distribution, the Charleston branch had provided substantial discounting facilities (exceeding $3 million at the end of 1831).²⁴

The closing of this branch created a particularly strong inducement to state chartering and to financial capital investment.

Had the Second Bank remained in place, both parts of the dual banking system would have grown. To estimate how much of the observed growth in state bank credit in frontier areas can be attributed to the removal of the Second Bank, it is useful to conduct a counterfactual exercise, calculating what this growth would have been if the Second Bank had continued to exist (see Table 9). Applying this alternative growth rate generates the theoretical level of state bank loans per capita in the Southwest and the Northwest in 1836 (row 1). Adding what would then have been Second Bank loans per capita (row 2)—assumed to be equal, in 1836, to the observed level of 1830—and multiplying the per capita bank loan estimate by the (estimated) 1836 population (row 3) produces the level of total commercial bank loans (row 4). The difference between the observed level of commercial bank loans in 1836 (all from state banks) and what would have been the level in a dual-banking system can be attributed to the removal of the Second Bank.

This counterfactual method contains a bias against finding support for the hypothesis that the removal of the Second Bank had a significant effect: it assumes that there would have been as much population movement into frontier areas even if there had been no change in banking institutions, and this assumption consequently amplifies the hypothetical loan levels. The method also incorporates the assumption that all growth in credit per capita would have come from the state banks, not from the Second Bank—an assumption that can be justified on the grounds that growth in Second Bank loans would have been limited by the Bank's constant level of capitalization, set by Congress.²⁵

As a result, in the counterfactual scenarios the state-chartered segment of the dual banking system grows proportionately more than the federal segment. In all but one of these scenarios, the state-chartered banking system's market share rises to levels approximating those observed in the older regions in 1830 (see Table 4).

Clearly, the outcome of this counterfactual exercise rests critically on the growth rate assumed for state bank loans. Thus, two different scenarios may provide a more reliable conclusion than just one. One way to project growth in state banking on the frontier is to assume that state banking in the Southwest would have grown as fast as state banking in its more mature counterpart, the Southeast, grew in the dual banking environment of the 1820s, and that state banking in the Northwest would have followed the growth pattern of the Middle Atlantic states in the 1820s (from Table 4). Under this “older-state” assumption, the removal of the Second Bank would explain $79 million of the $122 million in commercial bank loans observed in the Southwest in 1836 and $12 million of the $26 million observed in the Northwest.

This figure of $79 million for the Southwest may be misleadingly large, however, because in the 1820s per capita loans grew much more slowly in the Southeast than in the Middle Atlantic (3.8 percent per year versus 8.3 percent per year). To modify this “older-state” scenario, therefore, one may assume that the Southwest would have been more likely to grow as Georgia did (25 percent average annual growth in per capita loans in the 1820s) than as the Southeast region grew as a whole. This alternative approach, reported in Table 9, attributes $63 million of $122 million to the removal of the Second Bank. A “fast-growth” approach projected the observed 60 percent average annual growth of 1830–1832 in the Southwest to both frontier regions. Even in this fast-growth scenario, a significant gap remains between the observed credit growth and credit growth in the counterfactual world ($55 million in the Southwest, 45 percent of the observed, and $7 million in the Northwest, 26 percent of the observed).²⁶

In sum, the counterfactual exercise shows that between 45 percent and 52 percent of the observed growth in bank credit in the Southwest and between 26 percent and 46 percent of the observed growth in credit in the Northwest can be considered excessive credit growth caused by the removal of the Second Bank.

THE CREDIT POLICY EFFECT

The closing of the Second Bank's branch offices also led to a qualitative change in commercial banking practices, as the dual banking system gave way to the state banking system. Although the types of commercial banks chartered by different states varied somewhat from each other, the credit policies of the state-chartered banks were generally more liberal than those of the Second Bank. Whereas the Second Bank had institutional obligations to the federal government, which chartered it, state-chartered banks had institutional obligations, explicit or implicit, to state governments and to stockholder-directors who contributed substantially to the banks' capitalization. These relationships drew state banks into areas of finance the Second Bank had avoided, notably land development and state public works projects.

Under its charter, the Second Bank had to provide free fiscal services to the national government, which had receipts and payments spanning a broad geographic area, both nationally and internationally. The Second Bank pursued a business strategy of creating synergies between its core lending business and its charter obligations. This strategy led the bank to specialize in trade finance. Thus, for example, in 1830, when the Second Bank provided 27 percent of all credit extended by the commercial banking system, it held over 60 percent of its earning assets in bills of exchange—short-term credit instruments that financed interregional and international trade.²⁷

As the federal government's fiscal agent, the Second Bank also facilitated federal borrowing and debt repayment; however, its charter prohibited it from making a loan larger than $50,000 to a state government.

After being appointed its president in 1823, Nicholas Biddle sought to improve the quality of the bank's portfolio. To its branches he issued a prohibition on real estate lending, arguing that the bank had “so large a part of the capital of the Bank locked up in real estate,” the result of the bank's having taken ownership of real estate in repayment of debts incurred during the credit expansion of 1818–1819, particularly at the western branches.²⁸

Biddle also sought to maintain arms-length relationships between borrowers and the bank branches' boards of directors.²⁹

The new banks that were chartered on the southwestern frontier were various combinations of “plantation banks,” internal improvement banks, and commercial banks.³⁰

“Plantation” or “property banks” conducted a banking business fundamentally different from that conducted by the Second Bank. In these banks, landowners presented mortgage deeds on their property as collateral for the new issues of state debt and became shareholders, entitled to borrow up to some fraction (typically half) of the value of their mortgaged property. The connection between the capitalization of these banks and their lending to shareholders for land acquisition and development was quite explicit. The bank's profits would be used to pay interest on its debts and to accumulate a fund for redemption of its bonds.³¹ A recent study concluded that at least “half to two-thirds of the portfolios of plantation banks … were long-term assets, mostly mortgages.”³²

The capitalization of these banks tended to be large compared to banks in other regions, and southern states borrowed heavily to assist in this capitalization.³³

Between 1830 and 1836, Louisiana, Alabama, and Mississippi, together with Florida (arguably also a “frontier” state although categorized in Appendix 1 with the Southeast), issued almost all ($28 million) of the debt issued for banking purposes, and state borrowing constituted 40 percent of all equity raised for banks in the Southwest.³⁴ Southern “bank bonds” were a vehicle for foreign and domestic wealth-holders to invest their savings indirectly in southern lands.

Commercial banking and real estate lending also engaged the new banks chartered by northern state governments.³⁵

Like the Second Bank, northern state banks provided trade finance to move agricultural products to market. Comprehensive balance sheet evidence on the composition of bank loan portfolios is not available, but the evidence for Ohio state banks shows that even the most purely commercial state-chartered banks held significantly less than 60 percent (the Second Bank benchmark) of their earning assets as bills of exchange, and that they held proportionately more of their earning assets as discounts or promissory notes, whose terms could exceed those of bills of exchange.³⁶ In 1839 the three large commercial banks in Cincinnati held between 39 percent and 54 percent of their portfolios in bills of exchange; smaller commercial banks in other towns and cities held between 7 percent and 48 percent. The State Bank of Indiana lent a large part of its funds, including its sinking fund for retiring the state bonds issued to capitalize the bank, to purchasers and developers of real estate. Between 1837 and 1842, the State Bank of Illinois held only 28 percent of its earning assets as bills of exchange and the remainder in longer-term loans, including real estate loans.³⁷

In his study of state banks in the Old Northwest, Carter Golembe concludes that real estate loans were common in Ohio, “even in towns that were able to provide banks with a great deal of business of a commercial nature,” and that the banks of Illinois, Wisconsin, and Michigan were “most heavily involved in real estate transactions, particularly of a speculative nature.” The Commercial Bank of Lake Erie, a Cleveland bank that was revived in 1831, “was in the mainstream of western banking institutions: issuing credit not only for mercantile transactions but also for real-estate purchases, both in Cleveland and in undeveloped lands sold by the state and federal governments in the mid-1830s.”³⁸

Real estate lending by western banks was variously conservative and speculative. The State Bank of Indiana and the Ohio Life Insurance and Trust Company followed a relatively conservative policy of restricting agricultural and urban real estate lending to improved lands, with low loan-to-value ratios, and in established urban areas. At the other extreme, the Bank of Illinois at Shawneetown and the Franklin Bank of Cincinnati directly helped finance the development of new towns at Massac, Illinois, and Jeffersonville, Indiana. A large shareholder of the State Bank of Illinois borrowed from that bank to speculate in town lots in Alton, Illinois. The Bank of Michigan financed investments in town sites in Detroit and western Michigan.³⁹

This real estate lending in the Northwest exposed banks to more liquidity risk than in the Southwest, because the Northwest banks were relatively more reliant on monetary liabilities to fund their assets and relatively less reliant on equity (see Table 3A).

An examination of how new Midwestern banks were organized and capitalized provides additional insight into the connection between state banking on the frontier and real estate investment. Between 1832 and 1834, Ohio state banking capital increased by $7.4 million, of which $5.3 million was issued by five of the largest institutions. Sixty percent of the assets of these five were owned by eastern investors, who formed group partnerships with the western elite to organize, capitalize, and direct these banks. In The Investment Frontier, John Haeger provides a detailed historical record of the work of one such investor group, led by Charles Butler, Lot Clark and Thomas Olcott, upstate New Yorkers who invested in both bank stock and western lands.⁴⁰

This group raised three fourths of the capital, all “foreign” owned (i.e., owned by Easterners), for the largest institution organized in the Northwest during 1830–1836, the Ohio Life Insurance and Trust Company.⁴¹ The eastern investors recruited two Ohio men, Elisha Whittlesey and Micajah Williams, to help secure the charter and garner the support of the political leadership. These Ohioans then followed the lead of their eastern partners by organizing and capitalizing other banks, such as the State Bank of Illinois, and investing in farmland and town sites farther west on the frontier.

The Dwight family of New York and Massachusetts played a similar role in the capitalization of the Bank of Michigan in Detroit and the Commercial Bank of Lake Erie in Cleveland. The cashier of the Bank of Michigan from 1825 to 1836 was Charles Trowbridge, who began his career as the Detroit agent and partner for Arthur Bronson, an eastern land investor with large western holdings, who provided finance for the ventures of the Butler-Clark-Olcott group as well. Like Whittlesey and Williams, Trowbridge later became an active land developer and promoter of town sites in western Michigan.⁴²

Frontier bank lending to the property sector amplified the land boom, either directly or indirectly. State-chartered commercial banks provided a significant portion of the financing of land purchases and land development on the southern frontier.⁴³

Between 1830 and 1836, the change in equity in commercial banks in Louisiana, Alabama, Arkansas, and Mississippi of $51 million was large enough to finance much of the $21 million in total sales of public lands in those states. Property banks in the South fuelled transactions volume in the primary land market directly, through their loans to their planter-owners for large-scale land acquisition and improvements.⁴⁴

On the northern frontier, the increase in bank equity ($14 million) between 1830 and 1836 was small in comparison with $34 million of total sales of public lands in Ohio, Illinois, Indiana, Michigan, and Wisconsin.⁴⁵

Seemingly, state banks did little direct lending to settlers buying land from the federal government; instead, they financed the public land purchases of western land investors, improvements to town sites, and the construction of commercial structures such as hotels, mills, warehouses, and wharves. Even if banks on the northern frontier financed only a small share of the activity in the primary land market, their urban real estate loans indirectly amplified the volume of land sales by stimulating settlers' demand for agricultural land around growing urban areas.⁴⁶

As farmers, planters and land developers looked to state-chartered banks for longer-term loans in 1830–1836, so did state governments, both where commercial banking expanded relatively rapidly (the Northwest) and where it grew relatively slowly (Pennsylvania, Maryland, and Virginia). By 1839 the State Bank of Illinois held $2.5 million in Illinois internal improvement bonds, not because it was required to by law but “to retain the good will of the legislature.” Similarly, the State Bank of Indiana made what started out to be short-term advances to the state's canal fund, eventually rolling them over into de facto long-term loans despite the discomfort of the bank's officers. State-chartered banks in Ohio were an important source of demand for the state bonds issued after 1836 to finance the expansion of the state's system of canals.⁴⁷

Again, bank lending had an indirect effect on the public land boom: frontier state spending on canals, partially financed by state banks, increased the expected return on western lands, further stimulating the demand for them.

Longer-term finance for land development and internal improvements would not have been so forthcoming from the old Second Bank, with its federal government orientation and “real bills” approach to portfolio management. Nevertheless, the removal of the Second Bank cannot account for all of the observed increase in the supply of long-term lending, because, had the Bank survived, the dual banking system would have continued to expand in frontier states, and with it the supply of longer term loans would have also grown. A counterfactual scenario based on the exercise in the previous section demonstrates the effect of the Second Bank's removal on the supply of longer-term loans. In this scenario, long-term lending is lower for two reasons: the total volume of loans is smaller, because the hypothetical survival of the Second Bank slows the pace of state bank expansion; and the share of long-term lending in total lending is smaller.

The exercise compares the actual increase in longer-term loans to a hypothetical estimated increase. Absent any systematic evidence on the composition of state banks' portfolios of earning assets, the actual increase must also be an estimate. I have based this estimate on the assumption that state banks held 50 percent of their loans as long-term loans in 1830 and 65 percent in 1836, and that western branches of the Second Bank held 40 percent of their loans as longer-term loans. The counterfactual scenario of a dual-banking world in 1836 is based on the assumption that state banks would have continued to hold 50 percent of their loans as long-term loans, the Second Bank 40 percent. Applying these loan-mix assumptions to the observed and counterfactual data on per capita loans and population from Table 8, using the “older-state” scenario for both frontier regions, leads to the conclusion that the removal of the Second Bank resulted in an additional $52 million of longer-term credit in the Southwest and an additional $11 million in the Northwest (Table 10).⁴⁸

These findings, focusing on longer-term lending as opposed to total lending as in the previous section, better reflect how much the removal of the Second Bank fed the land boom, because settlers, developers, and investors needed longer-term loans to acquire and improve property.

THE RESERVE-RETENTION EFFECT

In his history of the public land-office business, Malcolm Rohrbough presented the traditional historian's view that the Bank Veto and the removal of the federal deposits led state banks to expand loans “at a dizzy rate, while often filling their vaults with public monies deposited by the local (land office) receiver…(creating) a cycle of loan, purchase, and deposit that continued unchecked.”⁴⁹

Temin argued that this account could not have been accurate, because it implies that state banks' specie reserve ratios fell, whereas in fact they did not. However, the Treasury's monetary behavior did contribute to the disproportionate growth of reserves in frontier banks following the veto.

Frontier banks' reserves grew, first, because the reserves of the banking system as a whole increased, and, second, because these banks' share of total reserves rose. According to Temin's data, reserves of the commercial banking system first increased in 1830, when banks' share of a basically constant national specie stock rose. Bank reserves increased again after 1833 even though the banking system's share of the national specie stock fell, thanks to large increases in the national stock of specie as the national economy ran surpluses on its overall balance of payments.⁵⁰

Over this period, the national specie stock grew from $32 million to $73 million. Of this $41 million increase, $29 million accumulated outside banks and $12 million as reserves in the commercial banking system.⁵¹

At the outset of the 1833–1836 surge in commercial bank reserves, President Jackson issued an Executive Order mandating that federal receipts be deposited in designated state-chartered depository institutions instead of in the Second Bank. Between the fourth quarter of 1833 and the first quarter of 1834, the Second Bank's share of all federal deposits fell from 97 percent to 30 percent, then fell more gradually over the next two years, until by July 1836 there were no federal deposits at the Second Bank. As a result of the removal of Treasury deposits and the Second Bank's winding up its affairs as a federally chartered bank, the Second Bank's specie reserves fell by $8 million while those of state-chartered banks (other than the Bank of the United States of Pennsylvania) rose $20 million between 1830 and 1836.⁵²

During this same period federal deposits increased dramatically, as federal receipts from public land sales and import duties increased and as the newly debt-free federal government ran large budget surpluses. Between the beginning of 1833 and the end of 1836, $40.9 million in new federal deposits accumulated at state-chartered depositories; over the same period, $12.6 million of pre-existing federal deposits were drawn down at the Second Bank, for a net increase in federal deposits in the commercial banking system of $28.3 million (see Table 11). The large federal surpluses were left on deposit where new receipts were generated. As a consequence federal deposits accumulated disproportionately in federal depository institutions in the Middle Atlantic, Southwest, and Northwest. Of the $40.9 million in federal deposits that accumulated at state-chartered depositories, banks in the Middle Atlantic received 43 percent, banks in the Southwest 19 percent, and banks in the Northwest 22 percent (Table 11).

The Treasury's monetary behavior is one of many factors behind the disproportionate growth in frontier banks' reserves, including net external investment in western corporations (such as banks), in western land and land development, and in long-term securities issued by western state governments.⁵³

The two frontier regions together received 42 percent of the increase in state banks' specie reserves, but in 1830 their total share of the banking system's specie reserves was only 6 percent. State banks in the frontier areas received 32 percent of the increase in reserves, broadly defined to include interbank liabilities, and held 8 percent in 1830.⁵⁴ This growth in specie reserves explains why money creation by frontier banks was disproportionately high despite the fact that their reserve ratios were also high and did not fall significantly.

The disproportionate accumulation of federal deposits in frontier banks has been noted by others, but the point not already made in the literature is that under the dual commercial-banking system, even if the Treasury had kept the land office receipts in the frontier regions, the growth in federal deposits would have redistributed reserves from state-chartered banks to the western branches of the Second Bank.⁵⁵

This redistribution of reserves would have increased the share of reserves held by the segment of the commercial banking system with the more conservative credit policy. Instead, state banks in the receiving regions retained reserves, as federal deposits accumulated at local state bank depositories. This effect was larger on the northern frontier, where banks relied more on monetary liabilities to fund asset acquisition and where, by 1836, federal deposits constituted 37 percent of the stock of bank money, compared with 5 percent in the Southwest.

Until 1836 federal deposits were relatively concentrated in a few institutions in both frontier regions. In the Southwest three banks held 70 percent of all federal deposits at the end of 1836, when deposits were at their highest point: the Mobile branch of the Bank of the State of Alabama, the Union Bank of Louisiana, and the Commercial Bank of New Orleans. At the same time in the Northwest, the Bank of Michigan, the Franklin Bank in Cincinnati, and the Commercial Bank of Cincinnati held 43 percent of federal deposits in the region.⁵⁶

Balance sheet evidence for a group of nine frontier depositories, five in the north and four in the south, shows that federal deposits were an important source of funds for these frontier depositories.⁵⁷ Federal deposits made up 13 percent of all liabilities for the southwestern banks and 29 percent for the northwestern banks. For three of the northwestern depositories, federal deposits were as large as, or larger than, the banks' paid-in capital stock.⁵⁸

The growth in federal deposits and their concentration in a few institutions fuelled the land boom, however unintended that effect may have been. According to Golembe, individuals involved in organizing new banks actively sought federal deposits, which “seem to have been regarded by the receiving banks more as additions to their capital than as obligations which they might have to meet unexpectedly.”⁵⁹

Where they took this approach, western bankers were more likely to use federal deposits to fund longer-term loans for property and infrastructure improvements than they would otherwise have been. Further, the concentration of federal deposits increased the average size of each western depository's federal deposit, a condition that made it possible to extend larger land development loans and so increase the risk. Western bankers did not confuse federal deposits with reserves, as is suggested by some versions of the “traditional” account; but, to the extent that they confused federal deposits with capital, the growth in federal deposits fuelled the growth of longer-term land development loans in a way that a comparable increase in other state bank monetary liabilities (notes, deposits of individuals), or in federal deposits at the Second Bank, would not have.

In sum, after the veto the Treasury's management of federal government deposits reinforced the frontier land boom and the pattern of rapid money and credit growth on the frontier, by redistributing bank reserves to the same regional state banking systems that attracted much of the new bank equity. Had the Second Bank continued to serve as the federal government's fiscal agent, the same Treasury behavior would have redistributed bank reserves to the frontier branches of a financial institution that avoided loans for real estate development and state infrastructure.