The Question of Note Issue in American Free Banks

Howard Bodenhorn, Lafayette College, and

Michael J. Haupert, University of Wisconsin - LaCrosse

In 1912 Spurgeon Bell uncovered a paradox that has vexed banking and monetary historians ever since. He found that national banks failed to expand their note issues despite its apparent profitability. Subsequent writers refined Bell's original profit calculations and most have shown that, if anything, Bell's calculations understated the profit to be earned by banks through expanded note issues. Explanations of the paradox abound. Bell (1912) and Goodhart (1965) argued that banks failed to increase their note issues out of fear that circulation privileges would be revoked. James (1976) argued that regional interest rate differentials in the postbellum era made it more profitable for banks in some regions to focus on lending low-cost deposits rather than engage in higher-cost note issue. Champ (1990) and Kuehlwein (1992) focused on term-structure and holding period risks. Bond-secured note issue required banks to hold long-term assets as collateral against which short-term liabilities were issued. If redemptions increased significantly, banks might be forced to sell off bonds to redeem their notes. If such sales were required in a general panic, the bank would absorb a significant capital loss. Cagan and Schwartz (1991) believed that banks simply acted irrationally. Given the profits to be earned through additional note issues, bankers should have bid the price of bonds up to a level where note issue was no longer profitable. Despite all this work, the national bank note paradox remains unsolved; some might say insoluble.

In previous research (Bodenhorn and Haupert 1995a) we pursued the question raised by a cadre of writers examining the National banking era: why did the national banks issue so few notes? We examined this issue with regard to free banks, and found that like their national bank successors, they too passed up profitable note issue opportunities. In our quest to solve this paradox, we have followed the evolution of the literature, and further refined the latest profit calculations. While these new calculations reduce the foregone profits, they do not solve the problem.

In order to get a handle on this issue we now focus our attention on a different aspect of the bank decision-making process: the use of bond-backed note issues as a means of signaling a credible commitment made by the bank in its operations. The use of bonds by banks acts as a signal to the market about the separation of ownership and control over bank assets. This enhances the reputation of an individual bank by signaling a reduced likelihood of "wildcat" behavior by the bank, thus increasing the value of the bank's stock and aiding in keeping the bank's notes in circulation. This work builds on the research of Fama and Jensen (1983) and our earlier examination of bank returns during this era (Bodenhorn and Haupert 1995b).

Bank liabilities during the ante-bellum period consisted primarily of bank notes and deposits. The fact that banks issued demandable debt requires explanation. Mismatches in the maturities of assets and liablilities create the potential for bank runs which can be mitigated only by deposit insurance, the presence of a committed lender of last resort, or a general suspension of convertibility. A potential, and widely adopted, alternative was pure equity financing. That is, banks could simply lend the original capital supplied by shareholders in whatever base money or legal tender existed. This is similar to the "narrow banking" idea addressed by Barth (1991), and sources therein, wherein bankers would be required to invest all their capital in risk-free government debt and then lend only an amount equal to the market value of the bond holdings. Such a situation would protect note holders and depositors from loss so long as there were no significant changes in bond prices. This proposal is itself similar to, but not exactly the same as, both ante-bellum American "free banking" and the system put in place under the National Banking System (see Rockoff (1975), Rolnick and Weber (1983) and sources therein for a full description of American free banking).

By employing equity financing, banks would not face the maturity mismatch inherent to banking and the possibility of bank runs, though not bank failure, would disappear. Such a situation would be roughly analogous to private banking and money lending which has itself a long and venerable tradition in both the United States and Europe.

Bodenhorn and Haupert (1995b) have uncovered an additional conundrum surrounding the issuance of bank notes. They have shown that, as early as the 1840s, it was more profitable for banks to create deposits than to issue bank notes as part and parcel of their credit intermediation services. Despite the greater profitability of deposit creation, ante-bellum banks issued notes in large, but relatively (to deposits) declining amounts. If bank notes were less profitable, it would seem that a rational, profit-maximizing banker would have reduced the number of notes in circulation and increased deposits until the marginal profits of each were equalized. In fact, it was not until bank note issues were heavily taxed in the 1860s that bankers switched over en masse.

This seeming paradox disappears once agency costs are brought to bear. Bank note issues, in fact, solved two agency costs. Note issues aligned shareholder and manager interests; and they fostered outside, independent monitoring that was valuable to shareholders and creditors. Bank note issues in the face of seemingly high costs - potential runs and foregoing profits through deposit banking - then becomes analogous to the modern corporate dividend puzzle. The modern finance literature has long wrestled with the question of why firms pay dividends to shareholders while simultaneously financing investments with outside funding, i.e., issuing bonds. Given the costs of paying dividends and entering the bond market, a lower cost solution would be to retain past earnings and finance internally. Similarly for ante-bellum banks, a lower net cost solution would have been for banks to retain earnings and expand their operations with internally generated funds. Pursuing a policy of pure equity finance may have slowed the short-run growth of many banks; it would not have significantly slowed their long-term growth. Ante-bellum banks paid an average of about 6 or 7% of their paid-up capital in dividends each year and a continued growth of 6% per year can hardly be viewed as slow. Instead, shareholders demanded dividends and forced managers to continually enter the debt market through note issues where effective monitoring of management's policies occurred.

An agency cost explanation for note issue begins from a straightforward proposition: suppose managers are not perfect agents of the shareholders, but pursue their own self interests whenever they can (Easterbrook 1984, p. 652). Since managers are not (necessarily) residual claim holders, there may be a substantial divergence between their interests and those of the shareholders. In such cases, shareholders will find it beneficial to devise schemes which more closely align the managers' interests with their own.

By forcing bank managers to issue notes in providing credit intermediation services, shareholders overcame two potential agency problems: risk aversion by managers and who best to monitor. Investors with diversified portfolios concerned themselves only with the undiversifiable risks inherent in holding a particular equity. Bank managers, however, had a significant portion of their wealth invested in a particular bank, mostly in the form of firm-specific human capital. If the bank did poorly or failed, the value of the invested human capital may have fallen precipitously particularly if shareholders in other banks were aware of the manager's previous employment and its outcome. In the managerial labor market, a manager's previous performance and the bank's success or failure provided valuable information about a banker's managerial talent (Fama 1980, pp. 291-92). Although managers did not face immediate income declines as a result of poor performance (the wage was contracted in advance), future wages depended upon the bank's successes. Bankers, especially those with much firm-specific human capital, had a stake in the firm and may have been overly cautious in guaranteeing its continuation.

Risk averse bankers then would have chosen safer investment strategies with lower expected returns than shareholders preferred. As residual claimants, shareholders preferred riskier strategies with higher expected returns because they did not have to share the potential gains with creditors, but shifted some portion of the default risks onto them. Bank creditors, of course, understood this and controlled for it in various ways. Depositors required higher interest payments in the face of higher default risks. Note holders, too, demanded compensation and took it by accepting notes only at a discount from par. In demanding the risk premium, note holders equalized the expected marginal value of transaction and liquidity services provided by the bank and the expected costs of default.

Fama (1980, p. 292) contends that modern equity markets provide information by valuing management's risk/return tradeoffs. Equity holders purchase shares at a price which reflects its risks both present and future. While most shareholders have little incentive to monitor directly, the existence of public markets that efficiently price the firm's risks provide signals about management's performance. The free riding problem is solved if a sufficient number of potential investors believe management's actions are inefficient and can lead a takeover of the firm.

Ante-bellum equity markets, however, were thin and probably incapable of providing effective discipline. Debt markets, particularly that for bank notes, provided the obvious answer. Unlike equity markets, note markets were extensive and debt was actively traded. Like shareholders, note brokers purchased notes at prices reflecting both present and expected future risks and provided valuable information to both debt and equity holders about the risk/return tradeoffs being pursued by current management. Recent work by Haupert (1994) and Bodenhorn (1995) reveals that the market for those notes operated efficiently, and transmitted information concerning the reputation of individual banks. That is, the discount at which notes traded was in part a reflection of the reputation of the bank for credible service and commitment. Although bank note markets did not provide the disciplining device of potential takeover, the prices paid for notes informed shareholders of management's decisions and they could discipline managers directly at annual shareholder meetings by voting for directors willing to change the current management or indirectly by selling shares to others who were willing to do so.

An agency cost explanation then resolves the note issue/deposit conundrum. Although deposit banking was more profitable than note issue, deposits could not discipline bankers in the same way as note issues. Unlike bank notes which were freely alienable promises to pay a stated sum at any future date, deposits are generally not freely alienable rights to a specific sum (try cashing any two-party check even today, or try cashing a personal check out of state). Checks are promises to pay a particular individual a particular (often uneven) sum on or after a certain date. Because of their specificity, public markets have typically failed to appear in the trading of deposits. Asymmetric information problems would have become even more acute than they were in note markets. In addition to determining the default risk of the bank upon which the check was drawn, the prospective purchaser would have also had to determine the default risk of the check's drawer. Such problems were not insurmountable as active markets for bills of exchange attest. Like checks, bills of exchange ware promises to pay a particular (often uneven) sum to a particular person of firm at a future date. But unlike checks, bills of exchange were freely alienable and there was recourse for non-payment to all endorsers, drawers and acceptors of the bill. And information asymmetries were often overcome by having a prominent banker or merchant act as acceptor of the bill.

Demand deposit instruments never offered the disciplining device achieved by notes which explains why banks continued to issue notes even in the face of a superior from of credit intermediation. Several banking historians have argued that deposit banking displaced note issue only after state-chartered bank note issue was proscribed by federal banking acts passed during the Civil War and Reconstruction. An alternative explanation, and the one proposed here, is that note issue declined in importance only as equity markets grew in both size and sophistication thereby providing an alternative monitoring device for managers. Concurrent with deepening equity markets, of course, was increased bureaucratic oversight by the newly created Comptroller of the Currency and the federal treasury, both of which gradually replaced private with public oversight and monitoring. As with New York's Safety Fund experience, the presence of such public oversight may have diminished the expected value of private monitoring sufficiently so as to make it unattractive.

Previous studies of ante-bellum American banking have focused almost exclusively on the costs of multiple note issues. Cagan (1963, p. 20), for example, argued that:

. . . the overwhelming variety of circulating notes confused the public and abetted fraud. . . . The direct cost to society of transacting business with this confusing mass of paper, 'torn, greasy, issued by nobody knows whom,' can be measured by the expense of supplying and using bank note detectors . . . [and] no gains offset these costs, which therefore burdened the economy as a whole.

Even those with some sympathy for the operations of free markets in the supply of money and credit have tended to focus on the costs (Rockoff 1974, Rolnick and Weber 1983, 1988, Economopolous 1988, Friedman and Schwartz 1963). While most have shown the costs to be small, none has explicitly addressed the positive benefits arising from multiple issues. It is certainly true that multiple note issues created costs that were partially eliminated by the postbellum national banking system. There were, however, compensating benefits arising from the ante-bellum system. In a world where capital markets were thin and unlikely to provide adequate monitoring of bank management's behavior, bank note markets could and did provide it (Bodenhorn 1995). Specialized market makers priced bank debt in centralized, public markets and provided valuable information to otherwise uninformed traders. The presence of such markets and such monitors may, as well, have mitigated the ever-present possibility of bank runs.

While lessons can often be learned from past experience, one must avoid too sweeping generalizations. It is unlikely, of course, that the lessons learned from the history of multiple note issuers can have much direct relevance to current discussions on banking policy. It is notable, however, that several studies of the savings and loan crisis of the 1980s have argued that reform is necessary. And in many cases, the method of reform called for is to reintroduce market monitoring - not to replace publicly-sponsored deposit insurance, but to supplement it. The suggestions have been varied, and often contradictory, but have included such things as market value accounting, private reinsurers, true co-insurance, risk-adjusted capital requirements and the use of subordinated debt. The presumption underlying each suggestion is that markets can provide some effective monitoring above and beyond what public regulators can provide. If the ante-bellum experience can serve as a guide, it would appear that these presumptions are correct.