حل معمای صدور اسکناس های بانک ملی

Much of the puzzle of underissuance of national bank notes can be resolved for the period 1880–1900 (the period when detailed, bank-level data are available) by disaggregating, taking account of regulatory limits, and considering differences in banks’ opportunity costs cross-sectionally and over time. Banks with poor lending opportunities issued more, within regulatory limits. Banks tended to issue more when bond yields (the backing for notes) were high relative to lending opportunities. The profitability of note issuance was insufficient to attract entrants primarily or mainly for the purpose of note issuance. The observed lack of a general relationship between note issuance and reserve demand is inconsistent with the view that redemption costs from note issuance explain low note issuance in general. However, some variation in the propensities of urban banks to issue notes is associated with variation in reserve demand costs associated with the note issues of those banks. Generally, however, note issuance enjoyed economies of scope with deposit banking, including reduced costs of reserve requirements.

During the Civil War, the federal government began to charter national banks. These banks enjoyed the privilege of being licensed to issue national bank notes, which were default-risk-free liabilities of the banks, backed 111% by U.S. Treasury bonds deposited by issuing banks at the U.S. Treasury.3 The creation of these new banks, combined with a 10% annual tax on state bank note issues, soon resulted in the supplanting of state banks’ notes by the new national bank notes. Scholars have long puzzled over the observation that national banks did not take greater advantage of the authority to issue notes. The aggregate supply of notes never reached its maximum permissible level, despite calculations measuring the profitability of allocating capital toward bank note supply collateralized by bonds (e.g., as derived by Cagan, 1965), indicating that national bank note issuance was more profitable than the typical profit earned by allocating bank capital toward lending funded by a combination of deposits and capital. Friedman and Schwartz (1963, p. 23) wrote: Before 1890 the amount outstanding ranged around 20% of the possible maximum, by 1900 it had risen to about 28%, and by World War I to about 80%. The maximum was in fact approached only in the twenties, when for the first time U.S. bonds deposited to secure circulation and government deposits (which also required such security) nearly equaled the total of eligible bonds. Before 1905, the capital stock of national banks set narrower limits to their maximum possible note issue than did the total of eligible bonds, but the actual issue did not approach this lower limit either. Thereafter, the capital stock of national banks exceeded the total of eligible bonds and hence was not the effective limit on note issue. Yet, despite the failure to use fully the possibilities of note issue, the published market prices of government bonds bearing the circulation privilege were apparently always low enough to make note issue profitable except in the years 1884–1891. The fraction of the maximum issued fluctuated with the profitability of issue, but the fraction was throughout lower than might have been expected. We have no explanation for this puzzle. Friedman and Schwartz, 1963 and Cagan, 1965 argued that profits from note issue were large on the margin, because bond issues to back note issues remained cheap and because banks could easily leverage their capital devoted to those bond purchases.4 In their discussions of potential constraints on bank note issues, they pointed to the more than adequate aggregate supply of bonds, and while they recognized that regulations constrained bank note issuing relative to bank capital, they argued that bank capital was not a constraint because its aggregate amount exceeded the amount required for increased note issues. Cagan and Schwartz (1991) showed that the apparent excessive profitability of bank note issuance increased in the 20th century. The story typically advanced to explain low issuance of national bank notes conjectures hidden transacting costs faced by issuers. Authors such as Bell, 1912, Cagan, 1965, Goodhart, 1965, Cagan and Schwartz, 1991, Duggar and Rost, 1969, Champ et al., 1992 and Wallace and Zhu, 2004 have pointed to the possibility that redemption costs may have been large, and these hidden costs may explain bankers’ reluctance to issue despite the seeming profitability from expanding the supply of notes. Champ et al. (1992) argue that banks had an incentive to return national bank notes to their issuing banks (or to the Treasury) because national bank notes were not as good as greenbacks for purposes of satisfying national banks’ legal reserve requirements. They plot “redemptions” at the Treasury as a fraction of outstanding notes, and show that annual redemptions averaged about half of outstanding notes. This is the primary evidence offered in favor of their view that national banks were unwilling to hold each other’s notes, and routinely returned each other’s notes either to the Treasury or to the issuing banks, which imposed redemption costs on issuers. But the redemption evidence cited in Champ et al. (1992) suffers from some problems of interpretation. Much of the Treasury redemptions do not represent redemption demands by other banks to redeem notes because the notes were inferior to greenbacks as legal reserves. When one considers the stated causes of the reported redemption flows, it is not clear whether there were large redemption costs associated with those flows. The Annual Report of the Comptroller for 1890, for example, describes and quantifies the causes of the $67 million in “redemptions” that were received at the U.S. Treasury Redemption Agency for the period November 1889 through October 1890, which consisted in large part of (1) the replacement of worn out notes with newly printed notes ($24 million), (2) reductions in circulation related to legal requirements that required issuing banks to reduce their circulation ($21 million), and (3) reductions in circulation associated with insolvent and liquidating banks ($11 million). The remainder, about 20% of notes received by the Treasury, were not in any of these three categories; they were “fit for circulation, and [were] returned to the issuing banks.” Clearly, $32 million of the $67 million in notes were related to issuing banks’ reductions in circulation or bank liquidation. Most (nearly $24 million) of the remaining $36 million in notes appear to have been returned because they were worn out and needed to be replaced, leaving about $11 million in notes that were redeemed at the Treasury for other reasons (including, possibly, the hypothesized preference for greenbacks as bank reserves by banks receiving notes). Furthermore, there are reasons to doubt whether banks interested in swapping other banks’ notes for greenbacks would have chosen to return bank notes to the issuing bank or to the Treasury. Banks could pay out bank notes to the public instead. National bank notes were free of default risk and were perfect substitutes for greenbacks as cash in the hands of the public. Friedman and Schwartz (1963, pp. 21–22) point out that national bank notes virtually never traded at a discount relative to greenbacks. The public apparently was willing to accept national bank notes at par in lieu of greenbacks. Given the willingness of the public to accept national bank notes, it is not clear why notes would be presented at the Treasury or at any national bank for redemption, which would be more costly than just passing them on to the public. Of course, if the notes were worn out, the public might not accept them, and the notes could have to be returned to the Treasury, but predictable wear and tear might not entail very high costs of redemption for issuing banks.5 National banks had many ways of dealing with redemptions that could limit costs associated with disruptions to their operations or the need to maintain large cash balances to fund redemptions. They could instruct the Treasury to sell bonds on deposit at the Treasury to pay for redemptions, but they might be more likely to borrow in the interbank market (from a bank that could deliver funds to the Treasury as needed at low physical cost) to finance redemptions of worn out notes that would be immediately replaced. Presumably, in the case of worn out notes that were being immediately replaced by new ones, interbank lending would have been especially low in cost, given the knowledge that the issuing bank would soon be receiving fresh notes from the Treasury with which to repay the loan. Physical costs of moving notes back and forth were likely quite small.6 And even if notes could not be reissued immediately after being returned from the Treasury, this delay was not likely very costly, particularly since banks could still invest them in the interbank deposit market in the meantime, which yielded about 2% throughout our period. An alternative approach to explaining the low amounts of national bank notes issued was proposed by James (1978). He was the first to suggest that aggregate calculations, like those provided by Cagan and Friedman and Schwartz, might be providing a misleading picture of national bank note profitability. He showed that cross-sectional variation in the regional supply of bank notes was large and consistent with regional variation in the opportunity cost of note issuance (that is, regional variation in the profitability of bank lending). In James’s view, at least some of the puzzle of low bank note issuance was explained by the high profitability of bank lending in the South and West, where note issuance was relatively low. But James’s explanation was not a complete one. After 1874, there were no regional limits on note issuance, suggesting that banks in the East (where loan profitability was relatively low) should have substantially increased their outstanding notes. Why did the banks in the East not issue more notes? Hetherington (1990) showed that some of the time variation in the extent of note issue could be explained by changes in rules governing note issues. But that approach did not explain the puzzle posed by Friedman and Schwartz, 1963 and Cagan, 1965; like James’s (1978) explanation of cross-state variation in note issuance, Hetherington’s (1990) explanation of some of the variation in supply over time did not address the persistent underissue of bank notes: the aggregate level of bank notes remained far below its maximum despite the high apparent profitability of note issuing. In this paper, we are able to largely resolve the puzzle of note underissuance, at least for the period up to 1900, for which detailed individual national bank data are available. To do so requires a model of bank note issuing profitability, and a two-part empirical analysis. Section 2 describes the legal limits on note supply, and discusses the determinants of supply and demand for national bank notes. In the first part of our empirical analysis, in Section 3, we examine the behavior of existing banks, taking as exogenous their available amount of capital. By disaggregating data on national banks, and analyzing individual banks’ note issuing incentives and constraints, we show that the behavior of these banks is not puzzling: Banks with low opportunity costs issued the maximum amount of notes they could. Banks with high opportunity costs issued less than the maximum amount they could. Differences over time in propensities to issue notes are closely associated with variation in bond yields (the backing for note issues), ceteris paribus. In the second part of our empirical analysis, in Section 4, we consider implications of the opportunity cost view for entry and exit behavior. We show that exiting banks focused more on note issuing than average banks, and entering banks focused less on note issuing. In other words, lending and deposit taking, as opposed to holding government bonds and issuing national bank notes, appears to have been the business line that most attracted new entrants and improved bank survival.7 In Section 5, we consider whether redemption risk can provide an alternative or additional perspective to opportunity cost as an explanation for cross-sectional variation in note issuance. We find that, in general, bank note issuance does not raise redemption costs as measured by reserve demand. For urban banks, however, we do find some evidence of higher reserve demand associated with note issues, although this varies over time. Interestingly, that pattern of variation is associated with changes in urban banks’ residual propensities to issue notes (estimated in Section 3). Section 6 concludes that our results lend support to James’s opportunity cost theory of note issuing. A combination of legal restrictions on maximum note issuing and banks’ opportunity costs best explains the extent of bank note issuing in a manner consistent with bank profit maximization. Redemption costs, however, may have played a small role in discouraging urban banks from issuing notes in some places at some times.

For the period 1880–1900, we are able to explain much of the note issuing patterns of national bank notes when we disaggregate the data to the level of individual banks, take account of the limits banks faced on their maximum permissible note issues, and consider differences in opportunity costs of note issuing across banks. In 1880, 40% of national banks, in 1890, 5% of national banks, and in 1900, 21% of national banks were max- imum note issuers. Banks with low lending opportunities maximized their ability to issue notes but could not issue more than a certain amount. Other banks, with high lending opportunities, chose not to issue more notes. Years in which bond yields (the required backing for notes) were higher saw greater propensities to issue notes, holding constant lending opportunities. Models of redemption costs associated with issuing national bank notes (per se) do not explain the substan- tial cross-sectional variation in the extent to which national banks chose to issue national bank notes. The theory that redemption cost explains underissuance of notes is inconsistent with the observed lack of a general relationship between bank note issuance and excess reserve holdings. At the same time, the redemption cost explanation may have relevance for the issuance behavior of urban banks. For urban banks, low residual note issuance in some years (estimated in Section 3 ) is associated with higher reserve needs related to issuing national bank notes in some of those years, at least for one definition of reserves (NER). An examination of reserve demand indicates substantial economies of scope between note issuing, on the one hand, and deposit taking and lending, on the other hand. Those economies of scope probably included shared overhead costs as well as the ability to economize on the costs of maintaining mandatory minimum levels of reserves and capital when issuing notes. Combining deposit taking and note issuing allowed banks to make full use of capital and reserves that were legally required in support of note issuing but which exceeded warranted levels. National banks did not enter mainly to issue national bank notes. Indeed, over the period 1880–1900, new entrants focused less on note issuing, while banks exiting were more likely to be maximum issuers. Nor did banks raise new equity capital mainly to finance additional note issuance. Note issuing was profitable only for some banks, and only when combined with lending and deposit taking. Note issuing seems to have been a relatively unprofitable line of business for relatively successful bankers. During our period, understanding the low issuance of national bank notes is substantially improved by disaggregating data and thus avoiding misleading ‘‘representative bank ” analysis relating average bank behavior and average bank opportunities. Unfortunately, data limitations prevent us from applying this same approach to the period after 1900, when the apparent profitability of bank note issuance was especially high