Chapter 5 SCOTTISH FREE BANKING

The three previous chapters have presented a number of theoretical arguments for free banking. One must turn now to the historical side of this issue. This chapter is devoted to a look at Scottish free banking. Chapter 6 will examine the American experiment with multiple issuers of notes. These are probably the best-known and best-documented cases of free banking. However, the reader should be aware of a problem that persists with any review of these periods. There is much anecdotal evidence extant, but rather limited hard data—especially in the Scottish case. Therefore, some of the implications of the model presented earlier simply cannot be tested in any direct way, and the empirical conclusions that can be reached must in some instances be rather tentative because of the nature and age of the data. Despite such difficulties, the importance of these episodes makes their investigation imperative.

There was a time when knowledge of free-banking regimes had almost disappeared from economic history. This was, of course, prompted by the theoretical conviction that central banking was natural and necessary, because free banking was believed to be chaotic and inflationary. The recent rebirth of interest in the theory of multiple note issuers¹ has also inspired much historical investigation. Work by, among others, Hugh Rockoff (1975; 1985), Arthur Rolnick and Warren Weber (1982a; 1982b; 1985; 1986), Lars Jonung (1989), Kevin Dowd (1992), and Donald Wells and Leslie Scruggs (1986a) has significantly expanded the understanding of such systems. For example, it is now known that banking with multiple note brands has been much more common than most would think. There appear to have been at least sixty episodes of approximate free banking, that is, cases where banks were allowed to issue their own notes (Dowd 1992, 2).² Perhaps the most influential of all such examinations of historical free banking has been the work of Lawrence H. White on Scottish banking (1984a, 1984b, 1989a, 1991).

Friedrich Hayek had earlier presented a theoretical case for thinking that free banking would work. White’s Free Banking in Britain appeared in 1984 and argued that free banking had already worked in Scotland during an earlier period. A convincing “test case” for free banking seems to have been overlooked for generations.³ White’s conclusions have elicited intense interest, and deserve to be examined seriously and carefully.

WHITE’S INTERPRETATION

Free banks began to appear after the charter of the monopolistic Bank of Scotland expired in 1716. The first was the Royal Bank of Scotland in 1727. The competition that arose between the two rivals quickly brought benefits to the public in the form of “the cash credit account, a form of overdraft account” (White 1984a, 26). This expanded the bank’s note circulation at the same time that it “allowed an individual to borrow against his human capital at lower transactions costs and so enable him to undertake productive projects that otherwise would have been unprofitable” (White 1984a, 27). During the 1740s and 1750s, a number of small banking concerns appeared, but the most important entrant of the time was the British Linen Company (later the British Linen Bank) in 1746.⁴

There was nearly complete freedom of entry into banking during this era. After 1765, the only legal restrictions were (1) shareholders faced unlimited liability regarding creditors’ claims on the bank,⁵ (2) banks could not print notes of denominations smaller than one pound-sterling, and (3) option clauses were illegal. Despite these constraints, banking in Scotland prospered, and the economy grew. By 1844, there were “19 banks of issue in Scotland with 363 branches, providing one bank office for every 6,600 persons in Scotland, as compared with one for 9,405 in England and one for 16,000 in the United States” (White 1984a, 34). Even though there existed some economies of scale, there seemed to be no tendency for the system to collapse into a single entity. Moreover, economic growth seemed robust. Citing Rondo Cameron, White declares that “Scotland’s per capita income was no more than half that of England’s in 1750 but nearly equal it by 1845” (1984a, 24).

An important feature of the Scottish system was the weekly note-exchange. This developed spontaneously in the late 1760s. Banks realized that in order to maximize the market for their own notes, they would need to accept their rivals’ notes at face value. That is, if consumer X is presently holding notes issued by bank A, he is unlikely to open an account with bank B unless B is willing to credit him with the face value of the banknotes he possesses. This note-exchange process involved the weekly clearing of interbank holdings of one another’s notes. It served not only to increase the marketability of banknotes relative to specie, but also to check any attempt at overissue on the part of an individual bank (White 1984a, 20–22).

According to White, the virtues of the Scottish system were undeniable. Counterfeiting was “not a significant problem in the Scottish experience” (White 1984a, 40), for example. This he attributes to the care the free banks took in designing elaborate and distinctive notes and the rapidity with which—because of the note-exchange process—notes were returned to their issuer. By way of contrast, counterfeiting of Bank of England notes was a chronic problem in England, especially during the 1797–1821 suspension of specie payments. Noteholder losses in Scotland were also relatively small. The total loss to the public brought about by bank failures amounted to 32,000 pounds-sterling for the period 1695 to 1841, whereas the losses in England for the year 1840 alone were twice as great (White 1984a, 41). This was an innovative system as well. The British Linen Bank of Edinburgh established the first branch bank in the early 1760s. The Union Bank of Glasgow made the first public disclosure of a bank’s annual balance sheet in 1836.

Bank failure was not common according to White. The only major bank failure during the entire period was that of the Ayr Bank in 1772. It tried to do what could not be done in this system. It attempted to expand the circulation of its notes beyond the demand. Indeed, the Ayr Bank overextended itself to the tune of about 667,000 pounds-sterling. Nevertheless, this failure did not result in a general bank panic. In fact, the Edinburgh banks experienced a perceptible increase in specie demand for only one day. The reason was that, because of the rapid note-exchange mechanism, no other major banks were caught holding large amounts of the Ayr Bank’s notes. Furthermore, because of the unlimited liability provision, all the bank’s creditors were paid in full by the bank’s 241 shareholders (White 1984a, 31–32).

On top of everything else, the “Scottish free banking system proved far hardier during periods of commercial distress than did its English counterpart” (White 1984a, 44). Throughout the long Napoleonic Wars, for example, the Scottish economy allegedly experienced milder cyclical fluctuations than did that of England. This White attributes in part to its superior banking system. The empirical support for White’s conclusions consists of data on comparative failure rates for Scottish versus English banks over the period 1809–1830. This reveals that on average Scottish banks failed at the rate of 4.0 per thousand, whereas English failures were 18.1 per thousand (White 1984a, 48).

White admits the danger of basing broad conclusions on a single historical example. Nevertheless, he believes Scottish free banking to be illustrative of certain basic principles.

[B]ecause we lack knowledge of any other truly free banking systems of significance, the Scottish experience, interpreted in the light of our theoretical constructions, must largely inform our understanding of what we should generally expect from free banking. Though perhaps not conclusive, Scotland’s experience is certainly consistent with the hypothesis that monetary freedom is workable and self-regulating. (White 1984a, 137)

Finally, White notes that free banking ended not because of dissatisfaction on the part of either the consumers or the bankers in Scotland. Both seemed quite pleased with—even protective of—the system. The demise of Scottish free banking was imposed from above for political reasons. The English Parliament passed the Bank Charter Act in 1844 and the Scottish Bank Act in 1845. These effectively ended freedom of entry and the competitive issuance of notes, and put the Scottish banks under the aegis of the Bank of England.

A CRITIQUE OF WHITE

The increasing attention devoted to free banking in the last few years has certainly been both invigorating and long overdue. Moreover, much of the credit for this reanimated interest must be given to White for his research into Scottish banking. However, despite its stimulating content, White’s work is not without some questionable elements. The present section will enumerate some of the reasons for such skepticism.⁶ In general, it is not that White’s facts are false. The doubts that arise stem principally from the belief that White has either misinterpreted the facts or failed to include certain countervailing facts.

One should understand, first of all, that this controversy is not merely some trivial dispute over an arcane bit of history with no relevance for the present day. In many writers’ minds, the theoretical case for free banking has been intimately tied to the alleged success in Scotland. White himself contends that free banking in Scotland “provides unique evidence on the workability of monetary freedom” and “may also help us answer other questions concerning the stability or efficiency of an unregulated monetary system” (1984a, 137, 141).

Furthermore, White’s interpretation of Scottish banking has gained a wide circulation. Among those who have cited White favorably, one finds Milton Friedman and Anna Schwartz (1986, 49–51), George Selgin (1988a, 7, 81, 140; 1989, 450), Catherine England (1988, 795–96), Karen Palasek (1989, 400), Gerald O’Driscoll and Mario Rizzo (1985, 10, 225), Murray Rothbard (1983, 185), Gerald O’Driscoll (1986, 601–2), David Glasner (1989, 37), Dowd (1989, 153–57; 1992, 1), and Roger Miller and Robert Pulsinelli (1989, 211–12). As for objections, until recently, there had been few. Jack Carr and Frank Mathewson (1988), Larry Sechrest (1988; 1990; 1991), Rothbard (1988b), and Tyler Cowen and Randall Kroszner (1989) have posed challenges to White’s historical work, whereas Charles Munn (1985) and Charles Goodhart (1987) have offered short critical reviews.

The issues to be raised here are (1) bank failure rates, (2) economic growth, (3) note convertibility, (4) restrictions on small-denomination notes, (5) the Usury Law as a constraint on competition, and (6) the privileges of the chartered banks. The contention is that, although Scotland did allow a multiplicity of note issues, it nevertheless was not the close approximation to “pure” free banking that White seems to think it was.⁷

Much of the evidence presented in what follows will be drawn from the survey of Scottish banking written by S. G. Checkland (1975). This work White has called “S. G. Checkland’s authoritative chronicle of the industry” (1984a, 33). Moreover, White himself has declared that Checkland “is, of course, the authority on the facts” (letter to Peter Lewin and the author, April 30, 1986).

Bank Failures and Stability

Certainly one of the key dimensions along which one would want to measure the success of any banking system is the rate of firm failure.⁸ White obviously agrees for he bases most of his case for Scottish stability on the allegedly lower rate of failure experienced by Scottish banks relative to English banks (1984a, 48). Not surprisingly, he concludes from this that Scottish banks were substantially less failure-prone and calls the Scottish system one of “remarkable monetary stability” (1984a, 23).

First of all, White seems slightly to miscalculate the averages. From his own table, the figures work out to be 4.46 and 17.54, respectively, rather than 4.0 and 18.1. More importantly, White’s figures cover only a small portion of the period in question. He himself has stated that “the act of 1765 left Scotland with free banking” (1984a, 30). A longer time series would seem desirable, especially in light of the Ayr Bank failure of 1772. That episode White describes as minor insofar as the system as a whole was concerned (1984a, 32). Checkland agrees that little permanent damage was done (1975, 133–34), but does point out that “no less than thirteen Edinburgh private bankers fell with the Ayr Bank, never to rise again” (1975, 132). There were also several Scottish failures between 1773 and 1808. If one includes these data, one finds that for the period 1772–1830, the average annual failure rates per thousand for England and Scotland are 14.90 and 14.88, respectively (see Table 1). The rate for Scotland is thus not statistically different from that for England at the 99 percent confidence level.

White has recently attempted to rebut this point. He suggests carrying the series back to 1716 and thereby achieving a failure rate for Scotland of 7.98 (White 1991, 811–12). This, however, is inconsistent with his own declaration that the free-banking era started in 1765 (1984a, 30). Indeed, since White puts such great emphasis on the note-exchange process as a foundation of the Scottish success, one would think that he might identify the beginning of free banking with the development of that process, that is, about 1770.⁹ White also suggests that one should be more concerned with economic significance than with statistical significance and that raw failure rates per se are not necessarily very important (1991, 813). Both points are well taken. However, how does one determine what an “economically significant” rate of bank failure may be? For example, the average annual rate per thousand for U.S. banks between June 30, 1921 and June 30, 1929 was 22.23 (Upham and Lamke 1934, 247), higher than either the English or Scottish rates noted above. Furthermore, the 1920s are conventionally perceived as having been both prosperous and stable. Is 22.23 a “significantly” high rate of failure? Finally, if White did not think failure rates so important, why did he make the contrast between the two countries a prominent part of his defense of Scottish free banking?

Table 1
Bank Failures per Thousand, 1772–1830

Year	England	Scotland
1772	. . . *	451.6
1773	. . . *	0
1774	. . . *	0
1775	. . . *	0
1776	. . . *	47.6
1777	. . . *	0
1778	. . . *	0
1779	. . . *	0
1780	. . . *	0
1781	. . . *	41.7
1782	. . . *	0
1783	. . . *	0
1784	25.2	0
1785	. . . *	0
1786	. . . *	0
1787	. . . *	0
1788	. . . *	0
1789	. . . *	0
1790	. . . *	0
1791	. . . *	0
1792	. . . *	0
1793	17.9	90.9
1794	3.7	0
1795	. . . *	0
1796	6.6	0
1797	17.4	0
1798	12.8	0
1799	. . . *	0
1800	8.1	0
1801	10.4	0
1802	7.6	0
1803	14.6	0
1804	14.5	0
1805	11.4	0
1806	4.2	0
1807	5.8	0
1808	5.2	54.1
1809	5.7	0
1810	25.6	27.0
1811	5.1	0
1812	20.6	0
1813	8.7	14.3
1814	28.7	0
1815	27.3	9.0
1816	44.5	14.1
1817	4.0	0
1818	3.9	0
1819	16.5	0
1820	5.2	13.2
1821	12.8	66.7
1822	11.6	13.0
1823	11.6	0
1824	12.8	0
1825	46.4	12.0
1826	53.1	11.1
1827	11.9	0
1828	4.5	0
1829	4.4	11.4
1830	20.9	0

Sources: Lawrence H. White, Free Banking in Britain: Theory. Experience, and Debate. 1800–1845 (New York and London: Cambridge University Press, 1984) 48. Sidney G. Checkland, Scottish Banking: A History. 1695–1973 (Glasgow, Scotland: Collins, 1975) 132, 177–78. Leslie S. Pressnell, Country Banking in the Industrial Revolution (Oxford: Clarendon Press, 1956) 11, 537–38.

Notes: mean (England) = X₁= 14.90

mean (Scotland) = X₂ = 14.88

standard deviation (England)= σ₁ = 12.09

standard deviation (Scotland)= σ₂ = 60.01

observations (England)= n₁ = 37

observations (Scotland)n₂ = 59

To test the hypothesis that X₁ = X₂,

Therefore, one cannot reject the hypothesis at 99 percent confidence level.

In contrast, using White’s figures,

X1 = 17.54		σ = 14.33,		n1 = 22
X2 = 4.46		σ = 5.98,		n2 = 22

Thus, one must reject the hypothesis at the 99 percent confidence level.

*No data are available.

To pursue the issue of stability, one may note that there occurred a number of financial crises during the free-banking period: in 1772, 1778, 1787, 1793, 1797, 1802–1803, 1809–1810, 1818–1819, 1825–1826, 1836–1837, and 1839 (Checkland 1975, 213–14, 403), nor does one get an impression of much stability if one examines wholesale prices for Great Britain during the 1765–1845 period (see Table 2).¹⁰ Checkland offers a summary judgment of the Scottish system when he says that “[i]n principle, it should have been capable of stability or, at least, of fairly easy contraction. In reality, it was not” (1975, 214).

Table 2
Wholesale Price Index for Great Britain, 1765–1845 (1770 = 100)

Year	Index
1765	106
1766	107
1767	109
1768	108
1769	99
1770	100
1771	107
1772	117
1773	119
1774	116
1775	113
1776	114
1777	108
1778	117
1779	111
1780	110
1781	115
1782	116
1783	129
1784	126
1785	120
1786	119
1787	117
1788	121
1789	117
1790	124
1791	125
1792	123
1793	135
1794	137
1795	160
1796	162
1797	148
1798	150
1799	174
1800	210
1801	217
1802	170
1803	173
1804	173
1805	189
1806	187
1807	183
1808	201
1809	216
1810	213
1811	202
1812	228
1813	235
1814	215
1815	181
1816	166
1817	184
1818	194
1819	178
1820	160
1821	139
1822	123
1823	137
1824	142
1825	157
1826	139
1827	138
1828	134
1829	134
1830	132
1831	132
1832	128
1833	124
1834	121
1835	118
1836	132
1837	131
1838	137
1839	145
1840	143
1841	137
1842	124
1843	111
1844	113
1845	116

Source: Brian R. Mitchell, European Historical Statistics; 1750–1970. 2nd ed. (London: Macmillan Press, 1978) 388.

Economic Growth

What of White’s claim that “the period of Scottish free banking coincided with a period of impressive industrial development. . . . The growth of Scotland’s economy in the century prior to 1844 was more rapid even than England’s” (1984a, 24)? This conclusion is based on the statement of economic historian Rondo Cameron to the effect that “it would not be unreasonable to infer . . . that in 1750 per capita income of Scotland was no more than half that of England, but that by 1845 it very nearly equaled England’s” (1967, 94). This inference is not based on unambiguous data. Cameron himself admits that “there are no separate statistics of national income for Scotland for this period” (1967, 94). Conclusions of the sort in which Cameron and White indulge must, therefore, be based on indirect evidence. Admittedly, some such evidence does exist.

For example, both Richard Hildreth (1968, 16) and T. S. Ashton (1969, 70–75) suggest that the banking industry in Scotland contributed significantly to that country’s expansion. White cites Adam Smith as an additional proponent of this view (1984a, 24). However, this writer can find no such assertions in the Wealth of Nations.¹¹ Indeed, the opposite seems more nearly to be Adam Smith’s position.¹² Smith states that “Scotland, though advancing to greater wealth, is advancing more slowly than England” (1937, 189). He also makes references to the “imprudence” of Scottish banks (1937, 288) and “that excess of banking, which has of late been complained of both in Scotland and in other places” (1937, 302). It would seem clear that the issues of cyclical stability and economic growth will not be resolved until separate series on Scottish prices, interest rates, and national income are either discovered or generated. Until then, White’s conclusions must remain quite tentative.

Banknote Convertibility

White goes so far as to define free banking as “the unrestricted competitive issue of specie-convertible money by unprivileged banks” (1984a, ix). More recently, he has reaffirmed that redeemability is an essential characteristic of competitive inside money (1989b, 368fn). Thus, if convertibility was not, in fact, consistently practiced in Scotland, then one may conclude that a significant element of free banking was absent.¹³ Of course, it is well known that before 1765 immediate redemption did not always occur because banks sometimes invoked the “option clause”; that is, they delayed redemption in exchange for the payment of explicit interest to the noteholder. Did strict convertibility hold, however, after 1765? Damaging to White’s case is this declaration by Checkland:

The Scottish system was one of continuous partial suspension of payments. No one really expected to be able to enter a Scots bank . . . with a large holding of notes and receive the equivalent immediately in gold or silver. At best they would get a little specie and perhaps bills on London. (1975, 185)

Checkland also suggests that “much emphasis was laid on the loyalty of the banks’ customers—requests for specie met with disapproval and almost with charges of disloyalty” (1975, 184). Frank W. Fetter agrees: “To a large degree there was a tradition, almost with the force of law, that banks should not be required to redeem their notes in coin” (1965, 122). To this, Henry Meulen adds the observation that a Scottish bank usually “paid notes instead of gold to any depositor who might call, and thus was able to operate with a smaller reserve of gold than would otherwise have been necessary” (1934, 136). Additional comments along these lines may be found in both Checkland (1975, 186, 222, 438) and Meulen (1934, 129).

Curiously, White (1989a, 36) claims that statements such as the foregoing have been rebutted by Kevin Dowd (1989). In fact, Dowd agrees that “Scottish notes were imperfectly convertible, even after the passage of the 1765 Act” (1989, 156). What Dowd does dispute is that such inconvertibility represented a significant departure from free banking (1989, 156–57).¹⁴ Yet both White (1984a, 6–19) and Selgin (1988a, 94–96) have argued forcefully that free banks issuing debt-based, that is, specie-convertible, notes and deposit credits are constrained from overissuing such liabilities by the fact that these firms face rising marginal costs. Furthermore, said marginal costs rise largely because of liquidity costs, that is, the costs of acquiring and holding specie for the purpose of redemption. This redemption may occur in the course of either interbank or bank-customer transactions (the processes of adverse clearings and reflux).

In the absence of convertibility, free banks would experience a much-relaxed constraint on overissuance. For example, when discussing free banking in Michigan, Dowd seems to concede this when he states that “the suspension of convertibility removed the main check against over-issue, and so a monetary explosion was to be expected” (1989, 137). It would seem clear that convertibility is essential to any (specie-based) system that merits being characterized as free banking.

Small-Denomination Notes

In 1765, the British Parliament imposed on Scotland legislation that prohibited not only the option clause, but also the issue of notes smaller than one pound-sterling. The option clause has often been discussed—see Dowd (1991h), White (1984a), or Selgin (1988a), for example—but the prohibition of small-denomination notes seems to have received little attention.

Three aspects of this are of importance. First of all, one needs to realize that the one-pound note of 1765 had roughly the purchasing power of $180–$200 in the United States today.¹⁵ This figure may be achieved by discovering that British prices are presently ninety to 100 times the level of 1765 (Mitchell 1988, 719–34) and observing recent exchange rates of about $2.00 per pound. The implication is that after 1765 many day-to-day transactions could not be conducted in terms of banknotes; recourse to coins was necessary. Furthermore, the control of coinage rested with the Royal Mint and the Bank of England (Clapham 1958, Vol. II, 51–53). In other words, Scottish banks were systematically excluded from competition by means of notes for the business of those whose currency needs were relatively small in scale.

This restriction likely had two further effects. It may have served as a barrier to entry for small banks, since it could deny them the “niche strategy” of catering to small entrepreneurs and to the less wealthy consumers. Furthermore, since small-denomination notes always tend to circulate more rapidly than those of large denominations (White 1984a, 8), it would seem that the Act of 1765 must have reduced to some extent the effectiveness of the reflux process. That is, it may have raised the average period of circulation for Scottish banknotes and, thereby, increased the possibility of inflationary overissues. Consistent with this hypothesis, one finds Adam Smith’s observation in 1776 that in Scotland, “the circulation has frequently been over-stocked with paper money” (1937, 286). One may add to this the facts that (1) food prices fell from 1717 to 1750 but rose strongly in the latter part of the eighteenth century, as did coal, cattle, and grain prices, and (2) Scottish net exports declined after 1775 and were generally negative from 1780 to 1805 (Lythe and Butt 1975, 102, 103, 113, 116, 117, 162, 247). All of this suggests—but does not prove—the existence of an inflationary monetary expansion.

Interest Rate Ceilings

In 1714 the Usury Law, to which the Scottish banks were subject, established a legal maximum rate of 5 percent to be charged by financial institutions. The last remnants of this law did not disappear until 1854 (Clapham 1958, Vol. II., 224). Of these facts, there can be no doubt. However, White has questioned, first of all, whether the 5 percent ceiling applied to one of the key sources of revenue for Scottish banks: the discounting of commercial bills of exchange (letter to Peter Lewin and the author, April 30, 1986). It did indeed. The Usury Law was applicable to “the entire bill market” until 1833, when ninety-day bills were made exempt (Homer 1963, 205).

Also, in the same letter cited above, White wonders if this ceiling was ever a binding constraint. If one takes that phrase to denote a circumstance in which market rates of interest are driven above the maximum legal rate, then one must apparently answer in the affirmative. S. G. E. Lythe and J. Butt, while discussing Scottish finance in the eighteenth century, note that “the price for capital might be higher than the legal maximum bank rate” (1975, 155). Since consols¹⁶ issued by the British government were not subject to the Usury Law (Homer 1963, 205), one might take the yield on consols to be a reflection of market conditions. That yield exceeded 5 percent in the years 1781, 1782, 1784, and 1796–1799 (Homer 1963, 161–62). During the long Napoleonic Wars (1793–1815), effective market rates were often above the maximum legal rate (Homer 1963, 186, 205).¹⁷ Short-term market rates also rose above 5 percent during the years 1836, 1837, and 1839–1841 (Homer 1963, 208). However, it is unclear whether the latter posed an impediment to bank competition, since bills of exchange and promissory notes were exempt from the Usury Law after 1833 (Checkland 1975, 192, 443). It is possible that short-term rates were greater than 5 percent during part of 1826 as well: The average of such rates for that year was 4.5 percent (Mitchell 1988, 683).

It would seem that the interest rate ceiling must, in fact, have been a constraint on bank competition during at least part of the 1765–1845 period. Checkland concurs when he states that “the Usury Law limited competition for deposits” and that its effect on “any form of advance was seriously inhibitive” (1975, 432, 192). This assessment is echoed by Meulen (1934, 92).

Privileged Banks

One may recall White’s definition of free banking as a system of “unprivileged private banks.” Yet there were two tiers to the Scottish system: (1) three chartered “public” institutions (the Bank of Scotland, the Royal Bank, and the British Linen Bank) and (2) the various private banks and joint stock banking companies. Since the three public banks enjoyed limited shareholder liability while the others were all subject to unlimited liability, Checkland concludes that the former “were in a preferred position relative to all others” (1975, 235). The state had created these public banks and “continued to confirm their preferred position through their limited liability and through their public identity and perpetual succession” (Checkland 1975, 275). It would thus seem that the nonchartered banks faced a significant regulatory barrier to entry: unlimited liability.¹⁸ This did not prevent the formation of a number of private banking concerns, but it imposed a constraint on such firms that was not applicable to the three chartered banks.

White contests this. He asserts that unlimited liability must not have been a binding constraint on the private banks, because they “chose to retain unlimited liability in the 1860s and 70s even after limited liability became available to them” (1984a, 143). To argue thus is less than convincing. Institutional structures must be viewed contextually. The fact that Scottish banks of the 1860s seem not to have seen unlimited liability as an odious imposition does not prove that it was not considered to be such in, say, 1780 or 1810. By 1860, the tradition of unlimited liability may have become so deeply entrenched that to abandon it would have shaken consumer confidence. Despite this, unlimited liability may still have been a barrier to entry during the eighteenth and early nineteenth centuries. In addition, there is a powerful counterpoint that one must consider. Since the three public banks expended real resources in order to (1) obtain their charters and (2) prevent other banks from gaining charters, one must conclude that a limited liability bank charter was perceived as conferring some significant advantage upon its holder (Cowen and Kroszner 1989, 226).

A specific manifestation of said advantage was the fact that “there was a long-standing government instruction to the officers of the customs to accept only the notes of the chartered banks in payment of duties, and to ‘refuse the Notes of every other bank without exception’ ” (Checkland 1975, 186). In short, an artificial demand for the notes of the public banks was established by fiat. Yet one might wonder if the payment of customs duties was of a magnitude sufficient to produce a significant gain for those institutions. One possible indication is the proportion of total government revenues represented by customs duties. One finds that customs duties averaged 22.3 percent of annual government income over the period 1765–1801 and 27.6 percent from 1802 to 1845 (Mitchell 1988, 576–77, 581–82).¹⁹ The collection of customs duties—and therefore the benefit to the public banks—seems not to have been trivial.

Dependence upon the Bank of England

Several writers have argued that the Scottish system was not true free banking because it was crucially dependent upon the Bank of England. That is, it has been claimed that the private Scottish banks depended upon the three public banks, and those three in turn relied upon the Bank of England for their liquidity needs (Rothbard 1988b;²⁰ Sechrest 1988; Cowen and Kroszner 1989; Dow and Smithin 1992²¹). These arguments have been based on various statements by historians of the period.

For example, Checkland declares that “by 1810, the Bank of England, short of the state itself, was the effective final arbiter of the supply of liquidity, both for England and Scotland” (1975, 276). Meulen notes that “it transpired that at times when gold was being drained both from Scottish and English banks the Scottish bankers had not restricted their note issue, but had withdrawn gold from the Bank of England to support their credit system” (1934, 141). Furthermore, the 1810 Bullion Committee’s report stated that “the circulation of the Bank of England had an important influence on the circulation of the country banks and of the Scottish banks” (quoted in Fetter 1965, 50). Finally, “the three chartered banks of Scotland kept their reserves largely in deposits with the Bank of England” (Fetter 1965, 34).

White has offered a rebuttal to the foregoing that is quite plausible even if not conclusive (1989a, 20–34). He seems to make three key assertions. First, White points out that the evidence for the pyramiding of Scottish bank notes on a base of Bank of England notes is sketchy and circumstantial.²² There seem to be no unambiguous data that could resolve the question. Second, he grants that England and Scotland formed an integrated economic whole,²³ but denies that this means that the Scottish system was a satellite of the Bank of England. As he puts it, one would not conclude that the U.S. banking system was a satellite of the English system just because actions by the Bank of England had spill-over effects in America. Finally, White offers evidence that the Bank of England did not explicitly consider itself to be a lender of last resort during the Scottish free-banking period. He argues persuasively that there is an important distinction between an ex ante lender of last resort and an ex post source of liquidity. One might call the former structural dependence and take it to be indicative of central banking. One might call the latter circumstantial reliance and perceive it as being consistent with free banking.

SUMMARY

Lawrence H. White’s very valuable research into Scottish free banking has generated much interest and some controversy. His work has explicitly suggested that the Scottish experience can tell us a great deal about how a true free-banking system would perform. To many writers, Scotland has come to represent the paradigmatic “test” of free banking.

However, on closer inspection, one finds some serious flaws in the Scottish system. The failure rate (1772–1830) for Scottish banks was not lower than that for English banks. Banknotes were not consistently convertible into specie on demand. The prohibition of small-denomination notes not only curtailed mutually beneficial transactions between banks and their customers, but also may have diluted the constraints on overissue. The Usury Law limited competition in credit markets. The three chartered banks held privileged positions within the system, and Scotland seems to have avoided neither the inflation nor the numerous crises that plagued the English.²⁴ Doubts about White’s interpretation of the Scottish experiment with free banking seem justified.²⁵ Finally, it is interesting to notice that White himself seems to be retreating a bit from his original position. Instead of speaking of free banking as lasting from 1765 to 1845, he now appears only to defend the period 1810–1844 as exemplifying free banking (1989a, 15, 16, 35).

NOTES

Portions of this chapter are reprinted, by permission, from Cato Journal 10 (Winter 1991): 799–808.

1. The key works are probably Hayek (1978), White (1984a), Selgin (1988a), and Dowd (1989).

2. All of these included other restrictions on bank activity, however.

3. One of the rare exceptions was Vera C. Smith (1990).

4. These three large, chartered banks—the Bank of Scotland, the Royal Bank, and the British Linen Bank—were distinct from other Scottish banking concerns and were referred to as “public banks.” Whether these three were legally privileged is one of the controversies to be addressed.

5. This was true of all except the Bank of Scotland, the Royal Bank, and the British Linen Bank. The shareholders of those three were subject to limited liability.

6. The author would like to emphasize that he has great respect for White’s work in general, but is unable to dispel certain doubts about White’s view of Scottish free banking. Moreover, these doubts only very slowly displaced an initial enthusiasm for White’s conclusions.

7. Simply put, the question is “how good” a test of free banking was the Scottish case. This is important, but it is, admittedly, a question of degree rather than of kind.

8. Most assume that success is achieved by a low rate of failure. A few economists—particularly Murray Rothbard—argue that high failure rates are desirable as an indication of market restraint.

9. That is one of the reasons why the author started his series with 1772.

10.It would have been far preferable to present data for Scotland alone, but such a series seems not to exist.

11. Neither edition of the Wealth of Nations possessed by the author is that which is cited by White. It is possible that some small—but in this case important—difference exists between the editions consulted.

12. It would be fair to keep in mind that Smith was writing in 1776. His comments do not, therefore, necessarily apply to the mature Scottish system. On the other hand, White seems willing to consider 1716 as the beginning of free banking (1991, 811–12). From that perspective, Smith’s comments would apply.

13. Arguments in defense of option clauses and indirect convertibility within the context of free banking will be discussed in Chapter 7.

14. This is not surprising, since Dowd is not reasoning from the same model that White employs.

15. Cowen and Kroszner (1989, 224) cite an estimate of $200.

16. Consols are bonds that never mature but pay interest to the holder forever.

17. Considering the substantial inflation in Britain during the period, it is hardly surprising that this should be the case. See Brian Mitchell (1988, 720).

18. See Carr and Mathewson (1988) for more on this issue.

19. These figures are for Great Britain as a whole; separate series for Scotland do not exist.

20. Rothbard’s essay, though shrill in tone, nevertheless raises some worthwhile points. Most important of these is his claim that White has misinterpreted the theoretical debate over free banking. To the author’s knowledge, Rothbard is the only one who has challenged White on this issue.

21. Sheila Dow and John Smithin also argue that the Scottish system exhibited a strong tendency toward concentration. This they applaud, for they ascribe the degree of success the Scots enjoyed to the lack of competition in the conventional sense.

22. White allows for the possibility that such pyramiding might have occurred during the 1797–1821 period (1989a, 22).

23. White recognizes that his earlier work may have led readers to conclude that he believed Scotland to be an entirely autonomous economy relative to England (1989a, 33).

24. This brings one back to a key theoretical issue. Is a properly functioning free-banking system not capable of at least mitigating the effects of external shocks? The author would argue that true free banking is capable of this. To test the stability of the Scottish system in a clear fashion, one would need separate series for Scottish prices, interest rates, industrial production, employment, and national income. To repeat, these do not appear to exist.

25. The author’s position has long been that the theoretical case for free banking need not rely on evidence from ambiguous historical cases like that of Scotland.