read one of the two papers, write a report and answer questions

Financial History Review 3 (1996), pp. 49-68. Cambridge University Press.

A comparison of the stability and efficiency of the Canadian and American banking

systems, 1870-19251

MICHAEL D. B O R D O , HUGH ROCKOFF Rutgers University and the NBER

and ANGELA REDISH University of British Columbia

During the middle of the nineteenth century the United States and Canada embarked on two very different roads toward the goal of stable banking. The United States set up the national banking system (1863 and 1864), an attempt to achieve stability through imposing high capital requirements, a bond backing system for note issues, reserve requirements on deposits and other constraints. The idea of permitting branch banking and allowing market forces to operate was rejected. Canada, following Confederation, followed a very different path. Relatively few regulatory restrictions were imposed; instead Canada relied on the formation of large banks with interprovincial networks of branches to provide stability.

A financial economist from today transported to the mid-nineteenth century would expect the Canadian approach to be more appropriate. And in the end it was. The Canadian system survived the Great Depression undamaged, while the American system collapsed; and there have been no major bank failures in Canada since 1923. (There have been a substantial number of failures among trust compa- nies, especially since 1983.)2 In a previous paper3 we showed that this experience contrasts vividly even with that of the United States national banks (the strongest sector within the United States banking system) and, moreover, was not bought at the expense of the consumers of Canadian bank services.

1 A draft of this paper was given at the Colloquium 'Financial Institutions and Financial Markets in Twentieth Century Europe', Zurich, 27-28 May 1993, organised by the European Association for Banking History and the Verein fur Bankgeschichte (Schweiz und Fiirstentum Liechtenstein). For helpful comments on an earlier draft we wish to thank especially Neil Quigley and Ellis Tallman and an anonymous referee of this journal.

2 J. L. Carr, F. Mathewson and N . C. Quigley, Ensuring Failure: Financial System Stability and Deposit Insurance in Canada (Ottawa, 1994), pp. 1-4.

3 M. D. Bordo, A. Redish and H. Rockoff, 'The U.S. banking system from a northern exposure: stability versus efficiency', JoHma/ of Economic History, 54 (1994).

49

50 FINANCIAL HISTORY REVIEW

But our time-travelling economist would be disappointed if he or she expected the race to be decided quickly. For long periods before 1925, it could well be argued that the United States national banking system exhibited greater stability. It was only after the merger movement in Canadian banking (1900-25) reduced the number of banks and created a system characterised by a small number of large banks with thick interprovincial branch networks that the race was clearly won. Before the merger movement, freedom to branch allowed banks in Canada to become bigger relative to the Canadian banking system as a whole than could American banks relative to the American banking system. This created banks that could absorb substantial shocks, but it also created banks that could make major mistakes. Section I compares the stability of the two systems between 1870 and 1925 in terms of bank failures and system-wide disorders. Section II explains the differ- ences in stability over time in terms of branch banking and the shocks faced by the banking systems. Section III examines the effectiveness of the lender of last resort in both systems, while section IV takes a closer look at the merger movement and its effects on the stability of the system. Section V explores some additional evidence derived from bank portfolios and rates of return, and section VI reviews the main findings and discusses their implications for policy.

I

Two aspects of stability will be considered separately here: year to year variations in bank failures measured by the losses experienced by depositors and note holders; and the susceptibility of each system to liquidity crises.

With respect to losses on deposits and notes, Figure 1 shows deposit losses as a fraction of all deposits from 1870 to 1925 for national banks in the United States and for all banks in Canada. This is the appropriate comparison, at least to start with, because the national bank sector was stronger and weighted more heavily toward large money-centre banks. Loss rates were defined simply as total losses ultimately experienced by depositors of the banks that failed in a given year divided by total deposits in that year, expressed as a percentage. The number of years it took to collect on the assets, and to raise the additional amounts due from shareholders under the double indemnity rule that held in both countries, varied from bank to bank. Better measures, but not available, would be the present value at the time of failure of the losses actually experienced, or the difference between the nominal value of deposits and their market value at the time of failure.

Losses for Canada are from Beckhart,4 and total deposits are from Historical Statistics of Canada5 (series H20). Losses on United States national banks deposits are from the United States Federal Deposit Insurance Corporation, Annual Report 1934.

4 B. H. Beckhart, 'The banking system of Canada', in H. P. Willis and B. H. Beckhart (eds), Foreign Banking Systems (London, 1929).

5 M. C. Urquhart and K. A. H. Buckley (eds), Historical Statistics of Canada (Toronto, 1965).

CANADIAN AND AMERICAN BANKING

1070 1880 1890 1900 1910 1920

Canada US nationals

Figure 1. Deposit losses, i8?o-ig25 (percentage of total deposits; Canada and US nationals)

For the years 1870 to 1920 we simply divided the 'amount of losses' of'National Banks which did not reopen'6 by 'average of deposits reported on call dates'.7 This calculation omits some losses incurred by depositors in national banks that reopened. A feature of some reopenings was an agreement by depositors to waive some of their deposits, but the error from this source is likely to be small in the years covered. For the period 1921 to 1925 we took deposits of 'suspended' national banks, from which deposits of banks that reopened were subtracted,8 to obtain deposits of failed national banks. An obvious problem arises from a national bank that was temporarily closed, but reopened in a subsequent year. This should be less of a data difficulty before the 1930s, and should even up over time — the overestimate in one year will be offset by an underestimate the following year. To obtain deposit losses our estimate of deposits in banks that did not reopen was multiplied by 0.319, which is based on a sample of 267 'national banks which suspended during

6 United States Federal Deposit Insurance Corporation [hereafter FDIC], Annual Report 1934, PP- 92-3-

7 ibid., p. 89. 8 ibid., p. 94.

52 FINANCIAL HISTORY REVIEW

1921-1930 and did not reopen'.9 As for the period 1870 to 1920 the divisor is the 'average of deposits reported on call dates'.10

As can be seen in Figure 1, there were some losses for United States national banks in virtually every year, while in Canada there were long periods when there were no losses. But when failures did occur in Canada, they involved, on some occasions, much larger fractions of total deposits than was the case in the United States.11 A particularly striking example occurs during the 1890s, when the United States banking system experienced severe difficulties. Losses of national banks peaked at 0.27 per cent of total deposits in 1893, and gold payments were suspended for two months. No banks failed in Canada in 1893, and Canada avoided suspen- sion. But in 1895 the failure of one bank in Canada - the Banque du Peuple in Montreal — created losses equalling 0.86 per cent of deposits.

In the United States losses on notes (issued only by national banks) appear to have been negligible owing to the bond security provision of the National Banking Act. In most years the Comptroller12 shows sufficient cash deposited to redeem the outstanding notes of all failed banks. In a few years, however, there are small deficiencies and some notes may have gone unredeemed. In Canada noteholders lost in the long run in only two bank failures: the Mechanics Bank of Montreal in 1879 and the Bank of Prince Edward Island in 1881. But on those occasions the losses were sizeable compared with annual losses on deposits in both countries; the losses in 1879 amounted to about 0.32 per cent and in 1881 to about 0.33 per cent of total circulation in Canada. The bond security system in the United States, moreover, meant that noteholders were paid off quickly when a bank failed — as soon as the Comptroller of the Currency could realise the highly liquid bonds he held.

The comparative safety of the United States national banks, and especially the greater safety of national bank notes, was not lost on Canadian observers. During the late 1870s some sentiment existed in Canada to adopt a bond-secured note issue. However, various counter-arguments - perhaps the most persuasive being that a bond secured note issue would lack 'elasticity' - carried the day.

The revisions introduced by the 1890 Bank Act (the Bank Act regulated behav- iour of the chartered banks and was revised decennially), however, included a Bank Circulation Redemption Fund. Each bank was required to pay in an amount equal to five per cent of its average note circulation. Liquidators of an insolvent bank could draw on this fund, if the bank assets were not realised within 60 days. When assets were realised, reimbursement of the Fund was a first charge against the assets. Additionally, the notes of any bank that suspended were to bear interest at

9 ibid., p. 100. 10 ibid., p. 89. 11 As E. P. Neufeld, The Financial System of Canada (Toronto, 1972), p. 81, points out, the record of

the chartered banks was also 'more tarnished' than that of other Canadian intermediaries. 12 Annual Report of the Comptroller of the Currency 1920, 2, Table 37, pp. 8 0 - 1 2 3 .

CANADIAN AND AMERICAN BANKING 53

six per cent per annum until redeemed.13 Notes were also given a senior lien (ahead of deposits) on bank assets.14

Although Canadian banks had a higher loss ratio than the United States national system, they had a lower loss ratio than United States non-national banks, and a lower loss ratio than all United States banks. Loss data15 for non-national banks in the United States are far less reliable than for the national banks. State banking authorities did not always require detailed reports on bankruptcy proceedings, and some private commercial banks were not under any obligation to report to a government authority. The data plotted here are probably adequate for a broad brush comparison with Canada, or with the United States national banks, but not for more demanding uses. The data are based on surveys made by the United States Comptroller of the Currency.

For the years 1870 to 1920 we began with the liabilities of the 'state and private banks which did not reopen'.16 We then multiplied by 0.90 to get an estimate of the deposits - the figure the Federal Deposit Insurance Corporation used17 - to adjust for non-deposit liabilities such as bonded debt. Losses were obtained by then multiplying the deposit estimates by loss factors based on samples, each of which covered a group of banks failing over a number of years.18 To get the loss rate we then divided by 'estimated average deposits' of commercial banks 'other than national'.19 These data are reported on a biannual basis. For a given year we used the first of the pair of estimates that include data from that year. This decision may impart a slight upward bias in our estimates, perhaps offsetting in some measure the downward bias resulting from incomplete coverage.

For the years 1921 to 1925 we subtracted our estimate of losses on national bank deposits from the estimate of losses borne by all depositors reported by Friedman and Schwartz20 to get an estimate of losses at non-national banks. We then divided by the 'estimated average deposits of commercial banks other than national'21 to get the loss rate.

Figure 2 contrasts the loss rates of United States national and non-national banks, and Figure 3 contrasts loss rates of Canadian chartered banks and all United States banks. Although a number of Canadian failures still stand out, the worst Canadian years now have some rivals in the United States, particularly 1878, 1893 and 1907.

13 The Fund was held by the Minister of Finance and the banks received interest of three per cent per annum on their contributions. R. C. Maclvor, Canadian Monetary, Banking and Fiscal Development (Toronto, 1961), p. 77.

14 R. M. Breckenridge, The History of Banking in Canada (Washington, DC, 1910), p. 123. 15 Losses on United States non-national banks deposits are also from FDIC, Annual Report 1934. 16 ibid., pp. 92-3. 17 ibid., p. n o . 18 ibid., pp. 100-1. 19 ibid., p. 89. 2 0 M . F r i e d m a n a n d A . J . S c h w a r t z , A Monetary History of the United States ( P r i n c e t o n , N J , 1963),

p . 4 3 8 . 2 1 F D I C , Annual Report 1934, p . 89.

54 FINANCIAL HISTORY REVIEW

1870 1880 1890 1900 1910 1920

US national US non-national

Figure 2. Deposit losses, 1870-1925 (percentage of total deposits: US national and US non-national)

The high loss rate experienced by non-national banks in 1878 was surprising because this year is not identified as a crisis year by leading financial histories of the United States. Most financial histories appear to rely on Sprague's chronology.22

Sprague, however, may have neglected 1878 because of his focus on the national banking system. As shown in Figure 2, the difficulties in 1877 and 1878 were confined almost entirely to non-national banks.23 This source reports the results of the Comptroller's survey of national bank examiners, and the data may vary from state to state depending on the cooperation received by the examiners, how they defined a bank and the effort they put forward. The Comptroller's report, however, does confirm the large number of failures and the substantial liabilities of the non- national banks that failed in 1877 and 1878. It appears that failures were particularly heavy among mutual savings banks. Indeed, the FDIC Report24 shows 1878 to be the worst year for mutual savings banks in terms of loss rates between 1865 and 1934. The Boston mutual savings banks were especially hard hit in 1878. There were runs on a number of them and the Massachusetts legislature responded by

22 O . M . W . S p r a g u e , History of Crises under the National Banking System ( W a s h i n g t o n , D C , 1910). 23 T h e F D I C Report for 1934, w h i c h w e relied o n t o c o m p u t e t h e estimates o f d e p o s i t losses, is b a s e d

o n t h e C o m p t r o l l e r o f t h e C u r r e n c y ' s Annual Report (1896), p p . 52—7, for i n f o r m a t i o n c o n c e r n i n g n o n - n a t i o n a l b a n k losses.

24 F D I C , Annual Report 1934, p p . 1 1 2 - 1 3 .

CANADIAN AND AMERICAN BANKING 55

DO

C

1070 1880 1890 1900 1910 1920

Canada US all banks

Figure 3. Deposit losses, 1870-1925 (percentage of total deposits: Canada and US all banks)

simplifying the procedure by which mutual savings banks could temporarily suspend payments.25 In 1877 similar problems had afflicted savings banks in St Louis and Chicago. But, although losses were concentrated in certain cities, there were at least a few failures in a substantial number of states. The difficulties of the mutual savings banks may have been associated with the decline in commodity prices and the decline in railroad and utility stocks that occurred in 1877 and 1878. The entire episode deserves further study.

Table 1, which makes use of data from the FDIC's Annual Report for 1940, shows averages for several segments of 1865-1920. The estimates reported in 1940 may incorporate revisions of the estimates, particularly for non-national banks reported in 1934, which underlay Figures 2 and 3. (Unfortunately, while the Annual Report 1934 reported annual estimates, the Annual Report 1940 reported only averages for long periods.) The general picture, however, is similar. The Canadian system enjoys an edge for the whole period, but 1881 -1900 is an exception. Loss rates for Canada and for the United States banking system as a whole were similar over the period 1881-1900, but the loss rate for Canadian banks was higher than for United States national banks. The most important conclusion to be drawn from these comparisons is that losses on deposits and notes in Canada were on occasion quite large by 25 Bankers' Magazine, 32 (1877-78), pp. 826-7.

56 FINANCIAL HISTORY REVIEW

Table i. Losses

Years

1865-1880 1881-1900 1901-1920

on deposits; percentage of total deposits

Canada

o.oi a

0.16 O.OI

United States: national banks

0.06 0.08

O.OI

United States: all banks

0.21 0.15 0.05

Notes: aThis figure is for 1867 (Confederation) to 1880. In 1866 there was a major failure: the Bank of Upper Canada; if this failure were included, the Canadian average for 1865 to 1880 would be about 0.07. See R. M. Breckenridge, The History of Banking in Canada (Washington, DC, 1910), pp. 79-80, for a discussion of this failure.

Sources: Canada and United States national banks, see text. All United States banks: United States Federal Deposit Insurance Corporation, Annual Report (1940), p. 69.

United States standards. This presents a sharp contrast with the stability observed after 1925. W e turn now to a somewhat different aspect of stability: banking crises.

In the United States banking difficulties on five occasions (1873, 1884, 1890, 1893 and 1907 [Sprague dates]) led to severe stringencies in the money markets and on three occasions (1873, 1893 and 1907) to restrictions on the convertibility of deposits and notes with gold. These crises produced severe contractions in the stock of money and reductions in the quality of the stock of money (the ability of various components to perform as media of exchange) that in turn produced steep falls in the price level and contractions in economic activity. Canada, however, avoided this form of instability. Below we will consider the extent to which branch banking deserves the credit for this difference in stability.

II

T h e role of branch banking in mitigating instability in the banking system can best be understood by considering separately three classes of shocks: bank specific shocks, regional shocks and system-wide shocks.

With regard to bank specific shocks, bad luck may lead to a large percentage of assets of a single bank going bad at one time, and to the failure of the bank even as its neighbours remain solvent. T h e rules normally followed by prudent bankers, such as diversifying portfolios and demanding collateral for risky loans, are designed to minimise the probability of this type of failure, so that most failures of this type are likely to be the result of imprudent or fraudulent banking. Imprudent or fraudulent bankers were probably not more plentiful in Canada than in the United States. But the relative freedom of Canadian banks, compared with the United States nationals, to invest in a wide range of assets and to grow quickly to a large size by opening branches and acquiring other banks was hkely to make errors from mismanagement or fraud more costly.

Indeed, most of the major Canadian failures have been traced to gross misman-

CANADIAN AND AMERICAN BANKING 57

agement or fraud.26 The failure of the Mechanics Bank of Montreal in 1879 was attributed by Breckenridge27 to mismanagement and excessive note issue. This authority related the failure of the Maritime Bank of Canada in 188728 to its being a one-man bank that made excessive loans to a few favoured firms and that of the Banque Ville Marie in 189529 to gross fraud. The collapse of the Home Bank in 1923 was considered by Jamieson30 to be partly due to bad loans to the bank's directors. The failures of the Sovereign Bank of Canada in 1908 and the Farmer's Bank in 1910 were attributed by Jamieson partly to the attempt to add branches and attain an impressive size rapidly; a possibility, it is interesting to note, that was barred to United States banks.

Granted, there is a tendency to make management the scapegoat when a bank fails, and it is conceivable that in some cases managements of banks that failed did not act very differently from banks that survived. But the specificity of the criticisms suggests that in most cases management was to blame.

But why should the oligopoly of chartered banks that existed after the merger movement be entirely free of failures due to gross mismanagement or fraud? In other words, given that a fair number of Canadian banks failed before 1925 due to these causes, why should we not expect at least one or two failures owing to mismanagement or fraud after 1925? We suggest three possible answers.

First, the staff promotion pyramid within the large nationwide branch banks may have been so difficult to scale that incompetent managers were weeded out before they reached levels of management where their decisions could have endangered the entire institution, and may have discouraged short-sighted entrepreneurs from joining such banks in the first place. The large chartered banks, according to the conventional wisdom, would take on young men after graduation from high school and assign them to minor branches in remote areas of Canada. Those that did well would be given larger branches to manage and, after a number of years, would be brought back to the head office in Montreal or Toronto. After years of service they would eventually be given heavy responsibilities. In this way the banks developed a core of managers who felt they owed all of their success in life to their bank, and who responded with wholehearted institutional loyalty.

Second, as described by Jamieson,31 the large chartered banks developed a very sophisticated system, replete with fail-safe devices, to audit and control branches, and to make it difficult for individuals to do much damage; and third, the large chartered banks may have been lucky.

26 N e u f e l d , Financial System, p . 104, c o n c l u d e s t h a t 'loss o f c o n f i d e n c e in b a n k s a l m o s t always r e s u l t e d f r o m t h e i r h a v i n g m a d e i m p r u d e n t l o a n s a n d i n v e s t m e n t s o r f r o m suffering d e f a l c a t i o n s , a n d a l m o s t n e v e r f r o m e x t e r n a l forces o v e r w h i c h t h e b a n k s h a d n o c o n t r o l ' .

27 B r e c k e n r i d g e , History of Banking in Canada, p . 116. 28 ibid., p . 127. 2 9 i b i d . , p . 168. 30 A . B . J a m i e s o n , Chartered Banking in Canada ( T o r o n t o , r e v . e d n , 1955), p . 4 3 . 31 ibid., p a r t 11, passim.

58 FINANCIAL HISTORY REVIEW

The great strength of the Canadian banks was their ability to absorb regional shocks such as a decline in wheat prices that affected the western provinces, or the western United States. The Canadian banks were able to offset losses in one region with gains in another and to transfer reserves from head offices to the branches facing losses. Breckenridge comments on how the Canadian system handled a regional shock in 1882 in the following terms:

So serious were the losses there [Manitoba] ... because of the thoroughness with which the whole commercial community had been infected with the speculative virus, that three out of the seven Winnipeg managers were dismissed. Bad debts which would have swamped local banks, perhaps for all time, were taken care of by the Canadian banks which suffered them without other outward sign than reductions of capital, smaller additions to rest account, or lower dividends upon their stock.32

After 1900, even Canadian banks that did not possess nationwide branch networks were protected in some degree from regional shocks by the market for banks that existed in Canada. We discuss the legal change in 1900 that facilitated mergers in section IV. Consider a bank with branches in only one province. A regional shock would impose heavy short-term losses on the bank, and possibly alarm depositors. But the bank could then seek a merger with a larger bank with branches throughout the country. Although the larger bank would have to absorb some losses, it would acquire the branches, the goodwill and the local knowledge of the smaller bank. As Carr et al. show,33 this market worked extremely well in Canada and assured that only banks that were thoroughly rotten would have to close.

As an example, consider the Union Bank of Canada, which was acquired by the Royal Bank of Canada in 1925. The Union had 327 branches controlled from a head office in Winnipeg. It had taken a leading role in the development of western Canada, but had been struggling since 1923, when increases in its loan loss reserves were announced (probably for the same reason many United States banks were struggling — the postwar fall in world agricultural prices) and its management was afraid that the announcement of a large cut in dividends that seemed imminent might adversely affect the 'general standing' of the bank. Merger with a strong eastern bank was a reasonable solution.34 This was one of a series of mergers during a period marked by great uneasiness about the soundness of Canada's banks, and even runs, that might have degenerated into a panic.

In the United States, on the other hand, banking laws prevented the development of an interregional, and in many cases even an intrastate, market for distressed banks. A bank in Kansas suffering from loan losses and deposit withdrawals resulting from a decline in wheat prices could not look for salvation to a white knight from Chicago or New York.

By system-wide shocks, we mean an event that leads to currency withdrawals

32 B r e c k e n r i d g e , History of Banking in Canada, p p . 125—6. 33 C a r r et al., Ensuring Failure, p p . 27—36. 34 Jamieson, Chartered Banking, p. 68.

CANADIAN AND AMERICAN BANKING 59

throughout the banking system and may lead to a temporary suspension of con- vertibility. The event that triggers such a drain may be 'internal' - a shock to the domestic economy, such as the collapse of a major corporation, that damages the balance sheets of many banks and leads in turn to a scramble by the pubhc to convert deposits into currency. Or it might be 'external' - for example, a withdrawal of capital inflows from England prompted by an increase in the Bank of England's discount rate. In either case, branch banking played a role in protecting the Canadian system.

A major cause of banking crisis in the United States appears to have been the so-called 'pyramiding of reserves' in New York City. Country banks in the western and southern regions of the United States relied on correspondents in New York to provide reserves during times of heavy withdrawals.35 Presumably branch offices in western and eastern Canada relied upon head offices in Toronto and Montreal in the same way. But the commitment of a New York bank to its independent correspondents in the south and west was less strong and less visible to the depositor, than the commitment of head offices in Canada to their western and eastern branches. Moreover, on several such occasions the correspondents, having their assets tied up in a collapsing stock market, were unable to deliver. The holder of deposits in western and southern banks in the United States was therefore more prone to run to cash at the first sign of trouble, than his opposite number in Canada.

Not all of the relative financial instability in the United States, however, can be laid at the door of limited branching. One ongoing problem was the continuing debate over cheap money, particularly silver. The threat that the silver forces would gain the upper hand in the United States and force the United States off the gold standard may have made the system-wide shock in 1893 more severe.

In any event, it does appear that the United States economy suffered severely on the three occasions — 1873, 1893 and 1907 — when system-wide shocks produced a restriction of convertibility. The basic data are given in Table 2. Real and nominal GNP fell significantly with each crisis in the United States, and in each case the fall was larger than the corresponding fall, if any, in Canada.36 Indeed, during the first two episodes real GNP actually rose in Canada. Two factors seem to have been at work during the United States financial crises. First, as argued by Friedman and

35 M. D. Bordo, P. Rappoport and A. J. Schwartz, 'Money versus credit rationing: evidence for the national banking era, 1880-1914', in C. Goldin and H. Rockoff(eds), Strategic Factors in Nineteenth Century American Economic History: a Volume to Honor Robert W. Fogel ( C h i c a g o , 1992).

36 We have relied on N. S. Balke and R. J. Gordon, 'The estimation of pre-war gross national product: methodology and new evidence', Journal of Political Economy, 97 (1989) for estimates of GNP, rather than C. Romer, 'The pre-war business cycle reconsidered: new estimates of gross national product, 1869-1908\ Journal of Political Economy, 97 (1989), because the former appear to be closer methodologically to estimates of GNP for Canada in M. C. Urquhart, 'New estimates of gross national product, Canada, 1870-1926: some implications for Canadian development', in S. L. Engerman and R . E. Gallman (eds), Long-Term Factors in American Economic Growth, 51 (Chicago, 1986).

60 FINANCIAL HISTORY REVIEW

Table 2. The Canadian and United States economies during financial crises (percentage change)

1873-74 ^ 9 3 - 9 4 1907-08

Real GNP United States Canada Nominal GNP United States Canada Money (M2) United States Canada

-0.63 2.01

- 3 - 5 5 - 0 . 4 7

1.83 - 3 - 7 5

- 2 . 9 6

4-93

-8.42 -4.65

0.47

3-27

- 5 . 6 2 - 5 . 1 2

-7-77 - 4 . 4 1

- 1 - 3 9 13-97

Sources: United States: Real and Nominal GNP: N . S. Balke and R. J. Gordon, 'The estimation of prewar Gross National Product: methodology and new evidence', Journal of Political Economy, 97 (1989), p. 84; M2: M. Friedman and A. J. Schwartz, Monetary Trends in the United States and the United Kingdom: Tlieir Relation to Income, Prices and Interest Rates, 1867-1975 (Chicago, 1982), p. 122. Canada: Real and Nominal GNP: M. C. Urquhart, 'New estimates of Gross National Product, Canada, 1870-1926: some implications for Canadian development', in S. L. Engerman and R. E. Gallman (eds), Long-term Factors in American Economic Growth, 51, Conference on Income and Wealth (Chicago, 1986), p. 30; M2: M. D. Bordo and L. Jonung, The Long-run Behaviour of the Velocity of Circulation (Cambridge, 1987), pp. 154-5.

Schwartz,37 the banking difficulties produced a decline in the money multiplier which reduced the growth rate of the money supply, and hence aggregate demand. Second, as argued by Grossman,38 the financial crisis in the United States disrupted the credit allocation mechanism.39

There are some anomalies in Table 2. The rise in Canadian real GNP between 1873 and 1874, despite the fall in the stock of money and despite the financial

37 F r i e d m a n a n d S c h w a r t z , Monetary History. 3 8 R. S. Grossman, 'The macroeconomic consequences of bank failures under the national banking

s y s t e m ' , Explorations in Economic History, 3 0 ( 1 9 9 3 ) . 3 9 The implications of the differences in susceptibility to system-wide shocks for long-run macroecon-

omic stability, however, are unclear. On one hand G. Rich, Tlie Cross of Cold: Money and the Canadian Business Cycle, 1867-1913 (Ottawa, 1988), p. 157, concluded that 'Canadian GNP fluctuated less than its U.S. counterpart'. On the other hand, S. D. Williamson, 'Implications on financial intermediaries and implications for aggregate fluctuations: Canada and the United States 1870-1913', in O. J. Blanchard and S. Fischer (eds), NBER Macroeconomics Annual, 1989 (Cambridge, Mass., 1989), p. 332, relying on more recent estimates of US GNP (due to Romer, 'Prewar business cycle' and Balke and Gordon, 'Estimation of prewar Gross National Product'), and a different method for detrending the data, found, depending on the measure of US GNP used, that Canadian GNP was 11% or 56% more volatile than US GNP, and that the GNP deflator was 9% or 54% more volatile. Indeed, Williamson concludes that branch banking, the absence of reserve requirements on deposits and bond backing requirements on notes in Canada produced greater sensitivity to real shocks in Canada.

CANADIAN AND AMERICAN BANKING 6l

Table 3. Bank balance sheets, Canada and the United States, 1870-1919

Ratios 1870-79 1880-89 1890-99 1900-09 1910-19

Canada Loan: asset 0.717 0.706 0.696 0.722 0.640 Security: asset 0.013 0.021 0.071 0.087 0.110 Debt: equity 1.458 1.914 2.796 4.232 6.876

United States Loan: asset 0.487 0.563 0.589 0.546 0.567 Security: asset 0.253 0.169 0.117 0.164 0.168 Debt: equity 1.826 2.334 2.620 4.184 5-352

Sources: United States: United States Comptroller of the Currency, Annual Report; Canada: C. A. Curtis, 'Banking statistics in Canada', in Statistical Contributions to Canadian Economic History (Toronto, 1931).

distress to the south, is somewhat surprising. But note that over the whole cycle, 1873 to 1878, real GNP in Canada fell about four per cent. The very rapid growth of money in Canada between 1907 and 1908 is also surprising, and can be accounted for by the lender of last resort operations described in the next section, and to a large gold inflow that produced an increase in high-powered money sufficient to more than offset declines in the deposit-reserve ratio and the deposit-currency ratio.

The three episodes explored above are famous crises in the United States, but they were also international in scope so they should have had some impact in Canada. Nevertheless, it is fair to ask whether we would get similar results if we began with periods of distress in Canadian banking. The failure of the Home Bank in 1923 and the subsequent anxiety about the banking system comes closest to a banking crisis in Canada. According to Jamieson:40 'This [anxiety about the safety of the banks] was reflected in the dissemination of rumours, some of the wildest nature, and sporadic "runs" by depositors, which only a few of the very strongest banks escaped.' Several mergers of weaker with stronger banks followed in the wake of the failure of the Home Bank. But, while nominal GNP fell 1.19 per cent in Canada between 1923 and 1924, real GNP rose, although by an admittedly anaemic 0.73 per cent, and the stock of money in Canada rose 3.77 per cent. In the United States, by way of contrast, nominal GNP rose 2.64 per cent, real GNP rose 2.62 per cent and the stock of money rose 5.27 per cent. Evidently this episode does not compare in intensity with the more serious crises in the United States.

Overall, the conclusion that branch banking helped the Canadian banks cope with system-wide shocks seems to be well established. This conclusion is supported,

Jamieson, Chartered Banking, p. 65.

62 FINANCIAL HISTORY REVIEW

moreover, by Calomiris41 whose broad international survey of evidence on the relationship between branch banking and stability includes Canada. It is possible, however, to imagine system-wide shocks that are too big to be absorbed even by a system of large banks with nationwide branch networks. A lender of last resort, which we discuss in the following section, may be needed.

I l l

The Canadian banking system developed a lender of last resort only slowly and by the end of our period was still without either a central bank or deposit insurance. Yet an institutional structure had evolved that contained many of the features of a lender of last resort. This gradual evolution reflected the many protective barriers that insulated the Canadian banking system from such shocks. First, Canada was a small, open economy and the Canadian banks held deposits in New York as a form of contingent reserves. In addition, the asset-backed note issue of the Canadian banks created an elasticity of the Canadian currency which permitted an instan- taneous switch between deposits and notes. The significance of this elasticity became clearer when it was reduced in 1907.

The first occasion when the government acted as lender of last resort occurred in 1907. The ability of the government to do so rested on its issues of Dominion notes. These notes were legal tender notes issued by the federal government. The govern- ment was required to redeem them on demand and held 25 per cent reserves against issues up to C$30 million and 100 per cent gold reserves against any issues in excess of C$30 million.42 In 1907, at the instigation of the public (rather than the chartered banks), the Canadian government made an emergency issue of Dominion notes which it lent to the banks. The emergency issue reflected concern, primarily amongst farmers, of an incipient credit squeeze. The traditional explanation for the squeeze has been that the restriction of bank note issues to an amount less than paid up capital had become binding - a story which is consistent with the passage of legislation in 1908 that permitted the banks to increase their note issue to 115 per cent of paid in capital during the crop-moving season. However, Rich43 has argued that the crisis was brought about by the Canadian banks shifting funds to New York to take advantage of the very high interest rates prevailing there during the autumn of 1907, and restricting domestic credit to do so. In any case, the emergency Dominion note issues of 1907 represented the first time that the Canadian govern- ment had intervened to supply liquidity to the banking system. In 1914, at the onset of the First World War, the government passed the Finance Act which created a

41 C. Calomiris, 'Regulation, industrial structure, and instability in U.S. banking: an historical perspec- tive', in M. Klausner and L. J. White (eds), Structural Change in Banking (Homewood, 111., 1993), pp. 33-8.

4 2 R . C . M a c l v o r , Canadian Monetary, Banking and Fiscal Development ( T o r o n t o , 1961), p . 6 5 . 4 3 G . R i c h , ' C a n a d i a n b a n k s , g o l d , a n d t h e crisis o f 1 9 0 7 ' , Explorations in Economic History, 26 ( 1 9 8 9 ) .

CANADIAN AND AMERICAN BANKING 63

discount window for banks. It is clear, then, that by 1907 the Canadian government had begun to play a modest role of lender of last resort.

A more difficult question to answer is whether the Bank of Montreal, or 'the club' of large Canadian Banks, had also begun to play this role. In 1906 the assets and liabilities of the Ontario Bank were assumed by the Bank of Montreal, with other banks giving a guarantee, and in 1908 the assets and liabilities of the Sovereign Bank were taken over by a group of 12 banks including the Bank of Montreal. In testimony before the United States National Monetary Commission, Sir Edward Clouston, General Manager of the Bank of Montreal, was willing to agree that his bank had acted as a lender of last resort in these cases: 'in the case of the Ontario Bank and Sovereign Bank it was a very ticklish time, and if that run had been allowed to continue it might have spread, and it was done partially in self defence'.44

Carr et al., however, stress that the take-overs helped smooth the 'transfer of business' from the insolvent banks, although they also note the concern about 'externalities' associated with the closure of these banks.45 Ultimately, the point may be that there was a recognition on the part of the Bank of Montreal and the other large banks that they could be hurt by a general decline in confidence sparked by a bank failure, but this consideration could never be divorced from an analysis of the effect of a take-over on the profits of the banks.

The United States did not have a formal lender of last resort before the establish- ment of the Federal Reserve in 1914. A set of market-driven and official arrange- ments evolved to provide liquidity to the banking system in times of panic. These arrangements proved successful in allaying panics on several occasions - in 1884, 1890 and 1900 - but not on others.

One such arrangement was the issue of clearinghouse loan certificates, whereby the clearinghouses of New York, Chicago and other central reserve cities issued emergency reserve currency in the form of clearinghouse loan certificates col- lateralised by member bank assets, and even issued small denomination certificates that circulated from hand to hand as currency.46 A second was operations by the independent Treasury which on occasion conducted rudimentary monetary policy. The Treasury in times of stringency would use its powers as depository of fiscal revenues to transfer deposits to key commercial banks, speed up debt redemption and to arrange syndicates of private investors to provide timely liquidity.47 Finally, 44 U n i t e d S t a t e s N a t i o n a l M o n e t a r y C o m m i s s i o n , Interviews on the Banking and Currency Systems of

Canada ( W a s h i n g t o n , D C , 1 9 1 0 ) , p . 1 8 1 . 45 C a r r e t al., Ensuring Failure, p p . 21—2. 46 R. C. Timberlake Jr, 'The central banking role of clearinghouse associations', Journal of Money,

Credit and Banking, 16 (1984); G. Gorton, 'Clearinghouses and the origins of central banking in the U.S.', Journal of Economic History, 45 (1985); and G. Gorton and D. J. Mullineaux, 'The joint production of confidence: endogenous regulation and 19th century commercial bank clearing- h o u s e s ' , Journal of Money, Credit and Banking, 19 ( 1 9 8 7 ) .

47 R. C. Timberlake Jr, The Origins of Central Banking in the United States (Cambridge, 1987). Some states developed early deposit insurance schemes, but none were successful in preventing major panics; C. Calomiris, 'Is deposit insurance necessary? A historical perspective', Journal of Economic History, 50 (1990).

64 FINANCIAL HISTORY REVIEW

Table 4.

1895 1900

1905 1910 1915 1920

1925 1927

Canadian bank concentration measures, 1895—1927

Percentage of total bank

Top 3 banks

34-19 37-8i

37-39 38.08 44.14

51-95 65.90

68.61

Top 5 banks

43-85 48.46

46.03 48.56

55-89 66.11 78.91

81.99

assets

Top 10 banks

64.07

69.08

68.64 69.70 78.86 86.51

98.31 99.78

Herfindahl index

0.0648 (15)

0.0754 (13) 0.0775 (n) 0.0786 (13) 0.0891 (11) 0.1128 0.1656

0.1793

Source: B. H. Beckhart, 'The banking system of Canada', in H. P. Willis and B. H. Beckhart (eds), Foreign Banking Systems (London, 1929), pp. 330-3.

after the disastrous experience of the panic of 1907-08, the Aldrich-Vreeland Act was passed in 1908, which allowed groups of national banks to form a National Currency Association empowered to issue emergency currency based on members' collateral of commercial paper. Although only invoked once, it was successful in preventing panic at the outbreak of the First World War in 1914.48

IV

As we have had occasion to note above, the Canadian banking system was trans- formed by mergers and acquisitions between 1910 and 1925, ushering in a period of impressive stability. The causes of this consolidation are not completely understood, but there is general agreement that the Bank Act revision of 1900, which permitted a bank to acquire the assets of another bank without a special Act of Parliament, was at least a proximate cause. Prior to 1900 a merger was a 'time-consuming and frequently cosdy affair'49 while after 1900 approval of the Minister of Finance was sufficient to permit a merger.

Table 4 shows the percentage of total bank assets held by the top three banks, the top five banks, and the top ten banks by quinquenia. It also shows the Herfindahl index. The most dramatic change in the percentage of assets held by the largest firms is in column three: the percentage of assets held by the top ten firms. This was already high by American standards in 1895, 64 per cent, but by 1927 it was close to 100 per cent. The Herfindahl index is the sum of the squared shares, and handles mergers particularly well. The number in parentheses is the inverse of the

4 8 F r i e d m a n a n d S c h w a r t z , Monetary History, p p . 1 7 0 - 2 . 49 D. Bond, The merger movement in Canadian banking 1890-1920, some preliminary findings,

University of British Columbia Discussion Paper, 21, p. 5.

CANADIAN AND AMERICAN BANKING 65

Herfindahl index to the nearest whole number. It is the number of equal size banks that would yield the same Herfindahl index as the actual distribution. Between 1910 and 1925 the equal-bank-equivalent declined from 13 to six. By way of contrast, consider the United States in 1920. The three largest banks were in New York City: Chemical with 4.13 per cent of total national bank assets, National City Bank with 3.85 per cent and Chase with 2.41 per cent, for a total of 10.40 per cent of national bank assets which amounted to about 5.09 per cent of all commercial bank assets - a far cry from the 51.95 per cent of bank assets owned by the three largest Canadian banks.50 The Canadian banking system had taken its modern form.

The last failure among Canada's chartered banks was the failure of the Home Bank in 1923. By 1925 all of the Canadian banks, except two based in Quebec, had nationwide branch networks able to withstand the severe shocks that would hit the system in the 1930s.

V

Although stability is important, it cannot be the sole basis for judging a banking system. We also want to know how efficient the system is in providing loans and other basic banking services. In our earlier paper focusing on the period 1920-80, we measured efficiency by comparing for the two countries rates of interest charged on loans and paid on deposits and received on securities, and the overall rate of return on equity. For the earlier period now under review such comparable data for the two countries are not available so we are forced to use cruder proxies.

Table 3 brings together decadal averages of some key ratios for the Canadian chartered banks and the United States national banks.51 The loan: asset ratio during this period, we would argue, measures the ability of the banking system to supply loanable funds to the private sector. Evidently, the loan: asset ratio was consistently higher in Canada, in part because the security: asset ratio was consistently lower, although the differences narrowed somewhat during the decade of the First World War. The United States ratio was lower primarily because of regulations that allocated bank funds to the government or to suppliers of outside money in order to increase the safety of the system. The security: asset ratio was higher in the United States because of the bond security system instituted to protect national bank notes, and because national banks were required to purchase government bonds to secure their charters. The reserve ratio was also higher in the United States because of the required reserve ratios instituted to protect deposits.

The last line of each panel of the table is the debt (primarily deposits): equity ratios of the banking systems. In our previous paper we showed that the debt: equity ratio was consistently higher in Canada. We interpreted this as a reflection, ulti- mately, of the greater stability of the Canadian system: neither regulators nor

5 0 Annual Report of the Comptroller of the Currency 1920, 2 , p p . 156, 6 2 1 . 5 1 C a l o m i r i s , ' R e g u l a t i o n ' , p p . 5 8 - 6 2 , for a c o m p l e m e n t a r y d i s c u s s i o n o f t h e s e r atios.

66 FINANCIAL HISTORY REVIEW

1090 1900 1910

Canada ---US nationals

Figure 4. Rates of return to equity (Canada and the United States, 1890—1913)

depositors showed much concern when Canadian banks raised their debt: equity ratios. Between 1870 and 1919, interestingly, the situation was very different. In the first two decades the debt: equity ratio was actually slighdy higher in the United States, and a substantial difference in favour of Canada did not arise until the decade of the First World War. This picture is consistent with our earlier argument, that the Canadian system was not notably more stable than the United States national system and was not permitted notably higher debt: equity ratios.

Figure 4 compares the rate of return on equity in Canada (defined as the ratio of dividends and change in surplus to equity) with the rate of return on equity in the United States (defined as the ratio of net profits to equity) between 1890 and 1913. As is evident the two move on similar paths. This suggests that, despite the higher proportion of their portfolio that Canadian banks were able to devote to lending, the banking market was sufficiently competitive so that Canadian banks were unable to earn greater profits. This result contrasts with our findings for the post-1929 period. For that period we found that the rate of return on equity was generally higher in Canada. Admittedly, we used a different definition of rate of return on equity for the latter period so our results are not stricdy comparable. But assuming that the second period was different, the explanation would turn on the achieve- ment of stability. Stable banking permitted Canadian banks to increase their lever- age, which in turn produced high returns to equity after 1925.

In summary, a comparison of the balance sheets and earnings of the banking

CANADIAN AND AMERICAN BANKING 67

systems reinforces the basic picture: a Canadian system which was roughly on a par with the United States system at the start, but gradually pulling ahead in terms of safety and efficiency.

VI

Before 1925 the difference in stability between the Canadian and United States banking systems was smaller than it would be afterwards. True, measured by suscep- tibility to banking crises, the Canadian system was clearly stronger. But, measured by losses experienced by depositors and noteholders, the difference becomes prob- lematic; Canadian banks were more stable than United States non-national banks, but less stable than United States national banks.

Depositors and noteholders in United States banks experienced some losses in virtually every year between 1870 and 1925, while the Canadian system often escaped unscathed. But when a Canadian bank failed, the resulting losses often represented a relatively high percentage of total bank deposits. On two occasions, moreover, noteholders were not paid in full and on others noteholders had to wait an extended time to get their money. In the United States immediate redemption at par was assured by the bond security system of the national banks.

The Canadian banks were never forced to restrict convertibility of notes or deposits into high-powered money during international banking crises, while the United States banking system was forced to do so in 1873, 1893 and 1907. It should be noted, however, that it is not certain that the two systems faced system-wide shocks of the same intensity. For example, in the 1890s the political success of the free silver forces in the United States raised doubts about the commitment of the United States to the gold standard, a difficulty that the Canadian system was not forced to face.

The merger movement in Canada, however, radically altered the picture, by creating a banking system characterised by a small number of large banks with thick interprovincial networks of branches. From 1925 on, as we have shown in our previous paper, the Canadian banking system exhibited remarkable efficiency as well as stability. The key lesson from the present analysis is that the transition process was protracted and there were times when losses from bank failures were heavy. On some occasions imprudent entrepreneurs took advantage of the freedom to branch and merge by expanding their institutions too rapidly. To judge from the contrast between Canadian and United States banking in the period 1925 to 1980, however, it would have been a mistake to cut short the evolutionary process by imposing a system of highly restricted unit banks, such as the national banking system, on Canada, a proposal strongly advanced at the time.

The Canadian banking experience suggests that once an equilibrium in this industry was achieved, it enjoyed both stability and efficiency. However, the experi- ence of the United Kingdom and possibly other countries suggests a different outcome. In the United Kingdom case once the merger movement had largely

68 FINANCIAL HISTORY REVIEW

worked itself out by 1918, the resultant 'big five' banks formed a cartel, with the government's tacit assistance, which according to recent studies became very inefficient.52 Protected by exchange and capital controls and legal restrictions on potential competition by other financial intermediaries, it took a major change in legislation in 1971 (Competition and Credit Control Act) and the removal of external controls to rectify the situation. One can speculate that the Canadian banking system did not go the British route because of less intrusive government intervention; because of the proximity of the United States financial system; and because of the absence of extensive controls on capital movements.

These historical comparisons, we believe, have important messages for contem- porary policy-making. Currently, the United States is moving rapidly towards removing the centuries-old barriers to interstate branching. Once the forces of competition are unleashed, it is our belief that the United States banking system will follow a route similar to that taken in Canada during the first decades of this century, and in the United Kingdom somewhat earlier. Large banks will acquire smaller banks across states and regions, creating a much more concentrated banking system than now exists. In the process mistakes will be made, as happened in Canada, and large institutions may become insolvent. It will then be the responsi- bility of the monetary authorities to act to protect the payments system at large, but not to cut short the evolutionary process with new restrictions. Also, provisions will be required to protect small depositors from losses. Although the current system of deposit insurance does accomplish this, other serious problems associated with it suggest that other solutions may be required.

Our historical comparison of the two North American banking systems may also have lessons for the newly emerging countries of the former Communist bloc and for developing countries. It suggests that prohibitions against nationwide branch banking for whatever reason is a mistake. The benefits of long-run stability and efficiency outweigh the costs of concentrated economic power. The latter can be dealt with by permitting competition from foreign banking systems and domestic and foreign non-bank financial intermediaries. As in the Canadian case, it may take time to develop extensive branch networks - to find and train trustworthy and competent employees. In the interim, if unit banking is the prevailing norm, existing institutions should be encouraged to develop correspondents and ultimately branch networks.

52 F. Capie, 'Prudent and stable (but inefficient?): commercial banks in Britain', and L. Hannah, 'Effects of banking cartels', both in M. D. Bordo and R. Sylla (eds), Anglo-American Fitiaticial Systems: Institutions and Markets in the Twentieth Century (New York, 1995).