Banking Crises Theory And History

The theory and history of banking crises developed interconnectedly and almost simultaneously. It seems probable that the reason for this was that the theory developed in the nineteenth century, and was developed largely by those who had been or still were practical bankers, or, most notably in the case of Walter Bagehot, commentators on current financial arrangements and events.

Certainly until this century, only problems in the banking system were seen as financial crises. Crashes of financial or other asset markets were the consequence of prior 'manias', a result of human gullibility and folly, a proper subject of study by the disinterested observer but not requiring any policy action. This attitude is vividly summarized in the title of Charles McKay's (1845) classic, Extraordinary Popular Delusions and the Madness of Crowds.

Crises in the banking system, however, were regarded as serious, even dangerous, occurrences. Anna Schwartz (1986, p. 11), in a recent statement of this view, described such events as 'real' crises. 'Such a crisis is fuelled by fear that means of payment will be unavailable at any price, and in a fractional reserve banking system leads to a scramble for high-powered money.' In contrast to these 'real' crises are 'pseudo' crises. These involve 'a decline in asset prices, of equity stock, real estate, commodities, depreciation of the exchange value of the national currency; financial distress of a large non-financial firm, a large municipality, a financial industry, or sovereign debtors' (p. 24). Such loss of wealth causes distress, but is not in itself a financial crisis. A 'pseudo crisis' is simply an unusually large case of mistaken investment, and mistaken investments are inevitable in an uncertain world. A 'real' financial crisis is when the stability of the whole banking system is threatened.

Such 'real' crises have been quite rare, although more prevelant in recent times (Scandinavia, Japan, and in some developing countries such as Venezuela). Certainly, on this definition, episodes frequently described as international crises - such as Latin America in the early 1980s and Russia in 1998 - were not crises outside these countries, and were not always crises in the sense in which Schwartz uses the term inside these countries.3

According to Schwartz no such crisis has occurred in Britain since 18664 and in the United States since 1933.

Schwartz's definition says nothing of the cause of the crisis. In that regard, as in the definition itself, it follows in the tradition of Thornton (1802) and Bagehot (1873). Their approach aimed at clarifying a problem and then going on to propose a solution. A 'real' crisis in the Schwartz sense is dangerous because it can lead to an unanticipated and undesired collapse in the stock of money, and such an unanticipated squeeze will cause a recession, perhaps a depression. The monetary squeeze is produced both by a fall in the money multiplier (as cash shifts from the banking system to the public) and as bank deposits fall. To prevent this squeeze, Thornton, Bagehot and Schwartz, and other writers in this tradition, suggested the following course of action.

The central bank of whichever country experiences such a shock should lend freely on collateral. It should not restrict lending to the classes of security (usually quite narrow) that it would accept for discount in normal times. Advances should be made without limit, on demand, but at a rate of interest above the pre-crisis rate. These loans should be made to the market - that is, to anyone who brings in acceptable security. In addition (and argued in particular by Bagehot) it should be made clear that the central bank will act in that way should there ever be a crisis: this reduces the likelihood of runs because knowledge that the central bank will supply liquidity makes it seem less urgent to scramble for it.

What can trigger such a 'real' crisis? Palgrave (1894), under the heading 'Crises, commercial and financial', provides first a definition of crises and then a description of the development of several nineteenth-century crises: 'Times of difficulty in commercial matters are, when pressure becomes acute, financial crises.' His description of the events of 1825 is a good example:

The next serious crisis occurred in 1825, one of the most severe through which the commercial and banking [emphasis added] systems of the country had ever passed. At this date speculation ran very high, for the most part in loans and mining adventures, and other investments abroad. The foreign exchanges were so much depressed as to be the cause of a nearly continuous drain on the bullion of the Bank. Many and heavy banking failures, and a state of commercial discredit, preceded and formed the earlier stage of the panic. The tendency to speculation, and the undue extension of credit, was preceded, probably caused, and certainly favoured and promoted, by the low rate of interest which had existed for some time previously; and this low rate of interest was apparently prolonged by the operations of the Bank of England. (Palgrave, 1894, p. 457)

Palgrave gives several examples of such chains of events, and refers to Tooke (1838) and to Levi (1888) as providing many more details.

To summarize so far, the view developed in the nineteenth century, and restated in the twentieth by Schwartz and others, is that crashes in financial markets are not in themselves crises. They can lead to runs on the banking system, and thus produce 'real' crises. One can lead to the other by starting a scramble for liquidity.5 But to quote Palgrave (1894, p. 457) again, 'Commercial crises may take place without any reference to the circulating medium as has been exemplified in Hamburg and elsewhere.'

Were - and are - such crises random events? The famous Diamond and Dybvig (1983) model of a banking crisis would certainly suggest that they are. That model sees crises as being triggered by an unpredictable random shock - a sunspot theory of crises. There is of course criticism of that famous paper. But rather than simply looking at these, it is helpful to consider whether the classic banking crises they model actually were random. There is a good amount of evidence on this, examining both the seasonal and the cyclical pattern of crises.

Seasonal regularity has been noted both by present-day and by nineteenth-century writers. Miron (1986, pp. 125-40) observes that financial panics in the nineteenth century displayed a seasonal pattern in both Europe and the United States. For US data, a x2 test rejects at the 0.001 per cent confidence interval the hypothesis that crises were distributed randomly across the seasons. Examination of Kemmerer's (1910) listing of 29 banking panics between 1873 and 1908 shows 12 to be in the spring (March, April, May) and a further ten in September or December.

Kindleberger's (1978) enumeration of panics in Europe between 1720 and 1914 yields similar results. A x2 test rejects the hypothesis of uniform distribution across the year; as in the United States, a marked preponderance fell in the spring or the autumn (Miron, 1986).

The still accepted explanation for this seasonality is that given by Jessons (1866). He observed a seasonal pattern in interest rates associated with the agricultural cycle in asset demands. Reserve/deposit ratios for banks fell in the spring and the autumn when there was a seasonal upturn in the demand for both currency and credit. So it was in spring and autumn that banking systems were at their most vulnerable. The seasonal pattern in interest rates largely vanished when central banks started smoothing the interest rate cycle.6

Was there a cyclical as well as a seasonal regularity in the occurrence of crises? Seasonal regularity certainly need not imply cyclical regularity; it is necessary to look afresh at the data. Numerous writers in the nineteenth century noted the regularity with which 'commercial crises' occurred. Palgrave (1894) lists 1753,1763, 1772-73, 1783,1793, 1815,1825, 1836-39, 1897, 1866, 1875 and 1890. 'An examination of recorded years of acute commercial distress suggests periodicity. During the 140 years trade and banking have been carried on in war and peace, with a silver standard, with a gold standard, under a suspension of cash payments, in times of plenty, and in times of want; but the fatal years have come round with a considerable approach to cyclical regularity' (Palgrave, 1894, p. 466). Periodicity was also remarked on by Langton (1858), Jessons (1866), Mills (1868), and Chubb (1872), among others.

Table 15.1 shows the main 'commercial crises' listed by Palgrave (1894), with a brief description of each from a contemporary or near-contemporary source. Of these 'commercial crises', only that of 1847 was not at a business cycle peak according to the Burns and Mitchell (1946) chronology of British business cycles. This appears to suggest a close association with subsequent recessions.

Unfortunately it is not clear how much weight that finding can bear. The reason for this arises from how Burns and Mitchell determined their cyclical chronology. Their method was to inspect a large number of series for different aspects of the economy, and, on the basis of this inspection, reach a judgement on the length of each cycle. This causes a problem because one of the series they examined was for financial panics and they regarded the occurrence of one of these as suggesting the economy was at or around a business cycle peak. Further doubt about the finding is created by comparing the Burns and Mitchell chronology with that which Capie and Mills (1991) developed by estimating a segmented trend model. Capie and Mills's work goes back only to 1870, but between that year and 1907 (their last peak) they differ from Burns and Mitchell in the timing of three (out of five) peaks and one (out of five) troughs.

Turning next to the United States, analysis has been carried out by Gary Gorton (1988). His approach was to consider whether banking crises ('real' crises) are random events, 'perhaps self-confirming equilibria in settings with multiple equilibria', or alternatively, whether they were systematic, linked to 'occurrences of a threshold value of some variable predicting the riskiness of bank deposits' (p. 751). Table 15.2 provides the basic data for the early part of the period in the United States. Crises were usually at business cycle peaks, but were not by any means at every business cycle peak.

Deposit behaviour changed after 1914 (the founding of the Federal Reserve) and again after 1934 (the start of deposit insurance), but despite that, crises remained systematic, linked to the business cycle. 'The recession hypothesis best explains what prior information is used by agents in forming conditional expectations. Banks hold claims on firms and when firms begin to fail, a leading indicator of recession (when banks will fail), depositors reassess the riskiness of deposits' (Gorton, 1988, p. 248).7

It might be thought that Gorton's result gives support for the similar

Table 15.1 'Commercial crises', eighteenth and nineteenth centuries

Year(s) Commercial crises







1866 1890

Followed 'investments in machinery and in land navigation . . . Many houses of the most extensive dealings and most established credit failed'(McPherson, 1805)

'Severe pressure in the money market, extensive failures of banks in the North of England, great mercantile discredit' (Palgrave, 1894, p. 43)

'[T]he fall of prices, the reduction of private paper, and the destruction of credit, were greater and more rapid than were before, or have since, been known to have occurred within a short space of time' (Tooke, 1838, p. 10)

'[O]ne of the most severe through which the commercial and banking systems of the country had ever passed' (Palgrave, 1894, p. 168). Palgrave quotes Huskinson as saying 'that we were within a few hours of a state of barter'. This crisis followed a steady and very substantial fall in yields. Palgrave gives the price of 3% public funds.


3rd April

732 (the lowest)

1st July


3rd October



1st January


2nd April


28th April

974 (the highest)







14th February


'Though the period of pressure in 1825 was so short, it had been preceded by considerable and extravagant speculations in foreign loans and shares of companies, mining and commercial' (Palgrave, 1894, p. 180)

'[A] considerable period of speculative activity, fostered by a low rate for money, preceded this crisis also' (Palgrave, 1894, p. 200).

Palgrave also notes that Peel's Bank Act came into effect on 2

September 1844, and this 'took away from the directors [of the Bank of England] alike any power or any responsibility for the "regulation of the currency" so far as this consisted of their notes'. Interest rates fell, and there was, as in 1825, considerable speculation, this time in

'railways and other improvements at home' (p. 202)

'[I]t is very clear that during the years 1855 to 1856 the extension of credit was enormous and dangerous ... [In 1857] the reserve of the

Bank of England may be said to be continually at a danger point . . .'

Failure of Overend, Gurney and Co.

Baring crisis

Table 15.2 National banking era panics

National Bureau of Economic Research cycle Peak-Trough

Crisis date

Oct 1873-Mar 1879

Sept 1873

Mar 1882-May 1885

Jun 1884

Mar 1887-Apr 1888

No panic

Jul 1890-May 1891

Nov 1890

Jan 1893-Jun 1894

May 1893

Dec 1895-Jun 1897

Oct 1896

Jun 1899-Dec 1900

No panic

Sep 1902-Aug 1904

No panic

May 1907-Jun 1908

Oct 1907

Jan 1910-Jan 1912

No panic

Jan 1913-Dec 1914

Aug 1914

Source: Gorton (1988).

Source: Gorton (1988).

Palgrave/Burns and Mitchell view. But caution is necessary, for the structures of the banking systems of the two countries differed (and still differ) greatly. The importance of banking structure is discussed subsequently.

It would appear, therefore, that there is sufficient regularity in crises - regularity in the sense of their being associated with prior causal or at the least facilitating events - as to imply that it is a useful exercise to look for underlying causes and for ways of preventing crises. They are not, in other words, random events - such as meteor strikes or sunspots - over which we have no control.

The above does of course refer to classic liquidity crises. Banking crises in even the comparatively recent past were the consequence of a shortage of liquidity, not a shortage of capital. Even in the Great Depression in the United States, this was true. Now, this certainly represents a marked contrast with some recent events - notably the Japanese banking crisis and that in the Scandinavian economy.

Why shortage of capital emerged as a mid-twentieth-century problem is in large part a consequence of aspects of regulation and governance and is explained in some detail in the following section of this chapter.

Financial End Game

Financial End Game

How to profit from the global crisis and make big bucks big time! The current global financial crisis has its roots embedded in the collapse of the subprime markets in the United States. As at October 2007 there was an estimated loss on the subprime market of approximately 250 billion. If you want to come out on top, you have come to the right place.

Get My Free Ebook

Post a comment