On the So-Called ‘Failure’ of Free Banking in Australia : the Banking Crisis of 1893

The anti-free bankers sometimes rehash Hickson and Turner’s paper “Free banking gone awry: the Australian banking crisis of 1893” (2002) as a refutation of the theory of free banking while ignoring the many other successful episodes. In fact, some evidence provided by the paper contradicts their conclusion that the root cause of the 1893 crisis stems from the fact that “Australian banks opportunistically diverted dramatically increased deposits into overly risky assets without corresponding increases in equity capital”. Hickson and Turner ironically support the evidence brought by Selgin that the boom is caused “not by any increase in the bank money multiplier, but by injections of high-powered money from Austra-lian gold mines and from the British capital market”.

The figures 2 and 3 they provide did not even show a decrease in shareholders’ capital to deposits ratio and liquidity ratio during the 5 years, or even the 10 years before the 1893 crisis occurred. Hickson and Turner make the same error as Merrett did. In “The Experience of Free Banking” (1992), pages 67 and 68, Kevind Dowd wrote :

One of the key issues here is the banks’ liquidity and Merrett goes on to argue that the ‘inescapable conclusion is that the long decline in liquidity standards seriously undermined the banks’ ability to cope with the growing problem of higher risks’ (1989:77). However, as George Selgin points out

“the facts tell a different story. Merrett (1989, p. 75) reports that the aggregate reserve ratio … fell from .3217 in 1872 to .2188 in 1877; but his figures for later five-year intervals show no further downward trend … . Even the lowest figure compares favorably to those from other banking systems, both regulated and free. It is much higher than Scottish bank reserve ratios for the mid-nineteenth century … and about the same as ratios for free Canadian banks in the late nineteenth century and for heavily regulated US banks today.”
(Selgin 1990a: 26–7)

[…]

The other key issue is capital adequacy. The figures given in Butlin, Hall and White (1971: table 2) show a fall in the capital ratio from about 20 per cent in 1880 to 12.5 per cent in 1892, but these figures ignore the uncalled liability attached to bank shares, and a number of banks also had a contingent reserve liability which took effect if the bank went into liquidation (Merrett 1989:82). Merrett himself estimates that this extra capital resource amounted to ‘nearly 45 per cent of the conventional measure of shareholders’ funds’ (p. 81) which suggests that these capital ratios give a considerably understated impression of the ‘true’ capital adequacy of the banks. … The claim that the banks had allowed their capital ratios to fall to reckless levels is also difficult to defend in the light of Pope’s chart (1989: figure 8) on core capital adequacy. If this hypothesis were correct, we would expect the capital ratios of failed banks to show a distinct downward trend in the period before they failed, we would expect a model of bank failures to show that the capital ratio had a negative and statistically significant coefficient, and we would also expect there to be a major (and growing) difference between the capital ratios of failed and non-failed banks. In fact, the capital ratio of failed banks appears to rise in the two years prior to failure, and their capital ratio reaches a low point five years before failure and then recovers. Pope’s logit model of the probability of failure also shows that capital adequacy has a ‘correctly’ signed coefficient in only one out of four cases, and even that is only significant at the 10 per cent level (1989: table 1), so there is little evidence that capital adequacy ‘matters’ in the way that this hypothesis predicts it should. And note, finally, that the difference between the capital ratios of banks that were to fail and banks that were not is relatively small — under 3 percentage points, and usually considerably less — and shows no tendency to grow as the dates of the failures approach (1989: figure 8).

The decline in ‘Average equity/deposit ratio’ and ‘Average liquid assets/total assets ratio’ between 1862 and 1882 can be interpreted as an indication “that customers required lower reserves from older, trusted banks” (Erik Lakomaa, 2007) and this substantiates White’s argument that “The drop [in reserve ratios] may be attributed to lower costs of obtaining specie on short notice or to lower risk of substantial specie outflows” (Lawrence H. White,1995).

In “The Theory of Free Banking: Money Supply under Competitive Note Issue” (1988) George Selgin noted that, in a mature free banking system, the “commodity money does not circulate, its place being taken entirely by inside money”. As the free banking matures, people trust more in the banking system so that the banks see no necessity of holding such a huge amount of reserves. And as Selgin illustrates in this interview : “The Scottish people were in fact so trusting of their banks that they considered it a nuisance to be handed a gold coin rather than a Scottish banknote”.

Instead of reflecting the instability of the free banking, the decline in reserve ratios may reflect the high confidence of the public in the banking system — on the condition that the banks do not increase their amount of loans above what the public actually demand (see chapter 5 of “The Theory of Free Banking”).

Hickson and Turner then argue that “the average equity:deposits ratio for the surviving banks was significantly higher than that of the banks which failed or suspended”. But what does it reveal about the soundness of the banking system ? Actually nothing. It is commonly known that the less capitalized banks are likely the ones to be the more damaged by a financial crisis (especially the banks that have participated in the boom) regardless of the reserve ratios of the whole banking system. That’s not something new, and this tells us nothing about the so-called ‘instability’ of australian banks. Furthermore, the figures provided by Pope tell a very different story. In “The Experience of Free Banking” (1992), pages 67 and 68, Kevin Dowd cites Selgin and gives further details :

… Pope’s reserve figures also show a minor difference only — perhaps two percentage points — between the reserve holdings of failed Australian banks and those that weathered the crisis. This also suggests that ‘overexpansion’ was not the root cause of the banking collapse.”
(Selgin 1990a: 27)

[…] Pope’s logit model of the probability of failure also shows that capital adequacy has a ‘correctly’ signed coefficient in only one out of four cases, and even that is only significant at the 10 per cent level (1989: table 1), so there is little evidence that capital adequacy ‘matters’ in the way that this hypothesis predicts it should. And note, finally, that the difference between the capital ratios of banks that were to fail and banks that were not is relatively small — under 3 percentage points, and usually considerably less — and shows no tendency to grow as the dates of the failures approach (1989: figure 8).

Despite their claims to the contrary, australian banks did not overexpand their credit the years before the 1893 banking crisis. They failed to demonstrate that the collapse of the australian banks is due to overexpansion, since the main assumption of the authors is wrong. However, they disagree with Kevin Dowd’s explanation of the australian banking collapse in “The Experience of Free Banking” (1992) :

The bungling attempts of the Victorian Treasurer to pressurize the Associated Banks in to bailing out the weaker banks backfired at a critical point and needlessly undermined public confidence. (p. 70)

Hickson and Turner argue that the regulations may have helped to mitigate the severity of the crisis. In their own words :

The majority of banks that suspended were permitted to engage in financial reconstruction by converting some deposits into preference shares, converting some short-term deposits into long-term fixed deposits and raising new capital from shareholders. Dowd argues that the permission given by the colonial governments allowing the banks to reconstruct made it difficult for depositors to liquidate banks.

Furthermore, it was argued by The Economist that in order for a free market economy to function efficiently, liquidation should not be hindered by government interference and, based on this premise, it was maintained that the suspended banks should have been liquidated. It is also argued that the reconstruction mania may have forced some of the more established banks to suspend. Undoubtedly, the nature of the reconstructions meant that there were costs borne by depositors and shareholders, implying that there would have been no long-term malincentives associated with the reconstructions. However, liquidation may not have been possible because bank assets and collateral were unsaleable, and may even have had near zero values due to fire-sale losses. Furthermore, the banking system survived, which is the ultimate test of the efficacy of the reconstruction policy. Indeed, it was recognised by many depositors that the reconstructions were the best outcome that they could achieve.

A second government response to the crisis was the declaration made by the Victorian government of a five-day bank holiday on Monday, 1 May 1893. The Premier had declared the bank holiday after hearing that National Bank was about to suspend. He had hoped that the holiday would calm depositors’ excitement but, instead, depositor excitement ‘rose to fever heat, and it was hastily assumed that all the banks would have had to suspend had it not been for the government’s intervention’. The adverse effect of the holiday led the ‘Big Three’ to remain open with several other banks following suit. In this way, one can argue that the sounder banks were able to identify themselves to the banking public, while those that stayed closed were seen as unsound. The adverse effect of the bank holiday policy can only have further undermined depositor confidence because ‘the public became concerned chiefly at the incapacity of its rulers during so grave a crisis’.

The Victorian government had been urged as early as January 1893 to make bank notes legal tender and to give a partial deposit guarantee. However, it ignored these suggestions. In contrast, the New South Wales government in May 1893 passed three measures applicable only to banks with their head office in New South Wales. Their success led to Sir George Dibbs, the Colonial Secretary, being praised by the public and banks for ‘the promptness and energy with which the government had introduced legislative measures for the restoration of public confidence during the recent financial crisis’. Indeed, the actions of the New South Wales government may have actually prevented the ‘Big Three’ from failing and the crisis deteriorating even further.

The first measure assented to on 3 May 1893 was the Bank Issue Act which made bank notes a first charge on assets, gave the Governor of New South Wales power to declare bank notes legal tender and granted the government the right to inspect banks. Initially, this Act was not forced upon the banks but Coghlan was sent to encourage the five major banks operating in the colony jointly to accept the Act. However, only one bank wanted its notes made legal tender. Yet, on 15 May, as soon as the government learnt of the impending suspension of the Commercial Banking Company of Sydney, it declared the notes of the following banks to be legal tender whether their managements wanted it or not – Bank of Australiasia, Bank of New South Wales, City Bank of Sydney and Union Bank. This meant that these banks could print notes to cover their deposit outflows. They were acceptable to the public because they had legal tender status, were a first charge on assets and the designated banks had been vetted before legal tender status was conferred. It is interesting to note that none actually failed. Indeed, Coghlan states that, within a couple of days, deposit withdrawals had ceased and the crisis was brought to an end.

The second measure that the New South Wales government passed in late May 1893 was the Current Account Depositors Act. This empowered the government to advance to current account depositors 50 per cent of the sums owing to them by the suspended banks. The advances took the form of Treasury notes with legal tender status. This enabled commercial life to continue in some form, and reduced the strain upon the ‘Big Three’. According to Shann, ‘this was the traditional policy of choking a panic with cash’. It was argued at the time that the Current Account Depositors Act and declaring bank notes legal tender ‘theoretically objectionable as they may be, saved further disasters of such possible magnitude as would have involved the insolvency of the colonies within a measurable time’. A third measure taken by the New South Wales government to help calm the crisis was the declaration made by the Colonial Secretary that no other banks with their head office in New South Wales would be permitted to fail. The government was willing to act as a lender of last resort if necessary. The declaration was made by the Colonial Secretary following the collapse of Australian Joint Stock Bank on 21 April 1893.

Indeed, Sir George Dibbs, the Colonial Secretary, argued that such a measure should have been passed a quarter of a century earlier. The credibility of this declaration would have been somewhat undermined by the failure of the colonial government to prevent the suspension of Commercial Banking Company of Sydney in mid-May 1893. However, credibility would have been strengthened by imposing the Bank Issue Act upon the banks, and by the passing of the Current Account Depositors Act.

On the strength of the measures taken by the New South Wales government and their effect, depositors of banks headquartered in New South Wales would have become more assured as to the safety of their deposits. Significantly, the banking crisis in New South Wales was not as severe as that in Victoria. Therefore, an interesting issue is whether the Victorian government could have ameliorated the severity of the crisis had it adopted policies similar to those of the government of New South Wales. Correspondingly, one can also speculate whether the crisis would have occurred had such policies been adopted by both governments immediately after the failure of the Federal in January 1893.

But they failed to provide us a convincing argument for advocating banking regulation. Supporting unhealthy assets tends to create a serious moral-hazard problem, which lead to more frequent bank failures. They also failed to demonstrate that a regulated banking system would prevent this kind of crisis; they simply make the assumption that it is the case without being able to prove it. Their conclusion may be summarized as follows : “the free banking model did not prevent the 1893 crisis, so a regulated banking model cannot fail in this task”. Who knows, a regulated banking system may instead do even worse.

In conclusion, while Kevin Dowd may be wrong about the cause of the 1893 crisis, Hickson and Turner also failed to identify the cause of the crisis. The root cause of the collapse of australian banks does not appear to be clear at first glance. As Briones and Rockoff noted in “Do Economists Reach a Conclusion on Free-Banking Episodes?” (2005) :

Besides, conclusions reached in recently documented relevant experiences like the case of Australia are still highly controversial and probably require further research.

If the overexpansion cannot account for the boom in the late 1880s, one question still remain. How has the boom been financed ? Selgin, in “Bank Lending ‘Manias’ in Theory and History” (1990), may have the final words.

Pope’s annual data, presented graphically (1989, p. 20, fig. 8), are more plainly inconsistent with the overexpansion hypothesis. According to Pope’s chart, in the seven years preceding the crisis the average reserve ratio of the 13 banks that survived the crisis varied from about 0.17 to about 0.19, with no apparent trend. The average ratio of the 13 suspended banks actually rose steadily from about 0.15 to about 0.16. These figures are consistent with the decline in Australian lending rates during the period in question. […]

The boom was financed, not by any increase in the bank money multiplier, but by injections of high-powered money from Austra-lian gold mines and from the British capital market. [13] When injections from Britain came to a sudden end with the Barings Crisis of 1890, many Australian banks were obviously unprepared; but this did not make them guilty of having lent excessively and “irrationally” during the boom years. (pp 275-276)

Hickson and Turner ironically give some support for the evidence brought by Selgin (1990) when they wrote : “However, by 1891, the ‘zombie’ land mortgage companies were no longer able to survive as the inflow of British deposits dried up. Correspondingly, the commercial banks were forced to ration credit to land mortgage companies due to the increased cost of raising deposits in Britain. As a result, from 1891 to March 1892, in Melbourne and Sydney alone, deposit-taking building or land finance companies failed”.