RDP 1999-06: Two Depressions, One Banking Collapse 2. Two Depressions: Output and Banking

2.1 Output During the Depressions

In Australia, real GDP fell by around 10 per cent in the first year of both depressions – that is, 1892 and 1931 respectively.^[5] However, the depression of the 1890s was substantially deeper and more prolonged than the depression of the 1930s (Figure 1). Real GDP fell by a further 7 per cent in 1893, coinciding with the collapse of the banking system. Growth returned in subsequent years, although it was moderate and erratic. It was not until 1899 that the level of real GDP had surpassed the previous peak set eight years earlier. In contrast, during the 1930s depression, growth resumed in 1932, and by 1934 the level of real GDP had surpassed the previous peak of 1930. Because of the relatively high rate of population growth during the 1890s,^[6] the 1890s depression was even deeper than the 1930s depression in terms of real GDP per capita (Figure 2). Real GDP per capita fell by around 20 per cent over the 1890s, compared with a fall of only about 10 per cent over the 1930s.^[7]

The corollary of a deeper, longer depression is more substantial and sustained deflation. Figure 3 shows that, from 1891 to 1897, retail prices fell by more than 20 per cent. By comparison, the fall in the 1930s episode was smaller at around 15 per cent, and over a shorter period – from 1930 to 1933. Figure 3 also shows that there was a large fall in retail prices from 1890 to 1891, before the downturn in output. This was due almost entirely to falls in house rents which constitute 40 per cent of this retail price index. As we mention in Section 3.2, the property market turned down in the late 1880s and was an important factor leading to the collapse of the financial system.^[8]

It seems reasonable to assert that the rise in unemployment and the fall in employment would have been worse during the 1890s than during the 1930s. However, this is difficult to establish because comparable data on unemployment in the 1890s and 1930s depressions are limited. Statistics on the employment status of trade union members indicate that unemployment peaked at almost 30 per cent in 1932; no comparable data exist for the 1890s depression. One series that is available in both periods is the unemployment rate for members of the Amalgamated Society of Engineers, Australia. This indicates that although the peak rate of unemployment was higher in the 1930s depression (almost 26 per cent in 1931 compared with 16 per cent in 1894), unemployment returned to pre-depression levels more rapidly as the economy recovered.

2.2 1890s Banking Collapse versus 1930s Banking Problems

The 1890s depression was characterised by a severe financial crisis, whereas the financial problems during the 1930s were relatively mild by comparison. In the 1890s, more than half of the trading banks of note issue suspended payment and a

large number of non-bank financial institutions failed.^[9] This compares to the failure of only a few, mostly smaller institutions during the 1930s. Some indication of the depth of the crisis of the 1890s is provided in Figure 4 which shows the ratio of total bank credit to GDP. Over the course of the 1890s depression, bank credit to GDP fell from above 70 per cent at its peak to about 40 per cent by the turn of the century. In contrast, from a peak of about 45 per cent in the early 1930s this ratio had declined to 38 per cent by the beginning of the Second World War.

Clearly, the cycle in bank credit during the 1880s and 1890s was greatly exaggerated. If anything, the cycle in total credit was likely to have been even more pronounced during the 1890s than it was for bank credit because of the behaviour of building societies, finance companies and the ‘land’ and other ‘fringe’ banks. Data on credit provided by these financial institutions are difficult to obtain. However, data on assets of financial institutions show that building societies and finance companies grew extremely rapidly through the 1880s – their share of financial system assets rose from 12 per cent in 1885 to more than 21 per cent by 1892 (Figure 5).^[10] These institutions also lost market share very rapidly through the 1890s depression, especially building societies. In contrast, through the late 1920s and early 1930s, the asset shares by institution remained relatively stable. Banks lost some ground, but mainly to managed funds and funds administered by trustee companies – neither of which were a substantial source of credit.^[11]

The data shown in Figure 5 do not include a number of important financial institutions, including the land finance companies and institutions that are sometimes referred to as banks because they accepted deposits and provided cheque facilities. However, it is worth noting that both of these types of institution played a key role in the 1890s episode – they tended to be newer, less conservative institutions, more willing to lend for speculative purposes, and were more likely to have failed in the financial crisis of the 1890s.

2.2.1 Trading banks

The financial system in the years prior to the banking crisis of 1893 was dominated by the private trading banks of note issue. In the early 1880s there were 26 trading banks^[12] controlling around 90 per cent of the assets of the financial system (Figure 5). The growth of other financial institutions over the 1880s resulted in the market share of trading banks falling to less than 70 per cent by the early 1890s. Building societies and pastoral finance companies grew particularly quickly in tandem with the property price boom during the second half of the 1880s (Pope 1991, Merrett 1991 and Boehm 1971). By the beginning of 1893, there were 23 trading banks in operation.^[13]

The structure of the financial system underwent considerable change over the period between the two depressions. Consolidation among the trading banks was perhaps the most significant development. There were 11 amalgamations of trading banks between 1917 and 1927, leaving 10 trading banks at the onset of the 1930s depression. Trading banks also lost market share over the intervening years, mainly at the expense of savings banks, while building societies failed to regain the share of the financial system that they had lost in the crash of the 1890s (Figure 5).

The problems experienced by the banking sector during the depressions can be illustrated in two complementary ways: by examining the movement of deposits; and by describing the numbers and details of bank failures.

Trading banks suffered a loss of deposits during both depressions, although the loss was more rapid, larger and more sustained over the 1890s compared with the 1930s (Figure 6). In the two years to 1894, trading bank deposits fell by 15 per cent and did not reach a trough until 1898, by which time they had fallen a further 5 per cent. In contrast, from 1929 to 1931 trading bank deposits fell by less than 10 per cent and recovered rapidly thereafter.^[14]

The performance of the trading banks over the 1890s was even worse than suggested by the movement in aggregate deposits, since these data include deposits frozen as the result of the reconstruction of many trading banks.^[15]

During the financial turmoil of the late 1880s and early 1890s, the first institutions to experience problems were the land finance companies and building societies that had been established during the property boom. Pope (1991) suggests that between 1891 and 1893, 54 deposit-taking financial intermediaries closed their doors (with 60 per cent of these closing permanently). The first trading bank to fail in 1893 was the Federal Bank of Australia which went into liquidation in January, but the banking panic started in earnest with the suspension of payment by the Commercial Bank of Australia in April of that year (Merrett 1989).^[16] These two banks were both exposed to the property market through loans to the land finance companies. Initially, runs on the banks of note issue focused on these two banks, and those known to be similarly exposed to the property market. Although there was a widespread loss of confidence in the banks, Merrett (1991) suggests that at least initially, customers were able to differentiate between institutions and some depositors transferred funds from the weaker banks to the older, more established banks. These included the Bank of Australasia, the Bank of New South Wales and the Union Bank of Australia, which had tended to be more conservative in their lending practices through the boom period of the 1880s.

In total, 13 trading banks were forced to close their doors in the first five months of 1893 (Royal Commission 1937, paragraph 96). However, by early August, 12 of these banks had undergone a process of reconstruction and were able to reopen.^[17]

Although the details differed between institutions, reconstruction of the suspended banks generally involved the formation of a new limited liability company with the same name, writing off capital, converting some deposits into equity and deferring the payment of the remainder of deposits. Another important factor in the reconstructions was that shareholders were required to inject large sums of additional capital into the new company (this can be seen in the sharp rise in the ratio of paid up capital to assets for the trading banks, Figure 13, Section 3.5). Over the course of the resolution of the financial crisis, the 13 trading banks that had suspended payments were forced to write off an amount in excess of the initial value of their capital in 1893. They wrote off 40 per cent (£4.4 million) of their capital in the first year of the crisis. This proved to be inadequate and from 1894 to 1909 these banks were forced to write off an additional £7.4 million worth of capital. This was more than the £5.9 million of new capital which had been issued from 1893 to 1909.

At the time of their closure, the banks that suspended payment controlled around half of the total deposits of the trading banks in Australia. In general, the reconstruction process resulted in the immediate release of very small deposits, but for the majority of depositors, receipts were issued for the value of the deposit, to be paid some time in the future. Over 85 per cent of deposits in the suspended banks were repayable in cash, and the remainder were converted to securities such as preference shares.

Of the deposits repayable in cash, the majority were paid back between 1893 and 1901, however, payment was not finalised in some cases until as late as 1918 (Royal Commission 1937, paragraph 224). So although most depositors ended up being paid (with interest), there were considerable indirect losses in terms of reduced liquidity by having deposits frozen, as evidenced by the fact that many customers opted to sell their deposit receipts in secondary markets for less than face value (Royal Commission 1937). Moreover, some depositors incurred direct losses due to the failure of the Federal Bank and the City of Melbourne Bank which went into liquidation in 1895. These losses totalled around £4 million (about 4 per cent of total trading bank deposits in 1891).

In contrast to the 1890s experience, only three financial institutions suspended payment in the 1930s depression, none of which were trading banks. The largest of these was the Government Savings Bank of NSW (GSB), which had been experiencing pressure on deposits throughout 1930. Sykes (1988) suggests that it was not these withdrawals per se but rather political influences that resulted in the run on deposits that forced the closure of the bank. In the lead up to the NSW State election in October 1930, statements by the incumbent Nationalist Government predicted financial collapse if a Labor Government was elected. This caused public alarm, which was further inflamed when the Labor party won the election and the public became aware that the government had defaulted on the payment of interest due to the GSB. On 1 April 1931, the NSW Government also defaulted on interest payments due to British holders of government bonds and this triggered a run on deposits at the GSB which led to the closure of the bank in April of that year. In early May, the bank was temporarily reopened (with the backing of the Commonwealth Bank) in order to release deposits for individuals with ‘necessitous circumstances’. However, it was not until December 1931 that the bank was merged with the Commonwealth Bank.^[18]

The other two institutions to suspend payment were the Primary Producers' Bank of Australia and the Federal Deposit Bank Limited. Although the 1937 Royal Commission suggests that the Primary Producers' Bank was not in a hopeless position, it had lost almost 40 per cent of its deposits in 18 months and was liquidated. At the date of suspension, deposits equalled £1.2 million, less than half of 1 per cent of total bank deposits in 1931. After the sale of debts of the bank, creditors were repaid 98.75 per cent of their funds, but shareholders received nothing (paid up capital just prior to suspension had been valued at almost £½ million).

The Federal Deposit Bank was, in effect, more like a building society than a bank. After it was forced to suspend payment, it was taken over by the Brisbane Permanent Building and Banking Company Limited and arrangements were made to pay depositors in full. Shareholders were paid in shares of the new company (Sykes 1988).

In addition to the above closures, there were a number of mergers following depositor withdrawals and liquidity problems. The Australian Bank of Commerce merged with the Bank of NSW in 1931 after profits of the former had fallen by over 50 per cent from their level in 1930.^[19] The other merger was that of the State Savings Bank of Western Australia with the Commonwealth Savings Bank in 1932, due to the illiquid position of the former and the fallout from the suspension of the GSB.

The only other significant run was on deposits at the Commonwealth Bank itself after it took over the business of the GSB. This run was stopped by statements by the Chairman of the Commonwealth Bank that dispelled fears that the Bank was in financial difficulties (Royal Commission 1937). However, more generally, the Commonwealth Bank did not contribute to the more stable position of the financial system leading into the 1930s depression, either in terms of monetary policy, or in terms of playing a regulatory role in the banking system (Schedvin 1992; Commonwealth Bank of Australia 1936).

2.2.2 Savings banks

By 1871 there was at least one savings bank operating in each of the colonies (Butlin 1986). Savings banks were either government owned and run through post offices, or run by government-nominated trustees and commissioners. Compared with trading banks, savings banks controlled a small, though increasing, share of total financial system assets.^[20] Despite their size, the performance of the savings banks provides an interesting comparison with trading banks. Whereas trading banks suffered a sharp fall in deposits in the 1890s, some failed and even more closed their doors for a short time, savings bank deposits rose steadily through the 1890s, albeit from a small base (Figure 7). This difference between the two types of institutions through the 1890s may have reflected a perception that savings banks were safe havens, given their tendency to invest heavily in government securities and their implicit colonial government guarantees.^[21] Over the 1880s, savings banks invested 26 per cent of their assets in government securities, compared with 1.3 per cent for trading banks.

Even though the increase in savings bank deposits through the 1890s was not large in absolute terms, it highlights an important point – namely, that the runs on financial institutions were not driven purely by self-fulfilling expectations that all financial institutions would fail. Rather, runs were specific to those institutions which were unable to demonstrate their soundness in the face of increasing fragility across much of the financial system.

One of the significant structural changes between the two depressions was the gain in market share of the savings banks – by the late 1920s, savings bank deposits accounted for about 40 per cent of total bank deposits. From 1929 to 1931, savings bank deposits fell by 14 per cent (or £32 million), compared with a fall of 8 per cent (or £27 million) for trading bank deposits over the same period (compare Figure 6 with Figure 7). However, a significant part of the fall in savings bank deposits was accounted for by the GSB, which lost £14.5 million of deposits from June 1930 to April 1931; although, it seems that some of these deposits were placed into the Commonwealth Savings Bank (Royal Commission 1937, paragraph 351).

Apart from concerns regarding the GSB, the fall in savings bank deposits may have reflected the tendency of people to draw upon their savings during difficult times (Royal Commission 1937, paragraph 351). Because savings banks were in a relatively sound condition, these withdrawals did not seem to raise concerns regarding the viability of these institutions.^[22] In this way, the increased proportion of deposits held in savings banks appears to have been one of the factors that enhanced the stability of the financial system in the 1930s and limited the extent of runs on deposits.

In summary, although there were runs on banks during the 1930s, they were largely confined to smaller institutions. The major trading banks survived the 1930s relatively unscathed and there were no losses to depositors in the trading banks. This was not the case in the 1890s when there were direct losses to depositors due to the failure of the Federal Bank of Australia and the City of Melbourne Bank and indirect losses due to the freezing of deposits while banks were restructured. Also, a large number of non-bank financial institutions failed in the 1880s and 1890s. We have not considered the magnitude of these losses here, but they were significant and should not be overlooked in any comparison of the two episodes.^[23] Again this contrasted with the experience of the 1930s, during which there were only small direct losses to depositors due to the failure of the Primary Producers' Bank, in addition to indirect losses arising from the suspension of payment by the GSB.

Footnotes

The data prior to Federation are likely to be less reliable than post-Federation, although there is no reason to believe that the statistics prior to 1900 are biased in any general way. On occasion we comment on crucial data issues in the main text, though a detailed description of the data and sources is left to Appendix A. A few brief comments are, however, warranted at this stage. With regards to the precise timing of the depressions, there is some debate as to whether real GDP (available annually) is the best measure to use (Valentine 1984). This and other debates in the literature often depend on the interpretation of inadequate data, or on the validity (or otherwise) of aggregating data across the colonies/states (Boehm 1971). However, these problems are not of great concern to the main arguments of this paper. [5]

The population increased at an annual average rate of 1.7 per cent between 1891 and 1901 compared with an annual average rate of about 1 per cent between 1933 and 1947. These dates correspond to years in which censuses were conducted. [6]

A comparison of real GDP per capita over a longer period (not shown) confirms that the greater depth of the 1890s depression did not reflect more variable economic cycles over this earlier period. If anything, real GDP per capita was more variable in the two decades to 1930 than in the two decades to 1891. [7]

The implicit price deflator for GDP implies that deflation was actually more severe during the 1930s, although this is driven largely by the fact that export prices fell further in the 1930s. The retail price of groceries in Sydney suggests that deflation was not quite as deep in the 1890s as the 1930s, but was more prolonged; however, NSW did not appear to suffer as large a fall in output as Victoria during the 1890s. [8]

The colonial banking regulations allowed trading banks to issue their own notes, which were widely used as a medium of exchange. However, to do so they became subject to legislation which among other things required them to submit regular statistical returns. In this paper we define banks to be the note issuing trading banks and the savings banks. This excludes a range of institutions that are often referred to either as banks, ‘land’ banks or ‘fringe’ banks. While these institutions were an important part of the credit cycle of the 1880s and 1890s, data are not readily available. It appears that in terms of their nature and behaviour, these institutions were most like the building societies. [9]

We do not have data on assets held by building societies and finance companies prior to the mid 1880s. Their share of financial system assets at this time was not actually zero as is suggested in Figure 5. Nevertheless, the rapid growth of these institutions through the 1880s as implied by the data in Figure 5 is consistent with other evidence (for example, see Boehm 1971). [10]

These institutions provided some lending for mortgages, but this represented less than a quarter of their total assets (Royal Commission 1937). [11]

This does not include the Mercantile Bank of Australia which did not enter official returns until 1887, even though it was established in 1877. [12]

Boehm (1971) lists 22 Australian trading banks in operation at the beginning of 1893. We have included the Bank of New Zealand, which commenced operations in Melbourne in 1872. We did not include the Bank of South Australia because a major part of their business had been acquired by the Union Bank of Australia in 1892. [13]

Due to significant deflation, real deposits, and deposits as a share of nominal GDP, rose sharply during the early years of both depressions. However, the fall in nominal deposits weakened banks because their lending contracts were specified in nominal terms. [14]

Merrett (1993a) discusses some of the problems with the money supply data contained in Butlin, Hall and White (1971). Problems with these data arise from the treatment of deposits frozen in the suspended banks. Merrett presents a revised series showing that the fall in the money supply may have been larger than shown by the series we have used. This strengthens our finding that the fall in trading bank deposits was larger in the 1890s than the 1930s. [15]

Previously, the Bank of Van Diemen's Land went into liquidation in 1891 and the Mercantile Bank of Australia and the New Oriental Bank Corporation went into liquidation in 1892. [16]

It has been suggested that banks may have favoured reconstruction as a way of avoiding liquidation, irrespective of the resulting costs to creditors (Pope 1987). An alternative view is presented by Merrett (1993b) who suggests that the reconstruction schemes were necessary in order to end the runs on banks and minimise potential losses to creditors. Further, the Royal Commission (1937) suggests that of those banks that suspended only the Commercial Banking Company of Sydney did so unnecessarily. [17]

This followed protracted negotiations which began in May between the Commonwealth Bank and the NSW Government regarding the conditions under which the merger would take place. [18]

The Royal Commission (1937, paragraph 281) concludes from statements by the Chairman of the Australian Bank of Commerce that ‘…the bank would have experienced some difficulty in continuing to carry on its business without assistance’. [19]

Their share increased from 7 per cent on average over the five years to 1891 to 24 per cent on average over the five years to 1930. [20]

Indeed, the Victorian Government promised to guarantee deposits in the Commissioner's Savings Bank if it agreed to merge with the Post Office Savings Bank of Victoria (Murray and White 1992). [21]

One indication of the relative soundness of savings banks is the fact that their holdings of government securities averaged about 50 per cent of total assets over the 1920s, compared with an average of about 15 per cent for trading banks. [22]

An indication of the magnitude of these losses is provided by the fact that deposits in Victorian building societies fell by about 50 per cent (£2.7 million) from 1890 to 1892, and the number of registered building societies in Victoria fell from 70 to 56 over the same period (Boehm 1971). In 1891 and 1892, almost £3 million of Australian deposits were at risk following the suspension of the Melbourne ‘land’ banks (Boehm 1971); although details of the eventual losses are not readily available. [23]