RDP 1999-06: Two Depressions, One Banking Collapse 4. Real Macroeconomic Features of the Two Depressions

So far we have shown that the financial system prior to the 1890s depression showed clear signs of increasing instability, whereas prior to the 1930s depression the financial system displayed evidence of relative stability. In this section of the paper we examine whether variation in the financial pre-conditions was the primary reason for the difference in the performance of the real and financial sectors during the two depressions. To this end we consider the role that factors outside the financial system may have played. We examine a number of external and internal exogenous real factors before examining the importance of government policies. We conclude that those differences that did exist in terms of the performance of the real economy were either not significant enough, or were even working in the wrong direction, to explain the financial crisis.

4.1 Exogenous External Factors

Although there is some debate in the literature about the relative importance of internal and external factors in causing both the 1890s and 1930s depressions, there is little doubt that the external sector played a significant role in both.^[45]

The United Kingdom and the United States both experienced depressions in the 1890s and 1930s, although the 1930s depression was more severe in both countries. Real GDP fell in the UK by around 3 per cent from 1891 to 1893. Real GNP in the US fell by around 5 per cent in 1893 and by a further 3 per cent in 1894.^[46] By comparison, in the US, real GDP fell for four consecutive years, from 1930 to 1933, with a fall of 15 per cent in the worst year of 1932. The 1930s depression was less severe in the UK than the US, with output declining by 5 per cent from 1929 to 1931.

Relative falls in Australia's terms of trade across the two depressions reflect the greater severity of the world depression of the 1930s compared with the 1890s. The terms of trade fell gradually over most of the 1880s and early 1890s (Figure 15). Over the 11 years to 1894, the terms of trade fell by over 25 per cent. The falls in the terms of trade leading up to, and during the early years of, the 1930s depression were more dramatic. Between 1925 and 1932, the terms of trade fell by nearly 50 per cent; the fall in 1932 alone was a substantial 25 per cent (Jonson and Stevens 1983).

It is not surprising that the larger fall in the terms of trade during the late 1920s and early 1930s was associated with a more marked decline in nominal exports than was the case over the 1890s (Figure 15).^[47] This confirms Sinclair's (1965) observation that the influence of the external sector occurred mainly through capital flows in the 1890s depression.

In real terms, exports increased through both depressions – with similar and moderate increases over the five years following output peaks in 1891 and 1930 (Figure 15). It is true that real exports rose more rapidly in the first three years of the 1930s depression compared with the 1890s depression, although it is difficult to know how much of this difference can be explained by the devaluation of the exchange rate in the 1930s (Section 4.2.3). Firstly, the better performance did not last beyond three years, suggesting that special factors in the export sector may have been relevant. Secondly, the financial crisis of the 1890s may have made it difficult for producers and exporters to obtain working capital and trade finance.

In summary, the decline in world output was more substantial during the 1930s depression. Accordingly, the decline in the terms of trade was more pronounced for the 1930s episode than it was for the 1890s. Movements in nominal exports partially reflected these terms of trade declines. The rise in real exports was broadly comparable over the full course of both depression episodes. Therefore, it seems clear that the more substantial decline in real output during the 1890s depression cannot be explained by the size and impact of external real shocks. We now turn to the question of internal factors in the real sector.

4.2 Exogenous Internal Factors

4.2.1 Population

An underlying feature of Australia's economic development in the second half of the nineteenth century was the rapid growth of the population and workforce. The population increased at an average annual rate of 3.1 per cent from 1871 to 1881, and 3.4 per cent from 1881 to 1891 – rates well above all other countries of the Western world (Butlin 1958), and higher than Australia has experienced for any extended period since that time. This rapid population growth was a significant factor behind the high and sustained real GDP growth from 1871 to 1891.^[48] In fact, the fall in real GDP during the 1890s depression coincided with a slowdown in the rate of population growth to 1.7 per cent per annum on average over this decade. Population growth also slowed from the 1920s into the 1930s, although this was much less marked than during the 1890s depression.^[49]

Rapid population growth through the 1870s and 1880s was undoubtedly a factor behind the high level of building activity over this period. Hence, population growth contributed indirectly to increasing financial system instability prior to the 1890s depression. There is a relatively close positive relationship between annual population growth and residential investment across the three major colonies of NSW, Victoria and Queensland (Boehm 1971 and Butlin 1962). Population growth peaked in the first half of the 1880s in both NSW and Queensland, as did new residential capital formation. Population growth was increasing over the 1880s in Victoria, and peaked quite dramatically in 1888. This was also a peak year for new residential construction, and property prices in Victoria.

In the early 1890s, population growth rates turned down rapidly having peaked in the first half of the 1880s. This reflected continued falls in the growth rates in NSW and a rapid downturn in the growth rate in Victoria. Despite the trend decline in aggregate population growth rates, aggregate building activity (as a share of GDP) remained at a high level throughout the 1880s and into 1891. In this way, population growth had an important role in initiating the building boom, although it was not responsible for sustaining the boom for so many years (except perhaps in Victoria). Finally, population growth was not an entirely exogenous factor. During the early years of the 1890s depression, some of the fall in population growth should probably be attributed to an endogenous response by prospective immigrants to the rapid decline in living standards.

4.2.2 Weather conditions

Australia's reliance on pastoral industries for export income means that growth is influenced by climatic conditions. Other than 1888, seasons had generally been good leading up to the 1890s depression. However, Australia was hit by a severe drought between 1895 and 1903 (Boehm 1971), which led to a sharp fall in pastoral output as a share of GDP over this period (Figure 16). This may have been an additional factor in the relatively slow recovery from depression in the 1890s compared with the 1930s. However, the fall in the ratio of pastoral output to GDP in the three years following the 1891 peak in output was relatively minor. In other words, the drought cannot explain the relatively large fall in real GDP prior to 1895, and therefore, it cannot explain the collapse of the financial system in 1893.

4.2.3 The gold standard

Prior to 1931, Australia's exchange rate was determined by the gold standard, except for the period between 1914 and 1925 (Schedvin 1970). Under the gold standard, the exchange rate moved within a narrow range that was determined by the gold points. Gold points were the level of the exchange rate at which it became more profitable to import or export gold instead of buying or selling bills of exchange at bank rates. During the period leading up to the 1890s depression, the exchange rate generally moved within 1 per cent either side of par with sterling (Butlin, Hall and White 1971). The major departure from parity was a large depreciation of the buying rate in May 1893, as the surviving banks imported gold to help replenish their reserves after the banking crash.

After being suspended during the First World War, the gold standard was reinstated in Australia in 1925. As the external position worsened during 1929, the exchange rate came under pressure. However, the private trading banks, and the Commonwealth Bank, attempted to resist devaluation and tried to maintain parity with sterling throughout 1930 – in part out of the fear of creating inflation (Butlin and Boyce 1985). It was the Bank of New South Wales which eventually placed pressure on the other trading banks to devalue. The exchange rate was devalued in late 1930, and by the end of January 1931, £100 sterling bought £130 Australian (Schedvin 1970). Butlin and Boyce conclude that the decision of the trading banks to resist devaluation had led to very tight monetary conditions through 1930, which helped to add to the problems of the depression. Also, the policies of these banks (including the Commonwealth Bank) during 1931 and 1932 acted to delay the reduction in bank interest rates, thereby slowing recovery. Finally, the low level of foreign exchange reserves in 1930 (brought about by the failure to devalue earlier) prevented fiscal policy from taking a more active role; in fact, government spending had to be curtailed significantly (Section 4.2.4).

While it seems plausible that the devaluation in 1931 played some role in supporting demand after the initial downturn in output, it is difficult to determine the significance of this effect. Certainly, real exports grew more rapidly in the early years of the 1930s than during the 1890s (Figure 15), but this was not sustained beyond a few years.

Also, the ratio of nominal imports to GDP fell further during the 1930s depression than the 1890s depression. This may have reflected the relative price effect of the devaluation, that is, switching demand towards the non-traded sector. However, increased protectionism at about this time (Footnote 47) and the higher relative cost of imports, reflected in the large terms of trade decline, also played a role.

Another way of addressing this issue is to ask whether the 1890s depression might have been less severe if there had been some form of devaluation. No doubt this should have had some positive impact on growth. However, it seems very unlikely that devaluation would have been sufficient to prevent the financial collapse. Prior to the substantial fall in output in 1893, events in the early 1890s had already set a course towards the collapse of the banking system. Most telling of all, property prices had turned down as early as the late 1880s, and the crisis in the non-bank financial sector was well under way by 1891.

4.2.4 Fiscal and monetary policy

Federation and the formation of the Commonwealth Bank meant that there was greater potential for fiscal and monetary policy to act in combination to stimulate aggregate demand during the 1930s depression. We argued above that devaluation of the exchange rate in 1931 played some role in alleviating the impact of the significant downturn in the terms of trade. However, in subsequent years, policy is best characterised as having been only mildly expansionary.

The aggregate government budget had been in deficit prior to both depressions, but by less than 1 per cent of GDP. The deficit reached a high of only 1.3 per cent of GDP in 1893, whereas it reached 4 per cent and 3 per cent of GDP in 1931 and 1932 respectively. Even though it might appear as if policy had been loosened substantially in the early years of the 1930s depression, a closer examination of the details of fiscal and monetary policy changes suggests otherwise.

The fiscal deficits of the early 1930s appear to have been an endogenous response to a few special factors. Aggregate government expenditure actually fell from 1930 to 1932, by the same amount (17 per cent in nominal terms) as it did from 1891 to 1893. Within the expenditure categories, charges for debt increased in the 1930s, in part because of the impact of devaluation on debt denominated in pounds sterling (Barnard 1985).

Total tax revenue fell by 12 per cent from 1891 to 1893 and by 6 per cent from 1930 to 1932. Falls in tax revenue were driven primarily by large falls in customs revenue associated with declining import values – this loss was particularly large during the 1930s episode (despite increases in tariff rates). In the 1930s, the fall in customs revenue was partially offset by increases in income tax revenue and the imposition of a new sales tax. In this way, total tax revenues did not fall by as much as they did during the early 1890s.^[50]

Fiscal policy was not more expansionary during the 1930s in part because of the reluctance of the Commonwealth Bank to provide credit to the public sector (for details see Butlin and Boyce 1985). Much of the Federal Government deficit during the early 1930s appears to have been financed by the sale of treasury bills to the Commonwealth Bank (precise data is not readily available). The extent of this finance was limited by the Commonwealth Bank because of its fears regarding inflation over the longer term. To this end the Commonwealth Bank also pressured the Government to reduce the stock of outstanding treasury bills from 1932 until the end of 1935. In the meantime, the Commonwealth Bank sold some of its holdings of treasury bills to private banks at reasonably attractive interest rates, thereby helping to keep interest rates relatively high.

One of the changes in financial arrangements that occurred between the two depressions was the passing of note issue from the private trading banks to the Commonwealth Government. The relevance of the form of note issue on the stability of the financial system is unclear, although it seems that the existence of private notes per se was not a factor in the relative instability of the financial system leading up to the 1890s depression (Rohling and Tapley 1998). Merrett (1991) suggests that there is evidence of depositors moving their money to the stronger banks as confidence in the banking system deteriorated in the early 1890s. Moving deposits to other note issuing banks indicates that in itself, the existence of private notes was not enough to explain the banking crisis of the early 1890s. Indeed, the ability of banks to issue their own notes may have been a factor in limiting the severity of the banking crisis in NSW in 1893. The NSW Government declared the notes of certain banks to be ‘legal tender’, which allowed banks to meet withdrawals of deposits with their own notes. Therefore, customers faced the same risk whether their funds were in the form of a deposit with a certain bank or in the form of notes of that bank. It appears that this realisation by customers helped to limit runs on these banks.

Other actions of colonial governments during the 1890s did not prove so successful. The Victorian Government declared a bank holiday for five days at the start of May 1893. However, the stronger Melbourne banks ignored the regulation and opened for business. This helped alleviate the panic, but also drew further attention to those banks that shut their doors during the holiday.

To summarise, neither fiscal nor monetary policy can be characterised as truly expansionary during the 1930s (beyond the impact of the devaluation in 1931). Attempts to expand the fiscal deficit more vigorously were resisted by the Commonwealth Bank. In both the 1890s and 1930s depressions, fiscal expenditure fell by similar percentage amounts, while expenditure on debt payments increased during the 1930s depression. Also, in the 1930s the tax base was broadened with increases of existing tax rates and the imposition of new taxes.

Therefore, it would seem that the differences in government policy across the depressions were not sufficient to explain the significant variation in the performance of the financial sector. The same can be said of variation in real external shocks across the two depressions. In fact, the real external shocks of the late 1920s and early 1930s were much more substantial than during the early 1890s; although the impact of the significant decline in the terms of trade on output was at least partially offset by the expansionary effect of the nominal devaluation in 1931. Also, variation in the rates of population growth can help to explain why the building boom of the 1880s was much more substantial than that of the 1920s. But this does not help to explain the larger decline in real GDP per capita in the early years of the 1890s depression (compared with the 1930s). Finally, severe drought conditions were a factor in extending the 1890s depression beyond 1894, but were not relevant to the earlier decline in real output and the financial crisis of 1893.

Footnotes

See Sinclair (1965) and Valentine (1985) for a discussion of this debate. [45]

In the UK, real GDP growth had been stagnant since the early 1870s (Pollard and Crossley 1968) and Britain's major influence on the Australian economy was through the cessation of capital flows following the Barings crisis in 1890. Although the depression in the US was worse than the UK during the 1890s, trade flows between the US and Australia were still relatively small. There is no real consensus regarding the cause of the US depression in the 1890s. However, Friedman and Schwartz (1963) suggest that in the early 1890s there was a loss of confidence in the ability of the US government to maintain the gold standard. This caused expectations of a devaluation of the US dollar, which in turn caused funds to flow out of the US (this may have been exacerbated by the Barings crisis, which caused a general loss of confidence on behalf of British investors). The downturn was also fuelled by a crisis in the financial sector – the US stock market crashed in mid 1893 and a month or so later there was a widespread loss of confidence in the banks, leading to some suspensions and some failures. For a recent discussion of these events see Rockoff (1998). [46]

The degree of openness of the Australian economy was broadly similar over the two episodes. Therefore, a given-sized decline in the terms of trade should have had a similar impact. The sum of exports and imports, expressed as a percentage of GDP, suggests a trend decline in the degree of openness from the 1870s to the 1930s; although the average over the 1880s was similar to that over the 1920s – 35 per cent versus 34 per cent. Commonwealth tariffs imposed after Federation, and increased following the First World War, suggest that the economy had become more closed. For a history of protectionism in Australia see Anderson and Garnaut (1987), and Pope and Manger (1984). [47]

Population growth depends on factors which are both endogenous and exogenous to the Australian economy. Clearly population growth will impact on real GDP growth, just as economic prospects will influence prospective immigrants. Similarly, prospective immigrants will be influenced by exogenous conditions in their home countries. [48]

Population growth averaged 1.7 per cent per annum from the census dates of 1921 to 1933, and then fell to an average of 1 per cent per annum between 1933 and 1947. [49]

Government businesses were also an important component of both revenue and expenditure. Expenditure on Commonwealth and State government businesses fell by 38 and 35 per cent from 1891 to 1893 and from 1930 to 1932 respectively. However, the percentage fall in revenue from these businesses was twice as large in the latter period – 14 per cent versus 7 per cent – thereby helping to contribute to the more substantial budget deficit in the 1930s. [50]