RDP 9002: Public Sector Growth and the Current Account in Australia: A Longer Run Perspective 3. The Australian Experiences^[8]

In this section our aim is to present some “stylized facts” for the Australian economy using the framework introduced above. All National Accounts data are expressed in current prices as a proportion of current price GDP. We deflate by GDP to provide a yardstick for examining results. Nominal data rather than real data are used, partly because of data limitations, and partly due to unusual behaviour of some price deflators at points where data has been re-based. It is important to note, therefore, that some of the behaviour in the data will incorporate relative price changes as well as quantity changes. This is especially important for the trade balance data which includes terms of trade effects.

It is worth highlighting several features:

1. Taxation as a proportion of GDP has risen continually between 1960 and 1988/89. From the early 1970s to 1984/85 there was a large growth in the size of government in Australia. Higher taxation as a percent of GDP financed an increase in government current expenditure and transfers, as a percent of GDP. There has been a decline in government capital expenditure as a percent of GDP.
2. The increase in taxation has been focused almost entirely on the household sector.
3. Since 1983/84 the trend of government spending has been reversed, with the largest cutbacks in the 1987/88 and 1988/89 fiscal years. These cutbacks focussed equally on transfer payments and expenditure. The trend in taxation was reversed only in 1988/89.
4. The growth in government during the 1970s coincided with a trend deterioration in the trade balance (particularly after 1979/80) as well as rising private consumption expenditure and falling private saving (although interpreting the statistical discrepancy as unrecorded consumption reverses the consumption/saving story).
5. The fiscal cutback in the most recent few years has been associated with some improvement in the trade balance, a strong rise in private investment and a large rise in the statistical discrepancy.
6. Over the period since the early 1970s, Australia also experienced large movements in real and nominal exchange rates, and real and nominal interest rates.

a. Sectoral Balances

Figure 1 shows the trade balance, primary fiscal deficit and net-of-interest private saving/investment balance as defined in equation (2) as a proportion of GDP from 1961/62 to 1988/89. Several features of this graph stand out. In particular there was a rise in the primary fiscal deficit from less than 4 percent of GDP in 1970/71 to 10 percent of GDP in 1983/84. This was followed by a fall to 2–1/2 percent of GDP in 1988/89. Also noteworthy is the deterioration in the trade balance from an historically-large surplus of close to 4 percent of GDP 1972/73 to a deficit of 2 percent of GDP in 1988/89.^[9] The trade balance deficit through the 1980s was clearly larger than the deficit on average during the 1970s, although in 1988/89 the deficit of close to 2 percent of GDP was an improvement over the low point of 4 percent of GDP reached in 1981/82. The trade deficit was also similar, relative to the size of the economy, to those experienced during the mid 1960s.

The excess of private saving over private investment was relatively flat at around 3 percent of GDP during the 1960s but then rose to a peak of 10 percent of GDP in 1972/73. This was followed by a gradual decline to 2 per cent of GDP. The sharp fall in 1980/81 and 1981/82 reflected the surge in investment associated with the “resources boom”.

(i) Private Saving and Investment

The saving/investment balance is decomposed in Figure 2 into saving (as defined in equation 2), investment and consumption. Private consumption expenditure as a proportion of GDP has changed slowly over this period, falling by 6 percent of GDP from 1961–62 to 1973–74 and then rising steadily by 5 percent of GDP from 1974/75 to 1982/83 before again falling steadily by 5 percent of GDP to 1988/89. Private savings shows the inverse behaviour. Private investment has maintained its share of GDP, with a sharp rise associated with the resources boom in 1980/81 and 1981/82 followed by a sharp fall during the recession in 1982/83 and a strong rise in 1987/88 and 1988/89.

(ii) The Statistical Discrepancy

Since the sectoral balances as we have defined them in Figure 1 should add to zero, any discrepancy should be reflected in the statistical discrepancy shown in Figure 3. Several features of this figure stand out. The statistical discrepancy is clearly not random. There appears to be distinct trend changes in 1974/75 and 1982/83. The discrepancy is also highly negatively correlated with consumption (shown in Figure 2). The recent rise in the discrepancy of nearly 2 percent of GDP from 1985/86 to 1988/89 is equal to about half of the deterioration of the trade balance since 1961/62.

Given that this is such a large error in the data, it is worth attempting to allocate the discrepancy to the sectoral balances to see how different the economy would look. In Appendix A we reconstruct Figures 1 and 2 assuming that the discrepancy is unreported consumption. The effect of this assumption is quite dramatic and points to an important unresolved issue: what does the statistical discrepancy represent?

(iii) Government Taxation and Outlays

Another major component of the saving/investment balance is taxes. In Figure 4 we plot total taxes and the decomposition into personal income tax, company tax and other taxes. Several features of this figure are quite remarkable. Total taxes as a percent of GDP have risen steadily since the beginning of the period but noticeably faster since 1972/73, increasing by approximately 7 percent of GDP between this date and 1988/89 (or by 30 percent of their 1972/73 level). Company tax as a percent of GDP has fallen steadily since 1970/71 from 4 percent of GDP to 3 percent of GDP. At the same time as company tax has fallen, household income tax and other tax have risen dramatically from 1970/71, with income tax continuing the trend of the 1960s.

The final major component of the private saving/investment balance is net transfers from the public sector to the private sector. This is plotted in Figure 5 together with total government spending on goods and services (both current and capital expenditure) as a percent of GDP. Note that government is defined inclusive of state and local governments and public enterprises. The behaviour of transfers is similar to that for taxes until 1984/85, after which they fell as part of the fiscal cutbacks, while taxation continued to rise. The net of taxes and transfers therefore tend to cancel out in the saving data up until 1984/85, but becomes important for the behaviour after 1984/85.

As shown in Figure 5, government spending on goods and services rose consistently from 1961/62 to 1982/83. Movements in total government spending reflect change in the consumption component, with the trend rise in consumption slightly offset by the trend decline in public sector capital expenditure. There is a distinct jump of total government spending in 1974/75 to a new level 4 percent of GDP higher than 1961/62. Since the peak in 1982/83 there has been a fall of about 4 percent of GDP in government spending to 1988/89.

Throughout the period of general government expansion, the capital expenditure component of government has fallen as a proportion of GDP. In the recent period of fiscal tightening the cut has been the equivalent of 2–1/2 percent of GDP. Government consumption expenditure in 1988/89 was still about 2–1/2 percent of GDP above the level in 1972/73 and 4–1/2 percent above the level in 1961/62.

In Figure 6 we present two alternative measures of fiscal position. One is the primary fiscal deficit on which we have focussed so far, and the other is the PSBR, which is the usual focus of public debate. The difference between the two measures is interest servicing costs (which would tend to make the PSBR measure larger than the primary fiscal deficit) and in “other revenue”, interest received and depreciation provisions (all of which tend to make the PSBR smaller than the primary fiscal deficit). The general movements in the data are quite similar.

It is interesting to concentrate on the recent fiscal adjustment. In Table 1 we calculate different categories of government outlays and expenditure as percent of GDP based on the government accounts.

Between 1983/84 and 1988/89 the PSBR fell by 8.41 per cent of GDP. Of this fall, 2.86 per cent of GDP was due to increased revenue and 3.45 per cent of GDP was due to cuts in spending on goods and services. Cuts to other outlays such as transfer payments accounted for a further 2.79 per cent of GDP. It is important to note, however, that the fall in the final spending component did not occur until 1987/88. Between 1983/84 and 1986/87 the PSBR fell by 3.54 per cent of GDP, but the fall in the final spending component during this period was close to zero. In terms of the split between current and capital expenditure, the government accounts show that the cuts have been divided equally between capital expenditure and current expenditure. Much of the reduction in current expenditure, however, did not occur until 1988/89.

In terms of another measure of fiscal position – the debt accumulation equation – it is worth pointing out that between 1983/84 and 1988/89, the ratio of interest to GDP rose by 0.64 percent but through 1987/88 and 1988/89 the ratio began to fall due to declining government debt over this period. For the ratio of debt to GDP to reflect only the primary fiscal deficit, we require a cut in outlays equal to the rise in servicing costs. By 1988/89 the cut in government spending on goods and services of 3.45 percent of GDP was a net improvement in the spending side of fiscal policy (adjusting for higher servicing costs) of about 2.8 percent of GDP. The improvement in the trade balance in 1988/89 over the average from 1981/82 to 1985/86 is about 1/2 of one per cent of GDP. The reduction in spending places the fiscal adjustment in a much better light from the view of sustainability compared to the position up to 1986/87.

(iv) The Balance of Payments

In Figure 7 we plot the current account and the trade balance as a percent of GDP. The difference is primarily interest servicing costs, which have risen since 1982/83. As in the case of the fiscal deficit, the larger the interest servicing costs, the larger the trade balance surplus required to service the outstanding stock of debt. In the 1980s real interest rates have been positive, which places a greater debt servicing burden on the public sector and the economy. The size of the trade surplus and fiscal surplus required to prevent foreign debt or government debt from accumulating is larger in the 1980s than in the 1970s.

(v) Does the Twin Deficits Proposition Apply?

A comparison can be made between the current account and PSBR. These are the two series that are frequently linked by the “twin deficits” proposition. As noted on page 3 (based on the identity in equation 2), a change in the government's fiscal position will be associated with a change in the current account, provided that the private savings/investment balance remains unchanged. The relationship between the two deficits seem to be, at best, loose. Netting out the effect of rising real interest rates (which tend to worsen both the fiscal deficit and the current account) only slightly improves the relationship. It seems that in the case of Australia, the private saving/investment balance cannot be taken as given, as assumed in the twin deficits proposition.

Before we can interpret the behaviour of the sectoral balances from the national accounts data we need to examine other key macroeconomic variables. We focus attention on real output growth, real wages, inflation, nominal and real exchange rates and interest rates.

b. Real Growth

The behaviour of real output is shown in Figure 8. This figure measures the percentage change in output from the same quarter of the previous year. This smooths out quarterly fluctuations. Note that the 1970s was a period of lower average real growth than in the 1960s. The 1980s, which began with a severe slowdown in growth, is now a period of faster growth than the 1970s.

c. The Behaviour of Relative Prices

(i) Real Wages

Figure 9 shows the growth of the four-quarter percentage change in nominal wages (defined as average weekly earnings) and the GDP deflator. Several well known features of this Figure are the rise in real wages in 1974/75 after a levelling in the previous two years, and a further rise in real wages in 1981/82 after a levelling out. It is not shown here, but the decline in real wages is associated with strong employment growth, a substantial fall in the wage share and a correspondingly larger profit share.

(ii) Inflation and Real Interest Rates

In Figures 10 and 11 we plot the rate of inflation defined in terms of the CPI and GDP deflators respectively. We also need some measure of expected inflation to calculate an ex-ante real rate of return. The method of calculation of an expected inflation series is shown in Appendix A.

The CPI and GDP deflators have moved somewhat differently late in the period, with the CPI falling from around 12 per cent in late 1986 to around 8 per cent by 1988/89, while the GDP deflator has risen from around 5–1/2 per cent in 1984/85 to about 7 per cent in 1988/89. The different trends are even more pronounced when the expected inflation series are used. The difference between the two series primarily reflects changes in the terms of trade over this period, and partly definitional differences between the implicit deflator and the consumer price index.^[10]

In Figure 12 we plot the quarterly data for both a short-term and long-term real interest rate in Australia with expected inflation defined in terms of the GDP deflator. The short-term interest rate used is the return on 90 day bank bills. The long-term interest rate is the yield on 10-year government bonds. Each series is converted into an annualized return for each quarter. Real short-term interest rates defined in terms of the GDP deflator and the CPI series tell similar stories except for the recent few observations and therefore only one short-term real interest rate series is graphed. The 1970s was a period of very low and negative real short and long term interest rates. Rates rose steadily from the March quarter 1975 and peaked in the December quarter 1985 for short rates and three quarters earlier for long rates: they then fell until early 1988 by both measures. Since the March quarter 1988, real interest rates have again risen.

In an attempt to introduce taxes into the calculation of real rates of return, Figure 13 shows the annualized before-tax nominal interest rate, after-tax real interest rate and expected inflation.^[11] Although Australia has experienced relative high nominal interest rates during the period, real after-tax interest rates have been positive only since 1981. Both interest rate measures fell during 1987/88 but have risen again in 1988/89.

Figure 13 updates the Carmichael and Stebbing (1983) graph of the “Inverted Fisher Hypothesis”. Carmichael and Stebbing argued that in a regulated financial market with a zero nominal return to money balances, if financial assets are closely substitutable for money, then changes in expected inflation will be reflected in changes in real interest rates rather than changes in nominal interest rates on financial assets. If this is the case, then real returns to financial assets are unlikely to be a good approximation to the real returns to capital when inflationary expectations are changing. This is clearly the case in the period up to 1981. Carmichael and Stebbing also speculated that deregulation of financial markets would tend to dilute the relation between changes in inflation and changes in real returns to financial assets, and the graph gives some evidence to support this. What is the appropriate rate of return to use for investment and savings decisions? During the 1970s, a weighted average of returns to financial and real assets may be a better approximation of the real return to savings. The dramatic difference this could make can be seen from the U.S. evidence that the real return to capital remained positive during the 1970s even though the real return to financial assets became sharply negative.^[12]

(iii) Real Exchange Rates and Foreign Real Interest Rates

So far we have ignored the impact of the rest of the world on Australia except through the trade balance. Capital flows are also an important link. To gauge the behaviour of Australian interest rates relative to the rest of the world we present in Figure 14 the annual series for the Australian nominal interest rate together with the annual U.S. 90 day interest rate.^[13] We also present results for real interest rates in Australia and the U.S. in Figure 15. Nominal and real exchange rates relative to the U.S. are presented in Figure 16, and nominal and real exchange rates relative to Japan are presented in Figure 17.

First consider the behaviour of nominal interest rates in Figure 14. In a broad sense, uncovered interest parity seems to work pretty well in explaining interest differentials over the period, even despite the presence of capital controls until the early 1980s. Domestic and foreign interest rates are reasonably similar during the period of fixed exchange rates until 1972. Given the capital restrictions until the early 1980s, it is not surprising that domestic and foreign interest rates are reasonably similar in the period to around 1972. The two series diverge in 1972/73. The sharp appreciation of the nominal exchange rate in 1972/73 corresponded to the emerging fiscal deficit and a strong rise in world commodity prices. The interest differential widened and the nominal exchange rate depreciated until 1977. The second period of domestic interest rates above world interest rates began in 1981/82 and corresponded to a period of nominal exchange rate depreciation against the U.S. dollar. This partly reflected the higher inflation rate in Australia relative to the U.S. as well as a strong appreciation of the U.S. dollar until late 1985.^[14] The sharp widening of the interest differential coincides with a large depreciation of the exchange rate in 1985. The fall in interest rates during 1987 corresponds to a period of nominal exchange rate appreciation which was primarily due to strong commodity prices.

In Figure 16 we also plot the real exchange rate relative to the U.S. dollar. Ideally we should use a trade weighted basket but data limitations and the desire to examine interest parity conditions restrict us to focus on the U.S. dollar. The broad trends are similar to the movement relative to the yen and other major currencies – except in 1981–85. Note that a fall in both the real and nominal exchange rate is an appreciation. The commodity price boom in 1972 and 1973 was accompanied by a nominal and real appreciation. The real appreciation of 1972 was locked in until 1982 by the strong wage growth which offset the nominal exchange rate depreciation over this period. It was not until mid-1981 that the overvalued real exchange rate began to depreciate against the U.S. dollar, and not until 1985 against other currencies. Exchange rate changes were not passed into wage settlements, thus improving competitiveness in U.S. dollar terms from 1981 until 1986. Much of the real depreciation relative to the U.S. dollar was reversed by the nominal appreciation through 1987 and 1988. This rise in relative prices will show initially as an improvement in the trade balance (as we measure it), but over time as the real trade balance deteriorates, the trade balance will worsen. As a comparison, we present the real and nominal exchange rate relative to the yen in Figure 17. Note that the loss in competitiveness in 1988/89 is smaller when calculated in terms of yen.

Figure 15 plots real interest rates for Australia and the U.S.. This shows that broad trends in real interest rates are similar to those in the U.S. (was the inverted Fisher effect working in the U.S. as well?), although there are large deviations in the short run. The period before 1980 reflects the regime of fixed or managed exchange rates between Australia and the U.S.. The real interest rate differential was in Australia's favour from 1972 (i.e. our real rates were lower than the U.S. rates) which is hard to reconcile with the overvalued real exchange rate of the time. The greater flexibility of rate regime in the 1980s has allowed for a bigger divergence between nominal rates of return, but real rates have not diverged much. In 1982/83 and 1983/84, when the Australian real interest rate was below the U.S. real interest rate, it corresponded with a period of expected real U.S. depreciation. In the period after 1984/85 the high Australian real rates coincided with a period of expected real exchange rate depreciation. The large fall in real interest rates in 1987/88 coincided with the strongly appreciating real exchange rate.

This section has presented a variety of data on the Australian economy which was summarized at the beginning of this section. In the following figure we attempt to explain how it all fits together.

Footnotes

See Corden (1988) for an excellent description and interpretation of the experience between 1970 and 1985 and a comparison of the Australian experience with other countries. The focus in the current paper differs from Corden's in emphasis but not in overall conclusion. The reader is also referred to Gruen (1986) for a discussion of Australia's growth performance, which is not explicitly considered here. [8]

The peak of the trade balance surplus corresponded to a period of exceptionally strong commodity prices. If this is taken into account, the decline in the trade balance is not clearly apparent until 1979/80. [9]

Which series is appropriate for constructing a real interest rate depends on the purpose. If it is to reflect the real interest rate facing consumers in deciding whether to consume today or tomorrow, it should be a price index for the consumption bundle. On the other hand for a firm facing an investment decision the appropriate opportunity cost of producing today or tomorrow is the change in the GDP deflator. [10]

Defined as the average of the four quarterly observations on interest rates where each quarterly rate is measured as the rate at the end of the quarter. The method of calculating after-tax rates of interest is outlined in Morling (1990). [11]

See Bosworth (1982). [12]

The series is the annual average of the quarterly interest rate series. [13]

Smith and Gruen (1989) give a detailed analysis of the ex-post breakdown of this since 1985. The continuing large interest differential from 1985 suggests that market participants increasingly expected a decline in the Australian dollar which did not emerge. [14]