RDP 2001-02: Changes in the Determinants of Inflation in Australia 2. Inflation Performance

Australia's inflation performance over the past four decades is illustrated in Figure 1. Two measures of inflation are shown: the Treasury underlying series which, until recently, was the main measure of core inflation in Australia, and the Statistician's new inflation series.^[4] The shaded band indicates the current target of 2–3 per cent inflation over the medium term that has been the objective of monetary policy since 1993. There are several striking features of the figure. When viewed over the longer run, inflation has been highly variable, so that the 1990s emerge as a period of unusual stability. Also striking is that since the 1970s, there has been a tendency for each cyclical peak in inflation to be lower than the one before it. Since the early 1990s, inflation has been maintained at a rate not witnessed for more than a generation. Furthermore, in the period since the adoption of the inflation target, it is clear that inflation has been below 2 per cent for longer than it has been above 3 per cent.

Many of these features are not, however, unique to Australia. Comparing inflation performance in Australia with that in the OECD, it is clear that there is a fair degree of similarity (Figure 2).^[5] The timing of most major swings in inflation is roughly coincident, reflecting some commonality of shocks, rough correlation of business cycles and broad similarities in the stance of monetary policy. A trend reduction in inflation since the 1970s is common to many countries. So too is the shift to a regime of low, stable inflation and the tendency for inflation outcomes to be either below target or in the lower part of tolerance intervals. However, until recently, inflation in Australia tended to be higher, on average, than in the OECD and subject to greater oscillations.

Stevens (1992) provides a detailed episodic analysis of inflation in Australia for the period 1950–1991 and demonstrates that, as a small open economy, foreign shocks have played a prominent role in inflation outcomes. This is particularly evident during the 1970s when there were several major commodity price shocks. Under a fixed exchange rate regime, higher world prices were transmitted to domestic prices both directly, and indirectly through their effects on income. This partly explains why oscillations in inflation during the 1970s were higher in Australia than in the OECD generally. The adoption of a floating exchange rate (in 1983) has made it easier to insulate the economy from foreign shocks, but they remain a significant influence on domestic prices.^[6] They explain much of the difference between Australian and OECD inflation during the mid 1980s when, following a sharp fall in the terms of trade, Australia experienced a record depreciation of the currency.

While episodes of inflation in Australia have tended to be triggered by foreign shocks, domestic shocks have also played an important role, often amplifying the effects of external factors. In particular, outbreaks of wage inflation occurred in the early to mid 1970s, the late 1970s and early 1980s, that were more pronounced than in other OECD countries.^[7] These provided a powerful impulse to inflation, leading to a more exaggerated cycle in inflation in Australia than in most of the OECD.^[8]

While Australian inflation has since fallen, and converged with that of other industrialised nations, the inflation process has remained subject to shocks. Volatile currency movements continue, as do fluctuations in wages growth and there has been a pronounced and sustained upswing in real output. Certainly, the absolute size of these shocks has tended to be smaller in the past decade than previously. But notwithstanding this, the disturbances to inflation in Australia have been less than a casual reading of history would suggest. One very simple way of demonstrating this is to compare the variability of inflation with the variability of the inflationary shock. As Table 1 shows, the standard deviation of inflation fell significantly in the 1990s, while the standard deviation of import prices fell by a much smaller proportion, so that the ratio of the shock to the inflation outcome rose. A similar result can be found for the output gap and, to a lesser extent, wages in the manufacturing sector (often the wage leaders).^[9] At face value, this points to an increased resilience of the inflation environment to shocks.

So why does the experience of the past decade appear to be so different? The framework of inflation targeting has no doubt been helpful, providing a new price stability rule where others had broken down and an anchor to inflation expectations (Grenville 1997; Clarida, Gali and Gertler 1998; Taylor 2000). Furthermore, it has encouraged policy-makers to be more disciplined and forward looking.^[10] In doing so, it has demanded greater effort in forecasting inflation. But despite these efforts to account for influences on the inflation process, actual inflation has been surprisingly low.

The extent of this surprise in Australia can be summarised by an Assistant Governor of our Bank. Commenting on the fact that inflation has been below target more often than above it, he says ‘I cannot recall anyone predicting this outcome when we set out on the track of having an inflation target; in fact, I cannot recall anyone even contemplating it as a serious possibility. The scepticism we faced for quite some time over whether we would be able to prevent a return to high inflation seems like another world now’ (Stevens 1999, p 50).

Having highlighted the apparent increase in resilience to foreign and domestic shocks, in the following section we explore developments in some of the main influences on inflation, focusing on the experience of the past decade.

Footnotes

Following a regular five-year review of the Consumer Price Index by the Statistician, the measurement of the CPI was changed from the outlays approach (in which interest charges were included) to the acquisitions approach (in which they are not). For a description of these and other data see Appendix A. [4]

To permit a direct comparison with inflation in the OECD, CPI less interest is the measure of Australian inflation. (It peaks at a lower rate in the 1970s than our narrower measure of core inflation because it includes some items that experienced relatively low price rises in the September quarter 1974.) [5]

Gruen and Dwyer (1995) describe the mechanisms through which a terms of trade shock influences domestic prices under a fixed and floating exchange rate and present empirical evidence of the relationship for Australia and how it has evolved. [6]

The first reflected the government's intention to increase labour's share of income and establish ‘equal pay for equal work’, while in the late 1970s labour secured wage rises outside the centralised system. In the early 1980s, formal centralised wage fixing was abandoned but coordinated union campaigns resulted in a generalised wage rise. [7]

In contrast, the Prices and Incomes Accord secured generalised wage restraint during the second half of the 1980s, which helped to counteract the inflationary consequences of prevailing external shocks. [8]

For aggregate wages, the relationship appears to be unchanged. The stability of this relationship may reflect that much of a given rise in wages is an endogenous response to rising prices. [9]

These issues are explored comprehensively elsewhere (see Lowe (1997)). For a comment on how the target has conditioned Australian decisions on monetary policy specifically, see Stevens (1999). [10]