RDP 8906: A Random Walk Around the $A: Expectations, Risk, Interest Rates and Consequences for External Imbalance IV. Inflation and Interest Rates
October 1989
- Download the Paper 1.72MB
Figures 5 and 6 display, respectively, short-term nominal interest rates and 12 month ended CPI inflation rates for seven OECD countries over the period 1975 – 1989.^{[18]} The expected short-term real interest rate differential between each country j and the US, r^{j}_{diff}, is given by
where Eπ^{J} is expected inflation in country j. Figure 7 shows this differential using CPI inflation over the previous 12 months for Eπ^{J}.^{[19]}
Over the 59 months Nov 84 – Sept 89, this measure of the short-term real interest differential between Australia and the US was positive for all but four months, and averaged +2.6% p.a. An alternative measure of expected real interest rates can be derived using the realized inflation rate during the ex post 3 month period as a proxy for expected inflation. By this measure, over the nineteen quarters from Dec Q 84 to Jun Q 89 the short-term real interest differential between Australia and the US was positive for sixteen quarters, and averaged +2.4% p.a.^{[20]}
Frankel and MacArthur (1988) used this approach to measure the 3 month real interest differential between 24 countries and the US over the period Sept 82–Oct 86. Excluding the four closed economies in their sample (Bahrain, Greece, Mexico and South Africa), four of the remaining twenty countries (Austria, Denmark, Hong Kong and Switzerland) have an average real interest differential more than 2% p.a. below the US, while none have a real interest differential more than 2% p.a. above the US. As Frankel and MacArthur point out, this asymmetry occurs because of the high real interest rates in the US over this period. But the US continues to experience fairly high real interest rates.^{[21]} Hence, this comparison with Frankel and MacArthur emphasises how unusually high Australian short term real interest rates have been since late 84.
A different perspective is provided by Table 4 which shows the excess nominal return a US investor would have achieved by investing (and continually rolling over the investment) in Australian 3 month Treasury bills rather than US bills. Of course, the results include the exchange rate change which occurs between the purchase and the sale of the bills.^{[22]} Over thirty years, from Jan 60 to Sept 89, the average excess return from investing in Australian short-term bills has been negligible (0.3% p.a.). But the Australian real interest premium since late 84 would have returned a US investor a nominal excess return of 5.7% p.a. on an investment from Jan 85 to Sept 89 – because over this time the $A experienced real appreciation against the $US at an average rate of 3.0% p.a. Even from Jan 84 – which includes virtually the whole period of the $A float – the excess return has averaged 3.3% p.a. as over this time the $A appreciated against the $US at an average real rate of 1.2% p.a.
To the end of: | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
1964 | 1969 | 1974 | 1979 | 1984 | 1985 | 1986 | 1987 | 1988 | Sep 89 | |
From the start of: 1960 | 0.6 | 0.0 | 1.1 | 0.3 | −0.7 | −1.2 | −0.9 | −0.3 | 0.4 | 0.3 |
1965 | −0.5 | 1.4 | 0.2 | −1.0 | −1.6 | −1.2 | −0.5 | 0.4 | 0.2 | |
1970 | 3.4 | 0.6 | −1.2 | −1.9 | −1.4 | −0.5 | 0.7 | 0.4 | ||
1975 | −2.2 | −3.4 | −4.2 | −3.4 | −2.0 | −0.3 | −0.6 | |||
1980 | −4.6 | −5.8 | −4.2 | −1.9 | 0.8 | 0.3 | ||||
1984 | −7.1 | −9.4 | −4.5 | 0.3 | 4.7 | 3.3 | ||||
1985 | −11.7 | −3.1 | 2.8 | 7.9 | 5.7 | |||||
1986 | 6.3 | 11.0 | 15.3 | 10.8 | ||||||
1987 | 15.9 | 20.1 | 12.5 | |||||||
1988 | 24.5 | 10.7 | ||||||||
1989 | −5.4 |
Footnotes
The interest rates are 3 month Treasury bill rates for Australia, US and Canada; the 6 month Treasury bill rate for Italy, and 3 month Eurocurrency rates for Japan, West Germany and United Kingdom. For Australia, the CPI numbers are quarterly, while for all other countries, they are monthly. The Australian numbers make adjustment for the ‘Medicare effect’, which increases the estimated Australian inflation rate from Mar Q 84 to Mar Q 85. See the data appendix for data sources for this section. [18]
To compare like with like, we use the 3 month US Eurocurrency interest rate for i^{us} when evaluating the real interest differential for Japan, West Germany and United Kingdom. [19]
Estimated in the same way over the same period, the real interest differential on 10 year government bonds between Australia and the US averaged – 0.6% p.a. Since we have some evidence for a risk premium in 84–85, but not since then, it is interesting to evaluate the ex post real interest rate differential between Australia and the US over the fourteen quarters Mar Q 86 to Jun Q 89. It averaged 2.8% p.a. for 3 month Treasury bills and 0.2% p.a. for 10 year bonds. McKibbin and Morling (1989) examine alternative measures of the real interest rate differential for 90 day bank bills between Australia and US. From Dec Q 84 to Mar Q 89, using the quarterly change in the GDP (GNP) deflator for Australia (US) to estimate expected inflation, gives an average real interest rate differential of 2.3% p.a. Changing to a forecasting equation to estimate expected quarterly Australian GDP inflation (CPI inflation) leads to an average real interest differential over the period of 1.7% p.a. (2.5% p.a.). Finally, using a forecasting equation to estimate quarterly US GNP inflation and the forecasting equation to estimate expected quarterly Australian GDP inflation (CPI inflation) leads to an average real interest differential over the period of 1.9% p.a. (2.7% p.a.) – Steve Morling, personal communication. [20]
The ex post real US interest rate on 3 month Treasury bills averaged 5.4% p.a. from Sept 82 to Oct 86, 3.3% p.a. from Oct 84 to Jun 89, compared to an average of 2.0% p.a. from Jan 75 to Jun 89. [21]
We are grateful to George Fane for suggesting this Table. It is constructed using end year exchange rates and annual average yields on 3 month Treasury notes from Tables S.7 and S.11 in Norton and Aylmer (1988) and from more recent RBA Bulletins. [22]