RDP 9607: Towards an Understanding of Australia's Co-Movement with Foreign Business Cycles Appendix A: Empirical Estimates of Australia's Co-Movement With Foreign Business Cycles

Barry and Guille (1976)

Methodology Simple correlation between the Australian and world business cycles. Business cycle defined as deviation from trend through peaks in industrial production. Defined ‘world’ business cycle as trade-weighted industrial production. Sample: Quarterly, 1959–1974.

Results Australian business cycle lags by 2 – 3 quarters the world business cycle. Evidence that the lag has reduced over time.

Backus and Kehoe (1992)

Methodology Contemporaneous correlation of HP filtered logarithm of real output. Sample: Annual for more than 100 years.

Results Interwar years a high correlation (greater than 0.3) for Canada, Germany, Italy, Sweden, United Kingdom, and United States. Postwar a high correlation found only for Germany, Japan, Sweden, and United Kingdom.

Haslem, Hawkins, Heath and Tarditi (1993)

Methodology Test for ‘common features’ using Engle and Kozicki (1993) technique. Uses industrial production as the measure of the business cycle. Sample 1984:Q1 to 1993:Q1.

Results VAR(1) common feature between both the US and OECD business cycles and the Australian business cycle. A 1 percentage point increase in the OECD production growth leads to a 0.95 percentage point increase in the Australian industrial production.

McTaggart and Hall (1993)

Methodology 2 step error-correction model. First step between Australian and US GDP. Second step was to explain Australian GDP growth by the growth in US GDP and the error correction. Sample: 1967:Q1 to 1991:Q1.

Results Cointegration found between the level of Australian and US GDP. Coefficient on growth in US GDP is 0.5 and on the ECM −0.48.

Downs, Louis and Lay (1994)

Methodology Single equation GDP. Sample: 1972 to 1994.

Results Estimates US GDP growth has a larger and more significant (but still very small) impact on Australian GDP growth following 1983.

Gruen and Shuetrim (1994)

Methodology Single Equation Error-Correction Model. Sample: 1980:Q1–1993:Q4.

Results High contemporaneous impact of world output on domestic growth with a coefficient between 0.4 and 0.6 depending on the model and the measure of foreign activity. Coefficient on ECM is between −0.15 and −0.3. Finds small effect of terms of trade and real exchange rate on output growth.

Smith and Murphy (1994)

Methodology Johansen VAR. Sample: 1976:Q1 to 1990:Q1.

Results Attributes 2 per cent of the variance in GDP growth to international activity, 73 per cent to domestic demand and 15 per cent to real wage shocks.

Debelle and Preston (1995)

Methodology Unrestricted ECM between Australian GDP, OECD GDP and components of Australian GDP. Sample: 1971:Q2 to 1994:Q4.

Results Looks for cointegration. Finds little impact on components of GDP. Exports not cointegrated with OECD GDP. Some evidence of transmission through business confidence.