# RDP 9607: Towards an Understanding of Australia's Co-Movement with Foreign Business Cycles 2. The Australian and Foreign Business Cycles

Barry and Guille (1976) demonstrated correlation between the Australian and foreign business cycles.[1] They suggested that transmission was through the balance of payments. More recently, McTaggart and Hall (1993) and Gruen and Shuetrim (1994) have employed cointegration techniques to investigate the relationship between the domestic and foreign economic cycles to determine the long-run relationship between domestic and foreign activity. Gruen and Shuetrim (1994) provide the best empirical description of the correlation in business cycles when they estimate the following model of the Australian business cycle:

where y is Australian GDP, R is the real ‘cash’ interest rate, SOI is the Southern Oscillation Index to capture the effect of weather patterns on farm output, tot is the terms of trade, rtwi is the real trade weighted index of the exchange rate and yf is foreign GDP. The lower case variables are in logs and Δ is the change in the variable.

The Gruen and Shuetrim results are reproduced in Table 2. For our purposes, three results are important. First, the contemporaneous growth in foreign output is highly significant and large (0.4 or greater) in all the error-correction models of Australian GDP (2 through 7). Second, in the models which also incorporate the terms of trade and real exchange rate, the level of Australian and foreign GDP appear to be cointegrated. That is, a stable long-run relationship between the level of Australian and foreign GDP exists in the data. However, in two of the models (2 and 4), the terms of trade and real exchange rate do not have the expected signs and the probability of cointegration is reduced considerably if the terms of trade and real exchange rate are excluded.

OECD US Export markets Variable model: 1 2 3 1.61** (3.70) −27.25** (−3.90) −15.42** (−2.86) −31.71** (−4.01) −17.32** (−3.24) 15.20 (1.91) 9.75 (1.11) −0.027 {0.00} −0.057 {0.00} −0.035 {0.00} −0.033 {0.00} −0.021 {0.01} −0.035 {0.00} −0.037 {0.00} 0.011 {0.04} 0.017 {0.08} 0.011 {0.04} 0.010 {0.19} 0.007 {0.05} 0.017 {0.05} 0.012 {0.01} 0.030 {0.17} −0.060 {0.10} −0.069 {0.06} 0.004 {0.82} −0.023 {0.07} 0.022 {0.03} 0.047 {0.03} −0.012 {0.30} −0.29** (−3.69) −0.20* (−2.31) −0.31** (−4.12) −0.19** (−2.58) −0.19* (−2.49) −0.14 (−1.53) 0.35** (3.84) 0.24* (2.43) 0.38** (4.21) 0.23** (2.76) 0.16** (2.65) 0.12 (1.67) 1.22** (5.16) 0.84** (4.90) 0.60** (5.22) 0.40** (4.96) 0.50* (2.55) 0.55**(4.27) 0.47 0.65 0.56 0.67 0.56 0.53 0.47 0.09 0.34 0.46 0.37 0.47 0.10 0.35 85.24 {0.00} 45.58 {0.00} 86.68 {0.00} 39.94 {0.00} 6.79 {0.15} 11.80 {0.02} 3.50 {0.48} 15.07 {0.00} 9.91 {0.04} 10.34 {0.04} 3.02 {0.55} 16.92 {0.00} 18.85 {0.00} 22.26 {0.00} 17.32 {0.00} 23.71 {0.00} 17.25 {0.00} 20.02 {0.00} 0.16 {0.93} 0.28 {0.87} 2.44 {0.29} 0.59 {0.74} 3.26 {0.20} 0.17 {0.92} 0.46 {0.80} Notes: (a) Numbers in parentheses () are t-statistics. Numbers in brackets {} are p-values. Individual coefficients marked with *(**) imply that the coefficient is significantly different from zero at the 5%(1%) level. Standard errors are estimated using a Newey-West correction allowing for fourth order residual correlation. All variables in log levels and their differences are multiplied by 100 (so growth rates are in percentages). (b) The mean coefficient is reported for the real cash rate, the Southern Oscillation Index, the terms of trade and the real TWI to summarise the coefficients on these variables. The p-values are derived from chi-squared tests of the joint significance of the lags. (c) The SOI (Southern Oscillation Index) measures the sea level barometric pressure differential between Darwin and Tahiti. If the index is positive, trade winds are stronger and rainfall in Australia is more plentiful. If the index is negative, the trade winds are weaker and less rain occurs. The quarterly figures are the average daily value of the index throughout that quarter. # Reproduced from Table 2 in Gruen and Shuetrim (1994).

Finally, it appears the US-based model performs as well as or better than the OECD model, and substantially better than the model based on export-markets' GDP.

The striking feature of these results is not that foreign activity affects the Australian economy but how large and immediate the impact is.[2] Two further results support this finding. First, a simple unrestricted error-correction model of US and Australian GDP indicates that deviations from the long-run relationship between US and Australian GDP do not affect US GDP. This implies, as would be expected, that it is Australian and not US GDP which adjusts to remove the disequilibrium from the long-run relationship. The Gruen and Shuetrim relationship, therefore, reflects causation and not simply correlation. Second, if we abstract from the trending nature of the GDP variables by defining the business cycle with reference to an output gap, the results, reported in Appendix B, are very similar.

At odds with the ‘correlation implies causation’ view outlined above, are two papers which argue that foreign activity does not have large direct effects on domestic activity. Smith and Murphy (1994) find that foreign activity has little impact on Australian activity. Instead, they argue that variation in Australian GDP growth has been driven by domestic factors, namely real wage shocks and domestic demand. Downs, Louis and Lay (1994), using a single equation model of GDP growth, also conclude the direct effect of US activity on the Australian business cycle is small. However, they do acknowledge that the impact is larger and more significant since 1983. This period coincides with significant trade and financial market liberalisation, a fact which may point to likely explanations of the correlation in business cycles.

Explanations offered for the correlation in business cycles can be characterised by two schools of thought.[3] The first school focuses on the output and/or income effects of foreign business cycles. These affect activity either indirectly, through the terms of trade, or directly through Australia's exports.[4] The terms of trade mechanism in turn operates through two channels: an export supply response to changing export prices, and an income effect which leads to changes in domestic demand. The impact of the terms of trade on domestic activity through these channels may however have been blunted since the floating of the Australian dollar, since currency fluctuations have tended to be positively correlated with the terms of trade.[5] This leads to higher export prices being offset by the exchange rate and some of the increased demand being satisfied by imports rather than domestic output. As a consequence, while there is some evidence that the terms of trade affect Australian GDP, the effect does not appear to be large.[6]

Evidence concerning direct effects of foreign activity on exports is also unclear. Gruen and Shuetrim (1994) argue that because Australia's business cycle is better explained by US or OECD activity than by activity in Australia's trading partners, the transmission mechanism is not through exports. Debelle and Preston (1995) reach a similar conclusion using a disaggregated approach which fails to find a significant direct effect of foreign GDP on exports. A criticism of this approach is that, while a long-run relationship between Australian and foreign GDP may be expected, there is no obvious reason to expect the relationship between the components of domestic GDP and foreign GDP in aggregate to be stable. This is especially true if the export share of GDP is undergoing a structural shift.

The second school of thought seeks explanations based on common factors which affect different countries simultaneously. A number of possible explanations come within this general approach. One group of these focuses on the integration of domestic and foreign financial markets. Events which affect the large foreign financial markets (such as changes in monetary policy and expectations for growth, profitability and inflation), might flow through onto the financial markets of smaller countries; in this way, the smaller country would ‘adopt’ the business cycle of the country with the larger financial market.[7] Alternatively, foreign share markets might affect Australian activity through their direct impact on the cost of raising funds, which may depend, in part, on the performance of foreign share markets. This link could be strengthened by foreign ownership of Australian companies, since domestic subsidiaries may have greater access to low cost internal funds or equity finance when the parent company is highly profitable and its share price is high.[8]

Another explanation which centres on common causal factors is the impact of worldwide supply shocks. Examples include the positive oil price shocks of the 1970s, the negative oil price shock in 1986, and shocks to technology. Partly due to the difficulty in comprehensively identifying the shocks, this explanation has received little attention. Finally, Debelle and Preston (1995) suggest an influence through the effect of foreign business cycles on Australian business confidence. However, they acknowledge that an explicit role for business confidence in determining Australian investment is more difficult to identify.

Two of the linkages between the foreign and domestic business cycles appear particularly relevant from this review of the literature and are investigated further below. The first linkage is through exports, commonly regarded as the main channel through which business cycles are transmitted internationally. Given the dominance of this view and the increased openness to trade of the Australian economy, the next section considers the relationship between foreign activity and exports in detail. The second linkage, through share markets, is relevant due to the increasing integration of the Australian and foreign markets, and is considered in Section 4.

## Footnotes

Appendix A provides a synopsis of recent empirical work on the impact of foreign business cycle on the Australian economy. [1]

Gruen and Shuetrim estimate the contemporaneous impact of US GDP growth on the growth in Australian GDP to be between 0.4 and 0.6, depending on the model, which is consistent with McTaggart and Hall's (1993) estimate of 0.5. [2]

A summary of the literature on possible transmission mechanisms is provided in Appendix C. [3]

For example, see Pitchford (1992, 1993), Gruen and Shuetrim (1994), and Debelle and Preston (1995). [4]

See Blundell-Wignall and Thomas (1987), Gruen and Wilkinson (1991) and Tarditi (1996). [5]

Downes, Louis and Lay (1994) and Gruen and Shuetrim (1994) model the impact of the terms of trade on gross national expenditure and GDP. They conclude that the terms of trade has a large and significant impact on gross national expenditure but an insignificant impact on GDP. By implication, these results suggest any increase in exports due to a rise in the terms of trade is offset by a reduction in Australian output for domestic purposes. The increased domestic demand is then satisfied by imports. [6]

Fama (1981), Geske and Roll (1983), Kaul (1987) and Barro (1990) find that stock returns help predict future real activity. Furthermore, Canova and De Nicolo (1995) find expected US GNP growth helps predict European stock returns which in turn helps to explain future European GNP growth. [7]

Fazzari, Hubbard and Petersen (1988) and Froot and Stein (1991) argue that the cost of internal funds is less than external finance and show investment is sensitive to the availability of internal funds. Froot and Stein (1991) also argue that companies with higher relative wealth are more likely to engage in foreign direct investment. [8]