RDP 2004-11: Trade Openness: An Australian Perspective 4. Interpreting Australia's Low Openness Ratio

The regression results from Section 3 permit one to ask whether countries trade more or less than expected. Our specific interest is Australia. Using the 1996–2000 cross-section regression, the equation predicts an openness ratio for Australia of 41.8 per cent, which is only modestly above the actual ratio of 38.3 per cent.[28] That is, although Australia's openness ratio is relatively low by international standards, the regression results suggest that this ratio is about the level that one might expect. This is consistent with the finding in Table 2 that an Australian dummy variable is always insignificant in the cross-section and pooled openness regressions.

The estimates of the effect of various factors on openness also allow us to shed light on why countries trade as much or as little as they do. To this end, Table 7 reports how Australia compares with the sample average. The first column of data shows the difference between the Australian value and the mean value for the sample, divided by the standard deviation of the sample. Using a common unit, the standard deviation of each variable, this allows us to see whether Australia's difference from the rest of the world for each variable is large or small. The data show that Australia differs substantially from the sample average in terms of having a much larger land mass and having a less favourable (or more remote) location. In addition, Australia has a substantially higher income level and has a more liberal trade regime than most of the 120 countries in our sample, though its differences with the OECD average on these scores are much smaller. The second column is based on the 1996–2000 cross-section regression in Table 2 and shows whether a unit of one standard deviation of the variable has a relatively large or small effect on openness in the regression. Based on this measure, population is the most important variable for openness, followed by economic location and geographic size.

Table 7: Understanding Australia's Openness
  Australia's difference from average Parameter estimate for variable Implied impact on Australia's log(openness)
Relative to standard deviation of the variable
Log(total area) 1.82 −0.08 −0.14
Log(population) 0.38 −0.22 −0.08
Log(GDP per capita)(a) 0.21 na 0.03
Log(economic location) −1.68 0.16 −0.24
Trade policy index 0.97 0.07 0.05
Log(country price level) −1.09 −0.06 −0.07
Total impact na na −0.46

Note: (a) This variable captures the impact of both the unadjusted and the squared GDP per capita terms.

The last column in Table 7 is the product of the first two columns and provides an assessment of which variables are most important in explaining Australia's low openness ratio. The results suggest that an unfavourable economic location and substantial area have the largest effect in reducing Australia's openness, accounting for around three-quarters of the deviation from the sample average. Economic location is the most important variable in this regard, accounting for about half the deviation. For example, a simple counterfactual of imagining Australia near the middle of the North Atlantic Ocean, implying an improvement in our economic location to near the sample mean, would result in Australia's predicted openness ratio jumping from around 0.42 to around 0.51, or exports and imports rising from around 21 per cent of GDP to around 25–26 per cent.[29]

A final point to note, which is especially relevant for Australia's economic location, is that any measure of location or remoteness will change with shifts in the relative economic size of other countries. The experience of recent years indicates that with its relative proximity to Asia, Australia has benefited strongly from the growth of that region.[30] Looking ahead, the most obvious trend is likely to be the continuing growth of India and China, which in 2002 accounted for 38 per cent of the world's population but only 5 per cent of world GDP at market exchange rates. Australia's relative proximity to these countries suggests that its economic location is likely to improve in coming years as these economies grow.


Similarly, the individual cross-section regressions predict Australia's openness to be between 1 and 3 percentage points greater than its actual openness. Note that these regression results are based on the equations with the squared per capita GDP variable and price level variable, and so are not exactly comparable with the results shown at the bottom of Table 2. [28]

Battersby and Ewing (2003) have also noted that if Australia had the same location as the UK, Australian aggregate trade would increase by approximately 40 per cent. [29]

See Macfarlane (2003) for further details of the growth in Australia's trade with east Asia. [30]