RDP 2015-11: Unprecedented Changes in the Terms of Trade 1. Introduction

Over the past decade or more, the ongoing development of Asia has led to a surge in global resource commodity prices. There have been commodity price booms before, but this one, fuelled by an unprecedented era of high growth in China, has been by far the largest and most persistent.

A recurring question for commodity-exporting economies is the extent to which the recent increases in commodity prices will prove to be permanent.^[1] One view is that these fluctuations are the result of an unprecedented and prolonged shift in the demand for commodities associated with the economic development of Asia. According to this view, the level of commodity prices has reached a permanently higher ‘new normal’.^[2] Others argue that the increase is not permanent as higher prices will eventually induce increased supply, but with some lag. This view also implies a ‘new normal’ for commodity prices, but one associated with greater volatility due to the emergence of large new sources of global commodity demand coupled with the inelastic nature of short-run commodity supply curves, rather than a permanently higher level.^[3] Understanding whether a ‘new normal’ exists, what it entails and its implications is important for these economies.

In this paper, we take these competing hypotheses to the data. We first set up a model with multiple productivity trends that can capture steady-state drifts in relative prices. We then estimate the model on Australian data, allowing – but not requiring – both the long-run level of commodity prices and the volatility of shocks to commodity prices to change. A permanent change in the long-run level of commodity prices gives rise to a transition towards a new balanced growth path. The resulting economic dynamics differ greatly from those associated with a temporary change in commodity prices. By using a structural model that takes account of pre-existing trends in relative prices, we are able to distinguish between trends that belong to the balanced growth path, cycles around those trends, and fluctuations that originate from a transition towards a new balanced growth path.

We estimate the model on Australian data because Australia is a commodity-exporting economy that has benefited significantly from economic developments in Asia. Between 2003 and 2011, the foreign currency prices of Australia's commodity exports more than tripled (Figure 1). Over this period, broader measures of world commodity prices experienced similar increases, although the peak increase in Australia's commodity export prices was unusually large. This suggests that our analysis may be of interest to other commodity-exporting economies as well.

We find support in the data for both hypotheses. In particular, we find that the long-run level of Australia's commodity prices increased by 40 per cent in mid 2003 and that the volatility of shocks to commodity prices more than doubled soon after that.

An increase of 40 per cent in the long-run level of commodity prices is probably less than one would infer from visual inspection of the data. It is also less than the 100 per cent increase that one recovers from the estimation of single-equation reduced-form specifications. Our inferences, however, rely on more observable variables than single-equation estimates. In general equilibrium, a change in the long-run level of commodity prices affects quantities and prices throughout the economy. A permanent doubling of commodity prices is unlikely because it would imply behaviour of other economic variables that is greatly at odds with what we have actually observed.

The economic consequences of the 40 per cent increase in the long-run level of commodity prices that we estimate are, however, significant. To name a few: the commodity sector's share of exports increases from 35 to 52 per cent; consumer price inflation falls temporarily by 2 percentage points with tradeable inflation and non-tradeable inflation offsetting each other; the trade deficit widens by around 2 percentage points of GDP for a decade; and the real exchange rate appreciates permanently. Even if commodity prices ultimately settle far below their recent peaks, the Australian economy is likely to look different from how it did before the terms of trade boom.

Our work builds on that of Rabanal (2009) and Siena (2014), who set up models in which steady-state productivity growth differs between the tradeable and non-tradeable sectors of the economy. To these models we add capital accumulation with a differential trend in the relative price of investment goods as well as a commodity-exporting sector that takes the relative price of its output as given. Our work also relates to a large literature on the role of terms of trade shocks in open economies, to which we cannot do justice in the space we have here.^[4] Our work adds to this literature because we distinguish temporary shocks to the terms of trade – which have been the focus of the small open economy literature – from permanent shifts in the long-run level of the terms of trade. Our work is similar in spirit to that of Aguiar and Gopinath (2007), who exploit the information in consumption and net exports to identify trend growth. We use multiple observable series to identify permanent changes in the terms of trade process.

The rest of the paper is structured as follows. Section 2 discusses the model. Section 3 analyses the responses of the model to temporary and permanent changes in the terms of trade. Section 4 discusses our empirical approach, which involves calibration and estimation of date breaks and parameters. Section 5 describes the main results. Section 6 discusses the implications of these results for the Australian economy. Section 7 concludes.

Footnotes

Because for commodity-exporting economies the bulk of recent terms of trade fluctuations come from commodity prices movements, we use the terms commodity prices and terms of trade interchangeably. [1]

See Bernanke (2008), Bloxham, Keen and Hartigan (2012), Stevens (2011) and Yellen (2011). [2]

See Dobbs et al (2013). [3]

Instead, we point the reader to Ostry and Reinhart (1992); Mendoza (1995); Bidarkota and Crucini (2000); Bleaney and Greenaway (2001); Kent and Cashin (2003); Broda (2004); Blattman, Hwang and Williamson (2007); Medina and Soto (2007); Dib (2008); Jääskelä and Smith (2013); Charnavoki and Dolado (2014); and the references therein. [4]