In the past few years, the world has witnessed large swings in world relative prices, from oil, to metals, to food prices. These relative price movements are reflected in movements in the terms of trade, or the relative price of a country's exports in terms of its imports.

Conventional wisdom has it that terms of trade shocks represent a major source of business cycles in emerging and poor countries. This view is largely based on the analysis of calibrated business-cycle models.

Essentially, the result is obtained by first estimating a process for the terms of trade and then feeding it to an equilibrium business cycle model to compute the variance of macroeconomic indicators of interest induced by this type of disturbance. Then this predicted conditional variance is divided by the actual observed unconditional variance of the corresponding macroeconomic indicator to obtain the share of variance explained by terms-of-trade shocks. Consistently, this methodology arrives at the conclusion that at least 30% of the variance of output and other macroeconomic indicators is attributable to terms-of-trade shocks (Mendoza 1995, Kose 2002).

In this entry, we argue that the picture that emerges from structural vector autoregression (SVAR) analysis is quite different. We show that SVAR models predict that terms-of-trade shocks account for no more than 10% of business-cycle fluctuations in the majority of poor and emerging countries.

To identify terms-of-trade shocks, we assume that movements in the terms of trade are exogenous. This assumption, which has been embraced universally by the existing related literature whether empirical or theoretical, is motivated by the fact that the typical poor or emerging country is too small to affect world relative prices. Using data from 38 poor and emerging countries, we estimate country-specific SVAR models (Schmitt-Grohé and Uribe 2015). The data source is the World Banks' World Development Indicators (WDI) database. We define the set of poor and emerging countries as all countries with average PPP converted GDP per capita of less than 25,000 dollars of 2005 over the period 1990-2009. The SVAR system contains six variables, output, consumption, investment, the trade balance, the real exchange rate, and the terms of trade. The sample period is 1980 to 2011. The country selection is dictated by the requirement of at least 30 consecutive annual observations for all six variables.

Figure 1 presents the variance decomposition implied by the SVAR analysis in the form of a histogram. The horizontal axis measures the share of the variance of the cyclical component of real GDP per capita attributable to terms-of-trade shocks in percent. The horizontal axis is divided into six equally sized bins. The first bin contains countries for which terms-of-trade shocks explain less than 10% of the variance of output. The second bin contains countries for which terms-of-trade shocks explain between 10 and 20%, and so on. The vertical axis measures the number of countries in each bin. The figure shows that terms-of-trade shocks explain less than 10% of the variance of GDP in 19 out of the 38 countries. And only in 5 countries do terms-of-trade shocks explain more than 30%. Similar results obtain for the other variables included in the SVAR model. Specifically, for the median country terms-of-trade shocks explain 12% of the variance of the trade balance, 9% of the variance of consumption, 10% of the variance of investment, and 14% of the variance of the real exchange rate. These results suggest that there is a disconnect between theoretical and empirical models when it comes to gauging the importance of terms-of-trade disturbances in generating business cycles.

**Figure 1**. Share of variance of output explained by terms-of-trade shocks

Explaining this disconnect between empirical and theoretical models is an important item in the research agenda that lies ahead. Its resolution is likely to involve a combination of better empirical and theoretical models as means to interpret the data. For example, an improvement in the empirical model could stem from entertaining the hypothesis that commodity prices are a better empirical measure of the terms of trade than aggregate indices of export and import unit values – the measure underlying the present analysis. This is likely to be the case especially for countries whose exports or imports are concentrated in a small number of commodities. At the same time, the theoretical model could be amended by assuming that the government uses tax or commercial policy to isolate domestic markets from swings in world prices. In this case, movements in the terms of trade will elicit attenuated incentives to change the domestic allocation of output and absorption. A related reason why fluctuations in the terms of trade may not have large domestic effects could be the presence of nominal rigidities that introduce a wedge between domestic and world prices.

The jury is still out on whether future research will bring the empirical models closer to the predictions of the theoretical ones or vice versa, or whether the two will meet somewhere midway. Progress on this research agenda will add much needed light on determining more accurately how important terms-of-trade movements are in generating business cycles in poor and emerging countries.

## References

Kose, M A (2002), "Explaining business cycles in small open economies `How much do world prices matter?", *Journal of International Economics* 56, 299-327.

Mendoza, E (1995), The Terms of Trade, the Real Exchange Rate, and Economic Fluctuations, *International Economic Review* 36, February,101-137.

Schmitt-Grohé (2015), Stephanie, and Martín Uribe, How Important Are Terms Of Trade Shocks? NBER working paper 21253, June.