5. Is Latin America Saving Its Terms-of-Trade Windfall? A Metric

International Monetary Fund. Western Hemisphere Dept.
Published Date:
May 2013
  • ShareShare
Show Summary Details

On the back of a sizable terms-of-trade boom, Latin America’s fundamentals have improved markedly during the last decade. This has fed a sense of complacency that this time the macroeconomic response has been indeed different. Against this background, we propose a simple metric to quantify the terms-of-trade income windfall of the recent boom, and compare it with previous episodes. We find that while the recent terms-of-trade shock is not much larger than those observed during the 1970s, the associated income windfall has been far greater. Moreover, although aggregate saving increased more than in past episodes, the share of the windfall saved appears to be smaller. This suggests that stronger fundamentals reflect mostly the sheer size of the recent shock rather than a greater effort to save the windfall. Finally, our estimates suggest that, historically, using the windfall to increase domestic investment has been less beneficial to post-boom income than saving it in foreign assets. This raises questions about the recent weakening of external current accounts in Latin America.


Commodity exporting countries in Latin America have benefited strongly from the commodity price boom that began around 2002. Along with broadly prudent policies, the associated terms-of-trade boom allowed most countries to markedly strengthen their public and external sectors’ fundamentals (Figure 5.1). This has fed a sense that this time the macroeconomic response to the terms-of-trade shock has been different (and more prudent) than in past episodes. Whether that is the case remains an open question, which we address below.

Figure 5.1.Emerging Latin America: Terms of Trade and Selected Fundamentals, 2002–121

(Simple averages and 20th and 80th percentiles)

Sources: IMF, International Financial Statistics; and IMF country desks.

1 Emerging Latin America includes Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Mexico, Paraguay, Peru, Uruguay, and Venezuela.

2 Of goods and services.

The Terms-of-Trade Windfall—a Historical Perspective

To shed some light on this issue, we study episodes of large terms-of-trade booms from 1970 to the present, for a sample of 180 countries. The episodes are defined as those events where countries experience a cumulative terms-of-trade increase of at least 15 percent, and at least 3 percent per year on average. These simple thresholds identify 270 episodes, encompassing low-income countries, emerging market economies, and advanced economies. A first glance at the historical data shows that, while sizable, Latin America’s recent terms-of-trade boom has not been larger than those seen in the 1970s. Furthermore, the region’s recent boom is comparable only with those of oil-exporting countries in the Middle East and North Africa (MENA) region (Figure 5.2). The income impact of these terms-of-trade shocks, however, has increased over the course of the last four decades.

Figure 5.2.Emerging Latin America and Selected Regions: Terms-of-Trade Booms, 1970–20121

(Percentage change, cumulative during upswing)

Sources: IMF, International Financial Statistics; and IMF staff estimates.

1 Cumulative percentage change in terms of trade (of goods and services) from start to peak of each identified episode (that meets the criteria of at least 15 percent cumulative and 3 percent average increase). Episodes are grouped in 5-year windows according to the date of their first year. Dotted lines indicate group averages.

2 Emerging Latin America includes Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Mexico, Paraguay, Peru, Uruguay, and Venezuela.

A Metric of the Terms-of-Trade Windfall

To quantify the impact of terms-of-trade variations on aggregate income, we propose a simple metric. The metric focuses on the difference between actual real (gross domestic) income and a measure of real income at pre-boom terms of trade (see Figure 5.3 and Annex 5.1). This metric of income windfall takes into account the degree of trade openness of the economy and quantifies the ‘extra’ income arising from the terms-of-trade price-effects only (as a share of what income would have been had no shock occurred). It provides a lower bound estimate of the effect of the shock.1

Figure 5.3.Selected Latin American Countries–Recent Episodes: Real Domestic Income1

(Index T-1 = 1)

Source: IMF staff estimates.

1 Episode first year is reported in parenthesis.

2 Real gross domestic income, as defined in Annex 5.1.

3 Real gross domestic income projected at long-term growth rate (average of 1970–2012).

4 Calculations are based on official data.

This is a key innovation relative to other recent studies of the impact of external factors on Latin American economies, which have focused on the effect on output, rather than on income.2 The importance of focusing on income is evident from Figure 5.3. This is especially true for countries which output has not grown faster during the recent terms-of-trade boom than previously anticipated (e.g., Bolivia and Chile), in part because of the deployment of countercyclical policies.

A comparison of the cumulative income windfall across episodes points to a much larger effect in the recent episode than in past ones (Figure 5.4) on account of the higher degree of trade openness and the longer duration of the boom. Estimates of the income impact are sizable, implying an average increase in income of close to 15 percent per year during the recent episode. Furthermore, given the length of the latter, these estimates mean an average of about 100 percent of a year’s GDP cumulative windfall over the entire episode. Within the region, Bolivia, Chile, and Venezuela stand out as having benefited the most. Cumulative (annual average) windfalls for Venezuela reached 300 (30) percent of income and close to 200 (20) percent in the cases of both Bolivia and Chile. Not surprisingly, Brazil stands at the other extreme of the distribution, with significantly lower windfall estimates, reflecting a lower dependence on commodities and a smaller degree of trade openness.3

Figure 5.4.Emerging Latin America and Selected Regions: Income Windfall, 1970–20121

(Share of annual GDP)

Sources: IMF, International Financial Statistics; and IMF staff estimates.

1 Cumulative and annual average income windfall, as share of GDP, from start to peak of each identified episode (that meets the criteria of at least 15 percent cumulative and 3 percent average increase). Episodes are grouped in 5-year windows according to the date of their first year. Bars indicate cumulative values, dots indicate annual averages, and dotted lines indicate group averages for cumulative values.

2 Emerging Latin America includes Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Mexico, Paraguay, Peru, Uruguay, and Venezuela.

A Measure of Windfall Saving

Has Latin America Saved More of the Windfall This Time?

A glance at aggregate saving rates suggests that, compared with the past, the region’s response to the recent terms-of-trade episode has been more prudent (Figure 5.5).4 The median saving rate has increased by about 4–5 percentage points of GDP, as opposed to 2–3 percentage points in past episodes. This has been accompanied by a remarkable increase in the investment rate, in clear contrast with the past. Current accounts improved during the first stages of the episode, yet have deteriorated more recently as investment has outpaced saving, also in contrast to past episodes.

Figure 5.5.Emerging Latin America: Aggregate Saving, Investment, and Current Account during Terms-of-Trade Booms1

(Percent of GDP; medians and 25th and 75th percentiles)

Sources: IMF, International Financial Statistics; and IMF staff calculations.

1 Episode length is normalized and series are reported at fractions of the lifetime of each event.

At the same time, the public sector appears to have taken a more prudent approach to the use of the windfall than the private sector (Figure 5.6).

Figure 5.6.Emerging Latin America: Saving, Investment and Balance

(In percent of GDP, change from T-1, group simple averages)

Sources: IMF, International Financial Statistics; and IMF staff estimations.

These measures of average saving and investment rates, however, are affected by the size of the income windfall and do not inform us about how much of the windfall has been saved (i.e., the ‘marginal’ rates). To get a sense of the latter, we compute marginal saving (investment) rates, which measure the increase in saving (investment) as a proportion of the estimated income windfall (see Annex 5.1 for details).

Subsequently, we compute the marginal rates of domestic and foreign saving (i.e., how much is allocated to domestic capital formation and to foreign asset accumulation, respectively). Key findings are summarized as follows:

  • Latin America’s marginal saving rates have been lower in the recent episode than in past ones. This suggests that the ‘effort’ to save the windfall has not been necessarily stronger this time (Figure 5.7).

  • Furthermore, a glance at the dynamics of marginal saving rates (not shown) point to a gradual decline over time.5

  • At the same time, a growing share of the windfall is being allocated to domestic capital formation rather than to improving countries’ net foreign asset position (via a strengthening of the current account). This pattern may be consistent with a need to accumulate physical capital. Yet, it remains an empirical question whether domestic investment or foreign asset accumulation is preferable (in terms of increasing post-boom income) during periods of terms-of-trade booms. This is studied next.

Figure 5.7.Latin America and Other Groups-Current Episode: Windfall Saving1

(Cumulative, as a share of income windfall; medians and 25th and 75th percentiles)

Sources: IMF, International Financial Statistics; and IMF staff estimations.

1 Marginal savings rates as defined in Annex 5.1.

Windfall Saving and Post-Boom Income

We assess the effects of different saving patterns during the boom on the level of post-boom income by way of a cross-section econometric exercise.

The following specification is estimated using Ordinary Least Squares (OLS):

where RGNDYiPost denotes real gross national disposable income for episode i, measured 5 years after the boom’s peak, and ws stands for the corresponding windfall saving.

We use RGNDY, as it is a broad measure of real income, which includes net factor and financial income from abroad (the income balance of the external current account). As such, it takes into account the net income associated with changes in the country’s net external asset position.

Regressions include a number of country-specific controls X (e.g., terms-of-trade, pre-boom growth and de facto exchange rate regime) and global controls Z (e.g., U.S. real interest rates and world GDP growth), which are measured as averages over the 5 years after the boom. We also control for the level of real income at the peak of the boom. To assess the effects of domestic versus foreign saving, the regression is subsequently modified, decomposing the contribution of each of them, as follows:

where wsD and wsF denote windfall saving allocated to (domestic) investment and foreign assets, respectively. Finally, a Latin America dummy and its interaction with the windfall saving measures are included in alternative specifications to study if the effects have been different for the region.

Results reveal, as expected, that a higher windfall saving (during the boom) increases post-boom real income (Table 5.1, column 1). More important, the composition of the windfall saving matters. Results (column 2) point to a substantially higher payoff from allocating saving to foreign assets than to domestic investment (despite the fact that the sample is mostly composed of developing economies).6 These results appear to be particularly strong for Latin America (columns 3 and 4), likely reflecting a history of poor performance following terms-of-trade booms in previous decades. While conditions may be different this time, including because of more flexible policy frameworks, these results suggest that, in the current context, the deteriorating external current account balances in Latin America could be a source of concern worth monitoring.

Table 5.1.Effects of Windfall Saving on Post-Boom Real Income1
Dependent Variable:

Real Income 5 years after Peak
Windfall saving0.12 ***0.14 ***
Domestic windfall saving0.060.16
Foreign windfall saving0.13 ***0.14 ***
Dummy Latin America−2.67−1.46
Windfal saving * dummy LA−0.24 *
Domestic saving * dummy LA−0.43 *
Foreign saving * dummy LA0.00
Adjusted R20.630.640.640.64
Source: Authors’ estimations.Confidence level: (*) denotes 10 percent, (**) 5 percent, and (***) 1 percent.

Concluding Remarks

The chapter presented simple metrics of windfall income and saving that allow to compare terms-of-trade episodes across regions and time, and to assess the effects of saving the windfall. The analysis provides some interesting insights:

  • While Latin America’s recent terms-of-trade boom is of similar magnitude to those of the 1970s, the associated income windfall has been much larger. This reflects higher trade openness and longer persistence of the shock.

  • Sizable increases in aggregate saving rates in the recent episode, as opposed to past episodes, suggest a more prudent response this time around. At the same time, estimates of marginal saving rates suggest a weaker effort this time, thus implying that the observed improvement in fundamentals is mostly driven by the sheer size of the income windfall.

  • Notwithstanding the above, Latin America’s governments seem to have been more prudent in saving the windfall than the private sector.

  • Finally, econometric evidence suggests that, while savings pay off by increasing post-boom income, its allocation matters. In previous episodes, foreign savings appear to have delivered higher post-boom income than domestic savings. Hence, the current weakening of external current account balances in Latin America—even if driven by higher domestic investment—warrants a close monitoring.

Annex 5.1

The metric of terms-of-trade windfall presented here focuses on the ‘extra’ income arising from the price effect of the terms-of-trade shock (i.e., changes in output volumes are not attributed to this shock). As in Kohli (2004), real gross domestic income is defined as:7

where GDI (GDP) is gross domestic income (product); RGDP denotes real GDP; PY is the GDP deflator; and Pc is the consumer price index. It follows that RI^ = RGDP^ + PY^PC^, where X^ denotes the annual percentage change for any variable X. From the demand side components of the GDP deflator, the latter equation can be approximated as:

where PX,r = PX/P* and PM,r = PM/P* are country i’s export and import prices (expressed relative to the U.S. CPI); wX and wM denote the ratios of exports and imports (of goods and services) to GDP, and E is the exchange rate vis-à-vis the U.S. dollar. Our interest lays on the second term of the right-hand-side of equation (2), as this is a purely exogenous measure of the income windfall arising from changes in the terms of trade. This measure implicitly assumes that, in absence of the terms-of-trade shock, real income would have grown at the rate of growth of RGDP (i.e., ‘counterfactual’—see figure). It is therefore a conservative estimate of the windfall. Specifically, we compute the annual windfall income as:

where RI is our real income index, excluding real exchange rate effects, and RI* is the benchmark (‘counterfactual’) measure, in this case the RGDP index. Thus, the annual income windfall measures the vertical distance between real income and real income at pre-boom terms of trade; and the cumulative windfall the area between the two. Both are expressed as shares of real income at pre-boom terms of trade.

Once the income windfall has been computed, one can also measure the share of the windfall saved, or marginal rate. For this, we decompose the economy’s average saving rate into the ‘norm’ and the marginal saving rates as:

where S is the aggregate saving rate of the economy, s is the saving rate prevailing in the years preceding the terms-of-trade shock (taken as the norm) and sW is the marginal saving rate. Equation (4) can be re-arranged as:

Subsequently, relying on a similar concept of ‘norm’ for the investment rate and on the current account identity, the marginal saving rate can be decomposed into its domestic and foreign components, as follows:

The latter quantify how much of the extra saving is allocated to domestic investment and foreign assets, respectively (both expressed as shares of the income windfall).

The measure is a conservative estimate as it ignores the effect of these exogenous shocks on output levels. See Adler and Magud (2013) for estimates that include the latter effect as well.

Although countries such as Mexico and Uruguay may have benefited indirectly from the recent terms-of-trade boom of some of their neighbors, they have not faced sufficiently large terms-of-trade shocks to be considered cases of booms.

As the length of the episodes varies, their time windows are normalized for comparability and the series are reported at different fractions of the lifetime of each episode.

These dynamics may reflect changing perceptions of the persistence of this terms-of-trade shock, possibly being increasingly perceived as more persistent, and thus affecting saving and investment decisions over time. We do not assess the optimality of these changes in this chapter.

This result may reflect that the abundance arising from large terms-of-trade booms often lead to misallocation of resources, or that weak underlying current account positions end up being a drag on growth as terms-of-trade booms revert.

Gross domestic income differs from the concept of gross national disposable income (GNDY), as the latter includes the balance of income from abroad (i.e. GNDY = GDP + BI).

    Other Resources Citing This Publication