IV. Sustaining the Expansion and Raising Long-Term Growth

International Monetary Fund. Western Hemisphere Dept.
Published Date:
April 2007
  • ShareShare
Show Summary Details

The ongoing expansion in Latin America represents the most vigorous period of sustained growth since the 1970s. Even so, growth has lagged most other developing country regions. What would it take for investment and productivity growth in Latin America to catch up? And what can be done to make the current expansion sustained—that is, to avoid a relapse to earlier boom-bust cycles? These questions are addressed in turn.

Toward Higher Long-Term Productivity Growth

Over the long run, growth in Latin America has been low compared with other developing countries. This is overwhelmingly attributable to a history of macroeconomic instability and weak growth in labor productivity, which was only partly offset by increases in labor market participation (from below 60 to about 72 percent) and declines in dependency ratios. The low accumulation of physical capital has played a role in subdued labor productivity growth. Even more important, however, has been the low growth in “total factor productivity” (TFP)—that is, the efficiency with which labor and physical capital combine to produce output. TFP growth picked up starting in the early 1990s, but the gap with other regions, if anything, seems to have increased in recent years.12

As a share of GDP, investment in Latin America has remained lower than the developing country average. Investment contracted sharply in the aftermath of the debt crisis in the early 1980s and has since failed to recover significantly, in contrast with the experience of other regions. As a result, for the region as a whole, the contribution of capital accumulation to output per capita growth during 1970–2006 was only 0.3 percentage point per year, less than a fourth that in other developing countries (excluding China).

Output per Capita Growth and Its Sources, 1970–2006

(% per year) 1/

Source: IMF staff calculations.

1/ Unweighted averages. Output per captia growth is decomposed into contributions of labor productivity growth, changes in participation rates, and changes in dependency ratios.

Labor Productivity Growth and Its Sources, 1970–2006

(% per year) 2/

Source: IMF staff calculations.

2/Unweighted averages. Labor productivity growth is decomposed into contributions of physical capital accumulation and total factor productivity.

Investment to GDP Ratio

Source: IMF staff estimates.

Total Factor Productivity Growth

Source: IMF staff calculations.

1/“Large” countries are Argentina, Brazil and Mexico; “medium/small” are the remaining LAC countries in the sample. “Above median income” countries are those with PPP-adjusted per capita income above the sample LAC median income; “below median income” countries are the remaining LAC countries.

Alternative Investment-Productivity Scenarios

Source: IMF staff calculations

1/Investment ratio is the average since 1980.

Thus, achieving significant increases in per capita incomes will require higher investment and, more importantly, substantial improvements in TFP growth. To illustrate the challenges facing LAC countries, different scenarios were constructed in which per capita real incomes double over the next 20 years through different combinations of TFP growth and investment.13 These scenarios require considerably higher investment and TFP growth rates than seen in the region in the past—for comparison, if investment and TFP growth stayed at the 2001–06 regional average during the next 20 years, average per capita income would increase by only around 50 percent. More specifically, two of such scenarios are as follows:

  • In a high-productivity scenario, TFP growth would be 2 percent per year over the projection period. This is high by international standards: slightly above the average rate that has been achieved in developing countries (excluding China) since 1990 and more than twice the average rate for the LAC region during the same period. Even at this high rate of productivity growth, the region’s investment rates would need to rise significantly from current levels to an average of 24½ percent of GDP (from 20 percent in 2006).

  • With lower productivity growth, doubling per capita income over the next 20 years would require investment to rise to well above 30 percent of GDP. For example, if TFP were to grow by 1½ percent a year, still twice the average since 1990, investment would need to rise by an additional 8 percentage points of GDP (to 32 percent).

These scenarios are only indicative of the many possible combinations that could raise per capita income over the long run. However, they underscore the importance of policies that foster both higher investment and a more efficient use of resources. In practice, there is a two-way, positive association between high rates of investment and productivity growth. The modest pickup of productivity during the present recovery (Box 6) further highlights the need to deepen pro-growth reforms.

The question of what policy measures could raise productivity growth in Latin America today has attracted an intense policy debate in recent years. A full discussion is beyond the scope of this report (see the November 2006 Regional Economic Outlook, and Zettelmeyer, 2006, for a survey). Instead, the results of two simple simulations that quantify the effect of two classes of reforms are considered:

  • A reduction of labor market rigidities. In addition to its beneficial effect on productivity growth via more efficient resource allocation, this would raise potential growth during the transition to lower structural unemployment rates.14 The latter can be quantified by allowing factor utilization rates to vary in an aggregate production function.15 For instance, if unemployment rates were to fall permanently from their current levels to the average in advanced economies (6 percent), the boost in employment growth along the transition to lower unemployment rates would raise per capita income in 20 years’ time by 4½ percent, while real growth would rise by some ¼ percentage point per year. Reducing informality and underemployment, and promoting higher female participation in the labor market, would also have a favorable effect on growth through similar channels.

  • An improvement in education levels. A boost to Latin America’s human capital stock16 could significantly increase per capita growth. Simulations in which human capital enters explicitly as an input to the aggregate production function show that, if Latin America were to catch up with the current OECD human capital stock in 2035, rather than in 2050, per capita income levels in 20 years would be 3 percent higher; per capita income would be boosted by an additional 5 percentage points if the catch-up happened 10 years earlier. Faster rates of improvement would also help close an “education gap” that is likely to open up with the most dynamic developing regions, especially East Asian countries, that have recorded impressive advances in secondary completion and tertiary enrollment in recent years. High inequality in the region, reflected in lower enrollments of the poor and rural populations, and high dropout and repetition rates, would also need to be addressed (UNESCO, 2007). This fillip to growth would be even more powerful if accompanied by improvements in the quality of education, an area where Latin America lags (Box 7).

Box 6Productivity, Investment, and Employment in the Current Expansion

In the fourth year of the expansion, the cumulative increase in real GDP in Latin America and the Caribbean reached 20 percent, which is about two percentage points higher than in the previous two expansions.1 The contribution of total factor productivity growth to GDP growth in the last three recoveries has been roughly unchanged at just over 40 percent. More specifically, the main developments are as follows:

  • The ongoing pick-up in investment followed closely the pattern of the previous two expansions during the earlier stages of the recovery. In its fourth year, however, the recovery in investment exceeded that in the 1995–98 expansion. This reflected in part the lower level of investment at the beginning of the recovery—average investment was 17 percent of GDP at the beginning of the current expansion, more than 1 percentage point below investment at the beginning of the 1995–98 expansion.

  • The pace of productivity growth in the fourth year of the current expansion was slightly faster than in the 1990–94 expansion, with a cumulative increase in productivity of around 8½ percent (compared with 8 percent). The rise in productivity appears modest in light of the unusually favorable conditions during this recovery, including abundant liquidity in the global economy and improvements in the terms of trade. The latter improved by a cumulative 20 percentage points in the current recovery, in contrast with the previous recoveries, when it declined.

  • Similar rates of employment growth have been recorded in this recovery as in the two previous recoveries. A key difference is, however, that the current expansion has seen a reduction in unemployment rates—by an average of 2½ percentage points by its fourth year—whereas the previous two expansions resulted in slightly higher unemployment rates. The decline in unemployment in the current expansion has taken place from higher initial levels (11 percent) and has in turn offset lower growth in labor market participation.

Macroeconomic Developments during Recent Expansions

Source: IMF staff calculations.

1/Total factor productivity is corrected for changes in unemployment rates. Capital utilization is assumed to vary proportionally with employment rates.

Note: This box was prepared by Roberto Benelli.1 Expansions are identified with the 1990–2004, 1995–98, and 2002–06 periods. The beginning of the expansion (“year 0”) coincides with the trough of the previous economic cycle.

Box 7Education Quality, Human Capital, and Growth in Latin America

Education quality matters as much as, if not more than, educational attainment levels for improving a country’s stock of human capital and fostering economic growth (Hanushek and Kimko, 2000; Barro, 2001). Moreover, improving the quality of education can help reduce dropout and repetition rates, leading to higher levels of educational attainment.

While access to schooling is adequate in most of Latin America, students often do not perform well on critical skills (De Ferranti and others, 2003; Vegas and Petrow), forthcoming. The poor quality of education cannot be solely attributed to the region’s level of development: even after controlling for GDP per capita, Latin American countries perform worse than East Asian and OECD countries on international assessments. Concerns about low schooling quality are mirrored in survey responses by local business communities in the Global Competitiveness Index on their perceptions of the quality of math and science education in their countries (World Economic Forum, 2006).

Poor quality may also reflect inefficiencies in education spending, since outlays are not low by international standards (Clements, Faircloth, and Verhoeven, 2007). In many countries, efficiency could be improved by reducing the share of spending allocated to teacher salaries and increasing outlays for capital, textbooks, and classroom material (De Ferranti and others, 2003). In addition, incentives for good performance need to be strengthened. The experience of many countries in the region in linking education transfers to local governments and salaries to performance, and increasing the role of parents in the allocation of school budgets, is promising (De Ferranti and others, 2003).

In sum, the evidence implies that Latin America lags further behind East Asia and the OECD countries in human capital than is suggested by educational attainment data. Therefore, greater emphasis on improving quality, including by bolstering the efficiency of public education spending, is needed for Latin America to catch up with more advanced regions.

PISA Mathematics Outcomes and Real GDP per Capita

Sources: OECD (2001); OECD (2004); and World Bank, World Development Indicators.

Educational Attainment and Perceived Education Quality

Sources: World Economic Forum (2006), Global Competitiveness Index; and World Bank, World Development Indicators.

Note: This box was prepared by Priyadarshani Joshi.1 Mathematics scores for Argentina, Chile, and Peru are from the PISA 2000 international assessments. The PISA assessment tests assess abilities in reading, mathematics, and science. See also OECD (2001, 2004).

Sustaining the Expansion

In addition to low average productivity growth, a striking fact of Latin America’s economic history is the frequency and regularity with which growth, once underway, has suffered setbacks. Rather than taking the form of a smooth, sustained “hill” (Pritchett, 2000), expansions have often been short-lived, ending with crises or prolonged periods of stagnation. Latin American business cycles have tended to be volatile compared to advanced countries and developing countries outside the region (Aiolfi, Catão, and Timmerman, 2006). Large output drops have happened more frequently than in any other region except for Africa (Becker and Mauro, 2006). Long-run growth spells—that is, sustained periods of growth, with only transitory interruptions—have tended to be rarer than in other regions.

Frequency and Duration of Growth Spells 1/
RegionNo. of

No. of


% spells

lasting at least
10 years16 years
Industrial Countries372021.68055
Emerging Asia222020.77555
Latin America181412.64329
Sub-Saharan Africa433111.24513
Other developing201813.43928

In light of this discouraging record, what are the prospects for sustaining Latin America’s current expansion? At one level, an answer to this question can be inferred from Sections II and III, as well as previous analyses (see, particularly, Box 1 in the November 2006 Regional Economic Outlook). Macroeconomic volatility in Latin America has typically been a result of fiscal and external imbalances, which in turn reflected public and private overspending (Singh, 2006; Singh and Cerisola, 2006). Stabilization typically gave way to excessive capital inflows, credit booms, real exchange rate appreciation, and public and external debt accumulation, resulting in vulnerabilities to adverse shocks and, eventually, new crises. In light of this historical pattern, the current situation in Latin America looks greatly improved. With the region entering its fifth year of expansion, fiscal and external balances remain in surplus, and several traditional sources of vulnerability—particularly, relating to debt structure, exchange rate policies, and the financial sector—have been significantly reduced.

This said, the most recent expansion has occurred under exceptionally favorable external conditions. Would it withstand their deterioration? Section III provided a partial answer: growth in Latin America appears resilient in the face of moderate shocks, including a moderate deterioration in the terms of trade, but not to a significant deterioration in the international financing environment. It is possible that the empirical analysis in Section III exaggerates the current sensitivity of Latin America to such shocks, as it reflects the region’s behavior during the entire 1994–2006 sample period, and as such may not fully capture recent reductions in external vulnerability. However, there are also arguments in the other direction. Improvements in Latin American fiscal positions have mainly been revenue-driven, and little progress has been made in improving the composition of public expenditure—a key determinant of the durability of fiscal adjustments (Gupta and others, 2005). Furthermore, little progress has been made in reducing budget rigidities, which have traditionally compromised the quality of fiscal adjustment in response to revenue downturns (Alier, 2007). Entrenching lower macroeconomic volatility in Latin America is thus likely to require fiscal reforms.

Another approach to answering the question of what makes growth sustained is to examine the economic and political characteristics of economies undergoing long-lived expansions and compare them with economies experiencing shorter booms. These could include macroeconomic fundamentals, but also deeper structural characteristics that could have effects on economic policies, on the business environment, or on the ability of economies to cope with external shocks. Recent empirical work in this area (Berg, Ostry, and Zettelmeyer, 2007), based on the experience of 140 countries over five decades, identifies four main factors that seem to contribute toward sustaining growth over long spells:

Latin America and Other Regions: Changes in Selected Fundamentals

Sources: (Wacziarg and Welch 2003); Polity IV database; WEO; and WIDER Income Inequality database.

1/ Unweighted averages of most countries in the regions. Country coverage varies according to data availability.

  • First, trade and export orientation, in particular, liberal trade regimes and the absence of persistent current account deficits or overvalued exchange rates. The structure of exports also matters. Long-lasting growth spells tend to be associated with increases in the share of manufacturing in total exports over time.

  • Second, improvements in political institutions. Many countries—even autocracies—can get growth going. However, countries that are democratic, or that democratize over time, appear to be more likely to prolong growth spells.

  • Third, a more equal income distribution. Long growth spells are far more likely to occur in countries with relatively low levels of income inequality. A one-percentage-point higher Gini coefficient seems to increase by 5–10 percent the probability that a growth spell will end in any given year.

  • Finally, macroeconomic stability. Countries with high inflation or exchange rate depreciations tend to experience shorter growth spells. Macroeconomic stability, in turn, could be a result of better policy frameworks, a higher degree of social cohesion, and more developed financial systems.

Clearly, these are areas in which Latin America has traditionally fared poorly: during several decades through the early 1980s, the region was predominantly autocratic, inward oriented, and unequal, and suffered from high macroeconomic volatility. Much has changed in the meantime. The greatest progress has been achieved in the areas of democratic change and macroeconomic stability. Democracy is now firmly entrenched, and the region’s victory over hyperinflation is well known. More specifically, the last five years have been a period of much improved macroeconomic management. Progress has also been made in trade integration and export orientation, though the record in this area is mixed. While virtually all Latin American countries have liberalized trade to some degree, lack of competition and limited de facto openness remains a problem. Ratios of trade to GDP are still relatively low, export growth has underperformed other regions, and the share of manufacturing in total exports remains much lower than in Asia. Finally, economic inequality was on an upward trend during the 1980s and most of the 1990s. Very recently, there has been an improvement (Section II), but there is no question that Latin America remains highly unequal compared with other regions of the world (De Ferranti and others, 2004; Zettelmeyer, 2006).

Two main messages emerge from this discussion. First, with democratic societies across the region, macroeconomic stability, and more openness to trade, the prospects for prolonging the ongoing expansion are better than at any time in recent history. Second, the economic improvements underpinning more sustained growth are a recent development and have not yet been fully entrenched. In particular, efforts to create more equitable and less divided societies—which are better positioned to avoid drastic growth reversals in the future—are at an early stage. The desired progress in this area will take time, and require a combination of better education, labor and financial sector reforms, a nondistortionary tax system, and better targeting and quality of government expenditures (see the November 2006 Regional Economic Outlook). Reforms in these areas are also critical because of their direct effects on long-run productivity growth. In the meantime, macroeconomic and financial sector policies will need to remain vigilant, so that the region can grasp the ongoing opportunity for entrenching and raising growth, even in the event of less favorable external conditions than those that prevail today.

Main Economic Indicators

Output Growth (annual rate in percent)Inflation (eop rate in percent)Ext. Current Account (in percent of GDP)
1995–2004 Avg.200520062007 Proj.2008 Proj.1995–2004 Avg.200520062007 Proj.2008 Proj.1995–2004 Avg.200520062007 Proj.2008 Proj.
North America 1/–3.0–5.4–5.4–5.2–5.2
United States3.–3.4–6.4–6.5–6.1–6.0
South America 1/–
Central America 1/–5.0–4.8–4.8–5.0–5.2
Costa Rica4.–4.1–4.8–4.9–4.8–4.7
El Salvador3.–2.4–4.6–4.8–4.7–4.7
The Caribbean 1/––0.3
The Bahamas3.–10.1–8.8–10.9–13.4–11.5
Dominican Republic5.29.310.76.04.513.–0.9–1.5–2.4–2.2–1.6
ECCU economies 1/–15.8–22.1–21.4–18.9–21.1
Trinidad & Tobago7.77.912.
Memorandum item:
Latin America and the Caribbean 1/––0.2
Sources: IMF World Economic Outlook and IMF staff estimates.

See the September 2006 World Economic Outlook, Chapter 3, for more details on the methodology and related references. In this section, productivity estimates are based on a sample of 20 countries from the Caribbean, and South and Central America.

Doubling per capita incomes in 20 years is ambitious: average real per capita income grew only 80 percent in Latin America and the Caribbean between 1970 and 2006. As a point of reference, this would roughly correspond to the experience of non-LAC developing countries (excluding China) in the last 20 years.

The region has some of the most rigid labor market institutions in the world, which contribute to high unemployment rates (9 percent on average in 2006). See the November 2006 Regional Economic Outlook, Box 9, for an overview.

In these simulations it is assumed that capital utilization rates vary proportionally with unemployment rates.

Constructed as an average of a country’s educational attainment rates (defined as the shares of the adult population with primary, secondary, and tertiary education) weighted by average rates of return to the different educational levels.

    Other Resources Citing This Publication