Abstract

The region has weathered the recent market turbulence well so far and the economic outlook for 2007–08 remains generally robust. However, downward risks to the growth outlook have increased, and there are some signs that the improved fundamentals underpinning the region’s resilience may erode if policies do not strengthen. While inflation has remained generally low, it has recently been edging up in many countries. Fiscal and current account surpluses are shrinking as public spending and imports continue to rise strongly.

Overall Developments and Prospects

The region has weathered the recent market turbulence well so far and the economic outlook for 2007–08 remains generally robust. However, downward risks to the growth outlook have increased, and there are some signs that the improved fundamentals underpinning the region’s resilience may erode if policies do not strengthen. While inflation has remained generally low, it has recently been edging up in many countries. Fiscal and current account surpluses are shrinking as public spending and imports continue to rise strongly.

The sustainability of the expansion will depend crucially on external conditions, the strength of underlying fiscal positions, the resilience of domestic financial sectors, and containing inflation. New analytical work on these issues is summarized in this chapter and presented in more detail in the three thematic chapters that follow.

The LAC region continued to expand vigorously in the first half of this year after a strong performance in 2006, the best in many years for a number of countries. The shift in global conditions since midyear will likely slow growth somewhat next year from the 5 percent average projected for 2007. However, under the baseline for world growth in the IMF’s WEO, the expansion is forecast to remain relatively robust, at 4¼ percent on average in 2008—the fifth year of continuous growth in excess of 4 percent. Countries with closer links to the United States (such as Mexico and those in Central America) are expected to be more affected by the projected slowdown in the United States.

ch02fig01

Real GDP Growth

(Annual percent change)

Source: IMF staff estimates.

Economic activity has strengthened in several countries in which growth was relatively subdued last year, particularly in Brazil, where year-on-year growth reached a high of 5½ percent in the second quarter. In most other countries, the expansion remained vigorous and in some cases continued to set historical records. Even after the midyear financial market turbulence, commodity prices have stayed strong, supporting economic activity in commodity-exporting countries such as Peru and Bolivia. In Colombia and Uruguay, growth has been driven mainly by domestic demand but has also benefited from strong exports. A good 2007 harvest helped boost this year’s output in Argentina and Paraguay. Looking ahead, growth is expected to moderate next year relative to 2007 in most countries, reflecting the external environment and in some cases supply constraints. Growth in 2008 is expected to be in the 3–4 percent range in Mexico, Ecuador, and Uruguay; the 4–5 percent range in Brazil, Chile, Colombia, and Paraguay; and in the 5½-6 percent range in Argentina, Bolivia, Peru, and Venezuela.

ch02fig02

Contributions to Growth

(In percent, per year)

Source: IMF staff estimates.

Growth in Central America also remained strong in the first half of 2007. Although the slowdown in the United States, the region’s main trading partner, is expected to have a significant dampening effect, output is still projected at just under 5 percent on average in 2008. Macroeconomic and political stability and the implementation of Central American Free Trade Agreement-Dominican Republic (CAFTA-DR)—which has now been ratified by all Central American countries, following the referendum in Costa Rica in October—are boosting investment and the growth of nontraditional exports in the region. Further rises in remittances (albeit recently at slower rates, see Chapter 3) have bolstered consumption. The economic expansion has been particularly vigorous in Costa Rica and Panama.

In most of the Caribbean, growth has also been robust and is expected to remain strong by historical standards, despite some slowdown since 2006. Trinidad and Tobago is expected to continue to expand at more sustainable rates during 2007–08 (around 6 percent), as capacity constraints are emerging following several years of rapid growth. Growth in the Dominican Republic, which has been boosted by strong consumption and tourism-related investment, is expected to slow to 8 percent in 2007 and 4½–5½ percent in 2008. In the Eastern Caribbean Currency Union (ECCU) area, growth is projected at around 3½ percent during 2007–08, while Guyana’s growth is currently running at 5 percent, compared with negligible growth during 2002–05.

Output Growth

(In percent; annual rate)

article image
Source: WEO.

PPP-weighted average.

Recent natural disasters have extracted a high toll in human suffering, but their economic effects are likely to be limited, except in some Caribbean countries. In Peru, preliminary estimates for the impact of an earthquake in August 2007 suggest damages in the range of ⅓–½ percent of GDP. In Nicaragua, a hurricane caused significant dislocations in the sparsely populated northeast, but the aggregate economic impact appears contained so far. However, Hurricane Dean hit several Eastern Caribbean countries hard, particularly Dominica, where it caused damage of at least 15 percent of GDP. In Jamaica, hurricane effects are slowing production and raising food price inflation.

While recent financial turbulence has only modestly reduced the region’s baseline growth outlook, it has significantly increased downward risks. The main sources of macroeconomic risk for the region—discussed in greater detail in Chapter 3—stem from weaker external growth, the potential for a further tightening in U.S. and global credit markets, and the negative impact that weaker-than-expected global demand could have on commodity prices. In a few countries, including Argentina, Honduras, and Nicaragua, energy shortages could also create risks to the growth outlook. On the upside, domestic demand could expand faster than currently expected. The fan chart for the region’s growth prospects, based on the analysis in Chapter 3, summarizes the risks to the outlook. Significantly lower growth than currently projected remains a definite possibility. But a full-fledged recession in the region (with average annual growth falling to 1 percent or less) continues to be very unlikely.

ch02fig03

Growth Prospects for Latin America and the Caribbean

(In percent)

Sources: WEO; and IMF staff calculations.

Inflation remains contained by historical standards. Nonetheless, following an historical low in 2006 for the region as a whole, inflation rates have recently been edging up in many countries. This reflects both the cyclical expansion in the region and, in several countries, the impact of a global rise in food prices (see the October 2007 World Economic Outlook, Chapter 1). In some countries, little upward flexibility in nominal exchange rates has limited disinflation effects that might have operated through this channel.

ch02fig04

Inflation

(In percent, annual end-of-period)

ch02fig05

Cross-Country Distribution of Headline Inflation

(Based on 12-month percent changes) 1/

Sources: National authorities; and IMF staff estimates.1/ The chart show s the monthly cross-country distribution of twelve-month headline CPI inflation rates based on a sample of 22 LAC countries.

In Colombia and Uruguay, inflation has risen in the context of vigorous growth, and is approaching or exceeding the upper end of central bank targets. In Chile, Mexico, and Peru, headline inflation remains contained but food price inflation has forced it above central banks’ targets. Inflation also has been ticking up in some countries of Central America and the Caribbean, remains high in Argentina, has recently moved into the double digits in Bolivia (on a 12-month basis), and has stayed well in the double digits in Venezuela despite the relaxing of some supply constraints through imports. In Brazil, inflation remains below the midpoint of the target range but has also risen recently.

Inflation

(In percent; end of period rate) 1/

article image
Source: WEO.

End-of-period rates, i.e. December on December. These will generally differ from period average inflation rates quoted in the World Economic Outlook, although both are based on identical underlying projections.

PPP-weighted average.

Rising food prices since mid-2006 have contributed to inflationary pressures in many countries, with the median differential between 12-month food and headline inflation rates reaching 3¾ percent by August 2007. This has affected large countries (especially Brazil, Chile, Colombia, and Mexico) but also the smaller ones in Central America and the Caribbean. In response, some governments have taken steps to curb food-price pressures directly. For example, competition has been encouraged by liberalizing food imports further in Guyana, Mexico, and Uruguay. Other steps have included reducing the VAT rate (Venezuela), introducing temporary VAT exemptions (Uruguay), and prohibiting exports of some food products (Argentina). In Mexico, the government also negotiated with the private sector to limit price increases of several staple items.

Inflation has also been driven by aggregate demand pressures and closing output gaps (Box 2). Rapid import growth also indicates that domestic demand is rising faster than production capacity in many countries. A further indication that short-run supply constraints are beginning to bind comes from unemployment rates. These have declined steadily in the past few years. However, the pace of decline has recently slowed, suggesting that unemployment may now largely reflect structural bottlenecks (see the November 2006 Regional Economic Outlook: Western Hemisphere— henceforth, Regional Economic OutlookBox 9).

Latin America’s Cyclical Position

Since 2004, Latin America has been growing at its fastest rate (over 5 percent a year on average) since the 1970s, helping to reduce unemployment significantly in many countries. Robust economic activity has been widespread across the region and economic fundamentals point to continued vigor in the coming years. While growth has been broad based, inflation performance has varied more markedly within the region. For instance, in Argentina and Venezuela inflation rates are in the double digits, while in Brazil they have been below the central bank targets, albeit rising recently. This suggests the existence of resource constraints in some countries, although not in all.

This box presents estimates of the cyclical position of Latin American countries—the commonly used output gap, which represents deviation of actual output from its cyclically neutral position—and sheds some light on the extent of resource constraints in the region. An estimate of the region’s cyclical position informs the view on inflationary pressures, and also provides an input for fiscal structural balance calculations (see Chapter 4).

Measuring the output gap

The output gap is defined as the difference between actual (Y) and potential output (Y*) as a proportion of potential output, (YY*)/Y*. The concept of potential output should be understood as the production level at which the economy is in a cyclically neutral position. While not directly observable, the potential output of an economy can be estimated by different methodologies.

A commonly used method relies on applying a statistical filter on the output series to separate trend from cyclical variations. Econometricians have provided many different procedures to separate these two components, and here we use several of them. The filters used are the ones proposed by Hodrick and Prescott (1997), Baxter and King (1999), Christiano and Fitzgerald (2003), and the frequency domain approach of Corbae and Ouliaris (2002, 2006).

A well-known alternative method uses a production function framework to estimate potential output. In this framework, first, the contribution of observable inputs—capital and labor—to the production process is subtracted from output (using estimated weights for each factor) to produce a measure of total factor productivity. Then, a statistical filter, like the ones proposed by the authors listed above, is used to separate cyclical from trend variations in capital, labor, and total factor productivity. The potential output series is obtained by using a weighted average of the filtered production input series.1

Latin America’s cyclical position

Using the first method, statistical filters are applied to annual data for all Latin American countries from 1970 to 2010, using IMF staff forecasts as presented in the October 2007 World Economic Outlook, for the period 2007–10. The inclusion of these observations helps to attenuate the end-of-sample problem present in several of the statistical filters used here.

Two measures of the output gap in Latin America are then obtained: one by aggregating country-level information and another one by applying the filters directly to an aggregate output series for Latin America. As both series have similar cyclical properties, the chart shows only the latter measure. The output gap measure shows the expected troughs (during the 1982–84 and 1994 debt crises, for instance) and peaks. Using staff estimates for 2007, the data suggest that Latin America as a whole now faces a slightly positive output gap and a potential growth rate of about 4.7 percent.

This aggregate information masks cross-country heterogeneity. Among the large Latin American economies, Argentina and Venezuela present wider positive output gaps, corroborating the perception from rising inflation pressures in those countries that they have been facing significant resource constraints. Output gaps are also positive in Colombia and, in some estimates, in Mexico and Peru. The Brazilian and Chilean economies are estimated to be in cyclically neutral positions in 2006/07.

The production function approach could only be applied to a selected group of countries with readily available data. Also, the resulting estimates suffer from the end-period problem common in some filtering techniques, as forecasts for production inputs are not available. Nonetheless, output gap estimates using this approach tend to present the same cyclical properties as the first set of estimates discussed above. Important differences between the two sets of results include a more positive output gap for Argentina and negative, but shrinking in absolute size, estimates for Brazil and Chile.2

ch02fig06

Latin America: Cyclical Position

(In percent) 1/

Sources: WEO; and IMF staff calculations.1/ Potential output levels calculated as an average of all filtered series.

LAC Countries: Output Gap Estimates, 2006-07

(In percent of potential output)

article image

Statistical filters applied directly on output figures. Sample: annual data 1970-2010. The table shows a subset of all estimated filters. Abbreviations used: HP = Hodrick-Prescott; BK = Baxter-King. HP sets lambda at 6.25.

Filters applied to input series (capital, labor, and total factor productivity), which are then aggregated according to production function weights for each input.

Note: This box was prepared by Marcello Estevão, Maria Lucia Guerra, and Jeromin Zettelmeyer.1 The calculation assumes constant returns of scale and can be written as ln(TFP) = ln(Y) – α * ln(K)–(1–α) * ln(L), where TFP, Y, K, and L stand for total factor productivity, output, the capital stock, and employment. The parameter α varies somewhat across countries and, unless estimates were provided by a particular paper or by country authorities, was approximated by using the share of labor income in total value added, a method pioneered by Solow (1957).The results from the two approaches will tend to be similar as, ultimately, statistical filters are used to parse out cyclical fluctuations either directly on the output series or on the input series. However, small variations may appear depending on the relative trend estimates for each production input. Also, the production function approach can be extended to include a more detailed breakdown of inputs into variables with strong policy content (e.g., capital utilization, unemployment rate, and labor force participation).2 A third approach is currently being developed by IMF staff. It consists of backing out the size of the output gap using a small open-economy neo-Keynesian model consisting of equations for a Phillips curve, aggregate demand, uncovered interest parity, and an interest rate rule, as discussed in Berg, Karam, and Laxton (2006). The model is estimated using Bayesian techniques and an unobserved-components specification to identify the output gap, under the assumption that potential output is the production level consistent with stable inflation. An application for Mexico shows reduced inflation persistence in recent years, a moderate degree of pass-through from exchange rate depreciation to inflation, and a sluggish interest rate response to the cycle. Overall, the fit of the model reflects fairly well the recent disinflation episode in the country and suggests that the Mexican economy was at its potential in early 2007, a result consistent with most statistical filters used here.
ch02fig07
ch02fig07

Food Prices and Headline Inflation

Sources: IMF, International Financial Statistics ; national authorities; and IMF staff calculations.1/ The chart shows the monthly cross-country distribution of the difference between food and headline CPI 12-month inflation rates based on a sample of 22 LAC countries.
ch02fig08

Unemployment

(In percent) 1/

Source: EMED.1/ 12-month or 4-quarter rolling averages.

Financial Sector Developments

The recent global financial turmoil came at a time of robust growth in the region’s financial sector activities. Private sector credit in Latin America has been growing strongly for several years. Equity and local debt markets have generally seen increases in prices and trading volume, and derivatives markets and securitization have developed significantly in some countries.

During the global turbulence of July/August, market corrections were also observed across Latin America. These tended to be larger in countries with recent sharp market gains:

  • Following a peak around late July, equity prices fell between 10 percent and 28 percent across the region. Declines were particularly marked in Brazil, Argentina, Colombia, Mexico, Chile, and Peru. However, prices have since recovered, surpassing their pre-turbulence levels in several countries.

  • Similarly, while most currencies depreciated initially (between 1 percent and 14 percent in the larger economies), they have since tended to appreciate again.

  • Local bond prices fell and sovereign spreads rose, but in most cases by much less than in the crises of the 1990s and the early years of this decade. In Brazil, Chile, Colombia, Mexico, and Peru, for example, sovereign spreads rose by less than 100 basis points between mid-July and the peak of the turbulence, and had returned to within 30 basis points of their mid-July levels by mid-October.

  • Unlike in industrial economies, where the direct exposure to the U.S. mortgage market was higher and liquidity in money markets dried up, most countries in the region did not face liquidity problems in local markets.

Before the global turbulence, credit to the private sector was growing rapidly. This trend is expected to continue, but with some slowing as monetary policy has tightened in many countries and global credit conditions have become less favorable. In the first half of 2007, credit was up by over 40 percent from a year earlier in the seven largest Latin American countries, and by over 20 percent in the region as a whole. The marked expansion in private sector credit since 2004 has been driven mostly by household credit, which currently accounts for about 40 percent of private sector credit outstanding. More recently, corporate credit growth has accelerated and may have surpassed household credit growth in several countries. In contrast, credit to the public sector has further contracted, reflecting improved fiscal positions. Credit expansion in the region has been funded largely by robust growth in local private deposits.

ch02fig09

Latin America: Financial Market Indicators 1/

ch02fig11

CDS Spreads

Sources: Bloomberg, L.P., and JPMorgan.1/ Reflects latest available data.
ch02fig12

Credit Growth

Sources: National authorities; and IMF staff estimates.1/ Unweighted averages for Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Venezuela. For 2007, rates are based on latest available data.
ch02fig13

Household Credit Share in Private Sector Credit

(In percent; end-year) 1/

Sources: National authorities; and IMF staff estimates.1/ For 2007, rates are based on latest available data.

Cross-Border Claims of BIS Reporting Banks on Domestic Banking Sector

article image
Sources: Bank of International Settlements; and IMF, International Financial Statistics.

June data for 2006 for Venezuela.

ch02fig14

Private Sector Deposit Growth

(In percent; year-on-year) 1/

Sources: National authorities; and IMF staff estimates.1/ Unweighted average of 17 countries. For 2007, rates are based on latest available data.2/ Commercial and development banks only.
ch02fig15

Household Debt Burden

Sources: Argentina, Financial Stability Bulletin, 2007; Chile, Financial Stability Report, 2006.1/ Indebtedness defined as a ratio of loans to wage income in Argentina, and as a ratio of consumer debt to income in Chile.2/ Financial burden defined as interest and amortization, excluding amortization of loans shorter than one year in Chile.

While foreign borrowing by local offices of international banks has also played a role in countries such as Panama, Costa Rica, and Chile, foreign liabilities generally account for a small and declining share of banking systems’ total liabilities.

At this point, there do not seem to be clear signs of increased vulnerabilities as a result of the private sector credit expansion. While the capital adequacy ratio declined slightly in 2006–07, the nonperforming loan ratio has continued to fall, with increased provisioning and improved profitability. In Argentina and Chile, where data for household debt burden are available, indebtedness is within reasonable ranges. In Mexico, households have a net positive financial position of around 23 percent of GDP, contrasting with their near–net debtor position during the 1995 crisis (Guerra de Luna and Serrano Bandala, 2007). Finally, a new empirical analysis for identifying excessive credit booms, based on commonly used criteria supports the view that, so far, recent credit growth seems to be associated mostly with improved fundamentals (Chapter 5). This said, prudential indicators are of course backward looking and household debt data are limited in most countries. Furthermore, the aggregate picture may mask heightened vulnerabilities in financial institutions that have lowered credit standards in their pursuit of rapid expansion. The rapid pace of credit growth in some countries hence warrants enhanced regulatory oversight.

Financial Soundness Indicators for Latin America

(In percent) 1/

article image
Source: IMF (2007a).

Unweighted averages. Fixed sample of countries over time for each indicator. The number of countries varies by indicator.

Latest data available.

Nonperforming loans (NPLs) as a share of total loans.

Regulatory capital/risk weighted assets.

ch02fig16

New Listings in Latin America

Source: World Federation of Exchanges.

In many countries in the region, financial intermediation remains below that in other emerging market countries. However, the main markets have recently experienced growing trading volume, as well as higher valuations, and rises in corporate stock and bond issues. Derivatives activities in foreign exchange and interest rates have continued to expand in the larger local markets, with total outstanding volume reaching US$16 billion in the first half of 2006 for LAC markets, up from US$4 billion in 1999 (JPMorgan, 2007).

External Developments

Despite generally favorable terms of trade and improving export performance, a strong rise in imports is driving down the regional current account surplus. Notwithstanding recent episodes of portfolio inflows to a few countries, foreign direct investment (FDI) generally remains the main source of foreign capital.

The region’s external current account surplus is expected to contract significantly in 2007–08 after reaching a record high last year. High import volume growth, especially in Argentina, Brazil, Colombia, and Venezuela, is now far outpacing growth in export volumes. Furthermore, the gap between import and export volume growth is no longer expected to be offset by terms of trade improvements. Commodity prices for fuels, metals and agricultural products appear to have stabilized at high levels, and are expected to remain virtually flat on average in 2007 before declining modestly in 2008. While there continues to be great variation across countries in the level of the current account—with continuing surpluses in most commodity-exporting countries of South America, and deficits in Central America and, to a lesser extent, Mexico—the trend toward weaker current accounts is shared by most countries in the region.

External Current Account

(In percent of GDP)

article image
Source: WEO.

Dollar-weighted GDP average.

External Balance of Payments Developments

(In percent of GDP, unless otherwise indicated) 1/

article image
Source: IMF staff estimates.

Refers to LAC region as a whole.

Projections.

Capital flows to some countries—notably Brazil, but also Argentina, Colombia, Peru, and Venezuela—increased sharply before the recent outburst of global market volatility. Capital outflows showed a more mixed pattern, abating in Colombia and Peru, but picking up, for example, in Chile, driven by the accumulation of government assets abroad. High interest rate differentials in conjunction with appreciation pressures and limited or declining exchange rate volatility made some Latin American countries—particularly Brazil—attractive destinations for carry trades.

ch02fig17

Contribution to Changes in Trade Balance to GDP Ratios

(Percentage points, average per year)

Sources: WEO; and IMF staff calculations.1/ A positive export effect indicates an improvement in the trade balance due to export volume growth above real GDP growth. A positive import effect indicates an improvement in the trade balance due to import volume growth below real GDP growth. The TOT and RER effects are the contribution from changes in terms of trade and bilateral real exchange rates (based on GDP deflators) with the United States.
ch02fig18

Recent External Developments in the Largest Latin American Countries

(4-quarter rolling sums, billions of US dollars)

Sources: National authorities; and IMF staff calculations.1/ Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Venezuela aggregate. June 2007 aggregate does not include Peru.2/ Financial inflows (outflows) are the sum of portfolio and other investment inflows (outflows). Errors and omissions and the capital account are not shown.
ch02fig19

Latin America: Carry-to-Risk Ratio 1/

Sources: Bloomberg. L.P.; and Datastream.1/ Defined as the interbank interest rate differential relative to the one-week LIBOR rate, normalized by the unconditional standard deviation of nominal exchange rate (w.r.t. U.S. dollar) changes over a 90-day moving average window.
ch02fig20

Exchange Market Pressure Index in Selected Countries 1/

Source: IMF, International Financial Statistics ; and IMF staff calculations.1/ The index is calculated as % change in NEER + % change in reserves * (standard deviation of % change in NEER)/(standard deviation of % change in reserves), where NEER is the nominal effective exchange rate. All percentage changes and standard deviations are computed for the 12 months between June 2006 and May 2007.

Until mid-2007, surges in capital inflows and continued large current account surpluses in some commodity exporters led to significant exchange market pressures in several countries in the region. However, in only a few countries—especially Brazil and Colombia—did these pressures lead to significant bilateral appreciations against the U.S. dollar. Since the dollar has depreciated against the currencies of many other Latin American trading partners (particularly the euro zone), trade-weighted exchange rates were weaker. Indeed, nominal effective exchange rates have actually depreciated in most countries in the region this year (even before the onset of global market turbulence). At the same time, however, inflation differentials with trading partners were positive in most cases (Ecuador and Peru were the main exceptions). The net effect of these changes was a small appreciation of the real effective exchange rate in several countries, with larger appreciations—between 10 and 35 percent in the 12 months through end-June 2007—in Colombia, Brazil, and Venezuela.

ch02fig21

Bilateral vs. Multilateral Exchange Rate Movements

(In percent) 1/

Sources: IMF, Information Notice System ; and IMF staff estimates.1/ 12-month percentage changes through end-June 2007. An increase in exchange rates denotes an appreciation.
ch02fig22

Nominal and Real Effective Exchange Rates, and Inflation Differentials

(In percent) 1/

Sources: IMF, Information Notice System ; and IMF staff estimates.1/ 12-month percentage changes through end-June 2007. An increase in exchange rates denotes an appreciation.

The global financial shocks in July and August 2007 initially slowed capital inflows. Central banks in several countries, including Argentina, Brazil, and Peru, reduced reserve accumulation significantly during this period compared with preceding months (in Colombia, the central bank had already stopped intervention in May). Argentina intervened in the foreign exchange market to lean against depreciation pressures on the currency. With the region’s central banks generally much more comfortable with exchange rate fluctuation than in the past, currencies became a first shock absorber, depreciating significantly initially, although they soon began trending back to pre-turbulence levels.

ch02fig23

Exchange Rate Developments During the Recent Market Turbulence Episode

(July 20, 2007 = 100) 1/

Sources: Bloomberg, L.P.; and IMF staff calculations.1/ Bilateral exchange rates against the U.S. dollar. An increase denotes an appreciation. Reflects latest available data.
ch02fig24

Recent Reserves Accumulation

(Billions of U.S. dollars; change on a month earlier) 1/

Source: National authorities.1/ Net reserves for Peru.

FDI remains the main source of foreign capital for the region. However, it is still below the levels as a ratio to GDP reached in the late 1990s, when privatizations boosted FDI receipts for the region (Box 3). Inward FDI is now being partially offset by higher levels of outward FDI, as Latin American companies have increasingly begun to invest abroad. Following unusually low net FDI in 2006 as a result of such investments—culminating in the US$17 billion purchase of mining assets in Canada by a Brazilian company in October 2006—net FDI inflows are projected to rebound to 1½–2 percent of GDP in 2007 and 2008.

Trends in Latin America’s Balance of Payments

Like other developing country regions, Latin America experienced large current account deficits during most of the 1990s, financed by net capital inflows that helped governments fund fiscal deficits. This changed in 1998, when a “sudden stop” in private capital flows set in motion a process of rising regional current account balances and falling net capital inflows, which lasted until 2006. The Latin American experience of the past decade contrasts with that of developing Asia and—especially—Central and Eastern Europe. In Asia, current accounts also swung into surplus during the Asian crisis, but net capital flows recovered quickly, leading to large capital and current account inflows, and sizable reserve accumulation. In Central and Eastern Europe, current account deficits did not reverse and, in fact, continued to widen, financed by increasing net capital inflows.

Latin America’s current account reversal occurred in two phases. First, the emerging market crises of the late 1990s and the early years of this decade interrupted access of several of the largest countries to international capital markets and raised external financing costs throughout the region. Economic activity declined, fiscal adjustment took place, imports contracted, exchange rates depreciated in nominal and real terms, and exports picked up. Second, starting around 2003, import demand rose as Latin American economies began to recover, but export receipts grew more rapidly, bolstered in many countries by high commodity prices. The terms of trade improved on average by more than 20 percent between 2002 and 2006. In addition, private transfers—reflecting mainly workers’ remittances—contributed to stronger current accounts in several countries, particularly in Mexico and Central America.

Current account developments were not uniform across the region over the decade. Across most of South America and Mexico, external current accounts strengthened—in some cases, gradually, in others, more abruptly in response to the initial sudden stop of capital. By contrast, commodity-importing countries in Central America, which suffered terms of trade declines, continued to record large current account deficits, averaging around 5 percent of GDP, and to rely on net capital inflows (including aid) for their financing.

ch02fig25

Current Account Balances and Net Capital Flows in Latin America and Other Regions, 1990-06

(In percent of GDP)

Source: WEO.
ch02fig26

Current Account Balances, 1998 - 2006

(In percent of GDP)

Source: WEO.1/ Unweighted averages. Commodity exporters include Bolivia, Chile, Ecuador, Trinidad and Tobago, and Venezuela. Central America includes Costa Rica, Dominican Republic, El Salvador, Guatemala, Honduras, Nicaragua, and Panama.

A decline in net capital inflows between 1998 and 2006 was accompanied by important shifts in the composition of capital flows to Latin America.1

  • The initial decline in net flows was driven by a drop in capital inflows—particular debt and other portfolio inflows—as crises around the region made access to international capital more difficult or expensive. Inward FDI remained relatively strong, peaking in 2000 on the back of privatizations, before stabilizing at about 2½–3 percent of GDP.

  • In contrast, the continuing decline in net flows after 2002 is explained by a pickup in capital outflows, reflecting both higher outward FDI and higher portfolio investment abroad, in line with a worldwide trend toward greater international diversification. While gross inflows remained roughly unchanged as a share of GDP after 2002, their composition shifted, with increased lending and investment to the private sector, offset by repayments of external public debt.

As a result of the changes in both the quantity and composition of capital flows, Latin America’s net foreign asset position has strengthened significantly. With external debt declining, equity liabilities growing from a low base, and foreign direct investment liabilities broadly constant as a percentage of GDP, the region’s external liabilities have fallen from over 80 percent of GDP in 2003 to around 60 percent of GDP in 2006 (on a net basis, from 40 to around 25 percent).2 Furthermore, there has been a shift toward nondebt liabilities, and from public to private sector liabilities.

ch02fig27

Gross Capital Inflows and Outflows, 1998 - 2006

(In percent of GDP)

Sources: WEO, and IMF staff calculations.
ch02fig28
ch02fig28

External Assets and Liabilities, 1998 - 2006

(In percent of GDP)

Sources: Lane and Milesi-Ferretti (2006); and IMF staff calculations.
Note: This box was prepared by Roberto Benelli.1 Capital “inflows” are defined as the net purchases of domestic assets by nonresidents, and capital “outflows” as the net purchase of foreign assets by residents (with a positive sign indicating an increase in foreign assets held by domestic residents). As such, both inflows and outflows can take negative values.2 In spite of growing outflows (asset accumulation abroad), the value of the stock of external assets in domestic currency has declined slightly since 2003 as a result of domestic currency appreciation.

Monetary and Exchange Rate Policies

The environment facing monetary policymakers has shifted in recent months. In the first half of 2007, monetary policy was complicated by foreign currency inflows and, in some cases, by conflicting inflation and exchange rate objectives. The recent financial turbulence and slowing external growth require a careful balance between continuing concerns about inflation and the prospect that declining external demand may slow growth and perhaps weaken price pressures in Latin America.

In the first half of 2007, buoyant cyclical conditions and incipient inflationary pressures led a number of central banks to tighten monetary policy. Interest rates were raised in four of the five countries with formal inflation targeting (Chile, Colombia, Peru, and Mexico). Several other countries—including Argentina, Bolivia, the Dominican Republic, Paraguay, and Uruguay—expanded sterilization operations to rein in rapid growth in monetary aggregates.

However, sizable foreign exchange inflows ahead of the global financial turbulence tended to overwhelm sterilization efforts, as many countries (excluding Chile and Mexico) responded to the inflows by increasing discretionary intervention in foreign exchange markets. Intervention was motivated in part by concerns about real exchange rate appreciation and eroding competitiveness (Box 4), and in part by the desire to build reserves for precautionary reasons. As a consequence, foreign exchange reserves for the region as a whole increased by almost 40 percent in the 12 months to end-June 2007, to levels that now appear comfortable from a prudential perspective (Box 5). In spite of its wide use, foreign exchange intervention has generally not been very effective in containing real appreciation in Latin America (Box 6).

Real Exchange Rates and Competitiveness

Real exchange rates in Latin America have generally appreciated at somewhat faster rates than in other developing countries in recent years, particularly in relation to productivity growth. Combined with high import growth and comparatively subdued export growth, this has fueled concerns among some Latin American policymakers about eroding competitiveness relative to other developing country exporters (particularly in Asia). These concerns became particularly acute during the surge in capital inflows to the region in the first half of 2007, which put significant upward pressure on some currencies.

At the same time, there is ample evidence that exchange rates in Latin America continue to be broadly in line with fundamentals. First, real exchange rates have not yet returned to their mid-1990s levels, and in most countries are hovering around their average values since 1980. Furthermore, the recent appreciation is less marked if nominal effective rates are adjusted by changes in unit labor costs (ULCs). Among the large countries, it is primarily Brazil and Colombia that have experienced effective appreciations in ULC terms, while Argentina, Chile, Peru, and Mexico have experienced broadly stable—or even declining—ULCs relative to their main trading partners. Most important, fundamentals that typically determine equilibrium exchange rates—such as net external asset positions—have significantly improved since the beginning of this decade. As a result, recent country-level studies conducted at the IMF tend to conclude that real exchange rates in the region are generally appropriately valued.

Even if Latin American exchange rates are—for the most part—well aligned with fundamentals, specific countries may be facing pressures associated with competition from abroad (e.g., China and India), particularly in third markets such as the United States (see Freund and Özden, forthcoming). However, competition in third markets does not capture the overall impact of the rapidly growing Asian economies on Latin American exports, which is positive (Lederman, Olarreaga, and Soloaga, 2007).

ch02fig29

Real Effective Exchange Rates and Productivity

(1990 = 100)

Sources: World Bank, World Economic Indicators ; IMF, Information Notice System ; WEO; and IMF staff calculations.1/ CPI-based REER (weighted by PPP-adjusted GDP for the countries in the grouping).2/ The ratios of REER to country-specific indices of output per worker relative to an index of average output per worker in advanced economies.
ch02fig30
ch02fig30
ch02fig30

Real Effective Exchange Rates

(2003Q1 = 100)

Sources: IMF staff calculations based on data from national authorities; and IMF, International Financial Statistics.1/ ULC REER based on aggregate economy, or manufacturing when aggregate economy not available.

Note: This box was prepared by Roberto Benelli and Jeromin Zettelmeyer.

Optimal Prudential Reserve Levels

Recent reserves accumulation in Latin America has often been motivated by the desire to acquire a “cushion” that would protect the economy and smooth consumption in the event of a sudden stop in capital flows. However, maintaining liquid reserves also imposes quasi-fiscal costs. What level of reserves is optimal in light of these benefits and costs? In a recent IMF working paper, Olivier Jeanne and Romain Rancière (2006) answer this question in a simple utilitymaximizing framework.1 The result is a formula for the optimal level of reserves from a prudential perspective that states that optimal reserves should be larger: (1) the larger the size and output cost of a crisis (the “sudden stop” of capital inflows that the country wishes to protect itself from); (2) the higher the probability of the sudden stop and (3) the lower the cost of holding reserves. The latter is typically calibrated as the average spread between 10-year U.S. treasury bonds and the federal funds rate, while the output cost is based on past crises experiences. The need for foreign currency in a sudden stop depends on short-term debt, and foreign currency deposits net of banks’ liquid foreign assets (Gonçalves, forthcoming).

ch02fig31
ch02fig31

Estimated Optimal and Actual Reserve Levels

(In percent of GDP)

Source: Author's calculations.Note: The baseline optimal level of reserves is based on illustrative parameter values, including a probability of a sudden stop of capital inflows of 10 percent, a term premium of 1.5 percent, and a risk-aversion parameter of 2. The need for reserves and output loss are calibrated as in the original studies. The dashed lines indicate the range for the optimal level of reserves when the probability of a sudden stop varies between 5 and 20 percent.

IMF staff have separately applied Jeanne and Rancière’s formula to Mexico, Brazil, Peru, and Uruguay, using country-specific information to calibrate foreign exchange needs in a crisis and the likely output cost if a crisis occurs. Despite these differences, all four studies reached a similar conclusion: the gap between the optimal and actual reserve levels appears to have closed in recent years. The figure illustrates this general finding by making common assumptions (for country comparison purposes only) about the likelihood of a sudden stop, the cost of holding reserves, and the risk-aversion parameter.

The figure shows that the closing gap results from both reserve accumulation and falling optimal reserve levels as measured by Jeanne and Rancière’s formula. The fall in optimal reserve levels reflects declining short-term debt and net foreign currency deposits in recent years. As vulnerabilities have decreased, so has the optimal level of reserves calculated to be needed to self-insure against a “sudden stop.”

While optimal reserves estimates are notably sensitive to the crisis likelihood parameter (as shown by the dashed lines in the figure), the studies suggest that reserves have now approached comfortable levels. Nonetheless, a key message of the model is that the optimal level of reserves can rapidly increase if vulnerabilities return to past levels, warranting caution about future short-term foreign currency debt issuance and/or future increases in net foreign currency deposits.

Note: This box was prepared by Fernando M. Gonçalves.1 See also related work by Mulder and Bussière (1999); IMF (2000, 2001); Garcia and Soto (2006); and Jeanne (2007). The latter two studies go beyond the mitigating role of reserves and consider that reserves may also help prevent crises.

Effectiveness of Central Bank Intervention—Evidence from Colombia

How effective is central bank intervention in influencing the nominal exchange rate? While there is an extensive literature on this subject for advanced economies, less is known about the effectiveness of intervention as an independent policy tool in emerging markets.1 A major hurdle has been the lack of official, high-frequency data on central bank intervention (because of valuation changes, this cannot be inferred simply from changes in reserves). Moreover, it is often not possible to know, for example, whether authorities accumulate reserves with the intent of affecting the exchange rate or for other reasons, such as, self-insurance against external financial shocks.

Drawing on a new dataset on daily official statistics on foreign exchange intervention by the central bank of Colombia (Banco de la República), a recent study by Kamil (2007) examines the impact of discretionary intervention on the level and volatility of the nominal spot exchange rate in Colombia over the period September 2004 to April 2007. The case of Colombia is of particular interest for three reasons. First, Colombia has faced strong exchange rate appreciation pressures. Second, the period under study was punctuated by frequent, and at times large, discretionary purchases of foreign exchange to resist domestic currency appreciation. During these periods, intervention took place on almost 80 percent of business days. As a result, between January 2004 and May 2007, reserves almost doubled (from US$10.5 billion to approximately US$20 billion). Third, as shown in the first panel of the figure, the sample period considered in the study spans two distinctly different regimes with respect to the stance of monetary policy: one characterized by constant or falling interest rates (September 2004 to March 2006), and a second one by a tightening of monetary policy and an increase in nominal interest rates to reduce inflationary pressures (January to April 2007). This provides an opportunity to test the hypothesis that discretionary intervention to stem domestic currency appreciation is more effective when there is consistency between monetary and exchange rate policy goals.

Estimation results indicate that the effects of Banco de la República intervention varied sharply across the two periods. During the first period of unannounced, discretionary intervention (December 2004–March 2006), Banco de la República foreign currency purchases had a statistically significant, positive impact on the exchange rate level. However, while discretionary intervention contributed toward moderating the appreciation trend, its effect was economically small and short lived. For example, a US$30 million sterilized purchase of foreign exchange (the average daily amount of intervention within this period) would depreciate the value of the domestic currency by 0.32 percent. Further, intervention operations did not have a lasting impact on exchange rate dynamics: almost 70 percent of the contemporaneous effect was reversed in two days. Controlling for other factors affecting short-term exchange rate volatility, results also indicate that central bank intervention dampened the volatility of exchange rate returns during this period.

During the second period (January–April 2007), Banco de la República intervention did not influence the level of the exchange rate, even in the short term. Banco de la República’s intervention operations aimed at depreciating the currency were drowned out by offsetting movements in the EMBI spread and market reactions to higher-than-expected GDP and inflation announcements. During this period, there was a tension between monetary and exchange rate policy goals, as markets perceived that the policy of largescale foreign currency purchases was not consistent with meeting the Banco de la República’s inflation target (see second panel of figure). To stem the appreciation of the peso, the Banco de la República intervened aggressively, accumulating US$4.5 billion (38 percent of monetary base) in the first four months of 2007. At the same time, to cope with inflation pressures, it steadily increased its policy interest rate. But this had the consequence of attracting more capital inflows, thereby exacerbating appreciation pressures.

ch02fig32

Central Bank Intervention and Movements in the Reference and Interbank Interest Rates

ch02fig33

Evolution of Inflation and Inflation Expectations vis-à-vis the Inflation Target

Sources: Banco de la República; and author's calculations.

This tension between foreign exchange purchases and interest rate hikes opened the door to adverse market dynamics and one-way bets against the Banco de la República, which manifested itself in two ways. First, offshore entities built large long positions in pesos through the onshore forward market, as Colombia—with nominal interest rates 800 basis points above Japan’s—became one focal point for carry trade in Latin America. Second, consistent with Toro and Julio (2006), results indicate that trading volumes on the spot FX market increased abnormally on days when the Banco de la República intervened, even after correcting for changes in volatility and other macroeconomic factors. The Banco de la República decided to stop intervention in May, noting that continued intervention would have compromised achieving its inflation target (Banco de la República, 2007).

In sum, the results suggest that coherence between the intervention policy and inflation objectives is a critical factor in determining the success of discretionary intervention. While a government committed to resisting appreciation in principle has an unlimited supply of “ammunition,” inflation objectives can become a binding constraint.

Note: This box was prepared by Herman Kamil.1 Recent contributions using daily data for emerging markets include Guimarães and Karacadag (2004) for Turkey and Mexico and Disyatat and Galati (2007) for the Czech Republic. Canales-Kriljenko (2003) provides a summary of central bank intervention practices in developing countries.
ch02fig34

Reserve Accumulation from a Global Perspective

(12-month percentage change)

Source: IMF, International Financial Statistics.

Some countries went beyond intervention in attempting to reduce real appreciation pressures or offset their effect on competitiveness. Colombia established an unremunerated reserve requirement (URR) of 40 percent on external borrowing and gross portfolio inflows for six months (a similar URR has existed in Argentina since 2005). At the same time, it set tighter prudential limits on forward currency operations. Other actions comprised extensive administrative measures to contain inflation in Argentina, including limiting price adjustments in regulated industries, selective price agreements, and export restraints; and import tariffs and/or subsidized credit lines to help domestic producers deal with low-cost competition from abroad in Brazil and Colombia (the latter measures have so far remained small in scale).

Following the financial turbulence in industrialized markets, the main challenge facing monetary authorities in the region has shifted to assessing the impact of the deteriorating external environment against the backdrop of continued inflationary pressures in a number of countries. Although most economies in the region are operating at or above capacity—in itself a source of inflation pressure—a significant drop in U.S. economic growth could lead to slower growth in the region. In addition, central banks need to decide whether and how to react to rising food prices, which are likely to reflect international conditions in food markets—and thus a relative price adjustment—but could also be a sign of rapid domestic demand growth and cyclical conditions.

As central banks have weighed these trade-offs and individual country circumstances, monetary policy in the region has been cautiously tightened in some countries and put on hold in others. Notably, Chile and Peru raised rates in September and Mexico did so in late October, but Chile left rates on hold during its October policy meeting. Brazil continued its stepwise reduction of interest rates through early September but has since left rates unchanged.

Fiscal Policy

Primary fiscal surpluses—a key source of recent strength in the region—peaked in 2006 and are projected to decline to about 2 percent in 2007 and 1¾ percent in 2008, as revenue ratios stabilize or fall at a time when expenditure is still growing rapidly. Structural (cyclically adjusted) primary balances generally appear to remain in surplus. However, they are lower than current reported surpluses, and there is significant uncertainty about their size, particularly in the case of commodity exporters.

Primary fiscal surpluses in Latin America appear to have peaked in 2006 after strengthening for several years during which revenue growth outpaced rising expenditures. In 2007, expenditure has continued to grow rapidly—at a projected 9 percent in real terms on average—or faster than real GDP. But revenue-to-GDP ratios are projected to decline slightly, mainly as a result of flat or slightly lower commodity prices and/or production. As a result, primary surpluses will shrink this year to an expected regional average of about 2 percent of GDP, although with significant outliers. These include Chile, where the primary balance is projected to remain over 7 percent of GDP in line with the structural surplus rule, and Venezuela, where high real spending growth in 2006 and a decline in oil revenue as share of GDP could push the 2007 primary balance into negative territory for the first time in six years.

ch02fig35

Total Revenues and Primary Spending

(In percent of GDP) 1/

Source: IMF staff estimates.1/ Unweighted averages for 17 countries.

In light of this, it is important to assess how strong underlying fiscal positions are in Latin America. Are “structural” primary balances still in surplus? And how rapidly will they turn to deficits if spending growth continues unabated? These questions are addressed in Chapter 4, which analyzes the sources of revenue growth since 2002 and estimates how far primary balances would shrink simply as a result of a return to a neutral position in the economic cycle, and of the likely level of commodity revenues over the medium term. The main results are as follows:

  • Increases in commodity revenues have played an important role in the recent rise of revenues in the main commodity-exporting countries. These have been driven mostly by higher commodity prices, but also by increases in royalties and commodity-related taxes and, in some cases, by increases in production volumes.

  • In contrast, country business cycles have had only small identifiable effects on revenue ratios. In most countries, noncommodity revenues tend to rise roughly in proportion to economic activity, leaving the ratio to GDP approximately unchanged.

  • Tax policy has played a significant role in raising revenue in some countries—notably Brazil and El Salvador—while tax administration changes helped raise the revenue ratio in Costa Rica.

  • Finally, in some countries, including Argentina, Colombia, and Panama, revenues as a share of GDP have increased by more than can be clearly explained by these factors. A few other countries have had analogous unexplained declines in noncommodity revenue rates. In some, but not all, of these cases, the causes of these “residual” changes can be identified upon closer inspection of each country case.

Hence, the strength of Latin America’s underlying revenue positions today depends importantly on whether increases in commodity prices, as well as recent “residual” surges in noncommodity revenues are temporary or permanent. Although there is significant uncertainty in both dimensions, the analysis supports a general conclusion: structural primary surpluses are more modest than the unadjusted data indicate, particularly in commodity-exporting countries facing possible declines in prices over the medium term (see Chapter 4 for details).

At the same time, fiscal policy remains expansionary. As expenditure growth continues to outpace both revenue and GDP growth in many countries, fiscal surpluses—both actual and structural—are shrinking fast. In oil-exporting countries, in particular, noncommodity primary balances are worsening.1 Unless spending growth is curbed, many countries in the region are likely to return to structural deficits in 2008.

ch02fig36

Growth in Real Primary Expenditures

(In percent) 1/

Source: IMF staff estimates.1/ Unweighted averages.

In the last 12 months, a number of countries have initiated fiscal reforms, focused mostly on the tax system.2 Peru undertook several reforms aimed at broadening the tax base—particularly by rationalizing tax incentives—and lowering distortionary taxes, including through the stepwise reduction of the financial transactions tax. Colombia and Uruguay adopted broadly revenue-neutral tax reforms aimed at enhancing efficiency and also—in the case of Uruguay—at making the tax system more equitable, including by introducing a personal income tax. Mexico passed a fiscal reform package that includes a new minimum tax, improvements in expenditure management and revenue administration, and changes in intergovernmental fiscal relations. This is expected to raise about 2 percent of GDP over the next few years, to help cover declining oil revenues and new infrastructure spending. In Brazil, the authorities have announced plans for a comprehensive reform of the indirect tax system, merging all consumption taxes into two value-added taxes (a federal and a state-level VAT), and aiming for a significant simplification in tax administration. Several Caribbean countries—Antigua and Barbuda, Dominica, Guyana, and St. Vincent and the Grenadines—have also embarked on tax modernization initiatives, including the implementation of VAT systems.

ch02fig37

Changes in Fiscal Policy Stance, 2006-07

(Change in noncommodity primary balance, in percent of GDP relative to previous year) 1/

Source: IMF staff estimates.1/ More negative numbers indicate a more expansionary fiscal impulse. Oil exporters include Ecuador, Mexico, Trinidad & Tobago, and Venezuela. Other commodity exporters include Bolivia, Chile, Colombia, and Peru. Central America includes Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, and Panama. Other South America includes Argentina, Brazil, Paraguay, and Uruguay.

As a result of strong growth, some real appreciation, and continuing (albeit smaller) primary surpluses, public debt levels have continued their decline in 2006 and 2007, to about 50 percent of GDP in Latin America on a GDP-weighted basis and just under 40 percent on an unweighted basis (projections for end-2007). However, debt ratios remain above mid-1990s lows in several Latin American countries including Argentina, Brazil, Colombia, and Uruguay, and in the region as a whole on a weighted average basis. As discussed previously (see in particular, the November 2006 Regional Economic Outlook: Western Hemisphere, Box 1), debt structures have also improved across the board, with much less reliance on foreign currency debt and lengthening of local currency maturities. This trend has continued in 2007, with several countries—including Brazil, Chile, Mexico, and Peru—issuing long-term bonds in local currency, including in international markets.

ch02fig38

Public Debt: Domestic- and Foreign-Currency-Denominated

(In percent of GDP)

Source: IMF staff estimates.1/ Total public debt used when breakdown by type of currency not available.2/ Weighted averages for 17 countries, where available.

Social Policies

Poverty continues to decline but inequality remains high in the region. To tackle persistent inequality and high poverty, better targeting of social spending is essential.

There has been significant progress in reducing both extreme and total poverty rates in Latin America in recent years—the latter to around 38 percent of the population in 2006.3 Recent national data show a particularly impressive decline in Argentina, where the poverty rate has fallen by over 30 percentage points since 2002, reversing a large rise during the crisis period between 1999 and 2002. Brazil and Colombia have also achieved significant reductions—7½ and 10½ percentage points, respectively—since 2002. However, large differences in poverty levels remain across countries. In Chile, the percentage of the population below the national poverty line is below 14 percent, while it is in the 20–30 percent range in Argentina, Brazil, Costa Rica, Mexico, and Uruguay; around 40 percent in Peru; and over 50 percent in Bolivia.4 Income inequality, while persistent, has also decreased in most countries in the region since the late 1990s. Latin American countries still tend to have much higher inequality than countries of similar levels of development in Asia and Europe, possibly dampening the potential effect of economic growth on reducing poverty (see Box 7, and the October 2007 World Economic Outlook, Chapter 4). Furthermore, trend growth continues to be weaker in Latin America than in other developing country regions, particularly when the favorable external environment of recent years is taken into account.

Escaping Poverty in Latin America

Robust economic growth has helped to reduce poverty significantly in Latin America in recent years, reflecting both the strength of the expansion, by historical standards, and a greater impact of growth on poverty. Argentina has experienced the largest poverty decline, followed by Bolivia, Brazil, Chile, Colombia, Mexico, Peru, and Uruguay. Countries with low growth have experienced less impressive declines in poverty. Based on these recent national estimates, on average, a 1 percentage point increase in growth has corresponded to a 1.7 percent reduction in poverty in Latin America between 2002 and 2006 (population weighted). Similarly, studies using household surveys find that the poverty response to growth was about 1.2 percent between 2002 and 2005 (ECLAC, 2006). This is more than twice as high as during the 1990s, when a 1 percentage point increase in growth reduced poverty by only 0.6 percent on average. And, unlike in the 1990s, economic growth has tended to go along with lower income inequality. Hence, poverty reductions seem to have resulted from a combination of both higher growth and distribution improvements.

Why has recent growth been comparatively pro-poor and distribution-friendly? While it is hard to answer this question comprehensively, recent experience provides some clues. In contrast with earlier episodes, the current expansion has seen a significant reduction in unemployment rates across the region (see the April 2007 Regional Economic Outlook: Western Hemisphere, Box 6; and Loayza and Raddatz, 2006). There have also been improvements in the targeting of social spending, including through the creation of conditional cash transfer programs in countries such as Argentina, Brazil, Mexico, Chile, Colombia, and Uruguay. Chile’s success in halving poverty between 1990 and 1998, for example, has been attributed to a combination of growth and well-targeted social programs (World Bank, 2001); and recent studies of the impact of conditional cash transfer programs have shown that these transfers are better targeted than most social spending and contribute significantly to poverty reduction (Perry and others, 2006). Finally, workers’ remittances to Latin American countries have risen significantly in recent years. By effectively providing privately funded social safety nets and helping finance household investment and education, remittances could also have played a role in reducing poverty (Fajnzylber and López, 2007).

In spite of the recent successes, poverty in the region remains high, at about 38 percent in 2006. Further reduction will depend first and foremost on sustaining the current economic expansion while containing inflation, and thus breaking decisively with Latin America’s tradition of macroeconomic volatility. Moreover, poverty remains stubbornly high within particular socioeconomic groups. Indigenous peoples and female-headed households continue to be at the bottom of the income distribution. Countries with high rural populations face additional challenges as poverty is both much higher and less responsive to growth in rural than in urban regions. Finally, urban poverty, while comparatively low, has been rising in some countries. Measures that could help address these challenges include:

ch02fig39

GDP per Capita and Poverty, 2002-06

(Percent changes) 1/

Sources: National authorities; ECLAC (2006); and World Bank, World Development Indicators.1/ Poverty rate defined as percentage of population below national poverty line. Latest period corresponding to poverty data availability. For Chile and Uruguay refers to 2003-06; for Mexico and Bolivia refers to 2002-05; for Peru refers to 2004-06.
Note: This box was prepared by Priya Joshi.

As the expansion has consolidated, social spending has continued to rise. With greater macroeconomic stability, these outlays have also become less volatile. However, the region still has considerable scope to improve the efficiency of social spending; the quality of primary education, for example, continues to lag behind other regions of the world (Vegas and Petrow, 2007). And social spending continues to benefit primarily middle-and upper-income groups rather than the poor, owing to the high share of these outlays absorbed by higher education and social insurance. Indeed, social spending in many countries in the region has been regressive, benefiting the richer segments of society more than the poor in absolute terms.

ch02fig40

Poverty Rates and Reduction in Poverty 1/

Source: National authorities; and ECLAC (2006).1/ ECLAC data for Panama. 2006 refers to latest available data. Argentina refers to March 2007 data. 2004 may refer to 2001-04 data.
ch02fig41

Changes in Inequality, 1998/1999 to 2004/2005

(Gini coefficient; larger figures indicate greater income inequality)1/

Source: ECLAC (2006).1/ Closest year available to 1998/1999 or 2004/2005.2/ LAC (UW) refers to unweighted average, LAC (W) to weighted average.

Social spending may nonetheless reduce measured inequality in Latin America, because the poor’s share of the benefits from social spending tends to be larger than their share of pre-transfer income. As shown in a new study by IMF staff (Cubero and Vladkova-Hollar, forthcoming), this effect is generally positive although small (particularly in light of the high degree of inequality of the prefiscal policy income distribution). In Central America, for example, the combined redistributive effect of taxation and social spending has been to reduce the Gini coefficient by around 4.5 percentage points on average, with large variations—from 1.6 percentage points in El Salvador to 8 percentage points in Panama.

Targeted social assistance programs, which combine transfers to the poor with health or education-related goals, hold some promise for reducing poverty. Recent evaluations of the conditional cash transfer program Oportunidades in Mexico have demonstrated that the poverty alleviation potential of such programs can lead to improved health and education, as well as more capacity for investments, as credit constraints of poor beneficiaries are eased (Gertler, Martinez, and Rubio-Codina, 2007; Freije, Bando, and Arce, 2007). The Chile Solidario program, a social protection system that targets the poorest families, has been important in helping lower Chilean poverty rates from 19 percent in 2002 to below 14 percent in 2006. Programs of this type have also been implemented in other countries, including Argentina, Panama, Peru, and Brazil—where the Bolsa Família program has been expanded rapidly in the last four years to reach almost 12 million households. Other related social policy initiatives have also recently taken place in Peru—with a new national strategy to combat malnutrition among young children—Uruguay, and Nicaragua, which has launched a new social safety net program (Hambre Cero).

ch02fig42

Levels and Targeting of Social Spending

Social Spending

(In percent of GDP)

Source: National authorities.1/ Unweighted averages of 13 countries, averaged over 2003-06. Country coverage varies by category.
ch02fig43

Distribution of Social Spending Between the Rich and the Poor

(In percent of total)

Source: ECLAC (2006).1/ Unweighted averages. Country coverage varies by category. Data from household surveys between 1998 and 2003.
1

Changes in noncommodity primary balances—that is, changes in primary balances minus changes in commodity revenues—are a better indicator of fiscal contractions or expansions than primary balances per se, because changes in commodity revenue are mostly driven by payments by nonresidents, and as such do not have the usual direct impact on domestic disposable income (see the November 2006 Regional Economic Outlook: Western Hemisphere).

2

Several countries also initiated or completed changes to their pension systems. In Chile and Mexico, these reforms were aimed mainly at boosting competition among private pension fund administrators, and, in the case of Chile, at widening coverage. In Argentina, the government widened the scope of the public system, including by moving government employees and allowing other workers to switch to the public system.

3

Projections by the Economic Commission for Latin America and the Caribbean (ECLAC) show a decline in the overall poverty rate.

4

National poverty lines are not strictly comparable across countries, but an attempt was made to choose similar definitions when multiple definitions were available in a country. Using internationally comparable poverty headcount ratios (percentage of the population who live under $2 a day) leads to a broadly similar country ranking. Unfortunately, headcount ratios are not available after 2004; gauging recent trends hence requires the use of national data.