V. Financial Sector Development in Latin America: Recovery or Overheating?
- International Monetary Fund. Western Hemisphere Dept.
- Published Date:
- November 2007
In Latin America and the Caribbean, local credit and capital market activity have picked up strongly in recent years, with bank credit growing in double digits in several countries (see Chapter 2). As the external environment becomes more challenging, the question arises to what extent this increased intermediation mirrors the region’s rising resilience and not a renewed buildup in vulnerabilities.
This chapter examines the characteristics of recent financial sector developments in the LAC region to shed light on the sustainability of the region’s recent rapid credit and financial sector growth, focusing particularly on bank credit. The analysis addresses three questions. The first is whether rapid increases in bank credit and capital market activity have been associated with deepening and broadening of markets, as measured by underlying improvements in banking efficiency, market liquidity, and the number of new stock or bond issues. The second is whether the strong credit growth observed in several countries signals an excessive credit boom that cannot be sustained, or is more likely to reflect mostly improved macroeconomic and institutional fundamentals. The third issue concerns the key regulatory and institutional challenges remaining in the region, based on a survey of recent financial sector assessments.
Financial Deepening or Bubbles
Financial deepening strengthens an economy’s resilience to adverse shocks through stronger financial institutions, greater risk diversification, and more efficient allocation of capital, which also promotes economic growth. Stock and bond market development contributes to growth through the creation of liquidity, which makes investment less risky and more attractive while companies raise long-term capital (Levine, 1996). Well-functioning derivatives markets can contribute to this process through risk sharing and lowering the costs of managing complex risks.
Latin America’s financial markets continue to lag those in other emerging markets. Recent data show that this remains the case, as measured by bank credit–to-GDP ratios, market capitalization, and capital raised in equity markets. Market liquidity for both stocks and corporate bonds also remains relatively low in the region.
Domestic Market Capitalization
(In percent of GDP) 1/
Sources: Bank for International Settlements; and the World Federation of Exchanges.
1/ GDP weighted averages. Defined as the sum of local corporate bonds outstanding and domestic stock market capitalization.
Stock Turnover Ratio
(In percent) 1/
Source: World Federation of Exchanges.
1/ Unweighted averages. Defined as the ratio of value of domestic shares traded to market capitalization.
However, recent developments in the region indicate some genuine deepening and broadening in the LAC region’s financial markets. First, overhead costs, a measure of banking intermediation efficiency, have declined since 2001. Second, domestic stock and corporate bond markets in the larger Latin American countries have expanded in terms of total capitalization, although partly driven by higher asset prices (see Chapter 2). Third, markets have become more liquid on various measures, including the total values of stocks and bonds traded measured relative to market capitalization (i.e., turnover ratio), as a share of GDP, and relative to market volatility. Fourth, new stock and corporate bond listings across the region, although still low, have risen. They reached an annual average of 8 and 200, respectively, during the period 2004–06, compared with 4 and 155 during the previous three-year period. Finally, markets for exchange-traded foreign exchange derivatives and over-the-counter foreign exchange and interest rate derivatives have expanded in several countries (see Box 10).
Box 10.Derivatives Markets in Latin America
Well-functioning derivatives markets can provide significant benefits to corporations, financial institutions, and institutional investors by allowing them to improve risk management and lower funding costs. For example, greater accessibility and diversity of derivatives help alleviate exchange rate risk facing Latin American firms and facilitate commodity price hedging in some of the region's largest economies.
Partly reflecting these benefits, derivatives trading has surged in the larger economies in the region. Brazil, Mexico, Colombia, and Chile combined registered a daily trading volume of close to US$110 billion (notional) in 2006. Interest rate derivatives (swaps, options, and forward rate agreements) represented about 70 percent of total trading activity. Currency derivatives (FX forwards and swaps) have been growing rapidly in the wake of increased exchange rate and capital account flexibility as well as greater trade and financial integration. The use of other derivatives, such as those on local credit, remains limited.
With an average daily trading volume of US$46 billion in 2006, Brazil boasts the largest derivatives market in Latin America, followed closely by Mexico and Argentina. Trading activity in Brazil is dominated by exchange-traded interest rate and currency futures, while over-the-counter (OTC) derivatives trading (mainly interest rates) is relatively minor. In contrast, derivatives trading in Mexico and Argentina is focused on currency-based instruments (mostly on very short-term government debt) and commodities respectively.
In other countries in the region, derivatives trading is much smaller and occurs mostly OTC in either FX or fixed income assets. In Chile, trading is concentrated in onshore, short-term, OTC-traded FX derivatives contracts and in offshore nondeliverable forwards, motivated by limited foreign holdings in the domestic bond market, high transaction costs onshore, and regulatory limits on pension funds’ use of interest rate swaps and forwards. The derivatives market in Colombia is considerably smaller.
There is scope to develop these markets further, while strengthening regulatory oversight to mitigate potential downside risks. Reform priorities could include:
Strengthening the regulatory and legal frameworks for derivatives markets. Best practices include consistent capital rules, collateral requirements, and netting provisions; transparency-enhancing full balance sheet disclosure; accounting rules aligned with international accounting standards; a tax environment that creates a level playing field for all cash and derivatives trading; and provisions for short-selling.
Fostering exchange-based derivatives trading. In Mexico, MexDer offers a diverse set of derivatives instruments, and a well-functioning electronic settlement and clearing system enhances effective monitoring of trading activity and efficient execution of trades. In Brazil, BM&F has a solid track record in product innovation, asset diversification, and broadening of investor base. Both Colombia and Costa Rica are exploring possibilities to expand existing and create new exchanges for derivatives trading.
Encouraging a broad and balanced investor base for genuine hedging. For derivatives markets to be effective, especially in dollarized countries, the creation of complementary hedging interest is critical. Commercial banks with short-term liabilities (and long-term fixed-rate holdings) and institutional investors with long-term, foreign currency holdings have complementary term structures, with the latter acting as net suppliers of foreign currency or floating rate sellers. Removing limits or prohibitions on pension funds and insurance companies thus helps promote hedging and capital market liquidity in general. For example, the recent elimination of investment limits and transaction taxes for foreign investors in Mexico helped increase local bond market liquidity.
Credit Boom or Recovery
Recent rapid growth in bank credit needs to be viewed in light of the preceding long period of declining credit-to-GDP ratios (see the November 2006 Regional Economic Outlook). Moreover, to the extent that economic growth brings with it a certain degree of financial deepening, bank credit should be expected to follow a rising trend during an economic expansion such as the region has been experiencing. But as double-digit credit growth has continued in a number of countries across the region, it is important to evaluate whether this growth is sustainable. This is especially true for Latin America given the region’s history of susceptibility to “boom-bust” credit cycles.
In this section, two approaches were used to identify a potentially unsustainable credit boom: deviations of bank credit–to-GDP ratio from trend levels (Gourinchas, Valdés, and Landerretche, 2001) and a comparison of actual bank credit–to-GDP ratio with the estimated “potential” level, given a country’s macroeconomic and institutional fundamentals.
New Capital Raised from Shares
(In percent of GDP) 1/
Source: World Federation of Exchanges.
1/ Unweighted averages. Defined as amount of new capital raised through the sale of new shares issued by a new issuer, capital increases by already listed companies (reserved to previous shareholders), and Secondary Public Offerings (new shareholders subscribe to the shares).
Comparing Stock Prices during Market Turbulence Periods
(Percent change; local currency) 1/
Source: Bloomberg, L.P.
1/ Unweighted averages.
Bank Overhead Costs
(In percent of total assets) 1/
Source: Fitch's BankScope database.
1/ Unweighted averages.
In the first approach, the underlying trend is measured using a country-specific filter. A credit boom is identified when the deviation of the actual credit-to-GDP ratio from this estimated trend exceeds a given threshold. Threshold values are defined in terms of both absolute deviation (i.e., the simple difference between actual and trend credit to GDP) and relative deviation (i.e., the difference between actual and trend credit to GDP in percent of trend credit to GDP). The threshold values are selected so that they capture most of the past episodes of credit booms and/or banking crises (see Appendix).
In recent years, all but 6 of the 20 largest countries in Latin America and the Caribbean experienced increases in credit to the private sector relative to GDP. In 12 of these, sharp increases in credit pushed the ratio of credit to GDP over its trend level (see panel chart on actual and trend private sector credit). However, until 2006, no country experienced credit growth in excess of the “threshold” values for defining a boom. Last year, continued credit expansion pushed credit above one or both thresholds in five countries: Colombia, Honduras, Jamaica, Mexico, and Nicaragua. Past credit booms typically lasted several years before collapsing, although they suggest that continued debt expansion at this pace could warrant concern. It is thus too soon to conclude that these developments constitute an excessive or unsustainable “boom.”
Credit booms may also be identified on the basis of predictions of an economic model that links credit to fundamentals. The capacity for a country’s banking system to extend credit safely to the private sector depends both on economic conditions and on the quality of financial and public institutions. In this context, a credit “boom” could be viewed as credit growth that cannot be justified by economic fundamentals and institutional factors.
How do LAC countries measure up against other regions in the key determinants of financial development identified in the literature? As a number of authors have pointed out (de la Torre Gozzi, and Schmukler, 2006; and Braun and Hausmann, 2003), the still relatively low level of financial development is disappointing, given that several countries in the region have undertaken significant financial sector reform efforts since the 1980s. A regional comparison of underlying institutional determinants of financial development shows that while the LAC region is relatively strong in areas such as credit information and credit bureaus, it remains weak in areas such as creditor rights and governance. Moreover, even with the recent macroeconomic improvements, inflation and real interest rates also remain relatively high.
How empirically important are these weaknesses in holding back financial development in the Latin American and the Caribbean region? New econometric analysis using a panel of 79 countries over the period of 2001–05 shows that inflation, institutional quality (as measured by the World Bank’s governance index), and real deposit interest rates have significant effects on the ratios of private sector credit to GDP. For example, a 1 percentage point improvement in each of these dimensions is associated with an increase in private sector credit of over 1 percent of GDP.
Comparisons of Key Determinants of Financial Development 1/
The effectiveness of creditor rights protection, captured by the variable “legal origin,” is also shown to matter—legal reforms toward stronger protection of creditors and investors may be expected to lead to significant increases in private sector credit relative to GDP.18 Finally, productivity growth that leads to higher GDP per capita may also help boost credit intermediation. Given these results, where then do LAC countries stand relative to their “potential”?
A comparison of actual credit to GDP with the level that would be predicted by the model shows that many of the 20 largest LAC countries have been performing below their model-fitted potentials in recent years (see panel chart on actual and fitted credit). Of the five largest emerging markets in LAC, only Brazil has bank lending to the private sector at a level close to its “potential,” whereas Chile, Mexico, Peru, and Colombia have still to benefit from their relatively strong fundamentals. Under this criterion, there is no indication of excess credit growth in Colombia and Mexico. Four countries—Bolivia, Haiti, Honduras, and Panama—are found to have credit-to-GDP ratios higher than their model fitted potentials. However, in two of these cases—Bolivia and Panama—higher-than-predicted credit ratios reflect the slow unwinding of credit booms during the 1990s rather than recent credit expansion. In the case of Haiti, it reflects the collapse (and slow recovery) of fundamentals as a result of political and social strife.
Slow adjustment of actual credit to improvements in fundamentals—either as a result of reforms in the 1990s and this decade, or from postcrisis recoveries—could also be the reason why bank credit remains below “potential” in many Latin American countries. In addition, credit levels that appear to trail fundamentals could reflect factors that are not included in the analyses but nevertheless tend to reduce credit intermediation, such as financial transaction taxes, interest rate controls, and directed lending.
In summary, the two commonly used criteria for “excess” credit growth analyzed for the region suggest that much of the recent credit growth seems to be generally associated with improvements in fundamentals, and that there are not yet clear signs of general overheating in the region. This finding is consistent with the latest prudential indicators and equity market–based bank solvency estimates, which show that nonperforming loan ratios and banks’ estimated default probabilities have remained at low levels (see Chapter 2 and Box 9). However, the aggregate picture may mask heightened vulnerabilities in certain financial institutions that have lowered credit standards in their pursuit of rapid expansion. While the extent of such vulnerabilities is not clear owing to the difficulty in obtaining up-to-date information on the institutional distribution of credit growth, the rapid pace of credit growth in some countries warrants enhanced regulatory oversight and prompts measures to strengthen supervisory capacity.
Regulatory Responses and Options
A survey based on Financial Sector Assessment Programs (FSAPs) and IMF Article IV consultations for the region shows that, while considerable progress has been made in strengthening and developing the region’s financial sectors, there remains scope for improvement in a number of areas. First, enhancing regulatory independence and risk-based supervisory capacity remains a top priority in many countries, especially where credit has been growing rapidly. Second, significant increases in foreign participation in several local capital and credit markets make effective consolidated and cross-border supervision more important. Third, the increasing role of nonbank financial institutions such as finance houses and mutual funds present a new regulatory challenge. Fourth, improving bank resolution framework and financial safety nets (such as the clarification of lender-of-last-resort functions and the role of deposit insurance, and crisis contingency planning) will be important to sustained financial stability. Finally, the recent growth in the derivatives markets in several Latin American countries makes it important for regulators to be able to monitor and evaluate the risks associated with financial innovations. The experience of the relatively developed markets in the region (mainly Brazil, but also Chile and Mexico) points to the importance of regulatory measures, disclosure requirements, and documentation standards for compliance with capital rules and smooth functioning of the derivatives markets.
Progress in Financial Sector Reform
Source: IMF staff estimates. Based on information collected on the 20 largest countries in the LAC region.
Underpinned by institutional reforms and stronger policy frameworks, the region’s financial markets have deepened and broadened over the past several years. The analysis in this chapter suggests that recent rapid credit growth observed in the region mostly represents a continued catching up in the level of financial intermediation, following earlier crises and important financial sector reforms. However, the pace of credit growth indicates that heightened regulatory oversight may be needed in coming years. As governments in the region recognize, further reforms to deepen and broaden financial markets, and strengthen the institutional structure, will be important to bolster resilience going forward.
Two types of analyses are conducted in this chapter to give a range of indications for the extent that credit growth has exceeded a given trend level.
Filter-Based Trend Analysis
We derived the trend by applying a retrospective rolling Hodrick-Prescott filter for each country. This approach can be economically meaningful in that the threshold is selected based on past episodes of credit booms and/ or banking crises. Threshold values are defined in both relative deviation (i.e., the difference between actual and trend credit-to-GDP ratios relative to trend credit-to-GDP ratio) and absolute deviation (i.e., the difference between actual and trend credit-to-GDP ratio). The former adjusts for the different degree of financial deepening, and the latter for the different size of economy. A relative threshold of 18 percent and/or an absolute threshold of 3 percent capture a majority of the past credit booms or banking crises (e.g., 1994–97 Mexico crisis, 1994–99 Brazil crisis, 1981–83 Chile crisis, and 2002–03 Uruguay crisis).
The IMF’s International Financial Statistics data on banks’ claim on private sector and nominal GDP are used for the period of 1960–2006, and World Economic Outlook projections for nominal GDP are used for 2007 so that bank credit–to-GDP ratios are calculated as bank credit of the current year divided by the average of the current year GDP and the following year GDP.
Estimation of the “Potential” Level of Bank Credit to GDP
We included both economic fundamentals and policy and institutional variables that are found important for financial development in the literature. Djankov, McLiesh, and Schleifer (2006) found that both creditor protection through the legal system and information sharing institutions are associated with higher ratios of private credit to GDP, although the former is relatively more important in the richer countries and the latter in developing countries. A study by Huang (2007) found that trade openness and governance have positive effects on financial development, whereas civil legal origin, inflation volatility, and policy instability have negative effects on financial development. He also found initial conditions and geography important—countries with a smaller land area, higher initial GDP and population, more open trade policy, and stronger institutions have a higher level of financial development. A recent study by Dehesa, Druck, and Plekhanov (2007) found that higher credit-to-GDP ratios are associated with stronger creditor rights and lower inflation.
Sources: IMF, International Finance Statistics ; and IMF staff calculations.
1/ Credit is the deposit-taking institutions' claims on the nonbank private sector. The trend is obtained by fitting a Hodrick-Prescott filter for each country, with paramater set at 100.
Actual and Panel Fitted Credit, 2001-06
(Ratio to GDP) 1/
Sources: IMF, International Finance Statistics; and IMF staff calculations.
1/ Credit is the deposit-taking institutions' claims on non-bank private sector. The trend is model-fitted credit-to-GDP ratio.
We used GMM IV regressions to formally estimate the effects of various factors on the banks’ credit to private sector as percent of GDP. Because of the possible endogeneity from using the concurrent level of GDP per capita, it was instrumented using lagged values of explanatory variables, as well as initial GDP and population and an index of ethnic fractionalization. The first-stage regression results suggest that endogeneity is indeed a problem, and IV relevance test and overidentification test support the use of these instruments. Time dummies are included to capture changes over time in global markets and domestic banking structure.
The table on p. 63 shows results from the preferred regression specifications, which are generally robust to changes in specifications. The estimates are similar to those found in the literature, with coefficients generally having the expected signs.19
The estimations of the “potential” level of bank credit to GDP are based on a panel of 79 countries from 2001 to 2005. Annual data for PPP-adjusted real GDP per capita, which are measured in 2000 international dollars, and CPI indices are taken from the World Bank’s World Development Indicators database. Data for bank credit to the private sector relative to GDP are from the World Bank’s new database on financial development and structure. Data for deposit interest rates are taken from the IMF’s International Financial Statistics database; where deposit interest rates are not available, treasury bill rates are used. Data for trade restrictiveness are taken from the IMF’s trade policy information database, which includes ratings based on tariff structures and nontariff trade barriers. The ratings take the value of 1 to 10, with 10 being the most restrictive in trade regime.
The main data for institutional quality are from the World Bank and Huang (2007). Governance and institutional quality is the simple average of the six measures of institutional development: voice and accountability, political stability, government effectiveness, regulatory quality, rule of law, and control of corruption, based on extensive surveys conducted by Kaufmann, Kraay, and Mastruzzi (2007). Credit information and private and public credit bureau coverage (as percent of total adult population) are taken from the World Bank’s Doing Business database. Indexes of shareholder rights and creditor rights are based on La Porta and others (1998) and range from 0 to 6 and 0 to 4, respectively, with higher values being stronger rights.
The instrumental variables (real GDP per capita and total population in 1990, and index of ethnic fractionalization) are taken from the Penn World Tables 6.1 and Alesina and others (2003), respectively. The legal origin dummy for the French, German, and Scandinavian systems is based on the World Bank’s Global Development Network database. Data for the land area in square kilometers is based on Gallup, Sachs, and Mellinger (1999).20
|Relative Threshold = 18%||Absolute Threshold=3%||Banking Crisis 1/|
|Argentina||1980-82, 1998-2000||1980-81, 1998-2000||1980-82, 1989-90, 1994-95, 2001-02|
|Brazil||1988, 1994-97||1988, 1994-97||1990, 1994-99|
|Colombia||1984, 2006||1982-84, 1996-98, 2006||1982-87, 1999-2000|
|Mexico||1992-94, 2006||1992-94, 2006||1982, 1992-97|
|Honduras||1987-88, 1998-99, 2006|
|Boliva||1982, 1999||1982, 1993-94, 1997-2000||1986-87, 1994-95|
|Uruguay||1981-82, 1998-2002||1981-83, 1998-2002||1981-84, 2002-03|
|Ecuador||1983-86, 1997-98||1983-86, 1995, 1997-98||1982-84, 1996-2002|
|El Salvador||1985||1983-85, 1989, 1997-2000|
|Jamaica||1989, 1998-99||1982-83, 1988-89, 1998-2000, 2006||1994-2000|
|Nicaragua||1980-82, 1992-93, 1999-2000||1980-82, 1992-93, 1998-2000, 2006||1990-96, 2000-01|
|Panama||2001||1987, 1994, 1998-2000|
|Paraguay||1982, 1994||1994, 1997, 2001||1985, 1995-2000|
|Trinidad and Tobago||1985, 1987, 1998|
|Venezuela||1987, 1997, 2005||1986-87||1993-95|
(Dependent variable: private sector credit to GDP) 1/
|Log of real GDP per capita, PPP adjusted||0.11||0.00||0.10||0.00||0.11||0.00|
|Land area in log||-0.04||0.00||-0.03||0.00||-0.03||0.00|
|Credit information index||0.01||0.31||0.01||0.21||0.01||0.21|
|Real interest rates||-0.27||0.12||-0.27||0.11||-0.28||0.10|
|Governance and institutional quality||0.19||0.00||0.18||0.00||0.18||0.00|
|Civil legal origin dummy||-0.07||0.01||-0.07||0.01||-0.08||0.00|
|International Financial Center||0.09||0.23||0.08||0.27|
|Number of observations||370||370||370|
|OECD||Latin America and Caribbean||East Asia and Pacific||Middle East and North Africa||Europe and Central Asia||South Asia||Sub-Saharan Africa|
|Legal rights index||6.3||88||3.8||80||6.5||52||3.8||48||5.6||72||4.8||16||4.5||87|
|Credit information index||5.0||109||4.8||98||4.0||52||2.9||48||3.2||89||3.0||16||1.3||87|
|Private credit bureau coverage||8.8||88||10.9||80||4.9||52||10.0||48||2.5||72||0.2||16||1.5||87|
|Public credit bureau coverage||59.0||84||38.5||80||36.9||44||14.6||48||9.9||64||2.6||16||4.4||87|
|Shareholder rights index 1/||3.0||88||2.7||36||3.4||40||2.2||20||2.0||4||4.3||12||0.9||13|
|Creditor rights index 1/||1.9||88||1.3||32||2.8||40||2.5||16||2.0||4||3.7||12||1.3||13|
|Civil legal origin dummy||0.7||109||0.9||98||0.3||52||0.8||48||1.0||5||0.0||16||0.6||87|
|Log 1990 population||9.8||109||9.2||98||10.0||48||9.5||36||10.9||5||11.6||16||9.2||87|
|Log 1990 real GDP per capita||9.8||109||8.3||98||8.9||48||8.7||36||8.7||5||7.5||16||7.1||87|
|Log land area||11.8||109||12.5||98||11.5||52||12.4||48||13.6||5||12.8||16||12.2||87|
|Ethnic fractionalizaion index 2/||0.2||104||0.4||98||0.3||52||0.3||48||0.3||5||0.4||16||0.7||87|
|PPP-based real GDP per|
|Real interest rate||0.2||93||1.0||98||1.5||52||2.0||38||2.2||69||0.8||12||1.1||80|
Following La Porta and others (1998), common-law countries generally protect investors the most, and civil-law countries protect them the least.
Contrary to Huang (2007), trade restrictiveness is associated with higher credit-to-GDP ratio. While this may reflect the lack of variation in data, one possible explanation is that countries that are less open are also protective of their domestic banking sector, promoting domestic financing of trade.
Data for most of the time invariant variables are obtained from Huang (2007).