Latin American Structured Finance Scenario ... - Standard & Poor's

June 21, 2012Structured FinanceResearchLatin American Structured FinanceScenario And Sensitivity Analysis:The Effects Of Regional MarketVariablesPrimary Credit Analysts:Ignacio Estruga, Buenos Aires (54) 114-891-2106; ignacio_estruga@standardandpoors.comRenata Lotfi, Sao Paulo (55) 11-3039-9724; renata_lotfi@standardandpoors.comMaria del Sol S Gonzalez, New York (1) 212-438-4443; maria_gonzalezcosio@standardandpoors.comDaniel Castineyra, Mexico City (52) 55-5081-4497; Daniel_Castineyra@standardandpoors.comSecondary Contact:Erkan Erturk, PhD, New York (1) 212-438-2450; erkan_erturk@standardandpoors.comAnalytical Manager, Emerging Markets Structured Finance:Juan P De Mollein, New York (1) 212-438-2536; juan_demollein@standardandpoors.comTable Of ContentsKey Factors For The Latin American Consumer And Mortgage SectorsHow Do The Key Variables Affect Credit Quality?Historical Collateral Performance TrendsScenario Analysis Could Shed Light On Future Credit QualitySensitivity AnalysisTransaction-Specific Factors Cannot Be IgnoredRESEARCHThe Changing Global Economy And Its Impact On Latin AmericaRelated Criteria And ResearchAppendixwww.standardandpoors.com/ratingsdirect 1978790 | 300129047

Latin American Structured Finance ScenarioAnd Sensitivity Analysis: The Effects OfRegional Market VariablesThe economies of Latin American countries have steadily grown since withstanding the most recent global recession.However, the region is not immune to global macroeconomic changes. As such, these economies will likely continueto grow in 2012, but at a slower pace. The uncertain global backdrop could negatively affect the labor and exportmarkets, as well as domestic demand, which could stifle growth and ultimately affect ratings on structured financetransactions issued out of Latin America.Standard & Poor's Ratings Services is providing an analysis of the key economic variables that, in our view, wouldlikely affect the collateral performance and ratings on the most relevant asset classes for each of the region's mostactive structured finance markets: Brazilian and Argentinean consumer loans and Mexican residential mortgages.This article is related to "Global Structured Finance Scenario And Sensitivity Analysis: The Effects Of The Top FiveMacroeconomic Factors," published Nov. 4, 2011, in which we provided an overview of how structured financeratings across sectors changed over time based on their sensitivity to different economic variables. To provide themarket with more information on structured finance securities in specific countries, we plan to publish periodicscenario analyses on specific sectors.This study focuses on 90-plus-day delinquencies in the Brazilian and Argentinean consumer loan and Mexicanresidential mortgage markets from 2000 to 2011. We understand that multiple factors may simultaneously affect thecredit performance of underlying collateral in structured transactions. However, we have identified severalmacroeconomic factors and lending standards that we believe are most relevant to changes in credit quality, as wellas credit ratings on structured transactions.Analysis of data from 2000-2011 suggests an empirical link between the Brazilian consumer asset delinquency andunemployment rate, consumer leverage, credit growth, and share of short-term debt, as well as between Argentineanconsumer asset delinquency and GDP growth, unemployment rate, real salaries growth, and credit growth. Mexicanresults suggest that the local mortgage market delinquency is related to inflation, economical performance of thesecondary and tertiary sectors, and credit growth. Transaction-specific factors, such as servicer quality orcounterparty risk, also affect credit performance.For this analysis, we first examined the historical correlation and regression results between macroeconomic andmarket variables and the evolution of delinquency performance for the loans in each country. We then conducted ascenario analysis to determine the impact on asset performance based on the changes in the aforementionedvariables at three distinct stress scenarios: our current expectations or base-case for the rest of 2012, our recovery,and downturn trend scenarios. Lastly, we describe a rating impact analysis on hypothetical transactions consideringthe pool losses under these scenarios.Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 2978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesOverview• This scenario and sensitivity analysis on Latin American consumer asset and residential markets provides a more granularempirical analysis of regional structured finance securities.• This analysis focuses on the key economic variables that, in our view, would affect the underlying collateral performance andratings on the most relevant asset classes in each of the region's most active markets: Brazilian and Argentinean consumer loansand Mexican residential mortgages.• Data across 2000-2011 show that GDP growth, unemployment, credit growth, household leverage, wages, and price dynamicshave been key variables in these markets.• Under our base-case scenarios, for each of the analyzed markets, we expect credit performance to modestly deteriorate during therest of 2012 given our region economic outlook. We do not expect major rating shifts.• Our downturn trend scenarios for Argentina anticipate deterioration in portfolio quality by an average of 5x-5.5x, and as a result,lower ratings.• Data for Brazil and Mexico offer limited insight for the assets' future credit quality under extreme economic conditions since thecountries have benefited from a benign economic environment over the past decade.Key Factors For The Latin American Consumer And Mortgage SectorsWe focused our analysis on the relationship between macroeconomic factors, lending standards, and delinquencyrates in Brazilian and Argentinean consumer loans and Mexican residential mortgage portfolios. We collected datafrom each country's Central Bank and incorporates aggregated consumer and mortgage loans originated by all thefinancial institutions of each market. We use this data as a proxy for structured finance portfolio performance. Wenote that some of the data series have limitations. Notably, we had to estimate some monthly values that the CentralBank had not reported. It is our understanding that the data issue is most pronounced during major recessions. Forinstance, data on Argentinean consumer nonperforming loans suffered an interruption for some months during the2001-2002 crisis. To account for these missing values, we provided estimations based on the published values of theoverall variation in aggregated loan performance.In our study, we assessed the correlation between 90-plus-day delinquencies (loans that we believe have a highlikelihood of default) and the different economic variables. We also considered the correlation coefficients betweenthe economic factors themselves in order to determine the key variables that we consider to be the most influentialcontributors to consumer and mortgage portfolio performance. After performing the correlation analysis, weidentified uni-variate and multi-variate coefficients to assess the change in defaults based on the change in the givenvariable. (Please refer to Appendix section for details.)Based on our analysis, the following variables are, in our view, some of the key influential factors on Latin Americancollateral performance.Brazilian consumer loan performanceTo analyze performance of Brazilian consumer loans, we focused on personal and auto loan portfolios as defined byBrazilian Central Bank (Banco Central do Brasil). We conducted a separate analysis for both portfolios because autoloans tend to be more sensitive to adverse events than personal loans. Our analysis suggests a historical link betweenthe consumer delinquency performance and the following variables: unemployment, consumer leverage, 12-monthwww.standardandpoors.com/ratingsdirect 3978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market Variablescredit growth, and share of short-term debt.Argentinean consumer loan performanceTo assess performance of Argentinean consumer loans, we focused on aggregated consumer loan portfolios asdefined in the BCRA (Banco Central de la República Argentina) Financial Institution's reports. Our analysis suggestsa historical link between consumer asset delinquency performance and the following variables: GDP growth,unemployment, real salary growth, and credit growth.Mexican mortgage loan performanceTo analyze performance of the Mexican mortgage market, we focused on aggregated mortgage portfolio as definedby the Central Bank (Banco de México) reports. Our analysis suggests a historical link between the assetdelinquency performance and the following variables: inflation, real GDP, performance on secondary (defined as thesector of the transformation of raw materials) and tertiary (mainly services industry) sectors, and credit growth.Key variables help us better understand and assess the current and possible future state of Latin American structuredfinance securitized portfolios. We classify our key economic indicators under two categories: leading and parallel.Wages, GDP growth, and household leverage appear to be leading indicators because their variations historicallyprecede corresponding changes in loans credit quality. By contrast, we consider unemployment and credit growth asparallel indicators because they historically trend up or down at the same time as asset delinquencies.How Do The Key Variables Affect Credit Quality?GDP growthGDP growth is crucial to asset credit quality performance because it represents an overall measure of the country'seconomic health. Fluctuations in GDP growth reflect the creditworthiness of various economic agents, in our view,including borrowers of the loans backing structured finance transactions.UnemploymentUnemployment is also indicative of credit quality because it directly affects household creditworthiness, consumerconfidence and aggregate income level. It is, therefore, a good proxy for collateral pool performance in transactionsbacked by mortgages and consumer loans, such as residential mortgage-backed securities (RMBS) and someasset-backed securities (ABS).Inflation and salariesPrice dynamics and their impact on household salaries are relevant to our analysis because they directly affect thehouseholds' disposable income to honor debts. Inflation can hurt borrowers' creditworthiness because it mayincrease household spending.Credit growthCredit growth tends to capture the credit cycle. Usually, growing credit origination yields more available liquidityfor consumers to refinance debt, which could mask delinquency rates. Also, during periods of favorable andexpansionary credit, we believe financial institutions may adopt less stringent underwriting standards, which maybring borrowers with low credit scores to the credit market. This would likely prompt upswings in delinquency.Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 4978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesConsumer leverageConsumer debt is a leading indicator of delinquency behavior because when it is high, borrowers are more sensitiveto adverse events and a potential liquidity shortfall.Share of short-term debtThe interest rate charged on short-term debt (credit cards and overdraft lines) is very high in Brazil, Argentina, andMexico. Rises in short-term debt would hinder the consumer's ability to repay his/her debt. This may also reflect aselection of lower creditworthiness among borrowers.Property pricesAlthough we have not listed property prices as a key variable (because of information constraints), we consider it tohave a strong influence on structured finance credit quality. Property price increases and declines, respectively, lowerand raise leverage and, to some extent, risk in outstanding loans secured by property, such as those backing RMBS.Historical Collateral Performance TrendsMacroeconomic factors have varying effects on the collateral credit quality of Latin American consumer andmortgage loans that back ABS and RMBS. We have reviewed several years' worth of data to determine thecorrelations between economic variables and aggregate collateral performance.Overall, Latin American consumer and mortgage loans exhibited moderate volatility in performance during the lastglobal recession (see chart 1). Importantly, the availability of data on asset performance during times of crisis is verylimited since the level of stress throughout the regions is relatively small (see "Understanding Standard & Poor´sRating Definitions," published June 3, 2009). In this sense, the most relevant crises were:• Mexico, 1994, when the real GDP declined 15% during a period of nine months and there was no history onunemployment rate behavior; and• Argentina, 1998 – 2002, when the GDP declined 25% during a period of 21 months, and the unemployment ratereached 21%.Apart from that, there is not much data available on the relevant variables during the Latin America debt crisis in1981-1982.The relationship between the key macroeconomic factors and the credit quality of the Brazilian and Argentineanconsumer loans and Mexican mortgage loans follow the economic cycle. Periods of economic growth, for instance,tend to coincide with higher employment and rising property prices. Delinquencies are more likely to decline orremain stable under such benign economic conditions.www.standardandpoors.com/ratingsdirect 5978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 1BrazilSubsets of Brazil consumer loans have not performed uniformly. We believe the greater stability of personal loanportfolios over auto portfolios reflects the fact that payroll-deductible loans were historically about 65% of personalloans, 85% of which were to public workers, who enjoy greater job stability than those in the private sector.Labor market conditions have been solid in Brazil for the past years, and unemployment decrease has shown a highcorrelation with the decline in personal loan delinquencies (see chart 2).Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 6978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 2On the other hand, the auto loan industry delinquency levels rose even considering unemployment rate drops.Consequently, we anticipate upticks in unemployment could prompt delinquencies to increase to historically highlevels.We found that delinquency for auto loans were tied to credit lending growth (see chart 3). When credit lendingdecelerates, delinquency tends to go up (see chart 4).www.standardandpoors.com/ratingsdirect 7978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 3Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 8978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 4Consumer leverage and short-term debt are other key variables that could explain delinquencies among Brazilianconsumer loans. We found that consumer leverage is more linked to delinquencies among auto loans than personalloans, while short-term debt was more correlated to personal loans (see chart 5).www.standardandpoors.com/ratingsdirect 9978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 5ArgentinaThe overall health of the Argentina's economy helps to shape the performance of Argentinean consumer collateralthat backs Argentine deals ABS deals. In 2002, for instance, the republic went through the most stressful crisis in itshistory, in which political, economic, and social turmoil resulted in huge collateral credit quality deterioration.Nonperforming consumer loans peaked at 30.48% in December 2002, almost a 5.5x increase from December2001´s 5.63% rate. After the 2002 turmoil, the favorable global economy helped pick up the Argentinean economy,and in turn, the credit quality of the country's consumer loans. During the benign period between 2004 and 2007,the percent of nonperforming loans stabilized near 3.5%. Performance deteriorated steadily, though moderately,among Argentinean consumer loans during the global crisis between first-quarter 2008 and fourth-quarter 2009;nonperforming loans rose up to 5.57% during that time. Since then, even with the slow recovery of the globaleconomy, collateral has been improving, with nonperforming loans reaching a historical low-year average of 2.89%.Year-over-year changes in Argentina's quarterly GDP showed a negative correlation with credit quality of consumerStandard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 10978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market Variablesloans (see chart 6).Chart 6Looking at Argentina's unemployment rate from 2000 to 2011 shows a positive correlation with credit quality ofconsumer loans (see chart 7).www.standardandpoors.com/ratingsdirect 11978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 7Real salaries (inflation adjusted salary) growth with a six month lag of Argentina's workers showed an inverserelationship with credit quality of consumer loans (see chart 8).Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 12978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 8Credit growth shows an inverse relationship with credit quality of Argentinean consumer loans (see chart 9).www.standardandpoors.com/ratingsdirect 13978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 9MexicoMexico residential-mortgage performance has been sensitive to the macroeconomic environment. In fact,macroeconomic changes tend to immediately affect Mexican RMBS performance. After Mexico's 1994 economiccrisis, residential mortgage performance severely deteriorated in 1995, and delinquencies continued to rise, peakingat 29.53% in 1998. In addition, the world's financial and economic crisis of 2008 affected the Mexican residentialmortgage market almost immediately, during which the nonperforming loan ratio grew to 4.61% from 2.74%(though less harsh than the increase during the 1994 crisis).Year-over-year changes in Mexico's quarterly GDP exhibits an inverse relationship between GDP and residentialmortgage performance (see chart 10).Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 14978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 10Inflation showed a positive correlation with performance of Mexican residential mortgage market during 2000-2011(see chart 11).www.standardandpoors.com/ratingsdirect 15978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 11Secondary and tertiary sector performance exhibit an inverse correlation with residential mortgage delinquencies inMexico (see chart 12).Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 16978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 12Finally, the credit growth and the non-performing mortgage loans show a negative correlation (see chart 13).www.standardandpoors.com/ratingsdirect 17978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesChart 13Scenario Analysis Could Shed Light On Future Credit QualityAfter identifying the variables we discussed above, we further quantified the link between them and marketperformance during both benign and stressful economic periods.We were able to identify a recovery and downturn trends representing credit conditions that would help yieldpositive or stable loan performance, and worst case representing more stressed conditions that would likely result indeteriorating performance (see tables 1-3).Because Brazil and Mexico have benefited from a benign economy during the past decade, data since 2000 offerlimited insight for the assets' future credit quality under extreme economic conditions (for instance, conditionscomparable to the U.S. Great Depression). We believe that in order to address the impact of a severe economicscenario in asset performance we would need to refer to the results from the global structured finance market,although we understand that each country has its specificities and mitigating features.Table 1Brazil Recovery Trend (October 2009- October 2010)Unemployment rate Declined to 6.1% from 7.5% between Oct. 2009 and Oct. 2010Consumer leverage Increased to 20.1% from 18.3% between Oct. 2009 and Oct. 2010Credit growth12-month cumulative origination increased 22.3% between Oct. 2009 and Oct. 2010 for personal loans and78.3% for auto loansShort-term debt Decreased to 12.5% from 14.5% between Oct. 2009 and Oct. 2010Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 18978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesTable 1Brazil Recovery Trend (October 2009- October 2010)Consumer loans NPL(cont.)Decreased to 5.5% from 6.6% between Oct. 2009 and Oct. 2010 for personal loans and from 5.9% to 4.0% forauto loansBrazil downturn trend (September 2008- September 2009)Unemployment rate Equivalent to 7.7% in both Sept. 2008 and Dec. 2009Consumer leverage Increased to 20.4% from 18.3% between Sept. 2008 and Dec. 2009Credit growth12-month cumulative origination decreased 2.5% between Sept. 2008 and Dec. 2009 for personal loans and17.5% for auto loansShort-term debt Decreased to 14.9% from 15.2% between Sept. 2008 and Dec. 2009Consumer loans NPLDecreased to 6.7% from 7.0% between Sept. 2008 and Dec. 2009 for personal loans and increased to 6.0% from4.5% for auto loansSources: Brazili and Central Bank, Instituto Brasileiro de Geografia e Estatística (IBGE), and Standard & Poor’s.For the Brazilian consumer market, we expect aggregate delinquency rates, under our base-case scenario, to continueto increase modestly for the rest of 2012. We base this expectation on the consistent, but not explosive, increase inconsumer leverage and the receding growth in consumer lending. The extent to which delinquency rates will rise in2012, however, will depend on how the ongoing crisis in Europe and how conditions in the domestic labor marketplay out.Argentinean data since 2000 offer appropriate insight for assets' future credit quality under extreme economicconditions and, as a consequence, provides adequate information for calibrating a severe economic scenario for assetperformance, although some clarifications should be considered. Prior to the Argentinean crisis, an importantportion of the loans were denominated in U.S. dollars and their tenor used to be longer; consequently, we couldexpect less aggressive deterioration under an environment with the characteristics described below. The subsequentrecovery trend also provides an appropriate insight for calibrating recovery scenarios. According to "UnderstandingStandard & Poor's Rating Definitions," the described downturn scenario is commensurate with a 'AAA' stressscenario in Argentina.Table 2Key Factors And Argentinean Consumer Loans Credit PerformanceArgentina recovery trend (December 2002-December 2004)GDP Cumulative increase of 21.7% between Dec. 2002 and Dec. 2004.Unemployment rate Declined to 12.1% from 22% between Dec. 2002 and Dec. 2004.Real salary Cumulative increase of 96.95% between Dec. 2002 and Dec. 2011Credit growth Rose 20.46% between Dec 2002 and Dec. 2004Consumer loans NPL Declined to 10.61% from 30.5% between Dec 2002 and Dec. 2004Argentina downturn trend (Q4 1999-Q4 2002)GDP Cumulative decline of 25% between Dec 1999 and Dec. 2002Unemployment rate Increased to 21.5% from 13.8% between Dec 1999 and Dec. 2002Real salary Cumulative decrease of 60% between Oct 2001 and Dec. 2002Credit growth Declined 39% between Dec. 1999 and Dec. 2002Consumer loans NPL Increased to 30.5% from 5.2% between Jun. 2001 and Dec. 2002Sources: Argentine Central Bank, Instituto Nacional de Estadísticas y Censos (INDEC), Bloomberg, and Standard & Poor’s.For the Argentinean consumer market, we expect aggregate delinquency rates to increase modestly under ourwww.standardandpoors.com/ratingsdirect 19978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market Variablesbase-case scenario for the rest of 2012. We base our expectations on higher pressure on workers' real salaries givenprice dynamics, and on our expectation that the overall region economy slowdown will affect local GDP growth.Even though, we do not expect major ratings shifts.Performance and delinquency data during the Tequila crisis is limited or not comparable due to the characteristics ofthe loans that have changed as a result of the crisis. For example, prior to the Tequila crisis, mortgage loans inMexico were floating rate and could be denominated in dollars. After the crisis, interest rates grew substantially andthe peso depreciated, so this type of loan disappeared and was replaced by peso or inflation linked fixed-rate loans.Table 3Mexico Recovery Trend (1998-2002)GDP Cumulative increase of 4% between 1998 and 2002Unemployment rate 3% decline between 1998 and 2002Inflation 5% decline between 1998 and 2002Secondary sector Cumulative 14.8% increase between 1998 and 2002Tertiary sector Cumulative 10.4% increase between 1998 and 2002Mortgage loans NPL 3-4% decrease between 1998 and 2002Mexico downturn trend (1994-1995)GDP Cumulative decline of -7.2% between 1994 and 1995Unemployment rate 71.9% increase between 1994 and 1995Inflation 6.6x increase between 1994 and 1995Secondary sector 2.5% decrease between 1994 and 1995Tertiary sector 5.3% decrease between 1994 and 1995Mortgage loans NPL Increased to 14.1% from 5.6% between Dec. 1994 and Dec. 1995Sources: Banco de México, Instituto Nacional de Estadísticas y Geografía (INEGI), and Standard & Poor’s.During the rest of 2012, under our base-case scenario, we anticipate Mexican RMBS delinquency and default levelsto continue to increase but at a slower pace than what we have observed in the past 18 months. We believe this willreflect a more benign economy in Mexico that will eventually raise borrowers' disposable income and reduce theirindebtedness levels.Sensitivity AnalysisWe created four hypothetical transactions to evaluate the consequence of hypothetical movements in certainvariables presented in tables 1-8 from the Appendix. The analysis includes the potential impact on both collaterallosses and ratings for the applicable transactions.Brazilian ABS transactionTable 4 provides the characteristics of our created hypothetical Brazilian ABS transaction backed by personal loans.Table 4Hypothetical Brazilian ABS Personal Loan TransactionAssetsOriginal Pool BalanceCurrent Pool BalanceInitial Loss Projection 5%R$ 80 millionR$ 80 millionStandard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 20978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesTable 4Hypothetical Brazilian ABS Personal Loan Transaction(cont.)Liabilities Type Rating Balance ($) Credit support (%)Class A Senior brAAA 57.60 28.00Class B Junior NR 12.00 13.00Equity Equity NR 10.40For this particular pool, we specifically looked at the impact of changes in unemployment on our projected losses, aswell as on our rating on the most senior tranche while keeping all other variables the same (including securitiestarget yield, excess spread, expenses, and cash reserves, constant). Table 5 shows the potential loss projection andrating using the coefficients as listed in table 3 from Appendix below for the Brazilian market. By using thecoefficient, and since the variables are shown as natural logarithm, we are saying that based on available historicaldata, a 1% change in unemployment may change delinquency rate by 1.12%. However, this is just an estimationand does not reflect the actual impact.Table 5Sensitivity Analysis For Hypothetical Brazilian ABS Personal Loan TransactionAbsolute change in unemployment (%) 0 1 2 3 4 5Estimated loss (%) 5.00 6.10 7.20 8.40 9.50 10.60Rating impact brAAA brAA+ brAA brA+ brA- brBBB+We also created a hypothetical Brazilian ABS transaction backed by auto loans (see table 6 for characteristics).Table 6Hypothetical Brazilian ABS Auto Loan TransactionAssetsOriginal pool balance ($)Current pool balance ($)Initial loss projection (%) 580 million80 millionLiabilities Type Rating Balance ($) Credit support (%)Class A Senior brAAA 57.60 28Class B Junior NR 12.00 13Equity Equity NR 10.40For this particular pool, we evaluated the impact that changes in consumer leverage would have on our projectedlosses, as well as on our rating on the most senior tranche while keeping all other variables the same (includingsecurities target yield, excess spread, expenses, and cash reserves, constant). Table 7 shows the potential lossprojection and rating using the coefficients as listed in Appendix table 4 for the Brazilian market. By using thecoefficient, and since the variables are shown as natural logarithm, we are saying that based on available historicaldata, a 1% change in consumer leverage may change delinquency rate by 1.21%. However, this is just an estimationand does not reflect the actual impact.Table 7Sensitivity Analysis For Hypothetical Brazilian ABS Auto Loan TransactionAbsolute change in consumer leverage (%) 0 1 2 3 4 5Estimated loss 5.00 6.20 7.40 8.60 9.80 11.00www.standardandpoors.com/ratingsdirect 21978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesTable 7Sensitivity Analysis For Hypothetical Brazilian ABS Auto Loan Transaction (cont.)Rating impact brAAA brAA+ brAA- brA brBBB+ brBBBArgentinean ABS transactionWe considered an Argentinean ABS transaction backed by consumer loans with the characteristics listed in table 8.Table 8Hypothetical Argentinean ABS Consumer Loans TransactionAssetsOriginal pool balance ($)Current pool balance ($)Initial loss projection (%) 6118 million118 millionLiabilities Type Rating Balance Credit support (*)Class A Senior raAAA 105 11.00Class B Junior raA 13 0.00Equity Equity NR 17*Credit support does not include soft enhancement of 30% yearly excess spread.For this particular pool, we evaluated the impact that changes in unemployment would have on the projected lossfor the pool, as well as on our rating on the most senior tranche, while keeping all other variables the same(including securities yield, excess spread, expenses, and cash reserves, constant). Table 9 shows the potential lossprojection and rating using the coefficients as listed in Appendix table 6 for the Argentinean market. By using thecoefficient, and since the variables are shown as natural logarithm, we are saying that based on available historicaldata, a 1% change in unemployment may change the delinquency rate by 1.64%. However, this is just an estimationand does not necessarily reflect the actual impact.Table 9Sensitivity Analysis for Hypothetical Argentinean ABS Consumer Loans TransactionAbsolute change inunemployment (%)0 1 2 3 4 5 Worst-case scenario (up to 22%from 6.7%)Estimated loss (%) 6.00 7.60 9.30 10.90 12.50 14.20 31.00Rating impact raAAA raAAA- raAA- raA raBBB+ raBBB raCCCMexican RMBS transactionWe considered a Mexican RMBS transaction that has the following characteristics listed in table 10.Table 10Hypothetical Mexican RMBS TransactionAssetsOriginal pool balance MXN($)Current pool balance MXN ($)Initial loss projection (%) 6115 million110 millionLiabilities Type Rating Balance ($) Credit support (%)Class A Senior mxAAA 100 25Class B Junior mxA 10 5Equity Equity NR 5Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 22978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesFor this particular pool, we evaluated the impact that changes in inflation would have on our projected losses, aswell as on our rating on the most senior tranche, while keeping all other variables the same (including securitiesyield, excess spread, expenses, and cash reserves, constant). Table 11 below shows the potential loss projection andrating using the coefficients as listed in Appendix table 8 for the Mexican market. By using the coefficient, and sincethe variables are shown as natural logarithm, we are saying that based on available historical data, a 1% change ininflation rate may change the delinquency rate by 0.28%. However, this is just an estimation and does not reflectthe actual impact.Table 11Sensitivity Analysis For Hypothetical Mexican RMBS TransactionIncrease in inflation (%) 0 5 15 20 50Estimated loss (%) 8 10 13 14 22Rating impact mxAAA mxAA mxA mxBBB mxBBTransaction-Specific Factors Cannot Be IgnoredIn this report, we focus solely on the historical relationship between certain macroeconomic factors and theperformance of Latin American structured finance collateral. However, various transaction-specific factors are alsoimportant in explaining credit quality and rating trends.The below transaction-specific factors, which are relevant to credit quality, in our view, are not unique to LatinAmerican transactions.• Financial institutions, including banks, that act as counterparties in structured finance transactions could be adetermining factor on a rating on a transaction. A sudden default of a financial institution, without a designatedreplacement, could significantly affect credit quality, and in some cases, could have a somewhat widespread effecton structured finance transactions.• The ratings on some structured finance transactions may be linked to the ratings on the financial institutions thatissue the deals or their dependent entities. Therefore, downgrades of these entities could affect those on structuredfinance transactions.• Ratings on sovereign and structured finance securities have become increasingly linked because factors that affectsovereign rating performance usually affect the collateral performance of most structured finance securities.• The quality of servicers and backup servicers could affect the collateral performance and credit quality ofstructured finance transactions.• The legal and regulatory environment, market and transaction structures, and payment mechanisms that may beregion- or asset class-specific can also result in differences in credit quality and rating stability.The Changing Global Economy And Its Impact On Latin AmericaThe findings from our study indicate that GDP growth, unemployment, salaries and inflation, credit growth, andhousehold leverage have historically been linked with changes in credit quality for Latin American structured financecollateral performance.Uncertainty will characterize the global economy in 2012. Our recent analysis of the world economy suggests thatthe recessionary trends and financial stress in developed countries may have a ripple effect on emerging markets.www.standardandpoors.com/ratingsdirect 23978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesAlthough Latin America has more capacity than before to weather these stresses, we believe the region is stillvulnerable to the adverse global trends, which could lead to further collateral deterioration and more rating actions.Related Criteria And Research• European Structured Finance Scenario And Sensitivity Analysis: The Effects Of The Top Five MacroeconomicFactors, March 14, 2012.• Global Structured Finance Scenario And Sensitivity Analysis: The Effects Of The Top Five MacroeconomicFactors, Nov. 4, 2011.• Understanding Standard & Poor's Rating Definitions, June 3, 2009.• Latin America And The Caribbean: Holding Steady So Far, But Risks Loom, May 22, 2012.AppendixRegression and correlation analysisFor our analysis, we used aggregated 90-plus-day delinquencies (likelihood of defaults) reported by each country'sCentral Banks. We also used other economic data, such as each country's GDP and sector growth rate, theunemployment rate, consumer leverage, credit growth, short-term debt, inflation, and salary growth rate from eachcountry's statistics department.We assessed the correlation between each country delinquency and the different economic variables. We alsoconsidered the correlation coefficients between the economic factors themselves in order to determine the keyvariables that we consider to be the most influential contributor to consumer and mortgage portfolio performance.For a regression model, we considered most relevant variables regardless of any multi-collinearity problem in theregression model. Multi-collinearity results from high correlations among independent variables, such asunemployment and GDP growth, which makes it difficult to determine reliable estimates or regression coefficients.We identified uni-variate and multi-variate coefficients. These coefficients represent the change in default based onthe change in the given variable.Brazilian consumer loan performanceTo analyze performance of Brazilian consumer loans, we focused on personal and auto loan portfolios. Weconducted a separate analysis for both portfolios because auto loans tend to be more sensitive to adverse events thanpersonal loans. We adjusted the 90-plus-day arrears, as reported by the Brazilian Central Bank, by the portfoliogrowth (origination) in the previous three months.Particularly, we estimate the consumer leverage by adding the expected monthly payments on personal and autoloans, credit cards, and overdraft lines of credit. For credit cards, we considered a 25% monthly payment on theoutstanding balance, and for overdraft lines of credit, we considered the interest payment. Although we did notinclude monthly mortgage payments in our calculation, we understand that Brazil's mortgage industry is stillrelatively small, especially when compared with other credit products (such as personal and auto loans) offered toBrazilian households. Therefore, we don't believe adding monthly mortgage payments to our DSR calculation wouldsignificantly affect the result.Based on the pair-wise correlations, which are consistent with our expectations, we observed that the personal loansStandard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 24978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market Variablesdelinquencies are positively correlated with changes in the unemployment rate (0.89) and year-over-year short-termdebt (0.86) (see appendix table 1).We also observed that the auto loans delinquencies are positively correlated with changes in consumer leverage(0.61) and negatively correlated with changes in year-over-year credit growth (-0.85) (see appendix table 2).Appendix table 1Correlation Coefficients Between Economic Variables And Brazilian Personal Loans Delinquency 2000-2001Personal loansDelinquency six-monthlagDelinquency six-monthlag1UnemploymentUnemployment 0.9 1ConsumerleverageConsumer leverage -0.82 -0.93 112-month creditgrowth12-month credit growth 0.28 0.25 -0.27 1Short-termShort-term debt growth 0.87 0.92 -0.94 0.21 1Appendix table 2Correlation Coefficients Between Economic Variables And Brazilian Auto Loans Delinquency 2000-2001Auto loansDelinquency six-monthlagDelinquency six-monthlag1UnemploymentUnemployment -0.56 1ConsumerleverageConsumer leverage 0.61 -0.93 112-month creditgrowth12-month credit growth -0.85 0.25 -0.28 1Short-termShort-term debt growth -0.57 0.92 -0.94 0.21 1Appendix table 3Brazilian Personal Loans Delinquency – Six Months Lagged*- Regression ResultsRegression equation 1R square: 80.79% Coefficient Standard error T-stats P-value (%)Intercept 0.09 0.12 0.76 0.44Unemployment rate 0.47 0.16 2.92 -Short-term debt 0.81 0.19 4.15 -Regression equation 2R square: 72.29% Coefficient Standard error T-stats P-value (%)Intercept -0.35 0.11 -3.04 0Short-term debt 1.21 0.06 18.65 0Regression equation 3R square: 66.14% Coefficient Standard error T-stats P-value (%)Intercept -4.44 0.14 -31.88 0Consumer leverage -1.03 0.08 -13.19 0www.standardandpoors.com/ratingsdirect 25978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesAppendix table 3Brazilian Personal Loans Delinquency – Six Months Lagged*- Regression Results (cont.)Regression equation 4R square: 77.90% Coefficient Standard error T-stats P-value (%)Intercept 0.17 0.13 1.29 0.2Unemployment rate 1.12 0.06 20.23 0Regression equation 5R square: 0.04% Coefficient Standard error T-stats P-value (%)Intercept 0.09 0 27.09 0YoY Credit Growth -0.03 0.01 -2.24 0.03*For the regression analysis, we used the natural logarithm of data adjusted for seasonality. Credit cards and overdraft lines over total consumer debt.Appendix table 4Brazilian Auto Loans Delinquency – Six Months Lagged* - Regression ResultsRegression equation 1R square: 75.70%Coefficient Standard error T-stats P-value (%)Intercept -0.81 0.17 -4.62 0Consumer leverage 1.27 0.1 13.03 0YoY credit growth -0.55 0.05 -10.97 0Regression equation 2R square: 23.47%Coefficient Standard error T-stats P-value (%)Intercept -4.22 0.16 -26.58 0Short-term debt -0.57 0.09 -6.29 0Regression equation 3R square: 42.50%Coefficient Standard error T-stats P-value (%)Intercept -1.07 0.26 -4.03 0Consumer leverage 1.21 0.15 8.11 0Regression equation 4R square: 18.45%Coefficient Standard error T-stats P-value (%)Intercept -4.3 0.21 -20.2 0Unemployment rate 1.01 1.23 0 0Regression equation 5R square: 28.56%Coefficient Standard error T-stats P-value (%)Intercept 0.05 0 46.39 0YoY credit growth -0.02 0 -6.6 0Argentinean consumer loan performanceTo assess performance of Argentinean consumer loans, we focused on aggregated consumer loan portfolios asdefined in the BCRA (Banco Central de la República Argentina) Financial Institution's reports. Our analysis suggestsa historical link between consumer asset performance and the following variables: GDP, unemployment, real salary,and credit growth.Based on the pair-wise correlations, which are consistent with our expectations, we observed that the consumerloans delinquencies are positively correlated with changes in the unemployment rate (0.77), while consumer loansStandard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 26978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market Variablesdelinquencies are negatively correlated with changes in GDP growth (-0.77) (see appendix table 5-6). Although weconsider real salary growth and credit growth as key factors it is important to remark that their results should beconsidered limitative given that series information are only available since Q2 2002, were the recovery trend of theArgentine economy started. As a consequence, the estimated correlation and coefficients are not as representative asthe other listed factors.Appendix table 5Correlation Coefficients Between Economic Variables And Argentinean Aggregated Consumer Loans Delinquency2000-2001Delinquency rate GDP growth Unemployment rate Real salary growth YoY credit growthDelinquency rate 1GDP growth -0.77 1Unemployment rate 0.77 -0.97 1Real Salary growth -0.84 0.98 -0.93 1YoY credit growth -0.84 0.98 -0.9 0.97 1Appendix table 6Argentina Aggregated Consumer Loans Delinquency* - Regression ResultsRegression equation 1R square: 96.69%Coefficient Standard error T-stats P-value (%)Intercept 108.92 14.15 7.7 0GDP Growth (10.24) 1.26 (8.12) 0Unemployment rate 0.77 0.31 2.46 0.02Real Salary Growth 2.05 0.27 7.5 0YoY Credit Growth 0.75 0.28 2.7 0.01Regression equation 2R square: 59.12%Coefficient Standard error T-stats P-value (%)Intercept 37.14 4.33 8.58 0GDP Growth (2.78) 0.34 (8.16) 0Regression equation 3R square: 59.81%Coefficient Standard error T-stats P-value (%)Intercept -2.14 0.49 -4.4 0Unemployment rate 1.64 0.2 8.27 0Regression equation 4R square: 69.73%Coefficient Standard error T-stats P-value (%)Intercept 10.08 0.87 11.63 0Real salary growth -1.11 0.12 -9.48 0Regression equation 5R square: 70.00%Coefficient Standard error T-stats P-value (%)Intercept 13.2 1.26 10.46 0YoY credit growth -1.36 0.15 -9.04 0*For the regression analysis we used the natural logarithm of data. Available data series started after the 2001 crisis so results are limitative as they were only testedunder favorable scenarios.www.standardandpoors.com/ratingsdirect 27978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesMexican mortgage loan performanceTo analyze performance of the Mexican mortgage market, we focused on aggregated mortgage portfolio as definedby the Central Bank, Banco de México, reports. Our analysis suggests a historical link between the assetperformance and the following variables: inflation, real GDP, performance on secondary (defined as the sector of thetransformation of raw materials) and tertiary (mainly services industry) sectors, and credit origination growth.Based on the pair-wise correlations we observed that the Mexican residential mortgages loans delinquencies arepositively correlated with changes in the inflation (0.71), while delinquencies are negatively correlated with changesin GDP growth through its proxies, secondary, (-0.88), and tertiary sector growth (-0.89), and credit growth (-0.48)(see appendix tables 7-8).Appendix table 7Correlation Coefficients Between Economic Variables And Mexican Mortgage Loans Delinquency 2000-2001Delinquency rate Secondary sector growth Tertiary sector growth Inflation Credit growthDelinquency rate 1Secondary -0.88 1Sector growthTertiary -0.89 0.94 1Sector growthInflation 0.71 -0.57 0.6 1Credit growth -0.48 0.67 0.75 -0.07 1Appendix table 8Mexican Residential Mortgages Loans Delinquency* - Regression ResultsRegression equation 1R square: 80.95%Coefficient Standard error T-stats P-value (%)Intercept 29.17 3.27 8.92 0Inflation 0.76 0.21 3.59 0Real GDP -7.39 0.81 -0.06 0Regression equation 2R square: 79.14%Coefficient Standard error T-stats P-value (%)Intercept 41.43 3.4 1.19 0YoY Credit Growth 0.62 0.22 2.82 0Real GDP -11.25 1.07 -10.51 0Regression equation 3R square: 84.48%Coefficient Standard error T-stats P-value (%)Intercept 19.87 1.15 17.29 0Secondary sector growth -9.65 0.54 -17.67 0Inflation 0.91 0.1 7.31 0Regression equation 4R square: 87.42%Coefficient Standard error T-stats P-value (%)Intercept 16.29 0.51 31.95 0Tertiary sector growth -10.16 0.37 -26.87 0YoY credit growth 0.98 0.1 9.66 0Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 28978790 | 300129047

Latin American Structured Finance Scenario And Sensitivity Analysis: The Effects Of Regional Market VariablesAppendix table 8Mexican Residential Mortgages Loans Delinquency* - RegressionResults (cont.)Regression equation 5R square: 68.79%Coefficient Standard error T-stats P-value (%)Intercept 4.81 0.57 8.37 0Inflation 1.67 0.11 14.39 0YoY credit growth -0.96 0.1 -9.19 0Regression equation 6R square: 50.11%Coefficient Standard error T-stats P-value (%)Intercept 0.46 0.02 22.31 0Inflation 0.28 0.02 11.94 0Regression equation 7R square: 75.36%Coefficient Standard error T-stats P-value (%)Intercept 35.96 3 11.99 0Real GDP -8.34 766859 -8.94 0Regression equation 8R square: 78.60%Coefficient Standard error T-stats P-value (%)Intercept 2.08 0 836.49 0Secondary sector growth -0.06 0 -22.84 0Regression equation 9R square: 79.08%Coefficient Standard error T-stats P-value (%)Intercept 2.13 0 534.32 0Tertiary sector growth -0.1 0 -23.17 0Regression equation 10R square: 22.97%Coefficient Standard error T-stats P-value (%)Intercept 5.52 0.03 193.61 0YoY credit growth -0.21 0.03 -6.5 0*For the regression analysis we used the natural logarithm of data.www.standardandpoors.com/ratingsdirect 29978790 | 300129047

Copyright © 2012 by Standard & Poor's Financial Services LLC. All rights reserved.No content (including ratings, credit-related analyses and data, model, software or other application or output therefrom) or any part thereof (Content) may be modified,reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval system, without the prior written permission of Standard & Poor'sFinancial Services LLC or its affiliates (collectively, S&P). The Content shall not be used for any unlawful or unauthorized purposes. S&P and any third-party providers, as wellas their directors, officers, shareholders, employees or agents (collectively S&P Parties) do not guarantee the accuracy, completeness, timeliness or availability of theContent. S&P Parties are not responsible for any errors or omissions (negligent or otherwise), regardless of the cause, for the results obtained from the use of the Content, orfor the security or maintenance of any data input by the user. The Content is provided on an "as is" basis. S&P PARTIES DISCLAIM ANY AND ALL EXPRESS OR IMPLIEDWARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE OR USE, FREEDOM FROM BUGS,SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT'S FUNCTIONING WILL BE UNINTERRUPTED, OR THAT THE CONTENT WILL OPERATE WITH ANY SOFTWARE ORHARDWARE CONFIGURATION. In no event shall S&P Parties be liable to any party for any direct, indirect, incidental, exemplary, compensatory, punitive, special orconsequential damages, costs, expenses, legal fees, or losses (including, without limitation, lost income or lost profits and opportunity costs or losses caused by negligence)in connection with any use of the Content even if advised of the possibility of such damages.Credit-related and other analyses, including ratings, and statements in the Content are statements of opinion as of the date they are expressed and not statements of fact.S&P's opinions, analyses, and rating acknowledgment decisions (described below) are not recommendations to purchase, hold, or sell any securities or to make anyinvestment decisions, and do not address the suitability of any security. S&P assumes no obligation to update the Content following publication in any form or format. TheContent should not be relied on and is not a substitute for the skill, judgment and experience of the user, its management, employees, advisors and/or clients when makinginvestment and other business decisions. S&P does not act as a fiduciary or an investment advisor except where registered as such. While S&P has obtained information fromsources it believes to be reliable, S&P does not perform an audit and undertakes no duty of due diligence or independent verification of any information it receives.To the extent that regulatory authorities allow a rating agency to acknowledge in one jurisdiction a rating issued in another jurisdiction for certain regulatory purposes, S&Preserves the right to assign, withdraw, or suspend such acknowledgement at any time and in its sole discretion. S&P Parties disclaim any duty whatsoever arising out of theassignment, withdrawal, or suspension of an acknowledgment as well as any liability for any damage alleged to have been suffered on account thereof.S&P keeps certain activities of its business units separate from each other in order to preserve the independence and objectivity of their respective activities. As a result,certain business units of S&P may have information that is not available to other S&P business units. S&P has established policies and procedures to maintain theconfidentiality of certain nonpublic information received in connection with each analytical process.S&P may receive compensation for its ratings and certain analyses, normally from issuers or underwriters of securities or from obligors. S&P reserves the right to disseminateits opinions and analyses. S&P's public ratings and analyses are made available on its Web sites, www.standardandpoors.com (free of charge), and www.ratingsdirect.comand www.globalcreditportal.com (subscription), and may be distributed through other means, including via S&P publications and third-party redistributors. Additionalinformation about our ratings fees is available at www.standardandpoors.com/usratingsfees.Standard & Poors | RatingsDirect on the Global Credit Portal | June 21, 2012 30978790 | 300129047