تاثیر بحران مالی سال 2001 و پاسخ های سیاست اقتصادی در بازار وام مسکن آرژانتین

Following the 2001 financial crisis, the government of Argentina instituted economic policies to soften the adverse impact of the crisis on the economy. In this paper, we use loan-level data to empirically assess the impact of the currency devaluation and the economic response policies on prepayment and default patterns of residential mortgages in Argentina. On the one hand, our results reveal a significant higher prepayment rate of borrowers who are relatively wealthy or have a US$-denominated mortgage. On the other hand, we observe a significantly higher default rate of borrowers who are less wealthy or have Peso-denominated mortgage.

During the 1990’s, Argentina experienced an economic boom.1 Most observers attribute the Argentine success to the currency peg, control of the inflation rate and the broad market-oriented economic reforms (e.g., privatization, economic liberalization, and deregulation). The reforms encouraged the development of credit markets that facilitated the allocation of capital to the private sector. Capital flow into the residential housing market was especially significant—the unprecedented 16.2% average annual growth rate of mortgage credit subsequently led to the creation of a secondary market in 1996.2 By 1998, the Argentine economy showed early signs of contraction, partly afflicted by the economic turmoil and currency devaluation of neighboring Brazil. By November 2001, the Argentine economic recession evolved into a full-fledged financial crisis.3 The government defaulted on its debt, repealed the convertibility regime of the Peso and devalued its currency to 1.4 Peso per US$.4 To soften the impact of the crisis and to spread its burden between a large spectrum of economic actors, the Argentine government responded in January 2002 with a series of economic policies that significantly affected the financial situation of many households. The government converted US$-denominated bank loans into Argentine Pesos (ARP) at parity. At the same time, in a move designed to limit the loss incurred by savers but that had paradoxical effects, the government converted US$-denominated deposits into Pesos at the official exchange rate of 1.4 Pesos per US$. Furthermore, the government suspended foreclosure proceedings on defaulted mortgage loans. On the other hand, to alleviate the adverse impact of pesofication on creditors, the government first (January 2002) indexed all retail loans to the consumer price index and subsequently (November 2002) to the wage-linked index due to the substantial price surge caused by the currency devaluation and high interest rates. 5 Finally, the government had to float the Peso in January 2002. In this paper, we use a loan-level dataset to study the prepayment and default behavior of the Argentine mortgage borrowers in response to drastic changes in their household portfolio, triggered by the real life economic and financial stress scenario and the subsequent government policy responses. Specifically, we study the prepayment and default decisions of mortgage borrowers in response to the crisis (November 2001), the depreciation of the Peso (beginning January 2002), the indexation of the mortgage loans to the CPI (January 2002–September 2002), and to wage-linked index (November 2002–March 2004). The data used in our paper is unique in several respects. Our loan-level dataset is made up of individual mortgage loans that were originated between June 1998 and March 2000, with performance observed until March 2004. In addition to the traditional information captured at loan origination, we also observe if the loan was originated in US dollars (US$) or Argentine Pesos (ARP). The dataset also contains a very rich set of demographic variables (income, wealth, age, occupation, marital status, etc.). Using mortgage debt to assess the substitutability between the prepayment and default options is ideal because mortgage debt is much more sensitive to shocks since it has a longer maturity than other consumer debts. Moreover, following the 2001 financial crisis the Argentine mortgage borrowers faced wealth maximization incentives that were beyond the call and put options. The surge in Argentine mortgage prepayments and defaults that occurred during the post-crisis period simply cannot be explained by the factors shaping the traditional prepayment and default model as well as the actual prepayment and default behaviors in the US. We expect that the impact of the currency devaluation and the economic response policies on the prepayment and default decisions of the US$-denominated mortgage borrowers will differ from those of the Peso-denominated mortgage borrowers. Furthermore, the prepayment and default behaviors of the relatively wealthy/high-income borrowers will likely differ from those of the less-wealthy/low-income borrowers. In particular, the combination of soaring unemployment, declining income and sharp inflation should create a “poverty effect” leading to a surge in mortgage delinquencies/defaults, especially amongst the less wealthy or Peso-denominated mortgage borrowers. On the other hand, the Peso devaluation along with the pesofication as well as the indexation of borrowed capital and the appreciation of residential real estate should create a “wealth effect,” and thus lead to a surge in mortgage prepayment, especially amongst wealthy borrowers or US$-denominated mortgages. To preview our findings, once we control for all the traditional risk factors (LTV, interest rates, credit scores, etc.) predicted by the option theory of mortgage prepayment and default, we find that US$-denominated mortgages on average have a significantly higher prepayment rate and lower default rate in comparison to ARP-denominated mortgages. Moreover, the performance patterns tend to be heterogeneous among US$ and ARP mortgage borrowers along wealth and income characteristics. We find a 9.1% lower default rate and 15.7% higher prepayment rate amongst the wealthy borrowers (of both US$- and ARP-denominated mortgage). Similarly, a 10% rise in income lowered default risks by 9.8% and raised prepayment risks by 10.1%. Focusing on the impact of the financial crisis and the economic policy on the prepayment and default patterns, we find that following the crisis prepayment risks increased 17.8% and default risks rose 8.2% on average. For a 10% depreciation of the Peso, prepayment risks increased 5.2% and default risks increased 14.2%. Subsequent to the crisis, US$-denominated borrowers with wealth and high income were more likely to prepay but less likely to default. In general, the US$ mortgage borrowers’ prepayment rate rose 17.2%, while their default rate declined 3.1%. Next, we find that indexing the mortgage interest rates to the consumer price index (January 2002–October 2002) resulted in a 28.1% rise in the prepayment rate and 20.8% rise in default rate, with a significantly higher prepayment rate amongst those borrowers possessing more wealth. Finally, indexing the mortgage interest rates to the wage-related indices (November 2002–January 2004) resulted in a 14.3% higher prepayment rate and 13.6% higher default rate; these increases are uniform across income groups. These results reveal that subsequent to the wage indexation, the increase in default and prepayment trends have slowdown significantly, which suggests that higher income households may have engaged in rate-differential arbitrage and were able to maximize their wealth by continuing to make mortgage payments since the rise in their wages were higher than the rise in their interest payments. Thus, the mortgage market experienced a slowdown in the prepayment rates. On the other hand, low-income households—with no wealth and no fear of a foreclosure—continued to default on their loans, while those with wealth prepaid their loans. Another possible reason may be that by the late 2003, wages on average started to rise, causing the mortgage rates to climb and affecting borrowers’ ability to make monthly payments. These results highlight the mechanism and the heterogeneity of consumer response to the financial crisis and policy responses. They also raise serious concerns about the adverse impact of economic policy responses on the less-wealthy households. In particular, we find that instead of alleviating the financial burdens of all consumers, the Argentine government’s policy responses (pesofication, indexation, etc.) inadvertently discriminated against the relatively lower-income borrowers. On the one hand, the wealthy mortgage borrowers received a positive wealth windfall, which was in turn used to prepay their mortgage debt. On the other hand, the low-income mortgage borrowers, instead of receiving the same positive wealth benefit, endured “poverty effects” and defaulted on their loans. These outcomes reduced the ability and willingness of financial institutions to supply new credit. In turn, the viability and stability of the newly functioning mortgage-backed securities market in Argentina was weaken, which eventually led to a temporary shutdown of the housing market, with a negative multiplier effect for many economic activities and employment. The remainder of the paper is structured as follows. Section 2 provides a discussion of the Argentine financial crisis and the economic policy response. Section 3 discusses the performance of the Argentine mortgage market before and after the financial crisis. Section 4 reviews the mortgage prepayment and default literature and the econometric methodology. Section 5 describes the data and presents some descriptive statistics. Section 6 reports the findings for the default and prepayment behavior for US$- and ARP-denominated mortgages. Section 7 discusses the implications of the results on the MBS market in Argentina and the broader economy. Section 8 offers concluding remarks.

In December 2001, Argentina defaulted on her external debt and devalued her currency. In the subsequent months, the Argentine government pursued economic polices intended to soften the adverse impact of the Peso devaluation on the financial burdens of consumers and financial institutions. Specifically, the government converted all US dollar (US$) denominated debt into Argentine Pesos (ARP) at the pre-devaluation exchange rate of ARP1.00:US$1.00 and converted all US$ denominated deposits into ARP at the post-devaluation exchange rate of ARP1.40:US$1.00. Finally, the government announced a New Bankruptcy Law protecting debtors’ (over creditors’) rights - the law suspended the judicial foreclosure on all consumer or commercial loans. To shield the banks, the government indexed the mortgage interest rates, first to the CPI and then to the wage index. In this paper, we study the impact of both the currency devaluation and the subsequent economic policy responses on the Argentine consumers’ decision to prepay or default on their residential mortgage loan. We accomplish this with the help of a loan level dataset of US$- and ARP-denominated residential mortgages (originated between 1998 and 2000). In the absence of economic policies, we expect the default rates for both US$- and ARP-denominated mortgages to rise significantly. However, our empirical results show a sharp jump in the prepayment speed of US$-denominated mortgages and a sharp rise in the default rate for ARP mortgages following the crisis, which cannot be explained by factors shaping the traditional prepayment and default model or the traditional unobserved borrower characteristics. These results can only be explained by drastic changes in the household portfolio, triggered by real life financial and economic policies. The “pesofication” and the CPI indexation policies provided a one-time opportunity to wealthy and/or US$-denominated mortgage borrowers to prepay their loans. The suspension of the foreclosure laws, the 23% rise in unemployment, and a 40% drop in income provided an opportunity to the low income and/or ARP mortgages borrowers to default on their loans. Furthermore, the policies had significant impacts on the availability and price of credit and the functioning of the secondary mortgage market. The price of credit jumped by 5% points, while the availability of credit dropped by 50%. Finally, the secondary mortgage market also collapsed. The lessons learnt from the financial crisis and the subsequent economic policy decisions by the Argentine government can be summarized as follows. The policy decisions aimed at spreading the cost of the crisis amongst the whole spectrum of stakeholders, with an emphasis on protecting individual debtors, was in general well intentioned. However, in doing so, some fundamental mechanisms on which mortgage lending relied on were deeply affected: trust in the future value of financial instruments, effectiveness of security rights, creditworthiness of the government, intangibility of contracts, etc. Moreover, the policies resulted in disproportionately benefiting the higher income and wealthy individuals, while disadvantaged the lower income and unskilled consumers.