Abstract
This paper reviews policy tools that have been used and/or are available for policy makers in the region to lean against the wind and review relevant country experiences using them. The instruments examined include: (i) capital requirements, dynamic provisioning, and leverage ratios; (ii) liquidity requirements; (iii) debt-to-income ratios; (iv) loan-to-value ratios; (v) reserve requirements on bank liabilities (deposits and nondeposits); (vi) instruments to manage and limit systemic foreign exchange risk; and, finally, (vii) reserve requirements or taxes on capital inflows. Although the instruments analyzed are mainly microprudential in nature, appropriately calibrated over the financial cycle they may serve for macroprudential purposes.
VII. Macroprudential Instruments to Manage Foreign-Exchange Credit Risk
A. Motivation
Foreign-exchange (FX) credit risk can pose significant systemic risk to the financial system and the economy as a whole. FX credit risk is the risk faced by financial institutions associated with lending to unhedged borrowers that carry currency mismatches in their balance sheets. Experience in EMs—more recently in Europe—has shown that currency mismatch has been a prime vehicle for agents to take on credit risk, and has resulted in large exposures to systemic risk for the economy as a whole. In an environment of economic bonanza and rapid credit growth, banks may not be appropriately pricing or provisioning for the indirect exposures to FX risk. Banks’ currency risk from FX liabilities can be largely passed on to borrowers via FX loans; which could backfire if a large FX depreciation were to increase the debt burden of borrowers and lead to defaults.2
Although currency mismatches may be limited in the banking sector balance sheet; FX lending pose important credit risks that are difficult to assess. FX credit risks are very difficult to calibrate, and for financial institutions to internalize. The credit risk are associated not only with the direct exposure (when the bank extends FX-denominated credit to unhedged borrowers), but also with an indirect exposure (the bank extends domestic-currency lending to a borrower who is already exposed to FX risks). Lack of information about the corporates and households’ structure of debt obligations (currency and maturity) and financial exposures, including to derivatives, hinders also the effective assessment of FX credit risk. The most vulnerable sector may be the household sector; although the presence of FX deposits in dollarized economies may mitigate this risk to a certain extent.
Countries have implemented prudential measures to make financial institutions internalize FX credit risks associated with lending to un-hedged borrowing. Measures to limit FX lending in Emerging Europe before the global financial crisis through reserve requirements on FX liabilities and limits to FX exposure to banks or limits on FX lending as a percent of capital were not very effective as they were circumvented under easy external financing conditions. However, those measures were not specifically designed to deal with FX credit risks; as banks did not internalize the risks of lending in foreign exchange to unhedged borrowers. Recent measures focus on additional provisioning and capital requirements; but it is too early to assess their effectiveness. As a first step, authorities may request financial institutions to establish a system to identify and monitor FX risks, as well as internal procedures to manage such risks, and take corrective measures as required. But there is also a need to enhance supervision on this front, to ensure that financial institutions are accordingly evaluating the FX credit risks.
B. Instruments and Potential Benefits
Measures to limit FX credit risk enhance financial stability by ultimately reducing potential losses associated with large currency depreciation at time of financial stress. The measures help banks internalize FX risks associated to lending to un-hedged borrowers. Country experiences showed a focus on direct exposures; with no attempt to incorporate indirect credit risks (Annex I). The measures could also serve as a counter-cyclical tool if they help curb FX lending (at least) in the short run. The measures make FX lending more costly for financial institutions, and can act as a deterrent to extend FX loans.
Regarding quantitative measures, debt to income limits on borrowers may be more effective than bank credit limits as a percent of capital. In Romania; limits on domestic FX lending as a percent of capital had some temporary effects, but effects faded away as financial institutions found ways to raise capital at times of easy external financial conditions. Significant risks to financial disintermediation also rose as households and corporate search for alternative sources of funding. Limits on debt-to-income for individual loans extended in foreign currency could be more effective, and sometimes result in an effective banning of FX lending through specific instruments (i.e. mortgages). These limits are more stringent that those impose for lending extended in domestic currency, which somehow assess impairments on borrowers’ ability to service their debt under an implicit level of depreciation (see the case of Uruguay).
Additional capital or provisioning requirements seem more adequate to account for the FX credit risks. While the difference between provisioning and capital requirements may not be so clear, in line with general wisdom, one could consider a preference for provisioning to account for expected losses derived from FX credit risks—for example, by taking into account the identified direct FX exposure extended through credit. Additional capital requirements could then account for unexpected losses; from unidentified FX credit risk because it is unknown the extent of FX exposure to specific clients, or because of indirect exposures. Two countries (Peru and Uruguay) explicitly request additional capital requirements on FX credit risk. Peru establishes a capital add-on taking into account the FX exposure; while Uruguay establishes a higher risk weight for loans extended in foreign currency to un-hedged borrowers. The latter may be a more transparent way to establish additional capital requirements, because of the difficulties in calibrating the FX exposure. To be effective, measures need to result in a higher cost associated to FX lending to financial institutions.
As a first step, country authorities need to do more to compile detailed information about the financing structure of the corporate and household sectors in a systematic way, with emphasis on FX mismatches and exposure to complex financial instruments like derivatives. To properly internalize FX credit risks, information is required on FX position, FX sales and percentage of short-term liabilities in foreign currency from the borrowers, which may prove challenging. Lack of information about the corporates and households’ structure of debt obligations (currency and maturity) and financial exposures, including to derivatives, hinders also the effective assessment of FX credit risk; and exacerbated the impact of the global financial crisis. Limitations on the latter were evidenced in Brazil and Mexico in 2008, when large corporate losses materialized due to FX operations in the FX derivative market (see Annex I). In parallel, measures to encourage the adequate use of FX hedging among the household and corporate sectors would be welcome.
And enhanced supervision is required. Supervisory authorities need to fully understand/monitor how financial institutions identify FX credit risks; and stand ready to adapt regulation. In this regard, some homogenous framework across the financial system would be welcome, so that there is some benchmarking in the evaluation of FX credit risks across financial institutions risk evaluation systems.
Annex 1. Country Experiences in Managing FX Credit Risk
This annex surveys the experiences with specific prudential measures managing FX credit risks. Measure have not been widely use despite being a considerable threat to financial stability. The main challenge is to ensure that financial institutions have in place internal mechanisms to adequately qualify, define, and manage credit risks associated with lending to unhedged borrowers. In parallel, close monitoring and supervision would be crucial to ensure financial institutions internalize correctly FX credit risk. In parallel, measures to encourage the adequate use of FX hedging among the household and corporate sectors would be welcome.
As a first step, country authorities need to do more to compile detailed information about the financing structure of the corporate and household sectors, with emphasis on FX mismatches and exposure to complex financial instruments like derivatives. To properly internalize FX credit risks, information is required on FX position, FX sales and percentage of short-term liabilities in foreign currency from the borrowers, which may prove challenging. At times of easy external financing conditions, corporates and households tend to increase indebtness, in many cases raising FX mismatches because of more attractive financial terms.
A. Peru
Peru continues to be a highly dollarized economy, but has successfully pursued market-driven financial de-dollarization during the last decade. This has been possible thanks to macroeconomic stability, prudential policies to better reflect currency risks, and the development of the capital market in soles. However, as of October 2010, deposit and credit dollarization remains high at 47 percent and 44 percent, respectively.
Less than 20 percent of total credit is potentially exposed to FX credit risk, as of June 2010. According to the Superintendency of Banks (SBS), only 1½ percent of total credit is unidentified in terms of exposure to FX credit risks. Of the total credit outstanding, 18 percent is identified as exposed to FX credit risk. Under this category, mortgages report highest level, although the stock is very low and would not pose systemic risk at this juncture. Table 1 summarizes identified FX credit risk exposure by type of credit, as a percent of that sub-category total. The most striking feature is the sharp increase in SMEs lending in FX after the global financial crisis, likely due to more attractive financial terms.
Peru: FX Credit Exposure of Credit extended in Foreign Currency-Clients Classified as Normal
(In percent)
Peru: FX Credit Exposure of Credit extended in Foreign Currency-Clients Classified as Normal
(In percent)
| Banks | Commercial | Consumer | Mortgages | SMEs | Total |
|---|---|---|---|---|---|
| Dec-06 | 28.53 | 49.74 | 42.1 | 23.79 | 31.89 |
| Dec-07 | 26.14 | 56.55 | 59.28 | 44.08 | 32.44 |
| Dec -08 | 27.23 | 64.24 | 60.66 | 45.3 | 33.12 |
| Dec-09 | 29.14 | 65.87 | 60.11 | 45.35 | 34.49 |
| Jun-10 | 27.81 | 53.75 | 46.19 | 49.67 | 31.19 |
Peru: FX Credit Exposure of Credit extended in Foreign Currency-Clients Classified as Normal
(In percent)
| Banks | Commercial | Consumer | Mortgages | SMEs | Total |
|---|---|---|---|---|---|
| Dec-06 | 28.53 | 49.74 | 42.1 | 23.79 | 31.89 |
| Dec-07 | 26.14 | 56.55 | 59.28 | 44.08 | 32.44 |
| Dec -08 | 27.23 | 64.24 | 60.66 | 45.3 | 33.12 |
| Dec-09 | 29.14 | 65.87 | 60.11 | 45.35 | 34.49 |
| Jun-10 | 27.81 | 53.75 | 46.19 | 49.67 | 31.19 |
Financial institutions are asked to put in place internal mechanisms to qualify, define and monitor direct credit in foreign currency; including (i) identification of exposed and non-exposed clients to FX credit risk; (ii) requirements for extending credit in foreign currency and for excluding credit operations with associated FX risk; (iii) stress testing (at least, , two scenarios that embed real depreciation of, at least, 10 and 20 percent respectively); and (iv) corrective actions over changes in credit qualification or credit conditions. The bank’s Board must be informed about FX credit risk at least bi-annually with a summary of the aggregate FX credit risk exposure; of the potential losses (by type of credit) and an evaluation of the internal procedures that identify FX credit risks.
Some benchmarking for the identification of the FX credit risk may be welcome, as well as enhanced supervision in the context of a still dollarized economy. At the moment, the evaluation of FX exposure through credit is done according to the methodology of each financial institution. Some institutions take into account cash flows; others the capacity to pay under a certain exchange rate shock (10 percent, for example), others the income of the borrower to assess the capacity to service debt granted in foreign exchange. A more standardized approach would be welcome and will facilitate the supervision of FX credit risks in the financial system.
In terms of bank regulation, the Superintendence of Banks has taken two venues to incorporate explicitly FX credit risk exposures:
• Provisioning for FX credit risk, effective since 2006, applies to direct credit and financial leasing operations, except those with automatic guarantees. The provision applies to loans classified as normal, and is in addition to the general provisions. The provision requirements are: (i) 0.25 percent for credit operations covered with guarantees of rapid execution; (ii) 0.5 percent for FX credit with preferred guarantees; and (iii) 1 percent for the rest of FX credit. Financial institutions are, however, exempted from provisioning if minimum requirements and risk assessment practices (previous paragraph) are fulfilled. Provisioning associated with FX credit risks amount currently to about S/. 4 million (about 0.001 percent of GDP).
• An additional capital requirement, since July 2010. If possible, the financial institution incorporates the FX risk assessment in the overall credit risk assessment (through their internal rating). If not, the financial institution includes an additional capital requirement of 2.5 percent of total FX exposure. Of the two alternatives, financial institutions are currently applying the 2.5 percent capital add-on, as all of them are using the standard methodology for the identification of credit risks.3 According to SBS, the capital add-on amounts to about 1 percentage point of total capital requirements of the financial system.
B. Uruguay
As Peru, Uruguay continues to be a highly dollarized economy, despite successfully pursued market-driven financial de-dollarization during the last decade. Credit and deposit dollarization reached about 52 percent and 76 percent respectively, as of 2010Q3.4
In order to capture FX credit risk, capital requirements for credit/market risks on FX loans carry a differentiated risk weight. Loans to un-hedged borrowers carry a 125 percent risk weight (rather than 100 percent) in the calculation of the CAR.
Additionally, provisioning for loan losses is higher for FX loans, irrespective of whether the borrower is hedged or unhedged. In Uruguay, provisioning is required not only when the loan is past due, but also when a borrower show signs of difficulties to pay in the short/medium term. In practice,
• For commercial loans, banks need to assess the borrower ability to pay in case of a peso depreciation of 20 percent and 60 percent, therefore assessing FX credit risk. The bank should look at the borrowers fund flows and assess whether a devaluation of the peso would or not impact in the borrower’s ability to stick to the committed payments without restructuring their debt.
–In case a borrower ability were not substantially altered by peso depreciation of 60 percent, it could be classified as normal (provisioning of 0.5 percent).
–If the borrower could continue paying after a 20 percent devaluation (but not after a 60 percent devaluation), the loan requires a provision of 3 percent.
–And if the borrower could not pay without debt restructuring after a devaluation of 20 percent the loan, provisioning is raised to 7 percent.
• For consumer loans. If loans are granted in pesos, a nondelinquent loan should be classified as normal if the monthly projected payments do not exceed 30 percent of the borrower’s income. In case of exceeding that amount, the loan requires a loss provision of 20 percent. If a loan is granted in foreign currency the threshold on borrower’s income becomes 15 percent; which hides an implied depreciation of 50 percent for the peso.
The loan provisions may not aim to value loans perfectly but to act as a deterrent to FX credit risk. The penalization that consumer credit gets (20 percent loan loss provisions versus 3–7 percent in case of commercial credit) has had a larger effect than in the case of commercial loans (virtually banning mortgages in dollars to people with income in pesos, as 15 percent of income is not enough to pay for a house in a reasonable amount of time). The measures results in a virtual banning of credit in dollars to “nontradable sectors” in the case of commercial credit.
C. Romania5
During 2005, the authorities put into effect a series of prudential measures aimed at reducing the currency-mismatch risk associated with excessive foreign-currency lending. Although the measures focused on limiting banks’ foreign-currency exposure to unhedged borrowers and increasing the coverage and level of required reserves on foreign-currency liabilities; the authorities imposed a requirement limiting credit institution’ overall FX lending to un-hedged borrowers to less than 300 percent of banks’ own funds. The regulation was binding for 13 out of 39 banks at time of implementation.6 The authorities also tightened loan classification norms for credit institutions, explicitly requiring banks to consider FX risk when classifying their loans to individuals. The new regulation required banks to downgrade the classification of unhedged borrowers, regardless of their financial position or collateral. The latter measures resulted in an immediate increase in NPLs, from 8.1 percent at end-2004 to 9.4 percent in September 2005; forcing banks to increase provisions.
Despite a significant shift in the currency composition of credit growth in the short term, the effectiveness faded down over time. The y/y growth rate in FX credit fell from 56 percent in September 2005 to 30 percent in February 2006 and there was a dramatic shift away from FX loans in favor of local-currency lending, markedly for consumer lending. The 3-month FX credit flow went from a peak of 5 percent of GDP in August 2005 to 1.7 percent by end-December. Local-currency credit flows increased from 3.7 percent of GDP to 7.0 percent over the same period. Overall, credit declined from 9 percent of GDP to 5.3 percent in December.
And demand for credit remained strong, as lenders found a continued incentive to seek alternative channels of funding. Large corporate borrowers borrowed directly from foreign banks. Corporate access to foreign credit also contributed to the boom, rising from a net of 4 percent of GDP in 2005 to nearly 11 percent in 2007. The 300 percent capital binding rule resulted in some banks, especially foreign, increasing capital to resume FX lending. Overall 2005–08, the share of lending to households in FX rose from 44 to 59 percent, while the share of lending to nonfinancial firms in FX declined slightly, from 59 to 57 percent.
The major source of risk to the banking system stands through FX credit risks. The direct exposure of banks to FX risk through their net open positions is low, as their foreign currency borrowings are almost entirely offset through FX lending to households and nonfinancial firms.7 However, both household and corporate sector balance sheets face significant exposure to movements in the euro exchange rate and interest rates on euro loans. Banking sector’s vulnerability to exchange rate risk is greater than their lending in FX suggests. The exposure of the corporate sector to currency risk is greater than their local borrowing suggests; as large firms borrowed directly from abroad—the equivalent of nearly 11 percent in 2007.
D. Brazil and Mexico: “Playing” with Financial Derivatives8
Low currency volatility and the nominal appreciation trend observed in emerging countries before August 2008 led some corporations to increase their off-balance sheet foreign exchange exposure through derivative positions. As a consequence, a number of companies in Brazil and Mexico started betting against the depreciation of their currencies by selling foreign exchange options in the offshore market. These contracts allowed corporates to sell U.S. dollars at a favorable rate when the exchange rate rose above a “knock-out” price (i.e., the domestic currency appreciates), but forced them to sell dollars at an unfavorable rate if the exchange rate fell below a “knock-in” price (the domestic currency depreciates). The operation offered financing and currency trades at favorable rates, but with the drawback of having to deliver dollars at a loss if the domestic currency depreciated past a certain threshold.
The sharp currency depreciation observed in Latin America after mid-September 2008 resulted in elevated systemic risks. Large losses for some of the top companies in Brazil and Mexico materialized when the exchange rate triggered the “knock-in” provision, forcing them to sell double the amount of U.S. currency at the higher price.9 One month after the Lehman Brothers default, in Mexico and Brazil the currency depreciated by more than 30 percent. In Mexico, derivatives losses reached US$4 billion in the fourth quarter of 2008, while in Brazil; losses were as high as US$25 billion.10 The losses resulted in a strong elevation in the volatility and depreciation of the national currencies; and started to constitute a systemic credit risk because the companies could fail to pay to the banks. It also added stress to the strong restriction of liquidity in interbank operations and accentuated reduction of credit to productive firms in emerging market economies.
The complexity of such deals and the fact that they were done privately highlights the lack of transparency in these markets, as many of these companies did not disclose any information on their derivative positions. One result was a review of derivatives exposures across the region as policymakers realized that these exposures could pose systemic risk. Looking forward, policymakers would need to balance financial stability against market development in considering possible regulation of corporate derivatives risk. In Colombia, for example, the central bank established in May 2007 a maximum leverage position on forwards over the financial entities’ net worth, a measure that was widely criticized but later proved to reduce the impact of the crisis. In some cases, however, corporate derivatives have contributed to reducing financial vulnerabilities, as shown by the use of oil price hedge and currency swaps by the Mexican state-owned petroleum company (Pemex), which helped it to stabilize its 2009 budget.
Despite progress in compiling and disseminating balance sheet information for large enterprises, a systematic compilation on corporate and household balance sheet is crucial given the complex interactions between the financial and corporate sectors. Measures adopted in Brazil and Mexico since 2009 represent a good starting point. In Brazil, all financial institutions must register their exposures via derivative markets. Equally in Mexico, equity, long-term debt or equity issuers must document market, credit and liquidity risks associated with derivative contracts and assess its importance in the financial position and financial results of the company.
Prepared by Mercedes Vera Martin. The note has benefited from comments by C. Fernandez, M. Garcia-Escribano, and C. Tovar.
For a discussion on measures to limit FX positions in banks’ balance sheets, see accompanying note “Limiting Foreign Exchange Positions to Contain Systemic Risk,” prepared by C. Fernández-Valdovinos and Chris Walker (WHD).
Ultimately, Peru is moving toward a system in which banks would incorporate the FX credit risk in the overall assessment of credit risk (through higher internal ratings (like currently done in Chile).
Data on deposits excludes nonresident deposits.
The discussion in this section is drawn from “Credit Growth: Development and Prospects” (A. Tiflin, Selected Issues Paper, 2006 Article IV Consultation, IMF/06/169); and Romania’s financial sector stability report (IMF, 2010e); available at www.imf.org/external/pubs/ft/scr/2010/cr1047.pdf.
Additional measures included increases in foreign-currency reserve requirements, also to curb capital inflows.
See IMF (2010e) for a detailed analysis of Romania’s financial sector stability at www.imf.org/external/pubs/ft/scr/2010/cr1047.pdf
This section is drawn from Jara, Moreno and Tovar (2009).
The problem extended also to productive companies in India, China, and Korea. See Farhi and Zanchetta (2009) for details.
A major food retailer (Comercial Mexicana) sought bankruptcy protection in October 2008 with losses up to US$1.1 billion on non-deliverable forward (NDF) contracts it had made with international banks Gruma SA, the world’s largest maker of corn flour, and Alfa SAB, the world’s largest maker of aluminum engine heads and blocks, also suffered from considerable mark to market losses on derivative instruments during this period. In October, glass maker Vitro SAB announced that a large part of its $227 million of derivatives losses had come from natural gas forwards.