At present, only a few of the highly dollarized countries issue explicit regulations to limit banks’ exposure to currency-induced credit risk. Based on a survey of prudential practices in a sample of countries with different degrees of dollarization, there is only limited evidence of specific adaptations to manage credit and interest rate risks (Table 6). Interestingly enough, the less dollarized countries are the ones that have the more explicit regulations addressing dollarization risks. With one exception, no specific prudential measures to limit banks’ indirect exposure to currency risk were reported by the highly dollarized countries, perhaps because the high financial dollarization makes them difficult to implement and excessively costly in terms of financial disintermediation. By contrast, economies with low financial dollarization had various types of limitations in place, including an out-right prohibition to lend to borrowers in the nontradable sector (Chile); quantitative limitations on lending and a prohibition for households to hold foreign currency deposits (Mexico); and an outright prohibition of foreign currency lending other than on-lending of foreign credit in combination with an across-the-board prohibition to hold foreign currency deposits (Brazil).20
Risk Management Arrangements in Dollarized Economies
Risk Management Arrangements in Dollarized Economies
Liquidity Risk | |||||||
---|---|---|---|---|---|---|---|
Country Case |
Credit Risk | Liquidity/reserve buffers | Differential liquidity/ reserve requirements |
Limits on maturity mismatches |
Other | Lender of Last Resort in Foreign Currency |
Foreign Exchange |
Partially dollarized countries | |||||||
Argentina | No specific measures. | Liquidity requirements were lowered from 43 percent to 20 percent during a 16 percent deposit run in 1994. | No. | Limits on gaps from one week up to three months. | Central bank can increase requirements if concentration of liabilities is detected. | Instruments in foreign currency and preapproved credit lines. | Pre-2002, capital requirement for foreign exchange risk based on the per-currency exposure and volatility, except U.S. dollars. |
Armenia | No specific measures. | No. | No. | No. | Reserve requirements on foreign currency deposits to be deposited in domestic currency. | Central bank credit in foreign currency. | Symmetric limit of 25 percent of bank’s capital (5 percent in a less liquid single currency). |
Bolivia | No specific measures. | 10 percent liquidity requirement. Central bank uses foreign currency deposits as collateral for lender of last resort. | No. | No. | Zero percent reserve requirement over 2 years (no early withdrawal). | Central bank foreign currency credit, outright repurchase agreements, and cross-currency repurchase agreements. | Limit of 80 percent of bank capital minus fixed assets on net dollar assets and 20 percent on net liabilities. |
Bulgaria | No specific measures. | No. | No. | No. | Liquidity stress testing. | No. | Limit: 30 percent of capital, except euros. 100 percent risk weight for 2 percent net foreign position. |
Cambodia | No specific measures. | No. | No. | No. | One-month report on maturity gaps. | No. | Limit at 5 percent of the bank’s net worth. |
Peru | No specific measures. | Foreign currency reserve requirements lowered from 45 to 20 percent in 1998 following a creditor run. | Minimum liquidity requirement: 8 percent in domestic currency and 20 percent in foreign currency. | No. | Protocol covering deposit and credit runs. | Central bank foreign currency credit, swaps, and indexed deposit certificates. Protocol to cover runs. | Net dollar assets higher than −2.5 percent and lower than 100 percent of capital. |
Turkey | List of selected borrowers’ types. | No. | Liquidity requirements: 6 percent in domestic currency and 11 percent in foreign currency. | Limit equivalent to bank capital on net assets over 3 years. | 3 percent reserve requirement in foreign currency. | No. | Symmetric overall limit of 20 percent of bank’s capital. |
Uruguay | No specific measures. | No. | No. | No. | Reserve requirements: 10 percent up to 30 days, 4 percent for 30–180 days, and 2 percent over 180 days. | Central bank issues certificates of deposit in dollars. | Symmetric limit of 1.5 times equity. |
Zambia | No specific measures. | No. | No. | No. | Reserve requirements on foreign currency deposits to be partly deposited in domestic currency. | No. | Symmetric overall limit of 15 percent of bank’s capital. Capital requirement for 100 percent of open position. |
Fully dollarized countries | |||||||
Ecuador | No specific measures. | No. | No. | “Liquidity at risk” limit on negative cumulative gap vs. liquid assets. | Maturity gaps required starting in December 2002. | Repurchase agreements with central bank paper. | Value-at-risk minimum capital for foreign exchange risk by December 2003. |
Panama | No specific measures. | No. | No. | No. | 30 percent liquidity requirement, 20 percent if interbank deposits are significant. | The National Bank of Panama can grant credit to other banking institutions. | Minimum capital for foreign exchange exposures set internally by banks. |
Control group: Nondollarized countries | |||||||
Brazil | Foreign currency contracts are null and void, including lending, except on-lending of external loans. | No. | No. | No. | Liquidity stress testing. | No. | Limit of 60 percent of equity. Value-at-risk capital requirement on foreign exchange exposures |
Chile | Until recently, unhedged foreign currency borrowing was prohibited. | No. | Reserve requirements lower in domestic currency (3.6–9 percent vs. 13.6–19 percent in foreign currency). | Limits on 30-day and 90-day gaps. | Short-term liabilities exceeding 2.5 times capital have to be held as cash or central bank reserves. | Central bank foreign currency credit, promissory notes, and swaps. | Limit on the overall position of 20 percent of basic capital, differentiated by country risk. |
Mexico | There are maximum lending limits on the amount that banks may lend to individual borrowers. | Yes, in foreign exchange. | Liquidity coefficient in foreign exchange. | Limit on 60-day gaps. | Zero reserve requirements. | No. | Net assets or liabilities cannot exceed 15 percent of capital. |
Risk Management Arrangements in Dollarized Economies
Liquidity Risk | |||||||
---|---|---|---|---|---|---|---|
Country Case |
Credit Risk | Liquidity/reserve buffers | Differential liquidity/ reserve requirements |
Limits on maturity mismatches |
Other | Lender of Last Resort in Foreign Currency |
Foreign Exchange |
Partially dollarized countries | |||||||
Argentina | No specific measures. | Liquidity requirements were lowered from 43 percent to 20 percent during a 16 percent deposit run in 1994. | No. | Limits on gaps from one week up to three months. | Central bank can increase requirements if concentration of liabilities is detected. | Instruments in foreign currency and preapproved credit lines. | Pre-2002, capital requirement for foreign exchange risk based on the per-currency exposure and volatility, except U.S. dollars. |
Armenia | No specific measures. | No. | No. | No. | Reserve requirements on foreign currency deposits to be deposited in domestic currency. | Central bank credit in foreign currency. | Symmetric limit of 25 percent of bank’s capital (5 percent in a less liquid single currency). |
Bolivia | No specific measures. | 10 percent liquidity requirement. Central bank uses foreign currency deposits as collateral for lender of last resort. | No. | No. | Zero percent reserve requirement over 2 years (no early withdrawal). | Central bank foreign currency credit, outright repurchase agreements, and cross-currency repurchase agreements. | Limit of 80 percent of bank capital minus fixed assets on net dollar assets and 20 percent on net liabilities. |
Bulgaria | No specific measures. | No. | No. | No. | Liquidity stress testing. | No. | Limit: 30 percent of capital, except euros. 100 percent risk weight for 2 percent net foreign position. |
Cambodia | No specific measures. | No. | No. | No. | One-month report on maturity gaps. | No. | Limit at 5 percent of the bank’s net worth. |
Peru | No specific measures. | Foreign currency reserve requirements lowered from 45 to 20 percent in 1998 following a creditor run. | Minimum liquidity requirement: 8 percent in domestic currency and 20 percent in foreign currency. | No. | Protocol covering deposit and credit runs. | Central bank foreign currency credit, swaps, and indexed deposit certificates. Protocol to cover runs. | Net dollar assets higher than −2.5 percent and lower than 100 percent of capital. |
Turkey | List of selected borrowers’ types. | No. | Liquidity requirements: 6 percent in domestic currency and 11 percent in foreign currency. | Limit equivalent to bank capital on net assets over 3 years. | 3 percent reserve requirement in foreign currency. | No. | Symmetric overall limit of 20 percent of bank’s capital. |
Uruguay | No specific measures. | No. | No. | No. | Reserve requirements: 10 percent up to 30 days, 4 percent for 30–180 days, and 2 percent over 180 days. | Central bank issues certificates of deposit in dollars. | Symmetric limit of 1.5 times equity. |
Zambia | No specific measures. | No. | No. | No. | Reserve requirements on foreign currency deposits to be partly deposited in domestic currency. | No. | Symmetric overall limit of 15 percent of bank’s capital. Capital requirement for 100 percent of open position. |
Fully dollarized countries | |||||||
Ecuador | No specific measures. | No. | No. | “Liquidity at risk” limit on negative cumulative gap vs. liquid assets. | Maturity gaps required starting in December 2002. | Repurchase agreements with central bank paper. | Value-at-risk minimum capital for foreign exchange risk by December 2003. |
Panama | No specific measures. | No. | No. | No. | 30 percent liquidity requirement, 20 percent if interbank deposits are significant. | The National Bank of Panama can grant credit to other banking institutions. | Minimum capital for foreign exchange exposures set internally by banks. |
Control group: Nondollarized countries | |||||||
Brazil | Foreign currency contracts are null and void, including lending, except on-lending of external loans. | No. | No. | No. | Liquidity stress testing. | No. | Limit of 60 percent of equity. Value-at-risk capital requirement on foreign exchange exposures |
Chile | Until recently, unhedged foreign currency borrowing was prohibited. | No. | Reserve requirements lower in domestic currency (3.6–9 percent vs. 13.6–19 percent in foreign currency). | Limits on 30-day and 90-day gaps. | Short-term liabilities exceeding 2.5 times capital have to be held as cash or central bank reserves. | Central bank foreign currency credit, promissory notes, and swaps. | Limit on the overall position of 20 percent of basic capital, differentiated by country risk. |
Mexico | There are maximum lending limits on the amount that banks may lend to individual borrowers. | Yes, in foreign exchange. | Liquidity coefficient in foreign exchange. | Limit on 60-day gaps. | Zero reserve requirements. | No. | Net assets or liabilities cannot exceed 15 percent of capital. |
How Are Dollarized Countries Managing Dollarization Risks?
Some adaptations have been made to regulations on open foreign exchange positions. To limit banks’ direct exposure to foreign exchange risk, several highly dollarized countries have broadened the limits on open foreign exchange positions to facilitate dealing with temporary fluctuations in the currency composition of banks’ balance sheets and allowed banks to maintain long dollar positions (to protect banks’ capital-asset ratios from exchange rate fluctuations).21 In addition, some countries have introduced capital requirements against banks’ foreign exchange exposures. In contrast, consistent with the currency board’s commitment to preserve the exchange rate, some currency board countries (Argentina, prior to the crisis, and Bulgaria) have formally excluded from the calculation of banks’ open positions, the positions in the currency to which the exchange rate was pegged.22
While many highly dollarized countries have taken some prudential measures to limit liquidity risk, the extent and nature of the safeguards vary substantially across countries. High liquidity or reserve requirements on foreign currency deposits, held at least in part in the form of liquid dollar assets abroad or deposited at the central bank and backed, explicitly or implicitly, by international reserves, are often used as a liquidity buffer, to be lowered in times of crisis.23 Indeed, gross international reserves, of which about one-third originates from required dollar reserves deposited by banks at the central bank, covered nearly all foreign currency deposits in the Peruvian banking system during 2002. Together with sound fundamentals, this helps explain the stability of foreign exchange deposits in Peru during the recent Argentine crisis (see Figure 1 and Box 4). Reserve or liquidity requirements are often higher on dollar deposits than on local currency deposits, and often differentiated according to the maturity of the deposits, reflecting their differential exposure to liquidity risk. To enhance the effectiveness of these regulations, some countries, such as Bolivia and Uruguay, have complemented them with regulations to discourage early withdrawals.24 Other countries, such as Israel, have introduced minimum maturity requirements on foreign currency deposits in the banking system.25 While less effective in dealing with systemic runs, limits on maturity mismatches have also been introduced by a number of dollarized countries. Less-intrusive regulatory guidelines are generally also provided to orient banks’ liquidity risk management.26 In the case of branches of foreign banks, some countries, such as Paraguay, have imposed the provision of support from their main offices as a precondition for licensing.
Costs and Benefits of Norms to Control Liquidity Risk: The Case of Peru
This box estimates the joint costs and benefits of three prudential regulations aimed at limiting liquidity risk in the Peruvian banking system: (1) reserve requirements; (2) liquidity requirements; and (3) deposit insurance.
Prudential Norms
Reserve requirements: Required reserves are held as vault cash or deposits at the central bank, in the deposit’s currency denomination. In addition to a basic 6 percent rate that applies to all deposits, foreign currency deposits (FCD) are also subject to a 20 percent marginal requirement (down from 45 percent in 1998). Thus, average required reserves on FCD are currently 32.5 percent. The central bank does not pay interest on reserves up to 6 percent and pays the 3-month LIBOR rate minus 1/8 on foreign currency reserves above 6 percent.
Liquidity requirements: Banks are required by the supervisory authorities to hold liquid assets equivalent to at least 8 percent and 20 percent of all their liabilities maturing during the next 12 months, in domestic currency and foreign currency, respectively. Eligible assets include vault cash, deposits at the central bank, central bank certificates of deposit, deposits in first-rate foreign banks, and investments in securities negotiated in centralized markets and rated as investment grade by international rating agencies.
Deposit insurance: Banks are required to make quarterly contributions to a deposit insurance fund at a rate equivalent to 0.48 percent of insured deposits (U.S. dollars and Peruvian nuevos soles).
Costs
Both liquidity and reserve requirements affect banks’ profits, as liquid prime assets normally earn lower returns than less liquid assets and reserve requirements are remunerated at below market rates. The deposit insurance affects banks’ profits by increasing their expenditures. Assuming that, in the absence of liquidity or reserve requirements, banks would only hold liquid assets equivalent to 5.5 percent of local currency liabilities and 3.6 percent of foreign currency liabilities, costs would amount to 2.2 percent of liabilities in foreign currency (1.9 percent for reserve requirements, 1.1 percent for liquidity requirements, and 0.15 percent for the deposit insurance) and 0.4 percent of liabilities in domestic currency (0.08 percent for reserve requirements, 0.3 percent for liquidity requirements, and 0.1 percent for the deposit insurance).
Benefits
The marginal contribution of the liquidity and reserve requirements to limiting liquidity risk can be estimated by subtracting the liquid assets that banks would hold voluntarily from the required liquid assets, which amounts to about 18 percent of total bank liabilities. This is substantially above the maximum run experienced by the Peruvian banking system from 1993 to June 2002 (see Table 7). However, runs do not affect all banks equally as they are usually accompanied by a flight to quality. For instance, during the 1998 run, the maximum run experienced by a single institution (including reductions of short-term foreign lines of credit) amounted to 35 percent of its liabilities. Thus, the level of protection provided by the regulations would not have been sufficient to withstand the largest run at the individual bank level.
Creditor Runs in the Peruvian Banking System1
Includes recalls of short-term foreign lines of credit.
Only considers declines of over 2 percent of total bank liabilities. It is assumed that a run ends when liabilities grow or are stable for two consecutive months.
Creditor Runs in the Peruvian Banking System1
First Month of the Run2 |
Total Loss for the Banking System (in percent) |
Duration of the Run (in months) |
Maximum Monthly Loss for the Banking System (in percent) |
Maximum Loss for an Individual Bank (in percent) |
Standard Deviation of the Run in Individual Banks (in percent) |
Comments |
---|---|---|---|---|---|---|
December 1993 | 2.0 | 1 | 2.0 | 11.5 | 7.3 | |
November 1996 | 2.0 | 1 | 2.0 | 19.8 | 12.5 | |
August 1998 | 2.6 | 2 | 2.1 | 20.4 | 15.1 | Impact of the Russian crisis. |
December 1998 | 8.3 | 8 | 4.4 | 35.0 | 24.0 | Impact of the Brazilian crisis. Banco República was liquidated and Banco Latino was capitalized by COFIDE (public development financial institution). |
May 2000 | 2.2 | 1 | 2.2 | 16.9 | 6.6 | Impact of the election period. |
August 2000 | 6.0 | 7 | 2.0 | 15.9 | 23.3 | Impact of political instability preceding the fall of the Fujimori administration. Two banks were closed—Nuevo Mundo and NBK Bank. |
September 2001 | 5.0 | 6 | 2.2 | 29.0 | 30.9 | Impact of political uncertainty and the Argentina crisis. |
Includes recalls of short-term foreign lines of credit.
Only considers declines of over 2 percent of total bank liabilities. It is assumed that a run ends when liabilities grow or are stable for two consecutive months.
Creditor Runs in the Peruvian Banking System1
First Month of the Run2 |
Total Loss for the Banking System (in percent) |
Duration of the Run (in months) |
Maximum Monthly Loss for the Banking System (in percent) |
Maximum Loss for an Individual Bank (in percent) |
Standard Deviation of the Run in Individual Banks (in percent) |
Comments |
---|---|---|---|---|---|---|
December 1993 | 2.0 | 1 | 2.0 | 11.5 | 7.3 | |
November 1996 | 2.0 | 1 | 2.0 | 19.8 | 12.5 | |
August 1998 | 2.6 | 2 | 2.1 | 20.4 | 15.1 | Impact of the Russian crisis. |
December 1998 | 8.3 | 8 | 4.4 | 35.0 | 24.0 | Impact of the Brazilian crisis. Banco República was liquidated and Banco Latino was capitalized by COFIDE (public development financial institution). |
May 2000 | 2.2 | 1 | 2.2 | 16.9 | 6.6 | Impact of the election period. |
August 2000 | 6.0 | 7 | 2.0 | 15.9 | 23.3 | Impact of political instability preceding the fall of the Fujimori administration. Two banks were closed—Nuevo Mundo and NBK Bank. |
September 2001 | 5.0 | 6 | 2.2 | 29.0 | 30.9 | Impact of political uncertainty and the Argentina crisis. |
Includes recalls of short-term foreign lines of credit.
Only considers declines of over 2 percent of total bank liabilities. It is assumed that a run ends when liabilities grow or are stable for two consecutive months.
While adaptations sometimes have been made to safety-net arrangements, the scope for dollar liquidity support varies substantially across countries. Countries such as Armenia, Bolivia, and Peru have explicit arrangements to provide liquidity assistance in foreign currency. To enhance the credibility of lender-of-last-resort arrangements, whether in local currency or foreign currency, dollarized countries generally hold higher international reserves (Figure 7). However, the relationship between international reserves, as a share of M2, and financial dollarization is much less than proportional—a 10 percent increase in dollarization results in only a 3 percent increase in international reserves—and there is very wide cross-country dispersion.27 Central banks of several countries, such as Bolivia, Argentina, Chile, Peru, and Uruguay, have introduced debt instruments denominated in dollars, or indexed to the dollar, as a way to enhance flexibility in the use of international reserves, limit exchange rate volatility, and, in some cases, to facilitate the domestic recycling of dollar liquidity. With regard to deposit insurance, the countries generally do not make distinctions in the extent of coverage provided for domestic or foreign currency deposits. However, the insurance is normally paid out in domestic currency, at exchange rates that are set out in the regulation.
Average Dollarization and Official Reserves to M2, 1996–2001
(In percent)
Sources: National authorities; IMF, International Financial Statistics database; and IMF staff estimates.Is a Regulatory Tightening Needed?
The existing regulatory response to dollarization risk appears to be generally uneven and often insufficient. As outlined above, the regulatory response to liquidity risk varies widely across countries. In many cases, the authorities’ responses appear to be insufficient in scope and content. Regarding solvency risk, the countries that are most often exposed seem to be the ones less inclined to tighten their regulations.
No specific guidelines have yet been issued by the Basel Committee on Banking Supervision on how to counteract credit risk exposure by banks in dollarized economies. Instead, substantial flexibility within the legal and regulatory framework is given to supervisors and banks on how to address these risks. In part, this may reflect the view that dollarization does not introduce significant qualitative differences to the nature of the risks banks face and the way in which these risks must be addressed. Differences, if any, are seen as a matter of relative magnitude and emphasis. Dollarization risks must be taken into account in the ongoing supervisory process, as part of the normal risk assessment and management dialogue between supervisors and banks. Thus far, however, such circumstances have not been viewed as requiring fundamental adjustments of prudential norms. Consistent with this view, banks are expected to be able to correctly assess the risks they face and, perhaps with some prodding from their supervisors, adequately manage them, taking into account the inherent tensions between risk reduction and the corresponding costs entailed. As in any risk management strategy, the total elimination of risks may be deemed to be too costly, keeping in mind the limited probability of large adverse shocks.
Specific features of many dollarized economies may call for a more aggressive prudential stance against dollarization risks. The close association between dollarization risks, the monetary and exchange rate regime, and moral hazard argues in favor of more prudential activism. Because solvency risks derived from dollarization are systemic in nature, expectations of a government bailout in the event of a catastrophic exchange rate devaluation are widespread. To the extent that depositors expect a bailout, they have less incentive to require higher risk premiums on the deposits offered by the banks that lend primarily in dollars to local currency earners. In turn, because depositors will not reward them with lower risk premiums, bank shareholders have no incentive to adequately provide against such risks. Instead, they are better off not offering such a provision because it allows them to limit their losses in the event of a catastrophic depreciation (e.g., it enhances the option value of walking away). Bailout expectations similarly induce dollar borrowers to discount the real cost and risks of dollar borrowing.28 By penalizing the more prudent banks, competitive forces can help broaden across all banks the failure of at least some of the participants to fully internalize risk.
Similar moral hazard failures affect banks’ exposure to liquidity risk. The large international reserves held by central banks and abundant associated provision of liquidity support in the event of systemic runs provide free insurance benefits. Banks therefore have limited incentive to accumulate dollar liquidity on their own, if they know that all banks will be similarly affected and will need to be supported.29 Instead, competitive pressures will tend to penalize those banks that set aside more liquidity or otherwise take measures to limit the liquidity of their liabilities.
Agenda for Prudential Reform
The main objective of a regulatory tightening would be to internalize risk. This should eliminate moral hazard and level the playing field in favor of the local currency. In addition, it should increase the capacity of dollarized financial systems to withstand liquidity or solvency shocks, thereby enhancing the scope for monetary and foreign exchange policy. However, as emphasized below, the degree of dollarization and the implications for financial intermediation of a regulatory tightening, particularly regarding the scope for regulatory arbitrage, also need to be taken into account when implementing policy reform.
Solvency Risk
A flexible regulatory approach to limit currency-induced credit risk is preferable to one based on strict quantitative limits. As suggested by the observation of country practices, the range of options to manage currency-induced credit risk is very broad, going from doing nothing at one extreme to outright prohibition of foreign currency deposits or loans at the other extreme. Requiring banks to restrict their foreign currency lending to fully hedged borrowers—naturally hedged by foreign currency revenue or financially hedged by using appropriate hedging instruments—may be appropriate in countries where dollarization is marginal or as a middle ground option where tighter administrative restrictions on financial dollarization are already in place and need to be relaxed to enhance the scope for financial intermediation. However, for countries that are already highly dollarized, this may drastically curtail the scope for domestic financial intermediation. Even if the growth of local currency intermediation is appropriately encouraged and takes up some of the slack, it is unlikely to fully or rapidly substitute for dollar intermediation, in either volume or depth. Moreover, by limiting the scope for currency diversification in an increasingly integrated world, strict limits on dollar lending could thwart financial development and sound prudential management. By promoting offshore intermediation, it could also hinder the financial system’s ability to manage and supervise risk. By encouraging the best borrowers to find alternative sources of finance, it may, instead, have perverse effects on the banking system’s overall exposure to credit risk. In the more developed financial systems, quantitative limits on dollarization are likely to be mostly ineffective because currency diversification will take place in other ways, notably through derivatives.30
While a flexible approach that requires banks to internalize risk but does not thwart sound risk management is preferable, it raises important challenges. In particular, a proper instrument that is transparent, yet flexible, needs to be used. General provisions are preferable to normal regulatory capital in that they provide a more flexible solvency buffer. When general provisions are not part of the minimum regulatory capital, they can be used swiftly when needed—by converting them to specific provisions as credits become delinquent—without putting the bank in a situation of prudential undercompliance, which would require immediate recapitalization and might generate a credit crunch and an adverse market response.31 However, general provisions are meant to be used for losses that have already been realized (although not yet fully identified). This is not consistent with a value-at-risk approach that seeks to identify the maximum losses that could be incurred under a worst-case scenario, within a given confidence interval. For this purpose, additional capital requirements in the form of specific reserves proportional to the excess value-at-risk assumed by a bank when denominating loans to unhedged borrowers in foreign currency rather than in local currency are indeed more appropriate. However, to ease the impact of a large depreciation on the banking system and limit the potential for a credit crunch, these reserves should not be part of the regular capital adequacy requirements. Instead, transparent, preferably rules-based, arrangements should be set up to allow the reserves to be drawn down as warranted and gradually rebuilt once used. Adjusting the pace of the reserves buildup across time could have the additional benefit of smoothing out the credit cycle and, hence, limiting the prudential risks arising from it.
Defining appropriate reserve needs can be difficult in practice. The extent to which a borrower is unhedged requires a careful analysis of its cash flow under large hypothetical changes in the exchange rate. This exercise requires analyzing the currency composition of a borrower’s assets and liabilities as well as the extent to which its revenues are likely to be affected by an exchange rate depreciation.32 Letting banks define the reserves that need to be set aside for each loan, based on their own internal risk models and stress-test criteria provided by the supervisor, would be consistent with the second pillar of the Basel Capital Accord that emphasizes banks’ own risk assessment and management practices. However, the systemic nature of such simulations, in which macroeconomic ripple effects are bound to play an important role, may be difficult to reconcile with the more micro-oriented, idiosyncratic approach generally used by banks. Moreover, such an approach would not be practical for the smaller loans and less-sophisticated banks or supervisory environments. Thus, a first-pillar type of approach (i.e., a specific capital charge) might be needed in many, if not most, cases. The supervisory authorities could directly provide detailed reserve coefficients by type of loan, based on a systemic value-at-risk approach that identifies the maximum exchange rate depreciation that might reasonably be expected and for which provision will be made.33 This approach should be complemented by specific supervisory guidelines on how to assess borrowers’ ability to manage a foreign currency loan. In all cases, banks should disclose the extent of unhedged lending.
Regulatory adjustments to limit banks’ direct exposure to currency risk may also be needed. The supervisory authority could, for instance, recognize the need to preserve capital by allowing for a long, structural, and open foreign exchange position not subject to capital requirements and that reflects the currency composition of banks’ balance sheets. However, symmetrical limits could be imposed on the remaining foreign exchange position. In all cases, capital charges must be set on such foreign exchange positions.34
Liquidity Risk
With regard to liquidity risk, the aim should be to reduce the liquidity of banks’ dollar liabilities and increase that of dollar assets. Prudential regulations to limit the risk of dollar liquidity crises need to start from the realization that local assets, such as short-term loans or treasury bills, can turn illiquid under a systemic crisis situation, unless they can be freely used to obtain dollar liquidity from the central bank. Thus, a micro-oriented approach, such as maturity matching of assets and liabilities, is not appropriate to manage liquidity risk on a systemic basis. A preferable option is to generalize and, when appropriate, increase currency-specific liquidity requirements. The most liquid dollar liabilities should have the highest backing against liquid dollar assets held abroad or at the central bank or against dollar public securities that can be rediscounted or used for repurchase operations against the central bank’s dollar reserves abroad (song, forthcoming). Liquidity ratios can gradually decline as the maturity of banks’ dollar liabilities lengthens.35 As in the case of provisioning requirements, a proper balance needs to be found between costs and risks.36
Strong assurances of support from a bank’s parent group or readily available contingent lines of credit from first-line international banks might be accepted, at least in part, as a substitute for holding liquid assets abroad (Song, forthcoming). In the case of branches of foreign banks, strong, legally binding assurances of support from parent banks might be acceptable in lieu of liquid reserves when the parent banks meet appropriate criteria, such as high ratings by international rating agencies and sufficient size in relation to their local branch.37 By adjusting the policy response to the underlying risk, this should limit the overall regulatory burden on the banking system and therefore be less constraining to its development capacity. However, by favoring the branches of large international banks over indigenous banks, this could have broader consequences for the future of the banking system that would need to be carefully assessed.38
Dollar-lender-of-last-resort arrangements could be used to provide additional backstop liquidity support to banks. Unlike liquidity requirements, lender-of-last-resort arrangements by central banks are exposed to moral hazard as they fail to internalize risk and may further encourage dollarization by reducing the risks associated with dollar deposits.39 Yet, pooling external reserves at the central bank provides some economies of scale that can lower the overall cost of a foreign exchange liquidity buffer. A proper balance between incentives and costs therefore needs to be found in apportioning the liquidity buffer between the central bank and the commercial banks.40 In this context, the development of dollar public debt instruments that can be used for dollar-based monetary or foreign exchange operations can serve a useful purpose in highly dollarized countries. They can provide a means to recycle dollar liquidity and manage foreign reserves more effectively as well as to increase the responsiveness of local dollar rates to changes in risk perceptions, thereby limiting the need for liquidity support from central banks.
Simultaneously, measures may also be needed to promote price responses to changes in asset demands and to limit quantity adjustments. This can be achieved by promoting the secondary market trading of long-term bank liabilities and other public or private securities. A loss of confidence (e.g., an increase in country risk) can then be immediately reflected in asset prices, ensuring that they continue to be demanded. Moreover, asset holdings can be reshuffled from the more risk-averse investors to the less risk averse. In this context, regulations penalizing early withdrawals of bank deposits or other bank instruments and measures to help develop securitization and mutual funds may also be needed to promote the growth of variable-price savings instruments, such as securitized mortgages, mutual fund shares, and long-term certificates of deposit. Such measures should promote a gradual migration of dollar short-term liabilities from banks to institutional investors. While mutual funds, pension funds, and other institutional investors may, at least in financial markets with limited development, need to invest a large share of their assets back into the banking system, they are more susceptible to acquiring longer-term instruments than the average individual investor.41
Measures to promote the development of specialized foreign exchange hedges, such as forwards and options, may also be desirable. Similar hedging benefits can be obtained by replacing domestic dollar intermediation with local currency intermediation, complemented by a forward market in which investors and borrowers take opposite positions. The main benefit of using specialized risk instruments is that it eliminates liquidity risk, albeit not counterparty risk. While derivatives are unlikely to develop in nascent financial systems, they can take a pre-dominant role in more developed systems.42
Enhancing the Attractiveness of Local Currency
Measures to internalize the risks of intermediating in foreign currency need to be reinforced by those that enhance the attractiveness of the local currency and limit the risks associated with its use as an intermediation medium. To mitigate the costs of an aggressive prudential agenda to limit and internalize dollarization risks, the prudential reforms should be accompanied by an equally aggressive agenda to boost the attractiveness of the local currency, including its credibility and market acceptance. By lowering the costs and risks of intermediating in local currency, this can help promote a viable and more resilient local-currency-based path of financial development. While the inherent difficulties in achieving sustained dedollarization should not be underestimated, comprehensive measures that promote the attractiveness of the local currency and simultaneously internalize the risks of dollarization, by being mutually reinforcing, offer the best chances of gradual success. Countries should, over time, experience a “virtuous cycle” in which dedollarization enhances the scope for monetary autonomy, leading, in turn, to a further decline of dollarization.
A strong and visible commitment to protect the long-term purchasing power of the currency should help enhance its credibility. This means that when local inflation is high, relative to world inflation, above all else, further efforts will be made toward stabilization. However, as shown by the recent experience of many highly dollarized countries, stabilization may not suffice to restore the credibility of the currency, at least within a reasonably short time span. Confidence in the currency also implies reducing the odds that the monetary authorities may again lose control of monetary policy in the future. To address this problem, one option is to adopt a full-fledged inflation-targeting framework, backed by a strong and sustainable fiscal position.43 Yet to avoid destabilizing expectations and, in a worst-case scenario, triggering a banking crisis induced by deposit outflows and a sharp exchange rate depreciation, a change in the monetary regime may need to be cautious and gradual, ensuring that proper conditions are in place.
Measures to enhance the market acceptance of the local currency may also be needed. Direct financial repression affecting the local currency, if in place, should be eliminated. This may include removing administrative ceilings on local currency interest rates, high unremunerated reserve requirements, or other forms of implicit or explicit taxation. There is indeed good evidence that such financial liberalization was at the root of Egypt’s successful dedollarization.44 Measures to promote the development of the local currency money and bond markets may also be needed. This could require an effort to gradually develop markets for local-currency-denominated public securities, starting at the shorter end of the maturity structure. Consistent with this effort, the fiscal authorities must lead by example, in terms of structuring debt in a way that reduces risk. They must be willing to accommodate somewhat higher debt-servicing costs, at least during a transition period, and ensure the sustainability of the public debt under potentially adverse exchange rate scenarios. The effort to promote the development of the money and bond markets in local currency may also need to be accompanied by a strengthening of monetary management in local currency, aimed at smoothing out the day-to-day volatility of interest rates, consistent with a gradual freeing of exchange rates. To promote market deepening and diversification, banks and other financial intermediaries should be free to create their own financial instruments, provided the resulting risks are adequately managed and reported.
The introduction or further promotion of price-indexed instruments can complement efforts to promote local currency intermediation. While not fully immune to the prudential risks associated with abrupt changes in relative prices, such as changes in real wages, price-indexed financial instruments can nonetheless provide a prudentially superior alternative to dollar-denominated instruments for promoting financial deepening, particularly at the longer maturities.45 Moreover, they offer public debt managers a less risky alternative to dollar debt, which is also less costly than nominal local-currency instruments. In Chile, after the 1982 crisis, the use of the Unidad de Fomento (UF)—a unit of account indexed to the consumer price index—helped to promote rapid financial reintermediation and deepening. However, the introduction of price-indexed instruments needs to be accompanied by a decisive effort to stabilize so as to avoid indexation in the real sector and must be properly presented and explained so that it is not taken as a signal of the government’s lack of resolve to fight inflation. It must also be gradual, not the least because it must take place symmetrically on both sides of banks’ balance sheets.46 The successful development of price-indexed instruments may also require an aggressive issue of price-indexed public securities to facilitate market development, a reform of the tax code to ensure tax neutrality, and a review of the price index—in cases where the underlying consumption basket differs excessively from that of the average depositor, a CPI-indexed financial instrument may be relatively less attractive to the average investor than a dollar-indexed instrument.47
Measures that enhance the quality of the payments system in local currency can give it a competitive edge over the dollar. The demand for local currency instruments can also be enhanced by improving the quality and reliability of the payments services provided by the central bank. For example, good-quality notes and reliable, same-day check clearing in local currency can encourage bank customers to make payments in local currency. Similarly, a reliable real-time gross settlement (RTGS) system in local currency can facilitate wholesale payments and money market transactions. When central banks are already faced with a large component of the payments system taking place in dollars—that is, when the dollar payments system is systemically important, they may, for risk management purposes, have to provide minimum payments services in dollars as well as in local currency, such as dollar settlement on the central bank’s books.48 However, even in such cases, the use of the local currency can still be encouraged through better service or lower fees.
Practical Safeguards and Countermeasures for Offshore Financial Centers
Bank supervisors are often constrained in their attempts to tighten domestic prudential regulations because funds could be diverted offshore to circumvent the regulations. Effective supervision of offshores may be compromised, as supervisors are unaware of the true dimension of the activities undertaken. It is also often difficult to identify “sufficient evidence of control” to effect consolidated supervision and some offshore structures (parallel banks) are not usually subject to consolidated supervision.
Consolidated supervision is often the basis for extending onshore regulations to offshore affiliates. Access to information has special relevance in instances where a local financial institution has a licensed affiliate in an offshore jurisdiction, but where affiliate is actually managed from the home country and does not maintain a physical office in the jurisdiction. Parent financial institutions and parent supervisory authorities must be in a position to monitor risk exposure, including reputation risk, of the banks or banking groups for which they are responsible, based on the totality of their business wherever it is conducted. In these situations, host country laws should permit the home country supervisor to exercise effective consolidated supervision and to have access to the books and records of the affiliate that is legally licensed in an offshore jurisdiction. In addition, the legal framework of the home country needs to allow home supervisors to request uniform regulations to be applied to onshore and offshore entities.
Evidence of control (the basis for consolidated supervision) between financial institutions is often difficult to find. This problem arises not only when trying to define the scope of a banking group but also in relation to large exposures or when deciding if a “true sale” or a “true transaction” has taken place. Supervisors have been encouraged to look at the core characteristics of transactions rather than their legal structure. Moreover, they should presume that there is some evidence of control if transactions in the offshore financial center take place at the local bank’s premises or include the local bank’s name in their own name. Sometimes the solution adopted is to require that any equity investment in an entity located in a country with a favorable tax system has to be a “control stake” and therefore such a subsidiary must be subject to consolidated supervision (e.g., Brazil).
Conditional licensing: Recognizing that for affiliates operating in offshore financial centers there is a higher potential for restricted access to information, the home country supervisor should condition its authorization to establish an affiliate in such a jurisdiction based on a clear understanding with the institution that it will grant access to information needed for supervisory purposes.
Punitive capital charge: As a measure to ensure access to books and records for affiliates licensed in an offshore financial center, the home country supervisor could impose a punitive capital charge against the investment to such an extent that access is prevented. For a particular investment, instead of the requirement that the bank hold the minimum capital against the investment (e.g., 8 percent for many jurisdictions), the home country could impose a 100 percent requirement, as is done in Brazil.
Parallel banks: By definition, parallel banks share management and/or ownership links with a financial institution in a home country. Given the potential for inappropriate activities, supervisors have moved to impose a requirement that the bank in the home country consolidate into its operations (on an accounting basis) the activities of the parallel bank. Some supervisors either force a change in the group structure or directly include the parallel banks in the scope of consolidated supervision (e.g., Spain). The result is that the parallel bank is treated as a subsidiary for supervisory purposes and all typical consolidated supervision requirements would apply to the subsidiary.
Information sharing: As specified in Principle 25 of the Basel Core Principles for Effective Banking Supervision, host country supervisors must have power to share information needed by the home country supervisor for the purposes of carrying out consolidated banking supervision. Frequent arrangements include the use of memoranda of understanding or exchange of letters between supervisors that spell out clear parameters for information sharing and the authority to conduct on-site inspections. In this case, information on the asset side of the balance sheet poses fewer problems than disclosure of depositor information. There is a growing perception that information primarily related to liquidity (e.g., quantitative information on deposit concentrations or some specific situations involving a customer or groups of customers) should also be easily accessible. National laws usually prevent disclosure of information on individual deposit accounts. In practice, home supervisors may obtain information through the parent bank or by examining the bank themselves.
Implementation Issues
Tightening prudential norms without encouraging the growth of offshore or unregulated intermediation can constitute a major challenge. Measures to fully internalize dollarization risk and build prudential buffers might encourage the growth of unregulated intermediaries, particularly offshores. When local banks intermediate through offshore, rather than through domestic, branches, the risks incurred by the financial groups and the systemic exposure of the banking system as a whole remain basically the same. Yet the transparency of banks’ financial statements and the reach of the supervisor are drastically curtailed. Indeed, the need to recover control on their financial intermediaries led many countries with large offshore banks (including Costa Rica and Guatemala) to remove restrictions on domestic foreign-currency intermediation. Thus, a policy reversal toward tighter regulations will require that ways be found to limit the scope for regulatory arbitrage. As described in Box 5, these usually involve making the authorization of the offshore branches conditional upon a full consolidation with their onshore parents and the existence of comprehensive information-sharing agreements with the supervisory authorities of the countries in which the offshore branches are based.
The nature of the policy response to dollarization is likely to depend on the degree of dollarization, its trend, and the monetary and exchange regime. In countries where inflation is low and monetary control is good, where the possibility of an abrupt regime change is remote because the exchange rate is already flexible, and where dollarization is limited and stable, there may not be a need to alter prudential policies by much, if at all. In the intermediate case of countries where dollarization is still limited but increasing—or could increase due to unstable, or potentially unstable, monetary conditions or a more open capital account—a comprehensive, vigorous, and prompt policy agenda is needed, aimed at addressing prudential issues raised by dollarization. The extreme case of heavily dollarized countries where lack of currency credibility is deeply rooted calls for a more careful and nuanced assessment of the nature, pace, and strategy of policy reform.
The pace and timing of implementation of prudential reforms in highly dollarized countries needs to be carefully designed. In countries where most deposits and credit are in dollars, the necessary tightening of prudential norms would imply, in many cases, large capital additions and a reshuffling of banks’ balance sheets. If implemented too abruptly, this could exert an excessive burden on banks, worsen their financial situation, and lead to a credit crunch and a severe financial disintermediation. In countries where credit has already declined and banks have incurred large losses on their loans, following a slowdown in economic activity and exchange rate depreciations, such a prudential tightening may need to be phased in more gradually. Nonetheless, the opportunity to motivate the need for policy reform and to send a clear, up-front signal of forthcoming changes in policy orientations could be seized before the memories of the recent events fade away.