Journal Issue

Vietnam: Selected Issues and Statistical Appendix

International Monetary Fund
Published Date:
January 2002
  • ShareShare
Show Summary Details

II. Foreign Currency Depositsin Vietnam—Trendsand Policy Issues8

A. Introduction

15. As in a number of developing and transition economies, foreign currency deposits (FCDs) have become an important feature of monetary developments in Vietnam. Rising FCDs are typically associated with the increasing use by residents of foreign currency as a medium of exchange (for transactions) and/or as a store of value (for investment purposes), and are usually referred to as dollarization. However, there are no reliable data on foreign currency in circulation in Vietnam (although the amount involved is believed to be substantial), so that a comprehensive analysis of dollarization is not feasible. This chapter is focused on developments in FCDs as a proxy for dollarization. It reviews the FCD trends, analyses the causes, and assesses the benefits and policy challenges. The chapter concludes that in terms of policy implications, the gradually increasing (albeit moderate) dollarization in Vietnam appears to be largely in line with the country’s growing monetization and closer integration into the global economy. While this trend should not be resisted through administrative means, policy makers need to address the resulting challenges regarding prudential supervision, exchange rate management, and monetary policy.

B. Dollarization in Vietnam—Stylized Facts

16. Measured as the ratio of foreign currency deposits (FCDs) to broad money, dollarization in Vietnam first fell sharply from its peak in 1991 of 41 percent, following macroeconomic stabilization. Within two years, this ratio fell to 23 percent and stabilized in the range of 20 to 24 percent in the mid 1990s. Starting in 1997, the ratio has risen to more than 34 percent by August 2001. By international comparisons, while remaining in a medium range, dollarization in Vietnam has increased to significant levels (Figure II.1).

Figure II.1.Vietnam: Foreign Currency Deposits Indicators

Sources: IFS and Fund staff estimates.

1/ End June Data 2/ 1998 and 2000 data.

Vietnam: Dollarization, 1990-2001

Source: Vietanmes authorities; and Fund Staff Estimates.

Vietnam: Curaulalive Increase in Foreign Currency Deposits and Annual Remittances, 1995-2000

(In millions of U.S. dollars)

Sources: Vietnamese auihorities; and Fund staff estimates.

17. A variety of factors have influenced FCD developments. Hyperinflation in the early 1990s caused the sudden steep increase in dollarization ratios in 1991. Spillovers from the Asian financial crisis in 1997 were also associated with an uptick in dollarization. Specifically, spreads in cx-post real returns in favor of foreign currency over dong deposits peaked during 1998, primarily mirroring dong devaluations in the aftermath of the Asian crisis, and thus contributed to an increase in the dollarization ratio from 21 percent in September 1997 to 26 percent one year later. However, the increase in dollarization during 1999-2001, a period of relative macroeconomic stability, appears to be multifaceted, and four factors have played a role.

18. The first major factor is the reintermediation of foreign currency previously held outside the banking system. For the most recent upswing in dollarization, the strongest increase, by far, was recorded among households. In particular, during 2000, their FCDs were almost doubled to US$ 3.1 billion. Over the past 1½ years, households have accounted for 80 percent of the total increase in FCDs. (Table II. 1). Yet, in comparison, remittances have increased more gradually, and contributed only partially to the strong increase in FCDs (Figure II.1). Moreover, with tight controls in place on all capital account transactions, inward capital flows may have been limited. Overall, the sharp increase in households’ FCDs is believed therefore to reflect renewed confidence in the banking system, and was facilitated by the lifting of regulation on the opening of foreign currency accounts in 1999.9 Also pointing to the role of increased confidence is the fact that nearly all the increase in household FCDs has been in the form of time deposits. Finally, the preference of foreign currency deposits among households is likely to reflect also the relative thinness of the financial market and the lack of alternative interest-bearing financial instruments in Vietnam (Veliov, 2001).

Table II.1Vietnam: Deposit Money Bank Deposits, 1999-2001 1/(In trillions of dong)
(By type of depositor)
Dong deposits37.841.845.549.759.458.059.862.3
Foreign currency deposits 2/
of which: demend deposits12.313.912.812.613.514.013.213.5
Households 3/
Dong deposits26.630.630.931.229.935.635.935.1
Foreign currency deposits 4/22.627.029.133.645.147.851.554.6
of which: demand deposits0.
Deposit substitutes (dong) 5/10.912.314.713.
(Change since beginning of the year)
Dong deposits4.07.611.821.6-
Foreign currency deposits 2/
Households 3/
Dong deposits4.
Foreign currency deposits 4/4.46.511.
Deposit substitutes (dong) 5/
(Foreign currency deposits by bank)
Vietcombank (VCB)18.921.222.825.129.534.235.238.1
Incombank (ICB)
Bank for Agriculture and Rural Development (BARD)
Bank for Investment and Development (BIDV)
(Foreign currency deposits by bank in percent of total)
(Percentage share in increase of foreign currency deposits since beginning of the year)
(In millions of U.S. dollars, unless indicated otherwise)
Memorandum items:
Foreign currency deposits
Demand deposits (in percent of total foreign currency deposits)
Sources: State Bank of Vietnam; and Fund staff estimates.

Based on monetary survey comprising 6 state-owned commercial banks (SOCBs) and 83 non-state credit institutions.

Adjusted for enterprise holdings of foreign currency deposit substitutes at VCB (about 45 percent of total).

Referred to as savings deposits in the monthly derivation tables on the banking system.

Adjusted for household holdings of foreign currency deposit substitutes at VCB (about 55 percent of total) and at BIDV (all deposits estimated to be held by households).

Adjusted for deposit substitutes of VCB and BIDV (i.e. dollar-denominated bonds issued by banks).

Sources: State Bank of Vietnam; and Fund staff estimates.

Based on monetary survey comprising 6 state-owned commercial banks (SOCBs) and 83 non-state credit institutions.

Adjusted for enterprise holdings of foreign currency deposit substitutes at VCB (about 45 percent of total).

Referred to as savings deposits in the monthly derivation tables on the banking system.

Adjusted for household holdings of foreign currency deposit substitutes at VCB (about 55 percent of total) and at BIDV (all deposits estimated to be held by households).

Adjusted for deposit substitutes of VCB and BIDV (i.e. dollar-denominated bonds issued by banks).

Vietnam: Foreign Currency Deposits, 1995-2001

Source: Vietanmes authorities; and Fund staff estimates.

19. The second main factor is the differentials in ex ante real returns and portfolio and risk diversification motives. In particular, periodic expectations of a dong devaluation, in part fuelled by the real appreciation of the dong intermittently since early 2000, tended to coincide with rising dollarization. With devaluation expectations, households and enterprises have sought protection of wealth and purchasing power in FCDs. With nearly 80 percent of all FCDs in the form of time deposits, such investment and diversification motives, as explained in more detail below, may have played a primary role as opposed to mere money demand factors typically associated with current return differentials. Moreover, enterprises, which hold the bulk of short-term dong deposits, are estimated to have switched out of dong into foreign currency deposits, particularly during the first four months of 2001, when the dong appreciated by a total of 2.2 percent in real terms.

Vietnam: Dollarization and Real Effective Exchange Rate January 2000-August 2001

Source: Vietanmes authorities; and Fund staff estimates.

20. The third factor is the strong export performance during 1999-2001, when quarterly exports nearly doubled, compared with the previous three-year period average (Figure II.2). This factor coupled with a lower surrender requirement (first reduced in 1999 from 80 to 50 percent and then in 2001 to 40 percent) has allowed many exporters, who are also importers, to retain a larger share of their foreign currency earnings in order to self-finance their growing import activities and reduce related exchange rate risks. It is also noteworthy in this context that, by end-August 2001, more than 40 percent of all FCDs were held with Vietcombank, the main bank for international trade transactions. Moreover, in recent months this bank accounted for up to two thirds of the increase in FCDs in the entire banking system.

Figure II.2.Vietnam: Foreign Currency Deposits, Exports, Interest Rates and Reserve Requirements

21. The fourth factor is the virtual elimination of the dong-dollar interest rate differential, which had previously been very much in favor of dong deposits. The rate differential was reduced from 3 percent in mid-1999 to the range of 0 to ½ percent since early 2000, making FCDs more attractive (Figure II.2). Dong deposit rates may have been lowered in part to help maintain banks’ profitability, particularly when nonperforming loans were mounting, and in part as a result of ample liquidity in the banking system at the end of 1999. Continued restrictions on banking competition may have also contributed to the overall lowering of dong deposit rates, as foreign banks are not permitted to offer dong deposits in excess of 25 percent of their capital to non-borrowing households and enterprises.

22. On the other hand, recent administrative measures to curb the increase in FCDs appear to have been ineffective. Aimed at making FCDs less attractive to banks and depositors alike, the State Bank of Vietnam (SBV) raised in three steps the reserve requirements for FCDs to 15 percent in May 2001 (Figure II.2). The ceiling on interest rates on FCDs of enterprises continued to be in the range from 0.5 percent a year for demand deposits to 3 percent a year for time deposits over six months. Moreover, foreign bank branches are prohibited from offering FCDs to Vietnamese households. Despite these steps, FCDs continued to increase rapidly during January 2000 to August 2001.

C. Dollarization and the Role of Currency Versus Asset Substitution

23. In order to explain dollarization trends, it is useful to distinguish between two motives for the demand for foreign currency. The first is currency substitution, which occurs when residents hold foreign currency as a means of payment. The second is asset substitution, when residents hold foreign currency denominated assets for investment purposes, rather than as a means of payment.

24. Currency substitution is typically attributed to periods of hyperinflation (as explained in more detail in the Box II. 1), when the cost of using domestic currency for transactions becomes very high. This phenomenon helps explain Vietnam’s experience in the early 1990s when dollarization peaked at 41 percent as a result of hyperinflation, but then declined to about 20 percent in 1995 following the return to macroeconomic stability. However, since then, dollarization has gradually increased even though inflation has remained low and the exchange rate relatively stable.

25. This in turn may be due to increased asset substitution which is determined by risk and return characteristics of domestic and foreign currency assets. Thus, even when the expected exchange rate is stable and inflation is low, foreign currency denominated assets may be preferred as they protect against a possible (but unexpected) return of macroeconomic instability. As suggested by the calculations in Box II. 1, Vietnam’s past episodes of inflation and real exchange rate volatility could form expectations on future volatility that result in hypothetical dollarization ratios broadly in line with the ratios observed in recent months.

26. Asset substitution could also lead to higher dollarization as a natural consequence of increased trade and economic integration (Figure II.3). As more openness tends to result in a stronger exchange rate pass-through, trade integration would lead to higher inflation volatility and lower volatility of the real exchange rate.10 As a result, domestic currency assets would become more risky relative to foreign currency assets, which in turn would induce investors to increase the share of foreign currency assets in their portfolios (see Box II. 1). Indeed, Vietnam experienced a rapid increase in economic integration with the rest of the world during the 1990s; the share of trade, as measured by the ratio of imports and exports to GDP more than doubled from 44 percent in 1990 to 94 percent in 2000. Also, the exchange rate pass-through tended to increase (as indicated above), with the contemporaneous correlation of quarterly inflation and changes in the dong-U.S. dollar exchange rate increasing from 0.06 for the period 1993Q1–2001Q2 to 0.33 for the past 3½ years.

Figure II.3.Vietnam: Openness and Financial Indicators

Box II.1.Currency Versus Asset Substitution—Two Basic Approaches to Explaining Dollarization

Currency Substitution

The standard analytical framework for currency substitution has been a money demand equation augmented by expected exchange rate changes and foreign interest rates.1 Such an approach aims at measuring the opportunity cost of holding domestic versus foreign currency. The degree of substitution is then determined by relative rates of return deriving from interest rate differentials between foreign and domestic currency and expected exchange rate changes. This approach has been useful in explaining large upswings in currency substitution during periods of macroeconomic turbulence, when return differentials or the opportunity cost to hold domestic currency are high. However, it is less successful in explaining the persistence of high rates of dollarization when macroeconomic stability returns and return differentials subside. For example, many Latin American countries have experienced continuously rising currency substitution ratios even after the return of macroeconomic stability (Savastano, 1996). By comparison, the experience in transition economies has been more mixed. Poland and the Baltic states experienced significant reductions in currency substitution ratios as inflation returned to single digit levels (Sahay and Vegh, 1995).

Asset Substitution

Based analytically on the minimum-variance portfolio approach (Ize and Levy-Yeyati, 1998), asset substitution would be determined by the volatility of inflation (increasing the riskiness of domestic currency assets) and the real exchange rate (increasing the riskiness of foreign currency assets in terms of domestic currency). Subjecting investors to a risk-return trade off, the attractiveness of domestic relative to foreign currency deposits increases ceteris paribus when inflation volatility declines relative to real exchange rate volatility. In other words, for given inflation volatility, a decline of the volatility of the exchange rate would increase asset substitution, as it would enhance the benefits of hedging wealth with foreign currency assets. Thus, dollarization levels could remain high in a period of macroeconomic stability and low return differentials if the expected volatility of inflation remained high relative to the expected real exchange rate volatility. Based on these considerations, and assuming past episodes of inflation and exchange rate volatility determine expectations, the minimum-variance approach to a portfolio of dong and foreign currency assets would imply an underlying rate of dollarization for Vietnam that could be as high as 50 percent. Assuming that past memories of hyperinflation are fading and taking into account only data for 1994 to the present (a period of relative macroeconomic stability), the implied dollarization ratio would be about 30 percent—broadly in line with the ratio observed in recent months.

Vietnam: Measures of underlying dollarization based on asset substitution(as of June 2001)
Variance (quarterly change in CPI, p)34.784,942.61
Variance (quarterly REER, e)29.5212.4113.77
Covariance (CP1, REER)5.981.361.01
Underlying dollarization (MVP, in percent) 1/533120
Source: Vietnamese authorities; IMF, Information Notice System; and Fund state estimates.

Based on Ize and Levy-Yeyati (1998), the minimum-variance portfolio allocations yield the following dollarization ratio d (foreign currency denominated assets to total assets): d = Var(p) + 2 Cov(p,e)/[Var(p) + Var(e) + 2 Cov(p,e)].

Source: Vietnamese authorities; IMF, Information Notice System; and Fund state estimates.

Based on Ize and Levy-Yeyati (1998), the minimum-variance portfolio allocations yield the following dollarization ratio d (foreign currency denominated assets to total assets): d = Var(p) + 2 Cov(p,e)/[Var(p) + Var(e) + 2 Cov(p,e)].

1 For a recent literature overview see Balino et al. (1999) or Vetlov (2001).

27. The more recent increase of dollarization in Vietnam appears to be due primarily to asset substitution and the increasing integration with the global economy.11 Data on FCDs corroborates this view. First, the ratio of demand deposits to total FCDs declined from 29 percent at end-1999 to 17 percent at end-August 2001. This shift coincided with the rapid increase in household FCDs during 2000, nearly all of which was attributable to time deposits. However, data on FCDs can only provide a limited assessment of the extent of dollarization in Vietnam. First, reliable estimates on the amount of U.S. dollar currency in circulation are not readily available.12 Second, information on cross-border deposits held at banks abroad, which can be considered close substitutes for FCDs, is incomplete at best. However, data from the BIS suggest that Vietnamese cross-border deposits played a minor role and were relatively stable during 1999-2000, increasing by US$24 million to some US$139 million at end-2000, which was equivalent to less than 3 percent of FCDs in the Vietnamese banking system. Third, due to statistical inadequacies, the measure of FCDs may include holdings of non-residents (Agenor and Khan, 1996; Buchs, 2000).

D. Benefits, Challenges, and Costs of Dollarization

28. The increase of FCDs in the Vietnamese banking system, particularly since late 1999, has clearly enhanced the opportunities for reintermediation and financial deepening. Moreover, it has strengthened banks’ financial performance. Overall, banks have been cautious in extending foreign currency loans domestically, and have placed the bulk of FCDs as deposits with banks abroad (Table II. 2). These deposits have provided banks with a stable source of income, particularly considering the burden of nonperforming loans on banks’ income position. It is noteworthy that one percentage point in the interest rate margin earned on FCDs at end-2000 would have been equivalent to about 3 percentage points in terms of return on banks’ equity.

Table II.2.Vietnam: Foreign Assets of the Banking System(In millions of U.S. dollars)
Foreign Assets3,4455,5427,7708,744
Deposit with banks3,1463,7645,8246,558
State Bank of Vietnam2,0001,8601,7792,085
Commercial banks1,1461,9044,0454,472
State Bank of Vietnam981,5621,7341,826
Commercial banks201215213360
Memorandum item:
Exchange rate (dong per U.S. dollar)13,89014,02514,50114,993
Source: Vietnamese authorities.
Source: Vietnamese authorities.

29. Allowing residents to hold FCDs also enhances the credibility of the macroeconomic policy stance in Vietnam, by giving a greater role to market forces and avoiding direct intervention in asset allocation decisions of the private sector. Because the cost of monetary indiscipline and intervention increases with the degree of dollarization, the government should thus also be more credibly committed to stronger financial policies.

30. Perhaps the most important challenge of dollarization concerns banks’ safety and prudential supervision. Dollarization may add to banking sector vulnerabilities, arising from a currency mismatch of deposit and lending activities, and from exchange and credit risks; the share of nonperforming foreign currency loans to total loans is roughly twice as high as that of dong loans (Figure II.3).13 Moreover, especially in the case of state-owned banks, the potential volatility of FCDs could give rise to liquidity problems and represent potential claims on Vietnam’s official reserves. The coverage of FCDs by gross official reserves has declined substantially over the past five years from over 140 percent at end-1996 to below 60 percent at end-August 2001 (Figure II.3).

31. Dollarization also implies a loss of seigniorage revenues for the SBV, to the extent that it arises from currency substitution. While some seigniorage is earned from bank’s required reserves in foreign currency at the SBV, it is lost from the foreign currency in circulation. As noted above, no reliable data are available on the amount of foreign currency in circulation. However, assuming the ratio of foreign currency in circulation to deposits in foreign currency is the same as that in dong, foreign currency in circulation would be equivalent to D 42 trillion, representing about 10 percent of GDP at end-2000. With monetary aggregates programmed to grow by about 20 percent on average, this form of seigniorage losses would amount to 2 percent of GDP a year.14

E. Policy Implications

32. Dollarization in Vietnam appears to be largely a result of the country’s deepening monetization and closer Integration with the global economy, particularly in recent years. Administrative measures to reduce dollarization should be avoided as this could give rise to renewed disintermediation. In particular, reserve requirements should be more uniform so as not to limit or distort the intermediation of banks, albeit within prudential bounds. Moreover, increased interest rate flexibility, applied to both foreign currency loans and deposits, would allow banks to more accurately price risks involved and improve resource allocation.

33. Accepting dollarization, however, means that the SBV will have to address new challenges in prudential supervision, as well as a broader set of parameters in designing the appropriate monetary and exchange rate policies. In particular, prudential regulation on net open foreign currency positions needs to be strictly enforced and the quality and maturity of banks’ foreign currency assets monitored carefully. While the limited role of foreign currency lending in Vietnam is reassuring given banks’ weak credit risk management, the liquidity and quality of banks’ foreign assets as a counterpart to FCDs will need to be monitored. Supervision needs to be strengthened to maintain the public’s confidence in the banking system. While systemic liquidity risks could be addressed by a higher reserve requirement on FCDs, this latter should not impede bank intermediation and thus add to a foreign liquidity squeeze. Moreover, the SBV should set aside all required reserves on FCDs and should not consider them freely usable gross official reserves.15

34. On balance, the exchange rate regime should be flexible, given the prevalent role of asset substitution in Vietnam. Fixing the exchange rate under asset substitution, as with higher capital mobility, would further limit the SBV’s scope for independent monetary policy.16 At the same time, the scope for using a flexible exchange rate as an instrument to reduce dollarization is limited over time. As the exchange rate pass-through is expected to increase with Vietnam’s growing openness, the riskiness of foreign currency assets would decline along with the volatility of the real exchange rate.


    AgenorPierre Richard andMohsin S.Khan1996“Foreign Currency Deposits and the Demand for Money in Developing Countries,”Journal of Development EconomicsVol. 50 pp.101118.

    BalinoTomas J.T. andAdamBennett andEduardoBorenzstein1999“Monetary Policy in Dollarized Economies” IMF Occasional Paper 171 (Washington: International Monetary Fund).

    BergAndrew andEduardoBorenzstein2000“The Choice of Exchange Rate Regime and Monetary Target in Highly Dol1arized Economies,”IMF Working Paper WP/00/29 (Washington: International Monetary Fund).

    BuchsThierry2000“Currency Substitution in the Russian Federation (1992-1997),”MOC-MOST: Economic Policy ill Transitional EconomiesVol. 10 pp. 95110.

    CataoLuis andMarcoTerrones2000“Detennmants of Dollarization: The Bunking Side,”IMF Working Paper WP/00/146 (Washington: International Monetary Fund).

    FischerStanley1982“Seigniorage and the Case for a National Money,”Journal of Political EconomyVol. 90 pp. 295313.

    IzeAlain andEduardoLevy-Yeyati1998“Dollarization of Financial Intennediation: Causes and Policy Implications,”IMF Working Paper WP/98/28 (Washington: International Monetary Fund).

    MuellerJohannes1994“Dollarization in Lebanon,”IMF Working Paper WP/94/129 (Washington: International Monetary Fund).

    SahayRatna andCarlosVegh1995“Dollarization in Transition Economies: Evidence and Policy Implications,”IMF Working Paper WP/95/96 (Washington: International Monetary Fund).

    SavastanoMiguel A.1996“Dol1arization in Latin America: Recent Evidence and Some Policy Issues,”IMF Working Paper WP/96/4 (Washington: International Monetary Fund).

    VetlovIgor2001“Dollarization in Lithuania: An Econometric Approach,”Bank of Finland Institute for Economies in Transition Discussion Papers No. I (Helsinki: Bank of Finland).

Prepared by Olaf Unteroberdoerster.

The new regulation entitles resident individuals to open and maintain foreign currency (time and savings) accounts, to hold foreign currencies received in a number of ways (including in the form of remittances), and to make various kind of payments in foreign currency (SBV Circular No. 01/1999/TT-NHNN7).

At the extreme, for example, if the exchange rate pass through were equal to one, the volatility of the real exchange rate would be zero, as exchange rate changes would be matched one-for-one by changes in the inflation rate.

Alternatively, one might attribute the persistence of a certain degree of dollarization to switching costs (Mueller, 1994). However, switching costs are considered high for currency, but they may be less significant for interest bearing deposits, which are used as a measure of dollarization here. Note that in many empirical studies, these two components have typically been lumped together in a measure of currency substitution or dollarization (Catao and Terrones, 2000) because of insufficient data on foreign currency in circulation.

Previous research in the Fund and elsewhere has found data from the U.S. Customs Currency and Monetary Instruments Reports (CMIR) to be informative in this respect. This data, which is not available for Vietnam, would permit to calculate cumulative net inflows as a proxy for the stock of U.S. dollar currency circulating in a given country. The accuracy of the estimate would depend, however, on the length of the cumulation period and the absence of outflows, which are unrecorded.

The interest rate ceiling on foreign currency loans, which was only lifted effective June 1, 2001 (nearly a year after the ceiling on dong loans was replaced by the base rate mechanism), might have prevented banks from accurately pricing risks associated with such loans, thereby contributing to the relatively high share of NPLs in this segment.

Similar calculations are discussed by Fischer (1982). Seigniorage losses tend to be highest for developing countries with rudimentary payment and financial systems and thus a high ratio of currency in circulation to GDP. Therefore, seigniorage losses in percent of GDP should be expected to decrease over time as the financial system becomes more sophisticated and reliance on cash transactions declines.

For this reason and in line with the standard definition of international reserves (BOPM5) to cover only readily available and unencumbered reserves, FCDs of deposit money banks held at the SBV are subtracted from gross official reserves when calculating net official international reserves (NIR) under the PRGF-supported program.

It should be cautioned, however, that to the extent currency substitution persists a reduced flexibility of the exchange rate remains justifiable. This is because currency substitution tends to increase exchange rate volatility. As a result, the exchange rate in a floating rate system will be more sensitive to expected changes in the money supply and other variables affecting the money market. For a more detailed discussion on appropriate exchange rate regime in response to various economic shocks under currency substitution see Berg and Borenzstein (2000).

Other Resources Citing This Publication