This section investigates currency substitution and dollarization in the Lebanese context during 1964–93, and particularly since the start of the civil war. To provide a more accurate insight into the public's asset portfolio response since the beginning of the war, the origins and extent of Lebanese currency substitution and dollarization are examined. This is followed by a look at capital flight and repatriation and their key determinants. Finally, the relative merits and demerits of pursuing a policy of de-dollarization of the Lebanese economy, as well as the requirements for achieving such a goal, are discussed.
When civil disturbances erupted in 1975, the immediate response of the public was a portfolio switch to liquid assets; the ongoing nature of the conflict and its accompanying high rates of domestic inflation then induced the abandonment of the Lebanese pound in favor of the U.S. dollar and dollar-denominated deposits. In the face of heightened political and economic risk in the 1980s, a spatial rearrangement of Lebanese wealth portfolios occurred, involving large-scale capital flight. While recent years of economic stability have prompted capital repatriation, there was little evidence of an increased desire by the Lebanese to hold pound-denominated assets between 1990 and the end of 1993 (Table 8).
Currency Substitution

Currency Substitution
| Money (M1) (1) | Demand Deposits(2) | Currency in Circulation(3) | Ratio LL Deposits to Total Deposits (4) | Annualized Change in LL/US$ (5) | |
|---|---|---|---|---|---|
| (In millions of Lebanese pounds) | (In percent) | ||||
| 1964 | 1,390.00 | 889.80 | 500.20 | 83.21 | -076 |
| 1965 | 1,510.00 | 961.60 | 548.40 | 81.75 | -0.06 |
| 1966 | 1,540.70 | 893.60 | 647.10 | 80.61 | 1.92 |
| 1967 | 1,569.90 | 782.20 | 787.70 | 80.15 | 2.36 |
| 1968 | 1,749.30 | 909.10 | 840.20 | 81.32 | -1.49 |
| 1969 | 1,657.30 | 787.00 | 870.30 | 79.33 | 3.10 |
| 1970 | 1,676.50 | 830.00 | 846.50 | 78.38 | 0.44 |
| 1971 | 2,001.50 | 1,080.00 | 921.50 | 77.86 | -1.26 |
| 1972 | 2,274.70 | 1,239.30 | 1,035.40 | 79.97 | -5.48 |
| 1973 | 2,618.90 | 1,389.90 | 1,229.00 | 76.37 | -14.43 |
| 1974 | 2,998.20 | 1,642.20 | 1,356.00 | 76.99 | -10.83 |
| 1975 | 3,835.90 | 1,587.70 | 2,248.20 | 75.15 | -1.11 |
| 1976 | 4,904.80 | 1,803.60 | 3,101.20 | 78.02 | 24.74 |
| 1977 | 5,061.50 | 2,299.60 | 2,761.90 | 77.65 | 6.87 |
| 1978 | 6,147.80 | 2,838.00 | 3,309.80 | 80.63 | -3.70 |
| 1979 | 6,683.80 | 3,151.70 | 3,532.10 | 69.43 | 9.72 |
| 1980 | 7,666.60 | 3,668.40 | 3,998.20 | 63.76 | 5.96 |
| 1981 | 9,005.10 | 4,359.70 | 4,645.40 | 58.64 | 25.55 |
| 1982 | 11,069.80 | 5,467.80 | 5,601.00 | 73.29 | 9.96 |
| 1983 | 12,945.00 | 5,867.70 | 7,077.30 | 73.65 | -4.54 |
| 1984 | 13,783.60 | 6,089.00 | 7,694.60 | 68.94 | 43.79 |
| 1985 | 20,154.20 | 9,852.00 | 10,302.20 | 64.90 | 152.14 |
| 1986 | 30,325.60 | 15,557.40 | 14,768.20 | 29.04 | 133.72 |
| 1987 | 68,890.00 | 29,541.00 | 39,349.00 | 8.08 | 485.34 |
| 1988 | 182,862.00 | 65,810.00 | 117,052.00 | 20.87 | 82.21 |
| 1989 | 287,186.00 | 92,526.00 | 194,660.00 | 33.07 | 21.37 |
| 1990 | 449,923.00 | 115,271.00 | 334,652.00 | 26.68 | 39,94 |
| 1991 | 689,405.00 | 202,007.00 | 487,398.00 | 32.42 | 33.54 |
| 1992 | 1,199,399.00 | 393,537.00 | 805,862.00 | 31.86 | 84.52 |
| 1993 | 1,213,887.00 | 498,918.00 | 714,969.00 | 32.34 | 1.67 |
Currency Substitution
| Money (M1) (1) | Demand Deposits(2) | Currency in Circulation(3) | Ratio LL Deposits to Total Deposits (4) | Annualized Change in LL/US$ (5) | |
|---|---|---|---|---|---|
| (In millions of Lebanese pounds) | (In percent) | ||||
| 1964 | 1,390.00 | 889.80 | 500.20 | 83.21 | -076 |
| 1965 | 1,510.00 | 961.60 | 548.40 | 81.75 | -0.06 |
| 1966 | 1,540.70 | 893.60 | 647.10 | 80.61 | 1.92 |
| 1967 | 1,569.90 | 782.20 | 787.70 | 80.15 | 2.36 |
| 1968 | 1,749.30 | 909.10 | 840.20 | 81.32 | -1.49 |
| 1969 | 1,657.30 | 787.00 | 870.30 | 79.33 | 3.10 |
| 1970 | 1,676.50 | 830.00 | 846.50 | 78.38 | 0.44 |
| 1971 | 2,001.50 | 1,080.00 | 921.50 | 77.86 | -1.26 |
| 1972 | 2,274.70 | 1,239.30 | 1,035.40 | 79.97 | -5.48 |
| 1973 | 2,618.90 | 1,389.90 | 1,229.00 | 76.37 | -14.43 |
| 1974 | 2,998.20 | 1,642.20 | 1,356.00 | 76.99 | -10.83 |
| 1975 | 3,835.90 | 1,587.70 | 2,248.20 | 75.15 | -1.11 |
| 1976 | 4,904.80 | 1,803.60 | 3,101.20 | 78.02 | 24.74 |
| 1977 | 5,061.50 | 2,299.60 | 2,761.90 | 77.65 | 6.87 |
| 1978 | 6,147.80 | 2,838.00 | 3,309.80 | 80.63 | -3.70 |
| 1979 | 6,683.80 | 3,151.70 | 3,532.10 | 69.43 | 9.72 |
| 1980 | 7,666.60 | 3,668.40 | 3,998.20 | 63.76 | 5.96 |
| 1981 | 9,005.10 | 4,359.70 | 4,645.40 | 58.64 | 25.55 |
| 1982 | 11,069.80 | 5,467.80 | 5,601.00 | 73.29 | 9.96 |
| 1983 | 12,945.00 | 5,867.70 | 7,077.30 | 73.65 | -4.54 |
| 1984 | 13,783.60 | 6,089.00 | 7,694.60 | 68.94 | 43.79 |
| 1985 | 20,154.20 | 9,852.00 | 10,302.20 | 64.90 | 152.14 |
| 1986 | 30,325.60 | 15,557.40 | 14,768.20 | 29.04 | 133.72 |
| 1987 | 68,890.00 | 29,541.00 | 39,349.00 | 8.08 | 485.34 |
| 1988 | 182,862.00 | 65,810.00 | 117,052.00 | 20.87 | 82.21 |
| 1989 | 287,186.00 | 92,526.00 | 194,660.00 | 33.07 | 21.37 |
| 1990 | 449,923.00 | 115,271.00 | 334,652.00 | 26.68 | 39,94 |
| 1991 | 689,405.00 | 202,007.00 | 487,398.00 | 32.42 | 33.54 |
| 1992 | 1,199,399.00 | 393,537.00 | 805,862.00 | 31.86 | 84.52 |
| 1993 | 1,213,887.00 | 498,918.00 | 714,969.00 | 32.34 | 1.67 |
Dollar currency in circulation in Lebanon serves as a store of value, a unit of account, and a medium of exchange, and the economy has, particularly since 1985, been operating with extremely high levels of both foreign currency in circulation and foreign deposits in the domestic banking system. However, in gauging the substitulability between Lebanese pounds and U.S. dollars, data are available only for bank deposits denominated in these two currencies, and for Lebanese pound currency holdings. Dollar currency in circulation in Lebanon is unobservable; no official data exist, and individuals are unlikely to respond accurately to queries about their dollar holdings.30
Currency substitution is defined here as the use of foreign currencies as a means of exchange; dollarization is defined here as the use of U.S. dollars as the unit of account and the use of dollar-denominated assets as a store of value (Calvo and Vegh (1992)). Giovannini and Turtelboom (1992) have discussed the factors determining currency substitutability in the context of the unit of account, store of value, and medium of exchange functions of money. Currency substitution in the Middle East has been previously examined for Egypt and the former Yemen Arab Republic by El-Erian (1988) and Elkahif and Kubursi (1991), and for Lebanon by Osseiran (1987).
Origins and Extent of Dollarization
The use of foreign currencies in any given economy varies positively with high and variable domestic inflation rates and uncertainty over domestic economic policies, both of which serve to lower the relative demand for domestic fiat money. Indeed, Melvin (1988) characterizes dollarization as a demand-side monetary reform, where economic agents substitute away from an unstable domestic monetary unit.
The extent of dollarization (defined as the ratio of foreign currency deposits (FCD) to total deposits (TD), FCD/TD) and the ratio of Lebanese pound currency holdings to Lebanese pound deposits (LL/LLDEP) over the period 1964–93 are documented in Chart 4. Three points stand out regarding the LL/LLDEP ratio: (1) the flight to liquidity in the wake of the banking crisis of 1967–68, which raised the ratio from 25 percent in 1965 to an average of about 39 percent between 1967 and 1969; (2) a jump in the ratio to 49 percent in reaction to the early days of the civil war in 1975 and 1976, declining thereafter to a low of 15 percent in 1985; (3) further war-induced jumps in the ratio in 1987 (to 34 percent) and 1990 (to 36 percent), with a rapid decline in the ratio in 1993 to near historically low levels (15 percent), as the population continued to abandon the Lebanese pound in favor of the U.S. dollar as a medium of exchange.
In the face of protracted hostilities, accelerating domestic inflation, and exchange rate devaluation, the preferred response of economic agents was to make dollar-denominated deposits the chosen store of value. This move (as shown by the FCD/TD ratio) is noticeable for the period 1978–81 and, more particularly, 1985–89 (Chart 4). While the FCD/TD ratio rose to 42 percent in 1981 and peaked at 92 percent in 1987, it remained high at about 67 percent during 1991—93. The Lebanese people appear to be viewing the continued run of primary fiscal deficits and the accompanying buildup of domestic debt as leading indicators of future rounds of deficit monetization, inflation, and depreciation. Chart 5 reveals the close link between the FCD/TD ratio and movements in the exchange rate, as the Lebanese moved their wealth out of pound-denominated assets into dollars and dollar-denominated domestic and foreign bank deposits. However, it should be noted that in measuring movements in the FCD/TD ratio, an upward bias could arise in the presence of a depreciation of the Lebanese pound. Conversely, a downward bias could be imparted to the extent that foreign currency deposits are moved offshore.
Currency Substitution and Dollarization
(In percent)
Source: International Monetary Fund, Intemotional Financial Statistics.Dollarization and Exchange Rate Movements
(In percent)
Source: International Monetary Fund, Intemotional Financial Statistics.Capital Flight and Repatriation
Dollarization and its close cousin, capital flight, are weak and strong forms of the same phenomenon—the desire of economic agents to minimize the risk associated with holding fiat currencies as a store of value. What distinguishes capital flight from dollarization (the reorientation of economic agents' portfolios of domestic assets from local to foreign currencies) is that the former occurs in response to domestically undiversifiable risk that has the potential to alter the value of domestic assets.
During the early 1980s, with continuing uncertainty over the stability of the domestic banking system, capital flight occurred and wealth was moved offshore into cross-border U.S. dollar deposits. As a measure of capital flight and its return to Lebanon, Charts 6, 7, and 8 present data on bank deposits held abroad during 1981–93, comprising the sum of cross-border deposits by Lebanese residents (line 7xrd of International Financial Statistics) and cross-border interbank claims by Lebanese banks (line 8xad of International Financial Statistics).31 Chart 6 presents the value of cross-border deposits in nominal and real (constant 1985) U.S. dollar terms—the rise in such deposits after the war shocks of 1981–82 and 1986–87 is clear, although real cross-border deposits peaked in 1990 and have been declining continually since then.


Real and Nominal Cross-Border Deposits
(In percent)
Source: International Monetary Fund, Intemotional Financial Statistics.1 In constant 1985 U.S. dollars.
Real and Nominal Cross-Border Deposits
(In percent)
Source: International Monetary Fund, Intemotional Financial Statistics.1 In constant 1985 U.S. dollars.Real and Nominal Cross-Border Deposits
(In percent)
Source: International Monetary Fund, Intemotional Financial Statistics.1 In constant 1985 U.S. dollars.While there are similarities between Lebanon and many South American countries in terms of the motivation for capital flight, Lebanese flight differed in that both Lebanese capital and labor departed during the 1980s, whereas South Americans largely moved only their capital offshore during the same period. Similarly, the return of Lebanese capital in the 1990s has been accompanied by net immigration on a much larger scale than is typically observed in South America.
Cross-border deposits as a percentage of Lebanon's foreign exchange reserves (excluding gold) and M2 are given in Charts 7 and 8, respectively, revealing the large jumps in these ratios in 1987 and (less so) in 1990, and declining thereafter. Indeed, at their peak in 1987 the ratio of cross-border deposits to reserves excluding gold and to M2 had reached the strikingly high levels of 2,767 percent and 297 percent, respectively. At the end of 1993, the corresponding ratios were 510 percent and 97 percent. Clearly, the economic and political stability in the period since the Taïf Accord has induced the repatriation of a large portion of the wealth of Lebanese residents that was sent overseas as bank deposits between 1985 and 1990. Although this stability resulted in a return of much of the flight capital, by the end of 1993 it had had relatively little impact on either the share of foreign currency deposits in total deposits or on the continuing use of the U.S. dollar as a means of exchange and a unit of account in Lebanon.


Cross-Border Deposits
(As percent of M2, including foreign currency deposits)
Source: International Monetary Fund, Intemotional Financial Statistics.
Cross-Border Deposits
(As percent of M2, including foreign currency deposits)
Source: International Monetary Fund, Intemotional Financial Statistics.Cross-Border Deposits
(As percent of M2, including foreign currency deposits)
Source: International Monetary Fund, Intemotional Financial Statistics.

Cross-Border Deposits
(As percent of M2, including foreign currency deposits)
Source: International Monetary Fund, Intemotional Financial Statistics.
Cross-Border Deposits
(As percent of M2, including foreign currency deposits)
Source: International Monetary Fund, Intemotional Financial Statistics.Cross-Border Deposits
(As percent of M2, including foreign currency deposits)
Source: International Monetary Fund, Intemotional Financial Statistics.Advantages and Disadvantages of Dollarization
The above analysis reveals that at the end of 1993 only about 32 percent of total deposits at banks were denominated in Lebanese pounds. Given such high existing levels of dollarization and the wide circulation of the U.S. dollar as a medium of exchange, a key issue for the Lebanese authorities concerns the advantages and disadvantages of striving to achieve either full dollarization (such as Panama) or full de-dollarization.
Full dollarization would preclude an independent Lebanese monetary policy but would not be much different from using the exchange rate as a nominal anchor for the economy. Full dollarization would also involve the loss of seigniorage revenues (which would then accrue to the United States), although the base for the inflation tax (holdings of pound-denominated currency) has shrunk considerably since 1975, as noted above. Table 9 shows the expected increase in the rate of the inflation tax (defined as the ratio of the change in the monetary base to GDP) owing to the collapse of the inflation tax base (see Fischer (1982)). The average annual revenue yield from the inflation tax in Lebanon during 1965–74 was 4.12 percent of GDP; during 1975–86, 8.83 percent of GDP; during 1987–93, 7.63 percent of GDP; and 6.91 percent of GDP for the full 1965–93 period. The inflation tax yield jumped dramatically in 1976 and 1982, years that were associated with heightened war activities and an accompanying decline in output; as a result of Arab aid under the Tunis Agreement, the tax was actually negative in 1979.
Average Inflation Tax

The ratio of the change in base money to nominal GDP.
Average Inflation Tax
| Base Money | Change in Base Money | Nominal GDP | Average Inflation Tax1 | |
|---|---|---|---|---|
| Year | (In millions of Lebanese pounds) | (In percent) | ||
| 1965 | 659 | 55 | 3,523 | 1.56 |
| 1966 | 834 | 175 | 3,867 | 4.53 |
| 1967 | 945 | 111 | 3,820 | 2.91 |
| 1968 | 1.127 | 182 | 4,273 | 4.26 |
| 1969 | 1,172 | 45 | 4,565 | 0.99 |
| 1970 | 1.199 | 27 | 4,866 | 0.55 |
| 1971 | 1.462 | 263 | 5,399 | 4.87 |
| 1972 | 1.643 | 181 | 6,365 | 2.85 |
| 1973 | 1.877 | 234 | 7.103 | 3.29 |
| 1974 | 3.131 | 1,254 | 8,137 | 15.41 |
| 1975 | 3,172 | 40 | 7,500 | 0.54 |
| 1976 | 4,229 | 1,057 | 4,099 | 25.79 |
| 1977 | 4,751 | 522 | 8.199 | 6.37 |
| 1978 | 5,384 | 633 | 8,799 | 7.20 |
| 1979 | 5.209 | -175 | 11,150 | -1.57 |
| 1980 | 6,197 | 988 | 14.000 | 7.06 |
| 1981 | 7,156 | 959 | 16.800 | 5.71 |
| 1982 | 10,147 | 2,991 | 12,599 | 23.74 |
| 1983 | 11.584 | 1.437 | 16,573 | 8.67 |
| 1984 | 13,170 | 1,586 | 28.171 | 5.63 |
| 1985 | 18.642 | 5,472 | 59,329 | 9.22 |
| 1986 | 25,138 | 6.496 | 108.096 | 6.01 |
| 1987 | 57,706 | 32,568 | 740,743 | 4,40 |
| 1988 | 195,952 | 138.246 | 1.356,000 | 10.20 |
| 1989 | 332.038 | 136.086 | 1,350.000 | 10.08 |
| 1990 | 509,070 | 177,032 | 1,973,000 | 8.97 |
| 1991 | 800,051 | 290,981 | 4,132.000 | 7,04 |
| 1992 | 1,521.861 | 721.810 | 9,499.000 | 7.60 |
| 1993 | 2,190,136 | 668.275 | 3,122.000 | 5.09 |
The ratio of the change in base money to nominal GDP.
Average Inflation Tax
| Base Money | Change in Base Money | Nominal GDP | Average Inflation Tax1 | |
|---|---|---|---|---|
| Year | (In millions of Lebanese pounds) | (In percent) | ||
| 1965 | 659 | 55 | 3,523 | 1.56 |
| 1966 | 834 | 175 | 3,867 | 4.53 |
| 1967 | 945 | 111 | 3,820 | 2.91 |
| 1968 | 1.127 | 182 | 4,273 | 4.26 |
| 1969 | 1,172 | 45 | 4,565 | 0.99 |
| 1970 | 1.199 | 27 | 4,866 | 0.55 |
| 1971 | 1.462 | 263 | 5,399 | 4.87 |
| 1972 | 1.643 | 181 | 6,365 | 2.85 |
| 1973 | 1.877 | 234 | 7.103 | 3.29 |
| 1974 | 3.131 | 1,254 | 8,137 | 15.41 |
| 1975 | 3,172 | 40 | 7,500 | 0.54 |
| 1976 | 4,229 | 1,057 | 4,099 | 25.79 |
| 1977 | 4,751 | 522 | 8.199 | 6.37 |
| 1978 | 5,384 | 633 | 8,799 | 7.20 |
| 1979 | 5.209 | -175 | 11,150 | -1.57 |
| 1980 | 6,197 | 988 | 14.000 | 7.06 |
| 1981 | 7,156 | 959 | 16.800 | 5.71 |
| 1982 | 10,147 | 2,991 | 12,599 | 23.74 |
| 1983 | 11.584 | 1.437 | 16,573 | 8.67 |
| 1984 | 13,170 | 1,586 | 28.171 | 5.63 |
| 1985 | 18.642 | 5,472 | 59,329 | 9.22 |
| 1986 | 25,138 | 6.496 | 108.096 | 6.01 |
| 1987 | 57,706 | 32,568 | 740,743 | 4,40 |
| 1988 | 195,952 | 138.246 | 1.356,000 | 10.20 |
| 1989 | 332.038 | 136.086 | 1,350.000 | 10.08 |
| 1990 | 509,070 | 177,032 | 1,973,000 | 8.97 |
| 1991 | 800,051 | 290,981 | 4,132.000 | 7,04 |
| 1992 | 1,521.861 | 721.810 | 9,499.000 | 7.60 |
| 1993 | 2,190,136 | 668.275 | 3,122.000 | 5.09 |
The ratio of the change in base money to nominal GDP.
The inflation tax remained high during 1986–93, owing to the large increase in currency substitution and the authorities' consequent need to raise the inflation tax rate by increasing the supply of base money. Lebanon's heavy reliance on the inflation tax has been slow to decline, largely because of the slow pace at which expenditure reductions and nonseigniorage revenue increases have occurred. This result also supports work by Khan and Ramirez-Rojas (1986) and Vegh (1989), who showed that the greater the incidence of currency substitution, which raises the elasticity of real money demand, the higher the optimal inflation tax for a given level of government spending. Moreover, the greater the use of foreign money, the stronger the inflationary impact of the monetization of any given fiscal deficit.
It should be noted that full dollarization is unlikely to be a useful substitute for fiscal discipline because it removes the “lender of last resort” role of the central bank. This role is likely to become important if commercial banks are pressured to lend to the Government to make up for forgone seigniorage revenues. On the positive side, full dollarization would, in principle, allow Lebanon to enjoy the interest rates and inflation rates of the United States, which would make it easier to finance the reconstruction efforts required to rehabilitate the physical capital stock of the economy.32 It would also lower the risk premium attached to domestic interest rates, which would in turn lower the financing costs of the existing internal debt.
What is the likely effect of dollarization on the success of a stabilization program? It can be argued that, by shrinking the inflation tax base, dollarization actually facilitates stabilization. It does this by minimizing the authorities' incentive to exercise their monopoly powers over the issuance of fiat money by monetizing fiscal deficits.33 Bufman and Leiderman (1993) find that even small increases in dollarization can have a significant impact on seigniorage revenues. This argument emanates from those who argue that competition among fiat currencies in an economy is beneficial because it minimizes the incentive for governments to engage in inflation-inducing policies.
Within the context of a flexible exchange rate system, a case has been made for currency substitution and dollarization as a beneficial outcome, as both induce price competition in the Lebanese economy between local and foreign currencies as stores of value and mediums of exchange: relatively greater use will be made of the low-inflation currency (see Girton and Roper (1981)). However, it is likely that too great a degree of currency substitution will make the use and maintenance of the exchange rate as a nominal anchor unsustainable by raising the volatility of the rates of growth of the currencies required to maintain rate parity (see Woodford (1990)).
Policies to Achieve, and Consequences of, De-Dollarization
In the few countries where de-dollarization has occurred (e.g., Bolivia, Peru, and Mexico in the 1980s), it has been involuntary. Typically, it entails confiscating dollar-denominated assets by exchanging them into the (usually inflating) domestic currency at less than the prevailing market exchange rate. In this sense, advocating a reduction in currency substitution as a policy goal is analogous to attacking the symptom of a disease rather than its cause, which, in Lebanon, is its fiscal deficits.
A government desiring to entice its citizens to voluntarily reverse the extent of foreign fiat currencies in circulation, the share of foreign currency deposits in total deposits, and the share of assets they hold abroad will need to satisfy several preconditions. First, the expected real return on assets denominated in Lebanese pounds will need to be raised or, more accurately, the spread between Lebanese pound borrowing and lending rates reduced.34 Second, this increase in relative returns will need to be brought about by a fall in the stock of domestic money.35 Third, fiscal and monetary policies need to be consistent and credible to dissuade the private sector from engaging in currency substitution. Fourth, expectations of future exchange rate movements should induce a shift in the composition of individuals' desired portfolios of domestic and foreign assets.
In the absence of a significant further reduction of the fiscal deficit, the economic cost of de-dollarization may be large if it raises real domestic interest rates to such an extent that it dampens investment and domestic activity. This is likely to be an important consideration in a Lebanese economy that is trying to rebuild its war-damaged capital stock. Without a tight fiscal stance, a clear potential conflict exists between the simultaneous achievement of the authorities’ goals of de-dollarization and rapid economic reconstruction. The low real interest rates, which would aid reconstruction, contrast with policies required to reassert the pound as the currency of choice in Lebanon.
The absence of reserve requirements on foreign currency deposits provides an incentive for continued dollarization—lowering the reserve requirements on Lebanese pound deposits and imposing a reserve requirement on U.S. dollar deposits would achieve the twin aims of providing the authorities with an instrument to control the dollar money multiplier and raising the relative attractiveness of Lebanese pound deposits. Although a ceiling is imposed on the amount of dollar-denominated deposits that banks can on-lend—a measure that potentially dampens growth in the money supply—it does little to diminish the relative attractiveness of U.S. dollar deposits.
Most important, provision of a stable macroeconomic and financial environment would aid the reflow of wealth into pound-denominated assets, and the Lebanese authorities have made a useful start in this area. In this sense, much of the discussion over the appropriateness of full dollarization or de-dollarization places the monetary cart before the fiscal horse. Clearly, it is very important that Lebanon's large fiscal deficits, as the root cause of its inflation, depreciation, currency substitution, and dollarization, be reduced substantially and as quickly as possible.
See Wilson (1992) for an analysis of the difficulties U.S. monetary authorities have in estimating the extent and direction of cross-border movements of dollars.
Capital flight could take forms other than bank deposits in the new host country, such as property assets and nondeposit financial assets, but it would be reasonable to assume that Lebanese residents would want to keep a large proportion of their foreign-based wealth relatively liquid to facilitate its repatriation to Lebanon when the economy stabilized. Note that the International Financial Statistics cross-border deposit data do not take account of transactions where deposits of Lebanese residents are held in the name of a foreign resident in a foreign bank. For the above reasons, the data on foreign currency deposits are probably a lower bound of the holdings of foreign money balances by Lebanese residents (see Agénor and Khan (1992)).
There would still be a margin reflecting country- and/or region-specific risks.
However, there is unlikely to be a complete guarantee that a system of full dollarization will not be reversed in the future (see the case of Liberia in the 1980s).
However, increasing returns to pound-denominated deposits will only temporarily increase the demand for the pound in the absence of a solution to Lebanon's fiscal disequilibrium.
Such a fall is unlikely at a time of de-dollarization. However, an increase in relative returns could be achieved with a rate of growth of demand for domestic currency greater than the rate of growth of its supply.

