Eight of the eleven countries that experienced chronic high inflation between 1960 and 1990 were in Latin America (Easterly, et al. (1995)). The inflation experience of these economies is well documented and has been analyzed extensively; see, for instance, Bruno, et al. (1991), Bruno (1993), Calvo and Vegh (1994), Dornbusch, et al.(1990), Dornbusch and Edwards (1991, 1995), Heymann and Leijonhufvud (1995) and Vegh (1992).
Some of the most representative studies of this literature are probably those of Ortiz (1983), Canto (1985), Ramirez-Rojas (1985), Bepkerman (1987), Canto and Nickelsburg (1987), Fasano-Filho (1987), Marquez (1987), Melvin (1988), Agenor and Khan (1992), Clements and Schwartz (1992), Rojas-Suarez (1992) and Savastano (1992). For a complete list of references, see the recent surveys by Calvo and Vegh (1992), Claasen and de la Cruz (1993) and Giovannini and Turtelboom (1994).
In Debelle and Fischer (1994) Sg is set equal to zero. No conclusions are changed, however, some of the results below are more analytically tractable.
The central bank actually controls the money stock, which is assumed to map directly into the inflation rate as shown below.
For a more comprehensive and systematic discussion of the analytical issues raised by dollarization in developing countries see Calvo and Vegh (1992) and Giovannini. and Turtelboom (1994).
Restrictions on the holding of FCD were eased (eliminated) in Argentina in the fourth quarter of 1978 (at the start of the “Tablita” episode); in Bolivia in October 1973 (nine months after a 68 percent devaluation ended a 16-year fixed exchange rate peg); in Mexico in March 1977 (following the 37 percent devaluation of September 1976); in Peru in the first quarter of 1978 (following a 3-month float that led to a 60 percent depreciation of the sol); and in Uruguay in October 1974 (when a 70 percent devaluation marked the start of a comprehensive liberalization program). For detailed analyses of these episodes, see Corbo and de Melo (1987), Edwards (1989), Ortiz (1983), Ramos (1986) and Savastano (1992).
FCD were declared inconvertible in Bolivia and Mexico in 1982 (in November and August, respectively) and in Peru in July 1985. In all cases the measure was accompanied by the reimposition of foreign exchange and capital controls that limited severely the issuance of new FCD, and by a large devaluation followed by a short-lived fixed exchange rate. Generally, the outstanding stock of FCD was forcedly converted into domestic currency at an exchange rate much lower than that prevailing in the parallel market. FCD were reallowed in Bolivia In September 1986, in Mexico In December 1985, and in Peru in September 1988. In the last two countries, however, a wide range of exchange and capital controls remained In place until much later, and FCD only regained full convertibility when those controls were abolished–in December 1987 in Mexico, and in August 1990 in Peru.
Analytically, the exclusion of bonds and other nonmonetary assets ensures focusing the attention on the degree of substitutability between “monies”–loosely defined here as those aggregates that provide “liquidity services”. In practice, however, the distinction between, say, a time deposit in foreign currency and a short-term bond denominated in foreign currency may be quite tenuous, especially in a high inflation environment and/or in a context of high capital mobility. In fact, as will be noted in Section III, the difficulties for distinguishing empirically the substitutability between “monies” from a more general set of arbitrage conditions among financial assets of varying degrees of liquidity denominated in different currencies is a perennial source of confusion in the empirical literature on dollarization and CS.
Partly because of these reasons, Lessard and Williamson (1987) consider that the series on foreigners’ deposits by country of origin reported in the U.S. Treasury Bulletin provide only a lower bound estimate of capital flight from Latin America.
Specifically, F*/M* is the ratio of the sum of FCD in domestic banks and foreign currency deposits abroad (FDA) to the sum of M2 and foreign currency deposits at home and abroad. All foreign currency deposits were converted Into domestic currency using the official exchange rate.
Except for foreign currency deposits abroad, all the series were obtained from documents and publications of the central banks of the respective countries. The IFS series on “cross-border bank deposits of nonbanks by residence of depositor” were used as the indicator of FDA from end-1931 onwards. For the earlier period (1970:1-1981:3) these series were spliced with those reported in Table CM-I-4 of the U.S. Treasury Bulletin–except for Bolivia, whose data are not included in the bulletin. An alternative series of FDA, based solely on the data reported in the U.S. Treasury Bulletin, was also used to construct F*. Compared to the spliced series depicted in Figure 1, this alternative estimate of F* was somewhat smaller in absolute terms, but otherwise displayed a remarkably similar behavior over the sample period in the four countries.
The financial repression that prevailed in most of these economies until the late 1970s–and in some cases well into the 1980s–makes it very difficult to obtain (construct) long, consistent and reliable series on more elaborate and appropriate measures of the opportunity cost of holding domestic money (such as the interest rate differential between monetary and nonmonetary assets denominated in foreign and domestic currency). The lack of these data precludes a more precise assessment of the relative influence of assets’ real rates of return on the pattern of dollarization in these countries during the sample period, and complicates even further the task of detecting the potentially different behavior of indicators corresponding to broad and narrow concepts of dollarization.
In Mexico, however, this dollarization measure did not reach the levels of the early 1980s when dollar deposits were reallowed. After reaching a peak of 18 percent in mid-1991, the share of FCD in M3 fell and stabilized around the 11-12 percent range in 1992-93.
For a detailed analysis of these stabilization episodes see Dornbusch (1995) and Kiguel and Liviatan (1995).
However, with the probable exception of Uruguay–whose role as an offshore financial center for the region strengthened significantly in the 1980s–such an omission is unlikely to introduce an important bias to this measure of dollarization. The high degree of macroeconomic instability and financial repression exhibited by these countries until the late 1980s not only did not attract inflows of foreign capital, but fueled recurrent waves of capital outflows (Lessard and Williamson (1987)).
See Savastano (1990), In a related model, Rodriguez (1993) also calls attention to the direct relationship between the share of foreign currency holdings maintained in domestic banks and the stock of international reserves.
Except for Argentina and Bolivia, where the period covered begins in the fourth quarter of 1979 and 1981, respectively.
The parallel market premium is equal to the ratio of the parallel market exchange rate to the official exchange rate, minus unity, and is expressed in percentage terms. The data on parallel market rates were obtained, primarily, from central bank documents and from the World Currency Yearbook (various issues).
Paper prepared for the volume The Macroeconomics of International Currencies: Theory, Policy and Evidence, edited by P.D. Mizen and E.J. Pentecost, Aldershot: Edward Elgar Publishing (forthcoming 1996). I am grateful to Carmen Reinhart, Julio Santaella, Ana C. Savastano, Carlos Vegh and Peter Wickham for valuable comments and discussions. All remaining errors are my own. The views expressed in this paper are solely mine and should not be attributed to the International Monetary Fund.