Paper for the Organization of American States
March 3, 1999
© Peterson Institute for International Economics
The authors are senior fellows at the Institute for International Economics. They benefited from the valuable research assistance of Erika Wada in preparing the tables for this paper. The views are their own opinions and do not reflect the views of the Institute's Board of Directors, Advisory Committee, or staff.
Over time, trade and financial policy issues have become increasingly connected, with the growing density of commercial transactions in the world economy. In this short paper, we highlight the policy intersections for the small economies in the Western Hemisphere—those arguably most concerned with the juncture between trade and finance. Our tables and remarks on small economies in the Caribbean and Central America (collectively referred to as the Caribbean Basin small economies). However, our findings may apply to other small economies in the Western Hemisphere.
Table 1 provides a snapshot of the small economies in the Caribbean and Central America. They range in size from micro island states like St. Kitts-Nevis to more populous Central American countries like Guatemala. Many of the economies have populations under one million; only a few have populations larger than five million. Some countries are the size of a US city; the land mass of others exceeds 100,000 square miles.
Small economies are generally very open: trade in goods and services together may exceed GDP, as illustrated in Table 4. They are also less diversified than larger countries. Many small economies depend heavily on one or a few commodities, or a few services such as tourism and banking, for the bulk of their exports. Some add labor-intensive light manufacturing into their production/export mix. All are "price-takers" in world markets.
Ratios of merchandise trade to GDP demonstrate their reliance on world markets: almost all the countries in Central America and the Caribbean have merchandise trade to GDP ratios above 50 percent; and the unweighted average ratio for this region in 1995 was 76 percent (see Table 4). Their ratios for services trade to GDP are also well above world norms, and on average have risen from about 35 percent in 1980 to about 41 percent in 1995 (again see Table 4).
Small economies are thus highly vulnerable to changes in foreign demand and supply, and to gyrations in foreign capital and exchange rate markets. During the past two years, events in far-flung markets around the world have adversely affected economic conditions in the Caribbean Basin. The sharp depreciation of East Asian currencies has heightened trade competition for small economies that export light manufactures to industrialized markets. Commodity exporters also have suffered as sharp recessions in Asia, and softening demand in the Americas and Europe, have led to continuing weakness in prices of primary commodities produced by small economies.
Equally important, global financial crises (and particularly developments in Brazil since January 1999) have increased the risk premium on investments in nearly all emerging markets, and thereby increased the cost of financing new productive facilities and servicing old debt. Small economies are particularly vulnerable to such foreign shocks, since their limited tax base and low domestic savings make them heavily reliant on foreign funds—be it loans from international financial institutions, workers remittances, portfolio and direct investment, and tourism receipts.
As shown in Table 5, almost all the countries in the Caribbean Basin run current account deficits that range in size from worrisome to extreme. Such large net current outflows require comparable net capital inflows or the draw down of reserves. Attracting these funds from abroad presents an ongoing challenge to national finance ministers and central bank officials—and may require tight monetary policies that also suppress demand in the home market.
Small open economies are not automatically carried forward on the tide of world growth. As the figures in Tables 2 and 3 indicate, real per capita GDP in many of these economies has been virtually stagnant during the past three decades. Several countries have experienced a real decline in per capita incomes (illustrated by a growth factor of less than 1.0 in Table 2). To be sure, real GDP growth has picked up in Central America and a few Caribbean islands in the 1990s, but overall these small economies have not recorded robust growth since the 1980s debt crisis. Special reasons can be cited for individual countries, especially civil wars in Central America. But beyond idiosyncratic explanations, a general disadvantage these countries share is their distance from large markets, their underdeveloped physical infrastructure, and the difficulties individual workers face in acquiring the skills necessary to prosper in the modern economy.
Trade and financial policies can make a positive or negative difference in national economic performance, but of course they are not the whole story. That said, we now survey the main intersections between trade and finance.
In most small economies in the Caribbean Basin, national development strategies, including trade integration initiatives, are constrained by the country's low savings rates (see Table 10) and heavy reliance on trade taxes for current government revenue (see Table 6). Two-thirds of the Caribbean Basin countries received at least 20 percent of their revenues from trade taxes over the period 1994-1996, and several of them relied on tariffs for more than half of their current revenues. These countries fear that lower import tariffs could weaken their fiscal accounts, if alternative sources of revenue are not found. Moreover, gross national savings ("gross" means before any allowance for depreciation of the capital stock) are less than 20 percent of GDP for most Caribbean Basin economies. Hence, most governments cannot rely on thrifty households and corporations to compensate for large public deficits.
For these countries, fiscal reform is a necessary complement to trade liberalization—and represents a major connection between trade and finance. These small economies need to broaden and diversify their tax base (e.g., greater reliance on value-added or general sales taxes), and improve tax collection and administration. Such reforms would not only accommodate new trade liberalization but would also provide greater incentives to savings and investment in the economy.
Exchange rate policy
Broadly speaking, the "north" and "south" poles of exchange rate policy, for a country that maintains a single exchange rate, are fixed and floating rates. (Capital controls and dual exchange rates are discussed later.) Between the polar choices for a single exchange rate, a country can choose among a continuum of six different exchange rate regimes. At one extreme, the country can adopt the currency of another, larger country. Monaco, for example, uses the French franc.
Among countries in the Caribbean Basin, dollarization has been widely discussed in the wake of the Asia/Russia/Brazil crises of 1997-99. Dollarization amounts to outright adoption of the U.S. dollar as the local currency. Of course to some extent U.S. dollars already circulate in the small Caribbean Basin economies. However, the available figures on the relationship between local currency and GDP (Table 9) indicate ratios that are in line with those observed in Mexico and Colombia. The implication is that the use of U.S. dollars in the small economies has not substantially displaced the use of local currency.
Under a dollarization policy, U.S. dollars would be the only currency in circulation. All bank accounts, government bills and bonds, private debts, and other monetary values would be stated in U.S. dollars. The national central bank would continue to regulate private banks and other financial institutions, but it would no longer issue a local currency. And the central bank might hold U.S. dollars and dollar assets of behalf of the government, both as a currency reserve and for investment and precautionary purposes. The central bank might also borrow dollars from the New York financial market. But it would not have the power to print a local currency.
With dollarization, interest rates on government bills and bonds, and interest rates paid on private savings accounts, business debts, mortgages and other financial instruments would closely correspond to U.S. interest rates for assets with the same credit risk. However, the national government would lose its ability to change the exchange rate in response to real shocks at home or abroad—for example, a hurricane, a crop failure, a dearth of tourists, or a financial crisis in Mexico. The government would also lose the ability to finance a public deficit by issuing bonds to the central bank in exchange for fresh local currency. The main benefits of dollarization would be lower inflation rates and lower interest rates over a period of time (these aspects are discussed later).
Slightly less severe than adopting the dollar, the country can enact a currency board. Under this exchange rate system, the currency board issues the national currency (notes and bank deposits) according to a fixed multiple of its own holdings of the designated foreign currency (e.g., 1 to 1, 2 to 1, etc.). When the balance of payments is in surplus, the stock of national currency increases; when the balance of payments is in deficit, the stock decreases. The Argentine currency board, for example, issues pesos 1 to 1 against its holdings of dollars; the Estonian currency board issues krooni 8 to 1 against its holdings of deutsche marks.
Next the country can peg its currency to another currency (or to a basket of currencies) either in a fixed ratio or allowing a small band of fluctuation. But the central bank can adjust the domestic money supply with a view to local economic conditions as well as its holdings of foreign exchange reserves. This approach is widely used among Caribbean countries (see Table 7). It is also used by Hong Kong (which pegs at about 7.8 HK dollars to 1 US dollar) and by Singapore (which pegs to a basket of currencies). As a variant, the country can adopt a crawling peg, in which the central rate moves slowly up or down (usually down) against a foreign currency, with a band of fluctuation allowed on either side. Until January 1999, Brazil's Real Plan envisaged a crawling peg in which the central rate for the real would decline against the dollar by about 7 percent per annum. With this system, the country's central bank has somewhat more freedom to adjust the domestic money supply.
Next, the country can adopt a managed floating rate. Here, the central bank intervenes from time to time to guide the exchange rate, in light of changing circumstances, but it does not advertise a central rate or a target zone (i.e., an adjustable band of fluctuation). With more flexibility in the exchange rate, the central bank has still more freedom to adjust the domestic money supply in light of local conditions. Since the late 1980s, Chile has pursued a managed float against the U.S. dollar.
Finally, the country can pursue a pure float. The central bank intervenes only on rare occasions when markets are "disorderly". For many years, this has been the approach of the Bank of Canada. After the peso crisis of 1994-95, the Banco de Mexico adopted the same approach. With a pure float, the central bank can focus nearly all its attention on domestic price stability and employment levels.
The choice of exchange rate system makes a great deal of difference to overall management of the economy. From an institutional standpoint, the central bank gains greater prominence as the nation's macroeconomic manager as the exchange rate system evolves from the fixed pole to the floating pole. The reason is that the central bank has more discretion to adjust the money supply, up or down, and to finance any public sector deficit by issuing fresh currency, without paying such close attention to the level of foreign exchange reserves.
The extra degree of policy freedom has advantages and disadvantages. The balance between advantage and disadvantage shifts depending on whether the economy is large or small.
Whatever the size of the economy, a movement in the exchange rate will exert a larger and faster effect on prices of traded goods and services than on wages. When the exchange rate depreciates, for example, exporting firms will earn more local currency for a given quantity of exports than previously; accordingly they will shift sales to the export market and mark up local currency prices. Similarly, firms that compete with imports will have more room to raise their local prices, because imports will be more expensive in local currency. However, wages will be slower to rise, because custom and contracts bring stickiness to wage rates.
While similar mechanisms are at play the world over, the pass through impact of exchange rate depreciation on the domestic price level (measured by the CPI) depends a great deal on the size of the economy. For a large economy, like the United States or Japan, movements in the exchange rate are barely detectable in the CPI. The reason is that traded goods and services are modest relative to the size of the overall economy. Firms are slow to adjust their local prices to reflect exchange rate changes because they face considerable competition from firms that have little to do with the international economy. Moreover, non-traded goods and services account for a big share of the economy. Prices in the non-traded sector are barely disturbed by movements in the exchange rate.1
For small economies, however, the exchange rate changes exert a strong pass through impact on the domestic price level. Most of the economy is oriented towards exporting and importing, and therefore exchange rate changes ripple through the entire price structure. The non-traded sector is relatively small, and therefore has little buffering effect.
Put succinctly, the smaller the economy, the more exchange rate depreciation serves to determine the domestic inflation rate. This effect can be seen in Table 8. The average depreciation of the domestic currencies of the small economies (annually over the period 1985-1998) is compared with their annual average "inflation premium". The inflation premium is defined as the annual excess of local inflation over U.S. inflation during this period. We tested the relationship between these variables by running a simple regression. We found that, for the small economies in the Caribbean Basin, the correlation coefficient is 1.10: in other words, domestic inflation in these countries since 1985 rose by 1.1 percentage point for each 1.0 percentage point of currency depreciation (see Regression 2 results).2 Likewise, interest rates on average are higher when the exchange rate depreciates over time. We found that interest rates in the small Caribbean Basin economies on average rise about 0.3 percentage points for each 1.0 percentage point of exchange rate depreciation (see Regression 3 results).
In addition to its year-to-year impact on the price level, the choice of exchange rate system importantly influences the depth of economic integration with the world economy. Consider this fact. Two-way trade between Montreal and Toronto (both within Canada), or between Houston and New Orleans (both within the United States) is approximately 10 to 20 times greater than two-way trade between city pairs across national borders, after allowing for differences of distance, city size, and per capita income. Statistics on services trade and capital flows within and between countries are far less available than statistics on merchandise trade. However, it seems probable that investment and service transactions between city pairs within a country are at least 10 times as great as transactions across borders, after allowing for other differences. There are many reasons that commerce within national borders is more intense. Among them are common laws and customs, and the ease of dealing between two businessmen with similar backgrounds. However, a common currency is part of the equation. This is illustrated by Regression 4, which indicates that, for the small Caribbean Basin economies, a fixed dollar peg has been associated with a higher services trade to GDP ratio.
Intense commercial relations are an important reason that per capita income converges upward in poor areas toward the level attained in rich areas, either within a country or between countries that are linked in an economic union (as in Europe). Various econometric studies indicate that income convergence over a long period closes the gap at a rate of about 0.5 percent per year.3 Income convergence does not require a common currency, but it is probably accelerated by a common currency. Regression 1 suggests that Caribbean Basin countries that pegged their currency to the U.S. dollar enjoyed significantly higher per capita income growth over the period 1970 to 1997 than Caribbean Basin countries that adopted various types of flexible exchange rate regimes. The explanation for this finding, of course, is that pegged exchange rates were strongly associated with lower inflation and interest rate premiums.
Other factors are also at play in raising or lower per capita income performance, and some of them are also reflected in Regression 1. Countries with fast population growth tended to have lower per capita income growth (with a coefficient on the margin of statistical significance). Countries with higher gross national savings rates in the 1990s, high slightly higher growth. The coefficient lacks statistical significance, but this may reflect the poor quality of savings data, and the limited period of coverage (see Table 10). The coefficient on the readiness indicator also has no statistical significance, but the indicator (Table 6) only reflects conditions over the last three years of the three-decade period of per capita income growth. In any event, it should be emphasized that the variables in Regression 1 explain only 50 percent of the country-to-country variance in growth performance over the period 1970-1997. This means that variables not indicated, along with pure good luck or bad luck, are also responsible for large differences in growth performance.
Capital controls and dual exchange rates
Some countries, including some of the small economies in the Caribbean Basin, use dual exchange rates to distinguish between capital flows and transactions in goods and services. Other countries, such as Chile, use different types of controls (for example, zero-interest reserve requirements for six months or a year) to regulate the inflow of capital. Chile's idea was to avoid an unwanted appreciation of the peso exchange rate (and possibly a "bubble" in its equity and real estate markets) created by an influx of "hot money" when world financial markets thought highly of its policies and performance. Conversely, Chile hoped to avoid a crash when sentiment changed.
Dual exchange rates and capital controls almost always give preference to merchandise trade, tourism and business services. In some cases, as with Chile, the intent is to retard the flow of "hot money" in and out of the country. In other cases, the intent is to permanently discourage capital movements and associated financial transactions (e.g., payment of interest, dividends and royalties).
Temporary capital controls, along the lines of the Chilean model, have acquired new esteem, or at least a fresh look, in the wake of the Asia/Russia/Brazil financial crises. However, dual exchange rates, or permanent capital controls, attract skepticism both because they act as a magnet for regulation and corruption and because of their negative impact on trade flows.
The regulation/corruption point is straightforward. When a country operates two different exchange rates, or controls flows, it creates huge incentives for private parties to make easy money by crossing the barrier. These incentives erupt in false invoices, phony transactions, and outright bribery. In turn, a regulatory bureaucracy must be created—usually in the central bank—to monitor all foreign exchange transactions.
The negative impact on trade arises in more subtle ways. It comes about because capital controls (and associated restrictions on interest, dividend and royalty payments) tend to discourage direct investment. Direct investment is critical for the local establishment of service providers—foreign-owned hotels, telecom firms, banks and the like. Direct investment is also essential for the local operations of multinational industrial firms. These multinational firms typically account for 15 to 30 percent of national imports and exports. Over a period of time, trade in goods and services associated with direct investment will be discouraged by dual exchange rates or capital controls.
In December 1997, members of the World Trade Organization signed the Financial Services Agreement (FSA), one of the important continuing negotiations from the Uruguay Round of Multilateral Trade Negotiations. The agreement entered into force on March 1, 1999 (despite the fact that a number of signatories had not yet completed their domestic ratification procedures). Only 6 of the 21 small countries included in our analysis have signed the FSA (see Table 6).
The FSA largely formalizes the status quo. Most of the key countries bound their existing liberalization of financial services, and all signatories agreed to WTO dispute settlement procedures. Very little additional reform was agreed in the key banking sector, and neither industrial nor developing countries opened their markets to any significant degree. A few emerging market economies did, however, liberalize their insurance and securities markets.4
In the wake of the Asia/Russia/Brazil crises, it is important to distinguish between the liberalization of financial services and the free flow of capital ("capital account convertibility" in IMF terminology). Although the two often go together, there is no iron link between them. A country can control the flow of capital, yet allow the unfettered operation of foreign banks, insurance companies and mutual funds. This is roughly the position of Bermuda. Conversely, a country can allow the free movement of capital, yet restrict the operations of foreign financial institutions. This is the position of Singapore and was the position of Mexico until 1999.
Moreover, the liberalization of financial services should not be associated with a financial-free-for-all. Instead, the regulatory regime should be strengthened. This means higher standards for banks and insurance companies (ensuring they have adequate capital, and that they write down non-performing loans), proper fiduciary standards for mutual funds (accurate reporting to shareholders, no skimming through high fees), and honest operation of stock exchanges (listing standards, limits on insider trading). Part of the improvement will be associated with the entry of world-class financial firms; part with more stringent local regulation. However, if financial institutions are allowed to open without regard to their credentials and without close regulation, the sort of debacle experienced in Thailand and other crisis countries will almost certainly ensue.
The benefits of a liberal, competitive, but properly regulated financial system can be substantial. For local households and firms, borrowing costs should be lower, deposit rates higher, and insurance premiums lower. Stock exchanges will be less susceptible to uninformed speculation and insider trading. Among other things, these benefits mean than local firms will find it easier to finance their import purchases and export sales (for example, through letters of credit), and they will have more opportunity to raise capital either through loans from bank and insurance companies, or through stock market issues. Without financial liberalization, local firms will forever operate at a disadvantage compared to foreign companies that have good access to the New York financial market.
To be sure, implementing such policies poses a daunting challenge for small economies, since they often lack the administrative and juridical infrastructure to apply and enforce financial regulations. To accelerate progress on these "second generation" reforms, small economies should seek both technical and financial assistance from regional development banks and national monetary authorities in industrial countries to help train bank supervisors and other financial regulators.
A key input for the provision of first rate financial services is the availability of high quality, low cost telecommunications. Unfortunately, monopoly telecom firms dominate the small Caribbean Basin economies. The absence of competition keeps prices high and restricts the range of services. With new technologies, the marginal costs of linking 10 minutes of voice or data connection between Santo Domingo and New York are not much higher than the marginal costs of linking New York and Washington. But the actual prices charged largely depend on the extent of competition in the local market.
The WTO Basic Telecommunications Agreement, signed in February 1997, is designed to bring meaningful telecommunications competition to all signatory countries.5 Thus, it is discouraging to note that only two Caribbean economies were full participants in the telecom agreement at its inception (see Table 6), though a few additional countries have since signed. An essential adjunct to the liberalization of financial services will be better and cheaper telecom service—and the WTO agreement is a good place to start.
1. The non-traded sector of a large economy acts, however, as a reservoir of resources (capital, labor and raw materials): when the exchange rate depreciates, and the traded sector expands, resources are drawn away; when the exchange rate appreciates, and the traded sector contracts, resources flow in.
2. Regression 2 examines the relationship between the inflation premium and exchange rate depreciation. The coefficient of 1.10 indicates that when the exchange rate is depreciated by 1.0 percentage points, the inflation premium increases by 1.1 percentage points. The sign of the coefficient indicates whether the two variables move in the same or opposite direction (positive or negative, respectively). The t-statistic of 8.7 next to the coefficient indicates that the relationship is statistically significant. The adjusted R-squared of 0.81, at the bottom of the table, indicates that the regression explains 81 percent of the movement in the inflation premium.
4. See Wendy Dobson and Pierre Jacquet, Financial Services Liberalization in the WTO, Institute for International Economics, June 1998. This summary is paraphrased from their press release.
5. See Gary Clyde Hufbauer and Erika Wada, editors, Unfinished Business: Telecommunications after the Uruguay Round, Institute for International Economics, December 1997.