Journal Issue
Share
Article

Exchange Rate Policy Attitudes: Direct Evidence from Survey Data

Author(s):
International Monetary Fund. Research Dept.
Published Date:
August 2008
Share
  • ShareShare
Show Summary Details

As in other areas of public policy, governments’ choices of exchange rate policies are conditioned by the preferences of their constituents. The nominal exchange rate regime and the level of the real exchange rate can have powerful effects on the private sector, and economic agents want government policies that favor them. A government that ignores its constituents’ concerns about the exchange rate will come under pressure to change course. Indeed, some of the most dramatic events in the history and recent experience of exchange rate policy have to do with the preferences of social groups: how they changed, conflict among them, how strong they are. The battle over gold in the late 19th and early 20th century—whether in the United States in the 1890s, the United Kingdom in 1925, or in Latin America throughout—was largely about which groups and sectors of the economy were likely to win, and lose, from being on the gold standard (Eichengreen, 1992; Simmons, 1994; Hefeker, 1995; Frieden, 1997; Broz, 1997). So too did the process of European monetary integration implicate powerful interests on both sides of the issue, as it continues to do in the accession countries of Eastern and Central Europe and the former Soviet Union (Eichengreen and Frieden, 1993; Frieden, 2002). Modern currency crises often begin with a government immobilized by contending demands to sustain a fixed rate and to devalue (Klein and Marion, 1997; Frieden, Ghezzi, and Stein, 2001; Leblang, 2002). Devising a politically viable policy response to exchange-market developments—including real appreciations and attacks on currencies—has proved extraordinarily difficult for governments from Argentina to Russia.

Just as it is important for policymakers to address the attitudes of powerful constituents toward the exchange rate, it is important for scholars to understand what these attitudes are and how they are expressed. Nonetheless, there is very little systematic empirical work on the policy preferences of major economic agents. Although papers by Collins and Giavazzi (1992), Gartner (1997), Gabel (1998), and Scheve (2004) address related macroeconomic topics—differences in individual attitudes toward inflation, the euro, and global economic integration—there is little work that directly addresses the exchange rate attitudes of business owners operating in different sectors of the economy. By contrast, there is a large body of work that examines sectoral and factoral attitudes toward trade policy (Balistreri, 1997; Scheve and Slaughter, 2001; Beaulieu, 2002; and Mayda and Rodrik, 2005).

This paper exploits a large cross-national survey, the World Bank’s World Business Environment Survey (WBES), to try to uncover the relationship between the economic activities of firms and their owners’ and managers’ attitudes toward the exchange rate (World Bank, 2000). The survey tells us only the extent to which corporate respondents regarded the exchange rate as “a problem,” which in itself is of limited use. However, we also know the currency regime prevailing in the country at the time of the survey, the level of the real exchange rate, and many things about the firms in question. By relating the prevailing exchange rate policy and the firm’s economic characteristics to how “problematic” the currency is perceived to be, we can draw inferences about the sources of attitudes toward the exchange rate.

We start with expectations from the theoretical literature on currency policy preferences. With respect to the exchange rate regime, we expect firms with substantial cross-border exposure to be particularly sensitive to currency volatility, and thus to be more satisfied with a fixed exchange rate. With respect to the level of the real exchange rate, we expect firms in the tradables sectors to be more satisfied when the currency is weak and more dissatisfied when it is strong.

We find, indeed, that owners and managers of internationally oriented firms prefer greater stability of the exchange rate: in countries with a floating currency, manufacturers are more likely to report that the exchange rate causes problems for their business than are producers of nontraded goods and services, who typically do not require foreign exchange. With respect to the level of the exchange rate, we find that exporters and manufacturers are more likely to be unhappy following an appreciation of the real exchange rate than are firms in domestically oriented nontradable sectors (services and construction). These findings conform to expectations about the cross-sectoral distributional effects of exchange rates.

I. Exchange Rate Policy, Politics, and Policy Preferences

The exchange rate is centrally important to economic activity, and government policy has a powerful impact on the currency. After all, the exchange rate is the single most important price in any economy, and it is a price that is routinely set, or at least targeted, by many governments. There is an enormous literature on appropriate currency policy, but unlike in the case of trade or fiscal policy, there is no simple welfare benchmark, so that debates typically involve different weightings of the trade-offs inherent in exchange rate policy choices. Supporters of flexibility confront opponents of volatility, but those who value the credibility and predictability of a fixed rate square off against those who dread its rigidity. A strong currency provides a powerful tool against inflation, and boosts national purchasing power; a weak currency gives national producers great incentives to sell into world markets. To paraphrase Jeffrey Frankel (1999), no single currency policy is right for all people.

Exchange rate policy is constrained by contending economic interests and policy preferences, which makes it important to understand these interests and preferences. A theoretical literature deduces attitudes from the distributional consequences of various exchange rate regime arrangements predicted by economic theory (Frieden, 1991), while empirical analyses have imputed attitudes indirectly from actual currency policies and legislative and other voting behavior (Eichengreen, 1995; Frieden, 1997; Frieden, Ghezzi, and Stein, 2001; and Frieden, 2002). But scholars have rarely been able to find ways of directly mapping economic position to exchange rate policy preferences.

There are two relevant dimensions of variation along which preferences may vary: on the regime by which the currency is managed, and on the level of the currency. In the first instance, the issue is whether to float or fix the exchange rate—and if to float, in which of the many possible ways.1 In the second instance, assuming the currency is not fixed, the question is the desired level of the real exchange rate.2 There is always the option to let the currency float completely freely, although developing countries have shown themselves reluctant to do this (Calvo and Reinhart, 2002; Levy Yeyati and Sturzenegger, 2005). On both dimensions, the attitudes of economic agents are likely to differ.

Attitudes Toward the Exchange Rate Regime: Stability and Credibility or Policy Flexibility?

In an open economy, the main advantage of a fixed-rate regime is to lower exchange rate risk and transactions costs that can impede international trade and investment.3 Volatile exchange rates create uncertainty about international transactions, adding a risk premium to the costs of goods and assets traded across borders. In addition, an exchange rate peg can enhance monetary-policy credibility. Theory and evidence suggest that fixing the exchange rate to the currency of a low-inflation country both promotes international trade and investment and disciplines monetary policy by providing an observable nominal anchor.4

But fixing the exchange rate requires that the government sacrifice its capacity to run an independent monetary policy. A floating exchange rate, on the other hand, has the great advantage of allowing a government to pursue an independent monetary policy. This independence provides flexibility to accommodate foreign and domestic shocks, including changes in the terms of trade and world financial conditions, and to affect the competitiveness of (relative prices faced by) the tradable goods sector.

In an open economy, then, economic agents confront a trade-off between two competing sets of values. On the one hand, a fixed rate brings stability and credibility, and on the other, it sacrifices flexibility.

The different valuation of this trade-off depends largely on the extent to which the economic actors are engaged in international economic activity (Frieden, 1991; Hefeker, 1997). Those heavily involved in foreign trade and investment—typically including exporters, international investors, the commercial and financial sectors, and those with substantial foreign currency liabilities—should favor exchange rate stability, because currency volatility is an everyday concern that makes their business riskier and more costly. By the same token, these groups care less about a loss of national monetary autonomy, because they typically do business in several countries, and can shift business or assets abroad if domestic conditions become unfavorable.

By contrast, groups whose economic activity is confined to the domestic economy benefit from a floating regime. The nontradables sectors (services, construction, transport) belong in this camp. They are not required to deal in foreign exchange and so are free of the risks and costs of currency volatility. They are highly sensitive to domestic macroeconomic conditions and thus favor the national autonomy made possible by floating.

Tradables producers may have reasons to oppose a fixed rate, as it eliminates the possibility of a depreciation to maintain or restore the competitiveness of tradables producers. This raises the second issue: the level of the real exchange rate.

Attitudes Toward the Level of the Exchange Rate

The real exchange rate affects the relative price of traded goods in both local and foreign markets. There is no clear economic efficiency argument for or against any particular level. A strong (appreciated) currency gives residents greater purchasing power, but also entails a loss of competitiveness for tradables producers. A real appreciation benefits consumers of imports and harms producers of goods that compete with imports (and exporters). So tradable (import-competing and exporting) industries lose from a currency appreciation, but domestically oriented (nontradable) industries and domestic consumers gain.

Of course, a real depreciation has the opposite effects, stimulating demand for locally produced tradable products, but raising the prices that consumers pay for foreign goods and services. Currency depreciations help exporting and import-competing industries at the expense of domestic consumers and producers of nontraded goods and services.

Thus the level of the exchange rate, too, involves two competing goals—stimulating local tradables producers and raising local purchasing power. The benefit of increasing the competitiveness of national producers comes at the cost of reducing the real income of national consumers and vice versa.

Two issues complicate theories of exchange rate policy preferences. First, exporters are likely to be torn between a concern for currency stability, on the one hand, and a concern for a favorable level of the exchange rate, on the other. These two concerns conflict, inasmuch as a fixed rate rules out adjusting the nominal exchange rate to improve the competitive position of exporters. Whether exporters favor stability over competitiveness, or vice versa, is likely to depend on such factors as the price sensitivity of consumers of exports, the ability of exporters to hedge against currency volatility, and so on. Second, and closely related, is the fact that tradable producers’ concern about currency movements depends upon how directly they are affected by changes in the exchange rate, which is a function of such things as pass-through, currency invoicing, and the importance of imported inputs (Campa and Goldberg, 1997 and 2005; Devereux and Engel, 2002). Generally, both exporters and import-competers in industries with high pass-through are more sensitive to the relative price effects of currency movements than those with low pass-through, because their prices respond more directly to changes in exchange rates. The policy preferences of exporters are thus likely to be contingent on a large number of factors. In both instances, the extent to which concern about volatility is more important than concern about the level of the exchange rate is largely an empirical question. A simplified picture of the expected preferences of firms would include:

  • Internationally exposed firms, including tradables producers, will prefer greater currency stability, hence a fixed exchange rate.
  • Tradables producers will prefer greater competitiveness, hence a depreciated currency.

In what follows, we assess the empirical relevance and accuracy of these expectations in a large, cross-national survey of firm managers and owners.

II. Data and Methods

To analyze attitudes toward the exchange rate, we use data from the World Bank’s WBES.5 The WBES was administered to owners and managers of over 10,000 firms in 80 countries in 1999, applying a common survey instrument to a representative sample of at least 100 firms in each country. We look at individual responses to the following question: “How problematic is the exchange rate for the operation and growth of your business?” Responses varied along the following ordered scale: 1 = no obstacle, 2 = minor obstacle, 3 = moderate obstacle, 4 = major obstacle. These individual responses represent the dependent variable of our analysis, which we will refer to as Exchange Rate Problem.

The variable Exchange Rate Problem exhibits significant variation both across and within countries. The average value of Exchange Rate Problem is 2.59, which indicates that the exchange rate represents a nontrivial con sideration for the average business. The average response in El Salvador of 2.58 is closest to the overall mean response. Hungary represents the lowest country average (1.61); managers in Ecuador, by contrast, were most likely to view the exchange rate as an obstacle to their business (3.74). This is understandable since Ecuador suffered a serious economic crisis in 1999. On the brink of hyperinflation and immersed in a deep financial crisis, Ecuador abandoned its currency in January 2000 and adopted the U.S. dollar as its legal tender. The average response in Thailand is also notably high at 3.63, which reflects the turmoil caused by the collapse of the baht in 1997 and the subsequent Asian financial and currency crises. The overall standard deviation of Exchange Rate Problem is 1.16, the between-country standard deviation is 0.60, and the within-country standard deviation is 1.00. Table 1 reports overall summary statistics. Tables 2a and 2b are the correlation matrices for our firm- and country-level data.

Table 1.Summary Statistics
VariableNMeanSDMinMax
Exchange Rate Problem75962.5851.16014
Manufacturing70610.3860.48701
Tradable70610.4540.49801
Agriculture70610.0690.25301
Exporter74230.3600.48001
Government-Owned75960.1200.32501
Size75961.7820.73513
REER Appreciation47–0.0190.115–0.3380.429
Log GDP/Capita677.4021.2824.76410.366
M3/GDP6739.68025.1100.002119.487
Log FDI Stock/Capita675.7721.7211.94610.159
Table 2a.Correlation Matrix(Firm-level variables)
Exchange Rate

Problem
ManufacturingTradableExporterGovernment-

Owned
Size
Exchange Rate
Problem1
Manufacturing0.0351*1
Tradable0.0464*0.8676*1
Exporter–0.0369*0.3333*0.3070*1
Government-Owned–0.0515*0.0594*0.0862*0.0699*1
Size0.00160.2175*0.2385*0.3072*0.2243*1
* indicates statistical significance at the 1 per cent level.
* indicates statistical significance at the 1 per cent level.
Table 2b.Correlation Matrix(Country-level variables)
Exchange Rate

Problem
REER

Appreciation
Log GDP/

Capita
M3/

GDP
Log FDI Stock/

Capita
Exchange Rate
Problem1
REER
Appreciation–0.13251
Log GDP/–0.23350.12821
capita
M3/GDP–0.20150.19550.4619*1
Log FDI Stock/1
Capita–0.29180.17660.8622*0.4474*
* indicates statistical significance at the 1 per cent level.
* indicates statistical significance at the 1 per cent level.

We want to assess how the respondent firms’ sectors, conditioned by the country’s exchange rate regime and the level of the real exchange rate, affect attitudes toward the currency. For this purpose, we need to classify countries by exchange rate regime. We employ two widely used classifications of de facto exchange rate regimes: Levy Yeyati and Sturzenegger (2005) and Reinhart and Rogoff (2004). Although these classification schemes differ in details, both attempt to capture the actual behavior of the exchange rate. Levy Yeyati and Sturzenegger, henceforth “LYS,” categorize countries as floats or pegs according to observed changes in the nominal exchange rate, the volatility of these changes, and the volatility of international reserves.6 Reinhart and Rogoff, henceforth “RR,” exploit the conditional probability of the exchange rate staying within a given range over a rolling five-year window, and use information about parallel (dual market) exchange rates in determining whether a regime continues from one year to the next. Our results are largely robust to these alternative regime classification technologies.

In Tables 3a and 3b, we report all our data by country according to the LYS “floating” and “pegged” exchange rate regime classifications. Table 3c provides a comparison of LYS and RR regimes in 1999, the year the WBES survey was administered. Note that RR classifies Indonesia as having a floating regime in 1999 but LYS do not. We drop Indonesia from the RR float sample because, in the aftermath of the currency crisis firms from all sectors would likely express dissatisfaction with the exchange rate. Indeed, the standard deviation of Exchange Rate Problem in Indonesia is 0.95, while its average value of 3.54 is one of the highest overall. Sectoral differences may be difficult to discern in this environment. A similar logic applies to Thailand and the Philippines, but because LYS and RR are consistent in their classification of these countries as floating exchange rate regimes, inclusion of these countries will affect the two samples in the same way.

Table 3a.Data by Country, Floating Exchange Rate Regime in 1999(LYS classification)
CountryExchange

Rate

Problem
ManufacturingTradable

(Manufacturing +

Agriculture)
AgricultureExporterGovernment-

Owned
SizeREER

Appreciation
Log

GDP/

Capita
M3/

GDP
Log FDI

Stock/

Capita
Albania2.5430.3230.3310.0080.1980.1261.3156.89655.6724.934
Belarus3.0850.2870.6090.3220.2740.2991.8726.96318.4834.745
Canada2.0200.2730.2830.0100.4750.0402.020–0.0129.94276.6508.658
Chile2.5250.4600.4830.0230.4140.0612.010–0.0478.44441.9647.958
Colombia3.2930.3810.4120.0310.3940.0202.343–0.0937.63334.2175.781
Czech

Republic
2.2860.1980.2300.0320.3280.1671.421–0.0188.54966.4697.442
Guatemala3.5530.4060.4060.0000.3200.0001.8357.40523.4385.677
Haiti2.8820.4290.4940.0650.2160.0291.6866.13532.8353.258
Honduras3.2640.5140.5290.0140.3300.0001.6926.82739.5175.252
Hungary1.6140.2770.3760.0990.3940.1981.5350.0228.29047.0597.727
Kenya1.7820.4530.5280.0750.8370.1642.2736.04239.8863.296
Madagascar2.2450.5240.6430.1190.4490.0201.7555.44323.9392.833
Mexico3.1820.5250.5250.0000.3940.0002.0200.0938.57332.4506.673
Pakistan2.9040.5050.5160.0110.4780.0641.851–0.0686.24846.7333.951
Peru3.0000.3860.4820.0960.2450.0092.0667.61128.1005.948
Philippines3.4500.4500.4500.0000.3200.0201.9500.0586.85163.7975.037
Poland2.2140.3060.3980.0920.4500.2091.675–0.0248.25534.7626.513
Russia3.1160.2740.4980.2240.0790.1441.704–0.3137.36520.2881.946
South Africa2.4340.4680.5320.0650.8860.0602.542–0.0538.00442.9377.048
Sweden1.7730.2890.2990.0100.5570.0821.742–0.0280.09648.4059.020
Tanzania2.2000.3440.3750.0310.2670.1431.7435.51121.2074.263
Thailand3.6340.5350.5760.0410.6450.0061.7440.0367.611101.9746.236
Turkey2.8300.4380.4580.0210.4110.1701.7350.0517.95138.8505.602
Ukraine3.0510.3790.4440.0650.2390.1641.617–0.2126.41713.9124.174
United States1.6320.2560.2910.0350.3410.0921.897–0.01010.36659.3348.131
Table 3b.Data by Country, Pegged Exchange Rate Regime in 1999(LYS classification)
CountryExchange

Rate

Problem
ManufacturingTradable

(Manufacturing

+ Agriculture)
AgricultureExporterGovernment-

Owned
SizeREER

Appreciation
Log GDP/

Capita
M3/

GDP
Log FDI

Stock/Capita
Argentina1.8110.3220.3560.0340.2950.0321.8740.0898.95525.9887.436
Armenia2.7900.2020.2580.0560.0780.2181.3950.0096.2609.1605.100
Belize1.7140.4120.4120.0000.2240.0821.388–0.0327.97548.4497.048
Bosnia and Herzegovina1.2110.5000.5630.0630.6060.3161.7476.76425.2314.205
Botswana1.2630.3580.3960.0380.5190.2111.8957.86224.0966.683
Bulgaria2.4020.5130.6500.1370.3450.3421.5730.0257.27447.4155.697
Cameroon2.2970.3780.4860.1080.6110.0541.973–0.0136.45014.1784.277
China1.7930.6090.6200.0110.3590.2391.804–0.0546.649116.6405.011
Côte d’Ivoire2.0000.3130.3540.0420.7390.0191.923–0.0096.46423.5494.754
El Salvador2.5780.5430.5570.0140.2870.0201.9127.61644.9035.687
Estonia1.8080.3750.3920.0170.5950.2081.6500.0408.11928.2457.490
Germany1.6950.1890.2000.0110.3580.0741.916–0.0329.98373.3347.959
Ghana2.5000.2500.3960.1460.4550.1401.980–0.0045.46822.6064.220
Lithuania1.8210.1890.2260.0380.2820.0571.1790.0777.97119.3396.373
Malawi2.2730.3180.3640.0450.6500.0912.0000.0175.01019.0583.296
Malaysia1.9770.5060.5290.0230.3790.0341.6670.0098.214119.4877.686
Namibia1.9650.1960.3040.1070.6800.0881.9657.47244.1016.633
Panama1.3000.5340.5690.0340.4630.0382.2758.20964.5787.794
Senegal2.1670.1360.1820.0450.5000.0001.3335.99823.2284.331
Singapore1.8200.3200.3200.0000.4600.0401.930–0.0689.908117.53510.159
Slovak Republic2.3360.2610.3530.0920.4780.1851.521–0.0268.14661.9516.043
Trinidad and Tobago2.3960.4580.5140.0560.4260.0691.6440.0178.61356.6358.506
Uganda1.8030.2750.3480.0720.4710.0851.549–0.0865.40613.7523.332
Zambia1.9500.2160.3510.1350.3420.1001.7500.0055.71017.7265.361
Zimbabwe3.0000.3430.4570.1140.6320.0142.0006.47641.3354.443
Table 3c.Comparison of Levy-Yeyati and Sturzenegger (LYS) and Reinhart-Rogoff (RR) Regimes, 1999
LYS FloatsRR FloatsLYS PegsRR Pegs
CountryER problemCountryER problemCountryER problemCountryER problem
Hungary1.614Hungary1.614Bosnia and Herzegovina1.211Bosnia and Herzegovina1.211
United States1.632United States1.632Botswana1.263Panama1.300
Sweden1.773Sweden1.773Panama1.300Germany1.695
Kenya1.782Kenya1.782Germany1.695China1.793
Canada2.020Uganda1.803Belize1.714Estonia1.808
Tanzania2.200Singapore1.820China1.793Argentina1.811
Madagascar2.245Tanzania2.200Uganda1.803Lithuania1.821
Poland2.214Poland2.214Estonia1.808Malaysia1.977
Czech Republic2.286Madagascar2.245Argentina1.811Côte d’Ivoire2.000
South Africa2.434Czech Republic2.286Singapore1.820Senegal2.167
Chile2.525Slovak Republic2.336Lithuania1.821Cameroon2.297
Albania2.543South Africa2.434Zambia1.950Bulgaria2.402
Haiti2.831Ghana2.500Namibia1.965El Salvador2.578
Turkey2.830Chile2.525Malaysia1.977Egypt2.742
Pakistan2.904Albania2.543Côte d’Ivoire2.000Peru3.000
Peru3.000Georgia2.767Senegal2.167Zimbabwe3.000
Ukraine3.051Nigeria2.825Malawi2.273
Belarus3.085Turkey2.830Ceroon2.297
Russia3.116Haiti2.882Slovak Republic2.336
Honduras3.264Ukraine3.051Trinidad and Tobago2.396
Mexico3.182Mexico3.182Bulgaria2.402
Colombia3.293Colombia3.293Ghana2.500
Guatemala3.553Indonesia3.354El Salvador2.578
Philippines3.450Philippines3.450Armenia2.790
Thailand3.634Thailand3.634Zimbabwe3.000
Note: RR floats include cases of managed floating (which includes a de facto crawling band that is narrower than or equal to ±5 percent and a moving band that is narrower than or equal to ±2 percent), and freely floating. We drop Indonesia from the RR float sample because, in the aftermath of the currency crisis, firms from all sectors found the exchange rate a problem in Indonesia in 1999, making it difficult to tease out any sectoral distinctions.
Note: RR floats include cases of managed floating (which includes a de facto crawling band that is narrower than or equal to ±5 percent and a moving band that is narrower than or equal to ±2 percent), and freely floating. We drop Indonesia from the RR float sample because, in the aftermath of the currency crisis, firms from all sectors found the exchange rate a problem in Indonesia in 1999, making it difficult to tease out any sectoral distinctions.

Figures 1 and 2 provide preliminary comparisons of managers’ concerns about the exchange rate across these regimes. Overall, managers report more problems with the exchange rate in floating regimes than managers operating under pegged exchange rates. According to the summary data in Figure 1, the average response among managers in floats is 2.78 vs. 2.04 for managers in fixed regimes. Figure 2 shows that 36 percent of managers in floating regimes find the exchange rate to be a “major obstacle” to their businesses but only 13 percent of managers in pegged regimes report the same level of concern. These distributions reveal that, ceteris paribus, managers find pegged regimes less problematic than floats.

Figure 1.Average Response of Firm Managers, by Exchange Rate Regime

Figure 2.Respondents Reporting Problems with the Exchange Rate, by Exchange Rate Regime

(In percent)

We want to test the hypothesis that internationally exposed firms prefer greater currency stability, and hence a fixed exchange rate. To do so, we need firm-level proxies for “international exposure.” The WBES asks respondents several questions about the characteristics of their firms; most important for our purposes, firms are asked to identify the sector in which they operate: manufacturing, service, agriculture, or construction.7 In a separate inquiry, managers are asked whether their firms export (yes or no).8 We use these responses to create three unique proxies for international exposure. Manufacturing takes on a value of 1 if the firm operates in the manufacturing sector; Tradable is a dummy variable for firms in either manufacturing or agriculture; Exporter indicates that the firm exports at least a portion of its production. As noted in Table 2a, Exporter correlates significantly with Manufacturing at 0.33 and with Tradable at 0.31.

A preliminary review of the data reveals variance across sectors in the degree to which exchange rate stability matters for firm managers. Figure 3 illustrates the average response among manufacturing firms vs. that of nonmanufacturing firms in floating and fixed exchange rate regimes. Manufacturers, on average, express greater discontent under floating regimes: the average response among manufacturing firms is 2.89 compared with 2.67 for nonmanufacturers. Although the average response is lower for both groups in fixed-rate regimes, the difference between the two groups is much lower in fixed-rate regimes; in fact, manufacturers are slightly happier than nonmanufacturers in fixed regimes (average values of 1.99 and 2.05, respectively).

Figure 3.Manufacturing vs. Nonmanufacturing Average Responses, by Exchange Rate Regime

Figure 4 compares the distribution of responses between manufacturing and nonmanufacturing firms operating in floating exchange rate regimes. Manufacturers express more dissatisfaction with exchange rates under floats than respondents operating in nonmanufacturing enterprises. Later, we use regression analysis to determine whether other factors contribute to these differences, but this preliminary evidence suggests that international exposure contributes to a preference for exchange rate stability among businesspeople.

Figure 4.Manufacturing vs. Nonmanufacturing Responses in Floating Regimes

(In percent)

Our next hypothesis relates to the level of the exchange rate, which we measure by the variable REER Appreciation; the percentage change in the real effective exchange rate between 1999 and 1998.9 Positive values indicate a real appreciation of the currency. We expect firms in the tradables sectors to be more concerned about the exchange rate as the real exchange rate appreciates. To evaluate this conditional proposition, we make use of interactions between our sectoral dummies and REER Appreciation.

Survey responses may reflect other firm- or country-level characteristics besides international exposure and sector of activity. We use the responses to two other WBES questions to control for firm-level factors: government ownership and firm size.10 We also include several country-level variables in our models: the log of GDP per capita, the log of foreign direct investment (FDI) stock per capita, and financial sector liabilities (M3/GDP).

Our first set of models aims to identify firm owners’ concern with the stability of the exchange rate. We estimate the following equation:

where the subscripts stand for firm i in country j. The dependent variable is Exchange Rate Problem, the response of firm i in country j. We divide the sample of responses according to the type of exchange rate regime under which the firms operate: floating or pegged. The variable of interest is Sector, which represents one of the following sectoral dummies: Manufacturing, Tradable, or Exporter. The vectors Firm and Economy are firm- and country-level controls.

To test our claims about sectoral attitudes toward the level of the exchange rate, we run interactions between the sectoral dummies and REER Appreciation. These models take the following form:

where the subscripts stand for firm i in country j. The dependent variable is Exchange Rate Problem, the response of firm i in country j. As before, we divide the sample of responses according to the type of exchange rate regime under which the firms operate: floating or pegged. Sector represents one of the following sectoral dummies: Manufacturing, Tradable, or Exporter. The variable of interest is the interaction Sector × REER Appreciation. We expect internationally exposed firms to be more concerned about the exchange rate as the real exchange rate appreciates. Firm and Economy are vectors of firm-and country-level controls.

Because our dependent variables are discrete, ordered responses, we estimate the equation with ordered probit models using standard maximum likelihood and heteroscedasticity-robust standard errors. Following Beck, Demirguc-Kunt, and Levine (2006), who model a different WBES survey response as a function of firm- and country-level variables, we allow for possible correlation of the error terms among firms within the same country using Stata’s “cluster” command. This technique maintains the assumption of independent error terms across countries, while allowing for within-country correlation of the errors, perhaps due to factors such as common linguistic or cultural interpretations of the survey.

To the extent that a government’s choice of exchange rate policy is influenced by the pressures of sectoral interest groups, as we believe, endogeneity bias is an obvious concern. For example, where manufacturing composes a large share of GDP, governments may be particularly responsive to manufacturers when setting exchange rate policy. But if the causal arrow runs unambiguously from firms’ preferences to exchange rate policies, the bias works against our finding sectoral differences in the data. That is, if manufacturers obtain the exchange rate policies they want, we should not observe a difference in the responses of manufacturing and nonmanufacturing firms. Thus, our results are likely to understate the effect of endogenous exchange rate policies on firms’ attitudes.

III. Results

Tables 47 contain estimates of the covariates of exchange rate attitudes among firms operating in floating and in fixed exchange rate environments, respectively. Tables 4 and 5 use the LYS exchange rate regime classifications but Tables 6 and 7 replicate the analysis using RR’s regime classifications. Our expectation is that, in floating regimes, firm owners in internationally exposed sectors will express greater concern with exchange rates than other sectors. Moreover, this concern will be muted in fixed regimes, as currency risk is limited. We use sector dummies to proxy for “internationally exposed”: Manufacturing, Tradable (manufacturing + agriculture), and Exporter. In Models 1–4, we include the following variety firm- and country-level control variables:

Table 4.Exchange Rate Attitudes in Floating Regimes(LYS classification)
(1)(2)(3)(4)
Government-Owned–0.330**–0.293**–0.293**–0.313**
(0.132)(0.127)(0.130)(0.137)
Size0.0420.0200.0290.025
(0.066)(0.064)(0.065)(0.054)
Log GDP/Capita–0.120–0.100–0.110–0.122
(0.098)(0.094)(0.094)(0.096)
M3/GDP0.0060.0070.0070.006
(0.007)(0.007)(0.006)(0.007)
Log FDI Stock/Capita–0.130***–0.152***–0.147***–0.134***
(0.048)(0.045)(0.045)(0.051)
Manufacturing0.187***
(0.056)
Tradable0.102*
(0.053)
Exporter0.092
(0.101)
Observations3,1082,9182,9183,049
Countries25252525
Pseudo R20.0340.0410.0390.035
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Table 5.Exchange Rate Attitudes in Fixed-Rate Regimes(LYS classification)
(1)(2)(3)(4)
Government-Owned–0.215–0.209–0.209–0.207
(0.154)(0.149)(0.147)(0.158)
Size–0.031–0.004–0.009–0.034
(0.068)(0.069)(0.069)(0.066)
Log GDP/Capita–0.159–0.161–0.161–0.149
(0.172)(0.163)(0.164)(0.170)
M3/GDP0.0000.0000.0000.000
(0.002)(0.003)(0.003)(0.002)
Log FDI Stock/Capita0.0260.0160.0160.027
(0.159)(0.148)(0.149)(0.157)
Manufacturing–0.068
(0.062)
Tradable–0.027
(0.066)
Exporter0.030
(0.119)
Observations1,9641,8231,8231,901
Countries25252525
Pseudo R20.0090.0110.0110.008
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses.
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses.
Table 6.Exchange Rate Attitudes in Floating Regimes(RR classification)
(1)(2)(3)(4)
Government-Owned–0.350**–0.326**–0.335**–0.340**
(0.149)(0.145)(0.147)(0.154)
Size0.0960.0640.0710.062
(0.074)(0.074)(0.075)(0.068)
Log GDP/Capita0.009–0.0090.0070.018
(0.195)(0.183)(0.186)(0.198)
M3/GDP0.0080.0090.0090.008
(0.006)(0.006)(0.006)(0.006)
Log FDI Stock/Capita–0.212–0.207–0.209–0.225
(0.150)(0.147)(0.149)(0.152)
Manufacturing0.197***
(0.060)
Tradable0.144***
(0.055)
Exporter0.174
(0.120)
Observations2,5072,3972,3972,450
Countries24242424
Pseudo R20.0290.0340.0330.032
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4 = “Major Obstacle.”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4 = “Major Obstacle.”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Table 7.Exchange Rate Attitudes in Fixed-Rate Regimes(RR classification)
(1)(2)(3)(4)
Government-Owned–0.396*–0.372*–0.373*–0.405*
(0.225)(0.214)(0.211)(0.230)
Size0.0750.0980.0930.094
(0.084)(0.084)(0.087)(0.087)
Log GDP/Capita–0.060–0.110–0.105–0.074
(0.247)(0.221)(0.222)(0.248)
M3/GDP–0.001–0.001–0.001–0.001
(0.003)(0.003)(0.003)(0.003)
Log FDI Stock/Capita–0.095–0.058–0.060–0.082
(0.269)(0.242)(0.243)(0.268)
Manufacturing–0.035
(0.072)
Tradable0.010
(0.077)
Exporter–0.100
(0.125)
Observations1,3111,2011,2011,281
Countries16161616
Pseudo R20.0140.0140.0140.014
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
  • Government-Owned (1 = yes, 0 = no)
  • Size (1 = small (5–50 employees); 2 = medium (51–500 employees); 3 = large (>500 employees)
  • Log GDP/Capita (average 1997–99)
  • M3/GDP (average 1997–99)
  • Log FDI Stock/Capita (average 1997–99)

In Tables 4 to 7, the coefficients for our variables of interest largely confirm our priors: in floating regimes, manufacturers and tradables producers are more likely to express concern about the exchange rate than firms in other sectors. In fixed regimes, by contrast, the internationally exposed sectors are less likely to report that the exchange rate represents a problem for their businesses, although the estimates are not statistically significant in the fixed-regime sample. Note that nontradables are the complement of the set of tradables, so the value of the estimated coefficient on nontradables is the same as for Tradable, but with the opposite sign. Thus, nontradables (services + construction) are less likely to be concerned about the exchange rates in floats.

The results on Tradable are not significant because attitudes toward a floating exchange rate among firms in agriculture appear not to coincide with those of manufacturing firms. Run separately (but not reported), the dummy for Agriculture is negative and significant in floats and positive (though not significant) in fixed regimes, which is exactly the opposite of the results for Manufacturing. We speculate that firm owners in the agricultural sector are less concerned about the exchange rate in floats because agriculture has long been highly protected in most national economies. In addition, much farm output may be in traditional products or goods that do not enter readily into world trade (perishable or delicate produce, for example). Trade barriers and other transport restrictions may thus insulate food producers from the vicissitudes of exchange rate variation in floating regimes.

The negative and significant effect of Government-Owned implies state ownership reduces managers’ apprehensions about the exchange rate. It is difficult to see this variable as a proxy for the nontraded sector, however, because the simple correlation between Government-Owned and services is small and negative (-0.062). In fact, state ownership does not correlate strongly or positively with any other stand-in for “internationally sheltered” sector. Therefore, we treat it as a control. One interpretation of the finding that state-owned firms are less concerned than other firms about the exchange rate is that they have a privileged relationship with the government that protects them from such forces. They are “sheltered” not so much from exchange rate pressures but from market and political forces of any kind, as the government stands ready to subsidize and protect them. To test this interpretation, we estimated models of firm responses to a separate WBES question about the security of property rights. Our findings (not reported) indicate that state ownership has a large and negative influence on property rights concerns among firms, and that this result is robust to a battery of firm- and country-level controls.

Our country-level control for the stock of FDI, Log FDI Stock/Capita, returns negative and (mostly) significant estimates in floating regimes. In other words, firms operating in floats with more FDI per capita are less likely to report that the exchange rate causes problems for their businesses. One interpretation of this result is that foreign investors are more diversified across countries and currencies and thus less sensitive to volatility in the local currency.

In Figure 5, we used the Clarify software from Tomz, Wittenberg, and King (2003) to provide substantive meaning to the ordered probit estimates from Model 2 of Table 4. We simulated the difference in the predicted probabilities of manufacturing and nonmanufacturing firms’ responses while holding all other variables in at their means. In this way, we isolated the impact of “international exposure,” as proxied by being in the manufacturing sector, on firm managers’ concern with the exchange rate. The effect is meaningful and significant for all categories of the response except “moderate obstacle.” For example, in floating regimes, the simulated difference in the probability that a firm in the sheltered, nonmanufacturing sector says the exchange rate is a “major obstacle” is 7 percentage points lower than that of a manufacturing firm.

Figure 5.Substantive Effect of Sector (Manufacturing vs. Nonmanufacturing)

(Clarify simulations run on Model 2, Table 4)

Tables 8 and 9 test our claims about sectoral attitudes toward the level of the exchange rate. To do so, we make use of interactions between our sectoral dummies and REER appreciation. One of our goals is to test the proposition that internationally exposed sectors are more concerned with the exchange rate as the currency appreciates. In addition, we also hope to uncover the reason why producers in internationally exposed sectors express more dissatisfaction with exchange rates in floating regimes than in pegged regimes. The interactions allow us to determine the extent to which the concern in floating regimes involves the level of the exchange rate, as opposed to its volatility.

Table 8.Real Appreciation and Exchange Rate Attitudes in Floating Regimes(LYS Classification)
(1)(2)(3)(4)
REER Appreciation (one year)1.3530.5220.4500.879
(1.134)(1.006)(1.038)(1.226)
Government-Owned–0.361***–0.353***–0.358***–0.340**
(0.137)(0.131)(0.133)(0.144)
Size0.0780.0520.0660.038
(0.075)(0.070)(0.071)(0.062)
Log GDP/Capita–0.220**–0.230**–0.224**–0.230**
(0.109)(0.102)(0.103)(0.111)
M3/GDP0.009*0.010**0.010**0.009*
(0.005)(0.005)(0.005)(0.005)
Log FDI Stock/Capita–0.227***–0.208***–0.216***–0.220***
(0.073)(0.065)(0.066)(0.080)
Manufacturing0.267***
(0.078)
Manufacturing x REER Appreciation (one year)1.168***
(0.358)
Tradable0.212***
(0.063)
Tradable x REER Appreciation (one year)1.249***
(0.330)
Exporter0.278***
(0.105)
Exporter x REER Appreciation (one year)1.599***
(0.615)
Observations2,3232,2582,2582,276
Countries16161616
Pseudo R20.0780.0840.0830.082
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Table 9.Real Appreciation and Exchange Rate Attitudes in Floating Regimes(RR classification)
(1)(2)(3)(4)
REER Appreciation (one year)0.147–0.568–0.648–0.217
(0.969)(0.838)(0.860)(1.141)
Government-Owned–0.435***–0.422***–0.436***–0.431**
(0.166)(0.154)(0.158)(0.168)
Size0.165**0.130*0.140**0.118*
(0.067)(0.067)(0.068)(0.066)
Log GDP/Capita0.0520.0680.0740.061
(0.241)(0.227)(0.232)(0.244)
M3/GDP0.012**0.013***0.013***0.012**
(0.005)(0.005)(0.005)(0.005)
Log FDI Stock/Capita–0.376**–0.395**–0.400**–0.386**
(0.182)(0.169)(0.174)(0.185)
Manufacturing0.275***
(0.083)
Manufacturing x REER Appreciation (one year)1.417**

(0.599)
Tradable0.223***
(0.069)
Tradable x REER Appreciation (one year)1.498***
(0.538)
Exporter0.279**
(0.112)
Exporter x REER Appreciation (one year)0.996
(0.841)
Observations2,1392,0692,0692,094
Countries19191919
Pseudo R20.0590.0670.0660.064
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.
Note: Ordered probit analysis of Exchange Rate Problem, a discrete, ordered dependent variable of firm managers’ responses to the question, “How problematic is the exchange rate for the operation and growth of your business?” (1=“No Obstacle,” 2 = “Minor Obstacle,” 3 = “Moderate Obstacle,” 4=“Major Obstacle”). Robust standard errors, clustered by country, are in parentheses. ***, **, and * indicate statistical significance levels of 1, 5, and 10 percent, respectively.

Tables 8 and 9 report the estimates of Equation (2) for firms operating in floating regimes, using the LYS and RR samples, respectively. We estimate the effects in floating exchange rate regimes because real appreciations are more likely in floats than in regimes classified as de facto pegs. Policymakers in de facto pegs have demonstrated a commitment to both nominal and real currency stability; indeed, the mean of REER Appreciation among countries coded as floating by LYS is -0.039, but it is just -0.002 among LYS pegs. We thus test the expectation that internationally exposed firms are more concerned about the exchange rate as the REER appreciates in floating regimes only.

In Table 8, Models 1–4, the signs of our internationally exposed sector variables are positive in every model, as are the interaction terms: in LYS floating regimes, manufacturers, tradables producers, and exporters all express greater concern about the exchange rate as the real exchange rate appreciates. Currency appreciation makes tradable goods less competitive both at home and abroad, which accounts for the positive and significant results on the interaction terms for manufactures, tradable producers, and exporters.

In Table 9, we estimate the same models using the RR sample and find additional evidence that sectoral attitudes are conditioned by currency appreciation. The interactions of manufacturing and tradables with real appreciation are positively signed and significant, indicating that firms in these internationally exposed sectors express greater concern with the currency as the REER appreciates. Although the same relationship holds for the interaction ofexporter and REER appreciation, the conditional effect is not statistically significant in this sample.

Figure 6 provides estimates of the magnitude of these interaction effects. Using Model 2 from Table 8, we simulated the change in the predicted probability of Exchange Rate Problem = 4 (“Major Obstacle”) as REER appreciation moves from its minimum to its maximum value, holding all other variables at their means. The simulations were performed separately for manufacturing firms (left panel) and nonmanufacturing (right panel) firms using Clarify.11

Figure 6.Exchange Rate Appreciation: Manufacturing vs. Nonmanufacturing

(Clarify simulations run on Model 2, Table 8)

Notes: These figures illustrate the change in the predicted probability that Exchange Rate Problem = 4 (a “major obstacle”) as REER appreciation moves from its minimum to its maximum value, holding all other variables at their means. The simulations were performed separately for manufacturing (left panel) and nonmanufacturing (right panel) firms using the Clarify software (Tomz, Wittenberg, and King, 2003).

For manufacturing firms, the predicted probability that a respondent will report that the exchange rate is a “major obstacle” increases by 25 percentage points (from 0.24 to 0.49) as real appreciation moves from its minimum (—0.31) to its maximum (0.09) value. By contrast, such a movement has far less influence on nonmanufacturing firms in floating regimes: the change in the probability of responding “major obstacle” is just 8 percentage points (from 0.27 to 0.35). There is also much more uncertainty around the point predictions for nonmanufacturing firms.

Our final concern involves the relative importance of volatility and appreciation to firm managers’ attitudes about the exchange rate. We found that in floating regimes internationally exposed firms are more concerned about the exchange rate than firms in sheltered sectors. We also found that internationally exposed firms are more concerned than sheltered, nonmanufacturing firms when the REER appreciates. But what we would like to know is whether the heightened concern among international exposed firms operating in floating regimes is caused by the inherent volatility of exchange rates in these regimes or by the tendency of floating regimes to experience real appreciations.

Some evidence for the magnitude of each effect can be extracted from the Clarify results presented above in Figures 5 and 6. If we compare differences in the predicted probability of manufacturing and nonmanufacturing firms reporting that the exchange rate poses a “major obstacle” across these two sets of regressions, we can get some sense of the relative magnitudes of volatility and level, respectively. From Figure 5, we see that, without regard to the level of the real exchange rate, the difference in the predicted probability that a firm in the sheltered sector reports a “major obstacle” compared with that of a manufacturing firm is 7 percentage points. From Figure 6, we estimate that, as REER appreciation ranges from its minimum to its maximum, the probability that a manufacturing firm classifies the exchange rate as a “major obstacle” is 17 percentage points higher than for a nonmanufacturing firm. With the caveat that the interaction simulations are based on a truncated sample and are subject to wide confidence intervals that diminish the statistical significance of the sectoral differences, it would appear that appreciation in floating exchange rate regimes is the more important problem for firms in the internationally exposed sectors.

Our results provide support for existing arguments linking the economic activities of firms to their preferences over exchange rate policy. Manufacturers express greater uncertainty under floating exchange rates than do firms in sheltered, nontradable sectors. Manufacturers and exporters are also more likely to disapprove of a real appreciation than are firms in nontradable industries. Whether and how these preferences are translated into pressures on policymakers, and into policy, is a matter for further analysis.

IV. Conclusion

Many analyses of the political economy of exchange rate policy start from the presumption that policymakers must take into account the exchange rate policy preferences of domestic economic agents. Such analyses typically posit that internationally oriented firms will be especially averse to the volatility associated with a floating rate, and will therefore prefer a fixed exchange rate. They also typically suggest that tradables producers will be particularly averse to the relative price effects of a real appreciation. Nontradables producers, conversely, are expected to prefer floating rates and an appreciated currency.

The empirical evidence we present here supports these expectations. Owners and managers of manufacturing firms, inherently more affected by the international economy, are more likely to regard the exchange rate as a significant problem in floating-rate regimes than those in the sectors that are not so exposed to cross-border transactions. This implies a preference of these more internationally engaged firms for a fixed exchange rate. Similarly, firms in the manufacturing sector and those with export interests are particularly concerned by a real appreciation of the exchange rate; firms in nontradables sectors are less likely to demonstrate such concern. This implies a preference of tradables producers, including exporters, for a relatively depreciated real exchange rate.

Although these results are hardly surprising, they do confirm theoretically grounded expectations about the policy preferences of important economic actors. Further work is needed to explore more nuanced and differentiated aspects of exchange rate policy preferences. For example, it would be important to know how such things as foreign currency liabilities, intra-firm trade, differential pass-through, and invoicing practices affect currency preferences. The evidence presented here is, nonetheless, a start toward a more rigorous and precise sense of the economic interests and attitudes with which policymakers contend.

REFERENCES

    BalistreriEdward1997“The Performance of the Hecksher-Ohlin-Vanek Model in Predicting Endogenous Policy Forces at the Individual Level,”Canadian Journal of EconomicsVol. 30No. 1 (February) pp. 117.

    • Crossref
    • Search Google Scholar
    • Export Citation

    BatraGeetaDanielKaufmann and Andrew W.H.Stone2003“The Firms Speak: What the World Business Environment Survey Tells Us about Constraints on Private Sector Development” in Pathways Out of Poverty: Private Firms and Economic Mobility in Developing Countriesed. byGary S.Fields and GuyPfefferman (Norwell, MassachusettsKluwer Academic Publishers).

    • Search Google Scholar
    • Export Citation

    BeaulieuEugene2002“Factor or Industry Cleavages in Trade Policy? An Empirical Analysis of the Stolper-Samuelson Theorem,”Economics and PoliticsVol. 14No. 2 (July) pp. 99131.

    • Crossref
    • Search Google Scholar
    • Export Citation

    BeckThorstenAsliDemirgüç-Kunt and RossLevine2006“Bank Supervision and Corruption in Lending,”Journal of Monetary EconomicsVol. 53No. 8 (November) pp. 213163.

    • Crossref
    • Search Google Scholar
    • Export Citation

    BrozJ. Lawrence1997“The Domestic Politics of International Monetary Order: The Gold Standard” in Contested Social Orders and International Politicsed. byD.Skidmore (Nashville, TennesseeVanderbilt University Press).

    • Search Google Scholar
    • Export Citation

    CalvoGuillermo A. and Carmen M.Reinhart2002“Fear of Floating,”Quarterly Journal of EconomicsVol. 107No. 2 (May) pp. 379408.

    • Crossref
    • Search Google Scholar
    • Export Citation

    CampaJose and Linda S.Goldberg1997“The Evolving External Orientation of Manufacturing: A Profile of Four Countries,”Federal Reserve Bank of New York Economic Policy ReviewVol. 3No. 2 (July) pp. 5381.

    • Search Google Scholar
    • Export Citation

    CampaJose and Linda S.Goldberg2005“Exchange Rate Pass-Through into Import Prices,”The Review of Economics and StatisticsVol. 87No. 4 pp. 67990.

    • Crossref
    • Search Google Scholar
    • Export Citation

    CollinsSusan M. and FrancescoGiavazzi1992“Attitudes Towards Inflation and the Viability of Fixed Exchange Rates: Evidence from the EMS” in A Retrospective on the Bretton Woods System: Lessons for International Monetary Reformed. byBarryEichengreen and MichaelBordo (ChicagoUniversity of Chicago Press).

    • Search Google Scholar
    • Export Citation

    DevereuxMichael B. and CharlesEngel2002“Exchange Rate Pass-Through, Exchange Rate Volatility, and Exchange Rate Disconnect,”Journal of Monetary EconomicsVol. 49No. 5 pp. 91340.

    • Crossref
    • Search Google Scholar
    • Export Citation

    EichengreenBarry1992Golden Fetters (OxfordOxford University Press).

    EichengreenBarry1995“The Endogeneity of Exchange Rate Regimes” in Understanding Interdependence: The Macroeconomics of the Open Economyed. byPeterKenen (Princeton, New JerseyPrinceton University Press).

    • Search Google Scholar
    • Export Citation

    EichengreenBarry and Jeffry A.Frieden1993“The Political Economy of European Monetary Unification: An Analytical Introduction,”Economics & PoliticsVol. 5No. 2 pp. 85104.

    • Crossref
    • Search Google Scholar
    • Export Citation

    FrankelJeffrey1995“Monetary Regime Choice for a Semi-Open Economy” in Capital Controls Exchange Rates and Monetary Policy in the World Economyed. bySebastianEdwards (New YorkCambridge University Press).

    • Search Google Scholar
    • Export Citation

    FrankelJeffrey1999“No Single Currency Regime is Right for All Countries or at All Times,”Essays in International Finance No. 215 (Princeton, New JerseyPrinceton University Press).

    • Search Google Scholar
    • Export Citation

    FriedenJeffry1991“Invested Interests: The Politics of National Economic Policies in a World of Global Finance,”International OrganizationVol. 45No. 4 (Autumn) pp. 42551.

    • Crossref
    • Search Google Scholar
    • Export Citation

    FriedenJeffry1997“Monetary Populism in Nineteenth-Century America: An Open Economy Interpretation,”Journal of Economic HistoryVol. 57No. 2 (June) pp. 36795.

    • Crossref
    • Search Google Scholar
    • Export Citation

    FriedenJeffry2002“Real Sources of European Currency Policy: Sectoral Interests and European Monetary Integration,”International OrganizationVol. 56No. 4 (Autumn) pp. 83160.

    • Crossref
    • Search Google Scholar
    • Export Citation

    FriedenJeffryPieroGhezzi and ErnestoStein2001“Politics and Exchange Rates: A Cross-Country Approach to Latin America” in The Currency Game: Exchange Rate Politics in Latin Americaed. byJeffry A.Frieden and ErnestoStein (BaltimoreJohns Hopkins University Press).

    • Search Google Scholar
    • Export Citation

    GabelMatthew1998“International Economics and Mass Politics: Market Liberalization and Public Support for European Integration,”American Journal of Political ScienceVol. 42No. 3 (July) pp. 93653.

    • Crossref
    • Search Google Scholar
    • Export Citation

    GärtnerManfred1997“Who Wants the Euro—and Why? Economic Explanations of Public Attitudes towards a Single European Currency,”Public ChoiceVol. 93No. 3 pp. 487510.

    • Crossref
    • Search Google Scholar
    • Export Citation

    GhoshAtishAnne-MarieGuldeJonathan D.Ostry and Holger C.Wolf1997“Does the Nominal Exchange Rate Regime Matter?”NBER Working Paper No. 5874 (January) (Cambridge, MassachusettsNational Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    HefekerCarsten1995“Interest Groups, Coalitions and Monetary Integration in the Nineteenth Century,”Journal of European Economic HistoryVol. 24No. 3 (Winter) pp. 489536.

    • Search Google Scholar
    • Export Citation

    HefekerCarsten1997Interest Groups and Monetary Integration: The Political Economy of Exchange Regime Choice (Boulder, ColoradoWestview Press).

    • Search Google Scholar
    • Export Citation

    KenenPeter1969“The Theory of Optimum Currency Areas” in Monetary Problems in the International Economyed. byRobertMundell and AlexanderSwoboda (ChicagoUniversity of Chicago Press).

    • Search Google Scholar
    • Export Citation

    KleinMichael W. and Nancy P.Marion1997“Explaining the Duration of Exchange-Rate Pegs,”Journal of Development EconomicsVol. 54No. 2 (December) pp. 387404.

    • Crossref
    • Search Google Scholar
    • Export Citation

    LeblangDavid2002“The Political Economy of Speculative Attacks in the Developing World,”International Studies QuarterlyVol. 46No. 1 (March) pp. 6991.

    • Crossref
    • Search Google Scholar
    • Export Citation

    Levy YeyatiEduardo and F.Sturzenegger2005“Classifying Exchange Rate Regimes: Deeds vs. Words,”European Economic ReviewVol. 49No. 6 (August) pp. 160335.

    • Crossref
    • Search Google Scholar
    • Export Citation

    MaydaAnna Maria and DaniRodrik2005“Why Are Some People (and Countries) More Protectionist Than Others?”European Economic ReviewVol. 49No. 6 (August) pp. 1393430.

    • Crossref
    • Search Google Scholar
    • Export Citation

    McKinnonRonald I1963“Optimum Currency Areas,”American Economic ReviewVol. 53No. 4 (September) pp. 71725.

    MundellRobert A1961“A Theory of Optimum Currency Areas,”American Economic ReviewVol. 51No. 4 (September) pp. 65764.

    ReinhartCarmen M. and KennethRogoff2004“The Modern History of Exchange Rate Arrangements: A Reinterpretation,”Quarterly Journal of EconomicsVol. 119No. 1 pp. 148.

    • Crossref
    • Search Google Scholar
    • Export Citation

    RoseAndrew2000“One Money, One Market: Estimating the Effect of Common Currencies on Trade,”Economic PolicyVol. 15No. 30 (April) pp. 945.

    • Crossref
    • Search Google Scholar
    • Export Citation

    ScheveKenneth F2004“Public Inflation Aversion and the Political Economy of Macroeconomic Policymaking,”International OrganizationVol. 58No. 1 (Winter) pp. 134.

    • Crossref
    • Search Google Scholar
    • Export Citation

    ScheveKenneth FKennethF and Matthew J.Slaughter2001“What Determines Individual Trade-Policy Preferences?”Journal of International EconomicsVol. 54No. 2 pp. 26792.

    • Crossref
    • Search Google Scholar
    • Export Citation

    SimmonsBeth1994Who Adjusts? (Princeton, New JerseyPrinceton University Press).

    TavlasGeorge1994“The Theory of Monetary Integration,”Open Economies ReviewVol. 5No. 2 pp. 21130.

    TomzMichaelJasonWittenberg and GaryKing2003CLARIFY: Software for Interpreting and Presenting Statistical Results Version 2.1 Harvard University. Available via theInternet: http://gking.harvard.edu.

    • Search Google Scholar
    • Export Citation

    VéghCarlos1992“Stopping High Inflation: An Analytical Overview,”IMF Staff PapersVol. 39 (September) pp. 62695.

    World Bank2000World Business Environment Survey (WashingtonWorld Bank).

*J. Lawrence Broz is an associate professor in the Department of Political Science at the University of California, San Diego; Jeffry Frieden is a professor in the Department of Government at Harvard University; and Stephen Weymouth is a Ph.D. student in the Department of Political Science at the University of California, San Diego. The authors thank Justin Wolfers and Barry Eichengreen for helpful comments. They also thank participants at the IMF’s Eighth Annual Jacques Polak Research Conference, Washington, November 15–16, 2007.
1Obviously, policymakers have a wide choice of regime, ranging from a completely free float to a variety of managed floats, degrees of fixity ranging from a target zone to a peg, and a currency board or dollarization. This discussion focuses on the extremes—hard pegs and pure floats—however, because the analysis of intermediate cases flows from the extremes, and the trade-offs described apply to the intermediate choices, albeit never as starkly as to the extremes.
2Under most regimes a government must decide whether it prefers a relatively appreciated or relatively depreciated currency. Free floats are rare, and by the same token, countries that opt for a pegged regime always have the choice of abandoning the peg.
3See Mundell (1961), McKinnon (1963), and Kenen (1969). A more recent survey is Tavlas (1994).
4See for example, the empirical results in Frankel (1995), Rose (2000), Végh (1992), and Ghosh and others (1997).
5For a discussion of the WBES project, see Batra, Kaufmann, and Stone (2003).
6LYS include an “intermediate” category, which we omit from our analysis because we have no strong theoretical priors about business elites’ attitudes in these regimes.
7We drop all firms identifying themselves as “Other.”
8Respondents that indicated they were exporters were also asked to specify the percentage of exports to total sales. Unfortunately, most firms did not respond to this inquiry, and so we are constrained to use a dummy variable to indicate whether a firm exports or not.
9The data are from the International Financial Statistics (IFS) and the Bank for International Settlements (BIS). The World Development Indicators provide data for Georgia, but coverage ends at 1998. Therefore, REER Appreciation for Georgia measures the percentage change in the real effective exchange rate between 1998 and 1997.
10Government ownership is a dummy variable that takes the value of 1 if a respondent indicates state ownership. Firm size is an ordered response: 1 = small (5–50 employees); 2 = medium (51–500 employees); 3 = large (>500 employees).
11Similar effects (not reported) were obtained substituting the interactions of REER appreciation with exporter and tradables.

Other Resources Citing This Publication