Dollarization in developing economies : did the dollarization in developing economies contribute to higher GDP per capita growth rates?

(1)

Faculty of Economics and Business

Dollarization in developing economies

Did the dollarization in developing economies contribute to higher GDP per

capita growth rates?

29 June, 2016

Abstract

This study investigates whether the process of currency substitution, also known as dollarization, has a positive effect on the per capita growth rates of GDP. A panel of data is constructed for 17 countries from the time period starting in 2002 and ending in 2014 and with the use of a panel data regression the effect of dollarization is estimated. Found was a positive but insignificant effect of dollarization on GDP per capita growth in the final regressions.

Bachelor Thesis Economics and Business Specialization: Economics

Name: Thierry Belt Student Number: 10587713

(2)

Statement of Originality

This document is written by Thierry Belt who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents

(3)

Introduction

Globalization and economic integration have never been on a higher level than in the last decade. The increasing worldwide interaction is an important determinant of today’s prosperity. It all started with the Bretton Woods System in 1944 after multiple countries experienced excessive volatile exchange rates during the 1930’s. Part of the Bretton Woods System were the International Monetary Fund (IMF) and the International Bank for Reconstruction and development, which were both established in 1944 to facilitate post-war reconstruction and secure financial stability (“About the Bretton Woods Institutions”, 2016). The establishment of these two institutions under the Bretton Woods System were seen as the milestones of the evolution of international financial integration. During the same period, the General Agreement on Trade and Tariffs (GATT) was founded. 23 countries agreed on moving from multiple domestic markets towards one common world market from the beginning of 1948 (“About the WTO”, 2016). From this moment, the international financial and goods market became more integrated and globalization had started to increase. At the end of 1994, after the Uruguay Round Negotiations, the GATT was replaced by the World Trade Organization (WTO) (The General Agreements on Tariffs and Trade, 1986).

During the last centuries the amount of countries increased rapidly, but the amount of currencies did not increase with the same pace. Why didn’t these two phenomena grow at the same rate? Some countries decided to abandon its own currency or not creating one in the first place and adopted another country’s currency as a means of payment and unit of account. This is also known as currency substitution. The United States dollar is the most frequent used currency because it is the world’s reserve currency, so this process is commonly referred to as dollarization.

Another currency is often adopted by developing countries to induce price stability and to lower uncertainty in the financial markets, in the hope of attracting investors. A few countries in Latin America and the Caribbean substituted their own currency, but also in Africa some countries adopted the dollar as their national currency. This phenomenon does not only occur in developing countries. In Europe for example, 19 members of the European and Monetary Union substituted their national currency for one common currency, the Euro.

Exchange rates, defined as the price of one currency in terms of another currency, play an important role in the economic interactions between countries. The exchange rates are determined in the international goods and financial markets. Differences in prices of tradeable goods and capital flows are reflected in these rates (Pilbeam, 1992). This important mechanism is removed when multiple countries adopt a common currency.

Goldfajn et al. (2001) have studied the effects of dollarization in Panama. One of their findings was that after the adoption of the U.S. dollar, the volume of trade with the United States tripled. Besides, the process of dollarization also led to more controlled inflation. Increased volume of trade and lower and more stable levels of inflation both led to better economic performance in the case of Panama.

The question that arises is whether the adoption of the U.S. dollar, by countries other than the United States, lead to higher growth rates of GDP in per capita terms compared to countries that maintained their own currency.

The structure of this paper is as follows. Theoretical consideration concerning the process of dollarization will be discussed in the first section. Section II will include the methodology and a detailed discussion on the variables used in the regression. The empirical results will be

(5)

discussed in section III. The fourth section will elaborate on the significance of dollarization. In section V the methodological issues and potential improvements will be examined. The paper will be completed with the concluding remarks in section VI.

(6)

Section I: Literary review

Alesina and Barro (2001) argue that the process of dollarization can be partly explained by the increasing number of small countries on the world map. For smaller countries the costs of creating an own currency are relatively high. For these “new” countries it is more beneficial to preserve the currency they were already using when the country was still part of a bigger entity, or to adopt an already existing currency.

What are the benefits of the system of dollarization? First of all, it lowers the information and transaction costs. Calvo (2001) argues that the increased transparency of prices lowers the information costs related to international trade. The model described by Merton (1987) states that an indirect barrier, like transaction costs, influences investor behavior. Besides, investors hesitate to invest in currencies that are not their own, because they value them as riskier than assets denoted in their own currency, while this isn’t necessarily true. This misconception of risks is called the home bias and is caused by information asymmetry (Ahearne, Griever, & Warnock, 2002). Adopting a foreign currency can reduce the information asymmetry between investors compared to a situation in which each country uses its own individual currency. On the one hand, information asymmetry leads to hesitance among investors, but this is only an indirect cost. On the other hand, direct costs arise when the uninformed party tries to get informed. According to Merton (1987) information costs consist of two parts; the cost of gathering and processing data, and the transfer of information from one party to another. The U.S. dollar is the most used currency in the world, because it is the world’s primary reserve currency. As a result, exchanging the U.S. dollar for another currency or vice versa is less costly in terms of time and money compared to the exchange between other currencies. Ultimately, when both countries use the U.S dollar, the information asymmetry is removed totally and the transaction costs are reduced and international trade connections become more interesting. The Mundell model also describes these transaction cost benefits, which have become more important in recent years, because of the increased economic integration and globalization (Calvo, 2001).

According to Sercu et al. (1995) transaction costs can lead to “no-trade regions”. These regions are defined as countries that would be better off by importing certain products rather than producing them themselves, but are obstructed to do so by trade costs. It could be the case that a foreign country can produce a certain product cheaper than the home country, but the lower prices are compensated by the additional transaction costs. This leads to a situation where the home country is producing the product itself, while the resources could be allocated more efficiently. These no-trade regions are obviously caused by all kinds of costs associated with international trade. In addition to transaction costs, one can think of insurance costs and shipping costs. Sharing a currency does not lead to the vanishing of all costs, but at least some no-trade regions caused by information and transaction costs will diminish or even disappear entirely.

Berg and Borenzstein (2000) argue that the adoption of the U.S. dollar reduces the risk premium on a country’s debt faced by investors, because the risk of devaluation is no longer present. A sudden depreciation of a currency can lead to extreme capital outflows. The removal of this risk leads to higher levels of confidence and the risk premium and thus the interest rate will decrease. Lower costs of borrowing stimulate investment, which in turn can lead to higher growth levels (Mankiw, 2003). This can lead to less uncertainty and more stable financial systems, which are often lacking in developing countries. This does not imply that investors face no risk at all, because the country’s default risk is still present. However, according to

(7)

Goldfajn et al. (2001) the correlation between currency and default risk should not be neglected. They state that the removal of the currency risk can in some cases even lead to an increase of the overall risk. Argued is that interest rates mostly reflect default risks instead of currency risks. The removal of the latter does in these cases not lead to a decrease of the interest rates. In the absence of exchange rate flexibility, it can put additional pressure on the government, leading to an increase of the default risk and the risk premium. Besides, it is still possible for governments to maintain a suboptimal fiscal policy, leading to budget deficits and eventually to investors leaving the country (Berg and Borenzstein, 2000).

When a country chooses to abandon its own currency, the possibility for seigniorage is removed as well. The process of printing money is often accompanied with higher levels of inflation, thus the loss of the process of seigniorage also reduces inflationary pressure. This is supported by Berg and Borenzstein (2000), who argue that the adoption of the U.S. dollar removes the possibility of an inflationary policy. This supports in turn also the development towards a more stable and investment-friendly environment.

According to the gravity model of international trade and Rose (2000), characteristics like sharing the same currency increases the amount of trade between countries. Rose (2000) states that sharing a common currency increases the volume of trade by a factor of 2-3.

There are of course also disadvantages related to currency substitution. First of all, the government loses part of its monetary policy (Alesina & Barro, 2001). The central banks are no longer able to print money whenever they think it’s needed. Seigniorage can only be examined by the home country of the currency, in most cases, the United States of America. Calvo (2001) also argues that the removal of one policy variable puts additional pressure on the remaining variables. The Mundell-Fleming model supports Calvo’s (2001) findings, because it states that monetary and fiscal policy are more effective under floating exchange rate regimes than under fixed regimes to restore the internal balance (Boughton, 2003). Sometimes, the exchange rate itself is used as an instrument. All these studies argue that the adoption of another currency leads to a loss in effectiveness of government policy. Calvo and Reinhart (2000) argue nonetheless that the argument of losing authority is mainly illusory.

Besides, when a major shock occurs (which is the most likely in developing countries), a country cannot decide to devalue the currency to restore the economy anymore. This means that countries can respond less quickly in the case of economic shocks and longer aftermaths are the result (Berg and Borenzstein, 2000).

There are not only economic reasons to oppose to currency substitution. It is argued by Berg and Borenzstein (2000) that a country’s currency is seen as a national symbol. This may lead to hesitation to the abandoning of the home currency.

Currency Boards (hard pegs)

A distinction should be made between the several forms of adopting another country’s currency. Instead of the substitution of a currency as described above, a country can also opt for a currency board. Within this form of currency substitution, the country does not totally abandon its own currency, but still a lot of monetary sovereignty is foregone. As stated by Mundell (1997), a currency board rules out exchange-rate changes, by fixing the exchange rate between the local currency and a foreign currency. This phenomenon is also called currency –or in this case dollar pegging. Most of the benefits and disadvantages of currency substitution also hold for a currency board. However, the national governments have to be more active to make sure that the market exchange rate equals the predetermined fixed rate. For example, 100% of the money of the central bank has to be covered by foreign-exchange reserves (Mundel, 1997).

(8)

Nonetheless, a currency board is less credible than currency substitution, since misconduct of the authorities can lead to pressure on the currency, because of carrier abandonment. According to Tsang and Ma (2002), there are two additional risks investors face under a currency board. First, they face the efficiency risk. This is explained as the monetary authority failing to match the daily spot rate with the predetermined fixed rate. Inadequate reserves or the lack of a clear convertibility undertaking on the monetary base can be the cause. Second, investors also face systemic risk, which can be explained as the risk of the government abandoning the currency board. The presence of these risks can put high levels of pressure on the currency. Whereas currency substitution removes all the currency risk, with a currency board, part of the risk is still present. The magnitude of this risk depends on the credibility of the government and the monetary authorities. Another difference concerns the revenue from seigniorage. This form of income for the government vanishes when a foreign currency is adopted. With a currency board, a government can still get revenues from printing money, because they are able to invest the home currency in foreign assets, resulting in a positive return (Schwartz, 1993).

Hypothesis

According to the gravity model of international trade, sharing the same currency facilitates the international interaction and trade between countries. The existence of absolute and/or comparative advantages allocates production resources more efficiently and the global welfare, measured as GDP per capita, will increase. Besides, maintaining the U.S. dollar as the national currency provides the financial system with stability, less uncertainty and less inflationary pressure. Finally, dollarization can remove the information asymmetry between investors and it can reduce the risk premium on a country’s debt. However, the abandoning of a country’s home currency can also lead to a loss of policy variables and thereby authority. Countries can also not benefit from any seigniorage revenue anymore. These drawbacks are nonetheless not expected to exceed the benefits. So, hypothesized is to find a positive effect of the process of dollarization on GDP per capita growth.

(9)

Section II: Methodology

The goal of this study is to analyze whether the process of dollarization has any positive effects on the GDP per capita growth rates. With the use of a panel data regression, the effect of dollarization is estimated. Compared to conventional cross-sectional or time series regressions, panel data result in a larger number of data points and less collinearity amongst explanatory variables. This results in more efficient estimates, because it can combine the cross-sectional dimension and the time-series dimension. Another benefit of a panel data analysis, is that individual or time specific effects can be captured (Hsiao, 2014). This is extremely important in this analysis, since data is gathered from different countries that all behave in their own way and are influenced by different factors. However, a panel data regression also contains some disadvantages; for example, the slopes and intercepts are a pooled average of the different countries in different periods of time. The actual slopes and intercepts can differ between countries or within countries at different points in time (Hsiao, 2014). Data will be collected on a set of 17 countries during the period starting in 2002 and ending in 2014. El Salvador and Panama represent the countries that fully adopted the U.S. dollar as their national currency in the dataset. Belize, Botswana, Ecuador, Oman, Venezuela and three independent members of the Organization of Eastern Caribbean States (Saint Lucia, Grenada, and Saint Vincent and the Grenadines) are examples of countries using a currency board. As already explained before, these countries are not fully dollarized, but they maintain a fixed exchange rate with the U.S. dollar. These countries will be included in the dataset as partially dollarized countries. Costa Rica, the Dominican Republic, Guatemala, Honduras, Jamaica, Nicaragua, and Trinidad and Tobago are all countries with a floating exchange rate regime and will be used as control countries in the data set.

The panel contains 17 separate longitudinal data bases covering different countries. The control variables that will be used to estimate the effect of dollarization will be discussed below. Solow Model

According to the Solow model of growth, output is determined by the three variables; labor, physical capital and human capital (Mankiw et al., 1990). Besides, knowledge, or the effectiveness of labor also affects the total output. Output (GDP) is increasing in all the variables. Higher amounts of labor (e.g. a large labor force) lead to higher levels of output than smaller amounts of labor. Since the Solow model has the structure of a Cobb-Douglas function, labor has to be accompanied with physical capital. The higher the level of physical capital in an economy, the higher the level of physical capital per worker and the higher the total production of the economy. The third factor that influences the output in an economy, is the level of human capital. It is argued that a higher level of human capital raises the productivity of a worker (Mankiw et al., 1990).

The data set will be built upon the fundamentals of the Solow model. Labor, physical capital and human capital will all be proxied by some variables included in the dataset. Besides the variable denoting the dollarization process and the variables suggested by the model of Solow, some other variables that influence GDP per capita growth will be included in the dataset as well. All the variables will be explained in further detail below.

Labor

As already mentioned before, a higher amount of labor increases the level of output according to the model of Solow. The amount of labor will be measured using the annual data on the total labor force gathered from the World Bank. To avoid the problem of non-stationarity, the first

(10)

difference of the total labor force will be used. The total labor force includes everyone in the age group 15-65 who is economically active and supplies paid labor for the sake of production. Investments

The second important pillar according to the model of Solow is the level of physical capital in the economy. The level of physical capital depends on the level of investments made in an economy (Mankiw et al., 1990). Barro (2003) has also investigated the main determinants of economic growth. Found was a positive and significant effect of investments on GDP. Barro (2003) used the ratio of investments to GDP to estimate the effect and magnitude of investments. The same ratio will be included in this data set. Data on gross capital formation as a percentage of GDP is gathered from the database of the World Bank. Gross capital formation contains all domestic investments on an annual basis. Other investments than domestic investments, e.g. foreign direct investments, influence GDP in their own way and will be treated as an individual independent variable. This variable will be discussed later on.

FDI

Borenzstein et al. (1998) studied the effect of foreign direct investment (FDI) on economic growth using data from industrial and developing countries. Suggested is that FDI functions as an important mechanism to transfer knowledge from one country to another. The contribution to economic growth is relatively higher than the contribution of domestic investment. More modern technology and better management skills result in higher efficiency of FDI compared to domestic investments. A low level of the human capital stock limits the absorptive capability of host countries. Wang (1990) and Findlay (1978) support these findings, concluding that the presence of FDI increases the rate of technological development. Borenzstein et al. (1998) also differentiate the inflows and the outflows. The inflows of investments would stimulate economic growth through enhancement of the process of technological development, while the outflows of investments wouldn’t affect the country’s economy in a negative way. For this reason, only the inflows of FDI (e.g. gross FDI) will be included in the dataset as an explanatory variable for GDP growth per capita. Furthermore, when the giving and receiving countries have similar levels of technology, a smaller effect of FDI on economic growth would be experienced compared to countries with different levels of technological development (Borenzstein et al., 1998). FDI will be included in the dataset as a ratio to GDP. However, FDI is only found to be effective if a certain level of technology is already present in the host country (Borenzstein et al., 1998).

To obtain better estimations of the effects of FDI and domestic investments on GDP per capita growth, a lag of two years will be introduced. The delayed effect of (foreign) investments is supported by Bar-Ilan and Strange (1996), who argue that the decision to invest and the reaping of the benefits of the project are rarely at the same moment in time. For example, the date of issuing a construction permit and the completion of an office building are estimated to be 18 to 24 months apart. However, according to the national income identity, investments do also have an immediate effect (Mankiw, 2003). For this reason, two regressions will be conducted; one regular regression and one regression including a lag of two years for FDI and investments.

(11)

Human capital

According to Colclough (1982), (primary) schooling contributes to the productivity of labor and it facilitates economic development. In the line of the extended Solow model; the level of human capital is increasing in the years of schooling of an individual, and output is in turn increasing in the level of human capital (Mankiw et al., 1990). According to the Solow model of growth, human capital is seen as a complementary for physical capital and it also affects GDP in a positive way (Mankiw et al., 1990). Human capital is defined by Mankiw (2003) as the knowledge and skills that workers acquire through their education. In the presence of technological externalities (e.g. knowledge spillover), the benefits of increased accumulation of human capital are even greater than the model of Solow predicts (Mankiw, 2003). Different countries can have different levels of human capital and therefore different growth rates. Besides, it is also argued that the labor force needs to have a certain level of education (i.e. human capital) for any positive effects of FDI. To capture the rate of education, the level of human capital is included in the model.

Benhabib and Spiegel (1994) argue that enrollment rates can be used as a proxy for the level of human capital. However, they also argue that these rates represent the level of investment in human capital rather than the level itself. Graduation rates and final grades give a better proxy of the level of human capital. However, a complete data set on these variables is very hard to construct. For this reason, the gross enrollment ratio for secondary education will be used as a proxy for human capital. All students already obtained basic reading, writing and mathematical skills during primary education. Secondary education adds more subject-or skill-oriented instructions provided by more specialized teachers to their knowledge. Secondary education forms the layer of further education and skilled labor. These rates are gathered from the database of the World Bank and are measured as the number of people in secondary education divided by the population of the age group that officially corresponds to secondary education and multiplied by 100. In theory, this variable can take values ranging from 0 to 100, but overage children and repeaters can lead to ratios higher than 100.

Political stability

Alesina et al. (1996) studied the effect of political stability on GDP per capita growth in a sample of more than 100 countries covering a time period of a few decades. One of their main findings was that in countries with a high propensity of government collapse, their definition of political instability, GDP per capita growth was significantly lower. Alesina et al. (1996) argue that political instability affects economic growth through increased uncertainty leading to lower investments. Political instability is also found to be persistent; government collapses in the past increase the probability of future collapses. Foreign investors are more hesitant to invest in the country in question and local investors are now more willing to invest abroad. Data on political stability is gathered from the Worldwide Governance Project (WGI) conducted in the name of the World Bank. This variable gives the score of a country’s political stability, ranging from 0 to 1. The lower bound indicates a low level of political stability, while the upper bound indicates a stable political environment.

(12)

Rule of law

A positive correlation between the rule of law and GDP per capita exists, according to Barro (2003). Assumed is, that an improvement of the rule of law leads to an increase of property rights. This increase results in a stimulation of investments, leading to higher growth rates. Data on the rule of law is also gathered from the WGI project conducted in the name of the World Bank. The variable gives the score of a country’s rule of law, ranging from 0 to 1. The lower bound indicates a weak rule of law, while the upper bound indicates a stronger rule of law. The correlation between the rule of law and GDP (per capita) growth is supported by Rigobon and Rodrik (2005).

Government consumption

Besides the rule of law, Barro (2003) has also studied whether other variables influenced economic growth. A negative effect of government consumption on economic growth was found. Former studies have measured these effects, by using the ratio of government consumption to GDP. This ratio will also be used in this data set as an explanatory variable. Trade openness

Barro (2003) has also studied the effects of the openness to trade on economic growth. Concluded was that the openness to trade influenced the growth rates in a positive manner. In the conducted regression, this was only found to be weak. However, since the gravity model of trade predicts an increase in the international trade when countries share the same characteristics like a common currency, the effect in the dollarized countries might be stronger. To estimate the effect, the ratio of trade, both imports and exports, to GDP will be included in the dataset.

Inflation

Borenzstein et al. (1998) and Barro (2003) both investigated the effects of the inflation rate on economic growth. Both studies found a negative and statistically significant effect of the inflation rate on GDP. Pollin and Zhu (2006) state that inflation rates higher than two or three percent, disrupt the efficient allocation of resources, because of more price instability. Annual inflation rates will be gathered from the databases of the World Bank.

Dollarization

According to a publication of the World Bank in 2000, the ratio of Foreign Currency Deposits (FCD) to broad money is a good proxy of the degree of dollarization in a country (Gomis-Porqueras, 2000). Therefore, this ratio will be included in the dataset as an explanatory variable. The higher the total amount of foreign currency deposits compared to broad money, the higher the degree of dollarization. This continuous variable can have values ranging from zero to one. Foreign currency deposits arise when economic agents decide to open a bank account with a foreign currency instead of the domestic currency. In highly dollarized countries, where the U.S. dollar is used as a means of payment, agents are more likely to deposit money in a foreign currency. Broad money is defined by the World Bank as the sum of currency outside banks, all kinds of deposits other than those of the central government, bank and traveler’s checks and other securities such as certificates of deposit and commercial paper. According to the definitions of monetary aggregates of the book of Mishkin et al. (2013), broad money fits the definition of M3 most perfectly. Data on FCD and broad money are retrieved from the database of the IMF and the World Bank respectively.

(13)

Including all the variables that affect GDP per capita growth, the resulting model is as follows:

𝐺𝐷𝑃_𝑖𝑡 = 𝑏₀+ 𝑏₁𝐷𝑂𝐿𝐿𝐴𝑅𝐼𝑍𝐴𝑇𝐼𝑂𝑁_𝑖𝑡 + 𝑏₂𝐹𝐷𝐼_𝑖𝑡−2+ 𝑏₃𝑃𝑂𝐿𝑆𝑇𝐴𝐵_𝑖𝑡+ 𝑏₄𝐿𝐴𝑊_𝑖𝑡 + 𝑏5𝐺𝑂𝑉𝐶𝑂𝑁𝑆𝑖𝑡+ 𝑏6𝑇𝑅𝐴𝐷𝐸𝑖𝑡+ 𝑏7𝐻𝑈𝑀𝐴𝑁𝐶𝐴𝑃𝑖𝑡 + 𝑏8𝐿𝐴𝐵𝑂𝑅𝑖𝑡 + 𝑏9𝐼𝑁𝑉𝐸𝑆𝑇𝑀𝐸𝑁𝑇𝑆𝑖𝑡−2+ 𝑏10𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁𝑖𝑡+ 𝑢

The 𝑖 denotes the country index; the 𝑡 denotes the year index. 𝐷𝑂𝐿𝐿𝐴𝑅𝐼𝑍𝐴𝑇𝐼𝑂𝑁 represents the degree of dollarization as measured by the FCD to broad money ratio. 𝐹𝐷𝐼 denotes the share of foreign direct investments of GDP. 𝑃𝑂𝐿𝑆𝑇𝐴𝐵 reflects the stability of the political environment using a scale ranging from zero to one. The rule of law is denoted by 𝐿𝐴𝑊 also using a scale ranging from zero to one. HUMANCAP denotes the enrolment ratio in secondary education as a proxy of human capital. The share of government consumption to GDP is reflected by 𝐺𝑂𝑉𝐶𝑂𝑁𝑆 and the degree of trade openness, as imports and exports combined as a share of GDP, is denoted by the variable 𝑇𝑅𝐴𝐷𝐸. LABOR proxies the amount of labor using the growth of the total labor force. INVESTMENTS reflects the ratio of total domestic investments to GDP. INFLATION reflects the annual inflation rate measured as the GDP deflator.

The lag concerning FDI and investments denoted by t-2 will only be used in the final regression. The other regressions make use of the simple subscript t.

This paper investigates whether the process of dollarization has any positive effects on the growth rates in per capita terms. Considering the literature, a positive effect of the process of dollarization on the growth rates per capita terms is hypothesized. The coefficient of dollarization is expected to be significantly larger than zero, leading to a rejecting of the null hypothesis:

𝐻0: 𝛽𝐷𝑂𝐿𝐿𝐴𝑅𝐼𝑍𝐴𝑇𝐼𝑂𝑁 = 0 𝐻1: 𝛽𝐷𝑂𝐿𝐿𝐴𝑅𝐼𝑍𝐴𝑇𝐼𝑂𝑁 > 0

The empirical results leading to the possible rejection of the null hypothesis will be discussed in the next section. Table 1 presents a quick overview of all the variables used in the regression, their definitions and their sources.

(14)

Table 1

Overview of the variables GDP per

capita growth

Annual percentage growth rate of (gross) GDP per capita based on the purchasing power parity (PPP).

Source: World Bank

Dollarization The sum of foreign currency deposits divided by broad money (the sum of currency outside banks; demand deposits other than those of the central government; the time, savings, and foreign currency deposits of resident sectors other than the central government; bank and traveler’s checks; and other securities such as certificates of deposit and commercial paper) measured annually.

Source: IMF, World Bank

FDI Inflows of foreign direct investments (investments minus disinvestments) as a share of GDP, measured annually.

Source: World Bank

Political stability

Political Stability and Absence of Violence/Terrorism measures perceptions of the likelihood of political instability and/or politically-motivated violence, including terrorism. Estimate gives the country's annual score on the aggregate indicator, in units ranging from zero to one

Source: WGI project (World Bank)

Rule of law Rule of Law captures perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence. Estimate gives the country's annual score on the aggregate indicator, in units ranging from zero to one

Source: WGI project

Government Consumption

General government final consumption expenditure includes all government current expenditures for purchases of goods and services (including compensation of employees). It also includes most expenditures on national defense and security, but excludes government military expenditures that are part of government capital formation. Measured annually as a ratio to GDP.

Source: World Bank

Trade The sum of exports and imports of goods and services measured annually as a share of GDP.

Source: World Bank

(15)

Human capital

The amount of children in secondary education divided by the population of the age group that officially corresponds to secondary education, and multiplied by 100 (Gross enrollment ratio).

Source: World Bank (UNESCO institute for statistics) Investments The ratio of gross capital formation to GDP measured

on an annual basis.

Source: World Bank

Labor growth

Growth rate of the amount of people in the age group 15-65 that is economically active and supplies paid labor (labor force).

Source: World Bank (International Labor Organization (ILO))

(16)

Section III: Results

In this section, the empirical findings of the study will be discussed. These empirical findings are displayed in table 2 on the next page. This table contains all the estimated coefficients and the corresponding p-values of the several regressions. In total, 4 regressions have been conducted. In the first regression only the variable denoting dollarization was included. The second regression also included the main determinants of economic growth according to the Solow model of growth. These variables proxy the amount of labor, physical capital and human capital in the economy. All the variables were used in the third regression, including FDI, trade, government consumption, political stability, rule of law and inflation. The fourth and final regression included the same variables as the third regression, with the only difference that FDI and investments were turned into lagged variables. Both are lagged for two years, denoted by t-2 in the regression equation.

The first regression displays the model with only the process of dollarization as an independent variable. The coefficient is found to be both negative and significant at the ten percent significance level, which contradicts the hypothesis. However, very little of the variance in the dependent variable is explained by the independent variable dollarization, as reflected by the 𝑅2_{of 0.0172}

In the following regression, also the variables retrieved from the Solow model were included. Labor, investments and secondary education (as a proxy for human capital) were all three found to be positive and highly significant, supporting the predictions made by the Solow model of growth. However, the coefficients of investments and secondary education were both found to be very small. Only labor seemed to affect the GDP per capita growth rates to a larger extent with a coefficient of 0.157 approximately. The coefficient of dollarization is no longer negative. The coefficient was found to be positive, although it wasn’t significant at any of the common used significance levels. The inclusion of the variables of the Solow model led to an increase of 𝑅2 to 0.6598

The third regression consisted also of the remaining explanatory variables. The variables retrieved from the Solow model still affect the dependent variable in a positive way, but the coefficients of investments and secondary education are still very low. As supported by the literature; FDI, political stability and the rule of law were found to affect GDP per capita growth positively. However, they were all non-significant at any of the common used significance levels. The coefficient of trade is negative, but is not significant as well. In line with the theory, a negative coefficient for government consumption was found. It was also found to be significant, but as already predicted by theory it’s influence on GDP is very weak. An increase of 10 percentage points of government consumptions decreases the GDP per capita growth rates by only 0.1% approximately. The same is true for the influence of the inflation rate on the dependent variable. Its coefficient is negative and highly significant, but its effect is very weak. The coefficient of dollarization increased to 0.103 approximately in the third regression. Although the p-value did decrease as well, the coefficient still isn’t significant at any common used significance level. The inclusion of the additional explanatory variables les to an increase of the 𝑅2 to 0.6894

No variables were added in the final regression, although some changes have been made. To get better estimates of the independent variables, lagged values of FDI and investments were used, because it is argued by theory that it takes time before the benefits of (foreign) investments can be reaped. The overall fit of the model decreased to a 𝑅2 value of 0.6302. The coefficient of FDI was found to be negative instead of a positive coefficient in the third regression,

(17)

contradicting the theory. Although the p-value lowered, it still is highly insignificant. Labor is still significant, but now only at the ten percent significance level and its coefficient also lowered. The coefficient of trade was now found to be positive, contradicting the theory, but with a p-value of 0.992 it is highly insignificant. No other change in significance and direction was found amongst the explanatory variables, compared to the previous regression.

Table 2

Estimated coefficients for the independent variables

(1) (2) (3) (4) Dollarization -0.1897215 (0.100)* 0.0360161 (0.642) 0.1030471 (0.246) 0.1076604 (0.265) Labor 0.1565012 (0.001)*** 0.1614288 (0.001)*** 0.0999415 (0.055)* Investments 0.0077944 (0.000)*** 0.0073197 (0.000)*** 0.0003734 (0.815) Secondary education 0.007463 (0.000)*** 0.0076451 (0.000)*** 0.0091725 (0.000)*** FDI 0.0005957 (0.742) -0.0014456 (0.442) Trade -0.0011936 (0.117) 0.00000809 (0.992) Government consumption -0.0141961 (0.005)*** -0.0207967 (0.000)*** Political stability 0.1150986 (0.312) 0.1281235 (0.300) Rule of Law 0.0716535 (0.451) 0.1232601 (0.242) Inflation -0.0031023 (0.002)*** -0.004136 (0.000)*** Constant 9.271304 (0.000)*** 5.816425 (0.000)*** 6.376634 (0.000)*** 7.232465 (0.000)*** Prob > 𝑐ℎ𝑖2 _0.1004 _0.0000 _0.0000 _0.0000 Observations 231 170 153 152 𝑅2 0.0172 0.6598 0.6894 0.6302

*significant at the 10% significance level, **significant at the 5% significance level, ***significant at the 1% significance level.

(18)

Section IV: Discussion

In this section, the empirical results concerning dollarization will be discussed in more detail. As found in the previous section, the process of dollarization has no significant effect on the growth rates of GDP in per capita terms, although a positive direction of the coefficient was found in the last three regressions.

First of all, with values for dollarization ranging only from zero to one, the coefficient seems to be very small. Using the coefficient of the third regression, the difference between a completely non-dollarized and a fully dollarized country would only be 0.1 percentage points in per capita growth rates. Even if the coefficient was found to be significant, dollarization still wouldn’t cause very large difference between countries.

It is also important to note that there are different reasons for countries to adopt the U.S. dollar. Panama, El Salvador and Ecuador are all countries with a high value for dollarization. However, their motives to obtain the U.S. dollar differ. Panama for example, was one of the first countries to obtain the U.S. dollar in the beginning of the twentieth century. This country has a long-term history with the United States, with the building of the Panama-Canal as one of the most outstanding events. Panama already has strong ties with the United States, making the adoption of the U.S. dollar a logical move (Quispe-Agnoli and Whisler 2006). The same is true for El Salvador to a smaller extent. The country is involved in several trade agreements with the U.S. and the United States is also its main trading partner. Making it also a rational decision to obtain the U.S. dollar. Ecuador on the other hand, did not adopt the U.S. dollar because of a lifelong history and connection with the United States. Ecuador started the dollarization process at the beginning of this century after the financial banking crisis of 1999 to stabilize the economy (Quispe-Agnoli and Whisler 2006). To summarize, the reasons why countries adopt the U.S. dollar may differ, and so may the implications.

Besides, the year that the economy went into the transition might be important as well. Where El Salvador and Ecuador started the process of dollarization during this century, Panama already obtained the U.S. dollar a hundred years ago (Quispe-Agnoli and Whisler 2006). If dollarization mainly effects economic performance in the long run, it might be true that no significant effect will be found in El Salvador and Ecuador. The opposite would be true, if dollarization only has short run effects. Since a panel data regression is used in this study, the displayed coefficients are a pooled average of the coefficients for all the countries at all points in time. This might clarify the insignificant coefficients.

Finally, the insignificant results can be explained by the fact that the process of dollarization might be considered a second best solution. If underdeveloped financial markets for example are the main reason a country does not experience much economic growth, the adoption of the U.S. dollar will not immediately lead to prosperity. The U.S. dollar might help to increase the confidence and stability in a country, but as long as the underlying problems aren’t resolved, the economy might not experience any increase in growth rates.

(19)

Section V: Methodological issues

The model used in this paper could support a lot of theory, but unfortunately, there might be a problem concerning the exogeneity of a few variables. One of the assumptions regarding the regression is that the independent variables are determined outside the model, that they’re exogenous. If the variables are determined within the model and the variables are correlated with the error term, the estimated coefficients are less consistent. Endogeneity is the result of a backwards and forwards causality between the independent and dependent variables (Stock and Watson, 2015). This case of simultaneous causality mainly occurs with (foreign) investment. Argued is not only that (foreign) investments foster economic growth, but higher growth levels also make (foreign) investments more interesting (Borenzstein, 1990). Besides, it is argued that insufficient economic performance also fosters government collapse and thus political instability (Alesina, 1996). Argued is that low economic growth creates dissatisfaction amongst the population, leading to higher pressure on the government. Finally, human capital might not be exogenous as well, because higher (growth) levels of GDP per capita may stimulate the investment in education simultaneously. Either FDI, investments, political stability as human capital might be correlated with the error term. To solve this problem, the coefficients could be estimated using two stage least squares (TSLS) and instrumental variables. Further research might be needed to construct more consistent estimations using this IV-regression.

Due to data limitations, coefficients had to be estimated using an unbalanced data panel. Not all data was available for all countries and every year, because for some variables, countries haven’t starting measuring or have already stopped measuring. Unfortunately, missing data leads to a loss of efficiency. When data is more widely available, an IV regression can be examined using a balanced panel data to estimate more efficient and consistent coefficients. A dataset consisting of more than 30 countries was constructed, but in the end only for 17 different countries there was data available for all the variables for at least one year.

Furthermore, there is some space to improve the measurement of human capital. In this study, the enrolment rates of secondary schooling are used as a proxy. As already mentioned before, secondary schooling is an import pillar in the process of accumulating human capital. It is, however, not complete. Total years of education or tertiary schooling might provide a better representation of human capital. An even better proxy of human capital would be to measure the educational level amongst the labor force instead of the total population. Since, only the labor force can turn human capital into output. Or, as suggested by Mankiw et al. (1990), the ratio between unskilled adults and skilled adults to measure the stock of human capital. These values are despite not widely available on every country and in every year in the dataset. For this reason, the gross enrolment rates of secondary schooling are used in this study.

The ratios of foreign currency deposits to broad money as a proxy for the degree of dollarization slightly exceeded unity in some cases. This was observed in the cases of Panama and Ecuador. In theory, broad money including foreign currency deposits cannot be smaller than just the foreign currency deposits alone. These irrational observations might be due to measurement errors.

(20)

Section VI: Conclusion

Several countries have foregone their individual currencies to adopt the United States dollar as a quick way to achieve stability and certainty. The literature states that the adoption of the U.S. dollar removes part of information and transaction costs and lowers the information asymmetry. It results in a more efficient allocation of resources and international trade will increase. It is argued that dollarization reduces the risk premiums on countries’ debt. This relation is however somewhat ambiguous, because other studies argue an increase of the risk premiums as a result of the process of dollarization. On the other hand, there are also drawbacks attached to the process of dollarization. National governments are no longer able to obtain any revenues from seigniorage and part of the monetary authority is foregone. The exchange rate cannot function as a policy instrument anymore, and the pressure on the remaining policy variables will increase. Besides, countries can also become less responsive to economic shocks as a result of the process of dollarization. A distinction should be made between full and partial dollarization. The latter is the case when a country uses a currency board. The degree to which a country is dollarized is measured using the share of foreign currency deposits to broad money. With the use of a panel data regression, it is tried to complement the theory with empirics. The results show a positive effect of the process of dollarization on the GDP growth rates in per capita terms in the final three regressions. However, in none of those three conducted regressions, the coefficient was found to be significant.

The fundamentals of the Solow Model of growth that form the pillars of this study are confirmed by the dataset. The measures for capital, labor and human capital; domestic investments, growth of the labor force, and the gross enrolment ratio respectively, were found to be positive and significant. The growth rates of the labor force affected the GDP per capita growthtothe largest extent.

The effect of FDI was found to be positive, but insignificant when current values were used. This might be explained by Borenzstein et al. (1998). They argue that the effect of FDI on GDP per capita growth can be non-existence for two reasons. First, a low level of the human capital stock limits the absorptive capability of the host countries. Second, FDI is only found to be effective if a certain level of technology is already present in the host country. These two factors might explain the insignificance of FDI. After the introduction of a lag of two years, the coefficient of FDI changed directions and became negative. However, it was still non-significant. It was argued that it takes time to process the investments. A project is not immediately finished after the decision to investment. It takes time before the benefits can be captured, because the investments have to be converted into usable capital (Bar-Ilan and Strange, 1996). However, the introduction of the lag for (foreign) investments did not add anything to the explanation of the effects on GDP per capita growth in this study. The coefficients in the final regression were not very plausible and the 𝑅2_{even decreased compared} to the third regression.

The degree of openness to trade as measured by the ratio of imports and exports to GDP was also found to be insignificant. This is somewhat in line with the study of Barro (2003), because there was only found a very weak effect of trade on GDP growth. Predicted was a bigger effect in this study, because of the statements of the gravity model of trade. Nonetheless, the coefficient was found to be very small and furthermore insignificant.

Although the directions of government consumption, the rule of law and political instability are aligned with the theory and the assumptions, they were insignificant. So, in this stage there’s not enough evidence to state that they differ significantly from zero.

(21)

Unfortunately, with the currently available data used in the dataset, there is not enough evidence to state that the adoption of the United States dollar by other countries than the United States, known as currency substitution or dollarization, has any positive effects on the growth rates of GDP in per capita terms. The empirics do not confirm the predetermined hypothesis as the null hypothesis cannot be rejected.

(22)

Reference list

About the Bretton Woods Institutions. (2016). Retrieved from:

http://www.brettonwoods.org/page/about-the-bretton-woods-institutions

About the WTO. (2016). Retrieved from:

https://www.wto.org/english/thewto_e/whatis_e/who_we_are_e.htm

Ahearne, A. G., Griever, W. L., & Warnock, F. E. (2004). Information costs and home bias: an analysis of US holdings of foreign equities. Journal of International

Economics, 62(2), 313-336.

Alesina, A., & Barro, R. J. (2001). Dollarization. The American Economic Review, 91(2), 381-385.

Alesina, A., Özler, S., Roubini, N., & Swagel, P. (1996). Political instability and economic growth. Journal of Economic growth, 1(2), 189-211.

Bar-Ilan, A., & Strange, W. C. (1996). Investment lags. The American Economic

Review, 86(3), 610-622.

Barro, R. J. (2003). Determinants of economic growth in a panel of countries. Annals of

economics and finance, 4, 231-274.

Benhabib, J., & Spiegel, M. M. (1994). The role of human capital in economic development evidence from aggregate cross-country data. Journal of Monetary economics, 34(2), 143-173.

Berg, A., & Borensztein, E. (2000). The pros and cons of full dollarization (No. 0-50). International Monetary Fund.

Borensztein, E., De Gregorio, J., & Lee, J. W. (1998). How does foreign direct investment affect economic growth? Journal of international Economics, 45(1), 115-135.

(23)

Calvo, G. A. (2001). Capital markets and the exchange rate, with special reference to the dollarization debate in Latin America. Journal of Money, Credit and Banking, 312-334.

Calvo, G. A., & Reinhart, C. M. (2000). Fear of floating (No. w7993). National Bureau of Economic Research.

Goldfajn, I., Olivares, G., Frankel, J., & Milesi-Ferretti, G. M. (2001). Full dollarization: the case of Panama. Economía, 1(2), 101-155.

Gomis-Porqueras, P., Serrano, C., & Somuano, A. (2000). Currency Substitution in Latin

America: Lessons from the 1990s (Vol. 2340). World Bank Publications.

Hsiao, C. (2014). Analysis of panel data (No. 54). Cambridge university press.

Mankiw, N. G. (2003). Macroeconomics (8th ed.). Palgrave Macmillan.

Mankiw, N. G., Romer, D., & Weil, D. N. (1990). A contribution to the empirics of economic

growth (No. w3541). National Bureau of Economic Research.

Matthews, K., Giuliodori, M., & Mishkin, F. S. (2013). The economics of money, banking

and financial markets. Pearson Higher Ed.

Merton, R. C. (1987). A simple model of capital market equilibrium with incomplete information. The journal of finance, 42(3), 483-510.

Mundell, R. A. (1997). Currency areas, common currencies, and EMU. The American

Economic Review, 87(2), 214-216.

Pilbeam, K. (1992). International Finance (4th_{ed.). Palgrave Macmillan.}

Pollin, R., & Zhu, A. (2006). Inflation and economic growth: a cross-country nonlinear analysis. Journal of post Keynesian economics, 28(4), 593-614.

(24)

Quispe-Agnoli, M., & Whisler, E. (2006). Official dollarization and the banking system in Ecuador and El Salvador. Economic Review-Federal Reserve Bank of Atlanta, 91(3), 55.

Rigobon, R., & Rodrik, D. (2005). Rule of law, democracy, openness, and income. Economics of transition, 13(3), 533-564.

Rose, A. K. (2000). One money, one market: the effect of common currencies on trade. Economic policy, 15(30), 08-45.

Schwartz, A. J. (1993, December). Currency boards: their past, present, and possible future role. In Carnegie-Rochester Conference series on public policy (Vol. 39, pp. 147-187). North-Holland.

Sercu, P., Uppal, R., & HULLE, C. (1995). The exchange rate in the presence of transaction costs: implications for tests of purchasing power parity. The Journal of Finance, 50(4), 1309-1319.

Stock, J.H., Watson, M.W. (2015). Introduction to econometrics (3rd_{ed.). Pearson education.}

The General Agreements on Tariffs and Trade. (July 1986). Retrieved from: https://www.wto.org/english/docs_e/legal_e/gatt47.pdf

Tsang, S. K., & Ma, Y. (2002). Currency substitution and speculative attacks on a currency board system. Journal of International Money and Finance, 21(1), 53-78.

(25)

Appendix

Below is presented a short summary of the data used in this regression. In the first table a summary of the data of the complete set of countries is provided. The second and third table make a distinction between the dollarized countries and the non-dollarized countries respectively. Each table contains for each variable; the total observations, the mean, the standard deviation, the minimum, and the maximum.

Complete data set Obs Mean Std. Dev. Min Max

Log GDP per capita 254 9.226629 0.5681657 8.153332 10.73854

Dollarization 231 0.2008217 0.3211856 0.0001061 1.227904 Log Labor 244 13.98216 1.591185 10.12663 16.47609 Investments 249 23.58664 6.025493 11.07783 46.28975 FDI 253 5.680139 4.614254 -1.270773* 23.17237 Secondary education 198 78.95051 15.51403 36.16367 120.3267 Inflation 254 6.984167 8.547919 -27.63333 45.94327 Political stability 232 0.7082328 0.1497382 0.38 1 Rule of Law 233 0.535794 .1716663 0.11 0.88 Trade 249 88.51028 25.02622 38.52093 158.3469 Government consumption 252 13.63124 3.731484 6.207438 24.91107

Dollarized countries Obs Mean Std. Dev. Min Max

Dollarization 133 0.3412205 0.3643759 0.0021131 1.227904 Log Labor 139 13.52726 1.905503 10.12663 16.47609 Investments 146 24.55307 6.190254 13.22337 46.28975 FDI 148 6.260145 5.653288 -1.270773* 23.17237 Secondary education 127 81.23688 14.00105 57.84692 111.2026 Inflation 149 6.377304 9.320548 -17.21635 45.94327 Political stability 134 0.7466418 0.1754648 0.38 1 Rule of Law 135 0.584 0.1802453 0.11 0.88 Trade 146 89.57796 26.96527 38.52093 158.3469 Government consumption 147 14.68004 3.631033 9.175027 24.97107

*FDI is measured as the net inflows (new investments minus disinvestments). A negative value implies a higher level of disinvestments compared to the new investments in a certain year.

(26)

Non-dollarized countries

Obs Mean Std. Dev. Min Max

Dollarization 98 0.0102803 0.0149093 0.0001061 0.0650969 Log Labor 105 14.58435 0.6746353 13.26653 15.7365 Investments 103 22.21673 5.19544 11.07783 36.06597 FDI 105 4.862607 2.294182 0.5324162 11.18749 Secondary education 71 74.8608 17.26426 36.16367 120.3267 Inflation 105 7.845335 7.269563 -27.63333 45.19113 Political stability 98 0.6557143 0.0797289 0.5 0.81 Rule of Law 98 0.4693878 0.1338395 0.25 0.75 Trade 103 86.99688 22.0271 49.14559 136.4898 Government consumption 105 12.16292 3.371148 6.207438 18.69974

Dollarization in developing economies : did the dollarization in developing economies contribute to higher GDP per capita growth rates?