Development aid and Dutch disease in Cambodia,
Lao PDR, Myanmar and Vietnam
Case study: Lao PDR
Quinn Veenstra
Master Thesis
Student number: 10564233 Faculty of Economics and Business MSc Economics Track International Economics and Globalization
Email: quinn.veenstra@hotmail.com Date: 15 August, 2017 Supervisor: Ms. N.J. Leefmans Second reader: Prof. dr. F. J. G. M. Klaassen Word Count: 10109
Statement of Originality This document is written by Student Quinn Veenstra, 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.
Table of contents
1. Introduction 4 2. Literature review on Dutch Disease and development aid 8 2.1 Theory on Dutch disease 8 2.2 Empirical literature 10 3. Data and Methodology 14 3.1 The model and the data 14 3.2 Tests for the data 17 4. Assessment of results 20 5. Case Study: Lao PDR 22 5.1 Lao PDR, spending effect and resource movement effect 22 5.2 Country specific circumstances 26 6. Conclusion 32 Appendices Bibliography1.
Introduction
Developed economies gave more than 150 billion dollars in official development assistance (ODA) to developing economies in 2015. These aid flows have continuously been increasing in the past few decades. ODA flows contain medium- and long-term concessional loans and grants from bilateral and multilateral sources. Obviously, development aid has a main goal of promoting economic development and welfare of the developing countries. However, improvement of the economic environment due to inflow of development aid in the recipient countries may not always be the case. The Fourth United Nations Conference on Least Developed Countries in 2011 concluded that 75% of the population in least developed countries (LDCs) still live in poverty. Furthermore, only three countries have been able to come out of this group in the past three decades. According to Burnside and Dollar (1997), White and Wignaraja (1992) and several other researchers, the past shows that development aid does not always turn out to be a success in terms of promoting economic growth. On the other hand, there are several studies that show a positive effect of aid on economic growth (Minoiu & Reddy (2010), Hansen & Tarp (1999), Gomanee, Morrissey & Verschoor (2003)). Thus, the results in the literature are very mixed. Some economists, such as Benjamin, Devarajan and Weiner (1989), even argue that ODA may have had negative effects on the economic environment of certain less developed countries due to the possible existence of the Dutch disease. The phenomenon of ‘Dutch disease’ can be an undesirable consequence of the inflow of official development assistance in developing countries. Recent research on the link between the Dutch disease and ODA flows mainly focuses on Sub-Saharan African countries, but a substantial part of ODA flows has been going to Asian countries as well, while Asian countries have not been represented as much in recent literature. The evidence on Asia being a major recipient of ODA flows in 2015 can be found in Table 1, which shows that Asian countries received almost 30% of total ODA flows. Due to the lack of research on Asian countries, this paper will take a closer look at the CLMV countries, otherwise known as, Cambodia, Lao PDR, Myanmar and Vietnam. These countries will be looked at in closer detail due to the fact that they have all been major recipients of ODA in Asia
in the last decades in both absolute and relative terms. Some data for the CLMV countries and their relationship with ODA flows in 2015 is shown below in more detail in Table 2. Moreover, research on Dutch disease in the CMLV countries due to ODA has not yet been done to my knowledge. Table 1: Position of Asia on ODA receipts in 2015 Group of countries Net ODA receipts (USD million) % of total Africa 51,036 33.46 America 10,087 6.61 Asia 45,546 29.86 Europe 6,847 4.49 Oceania 1,914 1.25 Unspecified 37,097 24.32 Total 152,527 100 Source: OECD Statistics on resource flows to developing countries http://www.oecd.org/dac/stats/statisticsonresourceflowstodevelopingcountries.htm Table 2: Position of CMLV on ODA receipts in 2015 Country Net ODA receipts
(USD million) ODA/GNI (%) % of total ODA to Asia
Cambodia 677 3.97 1.49 Myanmar 1169 1.99 2.57 Lao PDR 471 4.03 1.03 Vietnam 3157 1.73 6.93 Source: World Bank database on net ODA received and OECD statistics on resource flows to developing countries http://data.worldbank.org/indicator/DT.ODA.ALLD.CD?locations=MM http://www.oecd.org/dac/stats/statisticsonresourceflowstodevelopingcountries.htm
The Dutch disease was first mentioned by the Economist (1977), and referred to the difficulties the Dutch economy faced after the discovery of large gas reserves in the North Sea
in 1959. Due to the sudden boost in resources, exports increased a lot. As a result of the increase of gas exports, there was a huge inflow of foreign currency, leading to an increased demand for the Dutch guilder. This made the Dutch currency appreciate and loose its competitiveness in the international markets, as Dutch exports became more expensive and imports became cheaper.
The Dutch disease was primarily associated with the discovery, and thus an increase, of natural resource exports, but can also be associated with foreign direct investment, the increase of natural resource prices or in this case, foreign aid. Due to the fact that historical data for the CMLV countries under investigation are very scarce in the period before 1990, it is best to take the time period 1990-2014 for the investigation. A dynamic panel data analysis using a generalized method of moments estimator will be implemented to test what effect ODA flows have had on the real effective exchange rate (REER). A positive relationship of ODA flows with the real effective exchange rate indicates, that the REER appreciates after the inflow of ODA, thus possibly indicating Dutch disease. The precise relationship between these two variables will be discussed in more detail in the literature review.
The most recent extensive research on the link between development aid and Dutch disease is by Fielding and Gibson (2012). Their paper takes 26 countries in Sub-Saharan Africa as their sample for the period 1970 to 2009, and uses cross-sectional data to come to the conclusion that results vary significantly between countries. They state that four variables are of significance in what determines the exact impact of development aid on the real effective exchange rate. The variables are whether the country has a floating or a fixed exchange rate, the propensity to invest in non-traded goods production, the capital productivity and the aid dependence of the country in question. This will be taken into consideration in this paper, as Lao PDR will be studied in more detail to figure out what effect country specific characteristics may have had on the possible emergence of Dutch disease in this country. For the case study on Lao PDR the propensity to invest in non-traded goods production and capital productivity will be studied in closer detail. This is due to the availability of the data for this country and the lack of data for the other three countries.
This brings up the following research question: Did official development assistance cause Dutch disease effects in Cambodia, Myanmar, Lao PDR and Vietnam in the period 1990-2014 and what contribution did country specific characteristics have in the case of Lao PDR? The paper will be structured in the following way. Section II contains a literature review, describing the existing theory and empirical literature about the Dutch disease and its link to development aid. Section III presents the selection of the variables for the econometric model and the precise construction of the econometric model used to determine what effect development aid has on the emergence of the Dutch disease in the CMLV countries. This section will also include some econometric tests for the data. Section IV contains a presentation and an assessment of the results from the model. Section V contains the case study for Lao PDR, which consists of a descriptive analysis. Finally, Section VI provides an overall conclusion, and an answer to the research question.
2. Literature Review on Dutch disease and development aid
There has been extensive academic research on the topic of ‘Dutch disease’. Most of this research is based on the classical Dutch disease model, which considers a resource boom as the driving factor. However, research has also been done on the link between development aid and Dutch disease, with some papers having numerous countries as their research sample and some only looking at one particular country. All relevant previous research done on the topic of Dutch disease and development aid will be discussed in this section of the paper. First of all, a background of the theory concerning Dutch disease will be given, which will be followed by a review of the empirical literature.
2.1 Theory on Dutch disease
The first classical model describing the ‘Dutch disease’ was published by Corden and Neary (1982). The classic ‘Dutch disease’ model contains one non-tradable sector and two tradable sectors, of which one is booming and one is lagging. They argue that the economy can be affected in two ways by a sudden increase of natural resources. As one tradable sector is booming due to the resource boom, there is an increased demand for labor in that sector, which attracts production in that sector away from the lagging sector. This shift of labor is known as direct-deindustrialization. This effect is also described as the ‘resource movement effect’. The second effect is described as the ‘spending effect’. This effect occurs when there is an extra flow of revenues brought in by the booming sector, which leads to increased spending, which increases demand for labor in the non-tradable sector, at the expense of the lagging sector. This shift of labor is known as indirect-deindustrialization. As a result, there is an increase in demand for non-tradable goods, which raises the price of non-tradable goods. As prices in the tradable sector are set internationally, these prices cannot be changed (Corden & Neary, 1982). Consequently, the real exchange rate of the country in question appreciates, causing the exports to be more expensive for foreign countries to buy, and the imports for the home country to become cheaper. These changes in imports and exports both result in a worsened competitive position for the home country, thus hurting the economic growth of the country.
Although the model by Corden and Neary (1982) is based on a resource boom, a large inflow of aid can affect the economy in the same two ways as just mentioned. The inflow of aid will result in higher real household incomes, which triggers the rise of aggregate demand. Consequently, the non-tradable prices increase, which causes more resources to go to this sector away from the tradable sector. The second effect, the resource movement effect, can be studied by looking at the relationship between the prices of non-tradable goods and the prices of tradable goods. This will be of importance during the case study in section V. So in the short run, foreign aid will deteriorate the production of tradable goods through real exchange rate appreciation in the case of Dutch disease. However, it has to be noted that country specific circumstances, such as the capital productivity of the country in question can be of importance as to how foreign aid affects the real exchange rate. Following an increase in aid flows any investment will be more productive the higher the capital productivity. Thus a higher capital productivity will slow down the real exchange appreciation to a greater extent. Thus an increase of the inflows of aid into a country does not always automatically mean that the real exchange rate will appreciate. In many cases it is important to observe how the government in question deals with and distributes the received official development assistance. According to Hevia and Nicolini (2015) we should keep in mind that the Dutch disease in its first phases is not really a disease. The first signs of a real exchange rate appreciation are just the optimal response of prices and quantities to the relative price shock of non-tradable goods. However, the Dutch disease phenomenon comes up when the reaction to this change in relative prices is not handled in an adequate manner, and thus hurting the tradable sector.
2.2 Empirical literature
As mentioned, the Dutch disease does not only occur as a result of an increase in a country’s resources, but can also be associated with the inflow of official development assistance. This section will take a closer look at previous research on the link between foreign aid and the Dutch disease. Younger (1992) takes a look at Ghana and the relationship of development aid with the Dutch disease in the period 1981-1987. The paper illustrates that Ghana experienced an appreciation of the real exchange rate after the inflow of official development assistance, caused through domestic inflation rather than a nominal appreciation. Younger (1992) further mentions the problem that almost all of the aid goes towards the public sector instead of the private sector, resulting in the crowding out of the private sector, and aggregate demand for domestic products increasing. What followed was an attempt to counter the increase in aggregate demand by the Ghanaian authorities by a tight monetary policy, decreasing investments, and thus hurting the economy. He concludes that whether aid has had an effect on the emergence of Dutch disease is questionable, as government policies may have been the reason behind the economic slowdown. This indicates that the shifts of production between the different sectors may not always exclusively be a result of a real exchange rate appreciation. This paper shows why it is important to include the case study on Lao PDR, as country specific characteristics, such as government policies in this case, can have an important effect on the relationship between official development assistance and the real exchange rate.
Likewise, Sackey (2001) takes a look at the impact of aid on the real exchange rate of Ghana in the period 1962-1996. He finds that development aid leads to a depreciation of the real exchange rate, which goes against the possibility of the Dutch disease being present in the Ghanaian economy. In addition, Nyoni (1998) finds that aid leads to a real exchange rate depreciation in Tanzania for the period 1967-1993 using an OLS regression, thus also going against the proposition that aid may have caused Dutch disease. Although there was a significant increase of ODA flows, the real exchange rate depreciated sharply. Real exchange rate appreciation is the main channel through which aid affects the tradable sector, and as it is absent in this case, it contradicts the Dutch disease model.
On the other hand, there are also papers that show that development aid and the Dutch disease can be linked. Adenauer and Vagassky (1998) find a direct relationship between ODA flows and real exchange appreciation for four CFA countries during the period 1980-1993. They make use of the Generalized Least Squares method which takes heteroscedasticity and time series autocorrelation into account. Their model looks as follows:
ln(REER) = C + 𝐵 ln(RGDP) + 𝑋 GDIF + 𝑍 ln(TOT (-1)) + 𝑌 ln(RODA) + 𝑊 ln(RODA (-1))
REER is the real effective exchange rate, RGDP is the real GDP (which considers total factor supplies), GDIF is the average of the current and once lagged growth rate difference between the country under consideration and industrial countries (which takes into account the Balassa-Samuelson effect of differences in technological levels), TOT is the terms of trade, and RODA is the net received official development assistance. They suggest that, due to the increase in ODA flows, trade balances widened for these four countries. Furthermore, they obtained a positive statistical significant coefficient for both RODA and the lagged value of RODA, thus indicating that an increase in official development assistance results in the appreciation of the REER. This seems to support the idea of the Dutch disease. However, Nkusu (2004) argues that the authors did not put enough emphasis on the fact that the economic performance of the four CFA countries may have been affected by certain developments on the world market for their primary exports, and the appreciation of the French franc against the dollar during the late 1980s, as these countries had a fixed exchange rate with the franc. Real export figures could have helped to determine whether the trade balances were driven by declining world prices, declining trade volumes or both. The variables that will be used in this paper to search for Dutch disease in the CMLV countries will be based on the variables that Adenauer and Vagassky (1998) use in their paper. However, an important variable will be included in the model, which is a variable that represents the monetary policy, as this has been proven to also affect the real effective exchange rate.
In addition, White and Wignaraja (1992) find a direct relationship between total ODA flows and real exchange rate appreciation using an econometric model of real exchange rate
behavior for Sri Lanka in the period 1974-1988. The disappointing performance of the manufacturing sector is associated to the real exchange behavior of Sri Lanka according to the authors, thus supporting the Dutch disease theory.
Lartey, Mandelman and Acosta (2008) investigate the Dutch disease effects of remittances, rather than official development assistance. They use a panel data analysis for 109 developing and transition economies for the period 1990-2003 to test whether rising levels of remittances cause spending effects that lead to the appreciation of the real exchange rate and resource movement effects that favor the non-tradable sector at the expense of the tradable sector. They come to the conclusion that rising levels of remittances can have an important spending effect that culminates in real exchange rate appreciation. Furthermore, a resource movement effect also follows after an increase in remittances. This indicates that the share of services, which represents the non-tradable sector, in total output rises. On the other hand, the share of manufacturing, which represents the tradable sector, declines. These are both characteristics of the Dutch disease. In the case study for Lao PDR a descriptive analysis will be given by similarly looking at the figures for the tradable and non-tradable sector, while keeping the changes in aid flows in mind. This will give an indication of the presence of resource movement effects, and thus Dutch disease effects in Lao PDR.
Finally, Fielding and Gibson (2012) did research on the effect of aid on 26 Sub-Sahara African countries by using a vector auto regression (VAR) model consisting of time-series data. They come to their estimates by fitting either of two types of equations, depending on the exchange rate regime (fixed or floating) of the country under consideration, to the data using OLS. Using the parameters of the fitted model, it is possible to plot the response of each dependent variable in the system to a percentage point increase in the the aid variable. They include a table to their paper that summarizes the results by reporting the estimated response of GDP and the real exchange rate to a percentage point increase in foreign aid in the year of the increase (immediate impact) and 16 years after the increase (long-run impact). They came to the conclusion that there exists a considerable degree of heterogeneity across Sub-Saharan Africa. In most of the countries investigated, the inflow of aid leads to a real exchange rate
appreciation. However, the size of these effects differs significantly between countries, and some countries even experienced a depreciation of the exchange rate due to the inflow of aid. Fielding and Gibson (2012) argue that this shows why previous studies on Dutch disease have had so many varied results, when it comes to the investigation of one particular country or a small group of countries. They continue by doing a regression designed to explain the cross country variations in the size of the real exchange rate responses which were found by the VAR model. They use as explanatory variables: whether a country has a fixed or floating exchange rate regime, the propensity to invest in non-traded goods production, the capital productivity, and aid dependence. The variables are designed to capture some of the features of a recipient country that might be associated with the magnitude of the response of the real exchange rate to ODA inflows. According to the authors, these variables have a significant impact on whether or not aid has a Dutch disease effect and a significant impact on the size of the Dutch disease effect. The last mentioned paper by Field and Gibson (2012) shows that it is important to look at the country specific circumstances, as country specific circumstances, such as the just mentioned variables, could have an important influence on how official development assistance affects the real exchange rate. Therefore, Lao PDR will be examined in a closer manner to assess what exact policies could avoid the depreciation or appreciation of the country’s currency due to the inflow of development aid by using a descriptive analysis.
3. Data and Methodology
This section will go into further detail about the methodology and data that will be used to test whether there has been any Dutch disease effect in Cambodia, Myanmar, Lao PDR and Vietnam. Firstly, the model will be explained and presented. Secondly, the different variables will be explained in closer detail. Lastly, the different tests needed to describe the validity of the variables in the model will be presented. 3.1 The model and the data As mentioned earlier, dynamic panel data will be used for Cambodia, Myanmar, Lao PDR and Vietnam for the period 1990-2014. The Generalized method of moments estimator will be applied, to test for the relationship of ODA flows with the real effective exchange rate. The regression model that will be used will look as follows:
ln(REERit)= '()* 𝛽j ln(ODAGDPi,t-j) + 𝛽 ln(GDPCit) + 𝐶 ln(TOTit) + 𝐷 ln(RESMit) + 𝐸 GDIF + 𝜀it
REER stands for the real effective exchange rate, ODAGDP stands for the official development assistance inflows as a percentage of GDP, GDPC stands for the real gross domestic product per capita, TOT stands for the terms of trade, RESM stands for the foreign exchange reserves in millions of current US dollars, and GDIF stands for the difference in GDP growth rates between the country in question and the rest of the world. The equation above is based on Adenauer and Vagassky (1998) and Lartey, Mandelman and Acosta (2008). However, GDP per capita is used instead of the GDP. When GDP is weighted by population it gives an even more precise view of a country’s production. Furthermore, the variable for foreign exchange reserves is included as a proxy for monetary policy, as the lack of the inclusion of monetary policy was an important point of criticism on the paper of Vagassky and Adenauer (1998) by Nkusu (2004). The use of logs in the equation is widely used in the literature, thus also implemented in the equation in this paper.
Starting with the dependent variable, the real effective exchange rate. The REER is an index that represents the nominal effective exchange rate adjusted for relative changes in consumer prices, a proxy of cost indicators of the home country. The nominal effective exchange rate is an index that describes the strength of the home currency relative to a basket of currencies of trade partners of the country in question. As the real effective exchange rate is described as the relative price of domestic to foreign goods, an increase in the REER means there is an appreciation. The data for the real effective exchange rate are retrieved from the Bruegel database and are based on annual figures and 67 frequent trading partners for the country in question. The base year for each country is 2007, so the REER in 2007 is equal to 100.
Now to continue with the independent variable, which is the main focus of this paper, official development assistance (ODAGDP). The variable for official development assistance is measured by the net official development assistance and official aid received as a percentage of the gross domestic product. The data are retrieved from the United Nations database. A positive sign for the coefficient of the ODAGDP variable would mean that a percentage increase in official development assistance divided by GDP would increase the real effective exchange rate by 𝛽j %. This would indicate that there is a presence of the Dutch disease phenomenon. To capture the delayed effect of ODA flows on the real effective exchange rate, a lag length of 𝑝 is allowed for in some of the specifications. Figures for each country concerning the ratio of ODA to GDP are given in Appendix 1.
The rest of the variables are control variables and are known determinants of the real effective exchange rate. The first one is the gross domestic product per capita (GDPC). An increase in GDP per capita, causes an increase in incomes, and this results in an increased demand for non-tradable goods. Consequently, the real effective exchange rate will appreciate. This indicates that a positive relationship is to be expected between REER and GDPC in the model (Lartey, Mandelman & Acosta, 2008). The data for the real gross domestic product and the population of the country in question is retrieved from the World Bank database.
The second control variable is the terms of trade (TOT). The terms of trade is calculated by dividing the price of exports by the price of imports. Variations in the relative prices have clear
implications for the real effective exchange rate as an increase in the price of exports relative to imports, leads to an appreciation. This is due to the fact that an increase of exports relative to imports implies that there is an increasing demand for the country’s exports. This in turn means that there is an increase in revenue from exports resulting in an increase in demand for the country’s currency, which leads to the appreciation of the exchange rate (Lartey, Mandelman & Acosta, 2008). Thus a positive sign is expected for this variable. The base year for each country is the year 2000, so for each country the value for the terms of trade in the year 2000 is 100. The data for the terms of trade are retrieved from the World Bank database.
The third control variable is the international reserves in millions of current US dollars (RESM), which in turn represents monetary policy of the country in question. The Central Bank of the country in question could be able to limit a real exchange rate appreciation to some degree by buying up foreign exchange in exchange for domestic currency, and thus increasing its foreign exchange reserves (Oomes & Kalcheva, 2007; Ebrahim-Zadeh, 2003). This indicates that international reserves are an important determinant of the real effective exchange rate and thus is included into the regression. The data for international reserves are also retrieved from the World Bank database.
The last control variable is the difference in GDP growth rates between the country in question and the rest of the world (GDIF). A higher GDP growth indicates a higher wage growth, which leads to leads to real exchange rate appreciation. This captures the Balassa-Samuelson effect, which suggests that an increase in wages in the tradable sector of an emerging economy must also mean that the wages in the non-tradable sector will increase (Balassa, 1964; Samuelson, 1964; Vagassky & Adenauer, 1998). As a result, a positive relationship between GDIF and REER is to be expected in the model. Data for the growth rates of the sample countries and the world growth rates are retrieved from the World Bank database. Furthermore, the error term 𝜀it in the equation is assumed to be serially uncorrelated and
independent across all panels.
3.2 Tests for the data
As all variables in the model are now explained as to what they indicate and why they are of relevance in determining the real effective exchange rate, it is important to test whether the series are stationary at levels or stationary when first differenced. It is important to check whether the cross-sectional time series contain unit roots before any analysis has even begun. A series is stationary when historical relationships can be generalized to the future, thus when its probability distribution does not change over time. Otherwise, the series has a trend component which results in the series being nonstationary. If a series is nonstationary at level, it normally will be stationary at first difference, and then can be used in a panel data regression. The statistical properties of most estimators in a time series rely on the data being stationary in some degree (Stock & Watson, 2011). In Stata 14, for every perfectly balanced variable under consideration a Levin-Lin-Chu unit root test is done for both levels and first differences to test whether the variables are stationary or nonstationary. For the unbalanced variables, TOT and RESM (some countries had some missing data) a Fisher-type unit root test is done, again for both levels and first differences. The Fisher-type unit root test can deal with unbalanced panel data, whereas the Levin-Lin-Chu test can not. The results of these unit root tests can be found in Appendix 2. For the Levin-Lin-Chu test done for REER, ODAGDP, GDPC and GDIF, the null hypothesis states that the panels contain unit root and the alternative hypothesis states that the panels are stationary. For all individual variables in level terms that are done with this test, the null hypothesis can not be rejected, thus indicating that they all contain unit root. The test for their first differences, however, show that they are stationary at first difference so, in the regression, all these variables should be included as first differences. For the Fisher-type unit root test for TOT and RESM, the null hypothesis states that the panels contain unit root and the alternative hypothesis states that at least one panel is stationary. For the variable ln(RESM) it can be assumed that the variable is stationary at levels, while for ln(TOT) this is questionable.
Most of the series that are necessary to come to any results are non-stationary. In order to solve the issue of unit roots in the data series, difference Generalized Method of Moments
(GMM) will be used to estimate our regression model. Difference GMM automatically takes the first differences of all variables, thereby solving the issue of unit roots. Furthermore, difference GMM is a fitting method to this paper due to the possibility that some of the explanatory variables might be correlated with the error term. Difference GMM allows for the use of lagged differences or lagged levels of the explanatory variables as instruments for the endogenous variables.
For the lagged differences of the explanatory variables to be valid instruments they have to fulfill two conditions. Firstly, the differences of the explanatory variables and the errors are uncorrelated. Secondly, there should be no serial correlation in the errors. The validity of the instruments is crucial in determining whether the GMM estimator is consistent or not. If valid, the GMM coefficient will capture the immediate impact of the isolated exogenous component of the covariates on the dependent variable (Lartey, Mandelman & Acosta, 2008; Arellano & Bover, 1995; Blundell & Bond, 1998). To determine whether the instruments are valid, two specification test have to be done. These are a test of over-identifying restrictions and a test for second-order serial correlation in the error term. The first test is done by the standard Sargan test of over-identifying restrictions. When a number of moment conditions is greater than the dimension of the parameter vector, the model is said to be over-identified. Over-identification gives the possibility to check whether the model’s moment conditions match the data well or not. This test has a null hypothesis of that the instruments are overall valid, thus should be accepted. The second test is the Arellano and Bond (1991) test for second-order serial correlation. This test has a null hypothesis that states that there is no second-order serial correlation in the differenced error term, otherwise known as the residual of the equation in differences. If there would be a presence of serial correlation of the error term, it would follow a moving average process of at least order one (Lartey, Mandelman & Acosta, 2008). Both these tests, with the objective of proving the validity of the instruments, were done in Stata and the results can be found in Appendix 3. The Sargan test of over-identifying restrictions shows that the p-value is 0.7080, thus not lower than 0.05. This indicates that the null hypothesis, which states that the over-identifying restrictions are valid, can not be rejected. As
a result, the instruments are valid and therefore can be used in the model. The test for second-order serial correlation gives a p-value of 0.2335 for AR(1) and a p-value of 0.4305 for AR(2). Thus, in both cases the null hypothesis, which states that there exists no autocorrelation, can be rejected.
In Stata 14, two types of dynamic panel data estimations are possible. The Arellano-Bond estimation indicating the Difference GMM method, and the Arellano-Bover/Blundell-Bond estimation, indicating the System GMM method. If the results from the two tests both reject the null hypothesis the difference GMM method can be used. As Appendix 3 shows that both tests reject the null hypothesis, the Difference GMM method will be used in Stata on the panel data at hand.
4. Assessment of results
This part of the paper will present the results obtained from estimating our panel data model using the generalized method of moments estimator. Firstly, a table of results obtained from Stata will be presented, and then explanations of these coefficients will be given.
Table 3: Difference GMM estimation
Variable
Coefficient
Standard
error
z-statistic
p-value
lnODAGDP
0.0240***
0.0067
3.56
0.000
lnTOT
0.0198
0.0636
0.31
0.756
lnGDPC
0.4466***
0.0533
8.37
0.000
lnRESM
-0.0101
0.0081
-1.25
0.212
GDIF
-0.0051
0.0048
-1.07
0.286
Observations
96
Note: the dependent variable is the the logarithm of the real effective exchange rate. The instruments used include the lagged differences and second lagged differences of the logarithms of TOT, GDPC, RESM, and GDIF. The symbols ***, **, * denote the significance of the t-statistic at a 1, 5 and 10 percent level, respectively.
Table 3 above shows the results of the difference GMM estimation. The logarithm of the real effective exchange rate is the dependent variable. The results show that the coefficient of the
point increase in ODA/GDP will lead to a 0.0240% increase in the real effective exchange rate. Thus, this significant positive coefficient indicates that increases in ODA/GDP lead to an appreciation of the real effective exchange rate. This is a sign of the presence of the Dutch disease phenomenon in the CLMV countries.
The other significant variable in the model is the logarithm of the gross domestic product per capita. The gross domestic product per capita has a positive coefficient, indicating that a one percentage increase in GDP per capita results in a 0.4466% increase in the real effective exchange rate. This is in line with the theory, as an increase in income per capita raises the demand for non-tradable goods, which in turn causes the real exchange rate appreciation. The terms of trade variable has a positive coefficient, indicating that a one percentage increase in the terms of trade leads to a 0.0198% increase in the real effective exchange rate. This is in line with the theoretical background, which predicts a positive coefficient for this variable. However, the coefficient for the variable is not significant at a 5% significance level. The coefficients for the logarithm of the foreign exchange reserves in million USD and the growth differential are not significant, and do not have the expected sign as described in section 3.1.
5. Case Study: Lao PDR
In this section, a case study will be done on Lao PDR. The section will be split into two different parts. The first part will provide a general description of the economy of Lao PDR and a more in depth description of the spending effect, which looks at the relation between ODA flows and the real effective exchange rate, and the resource movement effect, which looks at the relation between ODA flows and the tradable to non-tradable output. The second part will be an analysis on country specific circumstances regarding the inflow of official development assistance in Lao PDR. All these analyses will be of a descriptive nature, using graphs, charts and other sources. 5.1 Lao PDR, spending effect and resource movement effect Lao PDR, a lower-middle income economy, has been one of the fastest growing economies in the Eastern Asia and Pacific region, with an average GDP growth of 8% in the last decade. In 2015, it was the 13th fastest growing economy in the world. Due to these high growth rates, the government faces the difficult task of distributing and translating the new and growing wealth into inclusive and sustainable human development. Furthermore, the general management of this new wealth is a difficult matter for the government. Although it is still a relatively poor country, the significant gains made in the economy and the social sectors led to continual improvements in human development, as poverty rates declined from 46% in 1992 to 23% in 2015 as stated by the United Nations Development Program (2017).
Furthermore, Lao PDR has been aid dependent in the last decades and still is to this day, as was shown in table 2 in the introduction and can also been seen in Appendix 1. ODA receipts consisted of 4.03% of the GNI for Lao PDR in 2015. This brings up the potential of Dutch disease. The following graph shows the relationship of the real effective exchange rate and the ODA flows. Dutch disease theory suggests that increase in ODA flows into a country results into the appreciation of the REER. For Lao PDR there exists a significant relationship between ODA flows and the appreciation of the REER, as this was found in the panel data analysis in section 4. Thus there seems to be a presence of a spending effect. By looking at graph 1, the REER and
ODA flows seem to be following the same pattern up until 2006. After 2006, the ODA flows increase rapidly till 2008 and then decrease slowly. The real effective exchange rate keeps on appreciating with the same trend up until 2014. What follows in subsection 5.2 is an explanation that will be given as to how the paths of the REER and ODA flows at first stayed at relatively the same trend, and then drifted apart in 2006, thus taking a closer look at what may have caused this drift apart. Graph 1: REER and ODA flows for Lao PDR
Note: The left vertical axis represents the values for the real effective exchange rate and the right vertical axis represents the values for ODA flows in millions
Source: Made by author using data from the UN, World Bank en Bruegel databases
Another important feature of the Dutch disease phenomenon is the resource movement effect, the effect of aid on the productive sector of the economy. This effect can best be captured by looking at the relationship of the tradable to non-tradable (TNT) output. According to the Dutch disease theory an increase in ODA flows to a country should lead to a shift of resources to the non-tradable sector and away from the tradable sector. It is important to also look at the TNT output, instead of just the real effective exchange rates, as this is considered as a more accurate approach to look at the real effects of the Dutch disease according to Lartey, Mandelman and Acosta (2008). This is now only done for Lao PDR and not for Cambodia,
0 100 200 300 400 500 600 0 20 40 60 80 100 120 140 160 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 REER ODA
Myanmar and Vietnam in section 3, because data on TNT output for the rest of the countries are very scarce or for some not even available. The tradable sector in the equation is measured by the share of the manufacturing sector plus the agricultural sector, and the non-tradable sector is measured by the share of the service sector in the economy. It was done in this manner as there exists an absence in data on traded and non-traded goods output. However, doing it this way still gives a clear view of the relationship of the tradable goods sector and non-tradable goods sector. In graph 2, the TNT output and ODA flows for Lao PDR is compared. A decrease in the value of TNT output indicates that the tradable sector is declining relative to the non-tradable sector, thus indicating movement of resources across the sectors. Graph 2 shows a continuing decrease of TNT output, while ODA flows keep on increasing. Thus there is evidence of a resource movement effect in Lao PDR in the period 1990-2014 according to graph 2, thus indicating the presence of a Dutch disease effect.
The government of Lao PDR has generally focused its efforts on the extremely poor in the country by directing social payments and poverty reduction measures to this part of the population. These government expenditures are backed up by the inflow of official development assistance, resulting in higher wages for a big part of the population. In 1990 approximately 61% of GDP was related to agriculture while in 2004 this had decreased to 39%, while the service sector went from 24% of GDP in 1990 to 41% in 2003. It can be argued that in the period 1990-2003 there was a significant resource movement effect away from the agricultural sector towards the service sector. However, in the period 2003-2014 the service sector only grew to 45% of GDP. This indicates that the resource movement effect stagnated. This can clearly be seen in graph 2 below.
Graph 2: Tradable to non-tradable output and ODA flows for Lao PDR Note: The left vertical axis represents the values for the tradable to non-tradable output and the right vertical axis represents the values for the percentage of ODA flows divided by GDP. Source: Made by author using data from the World Bank database and World Development Indicators Subsection 2 will go into further depth on the country specific circumstances. A more in depth description is needed as to how the data shown in graph 1 and 2, move the way they do. So the two graphs show that for both the T/NT output and the REER, ODA flows seem to have had a significant Dutch disease effect. Thus a resource movement effect and a spending effect seem to have been present. It is important to determine why ODA flows seem to have a significant effect on the decrease of the T/NT output, but the relation between the appreciation of the REER and ODA flows come to a halt after 2006. 0 100 200 300 400 500 600 0 50 100 150 200 250 300 350 400 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 T/NT ODA
5.2 Country specific circumstances
Fielding and Gibson (2012) stated in their paper that it is very important to have a closer look at country specific circumstances as these can be vital in explaining how the real exchange rate reacts to an increase in aid flows. Based on Fielding and Gibson (2012) this thesis will investigate the following variables that can have an effect on how the REER reacts to an increase in aid inflows: the propensity to invest in non-traded goods production and capital productivity. These variables might help a bit more in explaining the relationship of ODA flows with the real effective exchange rate in graph 1.
We start with a more in depth explanation of each variable and how they could be related with the REER and ODA flows. Firstly, the propensity to invest in non-traded goods production is an important variable in determining how aid affects the real exchange rate, due to the fact that aid that raises the capacity in the non-traded goods sector is likely to cause some real exchange rate depreciation. On the other hand, aid that raises the capacity in the traded goods sector is not going to have a direct effect on relative prices, as was mentioned in the literature review. Data and figures are not available to measure the exact proportions of aid directly invested into the non-tradable sector for Lao PDR. However, as a proxy one can also look at the gross fixed capital formation (GCF) to GDP. This shows how much of new value added into the economy is invested rather than consumed. The proportion invested in fixed capital is believed to be correlated with the marginal propensity to invest in non-traded goods capital (Fielding and Gibson, 2012). The following graph 3 illustrates the gross fixed capital formation to GDP, with REER and AID also included in the graph for the period 2000-2014.
Graph 3: GCF to GDP, REER and ODA flows for Lao PDR, 2000-2014 Note: The left vertical axis represents the values for the real effective exchange rate and the values for the gross fixed capital formation divided by GDP and the right vertical axis represents the values for ODA flows in millions Source: Made by author using data from the UN, World Bank en Bruegel databases Graph 3 illustrates that the gross fixed capital formation to GDP increased as ODA flows increased, and the REER appreciated. What can be said about this figure is that due to the rise in gross fixed capital formation to GDP, a relatively bigger proportion of investments went into non-traded goods. According to the theory the real effective exchange rate is supposed to depreciate, or at least appreciate at a slower rate, following increases in gross fixed capital formation, as this raises the capacity in non-traded goods. Thus, it has to be concluded that gross fixed capital formation does not sufficiently explain why the REER appreciated at a faster rate after 2006.
Now to continue with the second variable, capital productivity. As was mentioned in the literature review, following an increase in aid flows any investment will be more productive the higher the capital productivity. Thus a higher capital productivity will slow down the real exchange appreciation to a greater extent. For Lao PDR exact capital productivity statistics are not available, thus three country characteristics can be used that are said to be correlated with capital productivity, mainly GDP per capita, the openness of the country to international trade, and the institutional quality of the country (Fielding and Gibson, 2012). The first country characteristic said to be correlated with capital productivity is GDP per capita. A higher real GDP 0 100 200 300 400 500 600 0 20 40 60 80 100 120 140 160 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 REER GCF ODA
per capita will reflect a higher level of capital productivity, for a given propensity to invest. Lao PDR went from a GDP per capita of approximately 200 USD in 1990, to a GDP per capita of approximately 1980 USD in 2014, as can be seen in graph 4. This of course is a huge increase. Graph 4: Lao PDR GDP per capita, 1990-2014 Source: Made by author using data from the World Bank en Bruegel databases
This increase could explain why the REER did not appreciate as much after the huge relative increases of ODA in the period 2004-2008. Capital productivity levels may have been affected enough by the rising levels of GDP per capita to mitigate the appreciation of the REER, relative to the increases of ODA flows into the country. The second country characteristic said to be correlated with capital productivity is the openness of the country to international trade. This can be measured by exports plus imports divided by GDP. Grossman and Helpman (1991) have argued that countries that are more open to international trade can more readily absorb new productivity enhancing technologies, which should increase capital productivity, again mitigating the appreciation of the real effective exchange rate. Edwards (1998) and Alcala and Ciconne (2004) have published and discussed evidence on the link between openness to trade and productivity, and both conclude that there is a correlation between how open a country is to trade and the level of capital productivity. Thus this variable is useful in studying the capital productivity for Lao PDR. For Lao PDR, exports 0 500 1000 1500 2000 2500 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
plus imports divided by GDP went from approximately 36% in 1990 to 90% in 2014, as shown below in graph 5. With the world average being at 60% of GDP, Lao PDR is relatively open to international trade. According to the theory this would mean that the country should be able to absorb new productivity enhancing technologies, dampening the appreciation of the REER. With the same reasoning as with GDP per capita, this could indicate why the appreciation of the REER did not appreciate as fast since 2006 as ODA inflows were rising on a yearly basis. In the years before 2006 there were large increases in the openness to international trade, which could have led to the technological improvements that caused the ODA flows and the REER to shift away from each other from 2006 onwards. Graph 5: Exports plus imports divided by GDP for Lao PDR, 1990-2014 Source: Made by author using data from the World Bank en Bruegel databases
The third country characteristic said to be correlated with capital productivity is the institutional quality of the country in question. Acemglu, Johnson and Robinson (2005) present an empirical case on the link between political institutions and productivity. They find that some characteristics of governmental institutions are more likely to have a direct effect on productivity according to the literature. These are characteristics relating to the taxation of businesses and the formation of public economic policy. Fielding and Gibson (2012) use two indicators for institutional quality as a proxy for capital productivity and we use the same ones 0 10 20 30 40 50 60 70 80 90 100 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
as they use in their paper. The first indicator is government effectiveness, which measures ‘the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies’. The second indicator is regulatory quality, which measures ‘the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development’. These indicators are obtained from the World Bank’s Worldwide Governance Indicators. Graph 6: Government effectiveness and regulatory quality of Lao PDR, 1996-2014 Source: Made by author using data from the World Bank’s Worldwide Governance Indicators
Both indicators are measured on a scale from -2.5, indicating weak governance performance, to 2.5, indicating strong governance performance. From 2005/2006 onwards there has been an upward trend for both of the indicators. According to Fielding and Gibson (2012) this should mean that the improvement of the political institutions should have an effect on capital productivity and efficiency. The rise in these indicators, specifically since 2005/2006, could also explain why the real effective exchange rate kept on appreciating in a steady pace, while ODA flows were increasing and decreasing in a rapid pace on a yearly basis since 2006. -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 GE RQ
According to Fielding and Gibson (2012) the change in the indicators for government effectiveness and regulatory quality can take up to three years to see its effects on the relationship of ODA flows and the REER. In sum this section has shown, why it is important to take a closer look at the propensity to invest in non-traded goods production and the capital productivity of Lao PDR. Especially, the three country characteristics that are correlated with capital productivity have shown why the real effective exchange rate has moved in the way it has for Lao PDR.
6. Conclusion
This paper examined the impact of increases in official development assistance on the real effective exchange rate for Cambodia, Lao PDR, Myanmar and Vietnam. The main goal of the paper was to figure out if the aid flows have had any Dutch disease effects for this sample group of countries in the period 1990-2014. For the sample a dynamic panel model is set up that is estimated using a generalized method of moments estimator. The coefficient for ODA flows as a percentage of GDP has a positive significant value and indicates that a percentage increase in ODA flows to GDP will result in a 0.0240% appreciation of the real effective exchange rate. This shows that there is a presence of a spending effect, which is part of the Dutch disease phenomenon.
For Lao PDR a case study was set up to determine whether there is a presence of both a spending effect and a resource movement effect. Moreover, the case study was set up to have a closer look at whether the propensity to invest in non-traded goods production and the capital productivity of Lao PDR were of importance in determining how official development assistance flows have affected the real effective exchange rate. For Lao PDR it can be concluded, by looking at the graphs that there exists an overall positive relationship between official development assistance flows and the real effective exchange rate, which represents the spending effect. And there exists a negative relationship between the tradable to non-tradable output and official development assistance flows, which represents the resource movement effect. Both effects show that there was a presence of Dutch disease in Lao PDR for the period 1990-2014. By looking at the two country specific variables, more can be said about how and why the real effective exchange rate reacts the way it has done in the last decade. Firstly, looking at the propensity to invest in non-traded goods gives no clear image as to why the real effective exchange rate moves the way it does. However, by looking at the capital productivity, for which three different types of proxies are used, useful information can be extracted as to why the real effective exchange rate for Lao PDR has appreciated in the last twelve years. The real gross domestic product per capita, the trade openness and the institutional quality of Lao PDR can describe the relationship of official development assistance
flows and the appreciation of the real effective exchange rate. Each variable has been increasing for Lao PDR since 2006, which can explain why the official development assistance flows started having a different effect on the real effective exchange rate since 2006.
Data for many of the variables studied in this paper has not been available for very long for Cambodia, Lao PDR, Myanmar and Vietnam. Therefore, it was not possible to study a longer time period. Furthermore, only Lao PDR was looked at in closer detail as this was the only country for which data was available concerning the country specific circumstances. Further research will be able to study developing economies and the presence of Dutch disease in more detail as more precise data will become available throughout the upcoming years. However, research on the topic of Dutch disease should always be done on one specific country or a group of countries that are very similar, as countries that are different in their fundamentals are effected in different ways by the inflow of official development assistance.
Appendix 1: Charts describing REER and ODA/GDP historical data for Cambodia,
Lao PDR, Myanmar and Vietnam
0 1 2 3 4 5 6 7 0 20 40 60 80 100 120 140 160 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Real effective exchange rate and ODA/GDP
historical data for Vietnam
REER ODA/GDP 0 5 10 15 20 25 0 20 40 60 80 100 120 140 160 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014Real effective exchange rate and ODA/GDP
historical data for Lao PDR
REER ODA/GDP
Note: The left vertical axis represents the values for the real effective exchange rate and the right vertical axis represents the values for the percentage of ODA flows divided by GDP. Source: Made by author using data from the UN, World Bank en Bruegel databases
0 1 2 3 4 5 6 7 0 50 100 150 200 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Real effective exchange rate and ODA/GDP
historical data for Myanmar
REER ODA/GDP 0 5 10 15 20 0 20 40 60 80 100 120 140 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014Real effective exchange rate and ODA/GDP
historical data for Cambodia
REER ODA/GDPAppendix 2: Unit root tests for all variables
LLC test for series at levels
LLC test for series at first differences
lag
t-statistic
p-value
lag
t-statistic
p-value
ln(REER)
D.ln(REER)
0
-1.5539**
0.0601
0
-9.2863***
0.0000
1
1.0052
0.8426
1
-2.6400***
0.0041
ln(ODAGDP)
D.ln(ODAGDP)
0
-0.1855
0.5736
0
-10.8156***
0.000
1
1.4548
0.9271
1
-4.9426***
0.000
ln(GDPC)
D.ln(GDPC)
0
0.4065
0.6578
0
-6.3363***
0.0000
1
1.1488
0.8747
1
-1.4988*
0.0670
GDIF
D.GDIF
0
-5.6954***
0.000
0
-20.5637***
0.000
1
-4.8473***
0.000
1
-2.9401***
0.0016
Note: The null hypothesis states that the series contain unit root, thus the rejection of the null hypothesis indicates that the variable is stationary. The symbols ***, **, * denote the significance of the t-statistic at a 1, 5 and 10 percent level, respectively.
Fisher-type unit root test at levels
Fisher-type unit root test at first
differences
Lag
P-statistic
p-value
lag
P-statistic
p-value
ln(RESM)
D.ln(RESM)
0
24.0167***
0.0023
0
127.6018***
0.0000
1
2.6863
0.9525
1
42.2881***
0.0000
ln(TOT)
D.ln(TOT)
0
8.6399
0.3736
0
48.1010***
0.0000
1
8.8285
0.3570
1
42.7156***
0.0000
Note: The null hypothesis states that all panels contain unit roots, and the rejection of the null hypothesis states that at least on panel is stationary. The symbols ***, **, * denote the significance of the t-statistic at a 1, 5 and 10 percent level, respectively. The P-statistic indicates the statistic for the inverse chi-squared.
Appendix 3
Arellano-Bond test for zero autocorrelation in first-differenced errors Order z-statistic Prob > z
1 -1.1914 0.2335 2 0.78829 0.4305 Note: null hypothesis states no autocorrelation Sargan test of over-identifying restrictions chi2(65) 37.49986 Prob > chi2 0.7080 Note: null hypothesis states that over-identifying restrictions are valid
Bibliography
Acemoglu, D., Johnson. S. & Robinson, J.A. (2005) “Institutions as a fundamental cause of long- run growth” Handbook of Economic Growth, 1(1), pp. 385-472. Adenauer, I. & Vagassky, L. (1998) “Aid and the real exchange rate: Dutch disease effects in African countries” Intereconomics, 33(4), pp. 177-185. Alcala, F. & Ciconne, A. (2004) “Trade and Productivity” The Quarterly Journal of Economics, 119(2), pp. 613-646. Arellano, M. & Bond, S. (1991) “Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations.” Review of Economic Studies, 58(2), pp. 277-297. Arellano, M. & Bover, O. (1995) “Another look at the instrumental variable estimation of error component models.” Journal of Econometrics, 68(1), pp. 29-51. Balassa, B. (1964) “The Purchasing-Power Parity Doctrine: A reappraisal.” Journal of Political Economy, 72(6), pp. 584-596 Benjamin, N. C., Devarajan, S. & Weiner, R. J. (1989) “Dutch Disease in a Developing Country: Oil Reserves in Cameroon.” Journal of Development Economics, 30(1), pp. 71–92. Blundell, R. & Bond, S. (1998) “Initial conditions and moment restrictions in dynamic panel data models.” Journal of Econometrics, 87(1), pp. 115-143. Burnside, C. & Dollar, D. (1997) “Aid, Policies, and Growth.” World Bank Policy Research Working Paper No. 569252. Available at SSRN: https://ssrn.com/abstract=569252 Corden, M. W. & Neary, P. J. (1982) “Booming Sector and De-Industrialization in a Small Open Economy.” The Economic Journal, 93(386), pp. 825–848.Ebrahim-Zadeh, C. (2003) “Dutch disease: too much wealth managed unwisely” Finance & Development, 40(1). Edwards, S. (1998) “Openness, Productivity and Growth: What do we really know?” The Economic Journal, 108(447), pp. 383-398 Fielding, D. & Gibson, F. (2012) “Aid and Dutch disease in Sub-Saharan Africa.” Journal of African Economies, 22(1), pp. 1-21. Gomanee, K., Morrissey, O., Mosley, P. & Verschoor, A. (2003) “Aid, pro-poor government spending and welfare.” University of Nottingham: CREDIT Research Paper, No.03/03. Grossman, G. M. & Helpman, E. (1991) “Trade, Knowledge spillovers and growth.” European economic review, 35(2), pp. 517-526. Hansen, H. & Tarp, F. (1999) “Aid effectiveness disputed.” Journal of international Development, pp. 375-398 Harvey, C. (1992) “Botswana: Is the Economic Miracle Over?” Journal of African Economies, 1(3), pp. 335–368. Hevia, C. & Nicolini, J. (2015) “Monetary Policy and Dutch disease: The case of price and wage rigidity.” Federal Reserve of Minneapolis Working Paper, No. 726 Lartey, E. K. K., Mandelman, F. S. & Acosta, P. A. (2008) “Remittances and the Dutch disease.” Journal of International Economics, 79(1), pp. 102-116 Minoiu, C. & Reddy, S. G. (2010) “Development aid and economic growth: a positive long-run relation.” The Quarterly Review of Economics and Finance, 50(1), pp. 27-39. Nkusu, M. (2004) “Aid and the Dutch disease in Low-Income Countries: Informed Diagnoses for Prudent Prognoses.” IMF Working Paper, Africa Department.
Nyoni, T. S. (1998) “Foreign Aid and Economic Performance in Tanzania.” World Development, 26(7), pp. 1235–1240. Oomes, N. & Kalcheva, K. (2007) “Diagnosing Dutch disease: does Russia have the symptoms?” IMF Working Paper No. 0J/102 Sackey, H. A. (2001) “External Aid Flows and the Real Exchange Rate in Ghana.” AERC Research Paper No. 110 (Nairobi: African Economic Research Consortium). Samuelson, P. (1964) “Theoretical notes on Trade problems.” The review of Economics and Statistics, 46(2), pp. 145-154. Stock, J. & Watson, M. (2011). Introduction to Econometrics (3rd edition). Pearson Education Limited. White, H. & Wignaraja, G. (1992) “Exchange Rates, Trade Liberalization and Aid: The Sri Lankan Experience.” World Development, 20(10), pp. 1471–1480. Younger, S. D. (1992) “Aid and the Dutch Disease: Macroeconomic Management When Everybody Loves You.” World Development, 20(11), pp. 1587–1597.
