Twin deficits and economic growth in Asian financial crisis and global financial crisis period : a cross-sectional analysis

(1)

Twin deficits and economic growth in Asian financial crisis and

global financial crisis period: a cross-sectional analysis

Abstract

This paper investigates the correlation between the twin deficits, namely current account and government budget deficits, and economic growth in 49 developing and developed countries using cross-sectional analysis for the period before and after Asian financial crisis and before and after global financial crisis. There is a change in the results on before and after the Asian financial crisis. Government budget deficit is associated negatively with economic growth, while current account deficit does not significantly correlate with economic growth. For the global financial crisis period, the current account deficit correlates positively with economic growth, while government budget deficit does not correlate with economic growth.

Bachelor Thesis July 2016

Faculty: Faculty of Economics and Business Student name: Theresia Sukma Larasati

Student number: 10828087 Specialization: Economics

(2)

Statement of Originality

This document is written by Theresia Sukma Larasati 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)

Table of content

1. Introduction 4

2. Current account, budget deficits, and economic growth in theory and practice 1. Argument against Ricardian equivalence: why budget deficit matters 6 2. Thirlwall’s Law: how current account correlates with economic growth 7

3. Empirical evidence 8

4. Correlation between Asian financial crisis, current account deficit, and

economic growth 10

3. Data and methodology

1. The model 12 2. The data 14 4. Results 15 5. Limitations 23 6. Conclusions 24 7. Appendix 1 25 8. Appendix 2 26 9. Appendix 3 29 10. References 30

(4)

1. Introduction

Many countries have experienced current account deficits for some periods of time. United States has been on constant current account deficit since the mid of 1980s. India experienced it from 1981 until 2001 and again from 2006 onwards, while other Asian countries such as Korea, Malaysia, Philippines, and Thailand also suffered from current account deficits on 1990s. Moreover, it is not uncommon for current account deficit to be accompanied by government budget deficit. It is called twin deficits. In fact, it happened in United States and India for a long period of time. Government budget deficit occurred in United States from 1975 to 1997 and it happened again from 2001 onwards. In India, its government budget deficit has occurred from 1990 until today.

When several developing Asian countries experienced twin deficits in 1990s, Asian financial crisis happened in 1997-1998. Currencies of several Asian countries dropped significantly. The developing countries that had twin deficits were hit particularly hard by this crisis. The economic growth of those countries declined rapidly during this period. Indonesia had a massive drop in GDP growth during the crisis from 4.7 percent to -13.1 percent. Philippines also experienced a decline in economic growth from 5.2 percent to -0.6 percent. The post-crisis economic growth showed improvement for those developing countries. However, the growth after the crisis was still lower than the growth before the crisis. On the other hand, this crisis had less impact on developed countries. It can be seen from the growth of United States that only decreased by 0.1 percent during the crisis.

Another crisis occurred from 2008 to 2009. It was the global financial crisis where the housing bubble in United States burst and affected not only the financial market but also the real economic sector around the world. This crisis mostly affected developed countries such as United States and European countries. Their economic growth dropped sharply during the crisis, with the growth of United States decreased from 1.8 percent to -2.8 percent. However, those developed countries were able to improve the condition shortly after the crisis. This crisis hit developed countries more than it hit developing countries.

(5)

The countries that got hit hard by the Asian crisis had twin deficits during that period. The reason was that their model of growth, based on the twin deficits, was deemed to be not sustainable by international investors. This condition raises an important question: whether the model of growth of these countries changed following the Asian financial crisis. The same question applies to the condition after the global financial crisis. But, it is addressed to developed countries instead of developing countries.

Martin and Fardmanesh (1990) find that there is a negative correlation between government budget deficit and economic growth. Moreover, on these crises periods, one would naturally think that twin deficits would correlate negatively with economic growth. However, the government budget deficit can actually promote economic growth to some extent. Increasing government expenditure causes government budget deficit, but it is a form of expansionary fiscal policy, and if it works, it could enhance the economic growth of the country applying the policy.

This paper studies the correlation of current account deficit and government budget deficit with economic growth in both developed and developing countries. It tries to answer the questions how do both deficits correlate with economic growth and how the model of growth changes from before to after the crises periods. The empirical research will be conducted on the relationship between the twin deficits and economic growth, using real GDP growth rate as a measurement for economic growth. The study will be divided into two periods, before and after the Asian crisis and before and after the global financial crisis. The cross-sectional analysis will be used to observe the relationship, by regressing both deficits on real GDP growth rate with control variables such as real interest rate, household consumption, and real exchange rate. The findings show that both deficits correlate differently with economic growth. Current account deficit correlates positively, while government budget deficits correlate negatively with economic growth.

Sections of this paper following the introduction are as follows: Section 2 summarizes the existing literature and empirical evidence of the relationship between both deficits and economic growth. Summary of the importance and the relevance of Asian crisis and global financial crisis with this paper will be explained. Section 3 describes the

(6)

data and the model used for this thesis. Section 4 shows and elaborates the results of the empirical analysis, followed by limitations in Section 5, and finally, the conclusion will be drawn in Section 6.

2. Current account, budget deficits, and economic growth in theory and practice

2.1 Argument against Ricardian equivalence: why budget deficit matters

As stated by Mishkin, Matthews, and Giuliodori (2013), if households do not consider bonds as net wealth, there are two implications that follow. First, the impact of changes in wealth as government’s debt increases will disappear. Second, the expansionary effect of government spending program that causes budget deficit will be eliminated and the budget deficit will be restored to balanced budget. These two implications are very well connected and lead to the theory of Ricardian equivalence: households do not view bonds as net wealth if they take into account that in the future the taxes will need to be raised to repay the debt back. This argument was proposed first by the nineteenth-century economist David Ricardo and brought back in 1970s by Robert Barro.

The result of Ricardian equivalence depends on the following three assumptions. First, there is no liquidity constraint on households, means that households are able to borrow against the expected future income at the current interest rate. Second, the interest rate and time horizon faced by households and government are the same. And third, households have children or heirs and will incorporate the utility of the heirs and take it into account to their consumption behavior.

This theory implies that budget deficit is irrelevant. However, the consensus view does not support this argument. The changes in fiscal policy are apparently only partly offset by changes in private sector savings, which means that there is a non-equivalence in the real world, and this leads to the Ricardian non-equivalence not being a good representation of macroeconomic behavior, according to Mishkin, Matthews, and Giuliodori (2013). Moreover, Martin and Fardmanesh (1990) also provide several reasons why budget deficits may matter, particularly in less developed countries, as

(7)

opposed to the Ricardian equivalence. First, the discount rate for taxpayers may be greater than the rate that government can borrow due to the finiteness of human life, which leads to imperfections in capital markets. For example, human wealth cannot be used for collateral and people may not be able to borrow against their future income. Second, in countries where most of the deficit is financed by money creation, or by selling bonds to the Central Bank, there is no creation of direct future tax liability. However, the “inflation tax” caused by the money creation could particularly be a distorting way of financing the deficit, especially if it is accompanied by the attempts to reduce its effects via price control and regulation of capital markets.

2.2 Thirlwall’s Law: how current account correlates with economic growth

As stated by Thirlwall (1979), the neo-classical approach on why growth rates differ between countries concentrates on the supply side of the economy. According to this approach, growth rate differences are explained in terms of differences in the growth of factor supplies and productivity, in which demand adjusts to supply. Thirlwall challenges this approach by arguing that it is actually demand that drives the economic system, in which supply adjusts to demand. The reason why demand grows at different rates between countries may be the inability of economic agents, particularly government, to expand demand. He focuses on the constraints on demand, which in an open economy is balance of payments. Therefore, according to him, growth of a country is demand determined and balance of payments constrained.

The economic performance of most countries is closely related to the performance of the trade sector and balance of payments. As a rough rule of thumb, the growth rate of most countries can be approximated by the rate of growth of exports divided by the income elasticity of demand for imports. Thirlwall (1980) discusses not about countries, but instead about regions inside a country. In his theory, regional growth is demand determined, unlike the neo-classical theory mentioned above because of the inability to constrain the growth by supply when factors of production are freely mobile. Especially for regions with high capital and labor mobility, growth should be demand-determined. If the demand for output from a region is strong, labor and capital will move to that region, giving advantage to the particular region and disadvantage to another. He states that a region with common currency will

(8)

automatically be “deflated” if the planned growth of imports exceeds the planned growth of exports, which will show up as lower than potential growth rate. This mechanism can be seen in Harrod trade multiplier, which is the level of income equals the level of exports divided by the marginal propensity to import.

Thirlwall’s law does not directly state that current account deficit negatively affects economic growth. However, it does state that equilibrium in balance of payments is important in increasing growth rate. Since the current account is one of the most important factors in determining balance of payments equilibrium, it also needs to be in equilibrium condition. If the current account is not in equilibrium, it will likely affect the economic growth. Moreover, if it is in deficit, the effect can be strongly negative.

2.3 Empirical evidence

The discussion of whether current account deficit and government budget deficit correlate with economic growth in an adverse way has gathered attention since decades ago. However, the existing studies on this issue are quite few and partial in focus. There is no single research studying the effect of both current account and government budget deficit on growth. Several studies on the effect of government budget on economic growth do exist. But too few discuss about the effect of current account deficit.

Current account deficit and government budget deficit are known to be related, and the research on this issue has increased during the last two decades. A paper by Rosensweig and Tallman (1993) studies the causality relationship between fiscal deficit and trade deficit in United States. They conclude that government deficit leads to trade deficit. Such relationship can be called as twin deficit. On the other hand, Anoruo and Ramchander (1998) find that trade deficit cause fiscal deficit in five developing countries of Asia, which are India, Indonesia, Korea, Malaysia, and the Philippines. However, this paper will not discuss the relationship between the two deficits. Instead, it will discuss the correlation between the twin deficits and economic growth. The following paragraph will discuss the contributions studying the effects of fiscal deficits and current account deficits on growth.

(9)

Martin and Fardmanesh (1990) conducted a research on the relation between fiscal variables and economic growth using panel data analysis for 76 developing and developed countries. They focused on the impact of the share of government activities in GDP on growth by considering taxes, government expenditures and budget deficits. Previous studies on this issue had failed to consider the budget deficit in addition to taxes and government expenditures, which potentially misled the results, as stated by their paper. Using the data from period 1972-1981, they divided the countries into three groups, which are low-, middle-, and high-income groups. The aggregate results state that there is a strong negative association between budget deficits and growth, and the strongest results come from middle-income countries.

Cebula (1995) based his research on Martin and Fardmanesh (1990). He examined the impact of government budget deficit on economic growth in the United States over the 1955-1992 time period. Using the simple growth model, he provided an Instrumental Variables (IV) estimate that indicated the impact of budget deficit on economic growth. His result is that the growth rate of United States is a decreasing function of the budget deficit. This conclusion also supports the results of Martin and Fardmanesh (1990) research.

On the current account side, Atesoglu (1993) tested the Thirlwall’s law of economic growth in the United States by calculating an average balance-of-payments-constrained growth rate for a particular period using the Harrod trade multiplier and comparing these with the corresponding actual growth rates. The period studied is 1955-1970 and 1975-1990. The results reveal that the dynamic Harrod trade multiplier can provide a satisfactory explanation of variation in the long-term economic growth of the United States. This implies that Thirlwall’s law holds in this research, and current account equilibrium is important in increasing the economic growth.

Moreover, Loría and Fujii (1997), analyzed the balance of payment constraint to economic growth in Mexico over the 1950-1996 time period. They find that the worsening trade deficit in Mexico in 1996 represents the most important constraint to Mexico’s economic growth. To sustain the recovery of Mexico’s economy, not only the expansion of exports is required, but also the prevention of an excessive demand

(10)

for imports. This implies that a country should reduce its trade deficit in order to increase the economic growth. Therefore, this finding is consistent with the Thirlwall theories and the research by Atesoglu (1993).

Based on the existing studies, it can be seen that government budget deficit and current account deficit have a negative correlation with economic growth. However, these findings still cannot predict the correlation between both deficits and economic growth.

2.4 Correlation between Asian financial crisis, current account deficit, and economic growth

The crisis began in July 1997 when the value of Thai baht was allowed to float, thus decreasing its value by about 20%, as stated by Chowdhry and Goyal (2000). It then spread rapidly to Philippine peso and Malaysian ringgit. Not long after, Indonesian rupiah also fell by three-quarters of its value in August, which then followed by South Korean won that substantially depreciated in November (Barro, 2001).

Current account imbalances were said to have a potential role as a source of tensions in the financial markets, as stated by Corsetti et al. (1999). The US Deputy Treasury Secretary wrote in The Economist that “close attention should be paid to any current account deficit in excess of 5 percent of GDP, particularly if it is financed in a way that could lead to rapid reversals”. Several countries in the study of Corsetti et al. (1999) gave evidence on this statement.

As can be seen in Table 1 below, countries that had significant decrease in the value of currency also had significant current account deficits in 1990s. Thailand and Malaysia were the largest deficit countries in the sample and had experienced deficits for over a decade. The current account deficit in Thailand was over than 5 percent of GDP until 1996 and almost reached 9 percent of GDP in 1995 and 1996. Malaysia had smaller deficits but also almost reached 9 percent of GDP in 1991 and 1995. The Philippines, Indonesia, and South Korea also had continuous current account deficits from 1991 until 1997, with the deficit in some years in the Philippines exceeded 5

(11)

percent of GDP, while the deficits in Indonesia and South Korea were not as big as the other countries mentioned above.

Table 1

The table above is taken from Corsetti et al. (1999)

Note: The source of all the data in these tables is the international financial statistics of the International Monetary Fund (unless otherwise noted). The data for Taiwan are from various sources (Economist Intelligence unit reports, IMF's December 1997 World Economic Outlook and Asian Development Bank). The data for Singapore for 1997 are from the Economist Intelligence Unit country report, 2nd quarter 1998.

Large trade deficits were expected to be the cause of the large current account deficits. In Thailand, the trade deficits were higher than 4 percent of GDP in the 1990s. It even reached more than 7 percent of GDP in 1990 and 1995. It was more fluctuated in the Philippines, where the lowest trade deficit was 3 percent of GDP in 1991 and the highest reached 9 percent of GDP in 1994. Therefore, it can be concluded that the countries that severely affected by the crisis in 1997-1998 had large trade deficits and large current account deficits throughout the 1990s.

However, a large current account deficit can be perceived as sustainable if current and expected economic growth are high. For a given current account deficit to GDP ratio, higher growth rate implies a slower dynamics of the foreign debt to GDP ratio and increases the ability of the country to pay its external debt. GDP growth may reflect sustained capital accumulation rates driven by expectations of high profitability. Higher growth rate may also explain a transitory decline in saving rate, in anticipation of higher future income. Therefore current account imbalances that are caused by a transitory fall in private savings should not be a concern because future income growth will increase future savings.

(12)

3. Data and methodology

3.1 The Model

This paper will use cross-sectional regression analysis to study the correlation between both current account balance and government budget balance with economic growth. Cross-sectional analysis is a type of observational study that involves the analysis of data collected from a population at one specific point of time, which is cross-sectional data. It is a type of data collected by observing many subjects (such as individuals, countries, firms, regions) at the same point of time.

This paper aims to study the correlation between the twin deficits, namely current account deficit and government budget deficit, and economic growth in 49 developing and developed countries. The technique that will be used is the OLS regression model. According to Stock & Watson (2013), the OLS multiple regression with n regressors has the following form:

𝑌_! = 𝛽_!+ 𝛽_!𝑥_!!+ 𝛽_!𝑥_!!+ ⋯ + 𝛽_!𝑥_!"+ 𝑢_! 𝑖 = 1,2, … , 𝑛

This paper will analyze the following hypothesis: “Current account balance and fiscal balance are positively correlated with economic growth”. The null hypothesis that will be tested is as followed: H0: βCB = 0, βFB = 0, which refers to current account balance and fiscal balance having no effect on economic growth. The null hypothesis will be tested against its alternative, H1: βCB > 0, βFB > 0.

There are five independent variables that will be used in this study. Three are control variables, which are interest rate, household consumption, and exchange rate. Two are the main variables that will be analyzed, current account balance and government budget balance. The model is as follows:

∆𝐺𝐷𝑃! = 𝛼 + 𝛽!" ∙ ∆𝐶𝐵!+ 𝛽!"∙ ∆𝐹𝐵! + 𝛽!∙ ∆𝑟!+ 𝛽!"∙ ∆𝐻𝐶! + 𝛽!∙ ∆𝑋!+ 𝜀!

where GDP is the dependent variable, which is the real economic growth rate per capita in the countries analyzed. FB is the fiscal or government budget balance and

(13)

CB is the current account balance. Both are as ratio of GDP. The control variables chosen are r, HC, and X, which real interest rate, household consumption expenditure, and exchange rate respectively. The error term is indicated by ε, while i refer to the country observed. The term ∆ means percentage change in the variable from one year to the previous year.

The 49 countries will be divided into two groups, which are developing and developed countries groups. Developing countries group consists of 26 countries and the developed countries group consists of 23 countries. The division of the countries follows a research by Martin and Fardmanesh (1990), which divides the countries into three groups. According to their paper, the middle-income group is negatively affected by the government budget deficit and the high-income group is not affected by it. Following their conclusion, it can be expected that the developing countries group will be similarly affected by the government budget deficit.

However, the number of observation for each period is different because there are some countries that appear to be outliers. The outliers were removed by eliminating the GDP growth, current account balance, and government budget balance that have an absolute value larger than two times the standard deviation.

Several regressions will be run for four different years. There will be two different time frames, which are before and after Asian financial crisis in 1997, and before and after global financial crisis in 2008. The years observed are 1996, 2001, 2006, and 2010. Year 1996 represents the period before Asian crisis, year 2001 represents the period after Asian crisis, year 2006 represents the period before global financial crisis and year 2010 represents the period after global financial crisis. For each year, there will be two regressions run from both groups. In addition to the separate groups regressions, the regression on the aggregate of the countries will also be run in order to compare to the group regressions.

The purpose of doing before and after regressions on each period is to see whether the model of growth changes after the crises occurred. It is expected from the results that the economic growth of developing countries group would be more affected in Asian crisis period since some developing Asian countries mentioned in the previous section

(14)

were severely hit by this crisis. Similarly, the growth for developed countries would be more affected in global financial crisis period.

3.2 The Data

The dependent variable is economic growth, which is measured by real GDP growth rate per capita. The independent variables are government budget balance and current account balance. The data on GDP growth is a percentage growth rate of GDP per capita, while government budget balance and current account balance are in percentage of GDP. The control variables are real interest rate, household consumption, and exchange rate. The data on real GDP growth rate, government budget balance, current account balance, real interest rate, household consumption, and exchange rate is obtained from the databank in World Development Indicators section of World Bank, while the data for real effective exchange rate is obtained from Bruegel. This research uses percentage changes in the variables, therefore the data used are yearly data from the following years: 1995-1996, 2000-2001, 2005-2006, and 2009-2010. The year 1995-1996 and 2000-2001 data are before and after the Asian financial crisis in 1997, while data from year 2005-2006 and 2009-2010 are before and after the global financial crisis in 2008.

Dependent variable: real GDP growth rate per capita

GDP growth is the percentage growth rate of GDP, measured as the change of GDP from one year to the previous year divided by the GDP in the previous year. GDP per capita is gross domestic product divided by midyear population. GDP is the sum of gross value added by all resident producers in the economy plus any product taxes on the products minus the subsidies that are not included in the value of the products. This research uses real GDP growth rate, which is the nominal GDP adjusted for price.

(15)

Government budget balance is revenue including grants minus expense, minus net acquisition of nonfinancial assets. Current account balance is the sum of net exports of goods and services, net primary income, and net secondary income. Both of the deficits are in percentage of GDP. A positive number in the balances indicates a surplus and a negative number indicates a deficit.

Control variables: real interest rate, household consumption, and exchange rate

Real interest rate is the lending interest rate adjusted for inflation as measured by the GDP deflator. The lending rate used is the bank rate with short- and medium-term maturity. Instead of directly affecting the GDP growth, the interest rate affects investment in the sense that if the interest rate increases, the cost of borrowing become more expensive and therefore the investment rate decreases, and a decrease in investment could lead to a decrease of GDP.

Household final consumption expenditure is the market value of all goods and services, including durable products purchased by households. It also includes the expenditures of nonprofit institutions serving households. It is directly related to GDP. When household consumption increases, it implies that the GDP will also increase. Household consumption expenditure is expressed by percentage of GDP.

The real effective exchange rate is the nominal effective exchange rate divided by a price deflator. Nominal effective exchange rate is a measure of the value of a currency against a weighted average of several foreign currencies. This variable can change the volume of exports and imports of a country. When a country’s exchange rate increases, its currency will have lower value relative to other currency. This leads to increasing exports due to cheaper currency but also increases the cost of imports since other currencies become more expensive.

4. Results

This paper uses cross-sectional analysis and OLS regressions model to study the correlation between both deficits and economic growth. Several regressions were

(16)

performed; the first one that only used both deficits as regressors, the second until fourth which used both deficits and one control variable, and the last one that used all variables. These equations were repeated for the two different groups and two different periods.

As previously mentioned, the regressions carried out in this research studies the two periods of time, which are before and after the Asian crisis and before and after the global financial crisis. The presentation of the results is as follows: There are six tables; the first three present the results from the period of Asian crisis and the last three tables present the period of global financial crisis. Table 2 and Table 5 present the results for all sample countries. In Table 3 and Table 6, the results for developing countries are presented, and the results for developed countries are presented in Table 4 and Table 7. Column (1) until column (5) report the period before crisis, while column (6) until column (10) focus on the period after the crisis occurred.

The table below presents the coefficients with the t statistics value in parentheses below each coefficient for variables CB, FB, r, HC, X, and the constant term. The number of observation, N, the R-squared, and the significance of the model Prob>F are also presented.

Table 2 presents the results of the research for all sample countries on the Asian crisis period. It can be seen in the period before crisis that there is no sign of significance, both when the regressions performed with or without the control variables. It means that current account balance and government budget balance do not have any correlation with economic growth on this period. This is also the case with the three control variables.

On the other hand, the results from the period after crisis indicate that government budget balance has a strong positive correlation with economic growth, with its p-value lower than 1 percent. The coefficient of 1.098% on government budget balance in column (6) can be interpreted as: an increase of one percentage point in the government budget balance is associated with a 1.098 percentage point of higher growth.

(17)

Table 2 All sample countries

Before Asian crisis (1996) After Asian crisis (2001)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB 0.064 0.078 -0.148 0.057 -0.126 0.082 0.102 0.032 0.081 0.062 (0.36) (0.4) (-0.67) (0.31) (-0.51) (0.58) (0.65) (0.22) (0.56) (0.39) FB -0.215 -0.214 -0.206 -0.247 -0.252 1.098*** 1.108*** 1.162*** 1.099*** 1.179*** (-1.2) (-1.17) (-1.17) (-1.28) (-1.32) (3.52) (3.49) (3.73) (3.46) (3.65) R 0.013 0.006 0.016 0.025 (0.21) (0.1) (0.30) (0.46) HC -0.400 -0.349 -0.380 -0.399 (-1.6) (-1.29) (-1.38) (-1.39) X -2.818 -0.051 -0.269 0.404 (-0.5) (-0.71) (-0.05) (0.08) Constant -0.452 -0.480 -0.279 -0.357 -0.171 -0.772** -0.794** -0.625* -0.766** -0.661* (-1.22) (-1.21) (-0.74) (-0.85) (-0.39) (-2.25) (-2.24) (-1.76) (-2.10) (-1.74) N 37 37 37 37 37 38 38 38 38 38 R-squared 0.041 0.042 0.110 0.048 0.124 0.276 0.278 0.315 0.276 0.319 Prob > F 0.491 0.695 0.274 0.648 0.508 0.0035 0.011 0.0046 0.011 0.025 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

This means that higher budget deficit is associated with lower economic growth. This finding supports the research by Martin and Fardmanesh (1990), which found a strong negative association between government budget deficits and economic growth. The null hypothesis (βFB = 0) can therefore be rejected. However, the current account balance does not seem to be correlated with economic growth in after crisis period. The number of observation in the after crisis regressions is one amount higher than before crisis regressions because the number of removed outlier countries in before crisis period is more than in after crisis period.

Table 3 and 4 below present the results for developing and developed countries. It can be concluded that on the period before crisis, the current account deficits and budget deficits of both developing and developed countries did not have any correlation with economic growth because none of the variables regressed are significant.

(18)

Table 3 Developing countries

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB 0.064 0.072 -0.190 0.046 -0.129 0.111 0.226 0.060 0.101 0.168 (0.24) (0.21) (-0.56) (0.18) (-0.28) (0.53) (0.91) (0.27) (0.46) (0.60) FB -0.384 -0.387 -0.281 -0.522 -0.430 1.053* 1.095* 1.059* 1.059* 1.106 (-1.16) (-1.10) (-0.82) (-1.41) (-1.00) (1.81) (1.86) (1.79) (1.76) (1.77) R 0.004 0.004 0.072 0.071 (0.04) (0.03) (0.87) (0.80) HC -0.419 -0.297 -0.275 -0.242 (-1.17) (-0.71) (-0.68) (-0.56) X -8.049 -0.096 -1.451 -1.662 (-0.88) (-0.77) (-0.19) (-0.21) Constant -0.544 -0.559 -0.213 -0.211 0.016 -1.146* -1.343* -1.002 -1.132* -1.196 (-0.92) (-0.77) (-0.33) (-0.30) (0.02) (-1.91) (-2.08) (-1.55) (-1.82) (-1.65) N 21 21 21 21 21 19 19 19 19 19 R-squared 0.075 0.075 0.144 0.115 0.178 0.177 0.216 0.202 0.179 0.240 Prob > F 0.496 0.714 0.439 0.545 0.668 0.211 0.287 0.323 0.384 0.557 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

On the period after crisis, column (6) in Table 3 shows that an increase of one percentage point in government budget balance is associated to 1.053 percentage point of higher growth. The government budget deficit is negatively correlated with economic growth in developing countries. This finding indeed supports the research by Martin and Fardmanesh (1990). Moreover, the regressions on the developed countries show even stronger negative correlation between government budget deficits and economic growth. It can be seen from column (6) in Table 4 that an increase of one percentage point in government budget balance is correlated to 1.140 percentage point higher growth. Even though the crisis had more impact on developing countries, the results show that developed countries group has a stronger and more similar result to the all sample countries result with 1% significance level. As stated before, the difference in number of observation is because of the removed outliers.

(19)

Table 4 Developed countries

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB 0.268 0.251 0.178 0.268 0.146 -0.108 -0.109 -0.095 -0.117 -0.120 (1.08) (0.94) (0.50) (1.04) (0.36) (-0.53) (-0.52) (-0.46) (-0.54) (-0.53) FB -0.031 -0.025 -0.036 -0.033 -0.033 1.140*** 1.143*** 1.257*** 1.132*** 1.296*** (-0.17) (-0.13) (-0.19) (-0.17) (-0.15) (3.73) (3.57) (3.73) (3.56) (3.45) r 0.013 -0.006 0.0056 0.035 (0.21) (-0.07) (0.06) (0.34) HC -0.170 -0.231 -0.424 -0.556 (-0.36) (-0.34) (-0.86) (-0.96) X 0.419 -0.040 1.522 3.569 (0.06) (-0.47) (0.20) (0.42) Constant -0.713 -0.728 -0.691 -0.711 -0.676 -0.332 -0.329 -0.252 -0.381 -0.323 (-1.60) (-1.55) (-1.48) (-1.53) (-1.28) (-0.93) (-0.88) (-0.67) (-0.85) (-0.68) N 16 16 16 16 16 19 19 19 19 19 R-squared 0.086 0.089 0.096 0.086 0.115 0.520 0.520 0.542 0.521 0.553 Prob > F 0.558 0.761 0.740 0.771 0.925 0.003 0.010 0.007 0.010 0.042 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

Tables 5, 6, and 7 below present the regression results for global financial crisis period. Table 5 shows the results of the regressions for all sample countries. On the before crisis section, there are no regressors that appear to be significant.

On the period after the crisis occurred, when the regression was conducted with the restricted regressors of only current account balance and government budget balance, the result was not significant. However, when the three control variables were added to the equation, there was a significant result for the current account balance. Column (10) shows that an increase of one percentage point in current account balance is associated with a 0.416 percentage point decrease in economic growth. This means that current account deficit is positively correlated with economic growth. This finding is different from the expectation regarding the correlation between current account deficit and economic growth. It is expected that the deficit would have a negative correlation with growth, but the results show the opposite.

(20)

Before global financial crisis (2006) After global financial crisis (2010)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB 0.033 0.023 -0.009 0.030 -0.064 -0.287 -0.257 -0.375** -0.332* -0.416** (0.37) (0.28) (-0.08) (0.34) (-0.60) (-1.58) (-1.42) (-2.06) (-1.86) (-2.41) FB -0.172 -0.151 -0.170 -0.171 -0.139 0.268 0.422 0.146 0.302 0.321 (-0.91) (-0.82) (-0.89) (-0.90) (-0.76) (1.14) (1.67) (0.62) (1.32) (1.36) r 0.102* 0.137** 0.085 0.093* (1.93) (2.14) (1.54) (1.83) HC -0.088 -0.183 -0.673* -0.848** (-0.65) (-1.34) (-1.95) (-2.60) X 2.767 -2.016 -18.260* -21.409** (0.68) (-0.44) (-1.86) (-2.32) Constant 0.987*** 1.012*** 0.895*** 0.934*** 0.869*** 5.878*** 6.006*** 5.755*** 6.231*** 6.277*** (4.36) 4.63 (3.34) (3.88) (3.22) (8.81) (9.06) (8.84) (9.20) (9.92) N 40 40 40 40 40 49 49 49 49 49 R-squared 0.024 0.115 0.035 0.036 0.161 0.072 0.119 0.145 0.139 0.292 Prob > F 0.638 0.215 0.727 0.718 0.283 0.178 0.124 0.068 0.079 0.009 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

The regression results for the developing countries are presented on Table 6 below. There is also no correlation between both balances on economic growth before the crisis. However, there is a strong positive correlation between government budget balance and economic growth after the crisis. Real interest rate is the most relevant control variable in the equations, which is shown in the column (7) and column (10). Regression (7) can be interpreted as: an increase of one percentage point in the government budget balance is correlated with a 1.085 percentage point increase in economic growth. In column (10), current account balance appears to have a negative correlation with economic growth with 10% significant level. It means that government budget deficit correlates negatively with economic growth, while current account deficit correlates positively with growth.

(21)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB 0.001 -0.012 -0.065 -0.005 -0.135 -0.300 -0.191 -0.393 -0.415 -0.442* (0.01) (-0.12) (-0.47) (-0.05) (-1.03) (-1.15) (-0.82) (-1.40) (-1.52) (-1.81) FB -0.298 -0.245 -0.299 -0.305 -0.216 0.608* 1.085*** 0.553 0.718* 1.172*** (-1.15) (-1.00) (-1.14) (-1.15) (-0.88) (1.78) (3.17) (1.59) (2.07) (3.53) r 0.113* 0.162** 0.169*** 0.179*** (1.89) (2.18) (2.85) (3.21) HC -0.127 -0.236 -0.437 -0.569 (-0.75) (-1.46) (-0.92) (-1.45) X 3.103 -3.253 -18.723 -22.422* (0.61) (-0.5) (-1.30) (-1.86) Constant 0.999*** 1.022*** 0.859** 0.902** 0.873** 5.078*** 5.476*** 4.771*** 5.786*** 5.946*** (3.10) (3.37) (2.29) (2.47) (2.25) (5.42) (6.59) (4.78) (5.39) (6.38) N 22 22 22 22 22 26 26 26 26 26 R-squared 0.066 0.220 0.094 0.084 0.321 0.131 0.366 0.163 0.193 0.494 Prob > F 0.525 0.204 0.608 0.652 0.242 0.199 0.017 0.262 0.187 0.012 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

For developed countries, the results after the crisis are shown in Table 7 below. Column (6) shows that the current account balance has a strong negative correlation with economic growth. Moreover, when control variable final household expenditure consumption was added to the equation, the results changed significantly. Column (8) shows that current account balance and household expenditures correlate strongly and negatively with economic growth. Government budget balance also appears to be negatively correlated with economic growth. Therefore, the null hypothesis (βFB = 0) can be rejected with 10% significance level. This finding supports the research of Martin and Fardmanesh (1990) who found a negative correlation between government budget deficits and economic growth. However, the results for current account deficit are still quite odd and will need a further research.

(22)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB 0.250 0.234 0.235 0.404 0.383 -0.614** -0.588* -0.738*** -0.519* -0.604** (1.23) (1.15) (1.11) (1.76) (1.59) (-2.14) (-1.76) (-3.06) (-1.78) (-2.22) FB 0.068 0.167 0.069 0.007 0.116 -0.340 -0.334 -.650* -0.277 -0.580* (0.23) (0.54) (0.23) (0.02) (0.35) (-0.95) (-0.90) (-2.06) (-0.77) (-1.86) r -0.257 -0.272 -0.031 -0.044 (-1.08) (-0.86) (-0.16) (-0.29) HC -0.148 0.091 -1.549*** -1.582*** (-0.42) (0.20) (-3.15) (-3.27) X 19.477 17.905 -24.881 -26.162 (1.33) (1.17) (-1.27) (-1.61) Constant 0.912** 0.738* 0.786 0.951** 0.841* 7.411*** 7.322*** 7.882*** 7.083*** 7.419*** (2.67) (1.96) (1.71) (2.84) (1.84) (7.30) (6.23) (9.19) (6.85) (7.73) N 18 18 18 18 18 23 23 23 23 23 R-squared 0.103 0.172 0.114 0.204 0.261 0.188 0.189 0.466 0.251 0.541 Prob > F 0.443 0.436 0.625 0.348 0.541 0.125 0.253 0.007 0.131 0.014 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

This research finds that there is a change in the significance of some coefficients before and after the Asian financial crisis, particularly the coefficient of government balance. This might be because some of the countries affected by the crisis did not view the macroeconomic conditions correctly, which could lead to incorrect and ineffective policy actions. Therefore, any policy that those countries took would not be effective and would have no correlation with economic growth.

After the crisis happened, the recessions were much more severe than expected. According to Boorman et al. (2000), several Asian countries such as Indonesia, Philippines, and Thailand already had a set of fiscal policy measures before the crisis, which is a combination of tax rates, expenditure programs, and so on. However, they overstated the degree of fiscal adjustment associated with those set of policies, which resulted in larger fiscal deficits than originally programmed. This could be associated with the slow growth of those countries.

(23)

Some of the results of this research do not satisfy the base theories of this paper and do not meet the expectations. Even though the results on government budget deficit after the crisis period satisfy the base theory of this paper, which is a paper by Martin and Fardmanesh (1990), it still cannot be concluded whether the twin deficits correlate with economic growth. It is because the correlation of each deficit with economic growth is different. For instance, the results on after Asian crisis period show that while government budget deficit correlates negatively with economic growth, the current account deficit does not have any correlation at all due to the insignificant regression results. Moreover, the results of current account deficit are the opposite of what is expected and those results do not satisfy the base theories of this paper.

There are several possible reasons causing the unexpected results. This paper uses data from very different type of countries with very different growth level and policies, which leads to very different budget balance and interest rates. There are 49 observed countries in total, with 26 developing countries and 23 developed countries. However, as previously mentioned, outliers exist in each dataset for each regression. Therefore, those outliers need to be removed for better results. Removing outliers is also the reason of the difference in number of observation. This action led to a slightly better result, but since the sample countries became smaller, some of the results still do not meet the expectations. For a comparison, the regression results for all dataset including the outliers are presented in Appendix.

5. Limitations

There are several important limitations to this study, mainly in regards to limited data availability. Since this study uses cross-sectional analysis, the number of countries observed should be as much as can be found. However, the number of countries that provide data for all variables needed within the period observed is very limited. Some countries that are affected by Asian crisis such as Malaysia and Philippines also cannot be included because of the limited data availability on government budget deficits. Therefore, the number of sample countries is very small. It gets even smaller due to removing the outliers.

(24)

6. Conclusion

This paper studies the correlation between the twin deficits, which are current account deficits and government budget deficits, and economic growth in 76 countries that are divided into two groups and two periods. The two groups are developing and developed countries, and the two periods are Asian crisis and global financial crisis. Using annual data on real GDP growth rate as a measurement of economic growth, current account deficits, and government budget deficits, an OLS regression analysis was conducted while using several variables as control variables. Those are the real interest rate, final household expenditure consumption, and real effective exchange rate.

The results generated from all sample countries in Asian crisis period show that government budget deficit is negatively correlated with economic growth after the crisis occurred. Results from developed countries are stronger than from developing countries, which both generate the same results as the all sample countries results. There seems to be no correlation between current account deficit and economic growth in this period.

On the global financial crisis period, the regression results for all three groups show that there is a positive correlation between current account deficit and economic growth after the crisis. However, for both the developing and developed countries groups, government budget deficit is negatively correlated with economic growth. This finding confirms the research by Martin and Fardmanesh (1990) who found a strong negative association between government budget deficits and economic growth.

(25)

7. Appendix 1

In this paper, 49 countries are observed and are divided into two groups; developing countries and developed countries groups.

The following 26 countries are included in developing countries group: Algeria, Belarus, Belize, Chile, Costa Rica, Dominican Republic, Estonia, Guatemala, India, Indonesia, Israel, Kenya, Korea, Malta, Mauritius, Mongolia, Nicaragua, Oman, Peru, Russia, Seychelles, Singapore, South Africa, Sri Lanka, Thailand, Uruguay.

The following 23 countries are included in developing countries group: Bulgaria, Croatia, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Norway, Poland, Portugal, Romania, Sweden, Switzerland, United Kingdom, United States.

There are some outlier countries that needed to be omitted in order to get better results. The following countries are omitted in the before Asian financial crisis period: Algeria, Belarus, Bulgaria, Czech Republic, Iceland, Kenya, Latvia, Lithuania, Malta, Netherlands, Portugal, Seychelles.

The following countries are omitted in the after Asian financial crisis period: Algeria, Belize, Germany, Iceland, Israel, Luxembourg, Malta, Seychelles, Singapore, Sri Lanka, United States.

The following countries are omitted in the before global financial crisis period: Algeria, Belize, Bulgaria, Chile, Greece, Iceland, Kenya, Latvia, Romania.

The following countries are omitted in the after global financial crisis period: Algeria, Estonia, Mongolia, Oman, Russia, Singapore.

(26)

8. Appendix 2

Results for regressions including outliers in Asian financial crisis period

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB -0.179 -0.203 -0.186 -0.135 -0.176 0.011 0.090 -0.033 0.014 0.059 (-1.15) (-1.36) (-1.19) (-0.90) (-1.23) (0.09) (0.67) (-0.26) (0.11) (0.45) FB 0.002 0.137 -0.033 -0.031 0.060 0.541** 0.675** 0.415 0.539* 0.539* (0.02) (0.87) (-0.21) (-0.21) (0.39) (2.04) (2.48) (1.54) (2.01) (1.98) R 0.065** 0.068** 0.113 0.139** (2.27) (2.46) (1.67) (2.06) HC -0.180 -0.271 -0.516* -0.666** (-0.83) (-1.34) (-1.75) (-2.21) X 11.915** 10.819** 0.987 2.747 (2.35) (2.24) (0.13) (0.38) Constant 0.078 -0.244 0.228 -0.252 -0.335 -1.716*** -1.776*** -1.620*** -1.734*** -1.716*** (0.17) (-0.54) (0.47) (-0.56) (-0.71) (-3.57) (-3.75) (-3.42) (-3.43) (-3.61) N 49 49 49 49 49 49 49 49 49 49 R-squared 0.034 0.134 0.049 0.140 0.257 0.101 0.154 0.158 0.102 0.240 Prob > F 0.448 0.088 0.518 0.076 0.022 0.086 0.056 0.050 0.182 0.032 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

(27)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB -0.177 0.096 -0.224 -0.149 0.089 -0.069 0.082 -0.102 -0.072 0.042 (-0.67) (0.41) (-0.82) (-0.58) (0.34) (-0.41) (0.49) (-0.59) (-0.41) (0.24) FB 0.003 -0.198 0.035 -0.010 -0.195 0.425 0.738* 0.238 0.425 0.565 (0.01) (-0.72) (0.10) (-0.03) (-0.66) (1.13) (2.00) (0.56) (1.11) (1.34) r 0.153*** 0.159** 0.203** 0.212** (3.41) (2.64) (2.31) (2.32) HC -0.246 -0.068 -0.482 -0.479 (-0.73) (-0.22) (-0.95) (-1.00) X 12.224 -2.302 -0.803 -4.002 (0.172) (-0.23) (-0.07) (-0.37) Constant 0.302 -0.593 0.561 -0.267 -0.450 -2.555*** -2.892*** -2.332** -2.547*** -2.643*** (0.38) (-0.84) (0.64) (-0.31) (-0.54) (-3.25) (-3.92) (-2.83) (-3.13) (-3.38) N 26 26 26 26 26 26 26 26 26 26 R-squared 0.036 0.369 0.059 0.116 0.374 0.099 0.275 0.135 0.100 0.319 Prob > F 0.656 0.016 0.715 0.427 0.075 0.300 0.065 0.354 0.501 0.145 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

(28)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB -0.260 -0.2461699 -0.275 -0.115 -0.126 0.311 0.290 0.231 0.309 0.207 (-1.02) (-0.87) (-1.02) (-0.46) (-0.39) (1.37) (1.27) (0.94) (1.32) (0.81) FB 0.051 0.026 0.017 -0.046 0.089 0.810** 0.813** 0.868** 0.814** 0.852** (0.33) (0.10) (0.08) (-0.30) (0.36) (2.47) (2.48) (2.58) (2.40) (2.46) r -0.006 0.024 -0.119 -0.097 (-0.13) (0.44) (-1.04) (-0.77) HC -0.076 0.106 -0.358 -0.449 (-0.21) (0.24) (-0.91) (-0.83) X 11.847* 14.063* -0.838 6.144 (1.97) (2.01) (-0.11) (0.60) Constant -0.284 -0.225 -0.209 -0.203 -0.547 -0.779 -0.801 -0.728 -0.756 -0.906 (-0.53) (-0.31) (-0.32) (-0.41) (-0.77) (-1.66) (-1.71) (-1.53) (-1.42) (-1.65) N 23 23 23 23 23 23 23 23 23 23 R-squared 0.052 0.053 0.054 0.212 0.236 0.236 0.277 0.268 0.236 0.306 Prob > F 0.587 0.788 0.781 0.200 0.422 0.068 0.097 0.108 0.154 0.242 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

(29)

9. Appendix 3

Results for including outliers in global financial crisis period

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB -0.011 -0.013 -0.050 -0.013 -0.060 -0.287 -0.257 -0.375** -0.332* -0.416** (-0.16) (-0.19) (-0.59) (-0.19) (-0.70) (-1.58) (-1.42) (-2.06) (-1.86) (-2.41) FB -0.284 -0.317 -0.311 -0.268 -0.329 0.268 0.422 0.146 0.302 0.321 (-1.51) (-1.66) (-1.62) (-1.42) (-1.67) (1.14) (1.67) (0.62) (1.32) (1.36) r 0.058 0.046 0.085 0.093* (1.00) (0.76) (1.54) (1.83) HC -0.112 -0.131 -0.673* -0.848** (-0.79) (-0.91) (-1.95) (-2.60) X 4.591 3.978 -18.260* -21.409** (0.96) (0.78) (-1.86) (-2.32) Constant 0.936*** 0.994*** 0.824*** 0.837*** 0.766** 5.878*** 6.006*** 5.755*** 6.231*** 6.277*** (3.81) (3.94) (2.90) (3.14) 2.37 (8.81) (9.06) (8.84) (9.20) (9.92) N 49 49 49 49 49 49 49 49 49 49 R-squared 0.048 0.069 0.061 0.067 0.097 0.072 0.119 0.145 0.139 0.292 Prob > F 0.320 0.353 0.410 0.366 0.477 0.178 0.124 0.068 0.079 0.009 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

(30)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB -0.038 -0.0433 -0.130 -0.039 -0.142 -0.300 -0.191 -0.393 -0.415 -0.442* (-0.41) (-0.47) (-1.02) (-0.41) (-1.10) (-1.15) (-0.82) (-1.40) (-1.52) (-1.81) FB -0.224 -0.266 -0.300 -0.225 -0.344 0.608* 1.085*** 0.553 0.718* 1.172*** (-0.96) (-1.12) (-1.23) (-0.95) (-1.37) (1.78) (3.17) (1.59) (2.07) (3.53) r 0.067 0.059 0.169*** 0.179*** (1.07) (0.88) (2.85) (3.21) HC -0.193 -0.209 -0.437 -0.569 (-1.06) (-1.12) (-0.92) (-1.45) X 3.599 2.654 -18.723 -22.422* (0.61) (0.42) (-1.30) (-1.86) Constant 0.724** 0.801** 0.576 0.608 0.546 5.078*** 5.476*** 4.771*** 5.786*** 5.946*** (2.08) (2.26) (1.54) (1.52) (1.19) (5.42) (6.59) (4.78) (5.39) (6.38) N 26 26 26 26 26 26 26 26 26 26 R-squared 0.047 0.094 0.093 0.062 0.150 0.131 0.366 0.163 0.193 0.494 Prob > F 0.577 0.529 0.533 0.694 0.624 0.199 0.017 0.262 0.187 0.012 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

(31)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) GDP GDP GDP GDP GDP GDP GDP GDP GDP GDP CB 0.130 0.135 0.135 0.166 0.180 -0.614** -0.588* -0.738*** -0.519* -0.604** (1.09) (1.11) (1.04) (1.45) (1.42) (-2.14) (-1.76) (-3.06) (-1.78) (-2.22) FB -0.602* -0.516 -0.607 -0.454 -0.344 -0.340 -0.334 -.650* -0.277 -0.580* (-1.73) (-1.40) (-1.69) (-1.34) (-0.93) (-0.95) (-0.90) (-2.06) (-0.77) (-1.86) r -0.201 -0.264 -0.031 -0.044 (-0.75) (-0.92) (-0.16) (-0.29) HC 0.034 0.056 -1.549*** -1.582*** (0.10) (0.16) (-3.15) (-3.27) X 19.344* 19.939* -24.881 -26.162 (1.84) (1.80) (-1.27) (-1.61) Constant 1.552*** 1.391*** 1.587*** 1.397*** 1.239** 7.411*** 7.322*** 7.882*** 7.083*** 7.419*** (3.69) (2.92) (2.88) (3.44) (2.25) (7.30) (6.23) (9.19) (6.85) (7.73) N 23 23 23 23 23 23 23 23 23 23 R-squared 0.153 0.177 0.153 0.281 0.317 0.188 0.189 0.466 0.251 0.541 Prob > F 0.191 0.285 0.356 0.093 0.219 0.125 0.253 0.007 0.131 0.014 (t statistics in parenthesis) * p<0.1, ** p<0.05, *** p<0.01

(32)

10. References

Anoruo, E., & Ramchander, S. (1998). Current account and fiscal deficits: Evidence from five developing economies of Asia. Journal of Asian Economics, 9(3), 487-501. doi:10.1016/s1049-0078(99)80099-2

Atesoglu, H. S. (1993). Balance-of-Payments-Constrained Growth: Evidence from the United States. Journal of Post Keynesian Economics, 15(4), 507-514. doi:10.1080/01603477.1993.11489957

Barro, R.J. (2001). Economic growth in East Asia before and after the financial crisis (Working Paper No. 8330). Retrieved from National Bureau of Economic Research website: http://www.nber.org/papers/w8330

Boorman, J., Lane, T., Schulze-Ghattas, M., Bulir, A., Ghosh, A. R., Hamann, A. J., Phillips, S. (n.d.). Managing Financial Crises: The Experience in East

Asia. SSRN Electronic Journal. doi:10.2139/ssrn.243522

Cebula, R. J. (1995). The impact of federal government budget deficits on economic growth in the United States: An empirical investigation, 1955–

1992. International Review of Economics & Finance, 4(3), 245-252. doi:10.1016/1059-0560(95)90042-x

Chowdhry, B., & Goyal, A. (2000). Understanding the financial crisis in Asia. Pacific-Basin Finance Journal, 8(2), 135-152. doi:10.1016/s0927-538x(00)00011-1

Corsetti, G., Pesenti, P., & Roubini, N. (1999). What caused the Asian currency and financial crisis? Japan and the World Economy, 11(3), 305-373.

doi:10.1016/s0922-1425(99)00019-5

Loría, E., & Fujii, G. (1997). The Balance of Payment Constraint to Mexico's Economic Growth 1950–1996. Canadian Journal of Development

(33)

Studies/Revue canadienne d'études du développement, 18(1), 119-137.

doi:10.1080/02255189.1997.9669697

Martin, R., & Fardmanesh, M. (1990). Fiscal variables and growth: A cross-sectional analysis.Public Choice, 64(3), 239-251. doi:10.1007/bf00124369

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

banking and financial markets. Harlow, England: Pearson.

Rosensweig, J. A., & Tallman, E. W. (1993). Fiscal policy and trade adjustment: Are the deficits really twins? Economic Inquiry, 31(4), 580-594.

doi:10.1111/j.1465-7295.1993.tb00892.x

Stock, J & Watson, M. (2012). Introduction to Econometrics. Pearson Education Limited. 3rd Edition.

Thirlwall, A. (1980). Regional problems are “balance-of-payments” problems. Regional Studies,14(5), 419-425.

doi:10.1080/09595238000185371

Thirlwall, A. P. (2007). Regional Problems are “Balance-of-Payments” Problems. Regional Studies, 41(sup1), S89-S95.

doi:10.1080/00343400701232249

Data source:

World Development Indicators, World Bank

<http://databank.worldbank.org/data/reports.aspx?source=world-development-indicators>

Bruegel <http://bruegel.org/publications/datasets/real-effective-exchange-rates-for-178-countries-a-new-database/>