The Role of Macroeconomic Stability in Current Account Balances

The role of macroeconomic stability in current account balances has not been studied with a calculated index in the literature until now. It is aimed to nd out the role of macroeconomic stability in current account balances for the rst time in the study. The analysis is completed for the period between 1980 and 2016 for 97 countries. The macroeconomic stability is represented by an index which is created with ination rate, growth rate, unemployment rate and scal balance data of all the countries. It is found out that the macroeconomic stability is one of the important determinants of current account balances like institutional quality and nancial development. It has a negative and statistically signicant relationships with current account balances for four different country groups which are developing countries, all countries except industrial, all countries except industrial and African countries, and all countries. Results show that the macroeconomic stability is especially important for the developing countries rather than high income countries.


Introduction
Global imbalances exceeded 5% of the world GDP in 2008 from 2% in 1996. In 2018 it is still close to 3% of the world GDP (Fig.1). The most important change was the increase of USA's current account de cit. In 1997, it was less than 2% of GDP, and it increased to 5.8% in 2006. The current account de cit of the USA after nancial crisis decreased to 2.3% of GDP in 2009. The latest de cit of the USA is still 2.6% in 2019.
The USA is running current account de cits since 1992. China is running for 25 years of surpluses since 1994. Foreign trade of China increased quickly after the membership of the country in the World Trade Organization in 2001. Total trade between the USA and China was nearly US$559 billion in 2019.
However, that trade between two countries was unbalanced. USA started to run a huge trade de cit. The short-term borrowing, high level of unemployment rate, high level of increasing public debt, two digit increasing in ation rates, negative or declining GDP growth rates show a country's level of macroeconomic instability quite easily. Likewise, positive scal balance, current account balance surplus with decreasing debt levels, decreasing in ation rate in one digit, increasing growth rate are the signs of macroeconomic stability (Ames et al., 2001). Empirical studies also show that there is a positive relationship between macroeconomic stability and FDI in ows.

Macroeconomic Stability and FDI In ows
In several studies growth rate, in ation rate, total debt divided by GDP, exchange rate are found some of the determinants of FDI in ows representing macroeconomic stability. Onyeiwu and Shrestha (2004) nd for the period between 1975 and 1999, the economic growth and the in ation rate are some of the key determinants for FDI in ows for 29 African countries. Kok and Ersoy (2009) nd out that for the period between 1976 and 2005, the economic growth in developing countries has a positive signi cant relationship with FDI ows, and the total debt service/GDP and in ation have a signi cant negative relationship with FDI in ows for 24 developing countries. Ranjan and Agraval (2011) study foreign direct investment in ow determinants in BRIC countries: Brazil, Russian Federation, India andChina between 1975 and. They nd that macroeconomic stability is one of the determinants of FDI. In ation rate is also used for macroeconomic stability in this study. Chan et al. (2014) nd out that for the short and long term, GDP growth is one of the determinants of FDI ows for China. Shah (2016) studies the determinants of FDI ows for 43 African developing countries between 1990 and 2015. In his study in ation rate and direct exchange rate are used to proxy the macroeconomic stability. The result shows slow, predictable and steady depreciation of exchange rate with low in ation rate increase FDI ows. These studies mainly are completed for the developing countries. If macroeconomic stability is one of the determinants of FDI ows, it has to affect current account balances. If capital in ows increase when macro stability increases current account de cits have to increase as well. This study aims to understand if there is a relationship between macreconomical stability and current account balances.

Macroeconomic Stability Index
In literature, there are different ways to de ne macroeconomic stability. For example, in the eld of decentralization, price stability, which is measured by in ation as the proxy, is used for macroeconomic stability (Treisman, 2000;King and Ma, 2001;Neyapti, 2004;Shah, 2006 andThornton, 2007). Martinez-Vazquez and McNab (2006), Iqbal and Nawaz (2010), suggest using the Misery Index which is the combination of in ation and unemployment to proxy for measuring macroeconomic stability. Misery Index (invented by Arthur Okun) is calculated by taking the sum of the unemployment rate and the in ation rate for a given period. Bilan et al. (2019) use macroeconomic stability index which is the combination of ve different factors, in their study.
Macroeconomic stabilization is a very important subject for the policymakers in all countries, irrespective of their level of development. Although the indicators used to evaluate macroeconomic stability, are de ned differently, it increases the ability of the countries to prevent and absorb different external and internal shocks and increases economies' capacities to minimize their negative effects. Macroeconomic stabilization is therefore estimated in different ways, depending on the purpose of analysis and the aim of the governments' economic policies. These indicators shouldn't be regarded separately, as they are highly interdependent. For instance, high foreign debt is not necessarily showing low macroeconomic stability alone, if the debt is used for investment purposes which will increase the growth rates. In most of the cases, the macroeconomic stability is evaluated through ve basic indicators (Zaman and Drcelic, 2009): a) GDP growth rate, showing the overall potential of the economies to sustain a positive economic growth.
b) The unemployment rate, showing the capacity of the economies' full utilization of the human potential.
c) The in ation rate is generally measured by the Consumer Price Index (CPI), which shows the purchasing power of the citizens.
d) The scal balance (as a percentage of GDP), showing how the public resources are allocated.
e) The foreign debt, showing how many nancial resources are borrowed from foreign countries.
For macroeconomic stability, there must be a certain growth rate. The unemployment rate must be as low as possible. The in ation rate must be decreased to a level that does not cause any wealth or income redistribution. The government budget must be balanced and if possible, there must be a budget surplus. For our study, we will use a macroeconomic stability index which is combined of sub-indexes of the GDP growth rate index, the unemployment rate index, the in ation rate index and the scal balance (as a percentage of GDP) index. The indicators are calculated according to the UN methodology of international comparisons. It is a multi-criteria analysis. The original values are recalculated on a scale between 0 and 1. For translating the original levels into the new scale, any initial point X i which is between the minimum value of X min and the maximum value of X max , will have a correspondent Y i on the new scale.
The original values of the consumer price indexes are generally converted in natural logarithmic terms. The effect of the in ation cannot be considered as linear. In ation reduction from 400% to 40% is signi cantly less important than the decrease of in ation from 40% to 4%. In ation is 10 times lower for both situations, but 400% and 40% are still high rates of in ation. On the other hand, 4% of in ation can be assumed acceptable (Kolodko, 1993). We agree with this approach and use natural logarithmic terms for in ation.
The scale is reversed for three indicators, as a higher value for these indicators is equivalent to less stabilization. This is valid for the unemployment rate, the scal balance, and the in ation rate. For the growth rate the upper limit 10% and the lower limit 0% are selected. For the unemployment rate, the upper limit 25% and the lower limit %4 is selected. For the in ation rate, the upper limit 4.61 (ln 100) and the lower limit 0.69 (ln2) are selected. For the scal balance, the upper limit of 2% and the lower limit of -10% are selected. Zaman and Drcelic (2009) preferred to use 2.5% in ation for the analysis of the Serbian economy for the lower limit. Since most of the central banks use an in ation target of 2%, we use 2% instead of 2.5%. For the fractional component of unemployment, we use 4% instead of 5%. A lot of countries can reach that level without a high level of in ation. Based on these upper and lower limits sub-indexes are calculated.
Finally, by adding up these sub-indexes, the macroeconomic stability index is calculated for all countries. Maximum score for the macroeconomic stability index can be 4. Values will be distributed between 0 and 4.

Data And Methodology
It is aimed to nd out the role of macroeconomic stability in current account balances the rst time in the study. The analysis is completed for the period between 1980 and 2016 of 97 countries. The macroeconomic stability is represented by an index which is created with in ation rate, growth rate, unemployment rate and scal balance data of all the countries.
3.1 Data and model 7 macroeconomic variables that are added in the study. They are growth rate, terms of trade, real effective exchange rate, trade openness, net crude oil export, scal balance, and the relative income. Total private credit by nancial sector is the only nancial variable that is used in the study. Macroeconomic and nancial variables are added in the study starting from year 1980. 7 institutional variables are included in the study. The legal system and property rights index is available since 1980. Other institutional variables, voice and accountability, political stability and absence of violence, regulatory quality, control of corruption and rule law are available starting from 1996. The sources of the data can be seen appendix table 2.
To nd out the role of macroeconomic stability in current account balances, they are regressed onto a set of macroeconomic, nancial and institutional variables as well as macroeconomic stability index. First baseline results are estimated only based on the regression of macroeconomic factors. Then institutional and nancial factors are included. Eventually, the macroeconomic stability index is added to the base model. Cheung et al. (2013), Chinn et al. (2014), Altayligil and Çetrez (2020) have similar approaches. It is estimated as: CA represents current account balances, M represents macroeconomic variables, F represents nancial variable, I is the set of institutional variables and S is the macroeconomic stability variable.

Econometric methodology
Panel data analysis method is used to test the role of the macroeconomic stability in current account balances. Annual data is used in the model. Legg et al. (2007) use annual data in their analysis. Chinn and Prasad (2003) use both non-overlapping 5-year averages of the data and annual data as well. We are not just concentrated on the mid-term determinants of current account balances. We also include short term key variables like growth rate, the real effective exchange rate to see the full picture. So specially to capture the short-term effects better our model comprises annual data. The panel data set is unbalanced.
A lot of panel datasets may show signi cant cross-sectional dependency. Cross-sectional dependency must be checked at the beginning of the analysis. In the study, Paseran test (2004) is used to check the cross-sectional dependency among all the variables. First generation unit root tests must not be used when there is cross-sectional dependency among the variables. It is observed that most of the time that there is cross-sectional dependency among them. Only second-generation root tests consider cross sectional dependency. So, both second generation Paseran (2007) and rst-generation Fisher (Choi, 2001) panel root tests are used at the same time whether there is cross-sectional dependency or not.
The results of Paseran (2007) and Fisher (Choi, 2001) unit root tests show that the rst differences of all the variables are always stationary. Unlike, most of the time variables themselves are not stationary. OLS models are used with the rst differences of the variables. F results con rmed the usage of OLS models. Gruber and Kamin (2005), Chinn and Ito (2007), Cheung et al. (2013) express the variables as deviations from GDP-weighted averages. This is one way for controlling for unobserved heterogeneity (or common errors) by using OLS estimation of transformed models (Gromley and Matsa 2014). The other way to do that is to use the rst differences OLS models (Wooldridge, 2002).
So only the rst differences of all the variables are stationary and at the same time we want to control for unobserved heterogeneity, we decide to use the rst differences OLS model. Autocorrelation and heteroscedasticity for all the models are checked by using White (1980) and Wooldridge (2002) tests, individually. The only heteroscedasticity is identi ed in most of the models and OLS models are xed by Huber (1967), Eicker (1967) and White (1980) robust estimators.

Results
Before we add the macroeconomic stability index, we must create a baseline model with macroeconomic, nancial and institutional determinants. The rst step, estimation results for the macroeconomic variables are shown in Table 1. The last column shows the results of all countries. Institutional and nancial variables are included in the model step by step.

Baseline: Macroeconomic Institutional and Financial Determinants
Baseline model must be created before the estimation of the role macroeconomic stability in current account balances. The growth rate, scal balance, terms of trade, real effective exchange rate, trade openness, crude oil export, and nancial market development are estimated as the determinants of current account balances for different country groups for the baseline equation (table 1). Page 9/20 First differences of the variables are used. t-statistics are shown in parenthesis. *, **, *** indicate significance level at 10%, 5%, 1%, respectively. The constants' estimations are not shown.
There is a negative relationship between growth rate and current account balances for ve different country groups. When growth rate increases, the current account de cits increase for all country groups except industrial countries. Only for industrial countries, there is a positive relationship which means export supports growth rates for industrial countries. In line with expectations, there is a positive relationship between scal balance and current account balance which means, when the public de cit increases, the national savings reduce. There is a positive relationship for the terms of trade. Results support Harberger, Laursen and Metzler (1950) effect.
Real effective exchange rate has a negative sign. Results show an increase in real effective exchange rate leads to an increase in the current account de cit. Results support Marshall Lerner condition. Based on the results, there is a negative relationship between trade openness and the current account balances. Countries, when they are more open to international trade, will run higher current account de cits. But this factor is not signi cant for industrial and high-income countries groups.
The stage of economic development is found as another determinant of current account balances only for two country groups, developing countries and full sample without industrial countries. When the stage of economic development increases, current account de cit increases for these country groups. These results support the validity of Lucas Paradox for these country groups.
Results show countries with important crude oil export run higher current account balances, this relationship is not signi cant only for industrial countries and high-income countries. Financial market development is found to have overall negative and statistically signi cant relationships with current account balances for all country groups (table 2).

The Role of Macroeconomic Stability
Ben Bernanke's (2005) one of the suggestions is, developing countries must improve their investment climate by continuing to raise macroeconomic stability. Macroeconomic stability index is added to the base model. When macroeconomic stability gets higher more nancial capital out of the country is expected to enter the country. As a result, the current account de cits will increase. Negative relationship is expected between the macroeconomic stability index and current account balances for especially the developing countries. Altayligil and Çetrez (2020) nd a positive relationship between in ation rate and current account balances for some country groups. They use consumer price index to represent the macroeconomic stability.
Macroeconomic and nancial determinants are added to the base model. Institutional determinants are not added in the base model because there is not even one same institutional determinant for more than one country group (appendix table 4). Since the growth rate and macroeconomic stability index are highly correlated growth rate is excluded from the model. There is a negative relationship between macroeconomic stability index and current account balances. Higher macroeconomic stability for all country groups except industrial and high-income countries (table 3) increases the current account de cits. When economic stability gets better, this causes more capital in ow to these countries and ends up with higher current account de cits. First differences of the variables are used. t-statistics are shown in parenthesis. *, **, *** indicate significance level at 10%, 5%, 1%, respectively. The constants' estimations are not shown.
When the same regressions are re-estimated using each of the components of the macroeconomic stability index, in order to isolate which factors, drive the results. The test results suggest that unemployment is the most signi cant contributor among the four variables. It has a statistically signi cant positive relationship with current account balances for all country groups. The growth rate follows unemployment it has a statistically signi cant negative relationship with current account balances for all country groups except industrial countries. The in ation rate has no signi cant relationship.
Fiscal balance has statistically signi cant positive relationship with current account balances for all country groups except industrial and high-income countries. This is expected when scal balance is considered alone because an increase in the public de cit can decrease national savings if there is not a Ricardian offset from private savings and this may end up with a rise of current account de cits. Positive relationship has been found in many studies. As a proxy for the macroeconomic stability index negative sign may be expected. But as seen the rst effect is more important when scal balance is considered alone.
According to the results, macroeconomic stability is more important for developing countries. Macroeconomic stability is one of the determinants for all country groups except industrial and highincome countries. Better macroeconomic stability causes more capital in ows to these countries and ends up with higher current account de cits. So, to understand the effect better some of the emerging economies, like Poland, Russia, South Africa, Turkey, Brazil, and China are selected because they are some of the largest developing countries with high current account balances.
For the estimation, the equation for all countries except Industrial and African countries is used. The equation can simulate the direction of current account patterns reasonably well for the selected developing countries (Fig. 2) When the decompositions of the rst differences of current account balances are checked, it is seen that for Poland, Russia, Turkey and Brazil macroeconomic stability is one of the key determinants (Fig. 3). It has signi cant effect on current account balances of these countries. South Africa and China it has relatively less impact. It can be also seen the contribution of the other determinants in detail. Real effective exchange rate and oil export are two other key determinants for these countries.

Conclusions
The role of macroeconomic stability in current account balances has not been studied with a calculated index in the literature until now. It is aimed to nd out the role of macroeconomic stability in current account balances the rst time in the study. The analysis is completed for the period between 1980 and 2016 for 97 countries. The macroeconomic stability is represented by an index which is created with in ation rate, growth rate, unemployment rate and scal balance data of all the countries. It is found out that the macroeconomic stability is one of the important determinants of current account balances. It has a negative and statistically signi cant relationships with current account balances for four different country groups. Results support that the macroeconomic stability is especially important for the developing countries rather than high income countries.
The main aim of the study is to understand the role of macroeconomic stability on current account balances. The macroeconomic stability index is created and calculated for each country individually. For the rst time, macroeconomic stability which is represented by a new index is added to the study.
Macroeconomic stability is discussed by Bernanke (2005) as one of the potential determinants of current account balances. Macroeconomic stability which is represented with an index in the study, is found to have statistically signi cant relationships with current account balances for all the country groups except industrial countries and high income countries with negative signs. Higher macroeconomic stability increases the current account de cits. When economic stability gets better, this causes more capital in ows to these countries and ends up with higher current account de cits. It is also seen that for Poland, Russia, Turkey and Brazil macroeconomic stability is one of the key determinants of current account balances. South Africa and China, it has relatively less impact. It is found out that like institutional quality and nancial development macroeconomic stability is one of the key determinants of current account balances especially for the developing countries.
Declarations Acknowledgments Not applicable.
Authors' contributions MC and YA agreed on the content of the study. MC collected all the data for analysis. MC and YA agreed on the methodology. MC completed the analysis based on agreed steps. Results and conclusions are discussed and written together. Both authors read and approved the nal manuscript.

Funding
No funding is used.

Availability of data and materials
The datasets used and analysed during the current study are available from the corresponding author on request. First differences of the variables are used. t-statistics are shown in parenthesis. *, **, *** indicate significance level at 10%, 5%, 1%, respectively. The constants' estimations are not shown.