1
Faculty of Economics and Business MSc Economics
Track International Economics and Globalization
Does Financial Openness Affect The
Current Account Imbalance?
Kejiang Liu
10614222
Master Thesis
2018/July/8
Email: lkj19940908@126.com Supervisor: Drs. N.J. LeefmansSecond reader: Dr. D.J.M. Veestraeten Number of words: 11939
Statement of Originality
This document is written by Student Kejiang Liu who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are 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
Content
1. Introduction ... 4
2. Literature review: effect of financial openness on current account... 8
3. Theoretical Framework ... 10
3.1 Saving-Investment Approach ... 11
3.2 The effect of cross-border capital flows on countries’ industries ... 12
3.3 The effect of financial openness on export and import ... 13
4. Econometric frame work ... 13
4.1 The medium term regression model... 13
4.2 Description of variables ... 14
4.3 Sample selection ... 17
4.3.1 Period selection ... 17
4.3.2 Sample countries selection ... 18
4.4 Data collection ... 19
5. Empirical results ... 20
5.1 Regression model- financial openness measured by KAOPEN ... 20
5. 2 Regression model - financial openness measured by TOTAL ... 23
6. Conclusion and Discussion ... 26
References ... 29
Abstract
Current account imbalances have already been a serious problem towards several countries for a long time, and financial openness is considered as one determinant of the current account. This thesis collects data from 1997-2011 and uses a regression model to investigate the effect of financial openness on the current account. This thesis finds that financial openness has a significant effect on the current account which is conditional on the economic environment. Furthermore, the effect of financial openness on the current account is different in industrialized countries and emerging countries.
5
1. Introduction
Globalization has been the main trend of the world economy for several decades. On one side it brought a great welfare increase at the global level since production resources can be used more efficiently, however on the other side this also led to global current account imbalances problems, which means the current accounts of countries are persistently non-zero thus not in equilibrium. Figure 1 shows the current account as a percentage of the world GDP of several representative countries for the developed and emerging countries.1 It shows the global current account imbalances
raised rapidly both in the surplus and deficit countries since the middle of the 1990s, then dropped a bit after the 2008 Global Financial Crisis, before increasing again in recent years. This figure also reflects that global imbalances mainly refer to the US for the current account deficits, while on the other hand, the main contributors for current account surpluses are China and the south-east Asian emerging countries. The current account imbalances reflect distortions in domestic economics, and can hurt the stability of world economic growth (Blanchard and Milesi-Ferretti, 2012). Therefore it is important to investigate the factors that will affect the current account imbalances.
Figure 1. The current account as a percentage of world GDP
Source: The World Development Indicators (WDI) Data Base2, calculated by
author
1 SGP is Singapore, KOR is Korea, IDN is Indonesia, VNM is Vietnam, THA is Thailand, MYS is Malaysia, DEU
is Germany, JPN is Japan, CHN is China, USA is USA and GBR is Britain
2 See https://data.worldbank.org/ -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 SGP KOR IDN VNM THA MYS DEU JPN CHN USA GBR
Financial openness is considered as a factor that can influence the current account. Astley et al. (2009) and Bracke et al. (2010) suggested higher financial openness leads to higher cross-border capital flows, which are associated with the current account. As globalization has increasingly engaged capital markets, more and more countries especially emerging countries which used to be with low financial openness (high entry barriers for foreign financial business, and strict limitation of foreign capital) began to implement policies that liberalize capital flows. Figure 2 shows the gross foreign assets plus foreign liabilities as a ratio of GDP for countries with different development levels. One can see since the 1970s the ratio increased rapidly; although it dropped a bit around the Global Financial Crisis, it still has a main upward trend. The future world level of financial openness is likely to keep increasing and especially for the emerging countries who contribute a large part on the surplus side of global imbalances. With such a trend, it is important for policy makers and researchers to know the impact of a further increase in financial openness on the current account imbalances. So the research question of this thesis is as follows: what is the effect of financial openness on current account imbalances? Nevertheless, this thesis will also investigate whether the effect of financial openness on the current account is conditional on some economic environment related factors (in this thesis: financial development, legal development and government effectiveness).
Figure 2. The sum of gross foreign assets and liabilities as a ratio of GDP
Source: Dabla-Norris et al. (2015)
Surprisingly, although global imbalances are much debated in the literature, and financial openness is considered as a main determinant of the global imbalances, there are only few studies focusing on investigating the effect of financial openness. Chinn and Ito (2008)a suggested that such a lack of studies towards financial openness might
7
focusing on this question can be found (see Saadaoui, 2015; Chinn and Ito, 2008b;
Chinn et al., 2013), but these studies only used data until 2008b. This thesis will
contribute to the existing literatures by: using the most recent available data from 1997 to 2011 which including data after the Global Financial Crisis (GFC) and control the effect of the GFC by adding a dummy variable for the period after 2008. Furthermore, this thesis argues the effect of government effectiveness on the current account should be taken account. Government effectiveness is the quality of the formulation and implementation of the government’s policies, and the credibility of these policies. Chinn and Ito (2008)b stated that the economic environment may affect
the way how financial openness affects the current account, and this thesis argues government effectiveness should also be one part of the economic environment since government policies can affect the domestic economic development. So this thesis will control for the effect of government effectiveness on the current account. Furthermore, as noticed by Chinn and Ito (2008)b, the relationship between the current
account, financial openness, financial development and legal development might be non-fixed and rather conditional on some variables. So this thesis will also investigate whether government effectiveness will affect the association between financial openness and the current account. Nevertheless, the economies in industrialized countries and emerging countries differ a lot, for instance in terms of the industrial structure, public sector, people’s knowledge towards financial business, life expectancy and labor market. Therefore, the impact of changing financial openness might also differ a lot in the two types of countries, which means it is not proper to take both types of countries in one panel. This is the reason that the existing literature (see Saadaoui, 2015; Chinn and Ito, 2008b; Chinn et al., 2013) separates their sample
countries into an industrialized country group and an emerging country group. Therefore this thesis will select two sample groups for industrialized and emerging countries separately, and do specific research to investigate the causal effect of financial openness on the current account for each group of countries.
This thesis will use panel data of two groups of countries (an industrialized and emerging country group, including industrial 26 countries: Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Iceland, Ireland, Israel, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Slovenia, Singapore, South Korea, Spain, Sweden Switzerland, the UK and the US; and 27 emerging countries: Argentina, Brazil, Chile, China, Colombia, Ecuador, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Morocco, Pakistan, Peru, Philippines, Poland, Romania, Russia, South Africa, Sri Lanka, Thailand, Tunisia, Turkey, Ukraine, Venezuela and Vietnam). Data from 1997-2011 are used to investigate the causal effect of financial openness on the current account in the medium term by regressing the current account on financial openness and other control variables. The remainder of this thesis will introduce this model in details.
The remainder of the thesis is arranged as follows: Section 2 reviews the existing literature on the causal effect of financial openness on the current account. Section 3
will introduce the theoretically possible mechanisms through which financial openness can affect the current account. Section 4 introduces the regression model, variables selection, sample selection, the measurement of variables and the resources where the data used in this thesis is collected from. The empirical results will be presented in section 5. Finally section 6 will draw the conclusion and discussion.
2. Literature review: effect of financial openness on current account
This section will review the existing literature about the impact of financial openness on the current account.
Chinn and Ito (2008)b used data from 1971 to 2004, did a research and tried to
figure out the medium-term determinants of the current account. They collected data from 18 industrialized countries and 69 developing countries. They divided these countries into 4 sample country groups: industrialized countries, less-developed countries, less-developed countries excluding Africa and emerging countries. They used pooled OLS to regress the medium-term current account on a variety of factors which include government budget balance, net foreign assets, income level relative to the US and its square, dependency ratio of young and old, financial development, legal development, financial openness and interactions of each two of financial openness, legal development and financial development, term of trade volatility, GDP growth, trade openness and a dummy variable for oil-exporting countries. In their model all independent variables are converted into the deviation from their GDP-weighted average world level. In their result they found that the financial openness doesn’t have a significant coefficient in the full sample and all 4 sub country groups. However, in the whole sample, the interaction of financial openness and legal development has a significant positive coefficient at the 10% significance level, which means one unit increase in financial openness will increase the current account by (0.002*Legal Development) percentage point. In industrialized countries, the interaction of financial openness and legal development also has a significant positive coefficient at the 1% significance level, according to which one unit increase in financial openness shall increase the current account by (0.012*Legal Development). This result shows that instead of having a fixed effect, financial openness is more likely to have a variable effect on the current account which depends on other variables. In the other 3 sample groups: less-developed countries, less-developed countries excluding Africa and emerging countries, no interaction term related to financial openness has a significant coefficient. Combining this with the fact that the authors found financial openness doesn’t have a significant coefficient in these 3 sample groups, this research shows that financial openness cannot affect the current account in these three groups of countries. This research takes its sample period starting from 1970, which is early; during this sample period the world economic, financial market and industries had huge changes. (For instance, the reform and opening in China in the 1980s, the Digital Revolution in industries and the
9
innovations in the global financial system.) This fact means early data cannot provide sufficient explanation of recent financial and economic mechanisms.
Saadaoui (2015) did a research focused on the causal effect of financial openness on the medium-term current account, the author collected panel data from 1980-2003 for 18 industrialized countries and 21 emerging countries whose sum of GDP is a large fraction of the world GDP. In his research, a regression model of current account to GDP ratio on relative financial openness (relative to the weighted average world level) was set up to investigate the causal effect of the differently measured financial openness degree on the medium-term current account. Different from the research done by Chinn and Ito (2008)b, the author took account of fixed effects and regressed
the current account on factors including relative financial openness, relative population growth and relative output gap3 (All “relatives” mean relative to weighted
world average level). One interesting point is that besides the Chinn-Ito index which was used by both studies as the measure of financial openness, Saadaoui (2015) also used another de facto measure of financial openness to check the consistency of the regression result which is the relative gross foreign assets as a percentage of GDP (relative to the weighted average world level). According to the regression result， when financial openness is measured by the relative Chinn-Ito index, for industrialized countries, a 1 unit increase in the relative Chinn-Ito index will increase the current account to GDP ratio by 0.92 percentage point, while for emerging countries same increase in the relative Chinn-Ito index will decrease the current account to GDP ratio by 0.63 percentage point. When financial openness measured by relative gross foreign assets in percentage of GDP, for industrialized countries a 1 unit increase in relative gross foreign assets to GDP ratio will increase the current account to GDP ratio by 0.007 percentage point, while for emerging countries same increase in relative gross foreign assets to GDP ratio will decrease the current account to GDP ratio by 0.02 percentage point. One thing notable is in the regression all variables are the non-overlapping 4 year average value. This research shows that increasing financial openness has a positive effect on the current account in industrialized countries while it has a negative effect in emerging countries. The author suggested that this result is caused by the fact that higher financial openness allows investors in industrialized countries to make investments abroad more extensively and the emerging countries can receive the investment from abroad more extensively. Under the fact that emerging countries are usually with current account surpluses while developed countries are with deficits, the further liberalization on capital account can solve the global imbalance problem to some degree. However, unlike Chinn and Ito (2008)b, this study didn’t use many control variables. According to Chinn and Ito
(2008)b, government budget balance and net foreign assets both have significant
effects on the current account, and these two factors are considered correlated with financial openness. This means that although the author used panel data and took account of fixed effects, the regression model still suffers from an omitted variable problem, which then leads to a biased outcome. Furthermore, as the sample size is
limited, using a fixed effect model with 39 countries and only 6 cross-sections will significantly decrease the degree freedom of the regression model.
Chinn et al. (2013) did an almost similar research as Chinn and Ito did in 2008, which has been reviewed above. In this research the authors also used pooled OLS to regress the medium-term current account on a variety of factors. One main difference is the new research takes a sample for the period 1970-2008. And the other one is that this research adds two dummy variables for the years 2001-2005 and years 2006-2008. The former is the period when credit increased rapidly due to the robust economic expansion and the revolutions in financial institution, the latter is period just before the world financial system lost control and then led to the Global Financial Crisis in 2008. The result is also different from the early research, since in the whole sample, financial openness doesn’t have a significant coefficient, while the interaction of financial openness and legal development has a significant positive coefficient at the 1% significance level, which means a one unit increase in financial openness will increase the current account by (0.003*Legal Development). In industrialized countries, the financial openness has a significant positive coefficient at the 10% significance level, both of the interaction of financial openness and legal development and the interaction of financial openness and financial development have a significant coefficient at the 5% significance level. These coefficients shows that a one unit increase in financial openness will increase the current account by (0.008+0.012*Financial Development+0.028*Legal development) percentage point. In emerging countries, the interaction of financial openness and financial development has a significant negative coefficient at the 10% significance level, this means a one unit increase in financial openness will decrease the current account by (0.02*financial development) percentage point. Given the fact that some industrialized countries suffers a lot from the current account deficits and some emerging countries contribute a lot to the current account surpluses, the increase in financial openness and financial development in these countries can solve the global imbalances problem to some degree. One notable thing is that compared with the result of Chinn and Ito (2008)b,
with almost the same regression model, just adding 2 recent years can lead to a large difference in the regression result, this fact suggests that it is more proper to use an even more recent period in the sample. One more notable thing is the dummy variables; the 2001-2005 dummy variable has a significantly positive effect on the current account in all country groups while the 2006-2008 dummy variable just shows a significant positive effect in less developed countries.
3. Theoretical Framework
This section will introduce the theoretical mechanisms through which financial openness can have an effect on the current account.
11
3.1 Saving-Investment Approach
The current account can be seen as the national saving-investment balance as follows.
According to the Keynesian model:
Y = C + I + G + X – M (1) The national saving is the amount of national income that is not consumed by people or spent by government:
S = Y – C – G (2) Combining (1) and (2) gives:
S – I = X – M = CA (3) Where Y is GDP, C is private consumption, I is domestic investment, S is national saving, G is government spending, T is government’s tax revenue, X is export, M is import and CA is current account. The left side of equation 3 is the domestic saving-investment balance.
According to equation 3, the current account imbalance can be explained by the gap between the national saving and investment rate. So this thesis can analyze the effect of financial openness on the current account according to the saving-investment approach.
From this point of view it is clear that financial openness can affect the current account since it can change the capital flows across countries and therefore the countries’ saving-investment balances. Astley et al. (2009) suggested that capital is attracted by more productive investment opportunities in other words investments with higher rates of return. For instance, compared with developed countries, the emerging countries have large space for economic growth, and history has shown in recent years that they had much higher GDP growth than developed countries, therefore there are larger amount of investment opportunities in emerging markets and they can yield a higher rate of return than those in developed countries. In this case, higher financial openness in either emerging countries or industrialized countries will lead more capital to move from developed countries to emerging countries. According to equation (3) above, these capital flows will increase the domestic investment rate in emerging countries and decrease it in developed countries, and thus decrease the current accounts in the former and increase it in the latter.
Nevertheless, with much more mature financial systems, higher investment techniques, more efficient resources allocation and investment risk control for financial institutions, the financial markets in industrialized countries also have high attractiveness towards capital flows from emerging countries. Based on the
saving-investment approach, Bernanke (2005) and Clarida (2005) built up the global “saving glut” hypothesis in their research to explain the large current account deficit in the US during the 2000s. They stated that this large current account deficit in the US during the 2000s is caused by the underdeveloped financial market in Asian emerging countries: the increasing financial openness allowed them to export their high savings to the US. With a more efficient allocation of resources and higher ability to provide safe financial assets, the financial markets in developed countries are attractive, and can attract the savings in emerging countries, especially for the Asian emerging countries. After the Asian Financial Crisis, the investment rate in the Asian emerging countries collapsed due to the impact of the financial crisis while the saving rate still stayed at a relatively high level (Chinn, et al., 2013), this is caused by the fact that the capital owners in these countries lost confidence on the domestic investment due to the financial crisis, and prefer saving rather than investing their capital into the domestic market. With a higher financial openness, these capital owners can invest their savings to the financial markets of developed countries. This means increase in financial openness can reduce the savings rate in emerging countries, and reduce their current account. However, different from Bernanke (2005) and Clarida (2005), it is not necessary that these capital flows from emerging countries to industrialized countries will finally become the domestic investments in industrialized countries, because with limited domestic investment opportunities and highly developed cross-border financial businesses, large part of these capital flows will be transferred to other countries, in the form of second-stage investments.
3.2 The effect of cross-border capital flows on countries’ industries
As mentioned above, changes in financial openness can affect cross-border capital flows, for instance, increasing financial openness gives more opportunities to domestic industries to attract foreign investments, and this affects the expansion of countries’ industries. The changes in cross-border capital flows might have different effect on one country’s export industries and import industries, and therefore also change its trade balance and the current account. Because the effect of capital flows on industries is in the medium term, so the effect of financial openness should also take place in the medium term. This is the reason why this thesis focuses on the effect of financial openness on the current account in the medium term.
Furthermore, Reinhardt et al. (2013) stated that financial openness affect cross-border capital flows differently in industrialized and emerging countries, Bekaert et al. (2011) also found that the effect of financial openness on a country’s industry production is different by development level. This suggests that through both theoretical approaches mentioned above, the effect of financial openness on the current account might be different in industrialized and emerging countries, and it is more appropriate to divide the sample into industrialized and emerging country groups and do research specifically.
13
3.3 The effect of financial openness on export and import
According to equation (3) mentioned above, it is also clear that the current account can be explained by the gap between exports and imports, in other words, the consumption decision on export and import of goods and services. People make consumption based on their expectations for the economy and their budget constraints. This means a shock on productivity or change in financial resources can change people’s expectations and budget constraints. As mentioned above, a change in financial openness can have an effect on the industry structure and therefore can make a shock on productivity, this will lead to a change in consumer’s expectation and then affect the current account. Furthermore, Faruqee and Lee (2009) suggest that an increase in financial openness can expand the international lending and borrowing possibilities, and therefore change people’s budget constraints then change their consumption decision and finally the current account. For instance, the large capital flows to the US in the early 2000s led to an increase in the credit allocation to domestic consumers, the increased credit allocation then increased the consumption level and also the gap between imports and exports, as a result, the current account imbalance in the USA was worsened. According to the above, the change in financial openness can also lead to an impact on people’s consumption decision, and then affect the current account.
4. Econometric frame work
From the above it is clear that theoretically financial openness can have an effect on the current account. According to this theoretical analysis, this thesis builds up its econometric frame work.
4.1 The medium term regression model
To investigate the effect of financial openness on the current account, this thesis will use an OLS regression model developed from Chinn et al. (2013), which is as follows:
CAi,t = α + β1KAOPENi,t + β2 FDi,t + β3LEGALi,t + β4GEi,t + β5
KAOPENi,t*FDi,t + β6 KAOPENi,t*LEGALi,t + β7 KAOPENi,t* GEi,t + β8
FDi,t* LEGALi,t + zi*Xi,t + ui,t
Where CA is the current account, KAOPEN is financial openness measured in Chinn-Ito index, FD is financial development, LEGAL is legal development and GE is government effectiveness. X is a vector of other control variables, which includes factors that are considered to affect the current account and are correlated with
financial openness: Government budget balance (BB), Net foreign assets (NFA), Trade openness (TO) and the period dummy variable for the period after 2008 when the Global Financial Openness boomed. (D2008).
The research may face endogeneity problems. First, the current account may also have a reverse effect on financial openness, for instance, a country with a current account deficit may increase its capital account openness to absorb funds from abroad in order to reform the production and solve the current account deficits. However as tested by Saadaoui (2015), there are no such endogeneity problems. Chinn and Ito (2008)b, Chinn et al. (2013) both used an OLS regression model, so this thesis will
also use OLS regression. Second, the endogeneity problem could also be caused by the omitted variables, so this thesis will use factors that are generally considered to affect the current account as control variables.
As mentioned before, this thesis will also try a de facto measurement of financial openness in the regression model to have a more clear view on the effect of financial openness on the current account in the medium term. More details for the measurement will be introduced in the later part of this section. The sum of foreign assets and foreign liabilities is chosen as the de facto measurement of financial openness in this thesis. When financial openness is measured by the de facto measurement the regression model will be:
CAi,t = α + β1TOTALi,t + β2 FDi,t + β3LEGALi,t + β4GEi,t + β5
TOTALi,t*FDi,t + β6 TOTALi,t*LEGALi,t + β7 TOTALi,t* GEi,t + β8 FDi,t*
LEGALi,t + zi*Xi,t + ui,t
Where TOTAL is the sum of foreign assets and foreign liabilities.
One notable thing is the dummy variable D2008 is used to control the impact of the
Global Financial Crisis on the current account.
As this thesis is aimed to study the current account in the medium term, all variables except the GFC dummy variable will be non-overlapping 3-year averages of the corresponding annual variables.
4.2 Description of variables
In this section this thesis will introduce the variables used in the regression model in detail and explain why they should be put in the regression model. Also, the regression models introduced above include non-objective variables which cannot be directly observed and measured. They are: Financial openness, financial developments, legal developments, government effectiveness and trade openness. This section will also introduce how this thesis measures them in the empirical
15
research. Furthermore, in this research all variables except the Chinn-Ito index, legal development, government effectiveness and the dummy variable for the period after the Global Financial Crisis are measured as their percentages to national GDP.
The Dependent Variable and Independent Variable: First of all, as this research
focuses on the effect of financial openness on the current account, so the current account and financial openness are the dependent variable and the independent variable.
Government Budget Balance: According to the saving-investment approach
introduced in section 3, government budget balance can directly affect the current account since the government budget balance partly determines national savings. Furthermore, Chinn and Ito (2007) suggested that government budget balance can also affect people’s saving and investment decision, this means it can also affect the current account by affecting private savings and investments. Also, Chinn and Ito (2008) and Chinn et al. (2013) found that government budget balance has a significant effect on the current account. So government budget balance is used as one control variable in the regression model.
Net Foreign Assets: Net foreign assets should affect the current account because the
return of net foreign assets or the payment of net foreign liabilities is one part of the current account. Chinn and Ito (2008)b and Chinn et al. (2013) found that the stock of
net foreign assets has a significant effect on the current account. So the stock of net foreign assets is used as one control variable in the regression model.
Trade Openness: Trade openness means the degree of how free one country is for
cross-border trades, this factor by definition can have direct effect on the trade balance and therefore also the current account. So in the regression model trade openness is used as a control variable. As learned from Chinn and Ito (2008)b and
Chinn et al. (2013), the trade openness is measured by the sum of exports and imports of goods and services as a percentage of national GDP.
The Dummy Variable for the Period after 2008: Since the Global Financial Crisis
made a serious impact on international financial market, the financial system and people’s appetite towards investment in worldwide changed a lot compare to years before the crisis (see Barth, 2015; Balaban et al., 2013), so it is necessary to add this dummy variable.
Financial openness: Financial openness means the degree of how free one country is
for cross-border capital flows. As mentioned before financial openness is hard to measure, a proper measurement should be able to capture the full complexity of real-world capital control while many existing measurements failed to do this. In order to measure financial openness, this thesis will use the Chinn-Ito index which is initially introduced by Chinn and Ito (2006). They collected raw data from IMF's
Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER), and build up 4 dummy variables for 4 major types of restrictions on cross-border financial transactions: dummy variable for whether there exists multiple exchange rates (k1), dummy variable based on the restriction on current account transactions (k2), dummy variable based on restrictions on capital account transactions (k3) and dummy variable based on the requirement of the surrender of export proceeds (k4). According to these dummy variables they finally construct a financial openness index: the Chinn-Ito index, which is the first standardized principal component of k1, k2, SHAREk3 (the share of 5-year window of k3) and k4. A higher Chinn-Ito index means fewer restrictions on cross-border financial transactions, and therefore higher financial openness. The Chinn-Ito index keeps updating after be introduced in 2006 and now it measured the financial openness of 182 countries from 1970 to 2015. Furthermore, Quinn, Schindeler and Toyoda (2011) suggested that de jure measurement to some extent cannot reflect the actual capital flows caused by the change in restrictions on cross-border financial transactions, they argued in this case a de facto measurement should be more proper to measure financial openness, they also suggested the sum of foreign assets and foreign liabilities (TOTAL) is a proper de facto measure of financial openness and is widely used in literature researches towards financial openness. So in addition to measuring financial openness by Chinn-Ito index and do the research, this thesis will also use the sum of foreign assets and foreign liabilities to measure the financial openness.
Financial development: financial development means the development level of the
financial market, it will be measured by the percentage of private credit to country GDP as learned from Chinn and Ito (2008)b. A country with higher ratio of private
credit to GDP means a higher development level of its financial market. Highly developed financial market means high liquidity and high efficiency of capital, and therefore can have high attractiveness towards cross-border capital flows. This means it can affect both current account and the association between financial openness and the current account. So while use it as a control variable, the interaction of financial development and financial openness is also in the regression model.
Legal development: Legal development means the development level of one
country’s legal system, including how sound the legal system is and how strong its regulatory force is. Levine et al. (2000), Johnson et al. (2002) and Beck and Levine (2005) find that the legal development can affect the financial intermediaries, this means it can also affect the cross-border financial transactions and capital flows. In this thesis legal development will be measured by the estimated indicator of rule of law in Worldwide Governance Indicators. This indicator reflects how confident people are towards their legal system and how willing they are to abide their law, especially for the quality of contract enforcement and the courts. As cross-border capital flows are depending on the decisions of investors, this indicator should be a proper measure for legal development since it reflects people’s confidence in the legal system. Same to financial development, the interaction of legal development and financial openness is also in the regression model.
17
formulation and implementation of the government’s policies, and the credibility of these policies. It can have an effect on cross-border capital flows since the quality, implementation and credibility of one country’s policies can affect the investor’s decision towards this country. This means the effect of government effectiveness on the current account should be controlled for. Furthermore, as mentioned before government effectiveness should be one part of the economic environment, so the effect of government effectiveness on the association between financial openness and the current account should be taken into account. In this thesis government effectiveness will be measured by the estimated indicator of government effectiveness in Worldwide Governance Indicator. A higher indicator means the government effectiveness is higher in one country.
Due to their nature, the Chinn-Ito index, the estimates of legal development and government effectiveness range from negative value to positive value this will lead to complexity of interpreting the estimated coefficients of related variables. As learned from Chinn and Ito (2008)b, in order to avoid such complexity, in this thesis the
values of financial openness measured by Chinn-Ito index, legal development and government effectiveness will be adjusted so that their minimum value will be zero, in other words the adjusted value will range from zero to positive value.4
4.3 Sample selection
4.3.1 Period selection
As mentioned above, this thesis uses a sample of countries from 1997-2011, 1997 is selected as the beginning year because the world economic structure and the financial market is always changing, which means that although collecting sample from earlier periods can provide a larger sample, earlier data cannot provide sufficient explanation of recent financial and economic mechanisms and will lead to a meaningless study result. Furthermore, 1996 is the first year when the estimated indicators of legal development and government effectiveness in Worldwide Governance Indicator are accessible5, but in order to keep the non-overlapping period
structure of the sample, 1996 is not selected. The Chinn-Ito index is now accessible to 2015. However the sum of foreign assets and liabilities is only accessible to 2011, so 2011 is chosen as the end year of this sample. Considering the medium-term perspective, all explanatory variables except the dummy variable of the Global Financial Crisis will be taken in the form of non-overlapping 3-year averages of the corresponding annual variables.
4 All these variable will be adjusted to non-negative based on yearly value
5 Because Worldwide Governance Indicator measured their indicators every two year for their first three periods,
so in first three periods these indicators reflect their related variable in two years. This means in this paper, the WGI related indicators in 1997, 1999 and 2001 will take the value of the first, second and third period value in WGI.
4.3.2 Sample countries selection
The economic structure in industrialized countries and emerging countries can be largely different, for instance in terms of the industrial structure, public sector, people’s education level, life expectancy and labour market. This difference means the mechanisms of how financial openness could affect the current account might also differ a lot in the two types of countries. This difference can also be explained by the theoretical analysis in section 3:
First, according to the saving-investment approach, financial openness can affect the current account by affecting the decision of savings and investments. According to the analysis in 3.1, the effect of financial openness on the decision of savings and investments in industrialized and emerging countries is different, for instance, increase in financial openness lead emerging countries to export savings to industrialized countries.
Second, according to the analysis in 3.2, financial openness can affect the current account by providing the domestic industries more opportunities to attract the cross-border capital flows. Due to the fact that the industry structure in industrialized countries and emerging countries differ a lot with each other, the effect of financial openness on the domestic industry and the current account might also be different.
Third, according to the analysis in 3.3, financial openness can affect the current account by changing the expectation towards economic growth and the budget constraints of consumers, in other words the consumption decision. Considering the consumption habits in industrialized countries and emerging countries might be different, and the main domestic products in industrialized countries and emerging countries are also different (industrialized countries mainly produce high quality products and emerging countries mainly focus on cheap and basic products), it is possible that the increase in financial openness changes the consumption decision in industrialized countries and emerging countries differently, and then changes the current account differently.
The result from Saadaoui (2015) also shows that financial openness has completely the opposite effect on the current account in industrialized and emerging countries; this fact suggests that it is not proper to take both types of countries in one panel. Therefore this thesis will select two sample groups for industrialized and emerging countries separately, and do specific research on each group separately to investigate the causal effect of financial openness on current account for each type of country. After checking the nature and the data accessibility of each country, this thesis finally selects 26 industrialized countries (Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Iceland, Ireland, Israel, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Slovenia, Singapore, South Korea, Spain, Sweden Switzerland, the UK and the US) and 27 emerging countries (Argentina, Brazil, Chile, China, Colombia, Ecuador, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Morocco, Pakistan, Peru, Philippines, Poland, Romania, Russia, South Africa, Sri Lanka, Thailand, Tunisia, Turkey, Ukraine, Venezuela and
19 Vietnam).
4.4 Data collection
In this thesis data are collected from different sources because there is no database that can provide all the data of needed variables. These data from different database is merged to construct the whole panel data by tracing all variables needed to the country-year observations. The mnemonic and data source of the variable which will be used in the analysis is posed in Appendix 1 and the summary statistics of the variables in each observation group is posed in Appendix 2. According to Appendix 2, the average level of KAOPEN, TOTAL, legal development, financial development, government effectiveness and trade openness is much higher than in emerging countries. The graphs of all variables are posed in the Appendix 3, 4, 5, 6 and 7. According to the appendix one can draw the trend of the variables used in this research.
Current account: For most industrialized countries, their current accounts had
an increasing trend around the early 2000s, and then sharply decreased. For emerging countries, most of them had a rapid increase trend in their current accounts around the early 2000s, and then dropped sharply. It is not strange that countries don’t have completely same trends in variables: every country has its own independent policy system and economic environment.
KAOPEN: For industrialized countries, many of them stayed at a top high level
in KAOPEN in the whole period, this means these countries has already highly open for cross-border financial transactions in legal and political aspects with no space to increase the financial openness anymore. For other industrialized countries, they have a common increase main trend in financial openness. For the emerging countries, some of them stayed at a fixed level, some others had a fluctuation during the research period. For the remainder, they mainly had a common increase trend in KAOPEN, which means these emerging countries were keeping increasing their free degree towards cross-border financial transactions in legal and political aspects. According to all above, it can be found that becoming a country with high financial openness is the main trend of the world.
Total: For industrialized countries, almost all industrialized countries had an
increase trend in Total during the whole research period, which means the actual financial openness levels in these countries were keeping increasing. For emerging countries, most of them have an increase trend in Total, which means the actual financial openness in these emerging countries was keeping increasing.
Budget balance: The budget balance in both industrialized and emerging countries
stayed in fluctuation during the whole research period, but almost all country had a decrease in budget balance after the Global Financial Crisis. This decrease might be caused by after the crisis governments need to increase the expenditure to recover the domestic economic growth.
Trade openness: The trade openness in both industrialized countries and emerging
Global Financial Crisis. This shows globalization in trade is still one main trend of world economy.
Financial development: The financial development in both industrialized countries
and emerging countries has a main increasing trend. This shows the financial system in both type of countries were keeping developing.
Legal development and Government effectiveness: There is no common trend for
legal development and government effectiveness in either industrialized countries group or emerging countries group.
5. Empirical results
5.1 Regression model- financial openness measured by KAOPEN
To start with, this thesis runs the basic regression model with financial openness measured by KAOPEN, the result is posed in Table 1.
Table1. Basic regression model (Financial openness measured by KAOPEN)
Whole sample Industrialized countries Emerging countries
KAOPEN -0.7662299 (0.5668229) -2.887107 (4.886812) 0.1229021 (0.5814293) KAOPEN*FD -0.0104754 (0.0065082) -0.0219669* (0.0129653) 0.0389683*** (0.01275) KAOPEN*LEGAL 0.7168487 (0.7168343) -2.513838 (3.162776) -0.05120488 (0.8162769) KAOPEN*GE -0.1015578 (0.78727) 5.133725** (2.501588) -0.6845165 (0.9321291) FD -0.0123093 (0.0203433) 0.2135853** (0.0905916) -0.1336878*** (0.0296725) LEGAL -4.8802** (2.038964) 15.68075 (13.72877) -6.278846*** (1.992839) GE 1.977579 (2.378517) -18.90256* (10.24184) 3.255002 (2.21716) FD*LEGAL 0.0165659* (0.0090673) -0.0435329* (0.0225352) 0.0601596*** (0.0197288) BB 0.3078753*** (0.0774412) 0.301132*** (0.098753) 0.039511 (0.1061903) TO 0.0232129*** (0.0052833) 0.0169139** (0.0067356) 0.0348817*** (0.0089576)
21 NFA 0.0419365*** (.0051513) 0.0407036*** (.0069966) 0.0750651*** (.0114443) D2008 .8153912 (.654266) 2.946126*** (.9383287) -1.32493* (.7071601) _cons 4.857353*** (4.857353) -9.810061 (21.97283) 7.011613*** (1.721066)
Note: Standard errors in brackets. The symbol * means the coefficient is significant at the 10% level, ** for 5% level, *** for 1% level.
The regression result posed in table 1 shows that in the whole sample, KAOPEN has a negative coefficient, which means that a one unit increase in financial openness measured by the Chinn-Ito index will decrease the current account by 0.766 percentage point. However this effect is statistically insignificant. All the interactions related to KAOPEN don’t have a statistically significant coefficient for the whole sample. In total financial openness has no statistically significant on the current account. However as mentioned before the financial openness might affect industrialized and emerging countries differently. According to Table 1, all coefficients in industrialized country group and emerging country group are largely different with each other in both sign and size. This fact suggests that compared with simply doing research on the whole sample, which then will lead to a meaningless result, dividing observations into industrialized and emerging country groups and do specific research with each sample group is more appropriate, since due to their large differences in every aspect of economics and the financial market, the effect of financial openness on the current account should also be different a lot in industrialized and emerging countries. For the other variables in the whole sample, LEGAL has a significantly negative effect on the current account. BB, TO, NFA and the interaction of LEGAL and FD have significantly positive effects on the current account.
In industrialized countries, KAOPEN has a negative coefficient according to which a one unit increase in financial openness measured by Chinn-Ito index will decrease the current account by 2.88 percentage point. However this effect is not statistically significant. The interaction of KAOPEN and FD has significant negative coefficient at the 10% significance level, which means a one unit increase in financial openness measured by Chinn-Ito index will decrease the current account by 0.022*FD percentage point. This effect can be explained by the fact that higher financial development can attract more capital flows from other countries especially emerging countries when financial openness increases, and as a result the domestic investment rate increases, which leads to a decrease in the current account according to saving-investment approach. A higher financial development level can make this attractiveness more strong. The interaction of KAOPEN and GE has a significant positive coefficient at the 5% significance level, according to which a one unit increase in financial openness measured by Chinn-Ito index will increase the current account by 5.13*GE percentage point. This can be explained by the fact that the government is aiming at inducing domestic productivity and exports, the government
will formulate policies and run projects to reach this goal, increasing financial openness can provide more capital resources for these policies and projects, which means increase in financial openness can facilitating the productivity-inducing policies and projects runs by government, and therefore has a positive effect on the current account. In total a one unit increase in financial openness measured by Chinn-Ito index will increase the current account by 5.13*GE – 0.022*FD percentage point. For a country whose all variables at the average level of the industrialized country sample group, a one unit increase in financial openness measured by Chinn-Ito index will increase the current account by 11.85 percentage point. For the other variables in the industrialized countries sample, FD, BB, TO, NFA and D2008 have significantly positive effects on the current account. GE and the interaction of FD and LEGAL have significantly negative effects on the current account.
In emerging countries, KAOPEN has a positive coefficient, according to which a one unit increase in financial openness measured by Chinn-Ito index will increase the current account by 0.123 percentage point, however this effect is not significant. The interaction of KAOPEN and FD has a significant positive coefficient at the 1% significance level. According to which a one unit increase in financial openness will increase the current account by 0.039*FD percentage point. In total a one unit increase in KAOPEN will increase the current account by 0.039*FD percentage point. For a country whose all variables at the average level of the emerging country sample group, one unit increase in financial openness measured by Chinn-Ito index will increase the current account by 1.853 percentage point. This can be firstly explained by saving-investment approach, higher financial development level can let domestic people save less and invest more, and as a result the current account decrease because the saving account decrease and investment account increase. However, according to Bernanke (2005), higher financial openness in emerging countries will let these countries export more savings to industrialized countries, which means higher financial openness can let less saving transform to domestic investment, and therefore has an opposite effect on the association between financial development and the current account. Furthermore, according to what mentioned before, financial openness can lead to change in domestic industries, with higher financial openness emerging countries can attract more overseas investments with their higher marginal rate of return for investments. Although according to the saving investment approach the current account should be reduced directly, however these investments do have effects on the domestic industries, they help the expansion of domestic industries. As in emerging countries, industries are mainly export-oriented, the expansion of these industries can change the trade balance in the medium term. A higher development level for domestic financial market should also facilitating the effect of foreign investments on domestic industries since highly developed financial market can help these investments be allocated more efficiently. According to above it’s not surprising that in the medium term increase in financial openness can have a positive effect on the current account. For the other variables in the emerging countries sample, FD, LEGAL and D2008 have significantly negative effects on the current account. TO, NFA and the interaction of FD and LEGAL have significantly positive effects on the
23 current account.
One notable thing is that financial openness doesn’t show a fixed effect on the current account, but has an effect which is conditional on other variables according the result in table 1. This fact suggests that both how financial openness change the cross-border capital flows and how cross-border financial transactions change the domestic economy and industry is depending on the economic environment. As being just one aspect for the whole financial environment, it is not surprising financial openness has flexible effect on the current account which depends on the other aspects of the economic environment.
5.2 Regression model - financial openness measured by TOTAL
As announced before, financial openness can be measured by either a de jure or a de facto measurement. The Chinn-Ito index is a de jure indicator which focuses on the legislative and policy control on financial transactions by government. While de facto measurements focus on the actual but not legal or political free degree of financial transactions. According to Edwards (2005), the observed actual cross-border capital flows usually break the legal and political limitations put on them. Furthermore, in real life, general limitations like the acceptance of domestic markets towards cross-border financial businesses, the complexity of cross-border financial transactions and the efficiency of sub-processes when operating a cross-border financial business also affect the actual limitations on the cross-border financial transactions and therefore the actual financial openness. So in order to have a more clear view on the effect of financial openness on the current account, this thesis will also try measuring the financial openness by the TOTAL (the sum of foreign liabilities and assets), which is a de facto measurement for financial openness. The regression result of the basic regression model with financial openness measured by TOTAL is posed in Table 2.
Table 2. Basic regression model (Financial openness measured by TOTAL)
Whole sample Industrialized countries Emerging countries
TOTAL -0.0224982 ** (0.0102142) -0.0777143*** (0.0164276) 0.0611617*** (0.0194692) TOTAL*FD -0.0000829 *** (0.0000255) -0.0000603*** (0.0000277) -0.0003915 (0.0002462) TOTAL*LEGAL 0.012891 *** (0.0036288) 0.0212124*** (0.0050898) -0.0588425*** (0.0184033) TOTAL*GE -0.0037316 (0.0024385) 0.0055982* (0.0033677) 0.0618716*** (0.0251146) FD -0.0108671 (0.020151) 0.1821088** (0.0742443) -0.0738826** (0.033824)
LEGAL -5.546154 *** (1.215489) -0.7346977 (3.587729) -0.544195 (2.505821) GE 2.825964 ** (1.177704） 1.194773 (2.324615) -6.113128** (3.180599) FD*LEGAL 0.0141738 * (0.0077115) -0.0561444** (0.0242318) 0.0839384*** (0.0204717) BB 0.354339 *** (0.0727536) 0.3923925*** (0.0958852) -0.0399582 (0.1075536) TO 0.0311305 *** (0.0077717) 0.0094743 (0.0124594) 0.0246583** (0.010058) NFA 0.0373074 *** (0.0046895) 0.0308925*** (0.0048169) 0.0712102*** (0.0117347) D2008 0.9190399 (0.6378049) 2.963909*** (0.8836462) -1.683671** (0.7120827) _cons 5.299041 *** (1.41069) -1.234382 (8.133832) 2.513235 (2.320647)
Note: Standard errors in brackets. The symbol * means the coefficient is significant at the 10% level, ** for 5% level, *** for 1% level.
According to the regression result in table 2, in the whole sample, TOTAL has a negative coefficient at the 5% significance level, which means a one unit increase in financial openness measured by TOTAL will decrease the current account by 0.022 percentage point. The interaction of TOTAL and financial development has a significant negative coefficient at the 5% significance level, which means a one increase unit in TOTAL will decrease the current account by 0.000083*FD percentage point. The interaction of TOTAL and LEGAL has a significant positive coefficient, according to which a one unit increase in financial openness measured by TOTAL will increase the current account by 0.013*LEGAL percentage point. In total a one unit increase in TOTAL will decrease the current account by (0.022 + 0.000083*FD - 0.013*LEGAL) percentage point. For a country where all variables are at the average level in the whole sample, a one unit increase in KAOPEN will increase its current account by 0.0016 percentage point. For the other variables in the whole sample, LEGAL has a significant negative effect on the current account. GE, BB, TO, NFA and the interaction of FD and LEGAL have significant positive effects on the current account.
In industrialized countries, TOTAL has a significant negative coefficient at the 1% significance level, which means a one unit increase in TOTAL will decrease the current account by 0.0777 percentage point. The interaction of TOTAL and financial development has a significant negative coefficient at the 1% significance level, which means a one increase unit in TOTAL will decrease the current account by 0.00006*FD percentage point. The interaction of TOTAL and LEGAL has a significant positive coefficient at the 1% significance level, according to which a one unit increase in TOTAL will increase the current account by 0.021*LEGAL percentage point. The interaction of TOTAL and GE has a significant positive
25
coefficient at the 10% significance level, which means a one unit increase in financial openness will increase the current account by 0.0056*GE percentage point. In total, a one unit increase in TOTAL will increase the current account by (-0.00006*FD + 0.0056*GE + 0.0212*LEGAL - 0.0777) percentage point. In an industrialized country where all its variables is at the average level of the industrialized countries sample group, a one unit increase in TOTAL will decrease the current account by 0.00119 percentage point. For the other variables in the industrialized countries sample, FD, BB, NFA and D2008 have significant positive effects on the current account. The interaction of FD and LEGAL has a significant negative effect on the current account.
In emerging countries, TOTAL has a significant positive coefficient at the 1% significance level, according to which a one unit increase in TOTAL will increase the current account by 0.061 percentage point. The interaction of TOTAL and LEGAL has a significant negative coefficient at 1% level, according to which a one unit increase in TOTAL will decrease the current account by 0.059*LEGAL percentage point. The interaction of TOTAL and GE has a significant positive coefficient at the 1% significance level, which means a one unit increase in financial openness will increase the current account by 0.062*GE percentage point. In total, a one unit increase in TOTAL will increase the current account by (0.061 - 0.059*LEGAL + 0.062*GE) percentage point. In an emerging country where all its variables is at the average level of the emerging countries sample group, a one unit increase in TOTAL will increase the current account by 0.0465 percentage point. For the other variables in the emerging countries sample, FD, GE and D2008 have significant negative effects on the current account. TO, NFA and the interaction of FD and LEGAL have significant positive effects on the current account.
The results of the regressions with financial openness measured by different measurements are different. This is caused by the fact that these measurements focus on representing different aspects of limitation on cross-border financial transactions. As a de jure measurement, KAOPEN focuses on the legal and political limitations on cross-border financial transactions, while TOTAL measures it in an actual aspect. One main difference is that in Table 2 the legal development shows an effect on the association between financial openness and the current account. KAOPEN provides good information on legal and political restrictions on cross-border capital flows, so it can capture the change in cross-border capital flows in response to the change in legal and political financial openness, however as Edwards (2005) suggests there are also large amount cross-border capital flows which are out of the legal and political restrictions, KAOPEN cannot capture the change of these capital flows when the financial openness changed. TOTAL can capture the changes of the cross-border capital flows which are out of the legal and political restrictions, and legal development by definition does have an effect on these capital flows, so when financial openness measured by TOTAL, legal development shows an effect on the association between financial openness and current account. One thing clear is that when be measured by different measurements, financial openness shows different effects on the current account, and different aspects of financial openness affect the current account differently. This result goes against with the conclusion of Saadaoui
(2015), the author in his research found that either measured by KAOPEN or TOTAL increase in financial openness shall increase the current account in industrialized countries while decrease it in emerging countries, and the effect is fixed, combine this result with the fact that most part of the current account deficits are contributed by industrialized countries and most part of the current account surpluses are contributed by emerging countries, the author finally concluded further liberalization on capital account can help reduce the global imbalances. However Saadaoui (2015)’s research took sample from early periods, and in the author’s regression model the effect of LEGAL, FD and GE are not taken into account, so definitely this thesis’s result goes against with Saadaoui (2015).
6. Conclusion and Discussion
Current account imbalances have already been a serious problem towards several countries for a long time, and financial openness is considered as one determinant of the current account. This thesis collects data from different databases for the period 1997-2011 and uses an OLS regression model to investigate the effect of financial openness on the current account imbalances in the medium term. Considering the economic and financial fundamentals may differ a lot in industrialized and emerging countries, this thesis divided the sample countries into an industrialized country group and an emerging country group to investigate this causal effect separately in each group. What is added to the existing literature is that this thesis uses the most recent accessible sample period which includes the years after the Global Financial Crisis (GFC), controlling for the possible effect of government effectiveness and the GFC on the current account. Furthermore it also takes account of the effect of government effectiveness on the association between financial openness and the current account.
Besides measuring the financial openness by the Chinn-Ito index, which is a de jure measurement for the financial openness, this thesis also tried measuring the financial openness by the sum of foreign assets and foreign liabilities, which is a de facto measurement. This thesis finds that when measured by different measurements, financial openness shows different effects on the current account, this difference is mainly caused by the fact that the de jure measurement can only represent the legal and political restrictions on the cross-border financial transactions.
When financial openness is measured by the Chinn-Ito index, the research result shows that in the whole sample legal and political financial openness doesn’t have a significant effect on the current account. In industrialized countries, legal and political financial openness does have a significant effect on the current account, of which the sign and size depend on the government effectiveness and financial development level. Considering the distribution of these two variables, an increase in financial openness shall normally have a positive effect on the current account. This suggests that in the industrialized countries suffering from current account deficits like the US and the UK, an increase in legal and political financial openness can help these countries balance their current account. However as noticed by Saadaoui (2015), the financial
