University of Groningen
Faculty of Economics and Business
The influence of the Eurasian Customs Union on
trade of Belarus, Kazakhstan and Russia – Gravity Model for
Trade
M.Sc. International Economics and Business
2
Abstract
The objective of this thesis is to better understand and analyze the influence of the Eurasian Customs Union on bilateral trade between Belarus, Kazakhstan, Russia and other countries. The research question is the following: “How did a membership of the Eurasian Customs Union influence the trade between its member states – Belarus, Kazakhstan and Russia, and the trade of the members with third countries?’.
The empirical analysis is conducted using a gravity model approach and the database of bilateral trade monthly flows between the three countries of the Eurasian Customs Union and 40 other countries over the period of time from January 2000 to December 2015.
The findings indicate that such variables as GDPs, common language, common border and common history are highly significant and positively influence trade flows. The distance between the countries is significant and negatively related to the trade flows. The results show that the Eurasian Customs Union positively affects the volume of bilateral trade between Belarus, Kazakhstan and Russia and it negatively affects the volume of bilateral trade between the Eurasian Customs Union and third countries.
3
Contents
Abstract ... 2
Introduction ... 4
1. Background and literature review ... 9
1.1. The Eurasian Customs Union ... 9
1.2. The Relationship Between Customs Union and Trade ... 10
2. Methodology and Data ... 14
2.1. Gravity Model for Trade ... 14
4
Introduction
After the collapse of the Soviet Union, the newly established sovereign states have experienced a rapid liberalization of economies and implemented less protectionism and trade barriers. They had to adapt to the new world order – free market and international trade. The current international environment is globalized, where countries show growing interdependence and the border between international and national law is close to disappearing. However, at the same time, countries tend to show the behavior that leads to the regionalization. Consequently, rising regionalization leads to the formation of free trade agreements, customs unions and regional economic and political blocks. The Eurasian Customs Union and later the Eurasian Economic Union are recently created examples of such blocks.
5
Chart 1: Export destination countries of Belarus in 2009 (% of total exports in all products)
Source: The Observatory of Economic Complexity, Trade destinations: http://atlas.media.mit.edu/en/profile/country/blr/#Exports
Chart 2: Export destination countries of Kazakhstan in 2009 (% of total exports in all products)
6
Chart 3: Export destination countries of Russia in 2009 (% of total exports in all products)
Source: The Observatory of Economic Complexity, Trade destinations: http://atlas.media.mit.edu/en/profile/country/rus/
We can observe that shortly before establishing the Eurasian Customs Union not all of the member states were the main export destinations (and trade partners) of each other. Isakova et al (2015) argue that the main objective of the Eurasian Customs Union is to enhance economic integration between the countries, because the Eurasian Customs Union would bring potential for trade creation and diversion. However, as we can observe in the chart 2 and the chart 3, Kazakhstan and Russia do not have big shares of export with the other members of the Eurasian Customs Union. Kazakhstan exports only 9,1% to other members of the Eurasian Customs Union and Russia exports only 7%. When a customs union is created, there is an increase of trade between union members expected. Moreover, a customs union is expected to reduce trade with non-member countries. However, there is a possibility that trade within the union members will not necessarily increase as fast as their trade with the third countries will decrease. Furthermore, countries would tariff revenue due to reduction of trade with the third countries.
Some authors provide various reasons for the creation of the Eurasian Customs Union apart from the trade facilitation. An important reason for willingness of trade integration between Belarus, Kazakhstan and Russia would be an opportunity to enhance their own welfare and to reduce the profit of other countries, which are not the members of the Eurasian Customs Union (Isakova et al, 2015). This reason became especially
7
popular after 2014, when Russia faced the sanctions imposed by the European Union (European Union Newsroom).
Despite the economic magnitude of the Eurasian Customs Union, it is not a well-researched topic. Therefore, we indicate the need to research the changes and the effect of changing the trade flows between those countries after joining the Eurasian Customs Union.
The following research question will help to see if the economic integration through increasing the trade flows between three countries occur: ‘How did a membership of the Eurasian Customs Union influence trade between its member states – Belarus, Kazakhstan and Russia, and trade of the members with third countries?’
To answer the aforementioned research question, we employ a panel data analysis covering 43 countries over the period from January 2000 to December 2015. The focus of this study is on Belarus’, Kazakhstan’s and Russia’s bilateral trade and their trade with the third countries, which includes Armenia, Azerbaijan, China, Georgia, Kyrgyz Republic, Mongolia, Norway, Tajikistan, Turkmenistan, Ukraine, Uzbekistan, the US and 28 EU countries. These countries were chosen for various reasons. Firstly, Armenia, Azerbaijan, Georgia, Kyrgyz Republic, Tajikistan, Turkmenistan, Ukraine and Uzbekistan were chosen because of the common historical ties as member states of the Soviet Union. The European Union member states were chosen due to the noteworthy sharing of trade with Belarus, Kazakhstan and Russia. Moreover, some of the European Union countries have common borders and historical relations with the Eurasian Customs Union countries. China, Mongolia and Norway share common borders with at least one of the Eurasian Customs Union’s members. The US were chosen as one of the main trade partners with Russia (see Chart 3).
8
et al (2007), Kohl (2014) and additional independent variables, such as, common language, the common border and the common history due to their impact on trade flows.
9
1. Background and literature review
1.1. The Eurasian Customs Union
After the collapse of the Soviet Union, the post-Soviet countries have turned to the market economies and have liberalized their trade with other countries. Those post-Soviet countries had to reorganize their production and to adapt to free trade international trade. In 1991, the Commonwealth of Independent States Free Trade Zone Agreement was the first attempt to create trade cooperation based on the free trade agreement. It was initially planned to include Armenia, Belarus, Georgia, Moldova, Kazakhstan, the Kyrgyz Republic, the Russian Federation, Tajikistan, Ukraine, and Uzbekistan. However, after the long-lasting and unproductive political discussions, the Free Trade Zone never came into force, because Russia did not want to implement tariff-free trade of oil and gas (Tochitskaya, 2010). Ultimately, the idea was abandoned.
The next attempt took place in 1995. Belarus, Kazakhstan and Russia have attempted to create a customs union. This time it also had a support from other Eurasian countries. Kyrgyz Republic and Tajikistan have agreed to join it in 1996. It could have potentially become a successful case of cooperation between the Commonwealth of the Independent States. As the result, the customs union was reorganized into the Eurasian Economic Community in 2001. However, the newly established economic community have faced different willingness to protect or open trade to the international competition, due to the sufficient divergence in the structure of economies and different levels of economic development. The idea was eventually given up due to those differences and disagreements (Tochitskaya, 2010).
10
had already harmonized its’ own pre-union external tariffs to Russia’s. Shumylo-Tapiola (2012) sums up that Kazakhstan have experienced a stronger negative impact of the tariff, because Kazakhstan nearly had to double its own external tariff.
In January 2012, the Customs Union of Belarus, Kazakhstan, and Russia have removed the barriers of the capital and labour movement between the three countries (Dragneva & Wolczuk, 2014). In 2016, Belarus, Russia, Kazakhstan, Armenia and Kyrgyz Republic have joined the Eurasian Economic Community.
The further rapid integration of Belarus, Kazakhstan and Russia could be a result of the successful cooperation and the trade development through the Eurasian Customs Union. It is important to indicate that the impact of the Eurasian Customs Union on the trade between the countries. It is especially the case if we take into account that the three countries were not the main trading partners in 2009.
1.2. The Relationship Between Customs Union and Trade
11
The theory on the customs union was published in Viner’s (1950) pioneering work. Viner (1950) has shown that a customs union can result in a trade creation or a trade diversion. Since Viner’s work was published, these two terms have been widely used. Viner (1950) has argued that the trade creation is good and the trade diversion is bad if we look at it in terms of free trade. Moreover, he has assumed that any pre-union tariff is harmful for the trade. Later authors (Lipsey, 1957; Corden, 1972; Collier, 1979) have provided explanations regarding the trade creation and trade diversion, as well as have performed a deeper analysis of their advantages and disadvantages.
Trade creation occurs through the removal of internal tariffs within the customs union. When the trade barriers are lowered, it allows buying cheaper products from more efficient foreign producers. The trade creation involves a shift from high-cost domestic production to a lower-cost production in a partner country. Moreover, customs union legislations eliminate the administrative barriers and expand market access. This allows importing from the partners more easily (Viner, 1950). Corden (1972) has provided an example with two countries – country A and country B. The countries establish the customs union. Country’s B consumers start to choose cheaper imports from country A and the domestic high-cost production of country B decreases. This definitely enhances the trade between the countries (Corden, 1972). In this situation, the trade between country A and country B increases, and it also decreases between the members and the third non-member country C (Lipsey, 1957).
12
replaced by the same goods imported from a partner country, duty-free but at a higher real cost.
The effect that harms the trade is trade destruction. The reduction of the internal tariffs may lead to the decreased trade overall, due to the decreasing trade between the member states and the decreasing trade between the other countries. The trade destruction appears when some products cannot be supplied by the regional trade partners within the union (Cooper, 1965). For example, the aforementioned country A cannot provide the necessary products to country B, but these products from country C become relatively expensive for country B. Anderson (1979) argues that the newly created non-tariff trade with the members of the trade union influences the imports from the third parties to the member states. Moreover, the member states of the customs union introduce a common external tariff, which is an average of the previous tariffs. At this stage, some countries lower the tariffs, which may enhance their trade. However, some countries rise their tariffs, which increases the price of a product for domestic consumers and prevents the import of goods from the third countries. Therefore, higher tariffs lead to the reduction of the trade flows. As the result, the trade with the third countries is limited and the trade within does not increase (Isakova et al, 2015). This situation might arise after Kazakhstan have joined the Eurasian Customs Union; as mentioned before, Kazakhstan had to increase the external tariffs, which may affect Kazakhstan’s trade with the third countries after January 2010, especially considering that the third countries were the main trade partners of Kazakhstan before 2010 (see Chart 2).
As already discussed before, the protectionism activities used in the customs union may both enhance and harm trade flows between the members. However, Krueger (1997) indicates the advancement of customs union over the free trade agreement. Krueger (1997) indicates that both, the free trade agreement and the customs union, emphasize the trade creation due to the reduction of tariffs between the member states. Apart from that, countries that have the free trade agreement tend to keep their own trade policy to the third parties and do not diverse trade (Krueger, 1997).
13
the trade creation and trade diversion channels the trade between the members of the Eurasian Customs Union might increase, and trade with the third countries might decrease. On another hand, if the trade destruction takes place, the trade between the members of the Eurasian Customs Union might not increase, and it might decrease with the third countries.
We propose two hypotheses to find the answer to the research question ‘How did a membership of the Eurasian Customs Union influence trade between its member states –Belarus, Kazakhstan and Russia, and trade of the members with third countries?’.
The Hypothesis 1 reflects on the impact of the Eurasian Customs Union on the bilateral trade between its members: The Eurasian Customs Union positively affects the volume of bilateral trade between Belarus, Russia and Kazakhstan.
14
2. Methodology and Data
2.1. Gravity Model for Trade
The gravity model is useful for defining the relation between the implementation of customs union and bilateral trade flows. The gravity equation has emerged as the empirical way of measuring the effect of FTAs and customs union on the bilateral trade. Anderson and van Wincoop (2003) note that the gravity equation is the most successful empirical trade analysis used in past 40 years.
The main idea of the Gravity model was distinguished from Newton and his theory of the “Law of Universal Gravitation”. It proclaims that the force between two objects “derive” from two masses divided by the square of distance between the objects. In 1962, Jan Tinbergen have suggested to use this model to describe the international trade flows that are determined by a size of two economies and distance between them. The main difference from Newton’s law is that trade is inversely proportionate to the distance, whereas the gravitational force is inversely proportionate to the distance squared.
The general gravity law can be expressed as following:
𝐹𝑖𝑗 = 𝐺𝑀𝑖𝑀𝑗 𝐷𝑖𝑗
Where 𝐹𝑖𝑗 refers to the flow between country i and j. 𝑀𝑖 and 𝑀𝑗 are the mass of economies of country i and country j. 𝐷𝑖𝑗 describes the distance between country i and country j. 𝐺 refers to a gravity constant. The trade flows between countries is a dependent variable. Tinbergen (1962) found that countries’ incomes were positively related to the bilateral trade flows, but the distance between the countries was negatively related to the bilateral trade.
15
However, economists assume that only two variables of the gravity equation, GDP and distance, are not sufficient for defining the trade flows. Ghosh and Yamarik (2004) admit that “there is a lack of consensus in the gravity model literature on which other variables should be included in the extended gravity model equation”. Even distance might not be only a geographical measure expressed in kilometers. It also might be a language distance or a cultural distance. Cheng and Wall (2005) have involved four important variables that influence trade. The first one is the common language. The second determinant is the historical ties. Being a part of one country or a block, matters, because those countries may still have closer ties. The third determinant is a common border. The gravity model often includes more than one distance measure that affects the bilateral trade between the countries. Anderson and van Wincoop (2003) admit that there is a need to control for variables that influence the trade costs. These variables might increase or decrease the trade flows between the countries, because they influence the trade costs.
The four variables described previously were used in various studies on the international trade flows. For example, Suvankulov and Guc (2012) have used the standard explanatory variables such as the population, the GDP per capita and the distance. Moreover, they have used some additional variables, such as the common border, the language, the legal system, the single currency and the WTO membership. They have found that all of the coefficients were positive, except the one that refers to the distance.
16
impact on the trade at all, while only about one quarter of the agreements are actually promoting the trade.
To sum up the aforementioned determinants, the following widely used variables were chosen for our log-linear gravity equation – GDP, distance, membership dummy variable, common border, common language and common history indicator.
The log-linear gravity equation used in the following research is:
ln(𝑇𝑟𝑎𝑑𝑒𝑖𝑗𝑡) = 𝛽0+ 𝛽1ln(𝐺𝐷𝑃1𝑖𝑡) + 𝛽2ln(𝐺𝐷𝑃2𝑗𝑡) + 𝛽3𝑙𝑛(𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒𝑖𝑗) + 𝛽4𝐶𝑈𝑖𝑗𝑡 + 𝛽5𝐶𝐵𝑖𝑗𝑡 + 𝛽6𝐶𝐿𝑖𝑗𝑡 + 𝛽7𝑈𝑆𝑆𝑅𝑖𝑗𝑡+ 𝛽8𝑊𝑖𝑛𝑡𝑒𝑟𝑖𝑗𝑡 + 𝛽9𝐹𝑎𝑙𝑙 𝑖𝑗𝑡
+ 𝛽10𝑆𝑢𝑚𝑚𝑒𝑟𝑖𝑗𝑡+ 𝛽11𝐷𝑢𝑚𝑚𝑦1𝑖𝑗𝑡 + 𝜀𝑖𝑗𝑡
17
season, therefor we introduce these variables to find the effect of the season on trade. 𝑊𝑖𝑛𝑡𝑒𝑟 dummy variable value is 1 for winter months – December, January and February, and 0 for the rest of the year. 𝐹𝑎𝑙𝑙dummy variable value is 1 for fall months – September, October and November, and 0 for the rest of the year. 𝑆𝑢𝑚𝑚𝑒𝑟 dummy variables value is 1 for June, July and August, and 0 for the rest of the year. We introduce additional variable to test the hypotheses. 𝐷𝑢𝑚𝑚𝑦1 is a binary variable that refers to country pairs of Belarus, Kazakhstan and Russia. The value of the variable is 1 for pairs Belarus – Kazakhstan, Kazakhstan – Belarus, Belarus – Russia, Russia – Belarus, Kazakhstan – Russia, Russia – Kazakhstan. The value of he variable is 0 for country pairs that contain one of a non-member country, for example, Belarus – Belgium , Kazakhstan – China or Russia – United Kingdom. The main difference of 𝐷𝑢𝑚𝑚𝑦1𝑖𝑗𝑡 and 𝐶𝑈𝑖𝑗𝑡 that 𝐷𝑢𝑚𝑚𝑦1𝑖𝑗𝑡 value for countries does not differ by years. The value of 𝐷𝑢𝑚𝑚𝑦1𝑖𝑗𝑡 is always 1 for country pairs Belarus – Kazakhstan, Kazakhstan – Belarus, Belarus – Russia, Russia – Belarus, Kazakhstan – Russia, Russia – Kazakhstan. The value of 𝐶𝑈𝑖𝑗𝑡 is 1 for these pairs only after January 2010.
2.2. Data Selection
The panel dataset consists of observations for 43 countries, 3 of which are the members of the Eurasian Customs Union – Belarus, Kazakhstan and Russia, the rest 40 are used as a control group. The data are collected for the period from January 2000 till December 2015. The data are arranged as a panel dataset by country-pair, year and month. It consists of 24 192 observations. The number of the bilateral trade flows is 126, because between Russia-Russia or Belarus-Belarus, and Kazakhstan-Kazakhstan are not included, therefore, the total number of the bilateral trade flows is 126 (3*43-3=126).
18
price. The main limitation is that the GDP deflator reflects the prices of all domestically produced goods, including the tax revenue and services, and net export. We have used the GDP deflators of both countries – to find the real import we have used the GDP deflators of Belarus, Kazakhstan and Russia, and to find the real export we use the GDP deflators of all 43 countries. For each country we have used its own GDP deflator. However, we have to take into account that for some countries, for example, Russia, the GDP deflator may capture the oil prices. Nevertheless, the GDP deflator is more reliable that the Consumer Price Index (CPI), because the CPI is based on the basket of good and captures only products and services from the consumer’s basket. The GDP, however, reflects the changes of consumption patterns.
The GDP in constant 2010 USD is taken from the World Bank’s (2016) World Development Indicators. The constant series measure the real growth of series and are adjusted to the effects of inflation (World Bank)
The bilateral distance data, common border and language are taken from the CEPII GeoDist Database. This database was developed by Mayer and Zignago (2005). The bilateral distance is the weighted distance between the biggest cities with respect to the geographic distribution of a population. (Mayer and Zignano, 2005).
The advantage of the monthly data is that it provides more observations than the annual data. Especially considering that the Eurasian Customs Union came into force only in January 2010. The disadvantage of the monthly data is that it might be not normally distributed due to the seasonal changes in the trade; for example, if countries import or export different products based on their seasonal demand. In order to provide a better analysis we have introduced seasonal dummy variables that indicate seasonal impact on the trade.
2.3. Econometric Issues
19
Table 1 shows the descriptive statistics of the variables used in the analysis. The four variables are lnTrade, lnGDP1, lnGDP2 and lnDistance - the variables which originally were expressed in millions of USD and kilometers, are used in natural logarithms due to the skewness of data, because it might affect the mean and the analysis overall. For example, the GDP of USA is higher than the GDP of Cyprus. The highest value of lnGDP1 is Russia’s natural logarithm of GDP in 2014 – 28.15. The lowest value of lnGDP1 is Belarus’ natural logarithm of GDP in 2000 – 24.02. The highest value of lnGDP2 is the US natural logarithm of GDP in 2015 – 30.44. The lowest value of lnGDP2 is Tajikistan’s natural logarithm of GDP in 2000 – 21.67. The binary variables are CU, CB, CL, USSR, Dummy1, Winter, Fall and Summer. The mean values for the CU, CB, CL, USSR and Dummy1 differ, because there are some variations of 0 and 1 values between the country-pairs. There are more countries with a common border and historical membership in the USSR, therefore, the means of the CB and USSR are bigger than for the CL.
Table 1. Summary statistics
Variable Obs Mean Std. Dev. Min Max lnTrade 24,192 3.48 2.47 -6.88 9.09 lnGDP1 24,192 26.00 1.46 24.02 28.15 lnGDP2 24,192 25.63 2.02 21.67 30.44 lnDistance 24,192 7.80 0.60 5.86 9.23 CU 24,192 0.02 0.13 0 1 CB 24,192 0.18 0.39 0 1 CL 24,192 0.03 0.17 0 1 USSR 24,192 0.31 0.46 0 1 Dummy1 24,192 0.05 0.21 0 1 Winter 24,192 0.25 0.43 0 1 Fall 24,192 0.25 0.43 0 1 Summer 24,192 0.25 0.43 0 1
Source: own results
20
the common border. We can observe such low values because not all the countries of the Eurasian Customs Union share a common border or a common language. For example, Belarus and Kazakhstan do not share common border, but they share a common language - Russian. We can observe a negative correlation between the distance and the dummy variables (such as the CB, CL, CU and the USSR). Therefore, we can conclude that if a distance between countries increases, the possibility that those countries share a common border, common language or a common history decreases. A high correlation is observed between the Dummy1 and CU variables, because they refer to the same country-pairs; therefore, these variables will not dropped.
Table 2. Correlation table
Source: own results
We have also checked the data for the outliers. The Graph 1 shows the leverage against the (normalized) residuals squared. The leverage indicates how far is the independent variable value from the other observations, and large residual indicates the difference between the predicted and the observed value (Williams, 2016). Judging from the graph, we can see that a group of observations has a higher-than-average leverage. The high value of leverage in this observation group is due to high value of natural logarithm of Russia’s GDP. We do not observe other critical points that would appear in the right upper corner, therefore we do not drop the group of observations that have high leverage, because they are important for the study. Regarding residuals, the following points 110 indicate pair Belarus-Mongolia and the point 102 indicates country-pair Belarus-Ireland.
lntrade lnGDP1 lnGDP2 lnDistance CU Dummy1 CB CL USSR Winter Summer Fall
21 Graph 1. Test for outliers.
Source: own results
The Breusch-Pagan test for heteroscedasticity (see Figure 1) reflects that the p-value is less than 0.05, therefore, we reject the null hypothesis. The heteroscedasticity is present. To prevent biased standard errors, we need to apply robust standard errors for the following regressions.
Figure 1. Breusch-Pagan / Cook-Weisberg test for heteroscedasticity Ho: Constant variance
Variables: fitted values of lnTrade chi2(1) = 763.00
22
3.
Empirical Analysis
Table 3 shows the empirical results of gravity equation estimation for all pairs of the countries. The first column shows OLS coefficients for the basic variables of the Gravity model that are lnGDP1 (GDP of one of 3 countries) and lnGDP2 (the GDP of one of 43 countries), and lnDistance - distance between them. As we can see, the coefficients of the GDPs are positively associated with the trade flows, and the distance is negatively associated. All of the coefficients are highly significant. The second column reflects the model with the added CU dummy. It shows highly significant coefficients with the positive impact of the Eurasian Customs Union coefficient on the trade flows between countries. The column 3 reflects the same OLS analysis with the added common language dummy, common border and the USSR dummy. We can observe that all of the dummy variables have a significant and positive effect. When the common border, language and history are taken into account, the distance coefficient becomes smaller than before. It shows that the distance matters less if the countries share a common language, common border and history. Moreover, the coefficient of the CU dummy became smaller, therefore, it also matters less. From all of the dummy variables, a common history variable has the highest coefficient value that shows its impact on the trade flows and indicates that the trade flows between the post-Soviet countries that are not the members of the Eurasian Customs Union are bigger than between the countries that just share a common border.
Table 3. Results of Estimating Gravity Equation for all countries
23 CL 0.277*** (0.045) USSR 1.486*** (0.022) CB 0.491*** (0.024) Constant -33.97*** -33.84*** -39.81*** (0.195) (0.192) (0.189) Observations 24,192 24,192 24,192 R-squared 0.707 0.717 0.784
Robust standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
Table 4 shows the empirical results of gravity equation estimation for each of the Eurasian Customs Union’s country. As we can observe from the table the Eurasian Customs Union has a significant and positive influence on trade of Russia and Belarus. However, the Eurasian Customs Union coefficient appears as insignificant and negative in the case of Kazakhstan, which means that the Eurasian Customs Union does not have significant influence on trade of Kazakhstan. The Eurasian Customs Union has more influence on trade of Belarus.
Table 4. Results of Estimating Gravity Equation for each of the Eurasian Customs Union’s country
(Kazakhstan) (Russia) (Belarus) VARIABLES lnTrade lnTrade lnTrade lnGDP1 2.340*** 2.559*** 1.821*** (0.0411) (0.0566) (0.0584) lnGDP2 1.040*** 0.653*** 0.643*** (0.00836) (0.00603) (0.00820) lnDistance -3.069*** -1.687*** -1.154*** (0.0370) (0.0348) (0.0248) CU -0.0596 1.272*** 2.827*** (0.0989) (0.0815) (0.126) Constant -58.78*** -69.48*** -50.36*** (1.071) (1.594) (1.444) Observations 8,064 8,064 8,064 R-squared 0.719 0.657 0.555
24
Further, we have introduced the variable Dummy1, which indicates the country pairs of all three members of the Eurasian Customs Union. The Dummy1 value for the Eurasian Customs Union members is 1, and 0 otherwise. We have applied the new dummy variable to test the Hypothesis 1 and Hypothesis 2. We have run the regression analysis for the time period before January 2010, and the regression for the observations later than January 2010, when the Eurasian Customs Union came into force. This has allowed us to observe the influence of the Eurasian Customs Union in the gravity equation on the trade between three countries.
Table 5 shows the empirical results of estimating the gravity equation with the Dummy1. The Dummy1 coefficient is highly significant and positive before and after January 2010. There is a difference between Dummy1 coefficient’s value before January 2010 and after January 2010. After January 2010, the Dummy1 coefficient value becomes lower. It reflects that the trade flows between Belarus, Kazakhstan and Russia are lower after establishing the Eurasian Customs Union. Yet, it also indicates that the trade flows between the Eurasian Customs Union member states – Belarus, Kazakhstan, Russia, and the third countries were bigger after January 2010; because if we introduce another dummy to indicate the non-members of the Eurasian Customs Union, the sign would be negative, but coefficients would stay the same. The trade flows between the members of the Eurasian Customs Union and the third countries have increased after January 2010.
Table 5. Results of Estimating Gravity Equation with Dummy1
(OLS before January 2010)
(OLS after January 2010) VARIABLES lnTrade lnTrade
25
Observations 11,340 12,852
R-squared 0.746 0.722
Robust standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
In order to check for the seasonal impact on the trade, we have introduced the Winter, Fall and Summer dummies, respectively to the three seasons of the year. As we can see from the Table 6, the Winter and Fall variables have a significant impact on the trade flows between the countries. The Winter season has a negative impact on the trade flows, and the Fall season has a positive impact. Therefore, we can conclude that the Winter and Fall seasons have an impact on the trade flows. We will introduce the Winter and Fall variables in our later analysis.
Table 6. Results of Estimating Gravity Equation with seasonal dummies
(OLS) (OLS) VARIABLES lnTrade lnTrade
26
(0.196) (0.189)
Observations 24,192 24,192
R-squared 0.707 0.784
Robust standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
In Table 7, we can observe the results of estimating the gravity equation with the seasonal dummies and the Dummy1 before and after January 2010. We have included all of the variables: lnGDP1, lnGDP2, lnDistance, CB, CL, USSR, Fall, Winter and Dummy1. We can observe that the Dummy1 coefficient value becomes smaller after January 2010. This, once again, demonstrates that after January 2010, the impact on the trade was lower than before January 2010, and the introduction of the Eurasian Customs Union has lowered the trade between the member states. However, for the third countries, the impact on the trade is the opposite, and the trade between the Eurasian Customs Union countries and the third countries increases.
Table 7. Results of Estimating Gravity Equation with seasonal dummies and Dummy1
(OLS before January 2010)
(OLS after January 2010) VARIABLES lnTrade lnTrade
27 Fall 0.050 -0.171*** (0.034) (0.024) Constant -38.80*** -39.07*** (0.267) (0.263) Observations 11,340 12,852 R-squared 0.798 0.776
Robust standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1
28
Table 8. Results of Estimating Gravity Equation with Fixed Effects
(OLS) (FE)
VARIABLES lnTrade lnTrade
lnGDP1 1.032*** 2.491*** (0.007) (0.024) lnGDP2 0.583*** 0.440*** (0.005 (0.031) CU 2.204*** -0.377*** (0.074) (0.042) Constant -38.34*** -73.56*** (0.215) (0.555) Observations 24,192 24,192 R-squared 0.620 0.539 Number of group 126
29
4.
Conclusion
The main objective of this thesis was to identify the impact the Eurasian Customs Union has on the bilateral trade between the Eurasian Customs Union members – Belarus, Kazakhstan and Russia, as well as 40 other countries. Based on the gravity model of the trade, we have employed a panel data analysis which included such variables as the trade flows, GDPs that reflected masses of economies, distance between countries, common border, common language and the common post-Soviet historical tights. We have also introduced the seasonal variables in order to check whether the seasons have an impact on the trade between the countries.
The most important variable is the one that indicates a membership in the Eurasian Customs Union. In fact, we find that a membership of the Eurasian Customs Union has highly significant positive impact on trade of 3 member states. However, if we look more precisely, the Eurasian Customs Union has only significant and positive impact on trade of Belarus and Russia. The analysis did not show any positive and significant influence of the Eurasian Customs Union on of trade of Kazakhstan. Other variables that were expected to show a possible influence, i.e. masses of economies, a common language, a common border and a common history appear to be positively related to the trade between the countries. The distance between the countries has a negative impact on the trade flows. Moreover, another variable that indicates a common history showed bigger impact on trade flows than a membership in the Eurasian Customs Union.
30
indicates that the trade between the Eurasian Customs Union and the third countries experienced smaller negative impact after January 2010.
As it was already observed in the introduction, the main trading partners of Belarus, Kazakhstan and Russia are the third countries that are not the members of the Eurasian Customs Union. From the regression results of the period from January 2000 until December 2015, we have estimated that the trade within the Eurasian Customs Union was positively influenced and the trade with the third countries was negatively influenced, therefore it shows changing patterns of trade. Trade flows between partners are influenced positively, but trade with third countries shrink. The results can be referred to both trade creation and trade diversion described by Viner (1950), Lipsey (1957), Corden (1972) and Collier (1979). In this both cases trade flows between members of the one customs union increase, and trade with third countries decreases.
Regarding the additional seasonal variable, the results have shown that the trade in the winter months is smaller and is higher in the autumn months. The shrinking trade flows in the winter could be explained by the technical problems of transportation. The higher trade in the autumn months, then, might be due to the fact that it is the harvest period.
We can observe the surprising results if unobserved fixed effects are taken into account. There is a clear that the membership in the Eurasian Customs Union has a negative impact on trade between countries if unobserved fixed effects are taken into account.
31
Bibliography
ABDULLIN A., GALIAKBEROV A. (2014): “Theory and practice of regional integration based on the EurAsEC model (Russian point of view), Journal of Eurasian Studies, pp 116-121
AITKEN N.D. (1973): “The Effect of the EEC abd EFTA on European Trade: A Temporal Cross-Section Analysis”, The American Economic Review, Vol 63, No. 5, pp 881-892
ANDERSON J.E. (1979): “The Theoretical Foundation for the Gravity Equation”, The American Economic Review, Volume 69, No. 1, pp 106-116
ANDERSON J.E., van WINCOOP E. (2003): “Gravity with gravitas: a solution to the border puzzle”, American Economic Review, Volume 93, No. 1, pp 170-192
BAIER S.L., BERGSTRAND J.H. (2007): “Do free trade agreements actually increase members’ international trade?”, Journal of International Economics Volume 71, pp 72-95
BAIER S.L., BERGSTRAND J.H., VIDAL E. (2007): “Free Trade Agreements In the Americas: Are the Trade Effects Larger than Anticipated?”, The World Economy, Volume 30, Issue 9, pp 1347-1377
BALDWIN R.E. (1994): “Towards an integrated Europe”. CEPR, London
BERGSTRAND J.H. (1985): “The Gravity Equation in International Trade: Some Microeconomic Foundations and Empirical Evidence”, The Review of Economics and Statistics, Vol 67, No. 3, pp 474-481
BUSSIERE M., FIDRMUC J., SCHNATZ B. (2005): “Trade Integration of Central and Eastern European Countries: Lesson from a Gravity Model”, European Central Bank, Working Paper Series, No. 545
32
CHENG I.H., WALL H.J. (2005): “Controlling for heterogeneity in gravity models of trade and integration”, Federal Reserve Bank of St. Louis Review, Volume 87, Number 1, 49-63
COLLIER P. (1979): “The Welfare Effects of Customs Union: An Anatomy”, The Economic Journal, Volume 89, No. 353, pp 84-95
COOPER, A., MASSELL, B.F. (1965): “A New Look at Customs Union Theory”, The Economic Journal, Volume 75, No. 300, pp 742-747
CORDEN, W. M., (1972): “Economies of Scale and Customs Union Theory”, Journal of Political Economy, Volume 80, No.3, pp 465-475
DRAGNEVA, R., WOLCZUK, K. (2014): “Eurasian Economic Integration: Institutions, Promises and Faultlines”, LSE Ideas Report “The Geopolitics of Eurasian Economic Integration”,
DREYER, I., POPESCU, N. (2014): “The Eurasian Customs Union: The Economics and the Politics”, European Union Institute for Security Studies, Brief Issue, Volume 11
GHOSH S., YAMARIK S. (2004): “Are regional trading arrangements trade creating? An application of extreme bounds analysis”, Journal of International Economics, Volume 63, Issue 2, pp 369-395
ISAKOVA A., KOCZAN Z., PLEKHANOV A., (2015): “How much do tariffs matter? Evidence from the customs union of Belarus, Kazakhstan and Russia”, Journal of Economic Policy Reform
KROTOV, I. (2011): “Customs Union between the Republic of Belarus, the Republic of Kazakhstan and the Russian Federation within the framework of the Eurasian Economic Community”, University of Canberra, Australia and University of Munster, Germany, World Customs Journal Volume 5 No. 2, p 129-140
KOHL, T. (2014): “Do We Really Know That Trade Agreements Increase Trade?”, Review of World Economics, Volume 150 No. 3, pp 443-469
33
LEDERMAN D., OZDEN C., (2005): “Geopolitical Interests and Preferential Access to U.S. Markets”, World Bank
LIPSEY, R.G. (1957): “The Theory of Customs Unions: Trade Diversion and Welfare”, Economic, New Series, Volume 24, No. 93, pp 40-46
MAYER T., ZIGNAGO S., (2005): “Notes on CEPII’s distanced measures: The GeoDist database”, CEPII, WP No. 2011 – 25
SERLENGA, L., SHIN Y. (2004): “Gravity models of the intra-EU trade: application of the Hausman-Taylor estimation in heterogeneous panels with common time-specific factors”, University of Edinburgh, Mimeo
SHUMYLO-TAPIOLA, O., (2012): “The Eurasian Customs Union: Friend or Foe of the EU?”, Carnegic Endowment for International Peace
SUVANKULOV, F., GUC, Y. (2012): “Who is trading well in Central Asia? A gravity analysis of export from the regional powers to the region.”, Eurasian Journal of Business and Economics, Volume 5, Issue 9, pp 21-43
TINBERGEN, J., (1962): “Shaping the world economy: suggestions for an international economic policy”
TOCHITSKAYA, I. (2010): “The Customs Union between Belarus, Kazakhstan and Russia: and Overview of Economic Implications for Belarus”, CASE Network Studies & Analyses No. 405
VINER, J. (1950): “The Customs Union Issue”, Carnegie Endowment for International Peace, New York
WILLIAMS, R. (2016): “Outliers”,
https://www3.nd.edu/~rwilliam/stats2/l24.pdf, Retrieved December 2016
WORLD BANK (2016): World Development Indicators (Washington, DC: World Bank). Retrieved October 2016
WORLD BANK: “What is the difference between current and constant data?”,
34
EUROPEAN UNION NEWSROOM: “EU sanctions against Russia over Ukraine crisis”,
35
Appendix 1
Hauman test
Test: Ho: difference in coefficients not systematic chi2(2) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 17.38
Prob>chi2 = 0.0002
