The Economic Determinants of Post-Soviet Belarus: The Role of FDI in Economic Development

Adding to the vast literature readings, this study investigates the role of major determinants of economic development in Belarus from the period of 1992-2017. We employed the simple OLS regression and ARDL Bound testing approach to determine the economic indicators of Belarus. Our findings exhibit that imports and exports have been remained as major determinant whereas, FDI has not been a significant part since last three decades. Belonging to the pool of developing world, FDI has not been a determinant in Belarus while its main emphasis is on exports and imports. This study also implicates the Belarus’ initiative with other developing countries to increase FDI and reduce State-owned Influence. While ensuring the differences with other economies, government must attract the economically suitable FDI avoiding any possible negative effects of FDI on GDP.


Introduction
Since the dissolution of the Soviet Union in 1991, Belarus gained its Independence and thereafter has enjoyed periods of good economic performance. This period of economic prosperity coincided with maintaining an economic model based on state planning and state-owned enterprises. According the investment policy review of the republic of Belarus 2009, "FDI inflows has traditionally been very low, discouraged by frequent policy changes and an ambiguous policy stance toward private sector development. Since late 2007, in an effort to maintain high levels of economics growth, Belarus has launched reforms to develop the private sector which currently accounts for only around 25 percent of GDP. In this context, ambiguous goals have also been set to increase FDI inflows from 2.6 percent to 15-20 percent of total investment, and to move into the top 25 countries in the World Bank's "Doing Business" rankings". Since Belarus endeavored to change its policy regarding FDI, Investment has been observed above or close to 2007 levels as shown in Figure 1, which plots Belarus' FDI-GDP ratio. The reason for this rise is the undertaking of many FDI projects in the country. Notably, one of the largest and most promising of those projects is the China-Belarus Industrial Park (Great Stone Park). This paper will focus There is no doubt that the Great Stone international park serves as a catalyst for a favorable economic climate conducive for foreign direct investment. Initial estimates put investment by resident companies in the region of 1.4 -3.7 Billion dollars from 2016-2020. The administrators of the park aim to attract more than 200 high tech companies, and employ more than 120,000 people during this time period. This ambitious project is expected to have a positive impact on the Belarusian economy. The venture will employ thousands of people, improve Belarus's trade and introduce companies to the market of the customs union and common free market zone with a population of about 170 million.
Overall, there are some positive as well as negative effects of FDI on GDP. The positive effects include Trade Effects, Human Capital Contribution, Spillover effects, Competition level and Management and Governance Practices. The negative effects of FDI include, The Crowding Out Effect, Profit Repatriation and The Dual Economy Effect.

Positive FDI Effects
Foreign Direct Investment is believed to positively affect the receiving economy in a number of ways. According to Pelinescu (2009) there are four ways in which Foreign Direct Investment benefits the receiving economy. Firstly, Foreign Direct Investment introduces transnational companies which interact with the local economy by building production facilities and hiring workers. Secondly, the international affiliates of the transnational companies establish forward and backward linkages which stimulate production in local firms and provides a platform for the transfer of technology. Thirdly, affiliate companies can help local companies to improve their efficiency and effectiveness with competition. Finally, the entry of FDI projects can increase employment and income and have multiplier effects on the host economy. Some other positive effects are listed below.
Trade Effects FDI tends to boost factor productivity and efficiency while increasing the productive use of resources in the home country. This leads to an increase in Capital Stock which raises output levels. The home country can then improve its trade position in the global market by increasing Exports and possibly decreasing Imports from other countries.

Human Capital
When foreign Firms migrate to other countries, they provide a range of benefits for the workers in the home country. Foreign companies provide more places of work, which intern reduces the unemployment rate. They also provide higher wages and better working conditions for employees. Most importantly, the dissemination of technology from the Firms to local employees increases effectiveness and efficiency, providing them with new and better skills.

Spillover effect
This is effect is the results of the transfer of technology from the Foreign Company to the local economy. As the name suggests. This knowledge spills over into the host economy making it more efficient.
ijef.ccsenet.org International Journal of Economics and Finance Vol. 11, No. 2;2019 Competition level As a result of the spill-over effect, local companies will have more competition from foreign firms. This will raise the level of competition in the economy and will lead to higher levels of productivity and innovation.

Management and Governance Practices
The presence of foreign Firms, through the acquisition of local firms will change the management and corporate governance of the firms. This benefits the business environment by imposing systems and processes that are more efficient and transparent.

Negative FDI Effects
The crowding out effect As may be the result in a lot of developing countries, the presence of Foreign or Multinational Enterprises can cause emergence of monopoly power in the market. This can happen because local companies cannot match the product quality and efficiency of the MNEs or because local companies don not have access to as much resources and financing.

Profit Repatriation
This effect is self-explanatory. This happens when most or all of the profit gained by MNE is sent back to the MNE's country rather than have it circulate in the host country's economy.

The Dual Economy Effect
The Dual Economy effect occurs when there is a highly developed sector dominated by the MNEs and an underdeveloped sector dominated by the local companies. The country then depends more on the developed sector which tends to be capital intensive, while the underdeveloped sector is labor intensive. This disrupts the development of the country since most of its labor force is located in the undeveloped sector.
The Great Stone Industrial Park should have all of the aforementioned positive effects on Belarus' economy. Against this backdrop we investigate the relationship between Gross domestic product (GDP), Foreign Direct Investment (FDI), Exports , Imports, Gross capital formation (GCF), Population and Trade Openness. The purpose of this paper is to quantify the relationship between foreign direct investment and the Gross domestic product in Belarus. These results will be helpful to the Belarusian government in evaluating the benefits of the Great Stone Park. The rest of the paper is organized as follows. Section two reviews the literature. Section three analyzes the data. Section four presents the econometric methodology and results, and section five gives the conclusion.

Literature Review
The first observation noted is the different theoretical frameworks used by the authors in the literature. The popular framework is that of cointegration analysis, vector error correction and vector auto-regression. This type of analysis is preferred because it allows the author to investigate the backward and forward linkages between variables and, to give proper interpretation of non-stationary variables.
Rahman (2015) used the granger causality framework to investigate the relationship between FDI and GDP in Bangladesh. The data spanned 42 years from 1972-2013. The author found that there is a unidirectional relationship between FDI and GDP. It was found that GDP granger causes FDI.
Baig (2016) also uses granger causality frame work to analyze the relationship between FDI, GDP and Inflation in selected south Asian countries from 1991-2012. The author found that the relationship between FDI and GDP was not significant for countries like Pakistan, Sri Lanka and Singapore, however for Nepal the results show unidirectional positive relationship between FDI and GDP.
Hussain ( Mehboob (2014) estimates a production function to analyze the relationships between FDI, K, L and GDP for 59 selected countries from 1992-2010. The study revealed the presence of Unidirectional causality among Gross Domestic Product and Foreign Direct Investment Inflows and Bidirectional causality between the all variables of the production Function.
Still yet another methodology used in the literature is the simple regression and Dynamic regression. The drawback this framework is that it can only be used when variables are stationary which, is rarely the case in economics. In this framework, the magnitude of the slope coefficient is emphasized as there can be no directional analysis.
For the purpose of this study, the bounds testing approach to cointegration developed by peseran (2001) will be employed. This method is chosen because it does not need a lot of data.
Hymer (1976) propagates an interesting theory that sheds light on the literature. Hymer's theory posits that a corporation has a nationality and gives three reasons. The first reason is that a firm has a legal nationality. This provides boundaries within which the firm must operate. The second is the country of residents of the shareholders. The idea is that most shareholders may live in the same country and the firm is obligated to pay dividends in that currency. The third is the nationality of the managers, which is said to affect their loyalties. (the motivation of FDI).
Dunning (1977) introduced the OLI theory which stands for Ownership, Localization and Internalization. The theory looks at the three advantages of a Multinational corporation leading to FDI. Ownership refers to the specific knowledge capital such as managers, patents, technologies, marketing strategies etc. A company who has an abundance of this capital will be inclined to engage in FDI because the capital can be replicated in different locations without decreasing inefficiency. The advantages of localization which lead to FDI are quite attractive. They include producing close to consumers, saving transportation costs, finding cheap inputs among others. The internalization advantage deals with whether the firm contracts with the supplier in the foreign country at which time the supplier becomes an employee (Internalization) or whether it finds it best to leave the supplier separate.
Overall, there are some positive as well as negative effects of FDI on GDP. The positive effects include Trade Effects, Human Capital Contribution, Spillover effects, Competition level and Management and Governance Practices. The negative effects of FDI include, The Crowding Out Effect, Profit Repatriation and The Dual Economy Effect.

Data Sources
The study investigates the relationship between the log of Exports, Imports, Inward Foreign direct investment A careful analysis of the plotted data shows that all of the variables pretty much have an upward trend except population which has been steadily falling. It is easy to see that Belarus experienced a period of economic prosperity from around 2001 to 2012/2013. This period is characterized by upward trending GDP, EXPORTS, GCF, TRADE OPENNESS. These variables will be expected to have a positive impact on GDP as their levels increase. However, during that same period, imports increased and Population decreased. Although population growth slowed consistently, the reduction in growth is not significant to restrict the growth of GDP. It can also be seen that FDI did not play a big role in the development of Belarus during the period of economic prosperity since it does not have a definite upward trend. Upon graphical inspection, the variables which are expected to positively affect GDP are the four variables which were pointed out earlier in this paragraph. Population is expected to have little to no effect while FDI's impact is inconclusive and could go either way.  The summary statistics in table 1 reveal that Belarus generated an average of 35 Billion US dollars from 1992 -2017 with the maximum being 78 Billion and the minimum 12 Billion. This shows that Belarus has maintained an average level of GDP above its lowest level. Exports and imports also tell a similar story with maximums of 46 and 46.4 Billion, minimums of 1.9 and 2.5 Billion and averages of 17.9 and 20.7Billion. Likewise, since trade openness is calculated as a combination of Exports and Imports, it has also maintained an average level above its minimum with an average of 103%, a maximum of 141% and a minimum of 28%.
Although FDI has not been a major player in the Belarus economy, the statistics show that foreign interest was maintained. Within the period of study, FDI maintained an average of 905 Million, with a maximum of 4 Billion and a minimum of 7 Million. Next, we move to K or Gross Capital Formation. Any economist knows the importance of capital growth to a healthy economy. In Belarus' case, capital stock was maintained at an average of 11 Billion, with a maximum of 29 Billion and a minimum of 2 Billion. As a nation's population increases, and with it its labor force, one would expect that it should contribute positively to GDP growth. The descriptive statistics show that Belarus' population has also maintained an average level above its lowest value. To obtain a better understanding of the relationship between the variables we compute the correlation matrix. As expected, it shows strong positive correlations. Special emphasis is given to the correlation between LFDI and LGDP which is 0.80. These results are a pretext to the detailed investigation to follow. The study will go beyond this simple correlation analysis, however the expected signs on coefficients are expected to be positive except for IMPORTS.

OLS Estimation
In this study we make use of two econometric procedures. The first is a simple OLS estimation of the log of the four variables where the following model is assumed: LGDP

Autoregressive Distributed Lag Approach to Cointegration
In this study the ARDL bounds testing procedure, developed by Pesaran et al. (2001), is adopted. This approach is preferable to other single equation techniques like the Engle and Granger et al. (1987) two step estimator; because it assumes that all variables are endogenous. Also, the ARDL approach is applicable whether the regressors are purely I(0), purely I(1) or co-integrated We assume the same model as the OLS estimation as the general model for this procedure.
The first step of the bounds test is to estimate an unrestricted error correction version of equation (1) by OLS. The unrestricted error (equilibrium) correction model (ECM) proposed by Pesaran et al. (2001) applies the fundamental principles of the Johansen five error correction muliti-variance of which there are five possible models. The models are chosen based on the assumed components of the data. This is determined from a graphical inspection of the data. The possible models are as follows: Case I (no intercepts; no trends) This model is chosen if it is assumed that the data has no linear trends and no intercepts are desired.
Case II (restricted intercepts; no trends) ∑ This model is chosen if it is assumed that the data has no linear trends and if is fit to restrict the intercepts to a certain value.
Case III (unrestricted intercepts; no trends) ∑ This model is chosen if it is assumed that the data has no linear trends and the intercepts are not restricted to any values.
Case IV (unrestricted intercepts; restricted trends) This model is chosen if the data is assumed to have a trend, but is it found suitable to restrict the linear trends between certain values. Also, the intercept is unrestricted.
Case V (unrestricted intercepts; unrestricted trends) ∑ This model is chosen if the data is assumed to have a trend and the intercept is not restricted between values. Both the linear trend and the intercept have no restrictions placed on them in this model specification.
The bounds procedure for testing the relationship between two variables involves an F test on the joint null hypothesis that the coefficients of level variables are jointly zero. It should be noted that the F-statistics follow a non-standard distribution. Hence, instead of the conventional critical values, the F test involves two asymptotic critical bounds covering all possible classifications of the variables, i.e. whether they are I(1), I(0) or mutually co-integrated. The lower bound values assume that the variables are purely I(0), and the upper bound values assume that they are purely I(1). If the computed F-statistic exceeds its respective upper critical values, then ijef.ccsenet.org International Journal of Economics and Finance Vol. 11, No. 2;2019 there is evidence of a long -run relationship; if below the null of no cointegration, it cannot be rejected; and finally, if it lies between the bounds, inference is inconclusive. Once cointegration is established, the long-run and corresponding error correction models can be estimated. As a preliminary step to the empirical analysis, the order of integration of the variables is determined. LGDP -1.57 ( Note. Augmented Dickey Fuller Test to test for unit roots in the data at both level and first difference forms of the data. Table 5. Order of integration

Order of Integration
LGDP This is done to ensure that none of the variables are integrated of order two I(2). The Augmented Dickey -Fuller test is used to determine the level of stationarity. The null hypothesis of this unit root test states that the series has a unit root. Failure to reject the null hypothesis means that the series is not stationary. The results in tables 4 and 5 show that none of the variables are I(2). Therefore, the ARDL approach is undertaken.
In the first stage of the ARDL analysis, it is standard procedure to ascertain the existence of a long run relationship among the variables. The maximum lag length is set at 2 since the data being used is annual data. The results shown in table 6 imply that the null of no co-integration can be rejected. Therefore, there exists a cointegrating relationship between the four variables of interest. Note. Bounds Testing procedure based on a 6-parameter analysis. The F statistic is outside of the testing bounds which signifies a cointegrating relationship.
Given the existence of a co-integrating relationship, the long run conditional ARDL and the error correction models can be estimated. The Schwarz-Bayesian criterion is employed to determine the order of the ARDL. The chosen order is an ARDL (1, 1, 2, 0) model.  The OLS results seem to be quite counterintuitive for two of the variables. First of all, it can be clearly seen that only two of the variables are significant. These variables are DLEXPORTS and DLGCF. The signs coincide with expectations, positing that Exports and Capital Stock have a positive relationship with GDP. Indeed, a 1% increase in exports will increase GDP by 0.40%. also, a 1% increase in the Capital stock will increase GDP by 0.57%. Although all of the other variables are insignificant, their signs are not as expected. The results states that Trade Openness and Population growth have negative effects on GDP. The seeming puzzling effect of Population growth is explainable since Belarus' population has been steadily declining as GDP has been rising. In order to get a deeper understanding of the relationship between the variables, the ARDL approach to cointegration is employed. The first step of the ARDL process is to estimate the ARDL model chosen by the information criteria. The chosen model is an ARDL(1,0,0,0,0,0,1) model. This is then dissected into two parts, a short-run analysis and a long-run analysis.   Table 11 shows the shorn-run results for the ARDL model. One thing to notice is the trend variable which is highly significant. Secondly, the coefficient of TRADEOPENNESS is negative, suggesting a negative relationship with GDP in the short-run. A 1% increase in the TRADEOPENNESS position will decrease GDP by 0.9%. This of course is not surprising since imports are greater than exports from 1993-2017. The constant term is also significant at the 1% level, suggesting that if all the variables in the equation were zero, the value of DLGDP would be -2.04%.   Table 12 shows the long-run results of the ARDL model. The model reveals that capital stock, FDI and Population are not significant and therefore has no long-run relationship with GDP. However, the results show that a 1% increase in IMPORTS will increase GDP by 0.58%. It also shows that a 1% increase in exports will increase GDP by 0.45%. also, a 1% increase in the trade openness position will increase GDP by 0.02%.

Conclusion
This study employs the ARDL procedure to investigate the major determinants of Belarus' Economy. To investigate the determinants, we have observed short and long-run coefficients to see their impact on GDP for Belarus from 1992-2017.Other methodologies applied includes graphical analysis, correlation analysis and OLS.
The ARDL results posit that in either short-run or long-run, FDI has no relationship with economic development whereas, trade proxies (imports and exports) are found significant evidencing the major emphasis of Belarus on trade.
Since the study is limited to one country, it may not provide the evidence on other regional countries. To further check the robustness of investment deficiency in the portfolio of Belarusian economy, it is important to consider the presence of any linkage between trade and investments of the country. Moreover, it is also important to consider economic dynamics of other states which were once a part of the Soviet Union and broke away in the early 1990's for future research and construct the comparative analysis of these countries' economies.
Based on the results of our study, the implications for Belarus is to attract the foreign direct investments especially with the consideration of Great Stone Industrial Park that may assist in strengthening the economic development of the country. According to the results, since the dissolution of the Soviet Union, the FDI driven initiative has little to no effect on GDP in Belarus. In line with its policies to increase the amount of FDI in the country, one of the implications could be reducing the influence of state-owned enterprises. Secondly, strict supervision of the projects ensuring the optimum utilization of investments efficiently for economic growth of the country is highly required. Government must also seek to attract the right type of Foreign Direct Investment to avoid any negative effects such as the Dual Economy Effect, the Profit Repatriation Effect and the Crowding out effect. Domestic Market-seeking MNEs should be avoided and Export Oriented MNEs should be welcomed.
Since Belarus' Independence, FDI has played no significant role as a major determinant of the economic development. With the results achieved, it is necessary to improve the investment friendly environment ensuring the efficient economic growth of the country.