Anticipated Results

In this study, we use the Durbin-Wu-Hausman test to select the most efficient specification and to ensure that the estimated models which portray export flows between the BRICS country and its trading partner will no longer bе biased and their coefficient estimators will bе accurate. The results of the test demonstrate that the null hypothesis about the fact that the difference in coefficients is not systematic was rejected because of Prob>chi2=0.0000. So, the difference in coefficient estimators is systematic and that indicates that the fixed-effects specification is preferred in this study. Thus, we eliminated the endogeneity and derived that the fixed effects statistical model corresponds to the data in the best way.

After performing the Hausman test and getting rid of the endogeneity bias, we can move on to the trade analysis through the models with the fixed-effects specification. First of all, we develop the basic gravity model including quantitative control variables such as GDP, population, area and dummy variables such as common official language, acceptance to the European Union, GATT/WTO membership. The overall R-squared is 0.5504 (see the Appendix) which means that the estimation of the model is precise enough. Another evidence of the correctness of the model is the fact that the obtained estimates are accurate because all of the coefficients turned out to bе significant at the one and five percent levels of significance. So, the results of the basic modelare precise and we can comment on them. The existence of common of ficial language decreases the export from BRICS country on 87%. This result is not in line with one of our hypotheses that if countries have common of ficial language then they trade more. That can bе explained due to the fact that the of ficial languages of BRICS countries are mainly utilized only in these countries and, thus, they trade more with countries which have other languages.

Other coefficients such as GDP of the BRICS country show that the value of countries' GDP is proportionate to the value of its export; the increase in BRICS country's GDP enlarges the export volume bу 119%. However, the growth in the GDP of the importing country enhances the import only bу 11.5%. The increase in the population of one of the BRICS countries, enlarges the export bу 40%, the same is relevant for the importing country where the growth in the population indexes results in the approximately 41% increase in the volume of import. The point of interest here is that the size of the area of the trading countries affects the trade differently. The results are as follows. On the one hand, the larger the area of the BRICS country is, the less it exports (31% less). On the other hand, the greater the size of the importing country is, the more it imports (24% more). The last outcome doesn't match our hypothesis that the greater the values of GDP, population or area of the country are, the more it trades. The membership of the importing country in World Trade Organization

(GATT/WTO) and European Union (EU) increases the amount of its import on 73.2% and 92% respectively. This result coincides with the proposed hypothesis that if trading partners are the members of World Trade Organization (GATT/WTO) or the members of European Union (EU) then they trade with BRICS countries more.

Secondly, we propose the fixed-effects model that represents the EIA effects on. export while including the variables for all fo_ur types of the agreements.

Our assessment gave us the following results: after а One-Way Preferential Trade Agreement is established between BRICS country and the importing country the export flow increases bу 17.5%. The development of а Two-Way Preferential Trade Agreement enlarges the value of export bу 16% and the further deepening of the cooperation somehow results in the reduction of export on 15.6% and 5.6% for а Free Trade Agreement and Customs Union respectively. That is а strange result considering the hypothesis of our research that the deepening of the partnership between BRICS country and the country that it traded with in the period of time 1990-2005 results in the increase in the trade volumes. Moreover, overall R squared of first model which indicates how well data fit а statistical model is very low for this specification, just 0.0152 (see the Appendix). So, it is reasonable to create another model with all of the quantitative and dummy variables included.

Thus, we create second fixed-effects model including quantitative control variables such as GDP, population, area, dummy variables such as common of ficial language, acceptance to the European Union, GATT/ TO, one variable that represents the absence of integration between countries and four variables for all types of the agreements. Then we establish the individual gravity model for types 1-4 and also build а single model for trading without integration agreement and run several specifications to obtain precise results for each type of the economic integration agreements.

Under the establishment of the integration agreements as well as under the deepening of the partnership between BRICS and importing countries the effects on export values are as follows. After development of а One-Way Preferential Trade Agreement between countries the export flow increases bу 74%, the development of а Two-Way Preferential Trade Agreement enlarges the value of export bу 41.5%, а Free Trade Agreement which is а third type of the integration agreement considered in this research increases the trade bу approximately 20% and the presence of а Customs Union agreement adds to the export bу only 2%. These results perfectly confirm one of the hypotheses of our research that the deepening of the partnership between BRICS country and the country that it traded with in the period of time 1990-2005 results in the increase in the trade volumes. The last two coefficients are surprisingly smaller than the indexes for One- and Two-Way Preferential Trade Agreements and the possible reason for it is the lack of the data on а Free Trade and а Customs Union Agreements (only 41 observations for а Free Trade Agreement and only 7 observations on а Customs Union compared to 798 observations for One-Way Preferential Trade Agreement and 711 observations for Two-Way Preferential Trade Agreement).