The Sustainability of International Trade: The Impact of Ongoing Military Conflicts, Infrastructure, Common Language, and Economic Wellbeing in Post-Soviet Region

Abstract

The sustainability of international trade is subject to immense pressure. Apart from obstructed logistics, disruption of production chains and changes in demand, the sustainability of international trade is heavily affected by the sanctions caused by the Russia–Ukraine conflict. This paper studies the factors predicting sustainable international trade in the post-Soviet region. We hypothesize that ongoing conflicts, infrastructure, language integration, geographical proximity, common border, and economic wellbeing significantly impact international trade. Methodologically we rely on linear and hierarchical regressions estimating a set of gravitation models (N = 15 countries—104 trading pairs; 2010–2020). The results suggest that Russian as a primary language and the average density of road networks positively predict bilateral trade volume. The geographical distance, infrastructure differences, military conflicts, and, surprisingly, the pair-average GDP per capita diminish bilateral trade. Countries’ GDP mediates the effect of GDP per capita. The results are robust over time. The results present an important insight into sustainable international trade within the region affected by the numerous military conflicts in the past and the war conflict between Russia and Ukraine nowadays. The rebuilding of Ukrainian transport infrastructure is one of the essential measures from the country’s point of view and a factor supporting internationally sustainable food supply.

1. Introduction

The sustainability of international trade experiences immense pressure. Apart from troubles with logistics, the restrictions imposed during the COVID-19 epidemic disrupted production and trade patterns and decreased overall demand [1,2,3,4]. The revival of demand after the restrictions were abolished found international trade and travel flows unprepared [5,6,7]. The long-term conflicts in international politics led to a number of sanctions imposed on international trade flows and resulted in a military conflict between Russia and Ukraine. The conflict disrupted the trade flows even more, especially in food products and related products (such as fertilizers). The events endangered the sustainability of the food market, which was undergoing an immense transformation over the last decades [8,9,10,11], and questioned food safety. Ukraine, which was, in 2020, the fifth largest exporter of wheat in the world with exports amounting to USD 4.61B in wheat [12], partly lost its exporting capacities due to war restrictions, difficulties with agricultural production due to military interventions, and the price for energy resources and fertilizers. The breakage of trade flows with Russia led to the danger of an energy crisis in Europe. All these events elucidated the interdependence of the countries and the importance of sustainable international trade for economic development [13]. However, trade sustainability depends on the sustainability of underlying factors, some of which exhibit increasing volatility. This paper studies the factors affecting sustainable (at least ten years) bilateral international trade among post-Soviet Union countries.

The dissolution of the Soviet Union led to substantial economic, political and social disintegration of the former republics [14]. The disconnection from the informal center (Russia), the increase in nationalist tendencies, and the wish to join other trade zones (such as the EU) lead to the breakage of value chains in most sectors damaging the economies [15]. The increased competition made some industries obsolete, while the openness to the global world redirected part of the international trade outside of the region [14]. The local military conflicts and tensions made trade logistics within the area risky and, sometimes, even dangerous [16,17,18,19,20,21].

Yet, many factors predetermine the post-Soviet Union countries for trade integration. First, close geographical location and the same standards of product infrastructure inherited from the Soviet Union facilitate the logistics and certification [22,23,24,25,26]. Second, speaking a common language (Russian), though not always welcomed, facilitate mutual understanding [27,28,29,30]. Third, the trade connections and patterns created in the period of the Soviet Union can still prevail in some countries. Forth, the remnants of common consumption preferences can make some commodities produced in the region more welcomed either because of price or the type of goods [31,32,33,34].

This paper studies factors associated with trade integration of the Former Soviet Union countries. The literature lists many factors affecting international trade. The two most frequently tested ones are the geographical proximity of the countries [35,36,37] and the economic outputs of the countries [38,39]. Countries with higher GDP and geographically closer to each other usually trade more [40], as well as countries where people can speak the common language [28,41]. We suggest that in the region of post-Soviet Union countries, the ability to communicate in Russian can significantly predict the volume of international trade (similar to [30]). Moreover, international trade is significantly affected by the quality of infrastructure [24]. Given the territory, various landscapes, and climate conditions in the post-Soviet Union countries, the density of roads seems to be the relevant indicator for the quality of infrastructure.

Economic development is shown to be significantly associated with international trade. The direction of the relation may be twofold. First, the more economically advanced countries are more likely to have higher purchasing capacity, thus, can create higher demand for imported goods [42,43]. Oppositely, specialization, as accompanied by international trade, supports and predicates local economic development [44]. Being as it is, the divergence of the post-Soviet Union countries in economic outputs is likely to be one of the factors that are related to international trade.

To sum it up, the aim of this paper is to study the factors affecting international trade in the post-soviet region. We hypothesize that given the shared past, the factors affecting trade integration in the region can be listed as follows: economic development (measured both as GDP and GDP per capita), geographical proximity, ongoing conflicts, infrastructure, and common language. In the case of GDP per capita and infrastructure, we suggest that trade integration is affected by both the average levels and the difference between the trading partners.

Methodologically, we rely on the gravitation model of international trade [45,46] for one hundred and four couples of fifteen post-Soviet-Union countries over the period of 2010–2020. We run a set of yearly regression analyses and one hierarchical regression analysis to study the controversial effect of GDP per capita on international trade in the region. We also ran a mediation analysis to find the mediating variable that significantly influenced the impact of GDP per capita on international trade in the area.

The paper is structured as follows. First, we briefly summarize the factors related to international trade as studied in the literature. Second, we define the factors relevant to the post-Soviet Union region and formulate the hypotheses. Third, we present the methodology, data, results, and discussion. The last section concludes.

2. Trade Integration of Post-Soviet Union Countries

As a former part of the Soviet Union, the post-Soviet Union countries inherited several factors that predestined them for trade integration. First, the communist government viewed economic integration as a factor that helped to hold the country together and intentionally located parts of integrated production chains in different soviet republics [47,48]. Second, the countries share a close geographical position and, partly inherited, common production practices, though national differences grow in importance [37,49]. Third, the countries shared similar consumption tastes, which, after the break down of the Soviet Union, started to divert apart and yet had common roots. Forth, the countries shared similar standards in infrastructure, such as similar parameters of energy systems, similar technologies for power stations, similar width of railroads, etc. Fifth, the Soviet Union essentially eliminated the language barrier between the republics meaning that all the countries used the Russian language as an official language, and most of the population of the Soviet Union spoke Russian.

After the breakdown of the Soviet Union, many factors changed. The nationalist and separatist tendencies locked the countries to partial isolation [50,51,52] and ignited the inter- and intra-country ethnic conflicts. Within the surveyed countries, the conflict between Ukraine and Russia has the greatest impact on trade, so in a number of recent international trade between these countries has been disrupted, including transit [53].

Opening to the world redirected parts of international trade flows from the post-Soviet region to western countries, China, and others [15,54,55]. The economic downturns caused by the disruptions of the 1990s further decreased purchasing power in some countries leading to a relative decrease (in absolute value or in the rate of growth) of local demand for foreign goods. Similarly, the increased competition from the western world made some branches of industry, agriculture, or services obsolete, decreasing the export potentials of the countries [54,56]. Despite all this, in some parts of the region, trade survived and even enhanced, especially in agricultural export (Belarus, Ukraine, and Azerbaijan), raw materials (Russia and Ukraine), and energy products (Russia and Kazakhstan) [57,58].

3. The Factors Affecting Sustainable Trade Integration in the Literature

International trade is affected by many factors. First, the economic size of the countries (as measured by GDP or GDP per capita) was reported to positively affect international trade integration [59,60]. The direction of this association is likely to be circular. Countries with higher total GDP and GDP per capita have more goods and services to offer for export and have the higher purchasing power to welcome imported products [61]. Moreover, economic prosperity is usually associated with successful specialization [62,63], which further enhances international trade as the countries with high specialization need to import the products they do not produce.

Efficient trade requires a significant level of infrastructure and logistics as trade traditionally implies transporting the goods. The sufficient density of roads, railways, ports, and airports is substantial. Road net length and infrastructure are also essential for international trade, especially for countries without access to the sea. Hanousek and Kočenda [64] show that the ease with which goods can move across borders affects the time and transportation costs. In the case of energy trade, the pipelines, electric wires, and other connections significantly facilitate the trade. We can hypothesize that except for the absolute density of infrastructure, the difference in the density might be substantial for trade integration. Even though one country may possess perfect infrastructure, if the other country’s infrastructure is poor, the traded goods are destined to stay on the border if it ever riches it.

Trade is significantly affected by the geographic proximity of the countries and the fact that the trading countries share the same border. The border effect on international trade is well documented but not well understood. Though common border may be positive for trade in peaceful times (as the countries do not have to move the commodities through the territories of third countries), in case of military conflicts, when direct trade is not possible, the goods start to flow through the territory of the third countries (e.g., trade between Russa and Ukraine). Thus, the common border stops to have a positive effect. In general, the nature of the border effect resulting from trade barriers, past linkages, and comparative advantages is not yet understood [47].

Last but not least, the ability of people to speak the same language is advantageous for trade integration [27,28,29,30,41]. While in the western world, English, Spanish or French languages are often used as the language of trade, and the traditional common language was Russian in the post-Soviet Union area. Currently, Russian is the official language in four countries (Belarus, Kazakhstan, Kyrgyzstan, and Russia), though it is unofficially spoken in all countries. A common language is part of the integration process [65].

To sum it up, we hypothesize that the trade integration in the post-soviet region is affected by the following factors:

Hypotheses

The volume of bilateral trade of the post-Soviet-Union countries is

H1. 

Positively related to the joint economic size of the countries measured by multiple countries’ GDP.

H2. 

Negatively related to the distance between the countries.

H3. 

Greater if the countries have a common border.

H4. 

Positively related to the availability of infrastructure of both countries (namely the density of roads).

H5. 

Negatively related to the difference in the availability of infrastructure in the trading countries (namely the difference in the density of roads).

H6. 

Negatively related to the wars and other military conflicts in the countries.

H7. 

Positively related to the average economic performance of the countries (measured by GDP per capita).

H8. 

Negatively related to the difference of economic performance (the biggest trade is in the countries that are economically approximately at the same level).

4. Materials and Methods

4.1. Data and Indicators

The paper analyses the data of international trade of 15 post-Soviet Union countries over 2010–2020. Namely, we analyze the data from Russia (Russian Federation), Ukraine, Belarus (Republic of Belarus), Moldova (Republic of Moldova), Uzbekistan (Republic of Uzbekistan), Kazakhstan (Republic of Kazakhstan), Kyrgyzstan (Kyrgyz Republic), Tajikistan (Republic of Tajikistan), Turkmenistan (formerly the Republic of Turkmenistan), Georgia (formerly the Republic of Georgia), Azerbaijan (Republic of Azerbaijan), Armenia (Republic of Armenia), Lithuania (Republic of Lithuania), Latvia (Republic of Latvia), Estonia (Republic of Estonia). In total, we work with 104 couples of bilaterally trading countries per year over 2010–2020 and employ the following indicators.

The indicators for the volume of international trade for the couples of countries were adopted from UN Comtrade Database [66]. We employed the data of total export from country i to country j in the current USD. Given that the trade flows from country i to country j and back were significantly unequal, we computed averages of Exports of country i to country j and Exports of country j to country i. The final trade flow for the 2019 (the last year before the COVID epidemic) ranged 88,000–17,666,208,000 USD; M ± SD: 2,109,846,000 ± 668,560,000 USD.

The geographical proximity between 104 couples of bilaterally trading countries for each year (range: 123–3853 km; M ± SD: 2009 ± 1103 km) is measured by the air (flying) distance (also known as great circle distance) between capitals of countries (km, similarly to [67].

The ongoing conflicts scores (in 2019 ranging 1.41–3.07; M ± SD: 1.95 ± 0.46) were adopted from Institute for Economics and Peace (Global Peace Index [68]). We computed the average logarithms of conflict scores of trading couples for further analyses.

Lengths of the total road networks per km² for 104 couples of countries (range: 35–1378 km; M ± SD: 496 ± 454 km) were adopted from CIA’s World Factbook [69]). We computed the averages and the absolute differences in the lengths of roads of trading couples of countries for further analysis.

Over the last ten years, Russian as an official language has been adopted in Russia, Kazakhstan, Belarus, and Kyrgyzstan. In all the other post-soviet countries, the Russian language was used as an informal language to a greater or lesser extent. However, the attitude to this language varied greatly. We created a dummy variable that equaled one if at least one of the countries in the trading couple adopted Russian as an official language for further analysis. In trade literature, the common language is interpreted as a factor enhancing understanding and cooperation [30,41]. In the context of the post-Soviet Union countries, the prevailing use of the Russian language may also indicate the attitude to Russia and, possibly, belongingness to the common past.

The estimates of total GDP (range: 8300–1,687,448 millions USD; M ± SD: 112,183 ± 320,389 millions USD; 2019) and GDP per capita (range: 890.54–23,397.12 USD; M ± SD: 9000.36 ± 6996.9 USD, 2019) were adopted from the Human Development Indicators, World Bank database [70]. For further analysis, we computed Ln(GDP_i × GDP_j) for each couple of trading countries, as indicated in the model below. In order to account for GDP per capita, we computed the averages and absolute differences of Ln(GDP per Capita_i) and Ln(GDP per Capita_j). We did not adjust for inflation as the models below do work with time comparisons.

4.2. The Model

The classical gravitation models for international trade originated from the gravitation theory that indicated that the gravitation force between the two planets could be computed according to Newton’s law of universal gravitation, represented by the following formula:

$$\mathrm{F}=\mathrm{G}\frac{{\mathrm{m}}_1{\mathrm{m}}_2}{{\mathrm{r}}^2}$$

(1)

where

• F = force
• G = gravitational constant
• m₁ = mass of object 1
• m₂ = mass of object 2
• r = distance between centers of the masses

Similarly, the spatial interaction gravity models of international trade state that the volume of trade between two countries is proportional to their economic mass and negatively related to a measure of their relative trade frictions [45,46].

In the simplest formalized way, we can model the trade between countries i and j according to the following formula (ibid.)

$${\mathrm{Trade}}_{\mathrm{ij}}=\mathrm{A}{\left({\mathrm{GDP}}_{\mathrm{i}}{\mathrm{GDP}}_{\mathrm{j}}\right)}^{\mathsf{\alpha }}{\mathrm{D}}_{\mathrm{ij}}^{\mathsf{\beta }}{\mathrm{X}}_{\mathrm{ij}}^{\mathsf{\gamma }}{\mathsf{\epsilon }}_{\mathrm{ij}}$$

(2)

where

• Trade_ij is the volume of bilateral trade of the countries i and j
• GDP_i and GDP_j are the Gross Domestic Products of the countries i and j
• D_ij is the geographical distance between the countries i and j
• X_ij is a set of other variables representing trade frictions between the countries, namely

• Indicators of economic performance—the average GDP per capita of the trading couples and the difference in GDP per capita of both the countries
• Indicators of the infrastructure of both the countries—the average length of road networks per km and the difference in length of road networks per km
• Indicators of ongoing war conflicts (average index of ongoing conflicts of both countries)
• Russian language dummy (Russian as one of the official languages for at least one country in the trading couple of countries)
• Common border dummy—equal to 1 if the countries have a common border and 0 otherwise

Taking logarithmic transformation of Formula (2), one obtains

$$\mathrm{Ln}\left({\mathrm{Trade}}_{\mathrm{ij}}\right)=\mathrm{Ln}\left(\mathrm{A}\right)+\mathsf{\alpha }\mathrm{Ln}\left({\mathrm{GDP}}_{\mathrm{i}}{\mathrm{GDP}}_{\mathrm{j}}\right)+{\mathsf{\beta }\mathrm{LnD}}_{\mathrm{ij}}+{\mathsf{\gamma }\mathrm{LnX}}_{\mathrm{ij}}+{\mathrm{Ln}\mathsf{\epsilon }}_{\mathrm{ij}}$$

(3)

To estimate Formula (3), we run a set of yearly linear cross-sectional regression analyses for the years 2010–2020. The theoretical total number of couples of countries is given by the Formula (4)

N = n(n − 1)/2 = 15 × 14/2 = 105

(4)

where n = 15 countries in the region. We divide by two as we are indifferent as to which country was in the first place and which on the second in the couple.

We considered several options for the estimation of the model (3)—panel data analysis via fixed and random effect models, analysis of 5-year averages, and analysis of single years. The yearly analysis was chosen for the following reasons. First, we wanted to account for the year 2020 as the year of COVID-19 epidemy, which substantially disrupted international trade; thus, 5-year averages were not suitable. The panel data analysis was not chosen due to the insufficient availability of indicators for inflation suitable for the analysis of international trade flows. The classical inflation indicators, such as Consumer or Producer Price indices (CPI or PPI) or HDP deflators, have limited applicability to the overall international trade flows. Moreover, given the quality of data, the panel data analysis would reduce the whole sample to 91 observations, effectively excluding the Central Asian countries, as the latter do not report their statistics regularly. The methodological contribution of this study lies in the robustness of the results in time.

While estimating the models, we had to account for possible multicollinearity. Some of the explanatory variables had to be transformed to avoid multicollinearity. The resulting models reported Variance Inflation Factors (VIFs) at the edge of acceptable levels (we adopted the level of 10, though some sources chose the level 5; [71,72]). Although VIFs for some variables were relatively high (though they have never exceeded the edge of 10), most of the independent variables in the model were statistically significant. Given that, we can conclude that our results are reliable as multicollinearity, if high, rather increases the error term but does not bias the coefficients [73].

The models did not report any outliers above that three standard deviations. We did not have to remove the observations. The robustness of the analysis was manifested by the similar results achieved in 10 yearly analyses. The Pearson correlations of all the variables in Equation (3) are presented in Appendix A.

5. Results

The results of a set of 10 yearly regressions are presented in

Table 1. Factors predicting bilateral international trade in post-soviet union countries in 2010–2020. Results of the set of yearly, cross-sectional regression analyses (Formula (3)).

Year	2020			2019			2018
	B	Sig.	VIF	B	Sig.	VIF	B	Sig.	VIF
Constant	−28.760 ***	0.000		−27.339 ***	0.000		−26.845 ***	0.000
Ln(GDPi × GDPj)	1.304 ***	0.000	4.550	1.341 ***	0.000	3.770	1.301 ***	0.000	3.463
Ln(Geographical Distance)	−1.206 ***	0.000	3.580	−1.167 ***	0.000	2.401	−1.179 ***	0.000	2.396
Russian official language	0.689 **	0.006	1.579	0.693 **	0.006	1.569	0.460	0.056	1.589
Average Road Network Per km2	0.003 **	0.001	6.705	0.004 ***	0.000	5.016	0.003 ***	0.000	4.919
Difference Road Network Per km2	−0.001 **	0.001	4.239	−0.001 **	0.001	4.129	−0.001 **	0.003	4.112
Average Ln(Ongoing Conflicts)	−2.275 *	0.027	3.012	−2.278 *	0.014	2.493	−2.316 **	0.008	2.290
Average Ln(GDP per Capita)	−0.927 *	0.028	8.009	−1.360 **	0.001	6.232	−1.133 **	0.002	6.052
Difference Ln(GDP per Capita)	0.218	0.297	3.013	0.017	0.932	2.546	0.078	0.677	2.427
Common Border	−0.004	0.991	2.859	0.092	0.796	2.265	0.051	0.883	2.270
Sig	0.000			0.000			0.000
N	91			104			104
R2	0.864			0.856			0.860
Year	2017			2016			2015
	B	Sig.	VIF	B	Sig.	VIF	B	Sig.	VIF
Constant	−28.096 ***	0.000		−27.985 ***	0.000		−25.561 ***	0.000
Ln(GDPi × GDPj)	1.283 ***	0.000	3.683	1.204 ***	0.000	4.027	1.298 ***	0.000	3.612
Ln(Geographical Distance)	−1.150 ***	0.000	2.380	−1.048 ***	0.000	2.362	−1.032 ***	0.000	2.325
Russian official language	0.631 *	0.014	1.588	0.709 **	0.002	1.522	0.553 *	0.019	1.551
Average Road Network Per km2	0.003 ***	0.000	4.767	0.003 ***	0.000	4.783	0.003 ***	0.000	4.565
Difference Road Network Per km2	−0.001 **	0.005	4.118	−0.001 **	0.001	4.451	−0.001 **	0.004	3.896
Average Ln(Ongoing Conflicts)	−1.582	0.078	2.375	−1.166	0.186	2.712	−3.543 **	0.001	2.716
Average Ln(GDP per Capita)	−1.008 *	0.012	6.056	−0.694	0.092	7.308	−1.346 **	0.001	5.506
Difference Ln(GDP per Capita)	0.117	0.558	2.172	0.280	0.140	2.238	0.093	0.598	1.782
Common Border	0.052	0.888	2.290	0.175	0.598	2.286	0.345	0.306	2.186
Sig	0.000			0.000			0.000
N	104			104			102
R2	0.850			0.866			0.864
Year	2014			2013			2012
	B	Sig.	VIF	B	Sig.	VIF	B	Sig.	VIF
Constant	−25.558 ***	0.000		−20.202 ***	0.000		−19.728 **	0.001
Ln(GDPi × GDPj)	1.455 ***	0.000	3.594	1.198 ***	0.000	3.013	1.177 ***	0.000	3.158
Ln(Geographical Distance)	−0.897 **	0.002	2.323	−0.805 **	0.006	2.326	−0.816 *	0.012	2.336
Russian official language	0.980 **	0.007	1.563	1.182 **	0.001	1.509	1.468 ***	0.000	1.477
Average Road Network Per km2	0.004 **	0.001	5.001	0.003 **	0.003	5.159	0.003 *	0.013	5.682
Difference Road Network Per km2	−0.001	0.113	3.572	−0.001 *	0.049	3.477	−0.001	0.064	3.478
Average Ln(Ongoing Conflicts)	−4.838 **	0.001	3.249	−4.044 **	0.004	2.238	−4.602 **	0.004	2.363
Average Ln(GDP per Capita)	−2.295 ***	0.000	5.273	−1.586 **	0.003	4.551	−1.477 *	0.012	4.497
Difference Ln(GDP per Capita)	−0.063	0.799	1.679	−0.045	0.857	1.657	0.027	0.924	1.685
Common Border	0.039	0.939	2.178	0.046	0.930	2.167	−0.078	0.894	2.194
Sig	0.000			0.000			0.000
N	102			102			102
R2	0.748			0.744			0.710
Year	2011			2010
	B	Sig.	VIF	B	Sig.	VIF
Constant	−19.466 ***	0.000		−20.810 ***	0.000
Ln(GDPi × GDPj)	1.072 ***	0.000	3.190	1.055 ***	0.000	3.043
Ln(Geographical Distance)	−0.929 ***	0.000	2.278	−1.033 ***	0.000	2.303
Russian official language	0.842 ***	0.000	1.542	0.788 ***	0.000	1.589
Average Road Network Per km2	0.003 **	0.001	5.900	0.003 ***	0.000	6.047
Difference Road Network Per km2	−0.001 ***	0.000	3.466	−0.001 ***	0.000	3.470
Average Ln(Ongoing Conflicts)	−2.829 **	0.001	2.068	−1.014	0.095	1.826
Average Ln(GDP per Capita)	−0.853 *	0.012	4.796	−0.636 *	0.035	4.982
Difference Ln(GDP per Capita)	0.016	0.918	1.716	0.030	0.834	1.834
Common Border	0.392	0.249	2.158	0.192	0.510	2.142
Sig	0.000			0.000
N	101			101
R2	0.879			0.901

*** Significant at the 0.001 level (2-tailed). ** Significant at the 0.01 level (2-tailed). * Significant at the 0.05 level (2-tailed). The table presents the results of linear regression analysis for the gravitation model of international trade presented in Formula (3). Each separate sub-table presents a cross-section of a separate year as listed in the first row. The number of observations N differs, as data from some countries were missing. Namely, Tajikistan, Kirgizstan, and Turkmenistan reported their statistics with a considerable delay and thus were missing from the sample for 2022. Similarly, some of these countries were missing from the samples in the earlier years (2010–2015).

.

The results presented in Table 1 suggest that volume of international trade was positively related to the multiple GDPs of both countries and negatively related to the geographical distance between the capitals, though the existence of a common border was not statistically significant. These results are very common in gravity models of international trade [59,60].

Moreover, we found that infrastructure plays an essential role in both infrastructure indicators: the average density of infrastructure trading countries (similar to [64]) and the difference in the infrastructure across the countries. The average density of roads (per km²) was positively related to the volume of trade, while the absolute difference in infrastructure (measured by the density of the road network) played a negative role. Thus, it can be concluded that more international trade is reported in the countries which are approximately the same in the density of infrastructure. The too big difference in infrastructure decreases the volume of trade.

Expectedly, Russian as the official language proved positively related to the volume of trade. The adoption of Russian as an official language indicates the ability of the population to speak Russian and, most importantly, the lower tendency to stress local nationalism and break the ties (both economic and cultural) with Russia and other countries speaking Russian.

The ongoing conflicts, expectedly, proved to reduce international trade. The conflicts (especially military) divert the countries’ resources from peaceful activities to military ones and increase the uncertainty of international relations, including international trade.

Contrary to our hypothesis, the average economic performance of the trade couples measured by GDP per capita proved to harm international trade in the region. This phenomenon is still to be fully explained. We may hypothesize that more developed countries (as measured by the GDP per capita) have more possibilities to trade outside the region. In contrast, less economically developed countries are “locked” in their trade within the region. Thus, the bulk of trade of more advanced countries may be redirected elsewhere and is not reflected in the data we analyze.

Pearson correlations between the indicators of trade and average GDP per capita appeared to be positive and statistically significant (see Appendix A), as was originally expected. Thus, we concluded that the effect of moderation was present.

The moderating variable that switched the sign of the coefficient (and the direction of the relation) proved to be the joint indicator for the total GDPs of the countries (measured as Ln(GDP_i × GDP_j), the results are presented in

Table 2. Explaining the impact of GDP per capita on international trade. The results of hierarchical regression analysis. Adding Ln (GDP_i × GDP_j) reverses the sign of the coefficient of Average Ln(GDP per Capita).

	B	Sig.	VIF	B	Sig.	VIF
Constant	−27.339 ***	0.000		0.978	0.829
Ln(GDPi × GDPj)	1.341 ***	0.000	3.770
Ln(Geographical Distance)	−1.167 ***	0.000	2.401	−0.771 *	0.015	2.338
Russian official language	0.693 **	0.006	1.569	1.494 ***	0.000	1.468
Average Road Network Per km2	0.004 ***	0.000	5.016	0.001	0.530	4.559
Difference Road Network Per km2	−0.001 **	0.001	4.129	−0.002 *	0.019	4.122
Average Ln(Ongoing Conflicts)	−2.278 *	0.014	2.493	4.783 ***	0.000	1.584
Average Ln(GDP per Capita)	−1.360 **	0.001	6.232	2.076 ***	0.000	3.224
Difference Ln(GDP per Capita)	0.017	0.932	2.546	0.655 **	0.045	2.391
Common Border Dummy	0.092	0.796	2.265	1.603 **	0.005	2.018
Sig	0.000			0.000
N	104			104
R2	0.856			0.602

*** Significant at the 0.001 level (2-tailed). ** Significant at the 0.01 level (2-tailed). * Significant at the 0.05 level (2-tailed). The exclusion of Ln(GDPiGDPj) from the regression analysis different the sign of the coefficient for Average GDP per capita from negative to positive, thus presenting the mediation effect.

. Thus, the phenomenon should be interpreted as follows. Given the total countries’ GDPs, the effect of average GDP per capita on economic trade is negative in the countries of the former Soviet Union. Or in other words, given the total GDP, the more populous the country is, the larger the international trade.

In any case, the explanatory effect of average GDP per capita on trade is relatively small (it improves R² by 0.019 only) but statistically significant (see Appendix B). In 2019 (the year was chosen as the last year before the COVID-19 pandemic and the latest year with the biggest number of observations), the inclusion of GDP per capita in the regression model increased R² from 0.837 to 0.856, thus explaining 1.9% of the variance. The similar analyses for the data from other years in the sample provided similar outcomes and are omitted here for the conciseness of the paper.

Overall, out of eight tested hypotheses, five were confirmed, in two cases, we did not find any significant effect, and in one case, the effect was opposite to what we expected. Namely, we showed that the volume of international trade is positively related to the joint economic size of the countries measured by multiple countries’ GDP (H1), negatively related to the distance between the countries (H2), and positively related to the availability of infrastructure of both countries (namely the density of roads, H4), negatively related to the difference in the availability of infrastructure in the trading countries (namely the difference in the density of roads, H5), negatively related to the wars and other military conflicts in the countries (H6). The effects of difference in economic performance (the biggest trade is in the countries that are economically approximately at the same level, H8) and common border (H3) were not statistically significant.

Contrary to expected, GDP per capita was negatively associated with the volume of trade (H7). We explained this result as the outcome of the mediation effect, where the mediating variable proved to be the total GDPs of the coupled countries combined. We suggest two explanations. First, the negative association between the volume of trade and average GDP per capita might reflect the fact that more economically developed countries are more likely to trade outside of the region. Second, the negative association between the volume of trade and average GDP per capita might be related to the idea that GDP per capita is total GDP divided by the population. Thus, if GDP is kept constant, the negative effect of GDP per capita on the volume of trade might simply reflect the fact that more populous countries trade more.

6. Discussion

The world is going through unprecedented economic turmoil caused by the COVID-19 epidemic-related crisis, international political crises, and the disruption of international production links and trade. Under these conditions, the factors affecting trade are elucidated more clearly, similarly to the factors affecting international food, energy, and other sectors’ balances. In this regard, the post-Soviet Union region, originally integrated into world production and trade, loses its trade links to the rest of the world and within the region.

These processes are very unfortunate from the economic point of view, as sustainable international trade was frequently viewed as a driver of the economic growth of all the countries involved [13]. It also emphasizes the need to study the influence factors that enhance sustainable international cooperation.

This paper researched the factors affecting sustainable (at least ten years as presented in data) international trade within the geographically closed region of post-Soviet Union Countries. Though the countries experienced increasing dissolution from the common center (Russia), the trade flows seem to correspond to the established theories—the size of the economies, the infrastructure, and the common language significantly affected the volume of bilateral trade. The difference in infrastructure (the density of roads), the military conflicts, and the geographical distance negatively impacted the trade flows. Interestingly, the common border was not statistically significant.

The results evoke some ideas about the future tendencies for international trade development in the region and about the future reconstruction of Ukraine after the military conflict with Russia ends. Obviously, infrastructure will be important not only for Ukraine but also for the international food market.

This study presents several innovation points with respect to innovation the international trade presents, the innovation in international trade, and with respect to existing literature. First, international trade increases innovations in four ways: it increases the market size, improves competition, provides opportunities to utilize comparative advantage, and improves knowledge spillovers [74]. Thus, factors affecting international trade improve the overall innovation and economic performance. Second, any innovations in international trade bear consequences for not only the trade itself but also the overall institutions and wellbeing of the countries. For example, innovative green international trade tends to reduce ecological footprints, though the effect on economic wellbeing is unclear [75]. Third, our results partly correspond and partly differ from existing literature. Similar to [76], our results suggest that being close to other developed countries with good economic, infrastructural and institutional frameworks positively affects international trade. Long-term policies seem to play an important role. On the other hand, surprisingly, it was not important whether the countries had a common border.

The paper presents a study of one region only. Contrary to the World-Based general equilibrium approach based on primarily world-level general economic effects, as described in [77], this paper suggested that path dependencies, such as post-Soviet legacy, play a role in international trade. We concentrated exclusively on the processes within the block of post-Soviet Union Countries. However, from the perspective of [77], it would be interesting whether the tightness of the trade relations within the post-soviet countries still makes them serve as a coherent trading region. We suggest this topic for further research.

7. Conclusions

Sustainable international trade is one of the essential factors enhancing sustainable economic growth and overall wellbeing [13,78]. This paper studied the factors predicting the volume of international trade in the post-Soviet Union countries sustainably throughout 2010–2020. Though the countries have grown further apart over the last thirty years, the patterns of international trade correspond to theoretical predictions.

We conducted a set of spatial gravitation models for 2010–2020, one moderation analysis (to explain the unexpected relation between GDP per capita and the volume of trade, Appendix B), and a hierarchical regression analysis.

The results suggest that the volume of trade in the post-soviet area is positively related to the combined GDPs of the couped countries, adopting the Russian language as the official language in at least one of the countries, and the average density of road networks. We found that the volume of trade is negatively related to the proximity of the countries, the difference in the density of road networks, the index of ongoing conflicts, and, surprisingly, GDP per capita. As the correlation between the volume of trade and GDP per capita was positive, we looked for the moderating variable that switched the sign of the coefficient. The moderating variable proved to be the joint GDP of the countries. We suggest two explanations. First, more economically advanced countries might be more likely to redirect their international trade outside of the region of post-soviet countries. Second, the unexpected result possibly reflects the effect of the population of the country as given GDP, more populous countries are likely to trade more.

The results above gain increasing meaning in the context of current international affairs. The political disarray led to the military conflict between Russia and Ukraine, thus endangering sustainable international trade in food products and overall food sustainability. Ukraine, which used to be the fifth largest wheat exporter in the world in 2020, with exports amounting to USD 4.61B in wheat (OEC, 2022) [12], lost its exporting power. Given the military conflict, price increase, and lack of resources affecting agricultural production, the export potential of Ukraine in the future years will be compromised. Our results suggest that infrastructure development is the most important measure to regain export potential. It will be important from the country’s point of view and as a factor in increasing the sustainable supply of food products in the international arena.