**The United States' Clothing Imports from Asian Countries along the Belt and Road: An Extended Gravity Trade Model with Application of Artificial Neural Network**

#### **Danny Chi Kuen Ho 1,\*, Eve Man Hin Chan <sup>2</sup> , Tsz Leung Yip <sup>3</sup> and Chi-Wing Tsang <sup>4</sup>**


Received: 18 August 2020; Accepted: 7 September 2020; Published: 10 September 2020

**Abstract:** In 2013, China announced the Belt and Road Initiative (BRI), which aims to promote the connectivity of Asia, Europe, and Africa and deepen mutually beneficial economic cooperation among member countries. Past studies have reported a positive impact of the BRI on trade between China and its partner countries along the Belt and Road (B&R). However, less is known about its effect on the sectoral trade between the B&R countries and countries that show little support of the BRI. To address that gap, this study examines the changing patterns of clothing imports by the United States (US) from China and 14 B&R countries in Asia. An extended gravity model with a policy variable BRI is built to explain bilateral clothing trade flow. A panel regression model and artificial neural network (ANN) are developed based on the data collected from 1998 to 2018 and applied to predict the trade pattern of 2019. The results show a positive effect of the BRI on the clothing exports of some Asian developing countries along the B&R to the US and demonstrate the superior predictive power of the ANN. More research is needed to examine the balance between economic growth and the social and environmental sustainability of developing countries and to apply more advanced machine learning algorithms to examine global trade flow under the BRI.

**Keywords:** clothing trade; Belt and Road initiative; gravity trade model; panel data regression; artificial neural network

#### **1. Introduction**

Textiles and clothing industries have been driving the economic growth and development of low-income and developing countries like Bangladesh and Cambodia through improved trade, gross domestic product (GDP), employment, and foreign currency receipts [1]. As the United States (US) is the world's second biggest clothing market, any major changes in its GDP and trade policy would significantly affect clothing trade flows. In parallel, China, as the world's biggest clothing supplier, has a key role in shaping global clothing trade patterns. Worth noting is the roll-out of China's Belt and Road Initiative (BRI). It is a development strategy proposed by China in 2013 that aims to promote the connectivity of Asia, Europe, and Africa and to deepen mutually beneficial economic cooperation among member countries [2].

Despite the lack of US support and commitment to the BRI, any study of recent US–Asia bilateral clothing trade should not ignore the influence of this initiative, as promotion of unimpeded trade is a priority for the BRI. Given China's dominant role in the production and export of clothing products, major changes brought by the BRI will shape the sources of supply and patterns of global trade over time. With the establishment and improvement of trade-supporting infrastructure like power plants, highways, ports, and industrial and logistics parks in developing countries along the Belt and Road (B&R), new sources of clothing supply would emerge. Relocation of clothing factories from China to these countries could grow to take advantage of the relatively lower labor costs and improved infrastructure for trade facilitation. In this way, a win-win situation may be achieved. On one hand, developing countries could benefit from the expansion of their clothing sector, which contributes to export-led economic growth. On the other hand, countries with a large demand for clothing products could have more choices of supply. It is thus important to examine how clothing supply from China and other Asian countries has changed in the US market in the context of the BRI.

The current study's objectives are to (1) develop an extended gravity model to predict clothing imports of the US from China and 14 Asian countries under the BRI and to (2) compare the model's predictive power by panel data regression and artificial neural network (ANN) in the US's clothing imports from 1998 to 2019. This study is valuable as it contributes to the literature on global trade on two fronts. First, it addresses an important yet under-researched area of bilateral trade under the BRI. Although more empirical studies have focused on trade along the B&R, they tend to examine trade flows between China and its trading partners at the country level (e.g., [3–5]) and not trade between B&R and non-B&R countries at the sectoral level. As the BRI aims to promote unimpeded trade through better connectivity of infrastructure and facilities across geographical boundaries, developing countries that have joined the BRI would have a chance to build stronger links to global value chains that connect to high-profit markets that do not necessarily have to be part of the B&R region (e.g., the US). Improving trade not only within but also beyond the B&R region is particularly important for labor-intensive sectors like clothing because more jobs could be created for female workers and their welfare could be improved. Moreover, the entry barriers to the market are relatively lower than those of the industries that demand high-skill labor, advanced technologies, and large capital investment (e.g., new energy automobile). Despite the importance of integrating into global value chains and getting more orders from foreign buyers, little is known about the potential impact of the BRI on improving developing countries' exports to high-profit markets. The current study aims to fill this gap.

Second, this study applies a novel approach to ANN to analyze bilateral trade flows and demonstrates how ANN complements the conventional econometric approach. The gravity trade model is frequently used to explain global clothing trade patterns [6], and econometric models are built to fit the data. Most often, multiple linear regression of panel data is applied to examine the relative influence of various economic factors, such as a country's GDP and trade policy, on bilateral trade. More recently, advances in big data availability and affordable high computing power and online platforms have made ANN more accessible for researchers. The use of ANN in this study is relevant and useful not only because of its higher predictive power but also because of its ability to estimate complex trade relationships [4,7]. Although more Asian developing countries have joined the BRI, India is an exception as it has concerns about the expansion of Chinese political influence and interests across South Asia through the BRI [8]. The official Indian narrative of the BRI is not positive, and India's perceptions have been mainly shaped by geopolitical dimensions of the BRI rather than broader developmental aspects [9]. Moving beyond this one-sided view, it would be helpful to explore the BRI's effect on India's clothing exports if India would become a B&R country. To achieve the second objective, the study will develop a model of ANN based on the results of panel regression analysis and evaluate the two approaches based on the unseen data of 2019 exports values. Their predictive performance will be compared with reference to the models' forecast errors. Furthermore, a country's clothing exports can be estimated by the ANN when its B&R membership is changed (e.g., India

becomes a B&R country). This helps to explore the potential impact of the BRI on the exports of B&R and non-B&R countries.

The paper is structured as follows. Section 2 discusses Asia's clothing exports to the US under the BRI. Section 3 presents a literature review with a focus on a gravity model for trade estimation. Section 4 presents the methodology. Sections 5 and 6 present the findings and discuss the panel data regression model and ANN results, respectively. Finally, Section 7 concludes the study with implications for policymakers and future research directions.

#### **2. The BRI and Clothing Trade**

#### *2.1. Asia's Clothing Exports under the BRI*

Among Asian countries, China has been a leading clothing manufacturer and exporter since the nineties [10]. However, rising production costs and labor shortages in China have led many clothing manufacturers to relocate their labor-intensive production facilities from China to other, lower-cost, countries in the region such as Vietnam [11], Bangladesh [12], Cambodia [13], and the Philippines [14]. The BRI may present opportunities for many businesses to overcome some of the barriers to and risks of relocation. One of the BRI's major outcomes is infrastructure development across the "Silk Road Economic Belt" and "21st Century Maritime Silk Road", which helps to speed up product flows and provide efficient allocation of resources across markets. Improved connectivity of infrastructure and facilities can promote unimpeded trade across geographic boundaries, which are two cooperation priorities of the BRI.

Taking inspiration from the name and purpose of the ancient Silk Road connecting China and Europe for silk trading, the proposed economic corridors of BRI could bring opportunities and challenges to China, developing countries along the B&R, and their trading partners. In the six years since the launch of the BRI, China has signed 171 cooperation documents with 29 international organizations and 123 countries, and the total trade value between China and the B&R countries and regions has exceeded \$6 trillion USD from 2013 to 2018 [15]. In Asia, a growing number of countries have officially pledged support to the BRI by memorandums of understanding (MoU) or joint statements/communiques since 2013 (See Table 1 for the sampled countries).


**Table 1.** The year that the sampled countries joined the Belt and Road Initiative (BRI).

\* Source: Belt and Road Portal (eng.yidaiyilu.gov.cn).

China has been investing heavily in some mega infrastructure projects under the BRI, such as the Bangladesh-China-India-Myanmar Economic Corridor, a Sri Lankan port city, and an Indonesian high-speed railway, which are all designed to facilitate international trade. In 2020, China signed a number of new BRI infrastructure projects across Asia, including the construction of a railroad and deep-water port in Myanmar, a wind power plant in Vietnam, a biomass plant in Indonesia, and several railway projects across Africa [16]. With its implementation in full swing since 2015 [17] and as an ongoing endeavor, the BRI will continue shaping the global trade of different commodities and products including textiles and clothing.

It appears that the BRI benefits not only China but also developing countries that get the most inflows from foreign direct investment (FDI). For the clothing industry in Asia, the BRI could offer potential trading and expansion opportunities, where businesses with production facilities in China could be relocated to lower-cost B&R countries in Asia. It is worth mentioning that, since 2015, Vietnam's textile and clothing industry has witnessed a significant increase in FDI from South Korea (a B&R country) and the Greater China region (China, Hong Kong, and Taiwan), which injected more than tens of billions dollars in total to expand the production capacity in Vietnam [18]. The establishment of clothing production facilities in the regional B&R countries could boost their economic development by creating more jobs and improving labor welfare. Most importantly, these B&R countries could take the opportunity to build stronger links to global clothing supply chains and pursue export-led economic growth.

#### *2.2. The US's Clothing Imports under the BRI*

The US is the world's second largest clothing importer after the European Union (EU). The US's clothing imports have been growing overall, reaching a record high of 85.2 billion USD in 2015 (see Figure 1) [19]. In 2019, the US imported 83.8 billion of USD clothing products from the world, representing a 74% increase from 48.2 billion USD in 1998. Asia has been a major clothing supplier for the US market by value, with China as the biggest exporter, followed by Vietnam, Bangladesh, Indonesia, and India (see Figure 2).

**Figure 1.** The US's clothing imports from the world.

Although China has outranked Mexico since 2003 and become the biggest clothing supplier in the US market, its export started to fall after attaining a record high of 30.5 billion USD in 2015. A closer examination of the annual change of US clothing imports (see Figure 3) reveals that despite this, 2016 witnessed a 5.3% reduction in the US's annual clothing imports from the world and China's exports to the US dropped significantly by 8.7%. This pattern is also observed in 2017, where the US experienced a very small drop of 0.6% in its total clothing imports but China's exports to the US dropped by 3.2%. This pattern is in sharp contrast to Vietnam's clothing exports to the US. In 2016 and 2017, even when the US's total clothing imports dropped, Vietnam still attained an annual growth of 2.2% and 7% in its exports, respectively. This shows that Vietnam is able to expand its production capacity and capture a higher market share in the US, while China's clothing exports have been reducing from 2015 onward.

**Figure 2.** The US's clothing imports from the top five Asian suppliers.

**Figure 3.** Annual change (in percentage) of the US's clothing imports from the world and the top 5 Asian suppliers.

#### **3. Literature Review**

In this section, the theoretical framework of the gravity model for trade, recent studies using the gravity model for analysis of developing countries' textiles and clothing trade, and the configuration of an extended gravity model for clothing trade under the BRI are presented.

#### *3.1. Theoretical Framework of Gravity Trade Model*

The gravity model is the workhorse of the applied international trade literature. It has been frequently used to evaluate the impacts of various trade-related policies and factors [20], starting with Tinbergen [21] and Poyhonen [22], who found that the volume of trade between two countries is directly related to their economic size and inversely related to the geographical distance between them. In other words, countries with a larger economy tend to trade more, and greater distance, which is a

 × 

= A

proxy of transportation costs, hampers bilateral trade. The basic gravity model is represented by the following equation:

$$\mathbf{Y}\_{ij} = \mathbf{A} \frac{\mathbf{X}\_i \times \mathbf{X}\_j}{D\_{ij}} \tag{1}$$

where

*Yij* = Total value of trade between countries *i* and *j* A = Constant *X<sup>i</sup>* = GDP of country *i X<sup>j</sup>* = GDP of country *j*

*Dij* = Distance between country *i* and country *j*.

Anderson [23] provided a theoretical explanation for the gravity equation applied to commodities using a trade-share-expenditure system model. Later, Bergstrand [24] developed a microeconomic foundation for the gravity model and found empirical evidence supporting the proposition that the gravity equation is a reduced form of a partial equilibrium subsystem of a general equilibrium model with nationally differentiated products. Deardorff [25] showed that the gravity equation can be derived from the classic Heckscher–Ohlin model and is consistent with other trade models such as the Ricardian model. Evenett and Keller [26] evaluated gravity equations based on the imperfect specialization of production and found support from the increasing returns theory and Heckscher–Ohlin model. With solid theoretical foundations, the gravity model has been applied extensively in empirical studies of international trade.

#### *3.2. Empirical Studies of Gravity Model for Developing Countries' Textiles and Clothing Trade Analysis*

In the literature of sectoral trade, the gravity model has been applied to examine bilateral trade of textiles and clothing (e.g., [6,20,27–31]). The results of these studies support the proposition that greater GDP facilitates trade, whereas longer distance reduces trade. Depending upon the research objectives, past studies have developed extended (also called augmented) gravity models [32], which include (1) economic variables like the gross national product (GNP), per capita GDP, per capita GNP, consumer price index, FDI, rate of inflation, exchange rate, and membership in a free trade area; (2) geographical variables like common borders, landlocked, remoteness, land area, transport time, time difference, population size, and population growth; (3) social variables like common language, religion, and literacy rate; and (4) political variables like colonial link and political stability, among others [33].

As the textiles and clothing trade represents a major driver of economic growth for developing countries, a growing number of gravity trade model studies have focused on export countries like Bangladesh, India, Indonesia, and Pakistan. For example, Rahman et al. [34] examined a panel gravity model of Bangladeshi textiles and clothing export flows to 40 trade partners from 1990 to 2017 and found that GDP, per capita GDP, and real exchange rate of the importers as well as Bangladesh's WTO membership have a strong effect on Bangladesh's textile exports. Majeed et al. [35] found a positive impact of the EU's and the US's generalized system of preferences on Pakistan's exports of cotton and textile products to these markets from 2003 to 2014. Irvansyah et al. [36] examined Indonesian's exports of textiles and clothing products in key markets like the US, Japan, South Korea, and Turkey, whereas Chakrabarty et al. [37] focused on knitwear clothing exports from India to the US.

#### *3.3. Configuration of an Extended Gravity Model for Clothing Trade under the BRI*

Empirical studies that apply the gravity model to examine trade at product and sectoral levels under the BRI are growing. For example, based on the estimation of an extended gravity model using trade data at product-level during 2002–2016, Liu et al. [5] reported that cultural distance and institutional distance inhibit China's bilateral trade with the B&R countries. Zhang et al. [38] found positive impacts of trade facilitation on China's forest product exports to 13 B&R countries using transnational panel data from 2007 to 2016. Leng et al. [39] reported that China's wind energy product trade with the B&R countries has grown rapidly. Shahriar et al. [40] applied a commodity-specific gravity model to study China's meat exports to 31 trading partners from 1997 to 2016 and found a positive impact of the BRI on China's exports. Despite these studies having examined different products, they have the same focus on China's trade with the B&R countries. Less is known about the trade of B&R countries (other than China) with non-B&R countries like the US. To address this research gap, this study develops an extended gravity model featuring a policy variable BRI, which is expressed as the following log-linear equation:

$$\begin{array}{l} \log\Big(\mathsf{LISimport}\_{i\mid i}\Big) = \alpha + \beta\_1 \log\Big(\mathsf{GDP}\_{i\mid i} \times \mathsf{GDP}\_{j\mid i}\Big) + \beta\_2 \log\Big(\mathsf{D}\_{i\mid}\Big) +\\ \beta\_3 \log\Big(\mathsf{Extrete}\_{i\mid}\Big) + \beta\_4 \mathrm{Land}lock\_i + \beta\_5 \mathsf{WTO}\_{i\mid} + \beta\_6 \mathsf{BRI}\_{i\mid} + \varepsilon\_{i\mid t} \end{array} \tag{2}$$

where

α is the intercept;

*USimportijt* is the value of clothing (in USD) imported from country *i* (i.e., exporting country) by country *j* (i.e., the US) at time *t*;

*GDPit* is GDP in USD of country *i* at time *t*;

*GDPjt* is GDP in USD of country *j* (i.e., the US) at time *t*;

*Dij* is geographical distance (in km) between the capitals of countries *i* and *j* (i.e., the US);

*Exrateit* is official exchange rate of country *i* relative to the USD at time *t*;

*Landlock<sup>i</sup>* is a dummy variable with a value of 1 if country *i* does not have direct access to sea, otherwise 0;

*WTOit* is a dummy variable with a value of 1 if country *i* has joined the World Trade Organization (WTO) at time *t*, otherwise 0;

*BRIit* is a dummy variable with a value of 1 if country *i* has joined the BRI at time *t*, otherwise 0; ε*ijt* is the error term.

In the extended gravity model, four explanatory variables, official exchange rate, landlock, WTO membership, and BRI, are included in addition to GDP and distance. The dependent variable is the US's clothing imports (in USD) from Asian countries. Exchange rate is a key factor affecting clothing trade. In general, a weaker domestic currency stimulates exports. Depreciation of the domestic currency of Asian clothing suppliers against the USD is reflected by a higher value of *Exrateit*. That is, it requires more domestic currency to exchange one USD. It is expected that the sign of this variable is positive. Landlocked countries like Laos are constrained by their geographical limitations, i.e., no direct access to sea. Higher international trade costs are incurred because they normally depend on their transit neighbors' infrastructure for getting access to foreign markets. This problem is more acute when the cargos for external trade have to transit through neighbors' seaports. It is expected that the sign of the variable *Landlock<sup>i</sup>* is negative. WTO membership is of particularly importance to the growth of Asian countries' clothing exports because all quota restrictions on textiles and clothing products among WTO members were scheduled to be removed completely by 2005, as set out in the WTO's Agreement on Textiles and Clothing (ATC). It is expected that the sign of the variable *WTOit* is positive. Given that the BRI was proposed in 2013, Shahriar et al. [40] created a dummy policy variable with a value of one assigned from 2013 onward and zero otherwise. Different from their approach, the dummy variable of BRI is assigned a value of one for the export country from the year it joined the BRI and onward and zero otherwise in this study. This coding method can better capture the BRI influence on the bilateral trade of individual countries over time. It is expected that the sign of the variable *BRIit* is positive. For the variable of GDP *GDPit* <sup>×</sup> *GDPjt* , the expected sign is positive, whereas distance *Dij* is negative.

#### **4. Methodology**

#### *4.1. Dataset*

Using the proposed extended gravity model, this study estimates the value of the US's clothing imports between 1998 and 2019 from 15 countries in South/Southeast Asia including Bangladesh, Brunei, Cambodia, China, India, Indonesia, Laos, Malaysia, Nepal, Pakistan, Singapore, Sri-Lanka, Thailand, the Philippines, and Vietnam. Despite the fact that Timor-Leste is also a Southeast Asian country, it does not trade in the clothing industry and therefore is not analyzed. In contrast to other Asian countries, Myanmar is a special case that deserves examination in isolation because of trade sanctions imposed by the US during the study period. From 2004 to 2012, no clothing imports were recorded by the US from Myanmar. Myanmar is excluded from the sample. Since China initiated the BRI in 2013, the remaining 14 Asian countries joined the BRI at different times since then except India. The data are collected from multiple sources (see Table 2). There is no missing data or trade value with zero in the dataset. The values of dependent and four continuous independent variables are log-transformed and then standardized in the pre-processing stage such that their means become zero and standard deviations become one, as these variables have different units of measurement. No transformation is performed on the dummy variables.


#### **Table 2.** Data source.

#### *4.2. Panel Data Estimation Approach*

This study conducts a regression analysis with panel data through econometric and statistical software—EViews 10. Cross-sectional or pooled ordinary least squares (OLS) regression is often used to estimate the gravity trade model. Yet, biased results may be created by these estimation approaches [41]. This is because heterogeneity is not allowed in the error term for standard cross-sectional regression equations, thus yielding overestimated results. A panel estimation method with fixed effects (FE) and random effects (RE), on the other hand, could overcome the problems created by using the OLS approach. An advantage of using the panel data estimation method is that it can increase the volume of informative data in variability with less collinearity among the variables [42], which allows more degrees of freedom and efficiency. In this study, the panel data from 1998 to 2018 is analyzed to estimate the regression coefficients with pooled OLS, FE, and RE models. Poolability F test is performed for choosing between the pooled OLS and FE models. Hausman test is performed for choosing between FE and RE models. The best regression model is then used to predict the US's clothing imports in 2019. The out-of-sample forecast error of root mean squared error (RMSE) is computed and compared with that of the best ANN.

#### *4.3. The Configuration and Implementation of ANN*

The proposed ANN has three layers: input, hidden, and output. In the input layer, there are six features (the product of exporter's GDP and importer's GDP, distance between exporter and importer, official exchange rate, landlock, WTO, and BRI), whereas there is one target (prediction of clothing imports) in the output layer. The features of ANN are selected after panel data regression analysis is completed. Predictors that are not statistically significant at *p* ≤ 0.05 are excluded. The number of

neurons in the hidden layer (i.e., hidden neurons) is optimized by building various ANNs with hidden nodes of 3 to 15 (see Figure 4). The ANN with the best predictive ability is identified by comparison of RMSE of the testing dataset with unseen data across different networks. Similar to Dumor and Yao [4], this study uses Rectified Linear Units (ReLU) as the activation function. The ANNs are trained using the stochastic gradient descent optimizer with mean squared error (MSE) as the loss function.

**Figure 4.** The proposed artificial neural network (ANN) structure.

Instead of dividing the dataset into training and validation sets in one go (e.g., [4,7]), this study applies K-fold cross-validation for training and validation of each ANN. This method provides more robust models and combats over-fitting the model [43]. The 1998–2018 dataset with 315 observations is split randomly into five groups (folds) of equal size. One group is taken as a hold-out or validation set, whereas the remaining four groups form a training set. The model is fit on the training set and the fitted model is evaluated on the validation set. The evaluation score of RMSE is retained, and the model is dropped. This process is repeated five times. The mean of the five RMSEs are calculated for each trained ANN. The 2019 dataset with 15 unseen observations is used for testing of each trained ANN. That is to predict out-of-sample observations. The training dataset is divided into 32 batches, and 200 epochs are set to train each ANN with a learning rate of 0.01. The Keras Sequential model is used to implement the proposed ANNs in Python. The ANNs are created and trained in the Jupyter notebook environment on Google platform.

#### *4.4. Measures of the Model's Predictive Ability*

To examine the predictive power of gravity trade model, the conventional econometric analysis and the new approach of ANN are applied. Consistent with past studies (e.g., [4,7]), the prediction accuracy of regression model for panel data is measured by two metrics in this study: the coefficient of determination (R<sup>2</sup> ) and the RMSE. The magnitude of R<sup>2</sup> indicates the proportion of the variance in the clothing imports that is predictable from the independent variables. The higher the R<sup>2</sup> , the better the model fits the data. RMSE is the square root of the MSE, which is the average of squared errors between the predicted values and the actual values of clothing imports:

$$RMSE = \sqrt{MSE} = \sqrt{\frac{\sum\_{i=1}^{n} \left(\hat{\mathbf{y}}\_{i} - \mathbf{y}\_{i}\right)^{2}}{n}} \tag{3}$$

where *Y*ˆ *i* is the predicted export value, *Y<sup>i</sup>* is the actual export value, and *n* is the number of predicted export values. A smaller RMSE indicates higher predictive power of the model. This study compares the prediction performance of regression analysis and ANN by RMSE.

#### **5. Findings**

#### *5.1. Results of Panel Data Regression Models*

The results of the pooled OLS and year-FE models are shown in Table 3. The result of poolability test favors the year-FE model over the pooled OLS model (F(20, 288) = 4.29, *p* < 0.0001). And the result of Hausman test favors the year-FE model over the year-RE model (χ 2 (4) = 28.6, *p* < 0.0001). The year-FE model explains 74.01% of variance of the US's clothing imports.


**Table 3.** Results of panel data regression models.

Note: \*\*\* *p* < 0.001 and \* *p* < 0.05.

The sign of predictor coefficients of the year-FE model is consistent with expectation. The six predictors contribute significantly to the model, as the *p*-value of regression coefficients is smaller than 0.0001. As expected, larger GDP of both the US and Asian countries contribute to higher bilateral clothing trade (β<sup>1</sup> = 0.5256), whereas longer distance between them hampers the bilateral clothing trade (β<sup>2</sup> = <sup>−</sup>0.1307). Depreciation of domestic currency of Asian countries against USD promotes their clothing exports to the US (β<sup>3</sup> = 0.4048). However, the landlocked country (Laos in the sample) is disadvantaged in its clothing exports to the US (β<sup>4</sup> <sup>=</sup> −1.1717). The clothing exports of Asian countries grow more after they have joined the WTO (β<sup>5</sup> = 0.4038). The same pattern is observed after the Asian countries have joined the BRI (β<sup>6</sup> = 0.6539). The year-FE regression model attains the RMSE of 20.85 billion USD in the prediction of out-of-sample clothing imports in 2019.

#### *5.2. Results of ANNs*

As shown in Table 4, the mean values of RMSE decrease in the training and validation sets as expected when the number of neurons in the hidden layer (i.e., hidden neurons) increases. However, when the hidden neurons exceed 10, the predictions in the testing set become less accurate, as indicated by the rise of RMSE (>0.1824). The best model is identified when the ANN has 10 hidden neurons because it has attained the best prediction of out-of-sample clothing imports in 2019 with RMSE of 0.1824 (i.e., z-score on the transformed scale) or 2.29 billion USD.


**Table 4.** Results of ANNs.

#### **6. Discussion of Results**

The regression result shows a significant positive association between the BRI and Asian countries' clothing exports to the US. With the BRI as an ongoing endeavor in which more infrastructure projects are launched and completed and business opportunities continue to materialize, developing countries along the B&R can enhance their attractiveness for FDI in trade-led manufacturing and improve their competitiveness in global trade. The past few years have witnessed a growth in FDI from China injected into the textile and clothing industry in Asian countries including Cambodia, Bangladesh, and Vietnam [32,44]. This trend of relocation of clothing production has driven higher exports from these countries to the US.

Worth mentioning is the losing out of India to Bangladesh in clothing exports in the US market since 2008. Although India and Bangladesh are neighboring countries, their responses to the BRI are different—India has not signed a B&R MoU, whereas Bangladesh is a signatory country of the BRI. In the sample of this study, India is the only non-B&R export country. It is relevant to examine to what degree India would benefit from joining the BRI and, in particular, whether it would improve its clothing exports. The results of the ANN and panel regression analysis show that ANN has higher predictive power, as reflected by their RMSE (2.29 vs. 20.85 billion USD). ANN is applied to examine the change of India's clothing exports if it becomes a B&R country. That involves three steps. The first is to estimate India's exports value based on the unseen, real data of the six features (independent variables) in 2019. The policy variable BRI is coded as zero because India has not joined the BRI. The second step is to estimate India's exports value using the same dataset except that the value of the BRI variable is changed from zero to one. That is to reflect the change of India's B&R membership. The last step is to compare the two forecasted exports values. If there is an increase in exports, there is a potential for India to catch the trade development opportunity after joining the BRI. The ANN predicts that there is a 13.27% increase in India's clothing exports to the US when India becomes a B&R country.

To gather further support for the potential effect of BRI on trade development, the same analysis is performed on three key Asian clothing exporting countries, Bangladesh, Vietnam, and Indonesia. The unseen, real data of 2019 is used. In step one, the value of the BRI policy variable is coded as one because these countries have joined the BRI, whereas in step two, that value is changed from one to zero to reflect the disconnection of these countries with the BRI. In step three of the forecasts comparison, we see that if there is a reduction in exports, these countries would be economically disadvantaged if they cancel the B&R membership. The ANN results show a reduction of 5.38% in Bangladesh's clothing exports to the US when Bangladesh is no longer a B&R country. Similarly, if Vietnam and Indonesia drop the BRI, the reduction in their clothing exports is predicted to be as high as 40.58% and 30.37%, respectively, by the ANN.

Although the above scenarios are hypothetical, both ANN and regression results indicate the potential positive effect of the BRI on clothing exports of some Asian developing countries in the US market. To fully realize the BRI's potential in improving economic growth, developing countries need to enhance geographic, social, and economic factors for trade facilitation. For example, the distance between China and 62 B&R countries in geography (relative geographic distance), factor endowment (capital-to-labor ratio), culture (power distance, uncertainty avoidance, individualism-collectivism, and masculinity-femininity), and institution (measured by the World Bank's Worldwide Governance Indicators) have been found to affect China's exports from 2007 to 2016 negatively [45]. China's trade

agreement partnership and the BRI improve China's exports to 216 partner countries from 2010 to 2015 [46]. The connectivity of 30 B&R countries with China in policy coordination, facilities connectivity, unimpeded trade, financial integration, and people-to-people bonds have been found to contribute to their economic growth [47].

Future studies should expand their focus from the bilateral trade between China and the B&R countries to how developing countries can harness the BRI fully to pursue sustainable development through improving exports to high-profit markets in non-B&R countries like the US and Japan. In these studies, key issues of social and environmental sustainability should be addressed. Of particular importance is that the infrastructure projects funded under the BRI for trade facilitation should not be used intentionally or unexpectedly to fuel South-South competition, driving a new race to the bottom among developing countries along the B&R. That is, to attract FDI in labor-intensive manufacturing industries through improved trade-supporting infrastructure on one hand, and to secure orders from foreign buyers at the expense of local labor welfare through inadequate labor protections on the other hand [48,49]. More research on effective policies and measures, such as trade agreements with social clauses or provisions, that improve labor well-being of developing countries in the B&R context is needed.

Environmental degradation in the form of consumption of dirty energy, release of toxic chemical waste during production, and greenhouse gas emissions, among others, have been major concerns of buyers in developed countries and have growing impacts on the restructuring and operations of global clothing supply chains [50]. Developing countries along the B&R should be cautious about adopting the "pollute first, clean up later" growth strategy [51], which could result in permanent damage made to the natural environment and society that cannot be recovered fully even at high costs. A study of carbon emissions induced by exports and imports between B&R countries shows that China has become a pollution haven for 22 developed countries, and 19 developing countries have become China's pollution havens [52]. Future studies should identify a role model and examine effective mechanisms that developing countries along the B&R can follow and apply to strike a balance between economic growth and environmental sustainability.

Regarding the application of ANN on sectoral trade analysis, unlike past studies that have employed a large dataset (e.g., 4536 observations in Dumor and Yao [4] and 91,094 observations in Wohl and Kennedy [7]), only 315 observations (15 countries × 21 years) are used for the training and validation of ANNs in this study. Despite that, ANN has outperformed linear regression model in predictive performance of the US's clothing imports and corroborated results of past studies. ANN has great potential for use as an alternative method to predict bilateral trade. Without doubt, training a neural network with large datasets helps to avoid overfitting and generalize better. Yet, in some cases, due to various constraints, only a small dataset can be obtained. Future studies should explore using advanced algorithms of machine learning to achieve more accurate predictions with small datasets.

#### **7. Conclusions**

This study has expanded the empirical literature of global trade under the BRI. Different from past research that examined bilateral trade between China and the B&R countries, this study focuses on bilateral clothing trade between the US and 15 Asian countries along the B&R. An extended gravity model with a policy variable of BRI has been established to explain the clothing trade pattern from 1998 to 2019. Drawing upon the results of panel data regression and ANN, this study has two conclusions. The first is that there is a positive effect of the BRI on the clothing exports of some Asian developing countries in the US market. This finding is important because it supports the notion that the BRI could bring trade opportunity to developing countries not only by improving their bilateral trade with China, which has been revealed by past studies, but, more importantly, by enhancing the B&R countries' exports to non-B&R countries, such as the US, as shown in this study.

The second conclusion is that ANN outperforms a regression model in the prediction of the clothing exports of some Asian developing countries to the US. ANN also complements the regression model in analyzing the potential impact of policy change. As shown by the ANN results, there is a potential for India to improve its clothing exports to the US by joining the BRI. Moreover, there is a chance for some B&R countries, including Bangladesh, Vietnam, and Indonesia, to experience a reduction in clothing exports to the US to varying degrees if they drop their B&R membership.

The implication of these findings for policymakers is that developing countries in Asia could improve exports performance through participating in the BRI, which brings FDI to enhance trade-supporting infrastructure and expand and upgrade local production capacity so as to build stronger and deeper connections with global value chains and secure orders from foreign customers in high-profit markets. To fully realize the BRI's potential, policymakers need to identify country-specific barriers for building links to global value chains, which could be high costs and unstable supply of energy and key natural resources, insufficient high-skill workforce, weak labor rights protection, loose enforcement of environmental regulations, inefficient customs operations, outdated transport systems, inadequate information and communication technology infrastructure, poor governance and corruption, among other factors. Policymakers need to devise appropriate policies and measures to address the problems and work in close collaboration with other B&R countries and key stakeholders to co-create value for all in the pursuit of sustainable development.

This study is limited to analyzing conventional economic factors in the gravity model. Other factors that bring uncertainty, such as trade protectionism, unstable geopolitics, and social and environmental sustainability, and dynamics that shape global clothing production and trade should be examined in future research. Researchers are advised to employ more advanced machine learning methods in tandem with the conventional econometric approach to examine theoretical models that account for global trade flows at country and sectoral levels under the BRI. That helps to enhance our understanding of the BRI's role and impact on improving connectivity and promoting trade within and beyond the B&R region.

**Author Contributions:** Conceptualization, D.C.K.H., E.M.H.C., and T.L.Y.; methodology, D.C.K.H., E.M.H.C., and T.L.Y.; software, D.C.K.H. and E.M.H.C.; validation, D.C.K.H., E.M.H.C., and T.L.Y.; formal analysis, D.C.K.H. and E.M.H.C.; investigation, D.C.K.H. and E.M.H.C.; resources, C.-W.T.; data curation, C.-W.T.; writing—original draft preparation, D.C.K.H. and E.M.H.C.; writing—review and editing, D.C.K.H., E.M.H.C., and T.L.Y.; visualization, C.-W.T.; supervision, D.C.K.H. and E.M.H.C.; project administration, D.C.K.H. and E.M.H.C.; funding acquisition, E.M.H.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was partly funded by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No. UGC/FDS25/B01/17).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Big Data-Based Assessment of Political Risk along the Belt and Road**

**Xiaohui Sun <sup>1</sup> , Jianbo Gao 1,2,\*, Bin Liu <sup>3</sup> and Zhenzhen Wang <sup>4</sup>**


**Abstract:** Political risk assessment has become increasingly important in recent years, especially with the launch of the Belt and Road Initiative (BRI) and with Covid-19 still ravaging the world. This study aims to assess systematically the political risk of BRI countries during the period from 2013 to 2019 based on three big data sets, the Global Database of Events, Language, and Tone (GDELT), China Global Investment Tracker (CGIT), and Armed Conflict Location & Event Data Project (ACLED). It is found that to properly quantify the political risks for BRI countries, the type of events, "Material Conflict", and a variable characterizing the degree of cooperation/conflicts of the events, the Goldstein Scale, are of critical importance. Based on the chosen type of events and variable, we design a normalized variable to assess political risk of any country in any year so that comparison among different countries can be meaningly made. By decomposing political risk into two components, domestic and international, and examining the spatiotemporal evolution of political risk along the Belt and Road, we find that the sum of the number of BRI countries with the extremely high level and the high level of domestic, international, and (overall) political risk all reached the peak in 2015, and decreased thereafter, and that often the level of domestic political risk along the Belt and Road was higher than the international political risk. It is also found that a strong positive correlation exists between political risk and China's total investments and construction contracts along the Belt and Road during this period. The implications of this positive correlation are discussed. The analysis presented here may help to promote the sustainable development of BRI, and be extended to examine the risks associated with foreign investments other than BRI projects.

**Keywords:** political risk; assessment; big data; GDELT; Belt and Road Initiative (BRI); China

#### **1. Introduction**

Being a critical issue of business environment, risk assessment has been a hot research topic in recent decades. Especially with the launch of the Belt and Road Initiative (BRI) in 2013 by the Chinese President Xi Jinping, a subset of the issue, political risk assessment, has become increasingly important [1–5]. BRI is also known as One Belt One Road (OBOR), aiming to increase cooperation among participating countries [6]. As of Jan 2020, 138 countries and 30 international organizations have signed BRI cooperation agreements with China [7]. China's accumulated direct investment to BRI countries has reached USD 117.31 billion during the period from 2013 to 2019 [8], accounting for 11.60% of gross flow of China's outward foreign direct investment.

Since the launch of BRI, much research has been done to properly interpret BRI [9–12], carry out case studies of international projects instigated by BRI [13,14], study China's outward foreign direct investment to BRI countries [15–17], and study the influence of BRI on concerned countries or regions [18–25]. However, research into the political risk

**Citation:** Sun, X.; Gao, J.; Liu, B.; Wang, Z. Big Data-Based Assessment of Political Risk along the Belt and Road. *Sustainability* **2021**, *1*, 0. https://doi.org/

Academic Editors: Anna Visvizi and Federico Martellozzo

Received: 28 January 2021 Accepted: 30 March 2021 Published: 1 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

facing the many countries along the Belt and Road, especially China, has not gained sufficient attention. Understandably, political risks may impact on the sustainability of BRI and cause substantial loss of investment [26,27], and thus constitute a major challenge for any country which has investments abroad. As an example, Teheran-Qom-Esfahan high-speed railway project was terminated indefinitely as a result of the sanction on Iran's nuclear program in 2013 (from the USA, Europe, and the UN), which led to a tremendous economic loss of up to EUR 8.4 billion investment from China [28]. While many countries and regions around the world are still recovering from the global financial crisis of 2008 and adjusting to changed international investment patterns, the outbreak of COVID-19, which began in early 2020 and has intensified in recent months, has tremendously exacerbated the risk of instability. Indeed, tensions, mass protests, social unrest, economic collapse, and humanitarian crises have been reported in many countries and regions around the globe [29,30], and the global GDP growth has been forecasted to drop 4.4% in 2020 [31]. These issues may have a tremendous adverse effect on the sustainability of BRI. Therefore, systematic study of political risk along the Belt and Road has not only been important, but also pressing. Since a main component of BRI can be regarded as overseas investments and cooperation in this world, thus, systematic study of political risks for BRI countries may also shed light on general overseas investments. However, quantifying political risks is a difficult issue, as political risks have many manifestations, such as social unrest, civil disturbance, riots, political instability, terrorism, and even wars [32–34]. This difficulty motivates us to explore a big data-based assessment of political risk for BRI.

There is a vast literature on the assessment of political risk. While most research in the field focused on conceptual thinking [35,36], developing some rating indices [37,38], and quantifying risks based on small data [5,39–42], major efforts have yet to be made to use big data for assessment of political risk. Recently, an important and illuminating step has been taken from a research group based on big data using spatial statistical analysis [43]. The big data they have used is called the Global Database of Events, Language, and Tone (GDELT), one of the most comprehensive data set regarding news report in the world. GDELT has many advantages to make it valuable for analyzing political risk. In particular, GDELT has been covering almost all news about the events occurring in the world since 1979, in over 100 languages. By now, the number of events covered has exceeded 600 million, and the database is updated every 15 min. Each event has two actors, such as country A and country B (for example, the Nagorno-Karabakh Conflict between Azerbaijan and Armenia which started in 1988 and recurred recently). One of the most important and interesting attributes of the GDELT event data is that each event is assigned a number, called Goldstein Scale, which is in the range of −10 and 10 and quantifies the degree of conflict or cooperation between the two actors of the event. In Zhang et al.'s work [43], they basically used the number and location of four types of events, assault, protest, coerce, fight, as the proxy of political instability, social unrest, lack of democracy, and external conflict, to assess political risk of BRI countries. While enlightening, they produced some intriguing results, such as the level of political risk in Russia (particularly in Moscow and North Caucasus) is almost as high as that in Syria in recent years (more precisely, from Oct 2013 to May 2018). Is the political risk in Russia really this high, or the observation is due to some factors, such as Russia has been active in world affairs in recent years, and is thus rich in news?

To resolve the above and other issues, we will consider systematically how to assess political risk by using GDELT and other big data. More concretely, we aim to assess systematically the political risk along the Belt and Road during the period from 2013 to 2019, based on big data comprising GDELT, the China Global Investment Tracker (CGIT), and the Armed Conflict Location & Event Data Project (ACLED). We will focus on two important questions: (i) How can political risk of BRI countries be properly assessed? (ii) Are China's BRI investments and construction contracts largely in BRI countries with low levels of political risk? If not, what are the general characteristics of political risks associated with China's BRI investments and construction contracts?

In making efforts to answer the above questions, we made six contributions: (1) In trying to resolve why political risk measured by the number of events for "Protest", "Coerce", "Assault" and "Fight" in Russia is so high, we find that the basic reason is that the number of the type of events that are chosen for evaluating risks may be correlated with the total number of events that is covered by GDELT for a country, and the number of events can vary substantially for a country over time and among different countries around the globe in a fixed (short) time interval. For example, when international affairs are concerned, the more active a country is, the more news reports the country will get. In GDELT, the number of events is roughly proportional to the number of news reports. Realizing this, one can readily understand why Russia has the large number of events belonging to "Protest", "Coerce", "Assault" and "Fight"—this is because Russia has been very active in world affairs in recent years; in fact, Russia has the largest number of events among all the BRI countries. (2) Aiming to represent more pertinently and more comprehensively the events that may directly affect foreign investment, we select a new class of events called "Material Conflict" coded in GDELT, which consists of "Exhibit Force Posture", "Reduce Relations", "Coerce", "Assault", "Fight" and "Use Unconventional Mass Violence". (3) To facilitate comparison among countries that may have vast differences in national capabilities, geographical characteristics, cultural background, etc., we design a normalized quantify, the ratio between the sum of the Goldstein Scale of "Material Conflict" events and the sum of the Goldstein Scale of all the events. Clearly, using the Goldstein Scale of the events is more advantageous than directly using the number of events, since an event with the Goldstein Scale of −10 amounts to 10 events with the Goldstein Scale of −1. (4) To assess which type of political risk a BRI country is facing, we study domestic and international political risk which are two components of the political risk. (5) We examine the spatiotemporal evolution of political risk along the Belt and Road during the period from 2013 to 2019. (6) We find a strong positive correlation between the political risk and China's investments and construction contracts along the BRI during the period from 2013 to 2019. While this is quite the opposite of the ideal case that investment goes to countries or regions with as low political risk as possible, it nevertheless corroborates a general saying that chances are often associated with risks.

The remainder of the paper is organized as follows. Section 2 contains the literature review. Section 3 explains the data and methods. Section 4 presents the assessment results about the political risk along the BRI, and examines the correlation between the political risk and China's total investment and construction contracts in a period from 2013 to 2019. Sections 5 and 6 contain the conclusions and discussions, respectively.

#### **2. Literature Review**

Since multiple risks are involved in investments by multinational firms and international projects [44], in recent years, study on the correlation between the political risk and foreign investments [4,40,45–50] and the relevance of political risk to multinational firms [51–53] and international projects [37,54,55] are gaining increasing attention. These studies are especially important for the study of China's investments in BRI countries [56]. While much research on BRI has been done, including interpretation of BRI [9–12], case studies of international project [13,14], analyses of various risks (e.g., environmental risk, energy investment risk, investment risk etc.) [5,57,58], China's outward foreign direct investment to BRI countries [15–17], the influence of BRI on concerned countries or regions [18–25], and energy efficiency and environmental quality along BRI [59–61], recently, attention has also been paid to the analysis of political risk along the Belt and Road [3,5,43]. In particular, Morris [3] has provided three dimensions for the analyses of political risk, including geopolitical level, country level, and project level, and has called for a comprehensive understanding of political risk along the Belt and Road. Hussain et al. [5] considered political risk as part of challenges that China has to consider for its investment in a host country . To assess the risks facing China's investment to BRI countries, they employed the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method and proposed an

indicator system. Zhang et al. [43] combined big data sets and spatial methods to analyze political risk of BRI countries and China's investment and construction projects.

While interesting, existing studies on big data-based assessment of political risk along the Belt and Road are limited. In fact, as many countries and regions along the BRI are unstable and vulnerable, such as Yemen, Iraq, and Afghanistan, which have been struggling with social unrest, political conflicts, terrorism and even wars, China's foreign investment towards these BRI countries surely needs serious consideration, especially the factors that could impact greatly on the business environment along the Belt and Road and the sustainability of BRI. Indeed, many overseas projects by China have not been successful. For example, the Myitsone Dam in Myanmar, which is one of China's largest electricity infrastructure projects, was terminated by Myanmar's former military government [62]. As another example, the Hambantota Port project in Sri Lanka also highlighted the importance of political risk study [63].

Although the study of the political risk along the Belt and Road is important, relevant literature however is limited, therefore, in this section, we will take a review of the general studies on political risk that may not target BRI. This general literature is quite rich, and may be divided into two lines. One line has focused on conceptual discussion of political risk [35,64], which can be classified into two clusters [35,36,65]. One cluster has emphasized government interferences with business operations in different scenarios. Robock [66] has considered "government in power and its operating agencies" as one of groups which could generate political risk, such as confiscating certain properties of international operations. Similarly, Butler and Joaquin [67] indicated political risk is unexpected changes on the "rules of game" of business operate by sovereign host government, and this government action may cause more uncertain investment consequences. Besides the element of causing political risk, government intervention in business is regarded as one of the most serious political risk effects as well [1,68]. The other cluster has considered political events as generators of political risk in international business. Adverse outcomes may arise from political events, such as wide-scale strikes, bombings, riots, violence, changes in government [35,65,66]. As political risk may affect the outcome of foreign direct investments, Clark [40] has considered "the probability of politically motivated change". Likewise, Khattab et al. [36] proposed political risk to be regarded as the probability that a political event will occur, which may cause loss for companies and other investors. Consequentially, if political risk is considered as the probability that political events will result in loss of investment, the degree of political risk will be determined by the "size" of this probability [56].

The other line of research on political risk has focused on empirical analyses of political risk. Along this line, much effort has been made to develop proper indicators for political risk. As examples, Hussain et al. [5] proposed an indicator system for environment risk based on the "Technique for Order Preference by Similarity to Ideal Solution" (TOPIS) method, which contains four sub-indices for political risk. Chang et al. [55], based on a comprehensive review of literature, identified nine categories of political risks which contain a total of 29 political risk factors and three political risk consequences. Furthermore, efforts have also been made to develop several rating indices to address political risk through multifarious variables, such as International Country Risk Guide (ICRG) model [37], and the Fragile States Index (FSI) [38]. ICRG model, launched online by the Political Risk Services (PRS) group, has built a political risk rating system comprising 12 components of political risk with different ranges of values, as a means of assessing political stability in 166 countries [37]. These 12 components of the ICRG model, including the government instability, internal and external conflict, corruption and ethnic tension, law and order, democratic accountability of government, and quality of bureaucracy, were used by Busse and Hefeker [39] to examine their effects on foreign direct investment. Another rating index, FSI, uses a three-layer system to define "Political Indicator", where the 2nd layer uses by 4 groups of variables and the 3rd layer contains 12 detailed indicators to rank countries around the world [38]. Clark [40] proposed assessing political risk as a

cost in capital budgets, to value impacts of political risk on consequences of foreign direct investment. Butler and Joaquin [67] developed a model to isolate diversifiable and nondiversifiable sources of political risk, and analyzed the effects of political risk on returns and cost of capital. In contrast to developing a model, Howell and Chaddick [49] evaluated three political risk assessment models, the Economist Method, Business Environment Risk Intelligence (BERI), and Political Risk Services, with actual losses.

While literature on the assessment of political risk is rich, major efforts have yet to be made to assess political risk based on big data. Recently, an important and illuminating step has been taken by a research group based on big data using spatial statistical analysis [43]. The significance of this work is that it showed the potential of associating political events reported in mass media with risk. This potential lines well with many earlier studies showing that political events are often associated with political risk [1,50,65,69]. As event data can now be readily generated through machine-reading from texts, such as news reports, intelligence reports, press conferences, etc., we can hope that factors relevant to political risks will be more comprehensively identified in the future. Along this line, however, it is important to be reminded of a complexity emphasized by Clark [40] that the evolution of political risk may be involved in reaction to countless events.

Therefore, systematic assessment of political risks along the Belt and Road on China's foreign investment using big data has become increasingly important.

#### **3. Data and Methods**

#### *3.1. Data*

GDELT is the major data to be used here. It includes more than 600 million distinct events across all countries, during the period from 1979 to the present, covering 20 categories and over 300 subcategories. GDELT events are drawn from a wide variety of news media, both in English and non-English, from across the world, ranging from local to international sources in nearly every country, based on the Conflict and Mediation Event Observations (CAMEO) event coding ontology [70,71]. Each event has two actors (Actor1 and Actor2), such as country A and country B (for example, the Nagorno-Karabakh Conflict between Azerbaijan and Armenia which started in 1988 and recurred recently). One of the most important and interesting attributes of the GDELT event data is that each event is assigned a set of attributes, including the interval-level Goldstein conflict-cooperation scale value [72], called Goldstein Scale, which is in the range of −10 and 10 and quantifies the degree of conflicts or cooperations between the two actors of the event. GDELT can be downloaded from https://www.gdeltproject.org/ (accessed on 16 September 2020). To compute political risks using GDELT for countries along the Belt and Road during the period from 2013 to 2019, we use the computing platform provided by the Center for Geodata and Analysis, Faculty of Geographical Science, Beijing Normal University. We use the MATLAB-R2019b and Python to identify the time, Actor1, Actor2, types of events, number, and Goldstein Scale for the events covered by GDELT, and then compute variables of interest.

CGIT is used here for the purpose of examining the correlation between the political risk of the BRI countries and China's investments and construction contracts, covering 52 BRI countries. CGIT is the only open data set covering China's global investment and construction contracts comprehensively, and has covered 3400 large transactions and 300 troubled transactions [73]. These transactions are in the areas of energy, transportation, real estate, and other industries, and contain information about Chinese parent company, host country, and sector to which the investment or project belongs.

Besides, we shall also use a database called ACLED to better determine the risks a BRI country faces. ACLED is a conflict data set which records information of internal political conflict events with dates, geographical locations, types, actors, fatalities, etc. [74], which has recorded in real-time, covering almost a million unique events in over 150 countries during a period from 1997 to 2020. As only in 2018 and 2019 ACLED covers all BRI countries, we consider a country to have fatalities of more than 10,000 caused by armed conflicts

during the period from 2018 to 2019 to be war-torn. By this criterion, Afghanistan, Yemen, Syria and Iraq are war-torn. Investments to them, and to Syria (which is not covered by CGIT), will not be considered in the correlation analysis of this study.

While 138 countries signed BRI cooperation agreements with China as of Jan 2020 [7], we select the initial 63 countries along the Belt and Road since the launch of BRI in 2013 that have news reports covered by GDELT, so that their spatial evolution of political risks in the whole period from 2013 to 2019 can be examined. According to the official website of BRI (https://www.yidaiyilu.gov.cn/jcsjpc.htm (accessed on 16 September 2020) ), these 63 countries belong to 6 major regions: Northeast Asia, Southeast Asia, South Asia, West Asia and North Africa, Central and Eastern Europe, and Central Asia. They are listed in Table 1. The country codes referred to here follow those of GDELT.

**Table 1.** Regions and corresponding countries along the Belt and Road.


#### *3.2. Methods*

3.2.1. Choosing the Proper Type of Events—"Material Conflict"

We believe choosing pertinent types of events to evaluate political risk is of critical importance. While Zhang et al. used four types of events: "Protest", "Coerce", "Assault", and "Fight" [43], we will show in the Results Section that the type "Protest" is inappropriate to be included for evaluating political risk—the simplest reason one can immediately think of is that the nature of a protest in a democratic country or a non-democratic country is completely different, and thus including "Protest" will make comparison among BRI countries that are full of democratic and non-democratic countries impossible. Besides the remaining three types of events used by Zhang et al. [43], namely, "Coerce", "Assault", and "Fight", we will also include "Exhibit Force Posture", "Reduce Relations", and "Use Unconventional Mass Violence". These six types of events constitute one of the Quadclasses, "Material Conflict", in GDELT. The other three Quadclasses are "Verbal Cooperation", "Verbal Conflict", and "Material Cooperation" [71]. "Material Conflict" contains six groups of events, "Exhibit Force Posture", "Reduce Relations", "Coerce", "Assault", "Fight" and "Use Unconventional Mass Violence". The major characteristics of these 6 groups are summarized in Table 2.


**Table 2.** The Specific elements of "Material Conflict" covered by GDELT.

3.2.2. Choosing the Proper Variable—Goldstein Scale

Choosing the proper variable to assess political risk is also important. Zhang et al. [43] directly used the number of events to measure political risk. While this seems to be a viable choice, it is far from optimal, since different events have different degrees of importance. As we will show in the Results Section, using this variable could lead to hard to interpret results, such as the level of political risk in Russia (particularly in Moscow and North Caucasus) to be almost as high as that in Syria in recent years (more precisely, from Oct 2013 to May 2018). Recognizing this, here we choose Goldstein Scale of an event as the basis of our analysis. As we mentioned, Goldstein Scale is in the range of −10 and 10 and quantifies the degree of conflicts or cooperations between the two actors of the event—the score of −10 and 10 represent the strongest conflict and cooperation. Clearly, using the Goldstein Scale of the events is more advantageous than directly using the number of events, since an event with the Goldstein Scale of −10 amounts to 10 events with the Goldstein Scale of −1.

#### 3.2.3. Designing the Proper Measure for Assessing Political Risk

Based on the chosen type of events and variable, we design a normalized variable to assess the political risks of any country in any year so that comparison among different countries can be meaningly made. Without normalization (which is the case when the number of events is directly used for evaluating risks), comparison among countries with vast differences in national capabilities, activities in international affairs, geographical characteristics, cultural background, etc., is essentially impossible. Our normalized variable is the ratio between the sum of the absolute value of the Goldstein Scale of "Material Conflict" events representing political risk and the sum of the absolute value of the Goldstein Scale of all the events:

$$PR\_t^i = \frac{|\text{GS}(M)\_{it}|}{|\text{GS}\_{it}^{(+)} + |\text{GS}\_{it}^{(-)}|} \qquad (i = 1, 2, \dots, 63; t = 2013, \dots, 2019) \tag{1}$$

In the formula, *t* is the specific year belonging to the period from 2013 to 2019, and *i* is a BRI country. *PR<sup>i</sup> t* is an abbreviation of the level of political risk for BRI country *i* and year *t*. *GS*(*M*)*it* is the sum of Goldstein Scale of "Material Conflict" events for BRI country *i* and year *t*. *GS*(+) *it* is the sum of the Goldstein Scale values of all the events with positive Goldstein Scale for BRI country *i* and year *t*, and *GS*(−) *it* is the sum of the Goldstein Scale values of all the events with negative Goldstein Scale for BRI country *i* and year *t*. There are 63 BRI countries so that *i* is in the range from 1 to 63. We neglect three BRI countries, Georgia (GEO), Romania (ROU), and Slovenia (SVN), as they have too small number of events during the period from 2013 to 2019. In fact, they are in the exponential cut-off range in terms of number of events collected by GDELT, as is clearly shown in Figure 1. In our actual analysis, we will focus on the political risk of the remaining 60 BRI countries (that is, *i* will run from 1 to 60 in contrast with the specification in Equation (1).

**Figure 1.** Sum of the number of events covered by GDELT vs. the ranking (in log-log scale) for BRI countries during a period from 2013 to 2019, where the countries are ordered according to the descending sum of number of events. Red dots are BRI countries with the sum of the number of events less than 13,000, which are Georgia (GEO), Romania (ROU), and Slovenia (SVN) respectively.

It is often thought that political risk may be better considered from external and internal perspectives [56], since countless events associated with political risk can be classified into international and national levels [40]. This is implemented by checking the actors of the events. Based on this rationale, we will also consider political risk in two components, international and domestic political risk. We hope such an approach can help better assess which type of political risk a BRI country is facing.

#### **4. Results**

#### *4.1. The Type of Events "Protest" Is Inappropriate to Be Included for Assessing Political Risk*

As we mentioned, the nature of protest in a democratic country is entirely different from that in a non-democratic country, since protest has become a mode of public participation, a regular and even desired feature of politics in established democracies [75]. Considering that BRI countries are full of democratic and non-democratic countries, including events of protest when evaluating political risk therefore will make comparison among BRI countries impossible. This realization can be better appreciated by the following two straightforward analyses. First, India has the largest number of "Protest" events covered by GDELT among BRI countries during the period from 2013 to 2019, as is presented in Figure 2. Thus, including this type of event may lead one to mis-conclude India (and other, especially democratic countries with a lot of protests) to be in turbulent and even risky situations.

**Figure 2.** Number of "Protest" covered by GDELT vs. the ranking (in log-log scale) for BRI countries during a period from 2013 to 2019, where the countries are ordered according to the descending number of "Protest". The top5 countries are India (IND), Israel (ISR), Pakistan (PAK), Russia (RUS), and Palestine (PLE).

Second and more relevant to general overseas investments, we find that there is no correlation between the "Protest" and China's investments during the period from 2013 to 2019, as is shown in Figure 3. This highlights that the event type "Protest" is not correlated with China's investments at all. Please note that even if we measure political risk by Equation (1), the correlation between the "Protest" and China's investments during the period from 2013 to 2019 is still basically zero.

**Figure 3.** Scatter-plots between the proportion of Goldstein Scale of "Protest" and China's total investments and construction contracts for BRI countries during the period of 2013-2019, where (**a**) shows that the correlation coefficient for all BRI countries is only *r*<sup>2</sup> = 0.05 (*p* = 0.736), while the coefficient is decreased to *r*<sup>1</sup> = 0 (*p* = 0.977) when the war-torn countries, Yemen (YEM), Afghanistan (AFG), and Iraq (IRQ) (which were denoted by red), were excluded. The red regression line refers to *r*<sup>1</sup> = 0. The correlation coefficient was increased to *r*<sup>3</sup> = 0.18 (*p* = 0.264) but still almost zero, when four more countries, the United Arab Emirates (ARE), Lao (LAO), Singapore (SGP), and Kazakhstan (KAZ), which were denoted by green, were also removed; this is shown in (**b**) , with the green regression line referring to *r*<sup>3</sup> = 0.18.

#### *4.2. Using the Number of Events for Assessing Political Risk Is in Appropriate*

We mentioned that using the number of events for evaluating political risk may not be appropriate. The basic reason is that the number of the type of events that are chosen for evaluating risks may be correlated with the total number of events that is covered by GDELT for a country, and the number of events can vary substantially for a country over time and among different countries around the globe in a fixed (short) time interval. If this is the case, then a country with a large number of events covered by GDELT may be mis-classified as having high risks. To appreciate the idea, we show, in Figure 4, the total number of events BRI countries have and scatter plots between the number of the four types of events used by Zhang et al. [43] and the total number of events of the BRI countries. We observed that the number of events in Russia from Oct 2013 to May 2018 is the largest among all the BRI countries, reaching 8,554,758. The underlying reason must be that Russia has been very active in world affairs in recent years, and thus must have had huge number of events reported in the news media, which in turn have been collected by GDELT. In general, we can conclude that the more active a country is, the more events the country will generate, and the more news reports it will get.

More importantly, Figure 4b showed that the number of the four types of events chosen by Zhang et al. [43] to evaluate risks is strongly correlated with the total number of events (in log-log scale). These analyses clearly indicate that directly using the number of the four types of events shown in Figure 4 is not optimal for evaluating risks. In fact, one can readily see that even if one uses other types of events to assess political risks, directly using the number of events will still be far from optimal for evaluating risks.

**Figure 4.** (**a**) Ranked total number of events covered by GDELT during the period from Oct 2013 to May 2018 for BRI countries, where the countries are ordered according to the descending total number of events, with the top3 countries being Russia (RUS), Israel (ISR), and Syria (SYR); and (**b**) scatter plots between the total number of events and total number of 4 types of events (in log-log scale), including "Coerce", "Assault", "Fight" and "Protest", with the correlation coefficient *r* being as large as 0.99 and the *p*-value being less than 10−<sup>6</sup> .

#### *4.3. Spatiotemporal Evolution of Political Risk Along the Belt and Road*

We first discuss the temporal evolution of the political risk along the Belt and Road, then examine its two components, domestic and international political risk, and finally study the spatial evolution.

By computing the political risk defined in Equation (1) and the Probability Density Function (PDF) of the political risk during the period from 2013 to 2019, Figure 5 shows that the PDF of political risk, where a mean of 0.25 is indicated by a red vertical line. The PDF suggests us to define political risk in 4 levels. Concretely, the interval 0.2 < *PR* ≤ 0.3, which contains a large probability when the political risk falls within this interval, is defined

as the moderate level of political risk for BRI countries during the period from 2013 to 2019. The interval *PR* ≤ 0.2, which also contains a large probability similar to that for the interval 0.2 < *PR* ≤ 0.3, is defined as the negligible level of political risk. Two other intervals, 0.3 < *PR* ≤ 0.4 and *PR* > 0.4, both above average but containing much smaller probabilities than other two levels, are defined as high and extremely high level of political risk, respectively. This ensures that most countries along the Belt and Road are peaceful. Therefore, we find it appropriate to divide the political risk into 4 levels, including *PR* > 0.4, 0.3 < *PR* ≤ 0.4, 0.2 < *PR* ≤ 0.3, and *PR* ≤ 0.2. These 4 levels are called *Level* 4, *Level* 3, *Level* 2, *Level* 1, which represents extremely high, high, moderate, and negligible level of political risk, respectively. This classification will also be used when discussing domestic and international political risk below.

**Figure 5.** Probability Density Function (PDF) of political risk along the Belt and Road during the period from 2013 to 2019, where the mean value of 0.25 is indicated by the vertical red line.

It is instructive to examine the temporal evolution of the number of BRI countries at different risk levels from 2013 to 2019. This is depicted in Figure 6. For the extremely high level of (overall) political risk, the number of BRI countries is 7, 9, 9, 8, 6, 6, and 5, from 2013 to 2019, respectively. For the high level of political risk, the number of BRI countries is 7, 8, 14, 12, 9, 9, and 9, from 2013 to 2019, respectively. These numbers sum to 14, 17, 23, 20, 15, 15, and 14, from 2013 to 2019, respectively. Therefore, the sum of the number of BRI countries with the extremely high and the high level of (overall) political risk reaches the peak in 2015, and decreases thereafter.

For domestic political risk, the number of BRI countries with the extremely high level is 9, 11, 12, 12, 8, 6, and 7, from 2013 to 2019, respectively, and the number of BRI countries with the high level is 7, 11, 16, 15, 12, 13, and 14, from 2013 to 2019, respectively. These numbers sum to 16, 22, 28, 27, 20, 19, and 21, from 2013 to 2019, respectively. For international political risk, the number of BRI countries with the extremely high level is 4, 8, 8, 5, 5, 5, and 3, from 2013 to 2019, severally, and the number of BRI countries with the high level is 6, 4, 7, 8, 5, 6, and 5, from 2013 to 2019, severally. These numbers sum to 10, 12, 15, 13, 10, 11, and 8, from 2013 to 2019, severally. It is thus clear that the sum of the number of BRI countries with the extremely high and the high level of political risk either for domestic or for international political risk increases to the maximum in 2015, and falls from then on. This is similar to the temporal evolution of the sum of the number of BRI countries with the extremely high and the high level of (overall) political risk.

**Figure 6.** The temporal evolution of (**a**) political risk, (**b**) domestic political risk, and (**c**) international political risk along the Belt and Road from 2013 to 2019.

As the above discussions on the temporal evolution of the sum of the number of BRI countries with the extremely high and the high level of political risk all show that the number reaches maximum in 2015 for domestic, international, and (overall) political risk, it is instructive to explore the spatial evolution of political risk along the Belt and Road by focusing on 2013, 2015, 2017 and 2019. We first study the (overall) political risk along the Belt and Road, then analyze domestic and international political risk.

The spatial evolution of the political risk along the Belt and Road is shown in Figure 7. We find that Syria, Iraq, Yemen and Palestine exhibited the extremely high level of political risk in all these four years. They are followed by Afghanistan, Lebanon, and Israel, which exhibited the extremely high level of political risk in at least two of these four years. It should be emphasized that the nature of risks these countries faced was quite different. For example, Syria's risk has been mainly caused by the continued Syrian Civil War and attacks from the Islamic State of Iraq and the Syria (ISIS, a terrorist organization designated by the United Nations). The spillover of Syrian Civil War into Lebanon (2011–2017) impacted Lebanon greatly. In Iraq, the ISIS is also the main cause of risks, which even caused the Iraqi Civil War (2014–2017) with the Iraqi Forces. In contrast, Yemen has been struggling with the Yemeni Crisis (2011–present) and the Yemeni Civil War (2014–present). In fact, when the Yemeni Civil War erupted, Saudi Arabia made an armed intervention in Yemen. As for Afghanistan from 2001 to the present, there have been many wars and attacks, such as the assaults by Taliban and the Kabul attack. Palestine has been mainly struggling with the ongoing Israeli-Palestinian conflict.

For the high level of political risk, we find that Pakistan and India of South Asia, and Myanmar of Southeast Asia showed this level in all these four years, and the political risk of Pakistan even reached the extremely high level of political risk in 2013. Saudi Arabia, Egypt, Turkey, and Ukraine showed the high level of political risk in at least two of these four years. The high level of political risk exhibited by them were also mainly caused by religious conflicts, terrorist organizations, and complex relationships between them, including other

countries' interventions. For example, the loggerheads between Saudi Arabia and Iran and the Egyptian Crisis (2011–2014) were significant causes of the high level of political risk in this region. The Internal Conflict in Myanmar since 1948 has been the longest ongoing civil war in the world, which causes a series of insurgencies. For Pakistan and India, the India-Pakistan border skirmishes (2016–2018) and the India-Pakistan standoff in 2019 were the causes of the turbulent situations.

Please note that some sudden changes in the levels of political risk occurred in the following countries, with causes readily identifiable. In 2015, the level of political risk in Central and Eastern Europe was higher than in other times. This was caused by the European Refugee Crisis which was thought to have started in 2014 and reached the peak of crisis in 2015. Moreover, the Russo-Ukrainian War since 2014, which has been a protracted conflict between Russia and Ukraine mainly in the Ukraine regions of Crimea and Donbas, has caused a high level of political risk in Ukraine. Involved in this war, Poland made military responses, while Turkey encountered military actions by Russia. In Bangladesh around the same time, the political risk level not only became higher in 2015, but reached the extremely high level. This was a manifestation of the Bangladesh political crisis in 2015, a political turmoil between the Awami League (AL) and the Bangladesh Nationalist Party (BNP), a terrorist organization considered by AL. The crisis had led to many violent and even fatal attacks on the public. As in Philippines, the high level of political risk in 2017 was mainly due to an armed conflict between Philippine government security forces and the ISIS (more precisely, the Battle of Marawi).

**Figure 7.** The spatial evolution of the political risk along the Belt and Road.

Next, let us decompose political risk into domestic and international components. The spatial evolution of domestic political risk along the Belt and Road is presented in Figure 8. We find that the extremely high level of domestic political risk showed in Syria, Iraq, Afghanistan, and Yemen in all these four years. They are followed by Myanmar, Palestine, Lebanon, Israel, Egypt, and Pakistan, which showed the extremely high level of domestic political risk in at least two of these four years. The high level of domestic

political risk often exhibited in Iran, India, Myanmar, Philippines, Bangladesh, Turkey, and Saudi Arabia in at least two of these four years. Besides, sudden changes in the level of domestic political risk also appeared in 2015, with the political risk in Russia and many BRI countries of Central and Eastern Europe increasing to the high level. Major causes of domestic political risk in these BRI countries are thought to include conflicts due to religions, terrorism, civil wars and conflicts, actions by anti-government forces, social unrest, and the refugee problem. Besides, it is important to realize that some of the causes for the high level of domestic political risk are due to interplay between domestic and international events, such as the European Refugee Crisis.

**Figure 8.** The spatial evolution of domestic political risk along the Belt and Road.

The spatial evolution of international political risk along the Belt and Road is presented in Figure 9. Comparing with the domestic political risk along the Belt and Road shown in Figure 8, we find that the color becomes much lighter, meaning that overall the international political risk along the Belt and Road is much lower than the domestic political risk. While the extremely high level of domestic political risk showed in four countries, Syria, Iraq, Afghanistan, and Yemen, in all these four years, we find only one country, Yemen, reached the extremely high level of international political risk in all these four years. Besides Yemen, Iraq, Syria, Palestine, Israel, and Lebanon also exhibited the extremely high level of international political risk in at least two of these four years. The lesser level of international political risk, the high level, was found in Saudi Arabia and Afghanistan in at least two of these four years, followed by Pakistan, India, Turkey, Ukraine, and Bangladesh, which showed this level in only one of these four years.

**Figure 9.** The spatial evolution of international political risk along the Belt and Road.

#### *4.4. the Correlation between Political Risk and China'S Foreign Investments and Construction Contracts*

To examine the correlation between political risk and China's total investments and construction contracts for BRI countries during the period from 2013 to 2019, it is instructive to construct scatter plots between political risk and China's total investments and construction contracts for BRI countries during this period. This is shown in Figure 10. We find in Figure 10a that the correlation coefficient for all BRI countries is only *r*<sup>2</sup> = 0.16 (*p* = 0.28); it is increased to *r*<sup>1</sup> = 0.33 (*p* = 0.026) when the war-torn countries, Yemen, Afghanistan, and Iraq (which were denoted by red), were excluded. The correlation coefficient is further increased to *r*<sup>3</sup> = 0.58 (*p* < 10−<sup>6</sup> ) when four more countries, the United Arab Emirates, Lao, Singapore, and Kazakhstan, which were denoted by green, were also removed; this is shown in Figure 10b. Since the correlation coefficient is quite positive, by large, we can say that political risk and China's total investments and construction contracts for BRI countries during the period from 2013 to 2019 are strongly positively correlated. With investment, one certainly wishes substantial reward. The positive correlation between risk and investment thus highlights that reward and risk are highly entangled. As this is not ideal, we can ask, is there any way for us to break the "curse" of always accompanying investment with risk? The answer lies in excluding the four countries the United Arab Emirates, Lao, Singapore, and Kazakhstan. The risk levels in these four countries are rather low. However, China's investments and construction contracts in these four countries are quite heavy. One can readily perceive that if the number of such countries, i.e., low risk countries with substantial investments from China increases (for example, China's investments to low-risk countries such as Cyprus, Brunei, Moldova, Oman, Turkmenistan, and Belarus greatly increases), then the correlation between political risk and China's total investments and construction contracts for BRI countries may not only weaken, but becomes negative altogether. This would be the ideal case. Unfortunately, along the Belt and Road, the correlation between political risk and China's total investments and construction

contracts will basically remain positive, since there are a lot of high-risk countries but with heavy investments from China, including Pakistan, Indonesia, Malaysia, Bangladesh, Saudi Arabia, Egypt, India, Russia, Iran, Israel, Cambodia, and Turkey.

**Figure 10.** Scatter-plots between political risk and China's total investment and construction contracts for BRI countries during the period of 2013-2019, where (**a**) shows that the correlation coefficient for all BRI countries is only *r*<sup>2</sup> = 0.16 (*p* = 0.28), while the coefficient is increased to *r*<sup>1</sup> = 0.33 (*p* = 0.026) when the war-torn countries, Yemen (YEM), Afghanistan (AFG), and Iraq (IRQ) (which were denoted by red), were excluded. The red regression line refers to *r*<sup>1</sup> = 0.33. The correlation coefficient was further increased to *r*<sup>3</sup> = 0.58 (*p* < 10−<sup>6</sup> ) when four more countries, the United Arab Emirates (ARE), Lao (LAO), Singapore (SGP), and Kazakhstan (KAZ), which were denoted by green, were also removed; this is shown in (**b**), with the green regression line referring to *r*<sup>3</sup> = 0.58.

#### **5. Conclusions**

We aimed to gain insights into two important questions, (i) How can political risk of BRI countries be properly assessed? and (ii) Are China's BRI investments and construction contracts largely in BRI countries with low levels of the political risk? If not, what are the general characteristics of political risks associated with China's investments and construction contracts? In trying to resolve these two questions, we used a few big data sets, including GDELT, CGIT, and ACLED, to systematically assess the political risk along the Belt and Road during the period from 2013 to 2019. We made several findings: (1) the type of events, "Protest", is inappropriate to be included for assessing political risk, because the nature of protest in a democratic country is entirely different from that in a non-democratic country; (2) choosing the type of events, "Material Conflict", which includes "Exhibit Force Posture", "Reduce Relations", "Coerce", "Assault", "Fight" and "Use Unconventional Mass Violence", is more appropriate for evaluating the political risk; (3) using the number of events for assessing political risk is also inappropriate, since the number of the type of events that are chosen for evaluating risks may be correlated with the total number of events that is covered by GDELT for a country, and the number can vary substantially for a country over time and among different countries around the globe in a fixed (short) time interval; (4) using the Goldstein Scale of events is more advantageous than directly using the number of events, because an event with the Goldstein Scale of −10 amounts to 10 events with the Goldstein Scale of −1; (5) it is of importance to design a normalized variable to assess the political risks of any BRI country in any period of time, to facilitate comparison among different countries; (6) it is beneficial to decompose political risk into two components, domestic and international political risk, and then to assess which type of political risk a BRI country is facing.

By examining the spatiotemporal evolution of political risk along the Belt and Road during the period from 2013 to 2019, we observed that the sum of the number of BRI countries with the extremely high level and the high level for domestic, international, and (overall) political risk all reached the peak in 2015, and decreased thereafter, and that

overall the international political risk along the Belt and Road was much lower than the domestic political risk.

We found a strong positive correlation between political risk and China's total investments and construction contracts for BRI countries during the period from 2013 to 2019. While this is quite the opposite of the ideal case that investment goes to countries or regions with as low political risk as possible, it nevertheless suggests that if we want to achieve the ideal case, it would be necessary for China to choose to invest in countries and regions with low or even negligible political risks along the Belt and Road.

#### **6. Discussions**

While various kinds of traditional economic data will remain critical for assessing risks, it has become increasingly clear that big data, including massive media reports, offer an unprecedented opportunity to help to evaluate, manage and control risks. Yet, the challenge for achieving this goal is also enormous. To better know the potential of this viewpoint, in this paper, we tried to provide a new approach to assess political risk along the Belt and Road using GDELT. We showed that the "Material Conflict" types of events can represent pertinently and comprehensively the events that may directly affect foreign investment. Furthermore, we showed that the contribution of events in this category to risk is better quantified by the summation of the Goldstein Scale rather than by the number of events. These provisions, while simple, enable reasonable comparison among BRI countries. Clearly, the usefulness of these insights may not be confined only with BRI projects, but extended to general overseas investments.

Before pondering the potential future research topics, let us first discuss the caveats of the present study. There are quite a number. First, GDELT has under-reported a lot of interesting events, including those related to risks, in remote regions of the globe [76]. Unfortunately, this limitation is not a unique trait of GDELT. Rather, it is shared by all big databases based on news reports, and it does not appear that there is any way this limitation will go away soon. Second, it is very difficult to track future evolutions of an event or a cluster of events covered in GDELT, and thus it is not easy to evaluate long-time impacts of a specific event or a cluster of events. Third, this study only focuses on the national-level analyses of political risk. Fortunately, the last problem can be readily solved, since GDELT has provided geo-coordinates for each event, and thus in principle allows one to look into political risk associated with specific locations. The difficulties one may envision with a localized study is whether data for a chosen interested region may be large enough for meaningful statistical analysis.

Let us now ponder interesting future research topics that may be solved by analyzing GDELT. First, clearly it is interesting and worthwhile to extend the current study to assess risks associated with general overseas investments rather than just BRI projects. Second, it appears interesting and feasible to carry out a coupled study of politics and economics. Third, it may be useful to further divide the events chosen here for assessing political risk into a few different categories, then evaluate risks for each category, and finally synthesize the risks into a single risk index. Third, it may be interesting to use the risks along the BRI countries computed here as a reference to further study the spatial correlation of political risk among different countries, in the sense that many BRI countries may be bundled together due to a single event, such as the India-Pakistan border skirmishes (2016–2018).

Finally, we emphasize that the political risk identified here may not be equated to the actual risk a foreign investment may face. The risk identified here is better considered as the nominal risk. Part of this risk will be absorbed by a country because of its national collective power including the level of its economic development, research and development capability of its science and technology, government capacity, and resilience of its citizens. This calls for a completely new scheme to determine the actual risk by studying how the collective power of a nation affects the nominal political risk for a country identified here.

**Author Contributions:** Conceptualization, X.S. and J.G.; methodology, X.S. and J.G.; software, X.S. and B.L.; formal analysis, X.S.; data curation, X.S. and B.L.; writing–original draft preparation, X.S. and Z.W.; writing–review and editing, X.S., J.G. and Z.W.; visualization, X.S. and Z.W.; supervision, J.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the National Natural Science Foundation of China under Grant Nos. 71661002 and 41671532 and by the Fundamental Research Funds for the Central Universities under Grant No. 310432101. It is also supported by the National Key Research and Development Program of China, grant number 2020AAA0103402. One of the authors (JG) also benefited tremendously from participating the long program on culture analytics organized by the Institute for Pure and Applied Mathematics (IPAM) at UCLA, which was supported by the National Science Foundation.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data available in a publicly accessible repository.

**Acknowledgments:** We would like to thank the high-performance computing supported from the Center for Geodata and Analysis, Faculty of Geographical Science, Beijing Normal University [https://gda.bnu.edu.cn/] (accessed on 16 September 2020). We would like to thank anonymous reviewers for their valuable comments and suggestions for improving this paper.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Article* **Towards Local Sustainability of Mega Infrastructure: Reviewing Research on the New Silk Road**

#### **Hannes Thees**

Center for Entrepreneurship, Catholic University Eichstaett-Ingolstadt, 85072 Eichstaett, Germany; hannes.thees@ku.de

Received: 11 November 2020; Accepted: 11 December 2020; Published: 18 December 2020 -

**Abstract:** The Belt and Road Initiative is the leading project in the regions along the ancient Silk Road. This aims to revive the New Silk Road (NSR) as a transnational space towards an era of new regional integration and globalization. Despite the potential economic effects on a global scale, local sustainability remains questionable. Building upon the central engagement in infrastructure improvements, this article aims to investigate the role of local sustainability in research along the New Silk Road. Starting with 597 scientific articles, this article conducts a systematic literature review on four levels of concretization to characterize the research field of the New Silk Road, and to develop in-depth insights systematically. The results reveal a research focus on economic growth, which is lacking in environmental considerations and especially the socio-cultural dimension of sustainability on a local scale. Future directions in local sustainability should therefore include local stakeholders to build a joint understanding of sustainability by recognizing the characteristics of regionalism upon which manifold local support of mega infrastructure can evolve. Given these findings, the New Silk Road emerges as a field of study that calls for interdisciplinary research on different spatial levels.

**Keywords:** New Silk Road; local sustainability; mega infrastructure; systematic literature review; Belt and Road; sustainable development; local impact

#### **1. Introduction: The NSR and the Challenge of Sustainability**

The "Belt and Road" initiative is the largest development and globalization program worldwide [1]. In 2013, the People's Republic of China launched the initiative, which includes several overland corridors (Silk Road Economic Belt) and a maritime route (Maritime Silk Road). Connecting with the Ancient Silk Road, the Belt and Road aims to develop trade networks between Asia and Europe, but also towards Southeast Asia, Australia, Africa, and the Middle East. By 2019, 123 countries had officially joined the initiative [2]. Therefore, China promotes investments in different kinds of infrastructure in an export-led growth model, through various financial instruments. The countries or locations that are the focus of the Belt and Road investments are often involved in various global and local initiatives, and thus receive financial support from different countries. Therefore, here I apply the term "New Silk Road" (NSR) as a container for investments and developments in the respective regions that support international trade infrastructure.

Through the years of its operation, the Belt and Road has been subjected to various optimistic scenarios, but also criticism, as nowadays globalization is being challenged. People all over the world are critically assessing globalization because of protectionism, trade wars, or immigration stops, challenging the idea of globalism and a global community. This is combined with a shift in the political world order, wherein China is pursuing more integrated and inclusive globalization [3]. Beyond this, 2020 and the following years will challenge global connectivity and trade within the Belt and Road, as the global slowdown [4] during COVID-19 [5,6] could lead to a renationalization of value chains [7].

The impacts of globalization on local economies and societies are spatially different and influenced by local initiatives [8]. Although countries along the NSR see potential for economic development [9,10], the unclear local effects, a lack of transparency, fears around Chinese dominance, and the role of transit countries offer space for improvement [11–16]. Especially from a European perspective [17], the Chinese role in investments, infrastructure construction, and trade operations could hinder sustainable development. The considerable extent of infrastructure investments raises the complexity of the Belt and Road in the countries along the NSR; a complexity that requires us to steer between the Belt and Road, other international development projects, and local interests. Together with the need to balance infrastructure and its impetus for sustainability, the question arises:

#### *How Has the NSR Been Researched with Special Consideration to Local Sustainability and What Are the Future Directions of Mega Infrastructures in This Context?*

Given this question, the NSR serves as an exemplary research field. Regional studies, as well as multi-disciplinary research, should assist in finding pathways towards sustainability. Recent research on the NSR has been influenced greatly by edited volumes, for instance on globalization [18], transformation in Central Asia [19], financial implications [20], and geopolitics [21]. Within the many book publications, there has only been a little research on local or regional matters [22]. With the growing number of journal publications, the NSR is emerging as a diverse field of research.

Given the above local sustainability challenge of the NSR, this study aims to capture current research on the NSR and to gain insight into the discussion of local sustainability (Figure 1). First, the theoretical background addresses the importance of sustainability, within the specialties of mega infrastructure. Second, the methodology of this article is a systematic literature review [23,24], which is then also discussed in light of theoretical considerations. Finally, the study derives research pathways within the NSR to further shape local sustainability.

**Figure 1.** Research framework. Source: Own elaboration.

#### **2. Theoretical Background: Sustainable Implementation of Mega Infrastructure**

The conceptual options for the NSR are various. A range of concepts, theories, and disciplines are relevant as the NSR influences nearly all dimensions of life—including technology, economics, the environment, geography, businesses, politics, information and knowledge management, and socio-cultural aspects [1]. Such a holistic understanding meets the principles of sustainable development, even more so when the NSR is viewed from the perspective of the participating countries, exceeding the mere allocation of finance and the promotion of trade. In terms of sustainability, there are a number of linkages to discuss within the NSR. However, as the research field on the NSR is still emerging, this article concentrates on the fundamental linkage between infrastructure and sustainability, and puts the focus on the local scale. This limitation on a central theoretical linkage opens space for future research to occur in a more integrative manner. However, the spatial focus on the local scale determines the disciplinary positioning of this article in local or development studies. Complementary to this, transport geography is the starting point of the theoretical considerations, as the NSR is dominated by mega infrastructure—especially when recognizing the economic size of the receiving countries. This is all the more relevant as the NSR exceeds the perspective of single national infrastructure projects as it is combining a whole set of transnational mega infrastructures.

#### *2.1. Megaprojects and Their Challenges*

Megaprojects are often perceived as drivers for long-term development, as they allocate capital and workforce, but also technological knowledge. Megaprojects are defined by investments of more than one billion USD, and have a long lifetime of about 50 years [25,26]. Further criteria involve high complexity, specified knowledge, widespread impacts, or multi-stakeholder involvement [27]. Megaprojects may include all kind of projects, but are strongly related to physical infrastructure (mega infrastructure), greater industrial production, or resource extraction (see modernization theory, e.g., [28]). Such physical infrastructure is typically a public good, in which governments are highly involved, and ranges from power supply, telecommunications, water and sanitation supply, education and healthcare, to freight and public transport [29–32]. Transport infrastructure particularly includes high costs and a long duration of construction [33]. By taking a historical perspective, researchers have agreed that infrastructure and specific transport infrastructure enhance performance, accumulate capital, support knowledge creation, create opportunities of production and trade, and serve social and economic development [30,34–36]. This relationship is also stressed as an accelerator of growth in developing countries [29,34,37,38].

Nevertheless, the ongoing hunger to increase economic growth calls for even more massive infrastructure investments, causing significant challenges, leading to questions regarding the overall performance of megaprojects [39]. The risks of megaprojects include cost, demand, the financial market, and political risks [25]. Flyvbjerg [26] extended this categorization by claiming ten challenges, including also conditions such as a high number of participants, multi-nationality, diverging interests, increasing costs over time, and changing regularities. Further, he mentioned weak project leadership, knowledge integration, cultural differences, extraordinary technology implementation, difficulty in performance evaluation and planning, and the occurrence of an external and unplanned crisis. All in all, megaprojects face divergence of desired and realized outcomes, as they are often promoted in an over-optimistic way as political symbols [25,26,39]. This is also reflected by current infrastructure projects that have received criticism, such as the Nord Stream 2 pipeline, the Suttgart21 railway station, the Grand Ethiopian Renaissance Dam, the Fehmarnbelt crossing between Hamburg and Copenhagen and Desertec in the Sahara region.

From a theoretical perspective, researchers tend to focus on understanding the dynamics of megaprojects, and are thus addressing multi-disciplinary research agendas *"including management and organisation studies, but also history, anthropology, sociology, urban studies, engineering, and economic geography"* [39]. From the perspective of development studies, new international economics, according to Krugman [40], are suitable for mega infrastructure. In particular, transaction costs, capability, social conflict theory, handling of the interface, cost theories, power, and innovation theories are addressed [32,39]. Against the background of the high complexity, Söderlund [39] called for increased research on the management and functioning of cooperation and coordination [39]. Although the challenges and the cooperative nature of megaprojects are obvious, research on multi-country cooperation and on interests of private and public actors is limited, with the exception of research from Kardes et al. [32]. For the economic growth of a respective country, foreign aid or development assistance is not a necessary condition, according to several theories. Besides theories

of endogenous growth, development cooperation is connectable to several concepts and theories, such as modernization, Foreign Direct Investments, the reduction of disparities, multiplier and accelerator effects, and Rostow's growth model. Research might still be in the phase of giving economic reasons for megaprojects, which is visible in multiple research works on non-megaprojects and their impact assessment [34,36,41,42], productivity [35], trade relations [43], and performance [31]. However, research has lately engaged more in sustainability [29,44–46] in order to take responsibility for sustainably designing projects that can impact millions of people [26]. Recognizing this need, Söderlund [39] called for the need to rethink why megaprojects exist, and to include also a discussion of their soft effects [39].

#### *2.2. Framing Sustainable Development from a Local Perspective*

Local development can call on various research agendas and theoretical streams [47,48]. As regional studies are highly context-specific, they vary in terms of their sustainability-definition, which has recently been shaped by questions around inclusive development [49], sustainable development [50,51], and the transition towards sustainability in regions [52]. The manifold conceptual differences in sustainability can follow Sturup [53]: Sustainability is *"the property (a species, a process, a culture, a society etc.) or quality of being able to be sustained"*, which also implies a normative perspective. Sustainability includes a systemic perspective, including environmental, socio-cultural, and economic principles [54]. The term 'sustainable' is the *"measure of the degree to which something can be sustained"* [53,55]. Sustainable development follows the Brundtland definition of meeting the *"needs of the present without compromising the ability of future generations to meet their own needs"* [56] (p. 43), which is also based on the scarcity of resources [57,58].

Through its multiple perspectives and disciplines [59], the targets of sustainable development are wide-ranging, starting with environmental protection in the 1990s and moving towards increasing quality of life in the 2000s [60]. Further criteria of sustainable development are justice and equity in terms of recognition, process, procedure, and outcome, as well as respecting the limits of the ecosystem and promoting cultural identities [61]. This is supported by the concept of the triple bottom line, which calls for the harmonization of the environmental, social, and economic perspectives. Although sustainability has been frequently discussed in scientific discourse [54,62,63], it is still contested in its demarcation from sustainable development [50]. This article tries to follow the clear idea that sustainability is the goal in a system integrating economic, social, cultural, political, and ecological factors, and sustainable development is the implementation of measures from a long-term and multi-scale perspective. Practically, this means that local sustainability in mega infrastructure is the quality or goal to be achieved, supported by various development pathways that especially rely on sustaining local interests in a global system.

A significant challenge of sustainable development is its implementation [64]. Although initiatives such as the UN Sustainable Development Goals (SDG) seem to be broadly accepted and widely used, a conflict of interests exists for instance between economic sectors or regions. There are increasing calls for a discussion on regional sustainability, which connects global, regional, and local efforts [50,65] and asks the questions *"what is to be sustained, by whom, for whom"* [61]. Consensus exists about the need to have sustainability strategies at all spatial scales (principle of subsidiarity), and thus they have been implemented in nearly all policy documents [60], but broad stakeholder involvement is still needed in the end. A unique role obtained by residents is civic engagement, and bottom-up processes are central to starting and successfully implementing more sustainable initiatives [66,67]. Such initiatives also relate strongly to the concept of community development and endogenous growth in order to enhance the local culture and environment through sustainable production and consumption, with the target of improving quality of life, which is also accompanied by the empowerment of residents and local decision-making [68,69]. Recognizing the complexity and the multi-level and -disciplinary nature of sustainability, it has become a global focus and requires the joint action of the world community. However, the interventional nature of the concept of 'global development' is vivid: local development

itself is embedded in regional, national, or even global factors, and thus relies on joint infrastructure projects, trade, cultural exchange, diplomacy, and cooperation.

#### *2.3. Local Sustainability Assessment in Mega Infrastructure*

Against the background of increasing spending on megaprojects—specifically mega infrastructure—in order to keep the pace of economic development, solutions need to be found to assess such projects by a number of different dimensions [29,46]. Taking the local scale into focus, mega infrastructure and its impacts need to be evaluated frequently in terms of their local sustainability, and thus the value they bring to society [34], by asking *"Are megaprojects the right solution?"* [39]. This needs to be embedded in different political and developmental approaches; for example, the European nations consider the balance between development models (e.g., modernization or export-based development) more than Asian nations.

In practice, the attitude towards megaprojects is unclear. It remains uncertain, however, to which extent and how infrastructure investments create jobs, generate income, foster economic sectors, and facilitate local development, or even if social exclusion is increased or decreased [70,71]. Reflecting upon and researching the criteria of local, sustainable implementation of mega infrastructure leads towards an endless collection of intervening and loose criteria, of which each has been tested and applied only seldom in practice.

Practical evidence can be found through several infrastructure projects that affect all dimensions of sustainability. For example, there exists valuable international discourse and research on the environmental impact of the Channel Tunnel [72], the climate's long-term impact on New Zealand infrastructure [73], the impact of the Grand Ethiopian Renaissance Dam on water resources [74], port connectivity between Burgas (Bulgaria) and Alexandroupolis (Greece) [34], land use in the Polavaram River project (India) [75], international security linkages over Turkey's Ilisu Dam [76], the re-settlement of China's Three Gorges Dam [77], and the UN on Infrastructure and Human Rights [78]. Some of these projects have been highly criticized for their environmental or social impact. Although the Belt and Road projects are still under construction, criticism can evolve in some projects, leading to calls for learning from previous infrastructure projects and their sustainability assessments.

Central to assessing infrastructure is the analysis of its performance. The performance of mega infrastructure is highly shaped by complexity, time-duration, or its extensive impact on communities or ecosystems on several spatial scales [32,39]. Even during the planning phase, it causes high uncertainties in terms of forecasting potential effects [25]. Approaches to measure those effects have been provided by Dimitriou et al. [34], Fedderke and Bogetic [35], and Shen et al. [29], with special consideration around comparing the invested capital (private and public) and the economic performance. These assessments are in line with theoretical streams, such as modernization or dependences, and a number of theories and concepts that contribute to explaining certain elements of the complex relationships, such as spillover [79], productivity [35], regional cooperation [80], firm births [81], spatial distribution [82], competitiveness [83], local entrepreneurship [84], global production networks [85,86] or knowledge networks [86], and diversification or cluster management [38].

Calls are also increasing to include sustainability considerations in all phases of infrastructure planning, construction, and operation. The responsible parties need to decide how an infrastructure project assists in solving the issues of sustainable development in their related ecosystems [29,53,87]. In general sustainability studies, researchers have engaged in developing indicators, scenarios, and measurements of sustainability, but are still struggling with the availability, evaluation, and aggregation of data in a multi-dimensionality and interdisciplinary setting. Besides macroeconomic analysis, qualitative approaches also try to describe the behavior of actors in their surrounding socio-economic systems [65]. By applying complex measurements, sustainability can be broken up into a number of indicators, which cannot be assessed in their holistic surroundings [53]. In terms of the sustainability of mega infrastructure, evidence from research is fragmented or often missing altogether. An important step to address this was the launch of the Journal of Mega Infrastructure and

Sustainable Development in 2019 to fill this gap. Research from the Journal of Sustainable and Resilient Infrastructure is valuable, as well, but rather specific. On a practical level, institutions are engaged in deriving practical guidelines for project implementation. This practical collection of determinants to promote a more sustainable implementation of mega infrastructure stresses the role of planning and monitoring. The central determinants are:


In sum, the above argument follows the idea that mega infrastructure accelerates economic development, which needs to be transformed towards the goals and principles of local, sustainable development. Figure 2, therefore, serves as an exemplary discussion-grid that addresses the specialties of mega infrastructure in terms of sustainability dimensions, context, type of infrastructure, and supplementary determinants.


**Figure 2.** Dimensions of local infrastructure assessment. Source: Own elaboration, derived from Shen et al. [29], Ward et al. [92], the OECD [71], and Haughton and Counsell [60].

Reflecting upon the theoretical roots in regional studies, sustainability and transport geography, the assessment of the impacts of mega infrastructure on different local dimensions remains fragmented. There exist practical discourse and research that certainly assist in the planning of mega infrastructure, including within the NSR. Universally valid learnings are hard to find, as all projects are embedded in a unique local setting and are often stuck in a specific sustainability dimension. Future research at this interface of infrastructure and sustainability thus could engage in both integrated development approaches and tools for flexible sustainable planning.

#### **3. Methodology: Systematic Literature Review on the NSR**

Catching up the challenges of the NSR in its local implementation, research and practice need to develop an understanding of what sustainability along the NSR means, before developing guidelines or monitoring. With the aim of discussing local sustainability within research on the NSR, this systematic literature review (SLR) identifies research articles in a systematic and reproducible way. Consequently, specific search processes and search criteria were implemented (Figure 3, Table 1) and assisted in framing the research field [23,93–99]. Compared to a traditional narrative review, an SLR is less rigid and seeks to answer a specific research question [96,100,101]. There is manifold theoretical support for SLRs; scholars like Cooper [102] and Petticrew and Roberts [95] have explored the conceptual foundation, and Denyer and Tanfield [96] and Kitchenham [23] have provided detailed guidelines for conducting an SLR. There exist different types of systematic reviews [101], which range from explorative to confirming, as well as qualitative and quantitative approaches [103,104]. Qualitative SLRs often use qualitative content analysis for a guided exploration of the literature. In opposition, quantitative SLRs apply statistical methods, namely meta-analysis or bibliometric analysis, to evaluate the structures of a research field [99,101,103].

The increasing interest in SLR studies has promoted the establishment of the SLR as a research method. In terms of sustainability, various SLRs have been conducted; for example, on climate change adaption [104], behavioral patterns in climate change mitigation [105], sustainable tourism [94], sustainability transition [106] and performance [107], sustainable supply chain management [108], local sustainability assessment in forestry [109], green infrastructure [110], city logistics [111], and governance of smart cities [112]. While the SLR is widely accepted in research articles, different steps of analysis are predominant. Figure 3 provides a collection of operation steps that serve a qualitative SLR in a more holistic way, which also include pre- and post-considerations in the analysis [95,99,102,104].

**Figure 3.** The procedure of the qualitative Systematic Literature Review (SLR). Source: Own elaboration as an extension to Snyder [101], O'Neill et al. [24], O'Neill and Booth [113] and Tranfield et al. [99].

This article processes a SLR through NVIVO™ according to the provided steps in Figure 3. NVIVO™ is a software used for the qualitative analysis of data, which supports semi-automated as well as manual coding. Its specialty is the processing of multiple rounds of research and analysis, which is also reflected by different literature samples and levels. Therefore, this SLR begins with the meta-level and then continues to develop deeper insights into the lower levels of analysis (Table 1). This is why analysis on the micro- and meso-level applies a qualitative exploration of characteristics and sustainability dimensions. The general search criteria are: publications in scientific journals listed in World of Knowledge (WoK) and Science Direct (SD), publication date from 2013 (official start of the Belt and Road), and publication language English.



Source: Own elaboration.

The first screening stage identified 966 articles on the NSR (Table 2). A correction followed this query to eliminate duplexes, unwanted types of publication (abstracts, editorials, conference proceedings) or thematic misdirection. This correction led to a final meta-sample of 597 articles (Table 3), which represents the starting point for further analysis. The timely distribution of the articles shows the increasing interest in research on the NSR since its announcement (Table 3).





Source: Own elaboration.

#### **4. Results: Framing Research on the NSR**

The SLR had multiple rounds, which represented four levels of concretization. Equally, the presentation of the results starts with the meta-level and continues to explore detailed insights at the micro-level.

#### *4.1. Meta-Level: Is Sustainability of Relevance in the Research Field?*

The meta-level reflects a quantitative description of the research field by broadly including all journal publications in the keywords (Table 1). This chapter aims to evaluate the relevance of sustainability in the broad research field by referring to article keywords and abstracts.

The article keywords allowed to obtain a rough overview of thematic focuses. Based on the word stem, the following keywords were used frequently by authors (Table 4). Obvious keywords, such as Belt and Road, were excluded. Consequently, the keywords represent a lively mixture, including terms like *infrastructure* and *investment*, *regional scales* and *cooperation*, but also *sustainability* and *politics*. Nevertheless, the perception of the NSR as an economic development initiative introduced by China prevails.



Source: Own elaboration, articles = 597, keywords = 997.

Abstracts provide the space to formulate problem statements and research methodology, as well as to indicate the main results of the research. Based on word stems a word cloud was processed (Figure 4). In opposition to the keywords, the abstracts reveal that research on the NSR is often driven by the perspective of the countries along several corridors. *Belt and Road regions* or *Belt and Road countries* are terms which are often used in this regard. Although the term *Asia* is widely used, it is hard to identify a regional focus in these NSR studies. The thematic focus is equal to that of the article keywords: *Development*, as well as *economics*, *trade* and *investment*, play a crucial role. The role of a monitoring or impact assessment of the NSR cannot be neglected, as the keywords *e*ff*ects* and *impacts* reveal. An additional keyword which is visible in the abstracts is *sustainable*, highlighting that a number of research articles have been concentrated on sustainability lately.


**Figure 4.** Word cloud based on abstracts. Source: Own elaboration, articles = 597, words = 1000.

A large share of journals, such as Sustainability and the Journal of Cleaner Production, also confirm the sustainability orientation. Moreover, the journals reveal a possible focus on Eurasia (Eurasian Geography and Economics), Asia (Asian Education and Development Studies), in the Pacific (Pacific Economic Review), or more generally, emerging markets (Emerging Markets Finance and Trade).

In sum, the meta-analysis confirms the intense focus of the NSR on economic development, along with the countries of the respective infrastructure and trade corridors. The sustainability discussion around the NSR has gained increased attention through impact assessment. Beyond that, the NSR is presented as a project with different spatial layers and a need for international networks, relationships, or cooperation.

#### *4.2. Macro-Level: Which Thematic Clusters Evolve in Sustainability?*

The second level of analysis is the macro-level, which should explore clusters in sustainability discussions around the NSR. Therefore, a query with additional keywords (Table 1) represents the thematic focus of this study, and forms a new set of articles.

In general, the keywords of the full texts at the macro-level reveal that *institutions*, *banks*, *finance,* and *world* play a much more dominant role compared to at the meta-level. In more detail, a cluster analysis based on the Pearson Correlation for measuring the word distance in the full texts was conducted (Figure 5).

**Figure 5.** Clustering of articles and issues. Source: Own elaboration, articles = 162.

Tian et al. [114] and Wang et al. [115] discuss the effects of infrastructure investments in close relation to economic growth or development. A significant number of studies estimate the trade effects of the NSR, often in a quantitative manner, such as those by Chen et al. [116] or Baniya et al. [117]. Though it is frequently criticized as not being environmentally friendly, research on the NSR shows several approaches to environmental sustainability. Worth mentioning is the relation of *investments and environment*, calls to analyze the carbon footprint, and the effects of financial instruments following the central question: *"Does finance a*ff*ect environmental degradation?"* [118]. In addition to this question, researchers are strongly focusing on the cluster *energy*, which handles fundamental questions of energy

supply in rural areas, but also the implementation of green energy projects [119]. Besides discussions on the *environmental impact*, including monitoring or climate change, several authors have addressed the *SDG* at the interface of the three sustainability dimensions [120–122]. Still, sustainability challenges the *assessment of development* in general, but also at the local scale. Reflecting the macro-literature set, the spatial concept "local" is seldom applied [123–126]. Another literature gap exists in the exploration of socio-cultural and even socio-economic effects. Issues of health [127] and education [128] are discussed, but a comprehensive understanding of residents and their socio-economic surroundings is largely missing. From a more strategic and political perspective, the cluster on *global strategy* indicates several pain points of the NSR around geopolitical relations. In this vain, cooperation was analyzed between China and regions along the corridors [10,129,130], as well as the approach of global leadership.

A further indication is provided by analyzing the three sustainability dimensions. Figure 6 presents the number of coding references according to the most frequently used keywords in the macro-set that directly include the terms: economic, social (and cultural), and environment. Deriving from the concept of strong sustainability, the economy is at the center, followed by society and the environment. In addition to the cluster analysis, this net graph reveals a research focus on *economic cooperation* and the *development of corridors*, as well as *environmental degradation*, *quality* and *pollution*. The social dimension remains under-researched.

**Figure 6.** Net graph on sustainability dimensions. Source: Own elaboration, articles = 162.

In sum, the macro-level—comprising a specialized set of sustainability concerns along the NSR—indicates that issues of economic development are quite well researched. As a core of sustainable development, research so far has not sufficiently discussed the effects of the NSR on a local scale, including the residents as a local stakeholder group.

#### *4.3. Meso-Level: How Is Local Sustainability Researched?*

The analysis of the meso-level gives insights into the research characteristic of sustainability along the NSR. Therefore, a new set of articles evolves through the selection of relevant clusters; namely those that address the problem statement of local sustainability arising from the NSR impact. Those clusters are: *infrastructure, SDG, local development, assessing development, and economic development.* In general, keywords such as *development, countries, economics, region, infrastructure, policy* or *e*ff*ects* underline the

thematic focus of this meso-level. This selection is completed by a content alignment, which finalizes 58 articles of the meso-set. To characterize the research, this section presents a more qualitative analysis on scales, concepts and theories, as well as applied methods.

4.3.1. Scales: On Which Spatial Scales Does Research Discuss Sustainability?

The majority of the selected articles discuss the NSR in a cross-country analysis, involving sampling of up to 141 countries (Table 5). These analyses mostly have a quantitative and comparative character. However, there is also more specific sampling available, such as by selecting certain corridors or supranational regions; for example, Eurasia or Central Asia. In addition, specific national case studies have been conducted across Eurasia, but also in Kenya and India. Besides analyzing NSR-countries, China itself has often been studied, both at the sub-regional and national level. Subordinated spatial scales of research are hubs [131], special economic zones and industrial parks [132,133], or certain urban networks [134].


**Table 5.** Spatial scales.

Source: Own elaboration, articles = 58.

#### 4.3.2. Concepts: Through Which Theoretical Concepts Is Research Addressing Sustainability?

Table 6 and Figure 7 assign the found concepts to the dimensions of sustainability, including the political dimension and the respective linkages in between the dimensions. Worth mentioning is the high share of research at the linkage of economic and environmental concepts.




#### **Table 6.** *Cont.*

**Figure 7.** Linkages in sustainability research in addition to Table 6. Source: Own elaboration.

A research gap occurs in the social dimension. In general, the concepts reveal that precise theoretical approaches are missing, which makes a clear allocation towards several development streams (e.g., modernization, dependency) tricky. Partially, theoretical chapters are reduced to a minimum and represent a rather general explanation of the NSR, instead of a theoretical foundation, which leads also to an empirical analysis.

In the same way, research targets often remain broad, such as to research challenges and opportunities [162,170]. In fact, a significant share of the reviewed articles aim at understanding and measuring individual relationships, such as between growth and CO<sup>2</sup> emissions [141], trade and labor effects [166], infrastructure and (sustainable) economic development [114,115,169], or FDI and economic growth [146]. Only a few articles link the dimensions of sustainability [179].

4.3.3. Methods: Which Methods Are Applied to Research Sustainability?

Research on sustainability issues along the NSR is dominated by measuring economic effects and relationships between certain variables. Such research is supported by statistical data and the use of economic models and indicators (Table 7). Only a few articles apply spatial models or more qualitative approaches. All in all, the current research misses consequent empirical methods, as a significant share of articles remain descriptive and utilize the methods of a case study or statistical analysis as an empty framework.



Source: Own elaboration, articles = 58.

#### *4.4. Micro: How Is Sustainability Handled and Defined in the Local Context?*

The micro-level is based on a manual in-depth analysis of coded statements, which occurred at the interface of spatial scales (*local, regional, national*) and the respective sustainability issues (*sustainability, development, as well as the single levels of the environment, economy and society*). As many codings mention more than one sustainability dimension, quotations from the articles are presented together with color-coding. Therefore, Table A1 (Appendix A) assists us in understanding the foundations of the research questions on local sustainability.

The analyzed quotations and articles show that the NSR has the potential to foster sustainability in general, but also in local terms [114,120,159]. Its main strength may lie in its understanding as a global initiative [121], which goes beyond traditional aid models. The quotations mention the various relationships, which are strongly connected to sustainability. Led by economic concerns, which postulate the positive relationship between infrastructure and trade, and between trade and labor markets [166] or environmental sustainability [121], the research also indicates that top economic performance does not necessarily go hand in hand with sustainability performance [183]. Unfortunately, performance measurements remain vague and difficult to compare. Only seldom has current research considered the local scale. The high share of international datasets highlights this problem. Variables are the overall development statues, which influences FDI inflow, but also environmental quality [118,140], the industry structure [183], and the different spatial scales applied. For example, urban logistics face different challenges than rural logistics. It is certain that the NSR is going to change the overall spatial patterns [126] of infrastructure, specific sectors, and trade and living environments.

Still, challenges arise at the interface of resource usage and resource protection [171]. Hu et al. [184] call for recognizing local carrying capacities, while Tian and Li [114] demand a balance between trade and environmental issues. Some case studies indicate that environmental sustainability (also referring to the concept of strong sustainability) functions as the basis for further economic and social development [159,179]. In the case of transboundary water resources in Central Asia, Howard and Howard claim that *"countries need to recognise that the economic success of the "Silk Road Economic Belt" hinges on their ability to develop programs that can ensure the region's water resources are managed in a sound and sustainable manner."* [159].

Selected examples from the literature that used a case study method can further represent the local sustainability discussion. In the case of Kazakhstan, Daye et al. [123] found that the tourism sector is a likely winner from the BRI through creating job opportunities and overall prosperity. The positive attitudes of local stakeholders exceed opinions on the possible negative aspects, such as financial costs and indebtedness, or loss of local autonomy. With reference to the Russian part of the Ice Silk Road, Evseev et al. provided insight into the indigenous population in the arctic zone, which represents an especially vulnerable group. For the coastal infrastructure projects, the authors call for regional ecological and social stability through buffer zones. This emphasis on regulating and provisioning ecosystems opposes the need for technological improvements on infrastructure that could assist cleaner trade and manufacturing [121,140]. In opposition, the case of Algeria [120] underlines the necessity of road access as a development driver, and part of SDG goal 9. Therefore, the researchers found that during the Belt and Road projects, the access of the rural population to expressways increased significantly. The example of Georgia in the Caucasus region [126] reveals a highly complex picture of the NSR as a playground for various international interests. The case study shows a mixed picture of perceived benefits in infrastructure, but challenges in regional authority that limits self-determination towards sustainable development. It partially shows the uncertainty of stakeholders regarding the economic benefits of the projects.

In sum, the discussion on local sustainability is strongly connected to political considerations (authority, transparency, governance). On an operational level, a linkage between environmental and economic issues is present, which also shows a gap in the research around socio-economic and socio-environmental issues. Often, the social dimension is skipped by arguing that economic development will automatically increase the welfare and the quality of life of the residents. This gap may promote future studies that will question the social benefits [126], stress civil protests [178], and local autonomy [123,158] and local participation [145]. The need to critically evaluate the NSR and to strive for a more holistic approach to sustainability opens opportunities for further research.

#### **5. Discussion: Future Directions in Local Sustainability**

Building upon the theoretical introduction to various determinants (Figure 2), high complexity, risk, and uncertainty accompanies the sustainable implementation of various types of infrastructure over the four sustainability dimensions. Given a holistic setting, the local sustainability of mega infrastructure has only been addressed in a fragmented way in previous research. This article contributes to outlining future directions for local sustainability within the specific setting of the NSR. Therefore, this discussion addresses three sections for the support of local sustainability.

#### *5.1. Building a Joint Understanding of Sustainability*

Although there exists different approaches to sustainability or sustainable development, they have in common reliance on long-term perspectives, to include different scientific disciplines and to harmonize different interests. As a bottom-up process, it should also serve to foster the quality of life of the residents (Section 2.2).

The local sustainability along the NSR is hard to assess, as motivations of international donors and domestic authorities are unclear, or transparency is lacking at various stages of planning [135]. There is no doubt that the interests on the different scales can vary or even compete [185]. The joint

implementation of mega infrastructure in a sustainable way might only be possible if consensus is created, which is based on clear positions and responsibilities. This collective understanding of sustainability needs to go hand in hand with understanding and researching the basic variables of development: What are the effects of infrastructure? How can I monitor the effects? How can a region address the full potential of external investments? What is the role of local authorities? What are the benefits for locals?

The results of this study reveal that major relationships exist between infrastructure projects and economic growth. Researchers agree that infrastructure supports the latter (Section 2.1), but the effects of mega infrastructure as a public investment on sustainable development remain poorly understood [126]. However, economic growth is often considered alongside international trade and income effects for China. Though the concept of sustainable [186] or green trade [145] are gaining momentum, the local scale is frequently excluded. The aims to increase welfare and economic development, and achieve an inclusive and sustainable economy [121], often remain un-researched at the local scale. Attempts to create indicators and monitoring systems need to be extended here, and be included effectively in policy-making [122].

Taking the framework of the SDG provides the NSR with a *"strategic policy framework for pursuing societal prosperity without undermining environmental sustainability"* [175]. Although the SDGs are appreciated and to date have been widely implemented by institutions and companies [175], the NSR still needs to prove its international openness and the local sustainability of its implemented projects in the long term. Early studies have applied these factors in the context of the NSR [120,132]. The SDGs could help to recognize sustainable development as a holistic and integrated set of economic, social and environmental actions, paired with political responsibility [146,175,179]–even if the SDGs are not free of conflicts. Likewise, a focus on the overall setting of the SDGs could also increase depth in several sub-systems on several spatial scales; for instance, the evaluation of global environmental governance [125].

Finally, a joint understanding of sustainability lies at the core of many discussions in research in general (Section 2.2), but also in implementing the NSR; this should define the responsibilities, principles, and limitations as well. Especially in vulnerable regions, sustainable development is needed that reduces trade-induced emissions and preserves natural resources, which includes local authorities in decision-making and promotes labor markets and induces local welfare [185]. Such discussions should also take into account the many factors of mega infrastructure (e.g., Section 2.2). One needs to admit that the NSR is a critical geopolitical ground, which relies on power relations in bi- and multilateral settings, including political capitalism [187] and development cooperation [188]. Combined with power issues, the risk of debt and thus financial sustainability is of major concern for several countries [115,158], and probably hinders negotiations at eye level. Therefore, the choice of an adequate development model for nations or smaller local areas is of high importance.

#### *5.2. Supporting Mega Infrastructure Locally*

There is a theory that mega infrastructure could be a driver for economic growth if implemented in a holistic manner, which means to include the development of infrastructure hubs, diversification and related services, to provide a local workforce, to recognize entrepreneurial opportunities, and much more (Figure 2). Implementing infrastructure thus should exceed its singularity, and opens up local and regional dimensions. Research and discussions on recent infrastructure projects worldwide assist in finding sustainable pathways.

In the discussion on stakeholders in mega infrastructure, researchers are calling to include local communities in all stages of planning and implementation, to secure the sharing of benefits or to *"enable wellbeing and sustainable livelihoods"* [181]. This is complemented by the expectation that the NSR can follow an environmentally sustainable path if a cross-stakeholders pathway is followed, and a monitoring system assists the project implementation [155]. The concept of private-public partnerships (PPP) is also relevant in this case [189]. Traditional performance measures on infrastructure are inadequate for megaprojects of the extent of the NSR. Referring to the SDGs, mega infrastructure needs to show accountability. For example, an energy project such as a power station, wind farm, or solar array must show how it advances the goal of "*a*ff*ordable and clean energy*" [53].

When external investments push infrastructure, it requires local vigilance, self-confidence, and flexibility towards the own development plans, reflecting the concept of global development. The selection of a respective development model is a strategic decision; it is about growth, dependencies, access to markets, and a lot more. The NSR can be defined as such an investment-led development model [121] to foster trade networks. Primarily, highly fragile states depend on such investments [161]. Development cooperation is a central concept of this. Alonso and Glennie [190] added the convergence of developing countries to higher levels of income and wellbeing, as well as participation in international public goods. The aims of development cooperation, or even foreign aid, certainly have changed over the decades, with a tendency to affect more areas of social life and a clear goal-orientation. Therefore, foreign aid combines the motivation of donors towards human development, democratization, sustainable resource management, and poverty reduction, but also the aims of the receiving countries [188,191]. In addition to investments and capital transfer, neo-classical approaches highlight the countries' own policies; for instance, in adjustment programs and international development cooperation, converging in spillover theories and knowledge transfer [188]. The absence of evaluation leaves questions about the real effects of development aid on both the national and local scales [37]. In the case of Afghanistan and the NSR, researchers recommend reducing dependence on traditional development aid, and to focus more on sustainable aid models [161].

Several authors have recommended practical strategies to assist external investments, such as the creation of industrial parks along a transport corridor, linking local and international companies, involving the hinterland, and improving access to higher education and health services [88], which can be combined with tools of community involvement at a local scale [71]. This call for local adaption of the projects requires several preconditions, including financial impact assessment [135,145], international sustainability standards in domestic and international investments [177], support of inter-regional investments that promote the development of regional economies [171], and the promotion of ownership and control to the receiving countries [190]. Increasing human capital through education might be another prerequisite to participation by employing local workers. Often specific types of human resources are not available in sparsely populated areas, or the quality of the workforce in the least developed countries might not meet the requirements of mega infrastructure [71]. Further measures can be derived from the results of the SLR (Table 8).


**Table 8.** Local measures.

Source: Own elaboration.

Reflecting the theoretical contributions to transport geography or local development, infrastructure projects have researched pre-requisites which should be taken into account, and which probably forces regions to diversify their economic model by moving from mere primary and secondary sectors to the operations of service transport and the planning of respective projects.

#### *5.3. Introducing Governance Models and Regional Integration*

Throughout the last decades, the different global regions have seen widespread cooperation, which has culminated in regional integration—either in relatively informal or formal organizations, such as the EU or the ASEAN [195]. There is a number of reasons why states or regions cooperate in development, such as factor endowment or joint problem solutions, with the aims of promoting trade, knowledge gains, preservation of peace and security, and financial stability [196].

In the same way, the NSR changes international relations and governance modes, develops institutions, activates new regional cooperation [53], and finally promotes regional integration. Integration processes along the NSR are based on increased connectivity. In this regard, regions are bound together by a decrease in travel and transport time, or the establishment of new regional supply networks. Researchers also perceive these infrastructural connections as a prerequisite for sustainable mobility, which should be researched further [120,170,197]. Regional integration is also a matter of multi-level governance that focuses the decision-making processes and their coordination among several levels, including the local, regional, and global [198]. The popularity of multi-level governance has evolved, as it raises questions about nation-states, central governments, and also about other levels and actors, including NGOs and PPPs [199].

The governance model of the NSR is still challenged by limited transparency and power issues, which has led to calls for increased communication, dialog, joint research, and research-collaboration [122]. The main challenge of the NSR countries might be to balance different investment projects from different partners or donors. Research in this regard remains fragmented and highly context-specific, and efforts are thus underway to obtain a better overview. As the potential of international cooperation is clear [200,201], there are calls to introduce international organizations to link spatial scales [121] or Multinational Development Banks [202]. In terms of the NSR, multi-level governance can serve as a means to foster the stepwise integration of regions into global processes and networks, and to include local conditions and requirements. It lies in the center of regional integration and the overall concept of regionalism; that regional cooperation is also fostered in a national bottom-up process. This matches the findings on the need for local policies and international cooperation, and to rethink administrative decentralization in international projects. This likewise opens the discussion on supplementary or alternative development models in opposition to the Belt and Road. Fostering smaller, but independent regional cooperation networks and thus trade and infrastructure networks could also meet the recent challenges of COVID-19. Reinforced by COVID-19, the need for regional networks to be organized in a robust and resilient manner has increased. The World Economic Forum perceives regional integration as a strategy to meet the challenges of COVID-19 [187]. Regional institutions therefore could act as agents binding regions together. As a result of the shock to global supply chains, national governments were forced to provide basic necessities [203] and strengthen local supply chains instead. It is not clear yet whether regional integration was strengthened during COVID-19, as results from Africa show the opposite tendency due to breaking of regional integration protocols in favor of securing national supplies [204].

These approaches to regional integration might serve as an alternative development strategy for countries along the NSR; not to reject the Chinese BRI, but to carefully elaborate the benefits and challenges of all options, and to focus on bottom-up processes to maintain and to take responsibility for local development. This is all the more relevant as the NSR faces not only the challenges of a single national mega infrastructure project, but a complex set of intervened and transnational mega infrastructures. Nevertheless, the local scale is strongly connected to the global initiative, and vice versa. Therefore, sustainable development will require cooperation across nations and among regions

and corridors [155,166,177]. In the same way, China's economic development is closely connected to its partners and to world economies [121].

#### **6. Conclusions: Proposing a Research Agenda for the NSR**

This study builds upon the emerging field of the NSR, and especially gains insight into the role of local sustainability. Therefore, this article started by claiming gaps in the local implementation of the various mega infrastructures that build the NSR. In summary, this study contributes the following: first, a comparison between the principles of local sustainability in mega infrastructure (Section 2) and the current scientific discussion on the NSR (Section 4), revealing that challenges resemble theory along the NSR—although research on the NSR has focused on bridging local interests and international politics (Section 5). Second, the SLR shows the gap in sustainability implementation in infrastructure. Third, the different scales of analysis, and especially the in-depth review on the micro-level, point to a research agenda for the NSR in line with theoretical streams based on seven dimensions:


This article aimed to set up research directions for the NSR, with a special discussion of local sustainability. The NSR was analyzed through the method of SLR with a mixture of quantitative and qualitative factors. Based on these findings, this articles provide the starting point of a systematic analysis of local sustainability along the NSR, rather than providing an impact assessment. This contribution is limited by the range of an SLR, and the set criteria and focus. Thus, this article does not include further groundbreaking publications in the form of conference proceedings or edited volumes, and a detailed exploration of sustainability issues at a supranational or even global scale. The focus of this article lies in current NSR research, and evidence from former research on infrastructure assessment offers manifold theories and practical hints which need to be systematically included in future discussions. As an emerging research field, the NSR is about to enter the scientific discourse dynamically, and will probably provide further insights into local sustainability.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **Appendix A**

#### **Table A1.**Statements on sustainable development.






**Table A1.** *Cont.*


BR—Belt and Road Countries, O—Other, N—Countries, C—China, SN—Supranational, d—Data, e—economic and environmental modelling, s—spatial models, i—indexing and indicators, v—various.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
