Partnership Diplomacy and China’s Exports

Abstract

The use of active diplomacy tactics could explain the rapid growth of China’s trade. By collecting information from the official statements on China’s Ministry of Foreign Affairs website, we constructed a dataset of China’s partnerships to investigate whether partnership diplomacy influenced China’s exports during the period of 1995–2018. We found empirical evidence that partnership diplomacy had a significant and positive impact on China’s exports. The lag terms of the partnerships and the voting data from the United Nations General Assembly have been used as instruments to reduce potential endogeneity. Our mechanism analysis showed that the positive effect of partnership diplomacy on China’s exports followed the law of comparative advantage. By reducing bilateral trade costs, partnership diplomacy has mainly fostered exports in products and industries for which China is more productive and competitive, such as “machinery and electrical equipment” and “textile and articles thereof”. Partnership diplomacy is believed to bridge the ideological gaps between partner countries and enable them to pursue common interests. In light of this, partnership diplomacy could be regarded as a positive factor that has driven China’s export expansion.

1. Introduction

Since joining the WTO in 2001, China’s trade has been booming [1,2,3]. China has grown from an unimportant player to the largest exporter in the world. As shown in Figure 1, China’s total exports increased from USD 299.52 billion in 2001 to USD 2.22 trillion in 2018. Specifically, China’s exports to Asian countries increased from USD 101.64 billion in 2001 to USD 840.07 billion in 2018, exports to European countries increased from USD 74.63 billion to USD 514.52 billion, exports to African countries increased from USD 5.92 billion to USD 86.19 billion, exports to American countries increased from USD 111.13 billion to USD 715.95 billion, and exports to Oceanian countries increased from USD 6.20 billion to USD 61.95 billion. The existing literature has discussed a lot of potential contributory factors to China’s increase in exports. China’s exports can be affected by trade policies and tariff reductions [4,5,6], exchange rates [7,8], credit constraints [9,10], institutions [11], infrastructures [12], labor costs [13], etc. Intercountry conflicts and political tensions also play significant roles in China’s trade. For example, the New History Textbook event in 2005 and the Diaoyu Islands event in 2010 both harmed economic relations between China and Japan [14]. The Japanese invasion of China during World War II has also had a long-term impact on cross-border investments and trade between China and Japan [15]. Some trading partners have experienced significant reductions in their exports to China after meeting with the Dalai Lama [16,17]. To some extent, openness to the world has become a political choice for governments [18]. The impact of the US–China trade war has also drawn considerable attention over recent years [19,20].

Some articles have claimed that small shifts in political relations do not impact trade flows [21,22]. Given the high economic interdependencies and the sunk costs, many consumers, firms, and governments are reluctant to change their economic behaviors in response to political tensions. However, an extensive amount of research has confirmed that power and politics are essential determinants of international trade [23,24,25]. Governments still use trade to punish or reward trading partners. For example, the negative impacts of military conflicts and wars on trade are often significant and persistent [15,26,27]. Wars can even have negative impacts on neighboring countries that are not directly involved in the conflict [28]. Other negative events, such as meetings with the Dalai Lama [16,17], the Diaoyu Island conflict [14,29], and the US–France dispute over the Iraq War [30,31], could hurt international trade as well. Shifts in bilateral political relations can also bring more uncertainty to economic relations [32,33,34]. In particular, a few studies have investigated the effects of diplomatic activities on trade. For example, African politicians’ visits to China promoted China’s exports to Africa, especially the export of capital-intensive manufacturing goods [35]. State visits from Germany, France, and the United States improved their economic relations with trading partners [36]. The regular trade missions that are conducted by the Canadian government have generated tens of billions of USD in new business deals [37]. Embassies and consulates also provided foreign services to help build trade links [38].

A lot of the existing literature has examined the impacts of political relations on trade and some studies have confirmed the positive relationship between diplomatic activities and economic exchanges. Still, there has been little research on the estimation of the economic effects of partnership diplomacy. Partnerships can reflect lasting friendly political and diplomatic relations between countries. When in the honeymoon period, countries are more likely to establish or upgrade partnerships. The building of a global partnership network can be defined as partnership diplomacy. In the post-Cold War era, China has pursued a policy of non-alignment, and partnerships have occupied central positions in China’s diplomatic toolkit [39,40]. In 1993, China established its first partnership with Brazil. During the period of 1993–2018, China established 143 partnerships with 99 countries across the world. There were 50 partnerships with 32 Asian countries, 33 partnerships with 24 European countries, 29 partnerships with 23 African countries, 21 partnerships with 12 American countries, and 10 partnerships with 8 Oceanian countries. Since the National Congress of the 18th Communist Party of China (hereafter named the CPC), partnership diplomacy has entered a new stage of development. There were 78 newly established partnerships during the period of 2013–2018, which accounted for over half of China’s total number of partnerships since 1993. The 19th CPC National Congress report stated that China actively developed global partnerships and expanded the convergence of interests with other countries, which confirmed the importance of partnership diplomacy. It is believed that partnership diplomacy is an important part of China’s diplomacy in the new era, as well as the Belt and Road Initiative and the Regional Comprehensive Economic Partnership (hereafter called the RCEP). Both China’s exports and the number of partnerships have grown rapidly over recent years. The map shown in Figure 2 reports for each country the fraction of years between 1995 and 2018 for which there was a partnership with China. The map shown in Figure 3 reports for each country the total imports from China during the period of 1995–2018. We find many similarities between these two maps. Accordingly, we hypothesize that China’s partnership diplomacy could play a positive role in its export trade.

Even though there has been some studies about the economic effects of the Belt and Road Initiative and the RCEP, little is known about whether and how partnership diplomacy influences China’s trade. Our work aims to fill this gap. Our empirical analyses show that: (1) partnership diplomacy promotes China’s exports; (2) by reducing bilateral trade costs, this promotion effect is mainly driven by products and industries in which China has a comparative advantage, such as “machinery and electrical equipment” and “textile and article thereof”. This paper presents at least three contributions:

• The first contribution is to quantify partnership diplomacy. Based on the information from the official statements on China’s Ministry of Foreign Affairs website, we construct a dataset of China’s partnerships. This is the first well-defined and accurate dataset of China’s partnership diplomacy, which will help us to estimate its economic effects.
• The second contribution is to provide more empirical evidence that political relations can influence trade. From the perspective of international relations, we confirm a significant and positive effect of partnership diplomacy, which can reflect lasting friendly political relations, on China’s exports.
• The third contribution is to empirically prove that China’s global partnership network is a kind of economic diplomacy that fosters China’s trade. On the one hand, partnership diplomacy promotes China’s exports. It means that China’s partnership diplomacy is economically beneficial beyond purely nominal titles. On the other hand, this promotion effect follows the law of comparative advantage. Specifically, this effect is concentrated in products and industries for which China is productive and competitive. It means that partnership diplomacy promotes China’s exports by reducing bilateral trade costs, rather than direct market intervention or other administrative measures.

This article sheds light on the economic effects of China’s partnership diplomacy. Partnership diplomacy is supposed to bridge the ideological gaps between partner countries and enable them to pursue economic gains. By signing bilateral trade agreements, constructing free trade areas, expanding the use of local currency, strengthening financial cooperation, promoting connectivity in the region, etc., partnership diplomacy is an active factor that drives China’s trade growth.

The rest of this paper is organized as follows. Section 2 introduces partnership diplomacy and the dataset of China’s partnerships. Section 3 presents stylized facts and develops hypotheses. Section 4 reports our baseline results that the effect of partnership diplomacy on China’s exports is significantly positive and in line with the law of comparative advantage. Section 5 shows the robustness checks and discusses the endogeneity concern. Section 6 offers heterogeneity analyses. Section 7 concludes the paper.

2. Institutional Background

2.1. China’s Foreign Policies and Partnership Diplomacy

Since its founding in 1949, China has adjusted its foreign policy from “alignment” in the Cold War era to “non-alignment” in the post-Cold War era. In the Cold War era, there emerged a serious and continual confrontation between the imperialist camp led by the United States and the socialist camp led by the Soviet Union on the international stage. To uphold its sovereignty and independence, China experienced “leaning to one side” in the 1950s, “fighting with two fists” in the 1960s, and “one united front” in the 1970s. “Leaning to one side” means founding an alliance with the Soviet Union to oppose the US-led imperialist camp. “Fighting with two fists” means anti-imperialist (the United States) and anti-revisionist (the Soviet Union) at the same time. “One united front” aimed to unite as many forces as possible to fight against the Soviet Union. During this period, China normalized its relations with the United States. Since ideology and national security were the most critical issues, all these foreign policies were certain kinds of alignment in accordance with the existing international environment at that time. In the 1980s, independent from the Warsaw Treaty Organization and the North Atlantic Treaty Organization, China gradually gave up “alignment” and adopted the policy of “non-alignment”. The Cold War came to an end with the fall of Communism in Eastern Europe in 1989 and the disintegration of the Soviet Union in 1991. Peace and development became two major issues instead. To strive for a peaceful international environment for economic development, China began to build its global partnership network. In November 1993, China established its first partnership with Brazil. Then, China established partnerships with Russia, Pakistan, and Nepal in 1996, with France and Canada in 1997, and with Britain and South Korea in 1998. During the period of 1993–2018, China established 143 partnerships with 99 countries in the world.

Partnership diplomacy is a precious legacy of the “non-alignment” foreign policy. Based on the overall judgment and the scheme on international affairs of politics and economics, partnerships reflect one country’s trust in partner countries and willingness to share benefits with them. Partnerships are supposed to guarantee lasting friendly political relations in the long term. As the international situation changes dramatically and China emerges as a new power, whether and to what extent its partnership diplomacy brings economic benefits becomes a controversial question. On the one hand, becoming the second largest economy in the world, China attempts to extend its international influence and proactively pursue its global interests, including economic ones. On the other hand, with the fragile global economic recovery in the post-financial-crisis period, both developed countries and emerging economies pay more attention to economic opportunities brought about by international cooperation and coordination. As exchanges between countries are more and more frequent, partnerships are supposed to foster exports and investments. Partnership diplomacy is believed to be economically beneficial rather than symbolically meaningful.

2.2. Dataset of China’s Partnerships

We construct a dataset of China’s partnerships by collecting the information of the partnerships from the official statements on China’s Ministry of Foreign Affairs website. Since the first partnership with Brazil in 1993, China has built a global partnership network of 143 partnerships in 99 countries. We order these partnerships from lowest to highest according to the ordering of upgrading. Based on the extensive reading of the relevant joint statements, we find that governments always upgrade their partnerships based on the existing ones. Then, we define the existing one as a lower-level partnership and the upgrading one as a higher-level partnership.

Table 1. Coding for partnerships.

Category	Coding	Specific Name
partnership	1	friendly partnership
		new-type cooperative partnership
comprehensive partnership	2	innovative comprehensive partnership
		comprehensive cooperative partnership
		all-around cooperative partnership
		comprehensive friendly cooperative partnership
		all-around friendly cooperative partnership
strategic partnership	3	strategic partnership
		friendly strategic partnership
		innovative strategic partnership
		strategic cooperative partnership
		mutually beneficial strategic partnership
comprehensive strategic partnership	4	comprehensive strategic partnership
		all-around strategic partnership
		comprehensive strategic cooperative partnership
		global comprehensive strategic partnership for the 21st century
		comprehensive strategic partnership of coordination
		all-weather strategic cooperative partnership

Source from the Ministry of Foreign Affairs of the People’s Republic of China https://www.mfa.gov.cn/eng/ (accessed on 12 March 2022).

shows that, taking “comprehensive” and “strategic” as the keywords, 18 different partnerships can be classified into four categories: partnership (assigned the value 1), comprehensive partnership (assigned the value 2), strategic partnership (assigned the value 3), and comprehensive strategic partnership (assigned the value 4). For example, “comprehensive strategic partnership”, “all-round strategic partnership”, “comprehensive strategic cooperative partnership”, “global comprehensive strategic partnership for the 21st century”, “comprehensive strategic partnership of coordination”, and “all-weather strategic cooperative partnership” are categorized as the fourth type, “comprehensive strategic partnership”, which is assigned the value 4. Partnerships here are defined as an ordinal variable $partne{r}_{CHN,jt}$, with higher values meaning better bilateral political relations and deeper political mutual trust between China and its partner country j in year t. We take South Africa as an example to explain how to construct variable $partne{r}_{CHN,jt}$ (shown in

Table 2. Variable $partne{r}_{CHN,jt}$: taking South Africa as an example.

Year	${\mathit{partner}}_{\mathit{CHN},\mathit{jt}}$	Event
1995	0
1996	0
1997	0
1998	0
1999	0
2000	1	In April 2000, China and South Africa signed the Pretoria Declaration on partnership relations.
2001	1
2002	1
2003	1
2004	3	In June 2004, China and South Africa further defined their relations as strategic partnership of equality, mutual benefit, and common development.
2005	3
2006	3
2007	3
2008	3
2009	3
2010	4	In August 2010, China and South Africa established the comprehensive strategic partnership.
2011	4
2012	4
2013	4
2014	4
2015	4
2016	4
2017	4
2018	4

Source from the Ministry of Foreign Affairs of the People’s Republic of China https://www.mfa.gov.cn/eng/ (accessed on 12 March 2022).

). Before 2000, China had not established any type of partnership with South Africa, thus, $partne{r}_{CHN,jt}=0$. Since China and South Africa signed the Pretoria Declaration on partnership relations in 2000, $partne{r}_{CHN,jt}=1$. $partne{r}_{CHN,jt}=3$, since China and South Africa further defined their relations as a strategic partnership of equality, mutual benefit, and common development in 2004. Finally, $partne{r}_{CHN,jt}=4$ when China and South Africa established the comprehensive strategic partnership in 2010.

Table 3. The summary statistics of China’s partnerships in 2018.

Category	Value	Asia	Europe	Africa	America	Oceania
Partnership	1	0	0	1	1	0
Comprehensive partnership	2	6	5	5	0	0
Strategic partnership	3	5	11	5	4	6
Comprehensive strategic partnership	4	12	16	13	7	2
Average level		3.26	3.34	3.25	3.42	3.25

shows the summary statistics of partnerships in 2018. In 2018, China has established 6 comprehensive partnerships, 5 strategic partnerships, and 12 comprehensive strategic partnerships with Asian countries. The average level of China’s partnerships for Asian countries is 3.26. This is the first well-defined and accurate dataset of China’s partnerships, which will help us to estimate the economic effects of China’s partnership diplomacy.

3. Stylized Facts and Hypothesis Development

3.1. Partnership Diplomacy and China’s Exports

In this subsection, we talk about whether partnership diplomacy promotes China’s exports. Conflicts and political tensions hurt trade [16,29], while friendly bilateral political relations and frequent diplomatic exchanges, such as state visits [35,36] and trade missions [37], stimulate economic exchanges. Previously, when two countries decide to establish a partnership, they used to sign a joint statement. By going through these joint statements, we find that economic affairs are important parts of these joint statements. “Trade”, “export”, and “investment” are high-frequency words. For example, “trade” appears 10 times and “investment” appears 2 times in the joint statement between the People’s Republic of China and the Islamic Republic of Pakistan on strengthening the China–Pakistan all-weather strategic cooperative partnership in 2018 (The information is available at https://www.fmprc.gov.cn/mfa_eng/gjhdq_665435/2675_665437/2757_663518/2758_663520/201811/t20181104_519747.html (accessed on 18 August 2022).). Some joint statements even set clear expectations about trade, such as “the two sides should expand economic and trade cooperation in accordance with their respective development strategies, realize the goal of USD 160 billion for the 2017 bilateral trade volume” (The information is available at https://www.fmprc.gov.cn/mfa_eng/gjhdq_665435/2675_665437/2732_663468/2734_663472/201310/t20131006_517425.html (accessed on 28 June 2022).) and “the two sides should continuously make good use of the China-Mongolia intergovernmental cooperation mechanisms, and put into effect China-Mongolia mid-term plan of trade and economic cooperation to realize the goal of bilateral trade volume of USD 10 billion by 2020” (The information is available at https://www.fmprc.gov.cn/mfa_eng/gjhdq_665435/2675_665437/2742_663488/2744_663492/201408/t20140825_518223.html (accessed on 28 June 2022).). It is evident that partnership diplomacy is interest-driven rather than ideology-driven. Partner countries will work together to foster trade. Furthermore, Figure 4 shows a positive correlation between the fraction of years with China’s partnerships and China’s total exports during the period of 1995–2018. Note that we exclude the United States and Japan in samples to draw the scatterplot and run regressions because it is too intricate to discuss China’s relations with these two major powers within the framework of partnership diplomacy. As the largest and the second largest trading partners with China during the period of 1995–2018, the United States and Japan have not yet established any formal partnership with China. Including the United States and Japan may bring bias when estimating the effect of partnership diplomacy on exports. This scatterplot indicates that China tends to export more to these countries that have already established partnerships with China. That is to say, China’s partnership diplomacy may play an active and positive role in trade. On the basis of the above analysis, we develop our first hypothesis.

Hypothesis 1.

Partnership diplomacy can promote China’s exports to its partner countries.

3.2. The Law of Comparative Advantage

In this subsection, we talk about how partnership diplomacy promotes China’s exports. Conflicts and deteriorating political relations may raise trade barriers, create uncertainty, and block trade [41,42,43,44], while lasting friendly political relations may reduce costs of bilateral trade. By going through the joint statements on establishing partnerships, we find that a lot of specific measures are put forward to reduce bilateral trade costs and strengthen economic ties between partner countries. For example, the joint statement between the People’s Republic of China and Malaysia on establishing a comprehensive strategic partnership in 2013 stated that “…to encourage Chinese business to join in the development of northern Malaysia and the construction of the Kuala Lumpur-Singapore high-speed railway so as to promote connectivity in the region. To strengthen financial cooperation, to expand the use of local currency in trade and investment…” (The information is available at https://www.fmprc.gov.cn/mfa_eng/gjhdq_665435/2675_665437/2732_663468/2734_663472/201310/t20131006_517425.html (accessed on 28 June 2022).) The joint declaration between the People’s Republic of China and Mongolia on establishing a comprehensive strategic partnership in 2014 stated that “…enlarge the scale of bilateral currency swap, initiate as soon as possible the preparation of negotiating and signing bilateral trade agreement, and actively promote the establishment of the cross-border economic cooperation zone…” (The information is available at https://www.fmprc.gov.cn/mfa_eng/gjhdq_665435/2675_665437/2742_663488/2744_663492/201408/t20140825_518223.html (accessed on 28 June 2022).) The joint statement between the People’s Republic of China and the Islamic Republic of Pakistan on establishing an all-weather strategic partnership with cooperation in 2015 stated that “…Both sides should make overall plans for the construction of industrial parks along the China-Pakistan Economic Corridor and finalize negotiations on the China-Pakistan free trade area at an early date to lead the construction of free trade areas between China and the South Asian countries…” (The information is available at https://www.fmprc.gov.cn/mfa_eng/gjhdq_665435/2675_665437/2757_663518/2759_663522/201504/t20150422_520318.html (accessed on 28 June 2022).) By signing bilateral trade agreements, constructing free trade areas, expanding the use of local currency, strengthening financial cooperation, promoting connectivity in the region, etc., China and its partner countries are supposed to work together to reduce bilateral trade costs and increase openness. According to the law of comparative advantage, if the increase in China’s exports arises because of decreasing bilateral trade costs, then the surge in exports should be concentrated in products and industries in which China has a comparative advantage. The law of comparative advantage is the basic logic of standard international trade theory from Ricardian to Melitz [45]. With a decrease in bilateral trade costs, countries tend to export more goods that they have a relative cost advantage in producing and import more goods that they have a relative cost disadvantage in producing. That is to say, the more competitive the products or industries are, the more significant the promotion effects of partnerships on exports are. Accordingly, we develop our second hypothesis.

Hypothesis 2.

Partnership diplomacy promotes China’s exports by reducing bilateral trade costs with its partner countries. It implies that the surge in exports is concentrated in products and industries in which China has a comparative advantage.

4. Empirical Analysis

4.1. Empirical Specification

To test Hypothesis 1, we empirically estimate whether and to what extent partnerships can influence China’s exports based on the gravity equation [46]. The gravity equation assumes that trade flows are proportional to the size of trading countries’ economies and inversely proportional to their geographical distances, which is widely used for empirical studies of trade flows. Thus, our empirical specification is:

$$\ln EX{P}_{CHN,jkt}={\beta }_{1}partne{r}_{CHN,jt}+X_{jt}^{\prime }\gamma +{\delta }_j+{\eta }_{kt}+{\epsilon }_{CHN,jkt},$$

(1)

where subscripts $CHN$, j, k, and t are the indices for China, the partner country, product, and year, respectively. The explained variable $EX{P}_{CHN,jkt}$ denotes China’s exports of product k to country j in year t. We take the natural log of $EX{P}_{CHN,jkt}$ here. Data on China’s exports are obtained from BACI (This database is available at http://www.cepii.fr/CEPII/en/bdd_modele/bdd_modele_item.asp?id=37 (accessed on 16 April 2022).). The variable of interest $partne{r}_{CHN,jt}$ is an ordinal variable of China’s partnership with country j in year t, ranging from 0 to 4. $partne{r}_{CHN,jt}$ is an ordinal variable with four categories (partnership, comprehensive partnership, strategic partnership, and comprehensive strategic partnership). We order these categories from lowest to highest, but the spacing between the values may not be the same across the levels of the variables. For example, the spacing between the first and second category is different from the spacing between the second and third category. According to the Hypothesis 1, our primary coefficient of interest ${\beta }_1$ in Equation (1) is expected to be positive, which captures the average reduced-form impact of partnership diplomacy on China’s exports. $X_{jt}^{\prime }$ is a vector of control variables of time-varing country characteristics, including the natural log of gross domestic product (constant 2015 USD) $\ln gd{p}_{jt}$, the natural log of population $\ln populatio{n}_{jt}$, consumer price index (annual %) $cp{i}_{jt}$, trade openness (% of GDP) $opennes{s}_{jt}$, net inflows of foreign direct investment (% of GDP) $bo{p}_{jt}$, urban population (% of total population) $urbanizatio{n}_{jt}$, access to electricity (% of population) $electricit{y}_{jt}$, and mobile cellular subscriptions (per 100 people) $mobil{e}_{jt}$. Data on control variables are obtained from World Development Indicators (WDI) (This database is available at https://databank.worldbank.org/source/world-development-indicators (accessed on 16 April 2022).). Our dataset covers the period from 1995 to 2018. This specification includes both country fixed effects ${\delta }_j$ and product-year fixed effects ${\eta }_{kt}$. Country fixed effects ${\delta }_j$ capture all time-invariant country factors that may be correlated with China’s partnerships and exports. When country fixed effects ${\delta }_j$ are included, there is no need to control for those factors that are usually considered in a standard gravity equation, such as geographical distance, contiguous border, common legal system, etc. Product-year fixed effects ${\eta }_{kt}$ capture all time-vayring product characteristics. The individual product fixed effects and year fixed effects are absorbed by product-year fixed effects ${\eta }_{kt}$. Product fixed effects capture all time-invariant product factors, and year fixed effects capture all time-specific factors. ${\epsilon }_{CHN,jkt}$ is the error term. Considering the competitiveness and complementation among products or industries within a country, standard errors are clustered at the country level to address potential problems of heteroskedasticity and serial correlation.

According to Hypothesis 2, we add an interaction term $partne{r}_{CHN,jt}\times RC{A}_{CHN,kt}$ into Equation (1) to test whether the increase in China’s exports is greater in products and industries in which China has a comparative advantage. Then we obtain

$$\ln EX{P}_{CHN,jkt}={\beta }_{2}partne{r}_{CHN,jt}+{\beta }_{3}partne{r}_{CHN,jt}\times RC{A}_{CHN,kt}+X_{jt}^{\prime }\gamma +{\delta }_j+{\eta }_{kt}+{\epsilon }_{CHN,jkt}.$$

(2)

The relationship between partnerships and China’s exports is moderated by $RC{A}_{CHN,kt}$. The specification also includes the same vector of control variables, country fixed effects ${\delta }_j$, and product-year fixed effects ${\eta }_{kt}$ as in Equation (1). Note that the individual term $RC{A}_{CHN,kt}$ is absorbed by product-year fixed effects ${\eta }_{kt}$. We use the index of revealed comparative advantage $RC{A}_{CHN,kt}$ defined by Balassa [47] to capture China’s export competitiveness, which is given by:

$$RC{A}_{CHN,kt}=\frac{{\sum }_{j}EX{P}_{CHN,jkt}}{{\sum }_j{\sum }_{k}EX{P}_{CHN,jkt}}/\frac{{\sum }_i{\sum }_{j}EX{P}_{ijkt}}{{\sum }_i{\sum }_j{\sum }_{k}EX{P}_{ijkt}},$$

where $EX{P}_{ijkt}$ denotes country i’s exports of product k to country j in year t. $RC{A}_{CHN,kt}$ is the ratio of the export share of product k in China’s export basket to the same share at the worldwide level in year t, ranging from 0 to infinity. Specifically, we take $\ln RCA\_hs{6}_{CHN,kt}$, $RSCA\_hs{6}_{CHN,kt}$, and $d\_RCA\_hs{6}_{CHN,kt}$ as proxy variables of the revealed comparative advantage $RCA\_hs{6}_{CHN,kt}$ of China’s six-digit HS product k in year t and take $\ln RCA\_hs{2}_{CHN,kt}$, $RSCA\_hs{2}_{CHN,kt}$, and $d\_RCA\_hs{2}_{CHN,kt}$ as proxy variables of the revealed comparative advantage $RCA\_hs{2}_{CHN,kt}$ of China’s two-digit HS industry k in year t. $\ln RCA\_hs{6}_{CHN,kt}$ is the natural log of $RCA\_hs{6}_{CHN,kt}$ and $\ln RCA\_hs{2}_{CHN,kt}$ is the natural log of $RCA\_hs{2}_{CHN,kt}$, both ranging from negative infinity to positive infinity. The expressions of $RSCA\_hs{6}_{CHN,kt}$ and $RSCA\_hs{2}_{CHN,kt}$ are given by:

$$RSCA\_hs{6}_{CHN,kt}=\frac{RCA\_hs{6}_{CHN,kt}-1}{RCA\_hs{6}_{CHN,kt}+1^{\prime }}$$

and

$$RSCA\_hs{2}_{CHN,kt}=\frac{RCA\_hs{2}_{CHN,kt}-1}{RCA\_hs{2}_{CHN,kt}+1^{\prime }}$$

both ranging from −1 to 1. $d\_RCA\_hs{6}_{CHN,kt}$ takes a value of 1 if China exports six-digit HS product k with revealed comparative advantage (namely, $RCA\_hs{6}_{CHN,kt}>1$) and 0 otherwise. $d\_RCA\_hs{2}_{CHN,kt}$ takes a value of 1 if China exports goods in two-digit HS industry k with revealed comparative advantage (namely, $RCA\_hs{2}_{CHN,kt}>1$) and 0 otherwise.

In Equation (2), we allow the effects of partnerships to differ across products and industries depending on the extent to which China has a comparative advantage. According to the law of comparative advantage, if partnership diplomacy promotes China’s exports by reducing bilateral trade costs, then ${\beta }_3$ is expected to be positive. The positive impact of partnership diplomacy mainly works in products and industries for which China is productive and competitive.

4.2. Baseline Estimations

Table 4. Baseline: partnership diplomacy and China’s exports.

	(1)	(2)	(3)	(4)	(5)
	OLS	OLS	FE	FE	FE
partner	0.4409 ***	0.0980 ***	0.0582 ***	0.0703 ***	0.0703 ***
	(737.33)	(138.07)	(70.41)	(87.03)	(5.02)
lngdp		0.3440 ***	1.0250 ***	1.0722 ***	1.0722 ***
		(265.08)	(132.49)	(142.08)	(6.05)
lnpopulation		0.2577 ***	−0.3759 ***	−0.5182 ***	−0.5182 **
		(194.21)	(−34.35)	(−48.51)	(−2.56)
cpi		−0.0017 ***	−0.0009 ***	−0.0008 ***	−0.0008 *
		(−29.29)	(−19.13)	(−16.82)	(−1.75)
openness		0.0082 ***	0.0011 ***	0.0012 ***	0.0012
		(166.35)	(13.88)	(15.73)	(1.03)
bop		−0.0006 ***	−0.0008 ***	−0.0007 ***	−0.0007
		(−9.01)	(−14.06)	(−12.96)	(−1.11)
urbanization		0.0023 ***	0.0171 ***	0.0163 ***	0.0163 *
		(32.40)	(38.21)	(37.37)	(1.95)
electricity		0.0000	−0.0005 ***	−0.0031 ***	−0.0031
		(0.28)	(−2.99)	(−19.31)	(−0.99)
mobile		0.0078 ***	0.0029 ***	0.0028 ***	0.0028 ***
		(312.33)	(67.29)	(67.64)	(3.70)
Country FE	No	No	Yes	Yes	Yes
Product-Year FE	No	No	No	Yes	Yes
Product FE	No	No	Yes	No	No
Year FE	No	No	Yes	No	No
Cluster	No	No	No	No	Yes
Number of Countries	166	153	153	153	153
Observations	6,875,172	6,138,986	6,138,980	6,135,664	6,135,664
$R^2$	0.073	0.210	0.503	0.543	0.543

Notes: In column 1, China’s exports are regressed on partnerships directly. Column 2 includes a set of control variables. Column 3 includes country fixed effects, product fixed effects, and year fixed effects, while column 4 employs country fixed effects and product-year fixed effects. The robust standard errors of column 5 are adjusted for 153 clusters at the country level. Coefficients are reported with t-statistics in parentheses. *, **, and *** signify statistical significance at the 10, 5, and 1 percent levels, respectively.

reports baseline estimations testing Hypothesis 1 that China’s partnership diplomacy has a trade-promoting effect. In column 1, we regress partnerships on China’s exports directly without any control variables and fixed effects. We find that the coefficient on partnership diplomacy measure, $partne{r}_{CHN,jt}$, is 0.4409 and statistically significant at the 1% level. In column 2, we add a vector of control variables, including gross domestic product $\ln gd{p}_{jt}$, population $\ln populatio{n}_{jt}$, consumer price index $cp{i}_{jt}$, trade openness $opennes{s}_{jt}$, net inflows of foreign direct investment $bo{p}_{jt}$, urban population $urbanizatio{n}_{jt}$, access to electricity $electricit{y}_{jt}$, and mobile cellular subscriptions $mobil{e}_{jt}$. The coefficient decreases to 0.0980 and remains statistically significant at 1% level. In column 3, we include country fixed effects, product fixed effects, and year fixed effects. The coefficient on partnership diplomacy is 0.0582 and statistically significant. In column 4, product fixed effects and year fixed effects are replaced by product-year fixed effects. Product-year fixed effects absorb product fixed effects and year fixed effects for a panel dataset. As higher-level fixed effects, product-year fixed effects capture all time-varying product characteristics, time-invariant product factors, and time-specific factors at the same time. We still find a positive impact of partnership diplomacy on China’s exports, which is 0.0703. The R-square of 0.543 means that 54.3% of the sample variation in China’s exports can be explained by our specification. The relative magnitudes of the estimated coefficients and the R-squares from columns 1–4 show that it is necessary to employ all these control variables and fixed effects. In column 5, considering the competitiveness and complementation among products within a country, standard errors are clustered at the country level. It helps us reduce potential problems of heteroskedasticity and serial correlation. This is our final benchmark specification. Throughout the rest of the analysis, we run fixed effects regressions with standard errors adjusted for clustering at the country level. Compared with the estimates in column 4, all the estimated coefficients and the R-square remain the same, but all the t-statistics decrease. For example, the t-statistic on partnership diplomacy decreases from 87.03 to 5.02, but it is still statistically significant at the 1% level. Finally, we confirm that the average reduced-from impact of partnership diplomacy on China’s exports is 0.0703. This estimated result makes Hypothesis 1 hold and highlights the importance of economic interests in explaining partnership diplomacy.

4.3. Mechanism Analysis

To test Hypothesis 2 and obtain a better understanding of how partnership diplomacy influences China’s exports,

Table 5. Mechanism: reducing bilateral trade costs.

	(1)	(2)	(3)	(4)	(5)	(6)
	hs6	hs6	hs6	hs2	hs2	hs2
partner	0.0682 ***	0.0628 ***	0.0097	0.0654 ***	0.0646 ***	0.0408 ***
	(4.80)	(4.41)	(0.44)	(4.64)	(4.57)	(2.63)
partner × lnRCA_hs6	0.0574 ***
	(4.48)
partner × RSCA_hs6		0.1476 ***
		(4.01)
partner × d_RCA_hs6			0.1036 ***
			(3.57)
partner × lnRCA_hs2				0.0352 ***
				(3.55)
partner × RSCA_hs2					0.0796 ***
					(3.50)
partner × d_RCA_hs2						0.0502 ***
						(4.33)
lngdp	1.0730 ***	1.0731 ***	1.0734 ***	1.0723 ***	1.0723 ***	1.0720 ***
	(6.05)	(6.05)	(6.06)	(6.04)	(6.04)	(6.05)
lnpopulation	−0.5237 ***	−0.5217 **	−0.5194 **	−0.5207 **	−0.5200 **	−0.5169 **
	(−2.61)	(−2.60)	(−2.58)	(−2.57)	(−2.57)	(−2.56)
cpi	−0.0008 *	−0.0008 *	−0.0008 *	−0.0008 *	−0.0008 *	−0.0008 *
	(−1.73)	(−1.74)	(−1.74)	(−1.74)	(−1.74)	(−1.75)
openness	0.0013	0.0013	0.0013	0.0013	0.0013	0.0013
	(1.05)	(1.06)	(1.06)	(1.04)	(1.05)	(1.05)
bop	−0.0007	−0.0007	−0.0007	−0.0007	−0.0007	−0.0007
	(−1.12)	(−1.13)	(−1.13)	(−1.12)	(−1.12)	(−1.13)
urbanization	0.0167 **	0.0166 **	0.0165 **	0.0164 *	0.0164 *	0.0164 *
	(2.00)	(1.99)	(1.98)	(1.97)	(1.97)	(1.97)
electricity	−0.0031	−0.0031	−0.0031	−0.0031	−0.0031	−0.0031
	(−0.99)	(−0.99)	(−0.99)	(−0.99)	(−0.99)	(−0.99)
mobile	0.0029 ***	0.0029 ***	0.0029 ***	0.0028 ***	0.0028 ***	0.0028 ***
	(3.74)	(3.74)	(3.74)	(3.70)	(3.71)	(3.72)
Country FE	Yes	Yes	Yes	Yes	Yes	Yes
Product-Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Cluster	Yes	Yes	Yes	Yes	Yes	Yes
Number of Countries	153	153	153	153	153	153
Observations	6,135,664	6,135,664	6,135,664	6,135,664	6,135,664	6,135,664
$R^2$	0.544	0.544	0.544	0.543	0.543	0.543

Notes: Columns 1–3 take $\ln RCA\_hs{6}_{CHN,kt}$, $RSCA\_hs{6}_{CHN,kt}$, and $d\_RCA\_hs{6}_{CHN,kt}$ as proxy variables of the revealed comparative advantage of China’s six-digit HS product k in year t. Columns 4–6 take $\ln RCA\_hs{2}_{CHN,kt}$, $RSCA\_hs{2}_{CHN,kt}$, and $d\_RCA\_hs{2}_{CHN,kt}$ as proxy variables of the revealed comparative advantage of China’s two-digit HS industry k in year t. Coefficients are reported with t-statistics in parentheses. *, **, and *** signify statistical significance at 10, 5, and 1 percent levels, respectively. Standard errors are adjusted for clustering at the country level.

reports the estimated results of Equation (2). We add an interaction term $partne{r}_{CHN,jt}\times RC{A}_{CHN,kt}$ to test whether the estimated impact of partnership diplomacy on China’s exports is greater in products and industries in which China has a comparative advantage. If China exports more goods that it is more competitive in producing, it implies there are common efforts between partner countries to reduce bilateral trade costs. We discuss this mechanism in six-digit HS products (columns 1–3) and two-digit HS industries (columns 4–6). Specifically, columns 1–3 take $\ln RCA\_hs{6}_{CHN,kt}$, $RSCA\_hs{6}_{CHN,kt}$, and $d\_RCA\_hs{6}_{CHN,kt}$ as proxy variables of the revealed comparative advantage of China’s six-digit HS products. Meanwhile, columns 4–6 take $\ln RCA\_hs{2}_{CHN,kt}$, $RSCA\_hs{2}_{CHN,kt}$, and $d\_RCA\_hs{2}_{CHN,kt}$ as proxy variables of the revealed comparative advantage of China’s two-digit HS industries. As shown in Table 5, all estimated coefficients on interaction terms between partnerships and the revealed comparative advantage $partne{r}_{CHN,jt}\times RC{A}_{CHN,kt}$ from columns 1–6 are positive and statistically significant at 1% level. It means that the surge in exports is concentrated in products and industries in which China has a comparative advantage, not a comparative disadvantage. According to the law of comparative advantage in international trade, this result is consistent with the expectation that China experiences trade expansion sourced from decreasing trade costs and greater specialization in products and industries with high productivity. We provide evidence that the increased exports arise through partner countries’ common efforts to reduce bilateral trade costs by signing bilateral trade agreements, constructing free trade areas, expanding the use of local currency, strengthening financial cooperation, promoting connectivity in the region, etc.

5. Robustness and Endogeneity

5.1. Robustness Check

We test the robustness of our estimated results in

Table 6. Robustness check.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	hs2	Quantity	Dummy	Include	Diplomacy1	Diplomacy2	Winsor	1995–2000	2001–2018
partner	0.0200 *	0.0684 ***		0.0649 ***	0.0628 ***	0.0663 ***	0.0570 ***	−0.0003	0.0539 ***
	(1.72)	(4.83)		(4.64)	(4.49)	(4.68)	(4.39)	(−0.01)	(5.18)
1. partner			−0.0497
			(−0.52)
2. partner			0.0452
			(0.95)
3. partner			0.1557 **
			(2.27)
4. partner			0.3082 ***
			(5.17)
lngdp	1.1436 ***	0.8749 ***	1.0780 ***	1.0377 ***	1.0855 ***	1.0400 ***	0.9871 ***	0.8141 ***	1.0380 ***
	(12.96)	(4.24)	(6.06)	(6.00)	(6.02)	(5.84)	(6.30)	(3.41)	(6.18)
lnpopulation	0.1561	−0.2475	−0.4986 **	−0.5257 ***	−0.5800 ***	−0.4538 **	−0.4094 **	−1.0629 *	−0.4273 **
	(1.24)	(−1.16)	(−2.51)	(−2.66)	(−2.83)	(−2.19)	(−2.26)	(−1.89)	(−2.19)
cpi	−0.0011 ***	−0.0005	−0.0008 *	−0.0008 *	−0.0008 *	−0.0008 *	−0.0009 *	−0.0002 **	−0.0008
	(−2.93)	(−0.86)	(−1.80)	(−1.74)	(−1.73)	(−1.74)	(−1.97)	(−1.99)	(−0.51)
openness	0.0053 ***	0.0017	0.0013	0.0011	0.0004	0.0015	0.0012	0.0102 ***	0.0021 *
	(6.01)	(1.17)	(1.09)	(0.92)	(0.32)	(0.94)	(1.11)	(2.94)	(1.91)
bop	−0.0015 **	−0.0001	−0.0007	−0.0007	−0.0007	−0.0007	−0.0008	0.0016	0.0001
	(−2.21)	(−0.13)	(−1.10)	(−1.13)	(−1.05)	(−1.12)	(−1.20)	(0.70)	(0.11)
urbanization	−0.0003	0.0166 *	0.0170 **	0.0163 **	0.0183 *	0.0179 **	0.0148 *	−0.0027	0.0188 **
	(−0.06)	(1.93)	(2.08)	(2.11)	(1.95)	(2.11)	(1.97)	(−0.19)	(2.17)
electricity	0.0047 **	−0.0004	−0.0029	−0.0035	−0.0031	−0.0027	−0.0010	−0.0073	−0.0014
	(2.52)	(−0.13)	(−0.94)	(−1.13)	(−0.93)	(−0.84)	(−0.35)	(−1.28)	(−0.46)
mobile	0.0018 ***	0.0022 ***	0.0028 ***	0.0029 ***	0.0033 ***	0.0029 ***	0.0030 ***	0.0017	0.0027 ***
	(3.29)	(2.82)	(3.77)	(3.84)	(4.28)	(3.65)	(4.22)	(1.36)	(3.89)
Country FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Product-Year FE	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	No	No	No	No	No	No	No	No
Cluster	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Countries	153	153	153	155	141	141	153	122	153
Observations	62,389	6,127,468	6,135,664	6,328,372	5,786,381	5,964,467	6,011,132	719,178	5,416,486
$R^2$	0.448	0.545	0.543	0.557	0.544	0.544	0.518	0.444	0.550

Notes: In column 1, the unit of observation is in a two-digit HS industry, instead of a six-digit product. In column 2, quantity is used as an alternative measure of China’s exports. In column 3, we regard variable of interest $partne{r}_{CHN,jt}$ as a dummy. In column 4, we include the United States and Japan samples. In column 5, we remove those countries which newly established diplomatic relations with China during the period of 1995–2018. In column 6, we remove those countries which resumed diplomatic relations with China during the period of 1995–2018. In column 7, we winsorize outcome variables $\ln EX{P}_{CHN,jkt}$ at the 1% percentile. In column 8 and 9, we regress with the subsamples during the period of 1995–2000 and the period of 2001–2018. Coefficients are reported with t-statistics in parentheses. *, **, and *** signify statistical significance at 10, 5, and 1 percent levels, respectively. Standard errors are adjusted for clustering at the country level.

. In column 1, we aggregate export data to the two-digit HS industry level. The coefficient on partnership diplomacy is positive but only statistically significant at the 10% level. Still, partnership diplomacy has a significantly positive impact on China’s exports at the industry level. In column 2, the quantity is taken as an alternative measure of China’s exports. The coefficient of interest is similar in magnitude to the benchmark estimation in column 5 of Table 4 and remains statistically significant at the 1% level. In column 3, the variable of interest $partne{r}_{CHN,jt}$ is taken as a dummy to repeat the regression. As the estimated results in column 3 show, the coefficient on partnership (assigned the value 1) is −0.0497 but insignificant. The coefficient on comprehensive partnership (assigned the value 2) is 0.0452 but insignificant. Both coefficients on strategic partnership (assigned the value 3) and comprehensive strategic partnership (assigned the value 4) are positive and significant. It means that the positive impact is mainly driven by strategic partnership and comprehensive strategic partnership. Column 4 shows the results for a sample including the United States and Japan. The coefficient of 0.0649 is smaller than the coefficient of 0.0703 in the benchmark estimation. This suggests that the inclusion of the United States and Japan samples may generate a downward bias for the coefficient on partnerships and thus underestimates the impact of partnership diplomacy. The establishment of a partnership is based on the premise that these two countries already have formal diplomatic relations. Considering the viability of formal diplomatic relations during the sample period, we remove countries that newly established or resumed diplomatic relations with China. Column 5 shows the results for a sample without countries that newly established diplomatic relations with China during the period of 1995–2018. This includes Monaco, Bosnia and Herzegovina, Bahamas, Cook Islands, South Africa, Tonga, Timor-Leste, Dominica, Montenegro, Costa Rica, Niue, Malawi, South Sudan, Panama, Dominican, and El Salvador. For example, China and the Dominican Republic did not establish diplomatic relations until May 2018. Column 6 shows the results for a sample excluding countries that resumed diplomatic relations with China during the period of 1995–2018. It includes Niger, Central Africa, Guinea-Bissau, Northern Macedonia, Liberia, Grenada, Senegal, Chad, Gambia, Sao Tome and Principe, Burkina Faso, and Kiribati. For example, China broke off diplomatic relations with Burkina Faso for the Taiwan issue in 1994 and resumed diplomatic relations with it in 2018. Both coefficients of interest in columns 5 and 6 are positive and remain significant at the 1% level. Accounting for the potential problem sourced from outliers, we winsorize the explained variable $EX{P}_{CHN,jkt}$ at the 1% level in column 7. We check that the estimated results are robust for the winsorized samples. In columns 8 and 9, we check whether the estimated results are robust across time. Accession to the WTO in 2001 is a big event for China’s trade development. After WTO accession, China began to embrace openness and became an increasingly important role in world trade. Accordingly, we divide the sample period into two sub-periods: before 2001 and after 2001. The estimated coefficient on partnership diplomacy of −0.0003 is both economically and statistically insignificant for the period before 2001. There may be three reasons. First, China exported less before its accession to the WTO. According to the statistics from BACI, the export volume during the period of 1995–2000 only accounted for 3.02% of the total export volume during the period of 1995–2018. Second, China established partnerships with only nine countries before 2001. They are Brazil, Pakistan, Russia, Nepal, France, Canada, South Korea, Britain, and South Africa. Third, China progressively gained political and economic power after 2001. Consequently, the impact of partnership diplomacy on China’s exports during the period of 1995–2000 is not typical and representative. For the period after 2001, the coefficient of 0.0539 is positive and statistically significant. After the accession to the WTO, China’s economy and trade developed dramatically, which profoundly reshaped the global production network. Within the framework of partnership diplomacy, partner countries worked together to reduce bilateral trade costs and finally led to more trade flows. Only under the condition of “openness” can partnership diplomacy influence trade effectively. In conclusion, the positive and significant effect of partnership diplomacy on China’s exports is robust.

5.2. Endogeneity Concerns

We take at least two strategies to reduce the potential endogeneity problem. First, in order to obtain an unbiased estimate of interest, we try to control for all factors that may affect both China’s exports and partnership diplomacy as much as possible. We employ not only a vector of control variables but also country fixed effects and product-year fixed effects, which capture all time-varying country characteristics, time-invariant country factors, time-varying product characteristics, time-invariant product factors, and time-specific factors. Second, we take Two-Stage-Least-Squares (hereafter named the 2SLS) regressions to address potential endogeneity that arises from the reverse causal relationship between partnership diplomacy and trade. On the one hand, we hypothesize that well-established partnerships increase China’s exports to its partner countries. On the other hand, stronger commercial ties make trading partners maintain friendly political relations [48,49,50,51]. One primary concern is that there may have been a selection in targeting countries, in particular, the establishment and upgrading of partnerships are more common for those countries that trade more with China. The critical point in a 2SLS regression is to choose an appropriate instrumental variable (hereafter named the IV) that can sufficiently explain the potentially endogenous variable $partne{r}_{CHN,jt}$ but is not correlated to the error term in the second stage regression. Furthermore, there is an exclusion restriction for an appropriate IV. It suggests that an appropriate IV should have no direct impact on explained variable $EX{P}_{CHN,jkt}$. In other words, an appropriate IV should not affect explained variable $EX{P}_{CHN,jkt}$ through channels other than potentially endogenous variable $partne{r}_{CHN,jt}$. Accordingly, we employ the lag terms of the partnerships and the voting data from the United Nations General Assembly (hereafter named the UNGA) (the data are available from https://dataverse.harvard.edu/dataset.xhtml?persistentId=hdl:1902.1/12379# (accessed on 16 April 2022)) as IVs to reduce potential endogeneity in our empirical study.

We recode partnerships in the previous year, two years ago, and three years ago to obtain the variables of the lag terms $partne{r}_{CHN,jt-1}$, $partne{r}_{CHN,jt-2}$, and $partne{r}_{CHN,jt-3}$. The lag terms can be regarded as reasonable IVs for the current term $partne{r}_{CHN,jt}$. For the relevance with the endogenous variable $partne{r}_{CHN,jt}$, there is no doubt that what happened in the past could affect what happens in the present. The lag terms $partne{r}_{CHN,jt-1}$, $partne{r}_{CHN,jt-2}$, and $partne{r}_{CHN,jt-3}$ could sufficiently explain the current term $partne{r}_{CHN,jt}$ due to the serial correlation. For the exogeneity from the explained variable $EX{P}_{CHN,jkt}$, there is no direct relationship between the lag terms of the partnerships and China’s exports. The lag terms are also in line with the exclusion restriction. The lag terms of the partnerships can influence China’s exports only through the current term of partnerships. As reported in column 1–3 of

Table 7. Endogeneity: taking the lag terms and the voting data as IVs.

	(1)	(2)	(3)	(4)	(5)
	l1_Partner	l2_Partner	l3_Partner	Agree	All
partner	0.0871 ***	0.1066 ***	0.1335 ***	0.2398 ***	0.0795 ***
	(66.62)	(62.26)	(56.18)	(3.94)	(58.59)
lngdp	0.8956 ***	0.8940 ***	0.8743 ***	0.8585 ***	0.8878 ***
	(92.32)	(90.65)	(86.64)	(62.80)	(87.91)
lnpopulation	−0.4553 ***	−0.4407 ***	−0.4146 ***	−0.2971 ***	−0.4473 ***
	(−33.19)	(−31.61)	(−29.20)	(−5.07)	(−31.60)
cpi	−0.0004 ***	−0.0003 ***	−0.0008 ***	−0.0004 ***	−0.0008 ***
	(−6.84)	(−5.33)	(−6.33)	(−7.23)	(−6.47)
openness	−0.0007 ***	−0.0005 ***	−0.0003 ***	0.0001	−0.0005 ***
	(−6.83)	(−5.25)	(−3.16)	(0.28)	(−4.86)
bop	−0.0004 ***	−0.0004 ***	−0.0003 ***	−0.0003 ***	−0.0003 ***
	(−5.65)	(−5.03)	(−4.44)	(−2.99)	(−4.74)
urbanization	0.0178 ***	0.0169 ***	0.0154 ***	0.0097 ***	0.0175 ***
	(31.48)	(29.07)	(25.82)	(3.15)	(29.58)
electricity	−0.0018 ***	−0.0020 ***	−0.0020 ***	−0.0021 ***	−0.0018 ***
	(−8.97)	(−9.39)	(−9.55)	(−7.48)	(−8.35)
mobile	0.0028 ***	0.0027 ***	0.0027 ***	0.0024 ***	0.0028 ***
	(51.11)	(49.73)	(49.17)	(17.45)	(50.14)
Country FE	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Number of countries	153	153	153	153	153
Observations	6069545	5993277	5906755	6122171	5894935

Notes: We take the lag terms $partne{r}_{CHN,jt-1}$, $partne{r}_{CHN,jt-2}$, and $partne{r}_{CHN,jt-3}$ and the voting similarity index $agre{e}_{CHN,jt}$ as instruments to reduce endogeneity in columns 1–4, respectively. Column 5 takes the lag terms and the voting similarity index simultaneously as instruments. Coefficients are reported with t-statistics in parentheses. *** signify statistical significance at the 1 percent level. Standard errors are adjusted for clustering at the country level.

, the coefficients of interest in the second stage estimation are 0.0871, 0.1066, and 0.1335, which remain positive and statistically significant at the 1% level. All of them pass the tests of underidentification and weak identification. The 2SLS estimation with the IV of the voting data from the UNGA is reported in column 4. The voting data are a desirable IV for partnership diplomacy. Since its founding in 1946, the United Nations has conducted multiple rounds of voting on major international issues every year. The voting behaviors reflect one country’s state preference and political position [52,53,54]. Meanwhile, the similarity between two countries’ voting behaviors reflects political friendship or hostility. The voting similarity index $agre{e}_{ijt}$ is used as our IV to reduce potential endogeneity. $agre{e}_{ijt}$ is equal to the total number of votes where both countries agree divided by the total number of joint votes. The numerator here refers to agreement in votes recorded as 1, absences and abstentions as 0.5, and disagreement as 0. $agre{e}_{CHN,jt}$ denotes the voting similarity between China and country j in year t, which is the specific indicator of IV in our 2SLS estimation in column 4. For the relevance with the endogenous variable $partne{r}_{CHN,jt}$, the more similarities shared in voting behaviors, the more inclined China might be to establish partnerships with these countries. The likelihood that China establishes partnerships increases with the voting similarity. For the exogeneity from the explained variable $EX{P}_{CHN,jkt}$, there is no direct relationship between voting behaviors and trade. The voting similarity index is believed to be a comprehensive indicator of countries’ similarities in state preference and political positions. It covers at least six types of major international issues: the Palestinian conflict (19% of the total votes), nuclear weapons and nuclear material (13%), arms control and disarmament (16%), colonialism (18%), human rights (17%), and economic development (9%). The votes relating to economic developments only account for 9% of the total votes. The voting is the consequence of the games among countries, especially major powers. The IV of $agre{e}_{CHN,jt}$ is also in line with exclusion restriction. The voting similarities can affect China’s exports only through partnership diplomacy. The coefficient of interest in the second stage estimation is 0.2398 and remains statistically significant at the 1% level. It passes the underidentification test (Anderson LM statistic = 1740.281) and weak identification test (Cragg–Donald Wald F statistic = 1740.724). In column 5, the lag terms $partne{r}_{CHN,jt-1}$, $partne{r}_{CHN,jt-2}$, and $partne{r}_{CHN,jt-3}$ and the voting similarity index $agre{e}_{CHN,jt}$ are used as instruments simultaneously. The coefficient of interest in the second stage estimation is 0.0795 and still significant at the 1% level. It also passes the underidentification test, weak identification test, and overidentification test (Sargan statistic = 760.446). In summary, all coefficients of interest in 2SLS regression are larger than those in the benchmark. We find that partnership diplomacy still has an export-promoting effect when controlling for endogeneity.

6. Heterogeneity

6.1. Developed Countries vs. Developing Countries

There are large variations in social and economic development between developed and developing countries. It is interesting to estimate the heterogeneous effects of partnerships on China’s exports among countries. Due to weak infrastructure, unsound institutional systems, and low levels of economic development, developing countries are believed to suffer more from trade barriers. Partnership diplomacy makes partner countries pursue common economic interests and embrace openness. Thus, partnership diplomacy may play a more important role in developing countries that suffer more from trade barriers. In columns 1 and 2 of

Table 8. Heterogeneity: developed countries vs. developing countries.

	(1)	(2)	(3)	(4)
	HDI		OECD
	Developed	Developing	Developed	Developing
partner	0.0381 **	0.0930 ***	0.0315 **	0.0819 ***
	(2.22)	(5.06)	(2.29)	(5.00)
lngdp	1.0385 ***	1.0419 ***	0.9912 ***	1.1119 ***
	(3.33)	(4.36)	(3.71)	(5.29)
lnpopulation	−0.5869 *	−0.6060 *	0.8637	−0.5849 **
	(−1.69)	(−1.82)	(1.52)	(−2.56)
cpi	−0.0039 ***	−0.0004	−0.0106	−0.0006
	(−2.81)	(−1.22)	(−1.44)	(−1.49)
openness	−0.0030	0.0018	0.0020	0.0015
	(−0.68)	(1.34)	(0.54)	(1.18)
bop	−0.0014 **	0.0045	0.0007	−0.0008
	(−2.43)	(1.51)	(0.65)	(−1.14)
urbanization	0.0075	0.0222 *	−0.0049	0.0236 **
	(0.50)	(1.97)	(−0.38)	(2.19)
electricity	0.0254	−0.0022	0.1470 *	−0.0014
	(0.53)	(−0.67)	(2.03)	(−0.43)
mobile	0.0037 **	0.0016	0.0013	0.0033 ***
	(2.29)	(1.59)	(1.06)	(3.54)
Country FE	Yes	Yes	Yes	Yes
Product-Year FE	Yes	Yes	Yes	Yes
Cluster	Yes	Yes	Yes	Yes
Number of Countries	41	112	30	123
Observations	2,433,362	3,695,259	1,887,603	4,241,028
$R^2$	0.644	0.538	0.678	0.536

Notes: In columns 1 and 2, we define developed and developing countries based on the classification from the HDI. In columsn 3 and 4, we define developed and developing countries based on whether it is an OECD member. Coefficients are reported with t-statistics in parentheses. *, **, and *** signify statistical significance at 10, 5, and 1 percent levels, respectively. Standard errors are adjusted for clustering at the country level.

, we divide our sample into two subsamples of developed countries and developing countries based on the classification from the Human Development Index (hereafter named the HDI). The HDI is a composite measure compiled by the United Nations Development Programme to evaluate countries’ potentials for individual human development. It includes three major dimensions of human development: being knowledgeable, a long and healthy life, and having a decent standard of living. (The data are available at https://hdr.undp.org/data-center/human-development-index#/indicies/HDI (accessed on 24 May 2022).) When the HDI > 0.8, it is defined as a developed country for its high potential for human development; and otherwise defined as a developing one. For developing countries, the coefficient on partnership diplomacy is larger with a value of 0.0930 and statistically significant at the 1% level. Meanwhile, the coefficient is 0.0381 and statistically significant at the 5% level for developed countries. The point estimate for developing countries is more than twice of that for the developed countries. In columns 3 and 4, we use the classification from the Organization for Economic Co-operation and Development (hereafter named the OECD) to tell developed countries and developing countries apart. We regard OECD members as developed countries and the others as developing countries. Then, we repeat the regressions and analyses. The results are not substantively different from those in columns 1 and 2. We provide empirical evidence that China’s export expansion brought about by partnership diplomacy is most responsive to developing countries.

6.2. Different Continents

According to the criterion from China’s Ministry of Foreign Affairs website (the information is available at https://www.fmprc.gov.cn/mfa_eng/gjhdq_665435/ (accessed on 16 May 2022)), our sample includes 40 countries in Asia, 35 countries in Europe, 47 countries in Africa, 22 countries in America, and 9 countries in Oceania. Of these 153 sample countries, 91 countries have already established partnerships with China: 30 countries in Asia, 21 countries in Europe, 22 countries in Africa, 10 countries in America, and 8 countries in Oceania. It seems that China is more proactive in building a partnership network in Asia, Europe, and Africa. During the period of 1995–2018, China’s exports to Asia, Europe, Africa, America, and Oceania account for 36.43%, 40.05%, 5.75%, 13.69%, and 4.07% of the total, respectively. China mainly exports to Asia and Europe, which together account for 76.48% of the total. Both China’s partnerships and exports are very unevenly distributed among continents. Thus, we reasonably conjecture that the impact of partnership diplomacy on China’s exports may vary among continents. According to the estimated results in

Table 9. Heterogeneity: different continents.

	(1)	(2)	(3)	(4)	(5)
	Asia	Europe	Africa	America	Oceania
partner	0.0426 *	0.0451 **	0.1147 ***	0.1170 **	−0.0220
	(1.78)	(2.22)	(5.52)	(2.41)	(−0.78)
lngdp	0.6617 **	1.0560 ***	1.1296 ***	0.8700 *	2.1274 ***
	(2.04)	(3.36)	(7.45)	(1.76)	(7.48)
lnpopulation	−0.8732 ***	−1.1286	−0.4982	2.4909 *	1.1499
	(−3.28)	(−1.25)	(−0.83)	(1.80)	(1.22)
cpi	0.0012	−0.0008 **	−0.0011	−0.0056 *	−0.0001
	(0.36)	(−2.23)	(−1.28)	(−2.07)	(−0.02)
openness	0.0037	−0.0023	0.0012 *	0.0065	0.0029
	(1.10)	(−0.42)	(1.86)	(1.19)	(0.89)
bop	0.0056	−0.0012 **	0.0002	0.0082	−0.0102 **
	(1.16)	(−2.52)	(0.11)	(0.99)	(−2.77)
urbanization	0.0272 *	−0.0063	0.0096	0.0320	0.0431
	(1.79)	(−0.40)	(0.67)	(1.45)	(1.67)
electricity	−0.0024	−0.0283	0.0035	−0.0116	−0.0156 ***
	(−0.51)	(−1.57)	(0.61)	(−0.92)	(−4.93)
mobile	0.0019	0.0044 *	0.0026	−0.0017	0.0011
	(1.61)	(1.94)	(1.43)	(−1.06)	(0.79)
Country FE	Yes	Yes	Yes	Yes	Yes
Product-Year FE	Yes	Yes	Yes	Yes	Yes
Cluster	Yes	Yes	Yes	Yes	Yes
Number of Countries	40	35	47	22	9
Observations	1,794,873	1,876,055	1,301,443	885,358	225,360
$R^2$	0.561	0.662	0.548	0.643	0.777

Notes: In columns 1–5, we run regressions in the subsample of Asia, Europe, Africa, America, and Oceania. Coefficients are reported with t-statistics in parentheses. *, **, and *** signify statistical significance at 10, 5, and 1 percent levels, respectively. Standard errors are adjusted for clustering at the country level.

, only the coefficient of Oceania is negative but insignificant. The coefficients of Asia, Europe, Africa, and America are all statistically significant, but at different levels: Africa at the 1% level, Europe and America at the 5% level, and Asia at the 10% level. The point estimates from column 1 to 4 are 0.0426, 0.0451, 0.1147, and 0.1170, respectively. The impact of partnership diplomacy on China’s exports appears to be larger for Africa and America. In conclusion, the exporting-promoting effect of partnership diplomacy is mainly driven by Africa and America, which are geographically distant from China.

6.3. Heterogeneous Industries

The impacts of international relations on trade may vary across industries [16,17,35]. As in Hypothesis 2, we conjecture that partnership diplomacy should have a greater impact on China’s exports in industries in which China has a comparative advantage. Based on the two-digit HS industry criterion, we divide the sample into 21 groups and run a separate regression for each industry. As reported in

Table 10. Heterogeneity: different industries.

HS2	Industry	HS6	Share	$\mathit{\beta }$	$\mathit{SE}$	t	p
1	Animal products	01–05	0.64%	0.0444 **	0.0216	2.06	0.041
2	Vegetable products	06–14	1.14%	0.0588 ***	0.0155	3.80	0.000
3	Animal or vegetable fats and oils and articles thereof	15	0.04%	0.0624 ***	0.0218	2.87	0.005
4	Prepared foodstuffs; beverages; tobacco	16–24	0.93%	0.0470 ***	0.0149	3.15	0.002
5	Mineral products	25–27	1.72%	0.0568 ***	0.0150	3.78	0.000
6	Products of the chemical or allied industries	28–38	5.55%	0.0622 ***	0.0148	4.21	0.000
7	Plastics, rubber and articles thereof	39–40	3.82%	0.0813 ***	0.0171	4.75	0.000
8	Raw hides, leather, furskins and articles thereof	41–43	1.54%	0.0423 **	0.0197	2.14	0.034
9	Wood and articles thereof	44–46	0.70%	0.0580 ***	0.0170	3.42	0.001
10	Pulp; paper and articles thereof	47–49	0.95%	0.0586 ***	0.0144	4.06	0.000
11	Textile and articles thereof	50–63	12.11%	0.0890 ***	0.0174	5.12	0.000
12	Footwear, headgear, umbrellas and parts thereof	64–67	2.87%	0.0957 ***	0.0189	5.08	0.000
13	Articles of stone, ceramic and glass	68–70	2.00%	0.0801 ***	0.0161	4.98	0.000
14	Precious metal and articles thereof	71	0.42%	−0.0095	0.0176	-0.54	0.591
15	Base metals and articles thereof	72–83	9.05%	0.0667 ***	0.0165	4.03	0.000
16	Machinery and electrical equipment	84–85	43.17%	0.0715 ***	0.0169	4.22	0.000
17	Transport equipment	86–89	3.89%	0.0695 ***	0.0173	4.03	0.000
18	Optical, measuring, watches and parts thereof	90–92	2.90%	0.0560 ***	0.0165	3.39	0.001
19	Arms, ammunition and parts thereof	93	0.02%	0.0856 **	0.0378	2.26	0.025
20	Miscellaneous manufactured articles	94–96	6.53%	0.0533 ***	0.0175	3.05	0.003
21	Works of art, collectors’ pieces and antiques	97	0.02%	0.0681 **	0.0270	2.52	0.013
22	Reserved for special uses by contracting parties	98	–	–	–	–	–

Notes: We run regressions in the subsample of two-digit HS industries. Coefficients are reported with t-statistics in parentheses. ** and *** signify statistical significance at 5 and 1 percent levels, respectively. Standard errors are adjusted for clustering at the country level. Share denotes the share of each industry in China’s total exports during the period of 1995–2018.

, China’s export expansion is concentrated among the following industries: “machinery and electrical equipment” (43.17% of China’s total exports during the period of 1995–2018), “textile and articles thereof” (12.11%), “base metals and articles thereof” (9.05%), “products of the chemical or allied industries” (6.53%), and “transport equipment” (5.55%). Partnership diplomacy has a significantly positive impact on almost all industries, except “precious metal and articles thereof”. The coefficient of interest for “precious metal and articles thereof” is −0.0095 but insignificant. The effect of partnership diplomacy on China’s exports is greater for the following industries: “footwear, headgear, umbrellas and parts thereof” ($\beta =0.0957$), “textile and articles thereof” ($\beta =0.0890$), “arms, ammunition and parts thereof” ($\beta =0.0856$), “plastics, rubber and articles thereof” ($\beta =0.0813$), “articles of stone, ceramic and glass” ($\beta =0.0801$), and “machinery and electrical equipment” ($\beta =0.0715$). These six industries account for 63.99% of China’s total exports during the period of 1995–2018. The largest impact of partnership diplomacy is in “footwear, headgear, umbrellas and parts thereof”, whose estimated coefficient is 0.0957. For these two most important industries “machinery and electrical equipment” (43.17%) and “textile and articles thereof” (12.11%) in China’s exports during the period of 1995–2018, the impacts are also relatively sizable. The coefficients of interest are 0.0715 and 0.0890, respectively. Evidence from industry heterogeneity analysis support the Hypothesis 2 that the export-promoting effect of partnership diplomacy is mainly driven by industries in which China has a comparative advantage, such as “machinery and electrical equipment” and “textile and articles thereof”. Partnership diplomacy promotes China’s exports by reducing bilateral trade costs.

7. Conclusions

The use of active diplomacy tactics could explain the rapid development of China’s trade. In the new era, China has actively developed a global partnership network on the basis of equality and mutual benefit. Partnerships between countries guarantee lasting friendly political relations in the long term. This paper investigates the role of partnership diplomacy on China’s exports. According to the information from the official statements on China’s Ministry of Foreign Affairs website, we construct a dataset of China’s partnerships. Based on this dataset, we conduct some empirical analysis and provide evidence that partnership diplomacy can promote China’s exports. The significantly positive effect remains robust when controlling for endogeneity with IVs of the lag terms and the voting data. We also perform several sensitive tests for robustness as follows: aggregating export data to the two-digit HS industry level; taking quantity as an alternative measure of China’s exports; taking variable of interest $partne{r}_{CHN,jt}$ as a dummy; including the sample of the United States and Japan; removing the sample of countries that newly established or resumed diplomatic relations with China; using the winsorized sample at a 1% level; and dividing the sample period into sub-periods. Heterogeneity analysis shows that partnership diplomacy matters more in developing countries and countries in Africa and America. Most importantly, mechanism analysis shows that the positive effect of partnership diplomacy is stronger for products and industries in which China has a comparative advantage. According to the law of comparative advantage, it proves that partnership diplomacy promotes China’s exports by reducing bilateral trade costs with its partner countries. The export-promoting effect driven by industries such as “machinery and electrical equipment” (43.17% of China’s total exports during 1995–2018) and “textile and articles thereof” (12.11%) further supports the mechanism of reducing bilateral trade costs. In conclusion, partnership diplomacy is interest-driven rather than ideology-driven and plays a positive role in China’s export expansion.

Finally, there are at least two policy implications. First, continue to conduct partnership diplomacy and develop a global partnership network. This paper provides empirical evidence that partnership diplomacy strengthens the trade links between partner countries. As there are more trade frictions and ideological conflicts among countries nowadays and China becomes an emerging power in the world, partnership diplomacy plays a more proactive role in bridging the ideological gaps and enhancing communication and coordination between governments. It is believed that partnership diplomacy is economically beneficial for partner countries. Second, insist that partnership diplomacy promotes exports by economic measures rather than administrative measures. This paper confirms the mechanism of reducing trade costs and increasing openness. Within the framework of partnership diplomacy, China and its partner countries are supposed to sign bilateral trade agreements, construct free trade areas, expand the use of local currency, strengthen financial cooperation, promote connectivity in the region, and so on.