The Determinants of Export Performance:
Evidence from Chinese Cities
Name: Luxuan Zhou
Student number: 10826513
Date: 29 June 2015
MSc Business Administration: International Management
Final Version Master Thesis
Supervisor: Dr. Niccolò Pisani
Second Supervisor: Dr. Johan Lindeque
Statement of originality
This document is written by Student Luxuan Zhou who declares to take
full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original
and that no sources other than those mentioned in the text and its
references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the
supervision of completion of the work, not for the contents.
ABSTRACT
When reviewing the extant export literature, most of the studies investigate export behavior at either country level or firm level. The research topics are mainly about determinants of export behavior or drives of export performance. However, little available export study is conducted at city level. Based on the fact that there are significant differences among Chinese cities in terms of social and economic life, this paper is going to investigate export behavior at city level and examine the relationship between city characteristics and export performance as well as the moderating effects of investment in environment protection and infrastructure. The empirical study is based on the data from 278 Chinese cities. The results show that GDP is positively related to the total value of foreign trade export. Besides, the moderating effects of investment in environmental protection and infrastructure are significant on the relationship between GDP and the total value of foreign trade export.
TABLE OF CONTENTS
1. INTRODUCTION ... 5
2. LITERATURE REVIEW ... 7
2.1 Determinants of export at inter-country level ... 8
2.2 Determinants of export at intra-country level ... 11
2.3 Recent reflection on determinants of export ... 15
2.4 Research gap ... 17
3. THEORETICAL FRAMEWORDS ... 18
3.1 Gross domestic product ... 18
3.2 Trade openness ... 20
3.3 R&D intensity ... 22
3.4 The moderating effect of investment in environment protection ... 23
3.5 The moderating effect of investment in infrastructure ... 25
4. METHODOLOGY... 27
4.1 Sample and data collection ... 27
4.2 Variables ... 29
4.2.1 Dependent variables ... 29
4.2.2 Independent variables ... 30
4.2.3 Moderating variables ... 31
4.2.4 Control Variables ... 31
4.3 Statistical analysis and results ... 31
5. DISCUSSION... 40
5.1 Academic relevance ... 41
5.2 Policy implications... 43
5.3 Managerial implications ... 44
5.4 Limitation and suggestions for future research ... 45
6. CONCLUSION ... 46
AKNOWLEDGEMENT ... 49
REFERENCES ... 50
APPENDIX ... 55
LIST OF FIGURES Figure 1. Theoretical Frameworks ... 27
Figure 2. Moderating effect of investment on environment protect on total value of foreign trade exports... 36
LIST OF TABLES Table 1. Determinants of export at inter-country level ... 11
Table 2. Determinants of export at intra-country level... 15
Table 3. Descriptive Statistics; Means; Standard Deviation and Correlations ... 37
Table 4. Regression Results ... 38
1. INTRODUCTION
Export has always been the main focus and timeless topic in both international trade theory and macroeconomic development field. In addition to being a popular theme in the academic realm, it is also remarkably noticed by enterprises and policy-makers. Both of the scholars and policy-makers have recognized the considerable potential benefits deriving from export and they have made relevant strategies or policies to promote export accordingly (Katsikeas & Leonidou , 1996).
When it comes to entry mode strategy, firms are more likely to choose export since it is the easiest and most common approach to explore foreign markets at the first stage of internationalization (Johanson & Vahlne, 1979, 1990). There are also several empirical studies trying to evaluate export performance and examine the relationship between export behavior and firm performance. For example, Katsikeas et al. (2000) come up with a contingency approach to measure export performance by reviewing more than 100 studies. In addition, Bonaccorsi (1992) studies the impact of firm size on export intensity. Although Calof (1994) and Dhanaraj and Beamish (2003) make extra effort to explore the impact offirm characteristics and resources on export behavior, there is still little discussion on the influence of institutional differences at the country level to explain firms’ export behavior in different contexts.
Going beyond export research at the firm level, many scholars try to shift their unit of analysis to country and investigate the causality between export and economic growth in a specific context (Chow, 1987; Kavoussi, 1984; William 1981). Subsequent export studies take a step forward to evaluate export behavior in both developed and developing economies by identifying significant institutional
differences that will influence the export behavior (Li et al., 2013).
Based on the above discussion, we observe that current export literature mainly focuses on firm level and country level from the perspective of resource-, industry- or institution-based view. Little attention has been paid to export behavior at the city level. However, we should notice that city level study is also important in understanding international business. That is because the trend of globalization has penetrated from country level to city level. It can also be reflected in academic research that the literature in the field of urban studies as well as economic development has shed more light on how cities contribute to promoting international business by utilizing and developing their advantages. For example, the conceptualization of global cities put forward by Sassen (1991) has drawn more attention to the effect of cities on the process of globalization.
While import and export are two key indicators to estimate the economic openness of the target location (Harrison, 1996), current research has failed to figure out specific factors that determine export performance at city level. Therefore, this paper is going to tackle this research gap and investigate the relationship between city characteristics and export behavior. To make it more specific, we select 278 cities from China as research objects and try to find out the determinants and drivers of export behavior and performance at those Chinese cities.
There are favorable reasons for choosing Chinese cities as research objects to conduct this study. To begin with, China enjoys huge national territory area, which results in enormous geographic and cultural disparity among cities. Furthermore, the
institutional factors such as government policies are the most interesting and relevant to explain export behavior and performance in different cities. That is because although all cities grow and develop within the same national environment, there are still a lot differences with respect to resource endowment, economic growth and infrastructure construction in different cities due to institutional factors. For instance, the “Open Door” reform policy issued by Chinese government in 1987 has promoted overall national foreign business while bringing more inequality in resource distribution and economic performance among cities.
This research makes significant contributions to the literature. First, it fills up the research gap by studying export behavior at city level in the international business field and relates trade theory to urban studies. Second, it collects the data of 278 Chinese cities and conducts empirical study to test the hypotheses. Third, the findings carry out important implications for both export-oriented firms to make location decision and policy-makers to promote export behavior in specific cities.
This paper proceeds as follows. The next section will review the background literature and theories about export. The third section will describe theoretical framework and present the hypotheses. The following section will explain the approach to run empirical study and present the data. Last but not least, the fifth section will discuss the results and the final section ends up with conclusion.
2. LITERATURE REVIEW
Exporting is considered to be one of the oldest and most convenient forms of doing business aboard. That is because it requires the least resources and commitment
but allows more flexibility and involves less risk compared to other modes of entry such as license and equity-based investments (Cavusgil, 1984; Chu & Anderson, 1992). Accordingly, it has become one of the most popular research subjects within the academic discipline of international business and is favored by firms especially those of small-to-medium size from both developed and developing countries to explore the foreign market opportunities (Cavusgil, 1984; Coudounaris，Katsikeas & Leonidou, 2009). As a result of continuous endeavor contributed by renowned scholars, the export literature has achieved substantial progress in terms of quality improvement, empirical methodology and topic scope (Coudounaris，Katsikeas & Leonidou, 2009).
2.1 Determinants of export at inter-country level
By reviewing the development of export literature, one can recognize that there are three key evolution processes, namely traditional theory in 19th century, new trade theory (NTT) in 20th century and new new trade theory (NNTT) in 21st century. The traditional theory is mainly developed from Ricardo’s (1817) neoclassical model. He puts forward the concept of comparative advantages and argues that potential gains from international trades for countries are attributed to comparative advantages owned by one of the trade partner. For example, for one country which has overall advantages in two commodities, it will export the one with comparative advantages over its partner and import the other. Based on Ricardo’s (1817) neoclassical model, Ohlin (1933) enriches traditional trade theory by introducing factor endowment theory. He argues that the decisions of exporting or importing made by one country depend
on the scarcity and intensify of the factors owned by that country for making the products. Therefore, one country will export the products made of abundant factors and import those made of relatively scarce factors.
Both Ricardo (1817) and Ohlin (1933) explain the inter-industry export behavior under the assumption of perfect competition and believe the flow of goods is beneficial for the countries involved in the international business. However, the validity of perfect competition is challenged at the presence of specialization and intra-industry trade. To resolve the ambiguity, Krugman (1980) comes up with NTT which takes economies of scale, demand-supply model and technology differentiation into account and assumes international activities under imperfect competition. According to NTT, there are more microeconomic factors rather than production factors that stimulate the occurrence of international trade.
Although NTT takes a step further by providing richer insight on the motivation of international trade, the unit of analysis is still set at the country level. However, the drivers and patterns of export cannot be fully understood unless extending the unit of analysis from country level to firm level. While both traditional theory and NTT fail to link firm level characteristics to export behavior, the NNTT emphasizes the necessity of firm heterogeneity to account for variation in export performance among firms at intra-country level (Bernard et al., 2003; Helpman, Melitz & Yeaple, 2003; Melitz, 2003). Productivity is one of the most influential factors that distinguish exporting firms from non-exporting firms (Bernard et al., 2003; Melitz, 2003).
trade as a whole although the benefits are not even for every country (Subramanian & Wei, 2007). Thus, the patterns and intensity of competition have changed. In addition, the fluctuation of economy becomes more severe in recent years. It seems that the above three main international trade theories are not powerful enough to explain current export modes among countries. Alternatively, some scholars have tried to understand the motivation of export from other angles. Based on the recent export literature, free market mechanism (Gao et al. 2010; Shinkle & Kriauciunas, 2010; Yang & Malick, 2014), inflation (Greenaway et al., 2010; Yang & Malick, 2014) and financial crisis (Navarro et al., 2014; Yang & Malick, 2014) are identified as significant factors that will influence export behavior and performance at inter-country level.
Table 1. Determinants of export at inter-country level
2.2 Determinants of export at intra-country level
Based on the literature focusing on understanding about export behavior and performance at intra-country level, we raise the same reasoning on what are the drivers of export behavior and determinants of export performance at intra-country level. Instead of concentrating on the differences among countries, the intra-country level export studies are based on the assumption of heterogeneity among firms.
manager characteristics are associated with the variation in firms’ export behavior. He also identifies the managerial expectation of growth from export as the primary factor to promote the involvement of export activity. In addition, he suggests that the macro-level factors such as public policy are ineffective to promote export activity when the expected growth and market development estimated by managers are negative and unfavorable. Rather than treating external environment as a contingent factor when studying export behavior, Navarro et al. (2014) conduct an empirical study to verify whether external environment factors such as turbulence and the competitiveness of the foreign market will influence export behavior and performance. Provided the extent empirical studies on export stimuli are not systematic and well-organized, Leonidou et al. (2007) provide a more comprehensive insight on what induces smaller firms to export activities by reviewing representative empirical studies and critically ranking those simulating factors according to the degree of their impact. They have further summarized the drivers of smaller firms’ export behavior and labeled them into internal and external factors in their analytical review. In line with Cauvsgil’s (1984) findings, Leonidou et al. (2007) conclude that the expected extra sales and growth from export strongly motivate export behavior while ownership of unique products, the need for less relying on domestic market and uninvited orders from foreign clients exert equal high impact on inducing exporting.
In addition to the drivers of export behavior, it is also important to review the determinants of export performance. As it is mentioned above, Bernard et al. (2003) and Melitz (2003) begin to shift the unit of analysis from country level to firm level
and broaden our insight on export behavior by combining the consideration of sunk cost and heterogeneity residing in each firm. They link firm productivity to their export behavior and conclude that more productive firms will self-select into export activities. Based on their argument and going beyond that, Greenaway and Kneller (2007) find that the direction of relationship between productivity and export is contentious. They believe it is probably an accepted fact that ex-ante productivity determines the propensity of export behavior (Melitz, 2003; Martins & Yang, 2009). However, one could also argue from the opposite causality direction perspective that it is the exposure to export market that increases firms’ productivity i.e. ex-post productivity. Alternatively, Greenaway and Kneller (2007) provide a critical review by considering factors of firm heterogeneity, trade and FDI and find out that exporting firms are likely to become larger and more productive than non-exporters. Moreover, they recognize that FDI and export are complementary and thus export is not the only strategy for productive firms.
R&D is another significant firm attribute that raises scholars’ attention. While most of the literature deals with the relationship between export and R&D by assuming one can be explained by the other, Esteve-Pérez and Rodríguez (2013) extend our understanding by studying the interdependency between export and R&D. According to their empirical study, export and R&D is interrelated to each other and the success of export (R&D) is path dependent linking to the previous export (R&D) decision and investment. Furthermore, firm size is considered to be another stimulating variable to examine export behavior. Bonaccorsi (1992) contends firm size
is an effective parameter influencing the export behavior in terms of scale of economies, resource limitation and risk evaluation from managers. According to his reasoning, he criticizes the widely accepted belief that firm size is positively associated with export intensity and argues that network system and external resources are more necessary support for small firms to engage in export activities. Since export behavior cannot be explained by a single variable, a more comprehensive conceptual framework is required. Concentrating on emerging economies and based on resource-based view, Singh (2009) comes up with a more extensive framework to explain export activities by identifying the importance of firm size, R&D, marketing investment and business network.
Table 2. Determinants of export at intra-country level
2.3 Recent reflection on determinants of export
According to North (1991), it is not reliable to entirely focus on the internal environment to explain the economic activities since external environment especially the institution plays an important role in influencing the performance of economies. He argues that both informal institutions and formal institutions are vital factors that structure the interactions between different business participants. In line with North (1991) and linking to the knowledge of firm heterogeneity, Shinkle and Kriauciunas (2010) explore how export growth will be affected by firm size and age and what is
the role of the degree of free-market mechanism. They conclude that economic institutions do matter and it is more critical to combine institutional factors with firm characteristics to understand export growth. Following the same reasoning, Gao et al. (2010) elaborate on the previous literature and create a “strategy tripod” framework which integrates resource-, institution-, and industry-based perspective to uncover the drivers of export propensity and intensity as well as the determinants of export performance. Their study largely extent our knowledge of exporting with respect to the impact of institutional environment on emerging market, considering both domestic enterprises and foreign subsidiaries in one nation and indicating the financial performance from export activities.
At intra-country level study, Li et al. (2013) are among the few researchers that shed light on the importance of institutions and pay a closer look on the influence of quality of domestic legal institutions at the city level. What is worth mentioning, the notable advantage of their work is that they explore how location differences can make an impact on export performance by concentrating on 120 Chinese cities instead of country. Moreover, they try to link the institutional differences to market characteristics and further disclose the trade barriers can be caused by the quality of legal system such as information asymmetries. While institutional impact is strongly advocated by the above scholars, Bernard and Jenson (2004) find out the national subsidy on export promotion has little influence on enhancing export performance.
Although it is such a breakthrough in the academic research to focus on firm heterogeneity to explain export behavior, Yang and Malick (2014) recently question
on extant mixed findings on the relationship between firm level productivity and export behavior. They believe the mixed results in the empirical studies can be explained by the influence of country-specific macroeconomic factors, which again encourages us to concern with the variation in the external environment to explain export activities. Similarly, Greenaway et al. (2010) investigate the relationship between real effective exchange rate and export sales and demonstrate there is negative effect between them. Coming after the above reasoning, Yi (2014) narrows down the study to a Chinese province and argues that the determinants of export activities are context specific. He enriches the research by considering location advantages derived from institutional differences and agglomeration effect.
2.4 Research gap
By reviewing the export literature, it is easy to recognize both country level and firm level studies. However, the city level analysis is entirely neglected with the only exception of Li et al. (2013) who attempt to explain export differences at city level. It is one of the most important studies to help us identify the research gap in the export literature. Therefore, it is of great significance to fill up the gap and make more effort on the city level studies. Moreover, extant literature also provides the direction and hints for such research. For example, the variable of firm size can be translated into city size while the productivity can be interpreted as gross domestic product (GDP) of each city.
This paper is going to fill the gap in the export literature by studying determinants of export performance at city level. By doing so, this paper can not only
extend the scope of analysis of the extant export literature, but also link the trade theory to the urban studies discipline to provide a more comprehensive understanding on the relationship between trade growth and city development. Besides, the analysis is useful for export-oriented firms to select ideal location and helpful for policy-makers in pursuit of boosting export performance of one city. The theoretical framework will be explained in the following section.
3. THEORETICAL FRAMEWORDS 3.1 Gross domestic product
Productivity is considered to be one of the most outstanding determinants that induce export behavior at both inter-country and intra-country level study. According to Ricardo’s (1817) comparative advantage theory, export flows from a country which holds comparative advantages to another country. To be concrete, it means that those countries which can produce goods with less cost and better quality will fully utilize their strengths and export goods to countries which are not as good as them at producing those goods efficiently. In line with Ricardo’s comparative advantages theory, Ohlin’s (1933) factor endowment theory shares the same philosophy as his. More specifically, Ohlin (1933) believes that the export countries are able to make a certain product more efficiently than its trade partner because they possess some production factors in abundant quantity to increase productivity, such as labor and land. Thus, productivity is one of those comparative advantages that drive export propensity of a country. Moreover, from the perspective of Krugman (1980), productivity affects export activities indirectly by influencing economies of scale. He
argues that countries will export those goods with a large domestic market since it can realize the economies of scale to increase productivity.
When it comes to the firm level analysis, productivity is also identified as a significant factor that leads to firm heterogeneity and determines export behavior and performance. It is tested that there is positive relationship between productivity and firms’ export behavior. The more productive the firm is, the more likely it will expose to international trade (Bernard & Jensen, 1999a; Bernard et al., 2003). In Melitz’s (2003) dynamic intra-industry model, he further explains how firms will be selected and survives in the international trade. His empirical study not only confirms that productive firms will be more likely to enter international trade market than less productive firms, but also finds out that the least productive firms will be kicked out because of the inter-firm resource reallocation. Furthermore, Arnold and Hussinger (2005) collect the data of German manufacturing firms to demonstrate that productivity will facilitate export while export is not a stimulating factor to promote productivity.
According to the above empirical studies, input and output are the important variables to measure productivity. It can be referred to Melitz’s (2003) intra-industry model that higher productivity means producing the same quality at less cost or producing product of higher quality at the same cost. While the output of a firm represents the productivity of that firm, the GDP of one city reflects the productivity of that city. Xu (1996) conducts an empirical study on 32 economies to figure out the causality between export and GDP. Nearly half of the samples show that export
strongly promote GDP. However, Henriques and Sadorsky’ (1996) study on Canada concludes that real export, real terms of trade and real GDP are inter-related. In addition, the case of Canada verifies that it is the change in GDP will influence the change in export. Similarly, focusing on the city level study, we are going to explore the relationship between GDP and the value of foreign trade export in different Chinese cities.
H1: GDP is positively related to the value of foreign trade export at the city level.
3.2 Trade openness
The level of openness at the macro-level can make a difference on export performance at the micro-level since openness can reduce trade barriers and facilitate trade liberalization (Mansfield & Reinhardt, 2008). Although there are different measurements to evaluate the level of openness, this variable does play an important role in international trade.
From the perspective of macro-level analysis, Leonidou et al. (2007) identify openness with respect to relief of foreign protection rules and regulations as an attractive external stimulus for export. From their perspective, external stimuli refer to factors excluding those characterizing from firm-heterogeneity. Following the similar reasoning, there are other scholars trying to relate openness to institution-based view. They argue that formal institutions will make an impact on free market mechanism and intermediate institutions at the emerging countries such as China, which will further influence export behavior at the firm level (Gao et al., 2010). More specifically,
they believe institutional environment in terms of both formal institution and informal institution will influence firms’ international strategy and affect their export behavior. Moreover, the impact of institutional environment is more salient in emerging economies as those countries experience significant adjustments and changes in their economic transition (Gao et al., 2010; Shinkle & Kriauciunas, 2010).
Furthermore, the country level macroeconomic factors are even considered to be more reliable to explain the differences in export performance across countries (Yang & Malick, 2014). In their point of view, the extent mixed findings on the relationship between productivity and firm level export performance is due to the negligence of macroeconomic factors such as trade openness and financial reform at the country level. Their empirical study shows that the differences in post-entry productivity and effect of learning by export can be accounted by both country level and firm level factors.
Based on the above discussion, we believe trade openness also significantly affects export performance at the city level in China. Due to government control and regulation policy, there is considerate inequality in economic development among Chinese cities. For example, the “Open Door” reform policy aiming at facilitating international trade and attracting FDI has benefited most of the coastal cities while leaving inland cities falling behind at the same time. Therefore, it is assumed there is positive relationship between the level of openness and export performance.
H2: Trade openness is positively related to the value of foreign trade export at the city level.
3.3 R&D intensity
The impact of R&D on export activities significantly manifests at intra-country level. If comparative advantages are the premise for the occurrence of export behavior, investing in R&D is the effective way to sustain those advantages. As advocated by strategists of companies and scholars, R&D aims at improving and enhancing existing method and technology in order to maintain firm’s competencies (Franko, 1989; Ito & Pucik, 1993). The relationship between R&D intensity and export performance has already been verified by Hirsch and Bijaoui (1985). In their model based on empirical analysis, R&D intensity refers to creation of proprietary knowledge as well as relative innovative activities. Their regression model confirms R&D intensity is positively associated with export performance at firm level.
Faced with a lot of uncertainties in the external environment such as fluctuation in exchange rate and product imitation by other competitors, Ito and Pucik (1993) believe pursuing excellence in technology provides Japanese manufacturing firms a more reliable way to succeed in international trade competition. Based on the cross-sectional study of manufacturing firms in Japan, they find that R&D expenditures are positively related to export sales. Besides, they discover that the impact of R&D investments is accumulative to export performance, which is in line with Esteve-Pérez and Rodríguez’s (2013) findings. That is to say, previous R&D expenditures are related to subsequent R&D expenditure, which are beneficial to firms’ export sales over time.
explore the relationship between R&D and export. Their study confirms the neotechnology theory of trade which argues that R&D, as an indicator of technology, is a prominent factor to account for export activities. According to their empirical study based on the data of Chinese manufacturing firms, they find out that R&D is vital to the propensity of export activities as well as the intensity of export. Besides, they also prove the reciprocal relationship between R&D intensity and export activities.
Nevertheless, Lefebvre et al. (1998) interpret the influence of R&D intensity on export performance from another angle. Although they support the belief that R&D spending can be the indicator of the level of technology, they don’t believe the level of investment in R&D contributes to strengthen the competency of firms. What they critically emphasize is that R&D related capabilities resulting from various forms of R&D spending determine export performance on different dimension.
Similarly, it raises the question of whether R&D expenditure or intensity is also an important determinant of export performance at city level. Due to the availability of data, we refer R&D intensity to R&D expenditure. In order to test it, we examine the relationship between R&D expenditure of each city and their export levels.
H3: R&D intensity is positively related to the value of foreign trade export at the city level.
3.4 The moderating effect of investment in environment protection
Many developing countries chase for economic growth by the cost of deteriorating environment in their initial phase of development. According to an
empirical study conducted by Ye and Yu (2007), the dramatic development of export in China doses bring severe pollution problem. That is because developed countries may relocate their factories to developing countries for production since developed countries take the advantages of low environment protection requirement in developing countries. Following the reasoning, if there are more investments in environment protection, it implies that those developing countries may enter the higher phase of economic development. Building on this, Deng and Song (2008) find out that the relationship between environment pollution and GDP per capita can be described as a “U” shape. Therefore, it can be assumed that while there is more investment in environment protection, the economic growth can be expected to be faster. Since export level is also hypothesized to be related to GDP, we assume that the investment in environment protection will moderate the relationship of GDP and export level.
The concern of environment protection is also raised on the level of trade openness. Frankel and Rose (2005) question on whether trade does good for environment and try to clear out the causality. Their empirical study shows trade can significantly reduce the values of certain air pollution indicators. In addition, the harmful effect of trade on environment is subtle. Besides, Managi et al. (2009) also run a study to testify whether trade openness plays a role in improving environment quality. Their results show that the positive correlation happens to member country of Organization of Economic and Co-operation Development (OECD) while the adverse effect of trade openness on environment occurs in non-OECD countries. Based on the
above argumentation, we could posit that when there is more investment in environment protection, the level of trade openness also becomes higher. Consequently, we hypothesize that the moderating effect of investment in environment protection can occur to the relationship between trade openness and export level.
H4a: Investment in environmental protection positively moderates the relationships hypothesized in H1.
H4b: Investment in environmental protection positively moderates the relationships hypothesized in H2.
3.5 The moderating effect of investment in infrastructure
Export involves in different forms of mobility including capital mobility, goods mobility and labor mobility (Mundell, 1957). Since export is a mean of goods exchange, the convenience of goods mobility plays an important role in influencing export behavior. Traditionally speaking, export means goods flow from home country to host country. In the process of movement, it requires well-organized transportation infrastructure as well as the capability of electricity generating and telecommunication. In that sense, the quality of goods mobility will be promoted or impeded by quality of infrastructure such as network development of transportation. As identified by Gripsrud (1990), transportation cost can be a barrier and negative factor when considering exporting to a distant country.
Some scholars try to explore the interaction between GDP and investment in infrastructure. For example, focusing on the regions in Mexico, Looney and
Frederiksen (1981) find that investment in infrastructure does influence the GDP of different regions. What is worth mentioning, their analysis shows that whether the impact is positive or negative rely on the type of investment (i.e. economic or social overhead capital) and regions (i.e. intermediate or lagging). In addition, Canning and Pedroni (2004) also conduct a study aiming at figuring out whether infrastructure construction will influence economic growth in the long term. Their empirical results confirm the positive correlation between infrastructure provision and long-term economic growth. Therefore, we believe the quality of infrastructure moderates the relationship between GDP and export level.
According to the Report on the Work of the Government released by Chinese Government in 2015, it puts forward many policies to support and facilitate “Opening up” reform and advocates for maintaining trade openness. One of those policies emphasizes the sustainable importance of the quality of infrastructure and insists on investing on infrastructure such as transportation and public utilities. The quality of infrastructure has become one of the favorable factors to attract international trade. Taking transportation as an example, China covers huge geographic area, which makes transportation critically important for mobility and connecting the outside world. However, cities located in different areas enjoy different level of convenience and development in public transportation with respect to modes of transportation and quality of construction. Reflecting on the history of transportation development in China, policy-makers will invest transportation construction in the regions with more open trade environment first. The network of transportation will extend to other
relatively remote regions subsequently. Thus, we assume the quality of infrastructure can reflect the level of trade openness and furthermore moderate the relationship between trade openness and export level.
H5a: Investment in infrastructure positively moderates the relationship hypothesized in H1.
H5b: Investment in infrastructure positively moderates the relationship hypothesized in H2.
Figure 1. Theoretical Frameworks
4. METHODOLOGY 4.1 Sample and data collection
The research question of the thesis is to find out the determinants of export performance in 278 Chinese cities and we use a panel research design to test relationship between city characteristics and export performance at city level. The relationship between determinants and export value will be further moderated by two factors, including environment protection investment and infrastructure investment.
data of those cities. The reason for choosing Chinese cities as research subjects can be explained from North’s (1990) institutional perspective. Both informal and formal institutional differences appear obviously among Chinese cites, which makes it possible to study differences in export activities at city level. From the perspective of informal institutional-based view, China is a big country with respect to its huge geographic area and various regional cultures. On the other hand, due to the introduction of the “Open Door” reform policy issued by Chinese government in 1978 (Yi, 2014) and other institutional policies, the development gap among cities locating in inland regions and coastal regions has been enlarged in terms of both macro and micro economic performance.
To make it specific, there are four main city categories according to their sizes and economic performance, namely municipality, prefecture-level city, provincial-controlled division and sub-prefecture-level city. When deciding the sample of Chinese cities for this thesis, 4 municipalities and 274 prefecture-level cities will be chosen and studied. There are several reasons for selecting these two categories. Technically, the data of those cities is available on their official statistics websites and it is open and free to the public. Second, according to Gao et al. (2010), free market mechanism and industry export orientation can promote export propensity and intensity at firm level. Accordingly, we can assume that municipality and prefecture-level city are the most potentially relevant cities for studying their export behavior. That is because both of them enjoy superior geographical location and large city scale as well as receive more support from government for city construction and
infrastructure.
The data collection process will be described as follows. Since we will use panel method to conduct the research, we will collect the data of sample city for ten years, from 2004 to 2013. In order to capture city characteristics comprehensively, we collect the most relevant and representative variables from different aspects, including population size, macro and micro economic indicators, investment in urban construction, industries indicators, and environment indicators etc. The main data source is city statistical yearbook or city statistical bulletin on the statistical website of each city. Those data is conducted by the local bureau of statistic under the supervision of National Bureau of Statistics of China, which ensures the consistency and reliability of data for empirical study. Gao et al. (2010) comment that the database of China statistical yearbook covers nearly 70% of China’s total export level for a given time period. They also believe that multi-year characteristics of census data makes it possible to conduct panel research. Those reasons reinforce the notion that the source used to gather the data is authentic and creditable for the study.
4.2 Variables
4.2.1 Dependent variables
The dependent variable of the thesis is the export performance at the city level. To measure the export performance of each city, we will use the value of foreign trade export of each city to operationalize it. In order to make the study more reliable, we will collect the data for ten years. Reviewing the export literature at both country level and firm level, there are different measures to estimate export performance. For
example, Li et al. (2013) use several measures to reflect export performance at firm level, including export revenue, export revenue growth and export profits. Besides, Gao et al. (2010) capture export performance from two aspects. They use export proportion of its output to measure export propensity and use export sales proportion of total sale to measure export intensity. Based on their study, we will use the value of foreign trade export of each city to measure its export performance.
4.2.2 Independent variables
The independent variables of this thesis are the indicators of city characteristics. There are three independent variables in total to reflect the city characteristics.
The first independent variable is the value of GDP, which is used to evaluate economic development and productivity of each city. GDP is a widely used variable to study the correlation between itself and export performance at country level (Henriques & Sadorsky, 1996; Xu, 1996). Following the same reasoning, GDP is defined as one of the determinants to promote export performance in Chinese cities.
The second independent is trade openness. Reviewing the international trade literature, there are several measures to evaluate the level of trade openness. To be more specific, the value of export, the value of import and the value of inward and outward FDI are the most frequently used measures to reflect the level of trade openness. For example, Yang and Mallick (2014) use the total volume of export and import divided by GDP to measure trade openness. In addition, Gao et al. (2010) apply institutional indices of free market mechanism, published by National Economic Research Institute, to estimate trade openness. According to Liargouas and
Skandalis’ (2012) empirical study, trade openness plays a positive role in driving inward FDI. Thus, we will use the value of inward FDI to measure trade openness.
The third independent variable is R&D intensity. R&D expenditure is used to operationalize R&D intensity in the firm level study (Esteve-Pérez &Rodríguez, 2013; Ito &Pucik, 1999). Similarly, this thesis will use R&D expenditure of each city to measure R&D intensity.
4.2.3 Moderating variables
In order to study the relationship between city characteristics and export level more deeply, we use investment in environment protection and infrastructure as two moderators of this thesis. These two moderators are measured by the value of investment in environment protection and infrastructure respectively.
4.2.4 Control Variables
Following the previous export literature, there are three control variables of this thesis, including city size, male proportion and industry factors. The first control variable is city size, which is measured by the number of population of each city. It is a common control variable, which is controlled in the firm level study (Gao et al., 2010). The second control variable is male proportion, which is measured by the number of male divided by the whole population. The last control variable is industry factors, which is measured by the output of primary and secondary industry divided by total GDP of the city.
4.3 Statistical analysis and results
variables, moderating variables and control variables are presented in Table 3. Besides, the table also shows the correlation between each variable, which can reflect their multicollinearity. It is the first reliable step to run descriptive statistics analysis and check the multicollinearity before analyzing the data.
According to Table 3, the values of standard deviation of both independent and dependent variables are very big. That means there is large deviation among those variables, which is in line with our assumption and fit for the following regression analysis. For example, the standard deviation of total value of foreign trade export is 205.74, which confirms that substantial differences in export performance do exist among Chinese cities. It makes it meaningful to find out the determinants of export performance in those cities.
Correlation is a useful measure to tell how well sets of data are related although it cannot explain the causality between two variables (Field, 2013). He explains that when the value of correlation is above 0.5 or below -0.5, it implies the correlation between variables is quite strong. Based on the result of Table 3, the variables are not highly correlated between each other.
In order to test the hypotheses, we use the hierarchical multiple regression model to explore the relationship between designed variables. For investigating the hypotheses with dependent variables and independent variables only, the multiple linear regression model is used based on the continuity of values and our linear assumption between those variables. When considering the impact and explanatory power of moderators, the process regression model will be used to develop interaction
terms. The moderating variables and independent variables will be standardized to ease potential multicollinearity problems before developing interaction terms. Overall, there are two steps to run the regression analysis. First, we will test the model with control variables to detect their explanatory power on dependent variables. Subsequently, the models with independent and moderating variables will be tested to evaluate their additional effects on dependent variables.
There are nine models in total in this thesis. The first model is going to test the relationship between control variables and the dependent variable. Model 2, 3 and 4 will test hypothesis 2, 3 and 4 accordingly. In addition, Model 5 is going to integrate all the independent variables and dependent variable to observe which independent variable will influence the dependent variable the most. Furthermore, Model 6-9 will assess the hypotheses with moderators respectively. Table 4 and 5 will present the analysis results of the hierarchical multiple regression models. When interpreting the statistics, we will mainly discuss and consider three indicators including the value of R2, coefficient and Beta. First, the value of R2 indicates the fitness and predictability of the model. It can tell us how much the variance in dependent variable can be explained by independent variables or other variables in a model. If the value R2 is closer to 1, it implies the model can explain variance to a higher degree. Coefficient is one of the most important terms to determine whether the observed model is supported and whether the relationship between variables is significant. Only when coefficient is below .05, the result is reliable to support the hypothesis. In addition, the Beta standardized coefficient is responsible for explaining changes induced by
explanatory variables on dependent variables.
As stated above, Model 1 is going to test the explanatory power of control variables on dependent variables. The value of R2 is 0.128, which means the variance in dependent variable cannot be explained by control variables at all. However, we can recognize strong support for demographic indicators including population size (b=.203, p=.000) and male proportion (b= -.146, p=.000). Besides, significance also appears in the industry indicators. The GDP of primary industry (b= -.239, p=.000) is negatively related to export level. However, the impact of GDP of secondary industry is not significant.
Hypothesis 1 argues GDP is positively related to the value of foreign trade export for the sample cities, which is strongly supported by the analysis result. The coefficient (b=.882, p=.000) is significant for hypothesis 1. Thus, GDP is proved to be positively related to export level. Since R2 turns from .128 in control model to .601, GDP is more reliable to explain variance in export performance. However, R2 is back to .130 in Model 3 for hypothesis 2. It is also not supported with insignificant result (b=.060, p=.148), which means inward FDI is not related to export level. Model 4 for hypothesis 3 presents significant result (b=.321, p=.001) as well. Thus, hypothesis 3 is supported, which confirms R&D intensity is positively related to export level. In addition, Model 5 tests the explanatory power of all independent variable on depend variable at the same time. The results show that GDP has the most power over inward FDI and R&D intensity to explain the variance in the export level. However, what is worth mentioning is that the explanatory power of R&D (b=-.182, p=.061) on the
variance in export performance vanishes when combining GDP in the same model. When it comes to moderating variables, both two moderators exert significant impact on the relationship between GDP and export performance (hypothesis 1). However, the moderator effect on the relationship between inward FDI and export level (hypothesis 2) is subtle. To be more specific, Figure 2 shows the moderator effect of investment in environment protection on the relationship between GDP and export level. From the analysis results we are informed that the effect of investment in environment protection itself is highly significant on export level with the coefficient of .000. Besides, the interaction terms between GDP and investment in environment protection (b=.300, p=.000) reveals highly significance as well. Therefore, we can tell that investment in environment protection will positively moderate the relationship between GDP and export performance. As showed in Figure 2, when there is high investment in environment protection, the relationship between GDP and export performance is more significant. However, although the interaction terms between investment in environment protection and inward FDI is significant (b=15.20, p=.000), the moderator itself cannot explain the variance in the dependent variable. Thus, there is no convincing evidence to support hypothesis 4b.
Figure 2. Moderating effect of investment on environment protect on total value of foreign trade exports
As for hypothesis 5a, strong support is found for interaction term between GDP and investment in infrastructure (b=.700, p=.000). Together with significant impact of investment in infrastructure on export performance (b=.600, p=.000), the relationship between GDP and export performance is positively moderated by investment in infrastructure. The last model is rejected by the insignificant interaction term (b=.400,
p=.300). Therefore, investment in infrastructure does not moderate the relationship
between inward FDI and export performance. 0 2 4 6 8 10 High investment in envrionment protecttion Low investment in envrionment protecttion GDP E x p o rt s
5. DISCUSSION
The theoretical frameworks and empirical analysis combining to existing theories have expanded the understanding of export behavior at city level and brought profound implications to both policy-makers and firm strategists. While most of the previous export literature concentrates their studies either on country level or firm level, this thesis shifts the traditional units of analysis to city level and explores the determinants of export performance at city level. The results of empirical analysis are valuable in terms of academic relevance, policy implications and managerial implications. Overall speaking, the explanatory power of GDP is strong on the variance of total value of foreign trade exports at city level. R&D intensity is also positively related to the total value of foreign trade exports. It should be noticed that the significance of R&D intensity vanishes when integrating GDP into the same model. However, no significant evidence is found on the relationship between trade openness and export level.
In addition, the moderating effects investigated here are also worth mentioning as they have not been fully understood in the previous literature. The moderating effects of investment in environment protection and infrastructure on the relationship between GDP and export level are significant. Our results show that these two moderators positively strengthen the relationship between GDP and export performance. However, there is no evidence supporting the moderating effects on the relationship between inward FDI and export performance.
The academic relevance, policy and managerial implications, limitations and suggestions for future study will be discussed and explained in more details in the following sub-sections.
5.1 Academic relevance
This thesis contributes to existing export literature in several aspects. First of all, it fills up the research gap by shifting the unit of analysis from country level and firm level to city level. Second, it explores whether characteristics of cities can become the determinants and drivers of export performance at city level. The empirical analysis indicates significant factors that determine export behavior. Furthermore, it introduces the moderation of investment in environment protect and infrastructure into theoretical frameworks and examines their effects on export performance. While the independent variables are mainly derived from extant export literature at either country or firm level, the analysis of moderating effects of investment in environment protection and infrastructure provides fresh understanding on export literature. Last but not least, the empirical study collects the data of 278 Chinese cities to run the regression analysis, which enables the city level study in an emerging economy.
The impact of productivity on export performance has always been a central focus among existing country level and firm level export literature. According to the traditional trade theory, comparative advantages of one country, which can be referred to productivity, is the premise for the occurrence of exporting (Ricardo, 1817). When it comes to firm level study, Melitz (2003) develops a dynamic intra-industry model to illustrate how firm productivity will select certain firms to involve in exporting.
Besides, Xu’s study (1996) confirms the relationship between GDP and export performance at country level. While interpreting the analysis results of this thesis, we can find supportive evidence to confirm the relationship between GDP and export performance. Furthermore, GDP is the most influential factors to determine export performance compared to inward FDI and R&D intensity. This finding is line with Melitz’s (2003) argumentation that productivity selects exporting firm and affect their export performance.
Yang and Malick (2014) have questioned on the mixed results in the existing export literature and found that macro factors such as market mechanism and financial crisis are more reliable to explain differences in export performance across countries. Following the same reasoning, this thesis explores whether macro factors also exert the same impact on export performance at city level. The results partially support the above argument. It is found that GDP is positively related to export performance at city level. The relationship between R&D intensity and export performance is significant when excluding the effect of GDP. However, there is no significant evidence to validate the effect of trade openness on export performance. One possible reason for is that trade openness does not affect export performance directly but affecting it by influencing other factors.
When reviewing the export literature, it is easy to classify them into two categories, namely country level study and firm level study. It is also easy to identify that city level study is missing. However, city level study is also critical for fully understanding trade theory provided the fact that some cities outperform others in
terms of export performance to a great extent. The only one available city level study conducted by Li et al. (2013) finds that the quality of legal system in different cities will influence their export performance. They collect the data of Chinese cities to run their analysis. Their empirical study shows that institutional factors can result in the differences in the quality of legal system at city level, which will further influence export performance. This thesis follows their thinking and sheds more light on studying export behavior at city level.
Last, this thesis introduces the moderation of investment in environment protection and infrastructure into the theoretical frameworks. It not only extends the existing understanding of these two factors, but also links to the real situation of the data sample. There are great differences among Chinese cities in terms of their investment and received subsidy on environment protection and infrastructure. Previous literatures have identified the moderating effect on export at country level (Looney & Frederiksen, 1981; Ye & Yu, 2007). However, the investigation at city level is still missing. The analysis results of this thesis show that both two moderators affect the relationship between GDP and export level while there is no moderating effect on the relationship between trade openness and export level.
5.2 Policy implications
We believe the findings of this thesis have important implications for policy-makers in the following two aspects. First, policy-makers should be aware that GDP is a positive factor to increase export performance. The results show that higher level of GDP can predict corresponding higher level of total value of exports.
Therefore, if they want to boost export performance at specific cities, they need to promote the GDP at those cities. If policy-makers want to improve export performance by means of increasing R&D intensity, it should be noticed that the effect of R&D intensity on export performance vanishes when including the effect of GDP. Although trade openness does not influence export level significantly according to analysis results, it is still critical to encourage inward FDI. That is because trade openness may affect export level indirectly.
Second, the findings also confirm the moderating effects of investment in environment protection and infrastructure on the relationship between GDP and export performance. Recently, there are more negative reports about the environmental problems in Chinese cities especially those located in northern part of China. For example, the problem of sand storm is severe in Beijing, which has made some companies move out of city. Thus, it is vital for policy-makers to realize the importance of those moderators and make relevant policies to improve current situation.
5.3 Managerial implications
The findings also bring important implications for firm strategists. Where to locate has always been a critical consideration for the development and expansion of one company. The location strategy should be in line with firm strategy. According to the analysis results, the exporting performance can reflect the level of a city’s GDP and R&D intensity. On one hand, the exporting firm will benefit more if it locates in a city with better export performance. On the other hand, since the findings indicate a
positive relationship between R&D intensity and export performance at city level, a firm who wants to take advantage of knowledge spillover can choose the city with higher export level. Similarly, if environmental situation is an important criteria for firm’s operation, it can also select the city with higher export performance.
5.4 Limitation and suggestions for future research
There are also some limitations of this thesis that should be noticed. First, we use the construct inward FDI to measure trade openness. However, the analysis results show that the relationship between trade openness and export is not significant, which violates previous country level findings. We do not think it is reliable to deny the relationship at city level according to results of this thesis. Therefore, future research can explore whether trade openness will influence export performance indirectly at city level. Alternatively, future research can also use other proxies to measure trade openness such as trade barriers or taxation.
Besides, this thesis uses the investment in infrastructure to evaluate the quality of infrastructure and tests its moderating effect on export performance. We believe future research can specify the infrastructure to see whether investment in a specific infrastructure will improve export performance to a greater degree. For example, future research can focus on transportation infrastructure or telecommunication development to investigate their impact on export performance at city level.
Second, it is interesting to observe that the significant relationship between R&D intensity and export performance will vanish if the effect of GDP is integrated. Future
research can make more effort to investigate the relationship among GDP, R&D intensity and export performance.
Third, due to time and resource constraints, this thesis only selects several factors to explore their explanatory power on export performance at city level. However, we believe there are still other factors that will affect and make a difference on export behaviors. Future study can focus on institutional factors, including both informal and formal institutional factors. For instance, it is worthwhile to explore whether the cultures in different regions of one country play a role to promote export. It is also interesting to investigate whether the efficiency of bureaucracy in different cities will influence their export performance.
The forth limitation is due to the nature of secondary data. Although the thesis uses panel research design to run the analysis, there are considerable missing data for some variables in some years and some cities. Therefore, future study on Chinese cities can alter to other database in order to access to more data.
Last, the findings in this thesis provide a fresh perspective to understand export at city level. However, whether the results are universal to other cities is still in question since the sample of this thesis is limited to Chinese cities. We suggest that future research can conduct the study in other cities or emerged economies. In addition, future research can also compare export performance of cities from different developed countries or developing countries.
The purpose of this thesis is to find out the determinants of export performance in Chinese cities. Most of the previous literature focuses their units of analysis on country level and firm level, this paper fills up the research gap by shifting the unit of analysis and conducting export study at city level. Li et al. (2013) is one of the few who run the export study at city level. They try to investigate the relationship between quality of legal system and export performance in different Chinese cities. Following their reasoning, this thesis aims at figuring out significant determinants that will promote export performance at city level.
According to previous country level and firm level literature, some factors have been verified to be positively related to export behavior and performance. Countries with better GDP performance will involve in international trade (Henriques & Sadorsky, 1996; Xu, 1996) while the intra-industry model developed by Melitz (2003) argues that firms with high productivity will self-select themselves to export activities. Based on existing literature, this thesis has explored the effects of GDP, trade openness and R&D intensity on export performance at city level. Besides, we also integrate the moderation of investment in environment protection and infrastructure into the conceptual frameworks.
Since the purpose of this paper is to find out determinants of export performance at city level, it is important to choose suitable sample to conduct the analysis. Recognizing there are great differences in Chinese cities in terms of their geographic location and economic development, we decide to select 278 Chinese cities as our main data sample. Subsequently, we collect relevant city data from city statistical
yearbook or city statistical bulletin published on the statistical website of each city. In order to increase reliability and credibility, we use the panel research design and gather the data for ten year, from 2003 to 2014.
The thesis uses the multiple hierarchical regression analysis and develops 9 models to test the hypotheses. The results show that GDP is positively related to export performance at city level. The relationship between R&D intensity and export performance is significant on the condition that the effect of GDP is excluded. However, no supportive evidence is found to validate the positive relationship between trade openness and export performance. One possible explanation is that trade openness does not influence export performance directly. There may be other conditional factors that will affect the relationship. As for the moderators, we find that both investment in environment protection and infrastructure positively moderate the relationship between GDP and export level. However, the moderating effects are not significant on the relationship between trade openness and export performance.
These findings have important implications for both academic field and international business management. First, it extends our understanding on export theory by focusing the unit of analysis on city level. Second, policy-makers know how to boost export performance at city level while firms can choose ideal location based on their strategies.
