Political Connection, Local Protection and Domestic Market. Entry Barriers in China

Transcription

1 Political Connection, Local Protection and Domestic Market Entry Barriers in China Qun Bao, Ninghua Ye, Ligang Song 1. Introduction Trade theories assume that large entry costs exist for exporters and sales within one country s domestic market involve no or much less entry costs. As a result, exporters need to show higher productivity advantage to overcome the export costs across national borders, which is called the self-selection effect in trade literatures (Melitz, 2003), and many empirical studies supports the self-selection of exporters (see the surveys by Wagner, 2005; Greenaway and Kneller, 2007; Bernard et al., 2012). Such an assumption is quite plausible, since it s usually much harder to conduct international business than doing business in the home market. However, if a country s domestic market is not closely integrated, like the current Chinese market that is segmented by its provincial borders due to many institutional factors and local protectionist policies, we would expect that sales within the domestic market can also involve substantial market entry costs. Why Chinese domestic market is still highly segmented after decades-long market reform? What are the main sources of domestic market entry barriers in China? Unlike international trade barriers such as transportation costs and tariffs, barriers to domestic transaction are, to a large degree, created by intra-national protectionism and 1

2 the related tedious administrative procedures 1. China s case is more serious because of its unique institutional setting. China s fiscal decentralization has been a fundamental aspect of its reform programs towards a market economy, which aims at getting various levels of governments to be fiscally responsible with various forms of fiscal contracting systems and tax sharing system (Shen et al. 2012). Such reform has substantially prompted China s rapid regional economic growth (Zhang and Zou, 1998; Jing and Zou, 2005) through much improved incentives for China s local governments to implement various administrative intervention measures including protection of the local business in order to maximize its local fiscal revenues and employment. However, these protective measures have inevitably raised interregional trade barriers and are the fundamental causes for current market segmentation in China and for firms incentives to forge closer relationship with local governments to overcome those barriers for increasing cross-regional trade. It s Young (2000) who first raised the issue of domestic market segmentation. The author emphasizes that in a partially reformed economy like China, growing interregional competition leads local governments to impose a variety of interregional barriers to trade. Following the influential findings of Young (2000), many more studies have confirmed the existence of domestic market segmentation and trade barriers between regions (Naughton, 2003; Bai et al, 2004; Poncet, 2005; Fan and Wei, 2006; Holz, 2009; Xu and Fan, 2012). Herrmann-Pillath et al (2014) compared the effects of political factors and cultural differences on market segmentation and 1 For instance, Djankov et al. (2002) look at the procedures for registering a new firm in 85 countries, and find that government productivity has a significant impact on registering new firms. 2

3 concluded that trade barriers are indeed instigated by local protection, rather than caused by cultural and dialect differences. In other words, the cross-regional trade is likely to be hindered more by local administrative intervention than language and cultural barriers commonly seen in the international trade. 2 It is therefore our interest to ask how local protection affects firms domestic entry behavior in China. In a normal and competitive business environment, the sales activity and destination choices of firms will depend entirely on their own competitiveness. However, in the presence of trade barriers due to local governments administrative intervention, it can be expected that firms will resort to political lobbying, in an effort to eliminate the negative impact of local protection on their business by building a closer relationship with the administrative departments. However, to investigate how political connection affects firms domestic sales, we need detailed information about firms sale among different domestic destinations. For example, we need to know whether firms ever sell their goods to other provincial markets or prefer just to stay within their own provincial market. Compared with the detailed transaction data in export market collected by the Chinese Customs, it is very difficult to collect the statistics about firms domestic sale activities within China. To fill this gap, the World Bank provided some detailed survey data on firms sales destinations for 12,000 firms operating in China in This survey data 2 Local protection is definitely not unique to the Chinese economic development. In fact, either in developed or developing countries, local governments may intervene in the business environment for different reasons. This phenomenon has aroused widespread concern, such as the research of Sonin (2010) on Russia, Luong (2004), Kazakhstan, Das-Gupta (2006) for India, and Berdahl (2013) for Canada. The Chinese experience offers ample opportunities for studying this issue because of the unique feature of China's political system, including fiscal decentralization, economic tournament among regional economies, and the ways for official assessment on performance which all exacerbate the prevalence of local protection. 3

4 encompasses detailed market information including firms locations and destination for their products including sales to the local markets, other cities within provinces, other provinces as well as export markets, and importantly some proxies for capturing the political connection by firms. Using firms who only sell in the local market as the benchmark, we study those firms who enter the other cities or provincial markets by establishing closer political connection to facilitate their sales. In other words, it can be hypothesized that political connection helps firms to successfully enter domestic markets by lessening administrative intervention from local authorities in the targeted market. We also use the behaviors of exporters as a robustness check. As local protection doesn t matter to exporters, we would not expect that there is also a strong correlation between political connection and export activities among exporting firms. We also show the extent to which local protection is implemented in affecting firms sales in other markets. In the case of minor local protection, political connection mitigates the effect of local protectionism and helps firms enter other provincial destinations. However, once local protection becomes too strong, political connection fails to help businesses to break down the barriers across provinces. In other words, the effect of political connection is quite limited when the local protection becomes more serious. It implies that we can t exaggerate the role of political connection in facilitating domestic market entry, and reminds us of the fact that political connection can t be considered a fundamental solution to keeping off local protection and administrative intervention. 4

5 Our study helps to understand the motivation of domestic sale activities in transitional economies like China. Following Bernard and Jensen (1995), extensive literatures have examined the determinants of the international market entry among exporters (Wagner, 2007; Greenaway and Kneller, 2007; Bernard et al., 2012) 3. However, much less has been known for firms domestic sale behavior, especially for countries like China whose domestic market is so large and has a profound impact on firms business operation. This study thus provides useful evidence for understanding the entry behavior of firms in domestic markets. Our study contributes to the literatures on political connection. A large number of empirical studies prove that companies can indeed benefit from political ties, including more convenient access to bank credit (Joh and Chiu, 2004; Cull and Xu, 2005), seeking government-regulated preferential treatment (Kroszner and Stratmann, 1998), and seeking government's crisis rescue packages and policy asylum measures (Faccio et al, 2006). However, few studies have examined the effect of political connection on firms sale activities, especially for a geographically large transitional economy such as China. This study confirms that in an economy associated with market segmentation and local protection, political connection may help smooth the way for businesses entering into the remote markets. It echoes with previous studies that the emergence and prevalence of political connection is a reasonable and strategic reaction of firm especially private businesses in the presence of widespread administrative intervention. 3 Many studies also explored the reasons behind the rapid expansion of exports based on Chinese datasets (Lu et al, 2010; Manova and Zhang, 2012; Koopman et al, 2012; Ma et al, 2014.). 5

6 The arrangement of this paper is as follows. We describe the data source and estimation model in section 2; section 3 provides the estimation results on how political connection affects firms sale destinations; section 4 conducts the robustness tests; and the final section summarizes and concludes. 2. Data Source and Estimation Strategy Data used in this paper are from China Investment Climate Survey (2005) of the World Bank. The World Bank survey randomly chooses 12,400 firms located in 120 cities: Each sample city has 100 firms and the four municipalities (Beijing, Tianjin, Shanghai, and Chongqing) have 200 firms each. The unique advantage of this survey is that it collects detailed sales information in four types of targeted destinations, including locating cities, other cities in the same province, other provinces, as well as export markets. Thus it enables us to examine the determinants of firms entry decision in any targeted market. We first group the sample firms (exclusive between types) according to their sales destinations as the following four types: type 1 indicates firms with sales only in the city where they locate, type 2 firms with sales in other cities of the same province, type 3 firms with sales going to other provinces, and type 4 firms with exports. The first three categories are all domestic-oriented firms with type 2 and type 3 having sales across city and provincial administrative boundaries. Political connection is the core variable which we concern. In some previous studies of China s case, political connection is typically measured by whether the corporate management or board members have the party membership (Li et al, 2008), 6

7 or whether executives are the National People s Congress (NPC) or China People s Political Consultative Congress (CPPCC) members (Tian and Zhang, 2013), or whether managers and other senior executives are appointed directly by the authorities (Wang and Sheng, 2012). These proxies reflect firms affiliation with the governments to some extent. However, they do not speak for the efforts that firms make to enter other cities or provincial destinations. For example, if managers are directly appointed by the local authorities, it is more reflective of the political relations of firms at the localities, and firms are therefore more likely to adopt a localized business strategy rather than external expansion beyond their jurisdictions. In light of the potential drawbacks of these proxies, this paper uses "designated personnel for dealing with administrative businesses" in China's Investment Climate Survey (2005) to characterize the political connection. By designating personnel to deal with local governments which have been granted economic autonomy after the decentralization reform program was implemented in the 1980s and 1990s, it directly reflects firms needs and efforts of making the political connection to pursue their business interests. We would expect that to whom firms designate as their personnel to deal with administrative business are more likely to establish close ties with the authorities, and thus to overcome the entry barriers erected by the authorities more easily. A binary dummy variable gov is created, where the value 0 means absence of such personnel and 1 means presence of such personnel, indicating close political connection. We also use other measures of political connection as the robustness checks. 7

8 Table 1 provides the statistics about various types of sale behaviors of the sample firms. First, among the 3,727 sample firms, only 750 firms or 20.12% focus on sales within the home cities. It means that the vast majority of firms have a tendency to extend their sales to other markets. Second, 1,068 (approximately 28.66%) firms sell products only to other cities in the same province, while 1,113 (approximately 29.86%) selling their products to other provinces. Finally, 21.36% of the samples are firms engaging in exports. Table 1 also shows the relative importance of the political connection for various sale types. Generally, there is positive relationship between the scope of firms sale and the degree of their political connection. In other words, "the farther away that firm s products are sold", the more likely firms are to establish a robust relationship with the government as more distant markets could mean higher entry barriers for the firms. Specifically, only 17.47% of local businesses designate personnel in charge of administrative issues. Nearly 21% of companies targeted at other cities in the province have political ties, and among firms with sales in other provinces, the proportion increases to nearly 25%. Among exporters, more than 30% of the firms designate personnel to deal with administrative departments. Thus, we ask the question as to whether the closeness of political connection helps firms to enter distant market destinations. Control variables that may affect firms sales decisions are also incorporated to the model in estimations as follows. (1) lnp2003 is the average output: the logarithmic number of sales revenue divided by the total number of employees in the firm; (2) 8

9 ktl2003 is the capital-labor ratio: the logarithmic number of net fixed assets divided by the total number of employees; (3) scale2003 is the scale of the firm: the logarithmic number of main business revenue; (4) age is the age of the firm: the difference between the year of establishment and the year of study (2004); (5) rdr2003 is the proportion of research and development (R&D) investment in the main business revenue. (6) firms ownership dummies: private= {0, 1} represents private firms, hmte = {0, 1} represents Hong Kong, Macao and Taiwan firms, and fie = {0, 1} represents foreign-invested firms respectively. In addition, cities dummy variables and industrial dummy variables are also introduced to control for the region-specific and the industry-specific fix effects of firms sale activities. 3. Estimation Results on Political Connection and Market Entry 3.1 Firms entry into different destinations Table 2 reports the probit model estimation results of firms entering into various markets showing how political connection affects firms entry into different market destinations. The estimation results indicate that the regression coefficient of political connection is significantly positive for out-province sellers who enter other provincial markets, but not for out-city sellers and exporters respectively. The findings may suggest that close political connection does increase the probability of entry to other provincial markets, but this effect is not seen in selling products to other cities within the same province or to export markets. The estimation results here reveal the extent to which local protection matters in helping firms marketing performance. Firms usually face less administrative 9

10 intervention within the provincial range, and hence the sales to other cities in the same province are not clearly related to the political connection. It is understandable because the entry barriers at provincial borders in China are more palpable than those existed within provinces as one would expect. However, in the effort to open up markets in other provinces, firms may be confronted with various administrative obstacles out of local protectionism in other provincial destinations. As a result, the political ties with governments will help firms to noticeably increase the probability of overcoming the regional or provincial barriers to trade. Similarly, Herrmann-Pillath et al (2013) compared the impact of inter-city and inter-provincial boundaries on the linkage of two adjacent urban economies. The results show that the former has significant positive effects while the latter has no significant correlation with economic linkage illustrating that market entry barriers do exist at China s provincial borders. The finding in this paper confirms the same point by suggesting that intervention policies for local protection may be mainly implemented by provincial-level authorities towards market entry of businesses from other-provinces. The estimation results also show that the role of political connection in helping with firms overseas sales is not obvious. It implies that the frequent dealings of export-oriented firms with official departments do not help to explain much about firms export decisions and outcomes. Political connection does not play an apparent role in helping firms with reducing the costs of international trade for export-oriented firms. This finding too is as one would expect as firms engaging in foreign trade need to overcome the international transport costs, cultural differences, language barriers, 10

11 as well as a variety of tariff or non-tariff barriers which are largely beyond the influence of local governments. The next question we ask is the extent to which gov promotes the sales of firms in other provinces. Answering this question helps us to understand the effect of local protection in China. For example, even though the estimated coefficient of political connection is positive, if the actual contribution to sales is extremely limited, it may be the case that the local protection through administrative intervention may not affect Chinese firms domestic sales as significantly as expected. In other words, given the fact that local protection is popular in China, the actual effect of it on firms domestic entry may be quite limited. So, to understand better the role of political connection in the entry to the markets of other provinces, we further predict the probability of sales to other provinces for firms with political connection (gov = 1) and without political connection (gov = 0) respectively. We find that other things being equal, political connection will increase the probability approximately by 57.5% (= ( ) / 0.471). This calculation means that if a firm attempts to establish political ties with the local authorities, it will help to increase the probability of entering into other provincial market by more than fifty percent, which is not a trivial effect. It confirms that Chinese firms in general significantly benefit in market entry to other provinces from building ties with the authorities. For the impact of other control variables, it s not surprising to see that firms with larger scale and more R&D expenditure are more likely to enter remote markets because of the competitiveness that these attributes bring to these firms. From the 11

12 ownership point of view, relative to the state-owned enterprises (SOEs), private firms are more positive in developing external markets, posing a far higher probability of entering provincial markets and international markets, while HMTE and FIE show a higher tendency to export. To compare the importance of the political connection and other variables such as R&D investment and firm size, the marginal effect of the explanatory variables is further calculated. The results show that political connection has a marginal effect of 0.046, second only to firm size (0.063), but much higher than other corporate characteristics (i.e. R&D investment with a marginal effect of 0.024). It once again proves that the political connection plays a relatively more important role in promoting firms in market entry to remote or distant destinations. 3.2 Impact of Political Connection by Entry Type Table 2 also supports the importance of political connection to the choice of firms market entry type. To do that, we use the multinomial-logit model (MNL) to estimate the effect of political connection on various entry types. Specifically, we divide all firms into the following four types: k = { 1,2,3,4 }, with k = 1for LocalFirm firms who only sell in the home city market, k = 2 for ProvFirm firms who enter other cities markets within the same province, k = 3 for Domestic firms who enter other provincial markets, and k = 4 for Exporter. With k = 1(LocalFirm) as the base category, we use the MNL to model firms choices on their sale types: P ( K k) = = 1 4 exp2 exp ( β k X ) ( β X ) + = k k The explanatory variable vector X refers to political connection and other control variables. However, it is not straight forward to interpret the coefficients in the MNL 12

13 model in a satisfactory way. To better interpret the coefficients in the MNL model, we introduce relative-risk ratios (rrr) in our estimation model. For firm type i, the rrr for vector X, compared with the base category firm type 1 (ProvFirm), is derived as: P( K = i X + 1) / P( K = 1 X + 1) rrr i /1 = P( K = i X ) / P( K = 1 X ) The rrr i/1 for political connection shows that one unit increase in political connection will lead to relative odd of choosing to become firm type i is rrr i/1 times than what it was before. Thus, the value of rrr i/1 greater than unity indicates that an increase in certain explanatory variables leads to higher probability, and vice versa. As noted in the literature, the empirical application of the MNL model also needs strong assumptions, particularly the Independence of Irrelevant Alternatives (IIA) assumption, which requires that the choice of any option does not affect the relative probability of other options. If the assumption does not hold, one needs to use other estimation model specifications such as multinomial probit model. In the empirical analysis, we use the Hausman test in Hausman and McFadden (1984) to test the independence of irrelevance alternatives to show the validity of the assumption. The MNL estimation results in Table 2 show that for Domestic firms, the relative risk ratio of political connection is significantly greater than one, which means that political connection is indeed beneficial to the entry to other markets outside the province. Specifically, if designated personnel are in place for dealings with the authorities, then the relative probability of inter-provincial sales will be increased by time from the initial level. However, rrr is not significant in the statistical sense for ProvFirm firms and exporters, though the value is still greater than 1. For firms 13

14 targeted other cities in the province, the role of firm size and R&D investment are the most significant factors. For exporters, the most important determinant is firm ownership. For example, FIE s export probability is more than 37 times higher than SOEs. 3.3 Endogeneity of Political Connection and IV Estimation We are concerned about the endogeneity problem in capturing the role of political connection, which may lead to biased estimation results. There must be reasons why some firms choose to build close ties with the administrative authorities, and some don t. Hence, to deal with the potential endogeneity problem, we use the instrumental variables method by choosing proper instrumental variables for political connection. In this study, the right instrumental variable can be those firms who are committed to the establishment of good relations with the authorities but those good relations won t directly affect firms entry decisions (exclusion restriction). Specifically, we choose an instrumental variable (disp): whether the firm has disputes or fails to reach agreements with the government in the last three years. Obviously for firms ever had disputes with the authorities, it shows that they have weaker ties with the administrative departments. Things got even worse for those who couldn t reach an agreement with the government once they had dispute with the latter. Thus they should have stronger motivation to build close relationship with the government by designating personnel to deal with the government departments. On the other hand, it is difficult to say that there is any obvious correlation between disp and firms business entry behavior. 14

15 We adopt the usual two-step IV procedure to re-estimate the model, and the IV estimation results are shown in Table 3. Our main finding still holds true even if considering the endogeneity of political connection. For those firms who are subject to intervention measures for local protection, political connection indeed increases the probability of their access to other provinces markets, while this effect is not seen in their business in other cities in the province and export markets. 3.4 The Existence of Local Protection The above analysis confirms the effect of political connection on the domestic market entry of firms. However, it does not consider the heterogeneity of local protection. When examining the relationship between political connection and sales behavior, it is necessary to consider not only the presence of political connection, but also the constraints arising from administrative intervention that firms are subject to. For example, it s reasonable to expect that firms who face serious local protectionism would have a stronger demand for political connection than those without. So political connection should play a bigger role in overcoming regional barriers for those firms who are experiencing severe protectionism. The survey also asks "whether local protectionism affects the business operation and development of the firms". Such survey results enable us to deal with this heterogeneity issue. Based on firms answers to this question, indicators reflecting the impact of local protectionism borne by businesses are constructed. 498 or 56.5% of 881 firms with political connections report that local protection does not affect their business operation. We define govlpno=1 for those firms with political ties and free 15

16 from local protectionism. A total of 383 firms or approximately 43.5% believe that their business operation and development is really hindered by local protection, and we subsequently define govlpyes=1 for these firms. Table 4 presents the test results about the combined effects of political connection and local protection on firms offsite sale behavior. The estimation result supports the importance of local protection. For ProvFirm firms, the role of govlpno is negative, but not statistically significant, while govlpyes is significantly positive. It means that if firms are subject to local protectionism, political connection increases firms entry probability to other cities in the same province, and it s not true for those who aren t. In comparison with the results in Table 2, we find that for firms who targeted other cities in the province for sales, the role of political connection is mainly for keeping its businesses from the impact of local protectionism. Specifically, for firms who are subject to local protectionism, the relative probability of out-city sales is time higher thanks to political connection. The same result holds true for cross-province sales. Both in the entry probability model and in the MNL model, govlpno and govlpyes are both positive, but the former is not statistically significant, which means that political connection is an effective channel to increase the entry likelihood of other provinces only for firms who are indeed affected by local protectionism. For firms who are subject to the intervention of local protectionism, the relative probability of inter-provincial sales would increase by times. By comparison, both govlpno and govlpyes are not significant for 16

17 export. In other words, whether or not firms are subject to local protectionism, political connection does not have a significant effect on firms export decisions. It further shows that political connection plays a role in overcoming trade barriers between regions within the country, rather than in dealing with international trade barriers faced by exporters. 3.5 Does Degree of Local Protection Matter? We have shown the importance of political protection in helping firms to overcome the difficulties associated with local protectionism. However, the results only distinguish between firms who are subject to local protection and those who are not. It s our interest to ask further question as to whether the degree of local protection matters. By the degrees of local protection, the World Bank survey dataset asks firms to report how seriously they consider local protectionism affects their business operation and development among those 383 firms who are subject to local protection. Their answers are four types in ascending order (1, 2, 3, and 4). To test the impact of local protection by degrees, two new political connection variables are created. A total of 235 of the 383 firms have political connections and think a low degree of local protection (=1) of the firm, and they are referred as govlplow, while the rest 148 firms (degree = 2, 3, 4) as govlphigh referring to a high degree of local protection. It is our concern that whether political connection exerts a more serious impact on sales of these firms who experience severe local protection. As local protection primarily affects the sales behavior of firms in the domestic market, we focus on the role of local protection in affecting firms domestic sales. 17

18 Table 5 shows the new estimate results. Inconsistent with our expectations, for sales in other cities of the same province or sales outside the province, govlplow is significantly positive, while govlphigh is not significant. For firms who are severely hindered by local protection, the political connection no longer promotes their business development in other domestic markets as we expected. This result reminds us that the role of political connection is by no means omnicompetent for all firms. As a matter of fact, the role of political connection in market expansion is only limited to businesses moderately affected by local protectionism. Once firms business operation is intervened by high degree of administrative protection, it doesn t matter whether firms have attempted to establish close ties with the governments or not. In other words, it s difficult for firms to use political connection as an effective tool to overcome the regional trade barriers once these barriers become so high due to the severe local protectionism. We also introduce the interaction term gov * localpro (localpro=1, 2, 3, 4) to test whether the role of political connection is positively correlated with the degree of local protection. The results show that the estimated coefficient of gov * localpro fails to show statistical significance, which means that there is no simple linear relationship between the degree of local protection and domestic market entry. It indicates that political connection is not the panacea in helping firms to overcome market entry barriers, though conducive to easing the pressure from local protection to a certain extent. In the presence of too severe administrative intervention, political connection does not alleviate entry barriers brought by local protection. 18

19 4. Political Connection and Market Entry: Robustness Tests 4.1 Firms Ownership Heterogeneity As an economy in transition, the Chinese economy is clearly characterized by business operation heterogeneity between firms of different ownerships. SOEs enjoy more inclined support and preferential policies rendered by governments, and encounter fewer entry obstacles brought by local administrative protection. Private firms, however, face obvious discriminatory policies and grim administrative intervention, though the overwhelming evidence demonstrates their important role in macroeconomic growth and employment (Chen et al, 2005; Li et al, 2008). Apparently, due to the lack of policy support, private firms can only resort to political ties to resolve trade barriers from local administration. For example, Li et al. (2008) found that the party membership of private entrepreneurs can remarkably improve business performance, facilitate access to bank credit, and boost the confidence in the judicial system. It is therefore necessary to further distinguish the impact of political connection on the sales behavior of firms of different ownerships. Following the usual practice, firms are divided into SOEs, private firms and foreign-invested enterprises (FIEs). Table 6 shows that the role of political connection does vary with firms ownership types. For private firms, govlplow not only significantly increases the probability of entering the market in other provinces but also different cities of the same province. Similar to Li et al (2008), this study confirms that in the presence of local protection, private firms tend to resort to political connection to eliminate 19

20 market entry barriers, and the role of political connection remains significant even in the market within the same province. In contrast, this role does not exist for SOEs. Regardless of other cities or provinces, political connection does not influence the probability of their market entry, and this result does not relate to perceived degree of local protection. Finally, political connection also helps FIEs to significantly increase the probability of entering other provincial markets, but not, of entering other cities in the same province. In this sense, this study does demonstrate some significant differences in local protection effects on firms of different types. 4.2 Controlling for Travel Expenses One of the main arguments in the above analysis is that domestic entry barrier exists mainly due to local protection. However, except for administrative barriers under local protection, firms still face other normal transaction costs for market entry, such as for advertisement in remote markets and commercial communications with sales agents. Obviously, if ignoring these normal costs, the role of political connection in remote sales will be overestimated, since normal business costs may be attributed to the result of political connection. In order to control for the normal commercial costs of firms in market expansion, we introduce the indicator of travel expenses (traratio) in the estimation model characterized by the percentage of travel expenses to sales revenue from main business. The new estimation results reported in Table 7 show that the estimated coefficients and the statistical significance of govlpno, govlplow and govlphigh remain the same, indicating that the earlier conclusions reached are robust. 20

21 Meanwhile, in line with the expectations, traratio is significantly positive in the equations for markets in other provinces and in other cities of the province. The results indicate that with the normal commercial costs for sales being controlled for, political connection still has significant positive impact on offsite sales for those firms who are suffering moderate intervention from local protectionism. It means that regarding domestic market entry, firms need to bear the normal costs of sales across the regions, but also seek political ties to overcome administrative intervention arising from local protectionism. 4.3 Sale Probability vs. Sale Volume While we have shown the important role of political connection in increasing the entry probability to markets in other provinces, we wonder whether political connection also affects firms sales volume. Regional trade barriers caused by local protection are in general considered similar to the initial fixed costs of entry in the international trade, which raise the threshold for market entry. However, once firms successfully overcome these barriers, their sales in the destination market may essentially depend more on other things such as market size. In other words, given the market entry threshold under local protectionism, political connection is expected to influence more the entry probability, rather than firms sales volume. In this respect, by distinguishing the impact on entry probability and sales volume, it will help us better understand the entry barriers caused by local protection. It s worth mentioning that there are many samples firms with zero sales since only a small part of the sample firms have sales in a particular market, which may 21

22 lead to the sample selection bias. For this consideration, the self-selection model developed by Heckman (1979) is introduced to amend the self-selection phenomenon in the particular destination. According to Heckman (1979), due to the presence of self-selection in destination market entry, it is necessary to identify how firms are selected engaging in certain destination in the first market selection equation. In this paper, we use the type of license or certificate of registration as the identification variable, since the variable only affects firms market entry decisions rather than their sales volume. Table 8 provides the Heckman model estimation results about political connection and market sales, where equation (1) represents the probability of entry in a specific market and equation (2) represents the sales volume in a specific market. No strong evidence is seen that political connection increases the sales volume in remote markets, though it does facilitate domestic market entry. Regardless of sales in other cities or in other provinces, the estimated coefficient of political connection fails in the test of significance. It verifies that firms may ease entry barriers by virtue of political connection, but the consequent sales performance depends more on such factors as firms competitiveness. For example, in markets outside the province, firms with large size and more R&D investment usually have a larger sale volume. 4.4 Importance of Sales Targets - A Case of Government Orders Many studies on economic consequences of political connection have found that politically connected firms have more access to government orders (Faccio, 2006; Goldman et al, 2008). For example, based on the data on the United States listed 22

23 companies, Goldman et al., (2008) found that after the U.S. presidential election, businesses with links with the winning party are able to win more government orders. Does the same thing also happen in China that firms close ties with the authorities help them to obtain more government orders? To this end, this study further examines the relationship between political connection and business-to-government sales from the perspective of government orders. For example, for ProvFirm firms, we further divide them into two types: ProvFirm firms with government orders and ProvFirm firms without government orders. We expect that political connection would have a larger effect on ProvFirm firms with government orders than those without. According to the regression results shown in Table 9, govlplow is significantly positive for all types of firms. For sales to other provinces or other cities in the province, the absolute values of rrr of govlplow of firms with government procurement orders (2.940 and respectively) are much higher than those without (1.565 and respectively). It indicates that the dealings between firms and the authorities do include direct government orders. Meanwhile, for firms with no direct access to government orders, the coefficient of political connection remains significantly positive, which means excluding the impact of government orders, political connection is still able to increase the entry probability for those firms. 4.5 Other Measures of Political Connection Finally, we wonder whether our key finding suffers from the specific measure of political connection used. To do that, we use the following alternative measures of political connection as robustness checks as reported in Table

24 The first alternative measure is Days of General Manager (GM) spending with the authorities (govday). In similarity with designated personnel for political connection, the connection of firms managers with government agencies also reflects the closeness of political connection. To be specific, the World Bank survey asked "average days of General Manager dealing with the government in a month". govday is set to 1 if the time is greater than the median value of six days, and 0 otherwise. Further, the MNL estimation results indicate that, relative to the baseline, govday is significantly positive for Domestic firms. It means that from the perspective of govday, political connection also promotes market entry, again confirming the importance of political connection in impacting on firms business in distant markets. The second alternative measure is whether the general manager is appointed by the authorities (govapon). Apparently, direct appointment by the government speaks for closer ties and a sound relationship between firms and the authorities. However, the estimation results do not support the importance of govapon in offsite sales. The reason may be that direct appointment is usually seen in SOEs or partially privatized SOEs, so this indicator reflects the changes in the ownership of firms, rather than the efforts by the firms to expand the market. Additionally, for firms whose GM is directly appointed by the authorities, it may have strong incentives to stay in the local markets instead of expanding their business outsides. The third measure is Degree of closeness in relationship between firms and government departments. The World Bank survey also encompasses an evaluation of the relations with various government departments, covering tax (relatax), public 24

25 security (relapub), and labor and social security (relalab). For this consideration, these new indicators are introduced to measure the degree of political connection, but the results do not support the importance of such relations in offsite sales. Similar with the direct appointment of GM, if firms have close relationships with those government departments, it reflect that firms may focus more on their local business, and thus firms have less motivation to sell their goods to remote destinations. 5. Conclusions Considering the widespread trade barriers existing between regions in the domestic market in China, this paper examines in detail the impact of political connection on multi-destination sales decisions of the sample firms gathered from the World Bank Survey in The main findings are as follows. First, in the presence of local protection, firms political connection with local governments is indeed conducive to their entry in remote domestic markets, especially in other provincial destinations. In contrast, this role is absent for export-oriented firms, implying that political connection mainly eases domestic trade barriers brought about by local administrative protection. This core conclusion is robust and stands firm, taking into account the endogeneity of political connection and the control of firms normal travel expenses. Second, we also find that the role of political connection is closely related to firm ownership types. Private firms can better benefit from the close political connection with local governments, while there is no strong evidence on that the sales behavior of SOEs links to political connection. It means that local protection mainly affects the 25

26 market expansion of private firms, so that private firms are more inclined to seek political connection to overcome market entry barriers especially in markets of other provinces. Third, another important finding is that the role of political connection largely depends on the degree of local protection. Firms who are subject to local protectionism gain more from political connection, proving local protection is indeed an important source of domestic entry barriers. However, political connection hardly plays any role once local protection becomes severe, which reminds us that political connection is not a panacea for removing the domestic entry barriers. This paper provides empirical evidence for an overview of domestic market entry barriers in the process of transition in China. As a result of the increased competition of local authorities, the prevalent domestic market segmentation and regional trade barriers seriously hamper the business behavior of domestic market expansion, and it forces Chinese firms, especially the private firms to resort to close political ties with local governments to overcome such market entry barriers. However, recognizing the positive role in freer market entry, it is worthy of attention that political connection is by no means a fundamental solution to deal with administrative interventions which are still rampant at China s provincial borders. Firms need to spend additional resources and costs in developing and nurturing political connection with government officials and departments, thereby increasing their transaction costs for doing business, but also creating huge opportunities for rent-seeking activities to take place in government administrations. Removing local 26

27 administrative intervention in Chinese domestic market by building a unified and integrated domestic market will therefore help nurturing private entrepreneurship by engaging firms into more productive rather than unproductive or even destructive activities (Son and Song 2015). Such market integration also has high significance for long-term growth of the Chinese economy as it reduces firms transaction costs. References [1] Acemoglu D., Egorov G., and Sonin K., Political selection and persistence of bad governments. The Quarterly Journal of Economics, 125(4): [2] Allen F., J Qian, and M. J. Qian., Law, Finance, and Economic Growth in China. Journal of Financial Economics, 77: [3] Bai, C.E., Y. J. Du, Z.G. Tao, and S.Y. Tong., Local Protectionism and Regional Specialization: Evidence from China's Industries. Journal of International Economics, 63(2): [4] Berdahl L., (Sub) national Economic Union: Institutions, Ideas, and Internal Trade Policy in Canada. The Journal of Federalism, 43(2): [5] Bernard A and J. B. Jensen., Exporters, Jobs and Wages in U.S. Manufacturing , Brookings Papers on Economic Activity: Microeconomics, [6] Bernard A B, Jensen J B, and Redding S J., Firms in international trade. National Bureau of Economic Research. 27

28 [7] Bernard A, B Jensen, S. Redding and P. Schott., The Empirics of Firm Heterogeneity and International Trade. Annual Review of Economics, 4: [8] Brandt, Loren, Chang-tai Hsieh, and Xiaodong Zhu., Growth and Structural Transformation in China. In China s Great Economic Transformation, ed. Loren Brandt and Thomas G. Rawski, Cambridge and New York: Cambridge University Press. [9] Cai H, Treisman D., Does competition for capital discipline governments? Decentralization, globalization, and public policy. American Economic Review, [10] Chen C, Li Z, Su X., Rent seeking incentives, political connections and organizational structure: Empirical evidence from listed family firms in China. City University of Hong Kong Working Paper. [11] Chu, S. N. and L. Song., Promoting private entrepreneurship for deepening market reform in China: a resource allocation perspective. China and World Economy, 23, (1): [12] Clarke, D., P. Murrell, and S. Whiting., The Role of Law in China s Economic Development, in T. Rawski and L. Brandt ed. China s Great Economic Transformation. Cambridge University Press. [13] Cull R, Xu L C., Institutions, ownership, and finance: the determinants of profit reinvestment among Chinese firms. Journal of Financial Economics, 77(1):

29 [14] Das-Gupta A.,2006. Internal trade barriers in India fiscal check-posts. South Asia Economic Journal, 7(2): [15] Djankov S, La Porta R, and F. Lopez-de-Silanes, The regulation of entry. The Quarterly Journal of Economics, 117(1): [16] Faccio, M., Politically Connected Firms. American Economic Review, 96: [17] Fan, C. Simon and Xiangdong Wei, The Law of One Price: evidence from the transitional economy of China. Review of Economics and Statistics, 88 (2006): [18] Fan,J.P.H.,T.J.Wong and T.Zhang, Politically Connected CEOs, Corporate Governance and Post-IPO Performance of China's Newly Partially Privatized Firms. Journal of Financial Economics,84: [19] Fisman, R., Estimating the value of political connections. American Economic Review, 91: [20] Gilley, G., Breaking Barriers. Far Eastern Economic Review, [21] Goldman E, Rocholl J, So J., Political connections and the allocation of procurement contracts. Unpublished paper. [22] Greenaway D, Kneller R., Firm heterogeneity, exporting and foreign direct investment. The Economic Journal, 117(517): F134-F161. [23] Hausman. J and D. Mcfadden, Specification tests for multinomial logit model. Econometrica, 52: [24] Heckman, J.J., Sample selection bias as a specification error. Econometrica, 29

30 47(1): [25] Herrmann-Pillath C, Libman A, Yu X., Economic integration in China: politics and culture. Journal of Comparative Economics, 42(2): [26] Holz, C No Razor s Edge: Reexamining Alwyn Young s Evidence for Increasing Interprovincial Trade Barriers in China. Review of Economics and Statistics, 91(3): [27] Joh S W, Chiu M M., Loans to distressed firms: Political connections, related lending, business group affiliation and bank governance. Econometric Society 2004 Far Eastern Meetings. Econometric Society, (790). [28] Johnson, S., D. Kaufmann, and P. Zoido-Lobaton, Regulatory discretion and the unofficial economy. American Economic Review 88, [29] Koopman, Robert & Wang, Zhi & Wei, Shang-Jin, Estimating domestic content in exports when processing trade is pervasive. Journal of Development Economics, 99(1): [30] Krueger A O., The political economy of the rent-seeking society. American Economic Review, 64(3): [31] Kroszner R S, Stratmann T., Interest-group competition and the organization of congress: theory and evidence from financial services' political action committees. American Economic Review, 88(5): [32] Li H, Meng L, Wang Q, et al., Political connections, financing and firm performance: Evidence from Chinese private firms. Journal of Development Economics, 87(2):