WPS6474 Policy Research Working Paper 6474 Network Proximity and Business Practices in African Manufacturing Marcel Fafchamps MÃ¥ns Söderbom The World Bank Development Economics Vice Presidency Partnerships, Capacity Building Unit June 2013 Policy Research Working Paper 6474 Abstract Patterns of correlation in innovation and contractual sense are more likely to adopt similar practices. Indeed, practices among manufacturing firms in Ethiopia and for certain practices, adoption decisions appear to be local Sudan are documented. Network data that indicate strategic substitutes: if one firm in a given location uses whether any two firms in the utilized sample do business a certain practice, nearby firms are less likely to do so. with each other, buy inputs from a common supplier, or These results suggest that the diffusion of technology and sell output to a common client are used for the analysis. new business practices may play a more limited role in Only limited support is found for the commonly held spurring growth in Africa’s manufacturing sector than is idea that firms that are more proximate in a network often assumed in the present policy discussion. This paper is a product of the Partnerships, Capacity Building Unit, Development Economics Vice Presidency. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at marcel.fafchamps@economics.ox.ac.uk and mans.soderbom@economics.gu.se. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Network Proximity and Business Practices in African Manufacturing Marcel Fafchamps and MÃ¥ns Söderbom . JEL classification codes: O1, D2, D4. Keywords: Networks, technology diffusion, business practices, manufacturing firms, Africa. Marcel Fafchamps is a professor at the University of Oxford and a Co-Director of the Centre for the Study for African Economies; his email address is marcel.fafchamps@economics.ox.ac.uk. MÃ¥ns Söderbom (corresponding author) is a professor at the University of Gothenburg; his email address is mans.soderbom@economics.gu.se. The authors thank Matt Jackson, Yann Bramoullé, Alain de Janvry (Editor), three anonymous referees, and participants at seminars at the University of Gothenburg, Trinity College Dublin, the University of Wageningen, the Centre for the Study for African Economies conference Economic Development in Africa, Oxford, 2010, and the Nordic Conference on Development Economics, Copenhagen, 2011, for their very helpful comments. The authors also thank Magdi Amin for his encouragements and support, Gebrehiwot Ageba for his help with the geographical and social network data in Ethiopia, the World Bank for providing the data, and Rowena Chu and Michele Valsecchi for their high-quality research assistance. MÃ¥ns Söderbom thanks Sida for financial support. A supplemental appendix to this article is available at http://wber.oxfordjournals.org. Although technological upgrading and institutional innovation are critical for growth, these factors are particularly critical in Africa, where productivity has remained low. This fact begs the question of why productivity-enhancing innovations have not diffused equally to different countries or regions (Parente and Prescott 1994). Since Griliches (1958), the dominant model of technology adoption is one in which information about a more productive technology diffuses through the economy, and the new technology is subsequently adopted by individual firms. In this model, obstacles to the circulation of information, such as social or economic segmentation, delay technology adoption. Delays may also arise because of funding constraints or adoption costs, such as learning by doing, experimentation, and adjustment costs. As a result, pockets of backward technology may remain. This general view pervades much of the economic discourse on growth and development. A form of diffusion externality is either built into or hidden in all of the endogenous growth models in which technological innovation fuels growth (Parente and Prescott 1994; Romer 1990; Grossman and Helpman 1991; Aghion and Howitt 1992). The literature on the Industrial Revolution and the rise of the Western world describes how innovations in technology and business practices diffuse to neighboring enterprises, towns, and countries (North 1973; Mokyr 1990). Similarly, the literature on agglomeration effects ascribes a key role to the diffusion of innovative technology and business practices to nearby firms (Jacobs 1969; Fujita, Krugman and Venables 1999; Muendler, Rauch, and Tocoiand 2012). Analogous ideas underlie much of the literature on the productivity benefits from FDI and international trade (Casella and Rauch 2002; Tybout 2000). Supplier-client relationships are considered one important channel of diffusion among firms (Jacobs 1969; Rauch and Casella 2003). Another channel is competition between firms in the same market, especially foreign firms (Kraay, Soloaga and Tybout 2002). 2 Another strand of the economics literature has examined the diffusion of innovations within countries and regions. A shared assumption that underlies much of this literature is that by interacting, firms learn from each other about technological and institutional innovations that raise productivity. Although there is a rigorous body of research on technology diffusion among farmers (Griliches and Lichtenberger 1984; Young and Burke 2001), much of the existing literature on manufacturing firms in developing countries remains descriptive and relies principally on case studies (Sutton and Kellow 2010; Sutton and Kpentey 2012; Sonobe and Otsuka 2011). In this paper, we offer statistical evidence on the diffusion of innovations among manufacturing firms in Ethiopia and Sudan. 1 Our approach is to examine whether innovative business practices are correlated more strongly between firms that are relatively close in a network or a market sense. We find some evidence for a correlation in business practices, but the evidence is less convincing than one would expect if the diffusion effects were strong. Furthermore, we find evidence that along some dimensions (principally geographical distance), firms are more similar to distant firms than to nearby firms. This observation suggests that some adoption decisions are local strategic substitutes: if some firms adopt a certain practice, the incentive for other firms to adopt it appears to be reduced. This phenomenon is partly confirmed by noting that the practices for which we find evidence of strategic substitution, namely, R&D and vocational training to workers, are the very practices that are the most vulnerable to free riding by other firms. Overall, the evidence for diffusion and complementarities is weaker than one might expect given the emphasis in much of the current policy discussion on diffusion and agglomeration economies as a source of improved firm performance in Africa (Collier 2007; Page 2012). 3 This paper is organized as follows. Section 2 discusses the conceptual framework and some key methodological issues, section 3 describes the econometric testing strategy, section 4 provides information about the data, section 5 presents the econometric results, and section 6 concludes. <>I. Conceptual Framework: Diffusion in Networks Consider two economic agents, i and j , in a network. 2 The diffusion of a practice along the network means that i is more likely to adopt the practice if j has adopted it. This statement is equivalent to the assertion that the adoption decisions of i and j are strategic complements. To formalize this observation, let g ij = {0,1} denote a network link between two agents i and j and define the network matrix as G ≡ [ g ij ] , where g ii = 0 . Suppose that there are N agents. We follow Liu et al. (2012) and Bramoullé and Kranton (2011) in writing the payoff of agent i as follows: 1 2 Ï€ i = α i yi + γgi y + Ï?yi gi y − yi 2 , where yi denotes the action of agent i , y ≡ [ y1 ,..., y N ] is a vector of the actions of all of the agents, gi = [ gi1 ,...giN ] is a vector of the neighbors of i , the Greek letters are parameters, and the final term represents the cost of taking action yi , which is assumed to be quadratic for the sake of simplicity. Each agent chooses yi ≥ 0 to maximize the payoff Ï€ i . The first-order condition for an interior solution is yi = α i + Ï?g i y. (1) The parameters α i , γ , and Ï? are now straightforward to interpret: α i is a profitability parameter; Ï? indicates whether the actions are strategic complements ( Ï? > 0 ) strategic substitutes ( Ï? < 0 ), or neither complements nor substitutes ( Ï? = 0 ); and γ indicates whether there are positive externalities ( γ > 0 ), negative externalities ( γ < 0 ), or neither positive nor 4 negative externalities ( γ = 0 ). Note that it is possible for the externalities to be negative ( γ < 0 ) even if the actions are strategic complements ( Ï? > 0 ) and vice versa. The equilibria are action vectors y that solve the system of Kuhn-Tucker conditions, which combine the first-order conditions (1) with yi ≥ 0 ∀i ∈ N . The interior solutions y satisfy y = ( I − Ï?G ) −1 A, where A ≡ [α 1 ,...α N ] . When the actions are strategic complements ( Ï? > 0 ) and α i ≥ 0 for all i , a sufficient condition for an interior equilibrium is that Ï? is smaller than the largest eigenvalue of G . 3 If α i ≤ 0 for all i ∈ N , then there exists an equilibrium with y = 0 , but there may be other equilibria as well. 4 Bramoullé and Kranton (2011) characterize the equilibria that arise in network games with strategic substitutes ( Ï? < 0 ) and show that the equilibrium configuration ultimately depends on the lowest (that is, most negative) eigenvalue of G . With strategic substitutes, most equilibria have some agents choosing yi = 0 , and (some of) their neighbors choose a strictly positive yi (that is, the actions of neighbors tend to be dissimilar). In contrast, when actions are strategic complements, the actions of neighbors reinforce each other; thus, they tend to be similar (see also Jackson 2009). y ≡ y − E ( y ) = (I − Ï?G ) A −1 ~ These observations form the basis of our testing strategy: let ~ ~ where A = A − E ( A) . The covariance matrix of ~ y is (2) ( y ) = E (I − Ï?G ) A A′(I − Ï?G ) , Cov ( ~ −1 ~ ~ −1 ) where the α i s that enter matrix A are unobserved by the researcher. If the α i s are 5 ~~ ( ) independent and identically distributed, E A A′ = σ 2 I and the above expression can be simplified as follows: Cov( ~ ( y ) = σ 2 E (I − Ï?G ) (I − Ï?G ) . −1 −1 ) When the matrix G is sparse (that is, when few gij = 1 ), the ij elements of the matrix E[( I − Ï?G ) −1 ( I − Ï?G′) −1 ] that correspond to the existing links ( gij = 1 ) are approximately proportional to Ï? 2 . Other elements are functions of higher powers of Ï? and are much smaller than the elements that correspond to the linked pairs ij . In contrast, if Ï? = 0 and the α i s are independent and identically distributed, then Cov ( ~ y ) is a diagonal matrix, and yi , ~ Cov ( ~ y j ) = 0 for i ≠ j . Therefore, it is possible to test whether Ï? ≠ 0 by determining whether the values of y are more or less similar for linked pairs than for unlinked pairs. ~~ ( ) However, if E A A′ is not a diagonal matrix (that is, if the α i s are correlated), it is possible yi , ~ that Cov( ~ y j ) ≠ 0 even when Ï? = 0 . This possibility is an important caveat to keep in mind when interpreting our results; similar practices could be due either to strategic complementarity Ï? or to a correlation between the α i s (that is, a correlation in the profitability of taking action y between linked firms). Manski (1993) calls such correlations contextual effects. By similar reasoning, dissimilar practices may be due to strategic substitution or a negative correlation in the α i s. Strategic complementarity may arise for a variety of reasons. For example, the desire to imitate others or to conform to a social norm may be reinforced by peer pressure (Young and Burke 2001) and may result in complementarity. Another possibility is that the adoption of an innovation by others lowers the output price, which forces agent i to adopt the same innovation to remain competitive. In contrast, strategic substitution would arise if agent j 6 takes action y j and the incentive for individual i to take the same action weakens. For example, the possibility of free riding has long been recognized in experimentation; agents may wait for their friends and neighbors to experiment with a new technology before deciding whether to adopt it themselves (see Foster and Rosenzweig [1995] for an application to farming). The training of workers is another possible area where strategic substitution may be important: if firm j decides to train its workers, firm i may decide to try to poach them instead of training its own workers. In addition, a desire to avoid competition may be a driving force of strategic substitution. For example, if firm j decides to design its products for a high-end market, it could be optimal for firm i to tailor its products to a low-end market. <>Diffusion dynamics If information diffuses between linked agents, then, in the long run, we expect all of the connected agents to have the same information. The connection may be direct, whereby the agents are linked to each other, or it may be indirect, which implies that the agents are linked through others. This insight was initially formalized in the context of epidemiologic models on networks (see Jackson [2009] and Vega-Redondo [2006] for excellent summaries of this literature). It follows that when information has had time to percolate through the network, adoption patterns within a giant component depend exclusively on the distribution of the benefits from adoption, namely, the α i s, and on the local strategic complements and substitutes Ï? . If agents have dissimilar α i s or if Ï? < 0 , we expect sporadic adoption of business technology and practices, in which some agents adopt these things but others do not, although they all have the same information. In contrast, if agents have sufficiently similar α i s and Ï? ≥ 0 , we expect all of the agents in the same giant component to adopt similar technology and practices irrespective of whether they are directly linked. However, the latter expectation is not 7 true in the short run. If information circulates slowly, adoption decisions are more likely to be similar among agents who are directly linked. <>Business practices Thus far, we have discussed strategic complements and substitutes in general terms. Here, we briefly discuss specific business practices for which we have data, and we speculate about whether they are more likely to be strategic complements or substitutes for manufacturing firms in a developing country. 1) Technology: The adoption of more advanced equipment and machinery is likely to be a strategic complement within a given sector and region. Because regional firms in the same sector compete with each other, they must keep up with each other in terms of productivity. However, some firms may strategically choose to focus on niche products and markets that are poorly served by other firms to avoid competition (Fafchamps 1994). Such behavior may lead to differences rather than similarities in the technology decisions of firms in the same location. 2) Internal organization: Innovations in the internal organization of a firm should follow similar logic. If other firms gain a competitive edge by adopting a better organizational structure, competitors should follow suit. However, this maxim may not apply to firms that eschew competitive pressure by focusing on niche markets and products (see the previous item). 3) R&D: If firms compete through innovation, high R&D by some firms will induce others to invest in R&D as well. Therefore, we expect R&D to be a strategic complement unless firms can act as free riders by imitating the innovations of other firms or by choosing R&D strategically to avoid competition. 4) Vocational training of workers: If better-trained workers raise productivity, competition 8 between firms will lead them to train workers if new recruits are insufficiently qualified. However, firms may free ride and hire workers who have been trained by other firms instead of providing their own training. Thus, vocational training may be a strategic complement or a substitute. 5) Contractual practices: Because contractual practices involve other firms by definition, strategic complementarities in this area are likely to be stronger. For instance, if one firm imports from abroad or subcontracts part of its production, other firms may find it easier to import or subcontract in the same way. However, we cannot a priori rule out strategic substitution, such as if firms purchase inputs from the importing firm rather than importing these inputs themselves. 6) Reputational sanctions: Because reputation sanctions contain a strong public-good component, they are very likely to exhibit strategic complementarity. Indeed, the threat of exclusion from future trade has the strongest deterrent effect if all of the firms in the industry participate. Hence, the incentive to adopt a reputational sanction is highest when most other firms have already adopted it. The above discussion suggests that different types of proximity may have different effects. In principle, the strategic complementarities that arise from information exchange apply to all of the practices listed above. If information pertaining to technological, organizational, and contracting innovations circulates through supplier-client relationships, we expect such proximity to matter. The strategic complementarities that arise from competition should generate the strongest similarity among firms that share the same market, such as firms in a given sector and location. This observation is most relevant for technology, internal organization, and R&D because other channels of adoption diffusion are expected to be less important in these areas. If upstream and downstream firms face different competitors, which is 9 probable, the strategic complementarities that are driven by competition are expected to be smaller between firms that are located at different levels of the value chain. It follows that if we use geographical proximity as a proxy for competition, supplier-client proximity, which identifies different points on the value chain, may be associated with less similar practices. Thus far, we have discussed the adoption of practices. It is also possible to investigate payoffs directly, such as by analyzing firm performance and growth. In section 1 in the online appendix, we derive an expression for the covariance in profits across firms. 5 We show that if γ ≠ 0 (that is, if externalities are present), then positive externalities manifest as proximate firms that have similar performance. In contrast, negative externalities imply dissimilar performance. Furthermore, we show that even in the absence of externalities, firms’ performance may be similar because of a correlation in firm-specific conditions α i and α j ; these correlations are the so-called contextual effects. The potential presence of these contextual effects precludes the interpretation of correlated firm performance as evidence of externalities. <>Diffusion across heterogeneous firms Firms are heterogeneous, and diffusion patterns across firms are likely to depend upon enterprise characteristics. 6 For example, the scope for the diffusion of innovations between sectors may be limited if these sectors use technologies that are very different. Similarly, organizational practices that are suitable for large corporations may not be useful for microenterprises. In the model, this discrepancy is captured by differences between firms in the profitability parameter α i . If the adoption of new technologies and innovations is dichotomous, the likelihood of adopting can be expressed as λ (α i + Ï?g i y ) , where λ (.) is a logit or probit 10 function. Whereas firms with a low α i are unlikely to adopt irrespective of what neighboring firms do (that is, irrespective of Ï?g i y ), firms with a high α i are likely to adopt regardless of what others do. Thus, strategic complements and substitutes are very relevant for firms with intermediate values of α i ; for these firms, adoption may only be beneficial if neighboring firms adopt (if adoption decisions are strategic complements) or do not adopt (if these decisions are strategic substitutes). It is reasonable to assume that once they have been informed of an innovation, firms with a high α i would adopt it first and other firms would adopt it later owing to Ï?g i y effects. Therefore, we expect to observe the network-driven diffusion of innovation only among firms that are somewhat different, but the firms cannot be too different. Hence, the extent to which practices and technologies diffuse may vary, depending on the heterogeneity across firms. For instance, if all of the firms in sector A share a high α A for a particular innovation and the firms in sector B have a lower α B but a large Ï? , we expect all of the firms in sector A to adopt new innovations and technologies irrespective of whether they are linked. In contrast, we expect the firms in sector B to be more likely to adopt the same practices if they are linked to the sector A firms. In this example, although the correlation in adoption between firms within the same sector is not affected by network proximity, the correlation in adoption between firms in different sectors is stronger if these firms are linked. It is also possible that firms are heterogeneous within sector A ; whereas some firms may have a high α i and adopt new innovations and technologies, others may have a lower α i and adopt if and only if they have an adopting neighbor. Similarly, the firms in sector B may all have a low α i and may not adopt regardless of whether they are linked. As these two contrasting examples 11 illustrate, it is not entirely clear a priori what makes firms too similar or too different for network effects to affect diffusion. The economic importance of diffusion across heterogeneous firms is potentially high. For example, if Ï? is small in dissimilar firms, the diffusion of innovations will be more difficult in economies that are populated by very heterogeneous firms (much of sub-Saharan Africa has this characteristic). In such a context, not much should be expected from social networks and their ability to speed the diffusion of new ideas. Heterogeneity is also important from a methodological point of view because if we fail to take heterogeneity into account, we will underestimate the importance of networks for the subset of firms in which diffusion is occurring. As a result, we could erroneously accept the null hypothesis that networks play no role, and we must remember this point when we interpret our regression results. <>II. Testing strategy In this section, we outline the testing strategy that follows from the above reasoning. Each enterprise is a node, and we observe whether an enterprise i has adopted a practice yi . The vector g ij = (g1ij , g 2ij ,..., g Mij ) represents the supplier-client links between two enterprises i and j , and dij represents the geographical distance between them. We want to test whether two enterprises i and j are more likely to have a similar practice y if they are close in a network and geographical sense—that is, whether some or all of the elements of g ij are equal to one or if dij is small. For this purpose, we estimate models of the following form: y i − y j = g ij θ + ωd ij + | x i − x j | β + u ij (3) , where θ = (θ1 , θ 2 ,..., θ M ) is a vector of coefficients that are associated with network links, ω is a coefficient that reflects the relationship between the geographical distance and the outcome 12 similarities, | x i − x j | is a vector of absolute differences in the control variables x that is included to reduce omitted variable bias, β is a vector of parameters, and uij is an error term. 7 A negative θ m in (3) means that y is more similar when firms i and j have a link g mij = 1 . For the geographical distance dij , the interpretation of the sign of ω is exactly the opposite of the interpretation of the sign of θ m . Conversely, a positive θ m or a negative ω would mean that linked or nearby firms are more dissimilar. If y is more similar in proximate firms, then this occurrence is consistent with a situation in which adoptions by different firms are strategic complements; in contrast, if y is dissimilar in proximate firms, then adoption by different firms may be a strategic substitute. A positive β means that firms that share a similar x tend to have a more similar y. A negative θ m does not by itself imply network diffusion because firms i and j may have correlated technology and contractual practices for reasons other than network or geographical proximity, such as because they are subject to similar contextual effects (corr (α i , α j ) > 0) that are not adequately controlled for by the quantities | x i − x j | . If these unobserved contextual effects were more strongly correlated in linked firms, they would bias θ m below 0 . Hence, if we find a significantly negative estimate of θ m , the reason may be either diffusion or unobserved contextual effects. However, if θ m is positive or not significantly different from zero, the net effect of the diffusion and the contextual effects is likely to be positive or zero. There are two possible exceptions to the above scenarios. The first exception is when diffusion is rapid and all firms belong to a single connected network. In this case, our identification strategy will fail because the similarity of firms will depend exclusively on their α i s and not on 13 the distance between them. Hence, we will observe a zero θ even though diffusion across network links is taking place. The second exception is when the strategic complementarities and substitutes precisely offset each other. Although it is possible, this situation seems unlikely. If unobserved contextual effects can only generate positive correlations in technology and business practices, which is likely, then a nonsignificant θ indicates that the network diffusion is zero, and a positive θ suggests that the presence of strategic substitution affects adoption decisions. However, we cannot completely rule out the possibility that a negative correlation between practices could be the result of a negative correlation in the profitability of adoption α i . For instance, if an innovation such as subcontracting is profitable for upstream firms but not for downstream firms, then firms that are linked as suppliers and clients will have negatively correlated practices because suppliers, by definition, are upstream relative to their clients. Equation (3) is a dyadic regression. The dependent and independent variables are defined for every pair of firms i , j in the data, which implies that there are n × (n − 1) observations that underlie the regression ( n denotes the number of firms). Dyadic observations are not independent because the residual u ij is correlated with u ik . To compute standard errors that are robust with respect to the correlation in the error term across firms, we use the bootstrapping procedure that is described in section 1 in our online appendix. <>III. Data To implement our testing strategy, we use detailed firm-level data that were collected under the leadership of the World Bank in Ethiopia and Sudan. Virtually the same questionnaire and sampling strategies were used in the two countries. The data on the Ethiopian firms were collected as part of the Ethiopia Investment Climate Survey, which was implemented by the Ethiopian Development Research Institute in mid-2006. 8 The survey covered 14 major cities 14 located in seven regions of Ethiopia, and 42 percent of the observations came from Addis Ababa. The survey included firms with at least five permanent employees in four sectors: furniture, wood, and metal; food and beverages; leather and leather products; and textiles and garments. In all, 360 manufacturing firms were surveyed. The data from the Sudanese firms were collected as part of the Investment Climate Survey, which was launched in November 2007 and conducted by H&H Consultancy. This company is a Sudanese management consulting firm with expertise in conducting complex surveys. 9 Thus, the data were collected before South Sudan seceded from Sudan in July 2011; therefore, our sample includes firms in what is now South Sudan (see note 1). The survey covered 432 manufacturing firms in eight states, and most of these firms were private. The capital city of Khartoum accounted for 52 percent of the sample observations. No sector represented more than 20 percent of the sample; hence, the survey was diverse in terms of sector. The largest sectors were food and beverages (18 percent) and fabricated metal products (16 percent). Microenterprises were not covered. After deleting the observations that had too many missing values, we obtained a sample of 304 firms for Ethiopia and 401 firms for Sudan. 10 This subset of the observations formed our baseline sample. 11 Summary statistics are shown in table 1, and the variables that constitute our control vector are presented first. More mature firms and firms with higher-quality management should be more adept at recognizing the value of new technologies and business practices. Female ownership is included because female-headed businesses have been shown to be less growth oriented (de Mel, McKenzie and Woodruff 2009; Fafchamps 2003). We also include firm size, which is represented by the (log of the) total firm employment. The average log employment is 3.37 in Ethiopia (which corresponds to 29 employees) and 2.91 in Sudan (which corresponds to 18 employees). 15 Next, we report information on firm practices. Initially, we focus on the variables for which strategic complementarities across firms are a priori thought to be less strong, such as innovation. We end by considering the variables for which strategic complementarities are likely to be the strongest, such as reputation mechanisms. Within each category, adoption by a given firm may be correlated across individual practices, and this correlation may be positive or negative (if some practices are partial substitutes for each other). In this case, examining each practice separately yields inefficient inferences. To guard against this possibility, we follow the approach suggested by Kling, Liebman, and Katz (2007) and summarize the available information within each category using factor analysis. Thus, we construct an additional dyadic dependent variable from the first principal components. The factor loadings for each category are reported in table 1. The first variable that we consider is a dummy variable that indicates whether a firm introduced a new product in the year preceding the survey. Between one-third and one-half of the surveyed firms responded positively to this question. Approximately one-half of the firms invested in plants and equipment in the previous year in both countries. A nonnegligible proportion of the surveyed firms had spent money on R&D: 13 percent for Ethiopia and 23 percent for Sudan. In addition, we note some usage of information technology (IT), mostly in the form of email. At the time of the surveys, few manufacturing firms in Sudan or Ethiopia had a website. Information on labor management and investment in human capital is presented next. We find a higher ratio of nonproduction workers to total employment in Sudan than in Ethiopia, which suggests that Sudanese firms are less able to manage their workforce with a small number of clerks and managers. 12 In both countries, a substantial minority of firms had provided in-house 16 or external training to their workers, but the majority had not. The next panel of table 1 covers contractual practices. Firms were asked whether they imported inputs directly from abroad. Although buying directly from abroad requires trust, it is likely to improve the quality of the raw materials that are used in a firm’s production process. We find a difference between the two countries: landlocked Ethiopia lagged behind Sudan. Firms were also asked whether they sold on credit to any of their customers. A majority of manufacturing firms sell on credit to at least some of their customers, but a large minority does not. The data also show that subcontracting part of a firm’s production to other firms is rare. Next, we examine the extent to which the surveyed firms rely on reputation to enforce contracts with suppliers and clients. The respondents were asked five closely related questions: (i) If you have a dispute with a customer, will other customers find out? (ii) If another firm has a dispute with a customer, will you refuse to deal with that customer? (iii) If you have a dispute with a customer, will other firms refuse to deal with that customer? (iv) If you have a dispute with a supplier, will other suppliers find out? (v) If you have a dispute with a supplier, will other firms refuse to deal with that supplier? For each of these questions, we code y = 2 for “yes,â€? y = 1 for “maybe/do not knowâ€? and y = 0 for “no.â€? Hence, high values correspond to stronger reputation effects. The summary statistics presented in table 1 suggest that news about a dispute often travels to customers and suppliers. These statistics also suggest that the reputational sanction imposed on the customers and suppliers that are involved in a dispute is not severe; firms typically continue to deal with customers and suppliers that have been involved in a dispute. Similar results have been reported by Bigsten et al. (2000) and Fafchamps (2004) for African manufacturing. 17 A key module of the survey contains information about the names of the firms’ trading partners and their approximate geographical locations. The respondents were asked to name up to three clients and three suppliers. 13 Using the information from this module, we construct simple measures of network proximity between the firms in the two samples. Summary statistics for these measures are reported in table 2. We begin by constructing a dyadic dataset of unique firm pairs. For instance, because there are 304 firms in the Ethiopian sample, there exist 304 × 303 / 2 = 46, 056 unique enterprise pairs (i, j) in that sample. For each pair (i, j), we construct dummy variables that capture the different concepts of network proximity. When two firms are close in that network, we consider them to be linked. The most direct network proximity measure that we use is whether i and j buy or sell from each other. We are only able to identify a small number of such links in our data: 60 in Ethiopia and 5 in Sudan. The fact that there are so few upstream and downstream links among the sample firms is partly driven by the focus of the surveys on light manufacturing because clients are seldom manufacturers. We also construct dummy variables that indicate whether i and j have a common supplier or a common client. These types of links are more common: there are 481 supplier-based links and 273 client-based links in the Ethiopian data and 171 supplier-based links and 678 client-based links in the Sudanese data. These network proximity variables constitute the core of our vector g ij . The last proximity dummy is the distance d ij , which is defined as the log of the distance between i and j plus one. <>IV. Empirical Analysis Our objective is to test whether the outcomes and practices that are related to technology, human capital, contracting, and reputation are more similar among firms that are close to each other, either in a network sense or geographically. To that end, we estimate the parameters of 18 model (3). Our estimation technique is linear regression (ordinary least squares), and standard errors are bootstrapped to make them robust with respect to heteroskedasticity and correlation in error terms across firms. We refer to the presented results as baseline results. Additional results are available in the online appendix. <>Innovation and R&D We begin by investigating the association between geographical and network proximity and innovation and R&D. We construct dyadic dependent variables from dummy variables that measure whether firms introduced a new product in the previous year, invested in plants and equipment in the previous year, or conducted any R&D. A fourth outcome variable is constructed based on a firm-level measure of the extent of IT usage; the value of this variable is zero if IT is not used at all, one if the firm uses email, and two if the enterprise has a business website. 14 The dyadic regression results are shown in table 3, columns [1]–[4] for Ethiopia, and columns [6]–[9] for Sudan. In columns [5] and [10], we report results in which we use the first principal component of all four categories to construct the dyadic dependent variable. The estimated network proximity coefficients differ in the two countries. For Ethiopia, the dummies that track whether i and j trade with each other, have a common supplier, and have a common client are statistically nonsignificant. For Sudan, we obtain a negative and statistically significant coefficient of trade in the R&D regression (column [8]), and we obtain negative and significant coefficients (at least at the 10 percent level) of having a common supplier in the regressions for investment (column [7]), R&D (column [8]), IT usage (column [9]), and the first principal component (column [10]). Hence, network proximity seems to be associated with a more similar approach to innovation and R&D throughout the firms in Sudan, but not in Ethiopia. Some of these estimated effects are large; for example, the likelihood that 19 firms report the same answer (yes or no) to the question about whether money was spent on R&D is 34 percentage points higher for firms that trade with each other than for firms that do not trade with each other. However, because of the small number of direct links in the Sudanese data (see table 2), the estimated coefficients of direct trade should be interpreted with caution. Furthermore, we find that Sudanese firms with a common client tend to differ more than other firms with respect to R&D and IT usage. This finding is not consistent with the notion that network proximity tends to result in similar practices regarding innovation. Next, we consider the role of geographical distance between firms. For Ethiopia, the distance coefficient is negative in all five of the specifications shown in table 3, and this coefficient is statistically significant at least at the 10 percent level in four of these specifications. Hence, geographical proximity tends to be associated with greater differences in innovation practices. The results are similar for Sudan: the distance coefficient is negative and highly statistically significant in the models for R&D (column [8]), IT usage (column [9]), and the first principal component (column [10]). These results suggest that for technology, strategic substitution effects dominate strategic complementarities for firms that are located near each other. The control variables in these regressions have explanatory power. The estimated coefficients of the same-sector dummy are negative in all of the specifications except [6], and these coefficients are often statistically significant. This result indicates that, as expected, firms in the same sector tend to have similar innovation practices. Differences in firm size, which are measured as the absolute difference in the log of employment, are positively associated with differences in innovation practices in all of the specifications, suggesting that firms of similar size tend to adopt similar practices. There is also some evidence that managers of the same gender or with similar levels of education select similar innovation practices. The coefficients 20 of differences in managers’ experience or firms’ ages are mostly nonsignificant. <>Human capital and labor management Table 4 shows the results for our regressions on labor management and investment in human capital. We find no evidence that network proximity is associated with greater similarity in training decisions or labor management in firms. In fact, we obtain a positive and statistically significant coefficient of common clients in specifications [3], [4], and [6], which indicates that firms with a common client tend to have more distinct training policies than firms that do not share a common client. The estimated coefficients related to the distance between firms are negative in all of the specifications except [5], and these coefficients are statistically significant in four of the specifications (columns [2], [3], [4], and [6]). As with the results for innovation, this result implies that firms that are located close to each other tend to differ more with respect to their human capital decisions compared with firms that are far apart. This finding is consistent with strategic substitution. One possibility that is often emphasized in the literature on agglomeration effects (Henderson 1988; Glaeser et al. 1992) is that firms hire workers who have been trained by other firms. Thus, if there are more nearby firms that provide the necessary training, a given firm needs to train its workers to a lesser extent. Alternatively, strategic substitution may be driven by incentives to avoid local competition. For example, if two firms with similar human capital produce similar output, they will compete with each other if they are based in the same local market. By locating themselves in different places, both firms would face less competition and presumably higher profits. Another possibility is that firms that are located in the same place decide to differentiate their output, which may lead to differences in technology and human capital demand. Mechanisms such as these would result in the pattern 21 that we observe in the data of greater differences between firms that are located close to each other than between firms in distant locations. Furthermore, we find that, as expected, firms of similar size and firms in the same sector tend to be more similar with respect to their training decisions than firms of different sizes or in different sectors. The coefficients related to the other control variables, that is, differences in firms’ ages and in managers’ education, experience, and gender, are mostly nonsignificant. When they are significant, their coefficients are usually negative, which suggests that greater differences in these firm-level characteristics are associated with closer similarities in outcomes. <>Contractual practices Next, we investigate how the following three measures of contractual practices correlate across firms: whether a firm imports inputs directly, whether it sells on credit, and whether it subcontracts part of its production. The results are shown in table 5. For Sudan, we find a negative and highly significant coefficient on the dummy variable that indicates whether firms i and j trade directly with each other in the models for direct imports, selling on credit, and the first principal component. Thus, Sudanese firms that trade with each other tend to have more similar contractual practices. In addition, having a common supplier is associated with a greater similarity in direct imports, although this effect is only statistically significant at the 10 percent level. In contrast, for Ethiopia, the correlation between network proximity and the similarity in contractual practices is weak and nonsignificant in all of the specifications except for subcontracting, for which we obtain a positive coefficient related to having a common supplier (column [3]). 22 The estimated distance coefficients vary considerably across regressions. In two regressions, they are positive and significantly different from zero (direct imports and selling on credit in Sudan; columns [5] and [6]), which suggests that firms that are close to each other have more similar contractual practices than other firms. However, in two other regressions, the coefficients are significantly negative (direct imports in Ethiopia and subcontracting in Sudan; columns [1] and [7]). For both countries, distance is statistically nonsignificant in the regressions that model the difference in the first principal component. Thus, it is difficult to see a pattern here, which may be because the relative importance of strategic substitution and diffusion varies from one contractual practice to another. Regarding the control variables, the pattern is similar to what we observed above; whereas firms of similar size and in the same sector tend to have similar contractual practices, for other controls, the results are mixed. <>Reputation mechanisms Here, we examine whether there is evidence that network links facilitate the diffusion of information on contractual disputes between suppliers and clients. The theoretical literature has emphasized the role of the diffusion of information on contractual disputes along social networks in the development of modern market institutions (North 1990; Greif 1993). Consequently, we expect to find a strong correlation in answers from firms in the same networks. Using the five questions on the perceived consequences of disputes that are discussed in section 3, we code y i = 2 for “yes,â€? yi = 1 for “maybe/do not knowâ€? and y i = 0 for “no.â€? Then, we compute | y i − y j | for every pair of firms in the data. 15 The regression results are shown in table 6a and do not conform to theoretical expectations. Except for isolated cases in which a 23 network regressor is significant (columns [6] and [10], but with opposite signs), the social network variables are not significant. One possible explanation is insufficient power: the five categorical reputation variables may contain insufficient information to identify the social network coefficients. Additionally, no network variable is significant in the principal component regressions shown in table 6b; hence, combining the information contained in all five of them does not lead to better results. Furthermore, the coefficients related to the control variables are nonsignificant in the vast majority of cases. There are two possible interpretations of these findings: either the information about contractual disputes does not diffuse along the type of social networks that we have been able to measure, or this information diffuses so well that social links do not matter. One way to identify which of these two interpretations is more likely is to examine the coefficient of the distance variable. Even though information may diffuse rapidly along social networks within certain areas, the diffusion of information need not happen everywhere because strategic complementarities in diffusion create the possibility of multiple equilibria. If there are multiple equilibria, we expect to find that firms that are distant from each other perceive the consequences of contractual disputes differently. However, we do not find distinct perceptions of contractual disputes in distant areas. For Ethiopia, although the distance coefficient is negative and highly significant in three of the specifications shown in table 6a, it is positive and significant in the remaining two specifications. Dyadic differences in the principal component that is based on the five individual variables are negatively and significantly related to distance. For Sudan, the distance coefficient is negative and significant in two out of five individual regressions, and in the remaining cases, it is not statistically significant. These findings are difficult to reconcile with 24 the idea of widespread diffusion of contractual information among firms in the same location. If multiple equilibria are present, they seem to coexist within locations; thus, whereas some firms recognize that there are reputational consequences to contractual disputes, others in the same location do not. <>Firm performance and growth Thus far, we have focused on those business practices that may diffuse within networks. We have also investigated whether the above results are mirrored in labor productivity and growth rates, which are our measures of firm performance. The results for the specifications in which the dependent variable is defined as the absolute difference across firms in these performance indicators are shown in section 4 in the online appendix. For Ethiopia, we find little evidence that firms that are closer in the social-network sense have more similar performance. For Sudan, we find evidence that firms that share the same supplier have more similar performance than other firms. However, other results related to network links and geographical distance are weak. The overall conclusion from our analysis is that network links and geographical proximity are not strongly associated with any convergence in the performance of firms. <>Heterogeneous diffusion and networks Now, we return to the points raised in section I that are related to the diffusion patterns across heterogeneous firms. We question whether the reason we find only limited evidence of diffusion is that the firms are too heterogeneous. To investigate whether the evidence for diffusion is stronger among pairs of firms in the same sector, we interact our network and distance variables with a dummy that tracks whether firms i and j belong to the same industrial subsector, and we add these new interaction terms to the baseline specification. To minimize the number of explanatory variables, the same industry dummy is interacted with a 25 single network variable anylink ij , which is a dummy variable that is equal to one if there is any link between firms i and j . Such a link would amount to direct trade, a common client or a common supplier. The results for all of the outcome variables are shown in tables S2.E and S2.S in the online appendix. For Ethiopia, the sector-network interaction term is statistically nonsignificant in every specification, and the sector-distance interaction term is significant in only one specification (formal training; see table S2.E, column [8]; the term has a positive sign). For Sudan, the sector-network interaction term is statistically nonsignificant in every specification, and the sector-distance interaction term is significant in just one specification (direct imports; see table S2.S, column [10]; the term has a positive sign). These results suggest that sector heterogeneity is not the reason for limited diffusion. They also imply that strategic substitution is equally strong within each sector and across different sectors. We repeat this type of analysis, which focuses on firm-size heterogeneity instead of sector heterogeneity. To that end, we interact anylink ij and the distance variable ( d ij ) with a dummy that tracks whether firms i and j are of similar size, and we add these new interaction terms to the baseline model. 16 The results are shown in tables S3.E and S3.S in the online appendix. For Ethiopia, the size-network interaction term is statistically nonsignificant in all regressions. However, the size-distance term is negative in the vast majority of cases and is often statistically significant. This result suggests that strategic substitution is stronger across firms of similar size than across firms of differing size, which may be because geographically close firms strategically choose to differentiate themselves from each other to reduce competition. For Sudan, the network-size interaction term is statistically nonsignificant in all specifications, and the size-distance interaction term is significant in just three specifications (table S3.S, columns 26 [6], 11 and 20). On balance, we find little evidence that size heterogeneity is a likely reason for slow diffusion, and we note that the results for Ethiopia lend further support to the idea that strategic substitution may be important. <>Market differentiation within towns Finally, we investigate how the estimated coefficients of the geographical distance change if we add a dummy variable sametownij to the baseline specification. This dummy variable is equal to one if firms i and j are located in the same town and zero otherwise. We want to establish whether market differentiation within towns causes the result that a shorter geographical distance between firms is associated with greater differences in business practices. It seems plausible to suppose that strategic substitution is strongest within towns. If markets are localized such that firms in different towns pose no competitive threat to each other irrespective of the distance between these towns, events in town k will not affect the strategic decisions of firms in town l ≠ k . In this case, the relevant geographical circumstance is whether firms are in the same town; thus, conditional on sametownij , distance does not matter. Thus, by adding sametownij to the set of explanatory variables, we generalize the baseline’s functional form with respect to the effect of distance. The results that are based on this specification for all outcome variables are shown in tables S4.E and S4.S in the online appendix. For Sudan, the coefficients on sametownij are always negative whenever they are significant, which suggests that strategic substitution effects do not primarily operate within towns. For Ethiopia, the picture is more mixed: although we obtain positive and significant coefficients in five of the regressions shown in table S4.E columns [1]–[24], we obtain negative and significant coefficients in three regressions. Overall, the support for the idea that strategic substitution effects operate primarily within towns is quite 27 limited. <>V. Conclusions In this paper, we have documented empirical patterns of correlation in the adoption of innovation and contractual practices among manufacturing firms in Ethiopia and Sudan. Our empirical analysis is based on network data that indicate whether any two firms in our sample do business with each other, buy inputs from a common supplier, or sell output to a common client. We also exploit data on firms’ locations to investigate whether firms that are near each other tend to be more similar or dissimilar than firms that are geographically far apart. Our results can be summarized as follows: (i) for Sudan, but not for Ethiopia, there is some evidence that network proximity is associated with similar innovation strategies; (ii) for both countries, there is relatively strong evidence that firms that are located close to each other differ more with respect to innovation than firms that are far apart; (iii) there is no evidence that network proximity is associated with greater similarity in training decisions or labor management across firms; (iv) there is some evidence that firms that are located close to each other differ more with respect to training decisions than firms that are geographically far apart; (iv) for Sudan, but not for Ethiopia, there is some evidence that network proximity is associated with similar contractual practices; (vi) differences in contractual practices across firms are only weakly related to geographical proximity; (vii) there is no evidence that network proximity is associated with greater similarity in the perceived consequences of disputes; (viii) there is some evidence that geographical proximity is associated with greater differences in the perceived consequences of disputes; and (ix) except for supplier-based links in Sudan, differences in firm performance are only weakly related to geographical and network proximity. Overall, the strongest results are for innovation. 28 Thus, our results provide limited support for the commonly held idea that firms that are more proximate in a network sense are more likely to adopt similar contractual and technological innovation practices. Furthermore, we find some evidence that for certain practices, adoption decisions are local strategic substitutes. Hence, if one firm adopts, other nearby firms are less likely to do so. What should we make of these results? First, we again note that a correlation in practices does not imply diffusion because there may be unobserved contextual effects. Second, the evidence presented here does not imply that the diffusion of innovation between firms can never be important or even critical for growth. However, diffusion between firms should not be taken for granted. Many of the firms in our sample follow antiquated business practices even when some neighboring firms do not, which is consistent with the observation that firms in developing countries are often more heterogeneous than firms in developed countries (see, for example, Bloom et al. [2012] for evidence that the quality of management practices is more heterogeneous for the firms in Brazil, China, and India than in the U.S.). Third, it is possible that we searched for diffusion in the wrong place (that is, among existing firms) because it is possible that, in general, the diffusion of innovations takes place not because existing firms learn to imitate each other but rather because new firms emerge that adopt innovative practices. This interpretation is consistent with the findings reported in the exporting literature: although there is limited evidence that incumbent firms learn from exporting, there is ample evidence that firms that begin exporting are more productive than average, even when they are new entrants (Clerides, Lach and Tybout 1998; Fafchamps, El Hamine and Zeufack 2008). Fourth, we acknowledge that our data have certain limitations. One potentially important limitation is that the survey asked for a maximum of three clients and suppliers, which implies that we do not have complete coverage of all of the network links. In addition, it is likely that our network-link variables are measured with error, which may cause the network effects to be underestimated in 29 our analysis. Despite these caveats, we note that in several ways, the evidence for diffusion and complementarities is much weaker than one might expect given the emphasis in much of the current policy discussion on diffusion and agglomeration economies as sources of improved firm performance. 30 REFERENCES Aghion, P., and P. Howitt. 1992. "A Model of Growth through Creative Destruction." Econometrica 60 (2): 323-51. Bigsten, A., P. Collier, S. Dercon, M. Fafchamps, B. Gauthier, J.W. Gunning, A. Oduro, R. Oostendorp, C. Pattillo, M. Sö derbom, F. Teal, and A. Zeufack. 2000. Contract Flexibility and Dispute Resolution in African Manufacturing. Journal of Development Studies 36 (4): 1–37. Bloom, N., B. Eifert, A. Mahajan, D. McKenzie, and J. Roberts. 2012. "Does Management Matter? Evidence from India." Unpublished manuscript. Stanford University. Bramoullé, Y., and R. Kranton. 2011. "Strategic Interaction and Networks." Mimeograph. Duke University. Casella, A., and J. Rauch. 2002. "Anonymous Market and Group Ties in International Trade." Journal of International Economics 58 (1): 19–47. Clerides, S., S. Lach, and J. Tybout. 1998. "Is Learning-by-Exporting Important? Micro-Dynamic Evidence from Colombia, Mexico, and Morocco." Quarterly Journal of Economics 113 (3): 903–47. Collier, P. 2007. The Bottom Billion. Why the Poorest Countries are Failing and What Can Be Done About It. New York: Oxford University Press. de Mel, S., D. McKenzie, and C. Woodruff. 2009. "Are Women More Credit Constrained? Experimental Evidence on Gender and Microenterprise Returns." American Economic Journal: Applied Economics 1 (3): 1–32. Fafchamps, M. 1994. "Industrial Structure and Microenterprises in Africa." Journal of Developing Areas 29 (1): 1–30. Fafchamps, M. 2004. Market Institutions in Sub-Saharan Africa: Theory and Evidence. Cambridge, Mass: MIT Press. 31 Fafchamps, M. 2003. "Ethnicity and Networks in African Trade." Contributions to Economic Analysis and Policy 2 (1): article 14. Fafchamps, M., and M. Sö derbom. 2006. "Wages and Labor Management in African Manufacturing." Journal of Human Resources 41 (2): 346–79. Fafchamps, M., S. El Hamine, and A. Zeufack. 2008. "Learning to Export: Evidence from Moroccan Manufacturing." Journal of African Economies 17 (2): 305–55. Foster, A.D., and M.R. Rosenzweig. 1995. "Learning by Doing and Learning from Others: Human Capital and Technical Change in Agriculture." Journal of Political Economy 103 (6): 1176–209. Fujita, M., P. Krugman, and A.J. Venables. 1999. The Spatial Economy: Cities, Regions, and International Trade. Cambridge, Mass: MIT Press. Glaeser, E.L., H.D. Kallal, J.A. Sheinkman, and A. Shleifer. 1992. "Growth in Cities." Journal of Political Economy 100 (6): 1126–52. Greif, A. 1993. "Contract Enforceability and Economic Institutions in Early Trade: The Maghribi Traders' Coalition." American Economic Review 83 (3): 525–48. Griliches, Z. 1958. "Research Costs and Social Returns: Hybrid Corn and Related Innovations." Journal of Political Economy 66: 419. Griliches, Z., and F. Lichtenberger. 1984. "Intraindustry Technological Flows and Productivity Growth: A Reexamination." Review of Economics and Statistics 66 (2): 324–29. Grossman, G.M., and E. Helpman. 1991. "Quality Ladders and Product Cycles." Quarterly Journal of Economics 106: 557–86. H&H Consultancy. 2008. Sudan Investment Climate Survey I: Final Survey Report. Report submitted to the World Bank. Henderson, J.V. 1988. Urban Development: Theory, Fact, and Illusion. New York: Oxford University Press. 32 Jackson, M.O. 2009. Social and Economic Networks. Princeton: Princeton University Press. Jacobs, J. 1969. The Economy of Cities. New York: Random House. Kling, J.R., J.B. Liebman, and L.F. Katz. 2007. "Experimental Analysis of Neighborhood Effects." Econometrica 75 (1): 83–119. Kraay, A., I. Soloaga, and J. Tybout. 2002. "Product Quality, Productive Efficiency, and International Technology Diffusion: Evidence from Plant-Level Panel Data." Policy Research Working Paper Series 2759. Washington, DC: The World Bank. Liu, X., E. Patacchini, Y. Zenou, and L-F. Lee. 2012. "Criminal Networks: Who is the Key Player?" FEEM Working Paper No. 39. Milano: Fondazione Eni Enrico Mattei. Manski, C.F. 1993. "Identification of Endogenous Social Effects: The Reflection Problem." Review of Economic Studies 60: 531–542. Melitz, M.J. 2003. "The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity." Econometrica 71 (6): 1695–725. Melitz, M.J., and G.I.P. Ottaviano. 2008. "Market Size, Trade, and Productivity." Review of Economic Studies 75 (1): 295–316. Mengistae, T., and M. Honorati. 2009. "Constraints to Productivity Growth and to Job Creation in the Private Sector." Chapter 2 in Ethiopia Investment Climate Assessment. Washington, DC: World Bank. Mokyr, J. 1990. The Lever of Riches: Technological Creativity and Economic Progress. New York and London: Oxford University Press. Muendler, M-A., J. Rauch, and O. Tocoiand. 2012. "Employee Spinoff and Other Entrants: Evidence from Brazil." International Journal of Industrial Organization 30 (5): 447–58. North, D.C. 1973. The Rise of the Western World. Cambridge: Cambridge University Press. North, D.C. 1990. Institutions, Institutional Change, and Economic Performance. Cambridge: 33 Cambridge University Press. Page, J. 2012. "Can Africa Industrialize?" Journal of African Economies 21 (AERC Supplement 2): ii86–125. Parente, S.L., and E.C. Prescott. 1994. "Barriers to Technology Adoption and Development." Journal of Political Economy 102 (2): 298–321. Rauch, J., and A.Casella. 2003. "Overcoming Informational Barriers to International Resource Allocation: Prices and Ties." Economic Journal 113 (484): 21–42. Romer, P.M. 1990. "Endogenous Technological Change." Journal of Political Economy 98 (5) pt.2: S71–102. Sonobe, T., and K. Otsuka. 2011. Cluster-Based Industrial Development: A Comparative Study of Asia and Africa. London: Palgrave McMillan. Sutton, J., and N. Kellow. 2010. An Entreprise Map of Ethiopia. London: International Growth Centre. Sutton, J., and B. Kpentey. 2010. An Entreprise Map of Ghana. London: International Growth Centre. Tybout, J. 2000. Manufacturing Firms in Developing Countries: How Well Do They Do, and Why? Journal of Economic Literature 38 (1): 11–44. World Bank. 2009. Ethiopia - Toward the Competitive Frontier: Strategies for Improving Ethiopia's Investment Climate. Washington D.C: The World Bank. Vega-Redondo, F. 2006. Complex Social Networks. Cambridge: Cambridge University Press. Young, H.P., and M.A. Burke. 2001. "Competition and Custom in Economic Contracts: A Case Study of Illinois Agriculture." American Economic Review 91 (3): 559–73 34 TABLE 1. Summary Statistics Ethiopia Sudan Obs. Mean Std. dev. Loadings Obs. Mean Std. dev. Loadings 1. Firm characteristics Firm age (years) 304 17.93 16.1 401 15.21 14.1 Education of top manager(a) 303 2.71 1.20 399 2.92 1.25 Experience of top manager (years) 304 14.5 9.77 395 17.2 12.9 (b) Any female owner? 304 0.23 382 0.15 Log(firm employment) 304 3.37 1.66 399 2.61 1.14 2. Innovation and R&D Did the firm introduce a new product last year?(b) 304 0.35 0.70 391 0.48 0.58 (b) Did the firm invest in plants & equipment last year? 304 0.52 0.67 400 0.46 0.70 (b) Does the firm conduct any R&D? 304 0.13 0.72 388 0.23 0.74 IT usage (0 = nothing, 1 = email, 2 = website) 304 0.59 0.76 0.48 401 0.45 0.78 0.74 3. Human capital and labor management Ratio of nonproduction workers to total employment(c) 304 0.27 0.17 0.24 398 0.42 0.30 0.22 (b) Any in-house training of staff last year? 304 0.28 0.83 397 0.27 0.80 Staff sent to formal training course last year?(b) 304 0.28 0.84 398 0.12 0.80 4. Contractual practices Any direct imports of inputs?(b) 304 0.31 0.67 401 0.51 0.74 (b) Do you sell on credit? 304 0.53 0.65 401 0.64 0.73 Does firm subcontract production?(b) 302 0.12 0.33 382 0.09 0.22 35 TABLE 1 (continued) Ethiopia Sudan Obs. Mean Std. dev. Loadings Obs. Mean Std. dev. Loadings 5. Reputation mechanism If you have a dispute with a customer, will other 304 1.049 0.948 0.47 400 0.808 0.934 0.48 customers find out?(d) If another firm has a dispute with a customer, will you 304 0.457 0.815 0.67 401 0.783 0.954 0.65 refuse to deal with that customer?(d) If you have a dispute with a customer, will other firms 304 0.474 0.717 0.43 401 0.788 0.899 0.63 refuse to deal with that customer?(d) If you have a dispute with a supplier, will other 304 0.914 0.926 0.46 401 0.783 0.925 0.69 suppliers find out?(d) If you have a dispute with a supplier, will other firms 304 0.398 0.682 0.47 401 0.656 0.861 0.64 refuse to deal with that supplier?(d) Source: Authors’ computations based on data described in the text. (a) 1 = less than secondary, 2 = secondary, 3 = vocational, 4 = university. (b) 0 = no, 1 = yes. (c) Nonproduction workers include professionals, managers, administrators, and sales personnel. (d) 0 = no, 1 = maybe, 2 = yes. 36 TABLE 2. Dyadic Data Ethiopia Sudan Number of unique enterprise pairs 46,056 80,200 i & j trade with each other (number of pairs) 60 5 i & j have a common supplier (number of pairs) 481 171 i & j have a common client (number of pairs) 273 678 Average distance between i & j (kilometers) 282 421 Minimum distance between i & j (kilometers) 0 0 Maximum distance between i & j (kilometers) 876 1,770 Source: Authors’ computations based on data described in the text. 37 TABLE 3. Correlates of Dyadic Differences: Technology Acquisition Ethiopia Sudan [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] Did the firm Did the firm Does the firm IT usage First Did the firm Did the firm Does the firm IT usage First introduce a invest in conduct any (0 = nothing, principal introduce a invest in conduct any (0 = nothing, principal new product plants & R&D? 1 = email, 2 = component new product plants & R&D? 1 = email, 2 = component last year? equipment website) last year? equipment website) last year? last year? |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| i & j trade with each 0.0492 −0.0471 0.0957 −0.0703 0.0792 0.0147 −0.205 −0.339** 0.451 −0.302 other (0.126) (0.117) (0.108) (0.188) (0.240) (0.350) (0.326) (0.140) (0.572) (0.237) i & j have common −0.0175 −0.00152 0.0461 −0.0630 −0.0202 −0.0799 −0.154* −0.183*** −0.272** −0.310** supplier (0.0400) (0.0456) (0.0408) (0.0660) (0.0684) (0.0740) (0.0791) (0.0648) (0.109) (0.130) i & j have common 0.0666 −0.0259 0.0110 0.0602 −0.0645 0.0247 0.00539 0.123*** 0.229*** 0.178* client (0.0586) (0.0705) (0.0692) (0.0948) (0.0973) (0.0223) (0.0336) (0.0350) (0.0607) (0.101) log distance btw i & j −0.00478** −0.00129 −0.0104* −0.0173** −0.0154* 0.000372 −0.000603 −0.0122*** −0.0221*** −0.0176*** (0.00238) (0.00175) (0.00543) (0.00694) (0.00833) (0.00166) (0.00223) (0.00301) (0.00531) (0.00601) i & j belong to same −0.0323* −0.0561** −0.0218 −0.0567** −0.101*** 0.00460 −0.00129 −0.0133 −0.0955*** −0.0299 sector (0.0171) (0.0260) (0.0141) (0.0280) (0.0345) (0.0101) (0.0134) (0.0185) (0.0302) (0.0310) Abs diff firm age −0.000611 −0.00079*** −0.00100 −0.00187 −0.00322* 0.000121 0.000259 0.000361 −0.00318*** −0.00102 (0.000572) (0.000273) (0.00105) (0.00131) (0.00173) (0.000283) (0.000369) (0.00103) (0.00120) (0.00134) Abs diff managers’ 0.00863 0.0141 −0.00995 0.0644** 0.0475** 0.00567 0.0409*** 0.00756 0.0189 0.0664*** education (0.00906) (0.00956) (0.00833) (0.0286) (0.0226) (0.00616) (0.0144) (0.00927) (0.0167) (0.0235) Abs diff managers’ −0.000763 0.000110 −0.000577 −0.000880 −0.00312* −5.03e−06 −0.000301 −0.000411 −2.72e−05 −0.000124 experience (0.000897) (0.000386) (0.00132) (0.00151) (0.00169) (0.000287) (0.000265) (0.00116) (0.00218) (0.00240) Owners’ genders −0.00147 0.000586 −0.00558 0.0904** 0.00909 0.00113 0.0354* 0.0978** 0.435*** 0.241*** Differ (0.0186) (0.00596) (0.0285) (0.0454) (0.0482) (0.00629) (0.0187) (0.0396) (0.0909) (0.0751) Abs diff log 0.00623 0.0219** 0.0304** 0.189*** 0.121*** 0.0232** 0.0525*** 0.0394** 0.167*** 0.198*** Employment (0.00943) (0.00903) (0.0137) (0.0264) (0.0298) (0.00912) (0.0143) (0.0175) (0.0421) (0.0457) Source: Authors’ computations based on data described in the text. 38 Note: The table shows ordinary least squares results. A constant is included in all specifications. The numbers in parentheses are bootstrapped standard errors that are robust with respect to heteroskedasticity and cross-observation correlation in the error terms involving the same firms. Statistical significance at the 10 percent, 5 percent, and 1 percent levels is indicated by *,**, and ***, respectively. 39 TABLE 4. Correlates of Dyadic Differences: Human Capital and Labor Management Ethiopia Sudan [1] [2] [3] [4] [5] [6] [7] [8] Ratio of Any in-house Staff sent to First principal Ratio of Any in-house Staff sent to First principal nonproduction training of staff formal training component nonproduction training of staff formal training component workers to total last year? course last workers to total last year? course last employment year? employment year? |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| i & j trade with each other −0.0218 0.0807 −0.0451 −0.0514 −0.0859 −0.121 0.368 −0.200 (0.0334) (0.107) (0.133) (0.220) (0.0879) (0.263) (0.340) (0.336) i & j have common supplier 0.0136 0.0439 0.0113 0.0857 −0.000342 0.0313 −0.0195 0.0743 (0.0141) (0.0473) (0.0413) (0.0799) (0.0442) (0.0842) (0.0549) (0.147) i & j have common client 0.0161 0.0396 0.127** 0.215* −0.0403 0.0826** 0.0210 0.174 (0.0237) (0.0742) (0.0627) (0.124) (0.0288) (0.0375) (0.0713) (0.142) log distance btw i & j −0.00193 −0.0168*** −0.0153*** −0.0251*** 0.00320 −0.00803** −0.000397 −0.00612 (0.00159) (0.00299) (0.00291) (0.00824) (0.00195) (0.00315) (0.00434) (0.00847) i & j belong to same sector −0.00244 −0.0173 −0.0304* −0.0480* 0.00212 −0.0341* −0.00869 −0.0644* (0.00491) (0.0123) (0.0177) (0.0262) (0.00623) (0.0189) (0.0159) (0.0374) Abs diff firm age −0.000149 −0.000207 0.00180 0.00225 −0.000522* −0.000423 0.000882 0.000211 (0.000336) (0.000720) (0.00114) (0.00210) (0.000299) (0.000662) (0.000992) (0.00183) Abs diff managers’ education −0.00392* 0.0112 0.00644 0.0328 0.00381 −0.000336 −0.0145*** −0.0148 (0.00237) (0.0184) (0.0160) (0.0351) (0.00339) (0.00576) (0.00474) (0.0131) Abs diff managers’ experience 0.000213 −0.00167** −0.000966 −0.000834 0.000259 −0.00155*** 0.000637 −0.000751 (0.000477) (0.000803) (0.000838) (0.00165) (0.000598) (0.000501) (0.00108) (0.00185) Owners’ gender differ 0.00767 0.0181 −0.0120 −0.00539 0.0298* 0.0446 0.0847 0.204* (0.0111) (0.0258) (0.0177) (0.0403) (0.0176) (0.0375) (0.0542) (0.114) Abs diff log employment 0.0110** 0.0876*** 0.0987*** 0.230*** 0.0182*** 0.0567*** 0.0780*** 0.207*** (0.00490) (0.0155) (0.0143) (0.0309) (0.00569) (0.0170) (0.0205) (0.0499) Source: Authors’ computations based on data described in the text. Note: The table shows ordinary least squares results. A constant is included in all specifications. The numbers in parentheses are bootstrapped standard errors that are robust with respect to heteroskedasticity and cross-observation correlation in the error terms involving the same firms. Statistical significance at the 10 percent, 5 percent, and 1 percent levels is indicated by *,**, and ***, respectively. 40 TABLE 5. Correlates of Dyadic Differences: Contractual Practices Ethiopia Sudan [1] [2] [3] [4] [5] [6] [7] [8] Any direct Do you sell on Does firm First principal Any direct Do you sell on Does firm First principal imports of credit? subcontract component imports of credit? subcontract component inputs? production? inputs? production? |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| i & j trade with each other 0.0245 −0.0122 0.110 0.114 −0.423** −0.436** 0.0488 −0.813** (0.117) (0.105) (0.119) (0.252) (0.174) (0.177) (0.331) (0.380) i & j have common supplier −0.0186 −0.00130 0.0755** −0.116 −0.145* 0.0150 0.0479 0.0304 (0.0501) (0.0404) (0.0366) (0.0751) (0.0762) (0.0789) (0.0812) (0.135) i & j have common client 0.0838 −0.0139 0.0765 0.184 −0.0457 −0.00759 −0.0310 −0.0390 (0.0572) (0.0524) (0.0702) (0.124) (0.0422) (0.0430) (0.0651) (0.0998) log distance btw i & j −0.0123** −0.000254 0.00530 −0.00906 0.00814*** 0.00915** −0.00867** 0.00854 (0.00572) (0.00143) (0.00671) (0.00796) (0.00314) (0.00422) (0.00417) (0.00923) i & j belong to same sector −0.0397** −0.0171 −0.00192 −0.0368 −0.0299* −0.00602 −0.00572 −0.0459 (0.0160) (0.0155) (0.00908) (0.0286) (0.0157) (0.0137) (0.0131) (0.0279) Abs diff firm age −0.000424 0.000480 −0.00207*** −0.00417*** 0.000105 −8.22e−05 −0.000668 0.000303 (0.000804) (0.000556) (0.000636) (0.00145) (0.000240) (0.000582) (0.000563) (0.00107) Abs diff managers’ education 0.0299 0.000668 −0.0204*** 0.0373 0.0215* 0.00583 0.00164 0.0278 (0.0199) (0.00500) (0.00748) (0.0279) (0.0112) (0.00657) (0.00477) (0.0173) Abs diff managers’ experience −0.00152* 0.000485 −0.00178 0.00456** −9.80e−05 −0.000212 −0.000397 −0.000631 (0.000779) (0.000642) (0.00120) (0.00210) (0.000305) (0.000728) (0.000633) (0.00121) Owners’ gender differ 0.0457 0.00391 0.0235 0.0402 0.00257 −0.0164 0.0109 −0.00210 (0.0297) (0.00948) (0.0306) (0.0420) (0.00677) (0.0126) (0.0382) (0.0392) Abs diff log employment 0.131*** 0.00259 0.0150 0.138*** 0.0659*** 0.00494 −0.00398 0.0977*** (0.0150) (0.00473) (0.0132) (0.0312) (0.0145) (0.00781) (0.0123) (0.0264) Source: Authors’ computations based on data described in the text. Note: The table shows ordinary least squares results. A constant is included in all specifications. The numbers in parentheses are bootstrapped standard errors that are robust with respect to heteroskedasticity and cross-observation correlation in the error terms involving the same firms. Statistical significance at the 10 percent, 5 percent, and 1 percent levels is indicated by *,**, and ***, respectively. 41 TABLE 6a. Correlates of Dyadic Differences: Perceived Consequences of Disputes Ethiopia Sudan [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] If you have a If another If you have a If you have a If you have If you have a If another If you have a If you have a If you have customer firm has a customer supplier supplier customer firm has a customer supplier supplier dispute, will customer dispute, will dispute, will dispute, will dispute, will customer dispute, will dispute, will dispute, will other dispute, will other firms other other firms other dispute, will other firms other other firms customers you refuse to refuse to suppliers refuse to customers you refuse to refuse to suppliers refuse to find out? deal with deal with find out? deal with find out? deal with deal with find out? deal with customer? that that that that that customer? supplier? customer? customer? supplier? |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| |yi − yj| i & j trade with each −0.187 0.0330 −0.0502 0.124 0.0158 −0.704* −0.209 0.557 0.280 0.366 other (0.206) (0.178) (0.166) (0.207) (0.161) (0.373) (0.622) (0.515) (0.614) (0.637) i & j have common −0.0159 0.0321 −0.0611 −0.0335 −0.0490 −0.154 −0.0828 −0.0368 −0.105 −0.113 supplier (0.0914) (0.0896) (0.0663) (0.0914) (0.0745) (0.166) (0.164) (0.145) (0.141) (0.164) i & j have common 0.0480 0.0293 −0.0588 0.00750 0.0853 0.0580 0.0328 −0.104 −0.105 0.112** client (0.0795) (0.129) (0.0990) (0.0916) (0.111) (0.0557) (0.0669) (0.116) (0.111) (0.0475) log distance btw i & j 0.0110** −0.0198*** −0.0171*** 0.0104* −0.0268*** −0.00683** −0.00611** −0.000383 −0.00298 −0.00474 (0.00562) (0.00578) (0.00487) (0.00544) (0.00530) (0.00306) (0.00310) (0.00323) (0.00282) (0.00475) i & j belong to same −0.00832 −0.0163 0.00167 −0.0201 −0.0113 −0.00888 −0.0213 −0.00301 −0.00127 0.0275* sector (0.0272) (0.0258) (0.0187) (0.0254) (0.0186) (0.0213) (0.0281) (0.0194) (0.0186) (0.0164) Abs diff firm age −0.000181 −0.000755 −0.00166 0.000648 5.85e−05 −0.000854 0.000383 0.000530 −0.000695 0.00154 (0.000404) (0.00180) (0.00109) (0.000791) (0.00149) (0.000725) (0.00106) (0.000944) (0.000812) (0.00156) Abs diff managers’ −0.00403 −0.00790 −0.00932 −0.00511 −0.00802 0.000897 0.00325 0.00229 −0.00144 0.00723 education (0.00757) (0.0117) (0.0110) (0.00884) (0.0112) (0.00726) (0.00754) (0.00795) (0.00705) (0.00943) Abs diff managers’ 0.000140 −0.00130 0.00309 0.00114 0.00286 0.000672 0.000130 −0.00113 −0.000772 −0.00118 experience (0.000932) (0.00303) (0.00264) (0.00151) (0.00312) (0.000804) (0.000917) (0.000703) (0.000948) (0.00144) Owners’ gender −0.00783 −0.00611 0.0600 0.0184 0.0541 0.0298 0.00190 −0.0103 0.0157 0.0261 differ (0.0131) (0.0558) (0.0550) (0.0299) (0.0597) (0.0332) (0.0341) (0.0252) (0.0387) (0.0461) Abs diff log 0.00230 −0.00417 0.00893 0.0151 0.0102 0.0219 −0.0130 −0.00426 −0.0267*** −0.0311** employment (0.00661) (0.0180) (0.0150) (0.0132) (0.0190) (0.0157) (0.0113) (0.0114) (0.00875) (0.0133) 42 TABLE 6b. Correlates of Dyadic Differences: Perceived Consequences of Disputes, First Principal Component [1] Ethiopia [2] Sudan |yi − yj| |yi − yj| i & j trade with each other −0.0757 0.339 (0.146) (0.782) i & j have common supplier 0.00852 0.0100 (0.0841) (0.151) i & j have common client 0.0573 0.0579 (0.118) (0.0874) log distance btw i & j −0.0152** 0.00614 (0.00734) (0.00704) i & j belong to same sector −0.0130 0.0163 (0.0258) (0.0188) Abs diff firm age −0.000904 −0.00117 (0.00168) (0.000985) Abs diff managers’ education −0.00478 0.00139 (0.0133) (0.00863) Abs diff managers’ experience 0.00290 0.000487 (0.00340) (0.00169) Owners’ gender differ 0.00676 0.0369 (0.0613) (0.0481) Abs diff log employment 0.00317 −0.0111 (0.0164) (0.0136) Source: Authors’ computations based on data described in the text. Note: The table shows ordinary least squares results. A constant is included in all specifications. The numbers in parentheses are bootstrapped standard errors that are robust with respect to heteroskedasticity and cross-observation correlation in the error terms involving the same firms. Statistical significance at the 5 percent level is indicated by **. 43 NOTES 1 The Sudanese data used in this paper were collected in 2007. We began work on this paper in 2009, before South Sudan seceded from Sudan in July 2011. Thus, our sample includes firms in what is now South Sudan. Throughout the text, “Sudanâ€? means Sudan prior to the secession. 2 A network consists of links between the nodes in a finite collection of nodes (for example, firms). See Jackson (2009). 3 For this statement to be true, it is sufficient that Ï? is smaller than one over the maximum degree of any agent (Jackson 2009). 4 To illustrate the point, let N = 2 , α = −1 and Ï? = 2 . If y2 = 0 , then the y1 ≥ 0 constraint is binding and y1 = 0 . If y 2 = 1 then y 2 = −1 + 2 ×1 = 1 . Thus, we have two equilibria: ( y1 , y2 ) = (0, 0) and (1,1). 5 The online appendix can be obtained at http://soderbom.net/Fafchamps_Soderbom_Online_Appendix_2013.pdf and at http://wber.oxfordjournals.org. 6 Heterogeneity across firms has increasingly been recognized in the recent literature; see, for example, Melitz (2003) and Melitz and Ottaviano (2008). 7 There are two reasons to estimate (3) in terms of its absolute deviation and not in a covariance form as in (2). First, most of the outcome variables we investigate are binary; as a result, the only information they contain is whether yi = y j . In this case, (3) boils down to a linear probability model because yi − y j = 1 if yi ≠ y j and 0 otherwise. Second, in the more general case when y is continuous, model (3) captures the main idea behind (2) but offers the advantage of being more 44 robust with respect to outliers compared to using ( yi − y )( y j − y ) as the dependent variable. 8 See Mengistae and Honorati (2009) for details on the survey methodology. For a thorough report on the survey, see the World Bank (2009). 9 See H&H Consultancy (2008) for details on the survey methodology. 10 Maps of the survey locations are shown in section 3 in the online appendix. 11 For some of our outcome variables, there are missing values in the baseline sample. Therefore, some of our regressions will be estimated on a smaller sample than the baseline sample. 12 Fafchamps and Söderbom (2006) argue that the ratio of nonproduction workers to total employment indicates the ease with which firms manage their labor force. They show that many African firms have a high ratio of nonproduction workers to total employment despite the relative simplicity of their production processes. 13 Because the majority of firms (approximately 70 percent) list three names, there is truncation in the observed network because some existing links are not recorded. This problem may cause a downward bias in the estimated network effects. 14 Here, the three levels of usage are combined. The results for the alternative specifications that model email and website use are shown separately in tables S1.E and S1.S, columns [1]–[2], in the online appendix. The results are similar to those shown in table 3. 15 Columns [3]–[8] in tables S.1E and S1.S in the online appendix show the results for alternative specifications in which the reputation variables are defined as binary variables in the following way: yes = 1 and maybe or no = 0. The results are similar to those in table 5. 16 Firms are defined as having a similar size if the absolute log of the difference in employment is less than 0.2. 45 46 Online Appendix: Network Proximity and Business Practices in African Manufacturing Marcel Fafchamps and MÃ¥ns Söderbom Document description: This appendix contains additional material for the study "Network Proximity and Business Practices in African Manufacturingâ€? by Marcel Fafchamps and MÃ¥ns Söderbom. Table of Contents: Section 1: Additional material: Covariance in profits and bootstrapped standard errors Section 2: Names and definitions of variables used in the empirical analysis Section 3: Maps of survey locations Section 4: Dyadic regressions modeling firm performance and growth Section 5: Additional regression results: Extensions and robustness checks Section 1: Additional material: Covariance in profits and bootstrapped standard errors 1.1 Covariance in profits 1.2 Bootstrapped standard errors Dyadic observations are typically not independent since residual u_{ij} is likely to be correlated with u_{ik}.This complicates the computation of standard errors. In particular, robust standard errors must correct for cross-observation correlation in the error terms involving the same enterprises. To obtain consistent standard errors, we use bootstrapping, implemented as follows: (i) Draw a random sample s of n firms from the firm-level dataset with replacement; (ii) For this artificial sample of n firms, construct the corresponding n(n-1) dyadic dependent and independent variables; (iii) Estimate the regression for this sample and store the resulting parameter estimates; (iv) After repeating the process (i)-(iii) J times, use the standard deviations of the estimated parameters as estimates of the standard errors. Section 2: Names and definitions of variables used in the empirical analysis Variable names for dependent variables* Variable name Description Code newprod Did you introduce a new product last year? 0 = no, 1 = yes invdum Did you invest in plant or machinery last year? 0 = no, 1 = yes anyrd Does the firm do any research and development? 0 = no, 1 = yes it IT usage 0 = nothing, 1=email, 2=website pcf_tech First common factor for previous four variables suprat Ratio of non-production workers to total employment Percentage; see note (b) htrain Any in-house training of staff last year? 0 = no, 1 = yes ftrain Staff sent to formal training course last year? 0 = no, 1 = yes pcf_hk First common factor for previous three variables directimport Any direct imports of inputs? 0 = no, 1 = yes sellcredit Do you sell on credit? 0 = no, 1 = yes subcont Does firm sub-contract production? 0 = no, 1 = yes pcf_cont First common factor for previous three variables reput_D1 If you have a dispute with a customer, will other customers 0 = no, 1 = maybe, 2 = yes find out? reputD_2 If another firm has a dispute with a customer, will you refuse 0 = no, 1 = maybe, 2 = yes to deal with that customer? reputD_3 If you have a dispute with a dispute, will other firms refuse to 0 = no, 1 = maybe, 2 = yes deal with that customer? reputD_4 If you have a dispute with a supplier, will other suppliers find 0 = no, 1 = maybe, 2 = yes out? reputD_5 If you have a dispute with a supplier, will other firms refuse to 0 = no, 1 = maybe, 2 = yes deal with that supplier? pcf_repD First common factor for previous five variables lvademp Log( value-added per employee) Log USD dlemp Employment growth last 3 years Dlog growth1 Revenue growth last year Dlog growth3 Revenue growth last 3 years Dlog pcf_perf First common factor for previous four variables web Does the firm use the web? 0 = no, 1 = yes email Does the firm use email? 0 = no, 1 = yes reput_1 If you have a dispute with a customer, will other customers 0 = no, 1 = yes find out? reput_2 If another firm has a dispute with a customer, will you refuse 0 = no, 1 = yes to deal with that customer? reput_3 If you have a dispute with a dispute, will other firms refuse to 0 = no, 1 = yes deal with that customer? reput_4 If you have a dispute with a supplier, will other suppliers find 0 = no, 1 = yes out? reput_5 If you have a dispute with a supplier, will other firms refuse to 0 = no, 1 = yes deal with that supplier? pcf_rep First common factor for previous five variables * Note: In all regressions these variables are defined as absolute differences across any two firms. Variable names and definitions for explanatory variables supbuy i & j trade with each other samesup i & j have a common supplier samecus i & j have a common client same_sec i & j are in the same sector ldist log distance between i & j ad_fmage Absolute difference firm age ad_meduc Absolute difference managers’ education ad_mexper Absolute difference managers’ experience ad_mfemale Absolute difference managers’ genders ad_lemp Absolute difference log employment anylink Dummy = 1 if supbuy=1, samesup=1 or samecus=1 secdist Interaction: same_sec x ldist seclink Interaction: same_sec x anylink sizdist Interaction: similar size x ldist sizlink Interaction: similar size x anylink sametown Dummy = 1 if i and j located in same town Section 3: Maps of Survey Locations Figure S.3.1. Survey locations in Ethiopia Figure S.3.2. Survey locations in Sudan Section 4: Dyadic regressions modeling firm performance and growth Ethiopia Sudan [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] Log( value- Labor Revenue Revenue First Log( value- Labor Revenue Revenue First added per growth last 3 growth last growth last 3 common added per growth last 3 growth last growth last 3 common employee) years year years factor employee) years year years factor |yi-yj| |yi-yj| |yi-yj| |yi-yj| |yi-yj| |yi-yj| |yi-yj| |yi-yj| |yi-yj| |yi-yj| i & j trade with each other -0.0758 -0.0850 -0.0437 0.338 0.165 -0.534** -0.413 1.652 (0.284) (0.130) (0.121) (0.657) (0.424) (0.221) (0.676) (2.483) i & j have common supplier -0.106 0.0152 0.246** 0.00522 0.109 -1.651*** -0.277** -0.904*** -0.711** -0.480** (0.111) (0.0668) (0.113) (0.129) (0.142) (0.453) (0.122) (0.337) (0.350) (0.203) i & j have common client -0.182 -0.0136 0.195 -0.0119 0.113 -0.986 0.0900 -0.586*** 0.307 -0.00893 (0.136) (0.0929) (0.151) (0.202) (0.195) (0.660) (0.139) (0.203) (0.443) (0.216) log Distance btw i & j -0.0247** 0.0139 0.0179* 0.0193 0.0235 -0.0612 -0.0173 -0.0515** -0.0515 -0.0103 (0.0122) (0.0106) (0.0108) (0.0168) (0.0149) (0.0386) (0.0111) (0.0240) (0.0346) (0.0187) i & j belong to same sector -0.0819* 0.00508 -0.00350 -0.0243 -0.0130 -0.135 -0.0555 -0.0465 -0.0624 -0.0756 (0.0447) (0.0247) (0.0185) (0.0515) (0.0349) (0.156) (0.0394) (0.0680) (0.133) (0.0869) Abs diff firm age -0.00314 0.00267 -0.00260*** -0.00564*** -0.00397*** -0.00493 0.00112 0.00862 0.00458 0.00602 (0.00200) (0.00347) (0.000730) (0.00161) (0.00146) (0.00781) (0.00216) (0.00651) (0.00746) (0.00459) Abs diff managers’ education -0.00722 0.00380 -0.00917 0.00885 -0.0164 0.0460 -0.00466 -0.0333 -0.0201 0.0114 (0.0280) (0.0117) (0.0116) (0.0310) (0.0174) (0.0472) (0.0116) (0.0228) (0.0268) (0.0251) Abs diff managers’ experience 0.000671 7.15e-06 -0.00272 -0.00171 -0.00410 -0.00732 0.000766 0.00303 -0.00385 -0.00186 (0.00321) (0.00226) (0.00204) (0.00453) (0.00284) (0.00578) (0.00132) (0.00695) (0.00580) (0.00313) Owners’ gender differ 0.121 0.103 -0.0813** -0.220*** -0.110* 0.375 -0.00723 -0.161 0.0558 -0.138 (0.0932) (0.0669) (0.0389) (0.0686) (0.0582) (0.582) (0.0614) (0.134) (0.275) (0.0960) Abs diff log employment 0.0786** 0.00792 -0.0257** -0.0375 -0.0235 0.00718 0.0654** 0.0906 0.102 0.0611 (0.0311) (0.0242) (0.0117) (0.0360) (0.0223) (0.110) (0.0258) (0.0709) (0.114) (0.0600) Note: The table shows OLS results. A constant is included in all specifications. The numbers in ( ) are bootstrapped standard errors that are robust to heteroskedasticity and cross-observation correlation in the error terms involving the same firms. Statistical significance at the 10%, 5% and 1% level is indicated by *,** and ***, respectively. SECTION 5 Additional regression results: Extensions and robustness checks The regression tables in this Section of the online appendix shows OLS results. In these tables, the numbers in ( ) are bootstrapped standard errors that are robust to heteroskedasticity and cross-observation correlation in the error terms involving the same firms. Statistical significance at the 10%, 5% and 1% level is indicated by *,** and ***, respectively. Tables and Specifications: Part I: Results for Ethiopia Table S1.E: Results for specifications with dummy variables for web, email and perceived consequences of disputes Table S2.E: Results for the baseline specification with interaction terms (same sector x network variables) added Table S3.E: Results for the baseline specification with interaction terms (similar firm size x network variables) added Table S4.E: Results for the baseline specification with a dummy variable for location in the same town added. Table S5.E: Results for an alternative specification in which a single dummy variable for connectedness (anylink) replaces the three trading relationship dummies in the baseline specification Table S6.E: Results for the baseline specification with the geographical distance variable omitted Part II: Results for Sudan Table S1.S: Results for specifications with dummy variables for web, email and perceived consequences of disputes Table S2.S: Results for the baseline specification with interaction terms (same sector x network variables) added Table S3.S: Results for the baseline specification with interaction terms (similar firm size x network variables) added Table S4.S: Results for the baseline specification with a dummy variable for location in the same town added. Table S5.S: Results for an alternative specification in which a single dummy variable for connectedness (anylink) replaces the three trading relationship dummies in the baseline specification Table S6.S: Results for the baseline specification with the geographical distance variable omitted 2 Part I: Results for Ethiopia 3 Table S1.E: Results for specifications with dummy variables for web, email and perceived consequences of disputes (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES Web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0550 -0.125 -0.0650 0.00583 -0.0266 0.0706 -0.0304 -0.107 (0.113) (0.118) (0.121) (0.0965) (0.0935) (0.130) (0.0832) (0.180) samesup -0.0689* 0.00586 -0.0190 0.0244 -0.0518 -0.0270 -0.0229 -0.0649 (0.0392) (0.0524) (0.0484) (0.0464) (0.0434) (0.0486) (0.0447) (0.101) samecus 0.123* -0.0631 0.0385 0.00236 -0.0690 0.000110 0.0355 0.0340 (0.0676) (0.0733) (0.0403) (0.0684) (0.0560) (0.0499) (0.0741) (0.147) same_sec -0.0402* -0.0165 -0.00501 -0.00482 0.00721 -0.00622 0.00769 -0.000615 (0.0205) (0.0146) (0.0141) (0.0122) (0.0103) (0.0131) (0.00907) (0.0247) ldist -0.0251*** 0.00777 0.00192 -0.0107*** -0.0100** -0.000809 -0.0157*** -0.0310*** (0.00299) (0.00656) (0.00201) (0.00316) (0.00469) (0.00224) (0.00432) (0.00890) ad_fmage -0.000781 -0.00109** -6.10e-05 -0.000841 -0.00204*** 0.000619 -0.000481 -0.00268 (0.00110) (0.000534) (0.000213) (0.000912) (0.000638) (0.000539) (0.000841) (0.00177) ad_meduc -0.00391 0.0684*** -0.00364 -0.00498 -0.0127* -0.000439 -0.00838 -0.0173 (0.0138) (0.0200) (0.00223) (0.00639) (0.00771) (0.00543) (0.00636) (0.0139) ad_mexper -0.000455 -0.000425 -0.000157 -0.000943 0.00174 0.000806 0.00176 0.00239 (0.00122) (0.000638) (0.000390) (0.00160) (0.00177) (0.00114) (0.00209) (0.00426) ad_mfemale 0.0465 0.0439** -0.00101 0.00926 0.0263 0.00943 0.0201 0.0209 (0.0332) (0.0202) (0.00737) (0.0316) (0.0361) (0.0183) (0.0357) (0.0729) ad_lemp 0.0766*** 0.113*** 0.00243 -0.00226 0.00724 0.00813 0.00801 0.0109 (0.0194) (0.0150) (0.00408) (0.00951) (0.0103) (0.00889) (0.0114) (0.0214) Constant 0.265*** 0.167*** 0.497*** 0.407*** 0.283*** 0.446*** 0.246*** 1.135*** (0.0404) (0.0420) (0.0133) (0.0401) (0.0425) (0.0281) (0.0431) (0.0996) Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 4 Table S2.E: Results for the baseline specification with interaction terms (same sector x network variables) added VARIABLES (1) (2) (3) (4) (5) (6) (7) (8) newprod invdum anyrd it pcf_tech suprat htrain ftrain supbuy 0.0249 -0.0687 0.0572 -0.0735 0.0295 -0.0283 0.0631 -0.100 (0.136) (0.129) (0.114) (0.199) (0.252) (0.0368) (0.117) (0.138) samesup -0.0493 -0.0307 -0.00481 -0.0674 -0.0857 0.00506 0.0207 -0.0602 (0.0666) (0.0773) (0.0563) (0.0957) (0.115) (0.0201) (0.0654) (0.0671) samecus 0.0460 -0.0417 -0.0198 0.0581 -0.105 0.0109 0.0253 0.0795 (0.0666) (0.0769) (0.0739) (0.106) (0.108) (0.0233) (0.0807) (0.0725) same_sec -0.0452* -0.0402 -0.0229 -0.0522 -0.109** -0.00294 -0.0198 -0.0690** (0.0273) (0.0297) (0.0212) (0.0340) (0.0490) (0.00822) (0.0154) (0.0291) ldist -0.00553* -0.000262 -0.0104* -0.0170** -0.0158* -0.00195 -0.0169*** -0.0176*** (0.00288) (0.00201) (0.00563) (0.00698) (0.00926) (0.00157) (0.00309) (0.00274) ad_fmage -0.000611 -0.000780*** -0.001000 -0.00187 -0.00322* -0.000149 -0.000206 0.00180 (0.000574) (0.000272) (0.00105) (0.00131) (0.00173) (0.000335) (0.000720) (0.00114) ad_meduc 0.00864 0.0141 -0.00997 0.0644** 0.0475** -0.00392* 0.0112 0.00648 (0.00906) (0.00955) (0.00833) (0.0286) (0.0226) (0.00237) (0.0184) (0.0160) ad_mexper -0.000761 0.000113 -0.000573 -0.000880 -0.00312* 0.000214 -0.00167** -0.000962 (0.000896) (0.000386) (0.00132) (0.00151) (0.00169) (0.000478) (0.000802) (0.000839) ad_mfemale -0.00138 0.000628 -0.00547 0.0904** 0.00925 0.00769 0.0182 -0.0118 (0.0185) (0.00597) (0.0285) (0.0455) (0.0482) (0.0111) (0.0258) (0.0178) ad_lemp 0.00630 0.0218** 0.0304** 0.189*** 0.121*** 0.0110** 0.0876*** 0.0989*** (0.00943) (0.00903) (0.0137) (0.0264) (0.0298) (0.00490) (0.0155) (0.0143) secdist 0.00270 -0.00365 1.18e-05 -0.00101 0.00150 6.95e-05 0.000434 0.00819** (0.00473) (0.00472) (0.00335) (0.00484) (0.00723) (0.00140) (0.00291) (0.00403) seclink 0.0459 0.0349 0.0684 0.00469 0.0899 0.0116 0.0317 0.106 (0.0721) (0.0839) (0.0560) (0.116) (0.128) (0.0185) (0.0767) (0.0700) Constant 0.486*** 0.471*** 0.264*** 0.430*** 1.007*** 0.172*** 0.323*** 0.285*** (0.0190) (0.0220) (0.0424) (0.0594) (0.0721) (0.0135) (0.0362) (0.0363) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 5 Table S2.E continued VARIABLES (9) (10) (11) (12) (13) (14) (15) (16) pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 supbuy -0.152 0.0491 0.0313 0.0767 0.110 -0.0726 0.0131 -0.0891 (0.237) (0.124) (0.107) (0.116) (0.258) (0.209) (0.207) (0.173) samesup -0.0459 0.0147 0.0554 0.0322 -0.121 0.137 0.00529 -0.113 (0.115) (0.0556) (0.0588) (0.0594) (0.122) (0.131) (0.114) (0.0867) samecus 0.132 0.102 0.0231 0.0498 0.179 0.137 0.0147 -0.0886 (0.131) (0.0628) (0.0566) (0.0726) (0.145) (0.0989) (0.139) (0.102) same_sec -0.0846** -0.0578* 0.00718 -0.00606 -0.0531 -0.0331 -0.00254 0.0149 (0.0362) (0.0297) (0.0134) (0.0173) (0.0415) (0.0473) (0.0363) (0.0311) ldist -0.0272*** -0.0134** 0.00116 0.00511 -0.0101 0.00928 -0.0189*** -0.0162*** (0.00821) (0.00593) (0.00178) (0.00699) (0.00825) (0.00624) (0.00637) (0.00545) ad_fmage 0.00225 -0.000429 0.000481 -0.00206*** -0.00417*** -0.000194 -0.000751 -0.00165 (0.00210) (0.000803) (0.000557) (0.000636) (0.00145) (0.000405) (0.00181) (0.00109) ad_meduc 0.0328 0.0300 0.000648 -0.0205*** 0.0373 -0.00390 -0.00794 -0.00937 (0.0351) (0.0199) (0.00499) (0.00747) (0.0279) (0.00756) (0.0117) (0.0110) ad_mexper -0.000825 -0.00152* 0.000481 -0.00177 0.00456** 0.000127 -0.00130 0.00310 (0.00165) (0.000778) (0.000642) (0.00120) (0.00210) (0.000931) (0.00303) (0.00264) ad_mfemale -0.00505 0.0457 0.00375 0.0236 0.0402 -0.00813 -0.00607 0.0601 (0.0404) (0.0297) (0.00949) (0.0306) (0.0420) (0.0130) (0.0558) (0.0550) ad_lemp 0.230*** 0.132*** 0.00247 0.0151 0.138*** 0.00234 -0.00421 0.00890 (0.0309) (0.0150) (0.00472) (0.0132) (0.0312) (0.00657) (0.0180) (0.0150) secdist 0.00744 0.00414 -0.00510 0.000705 0.00357 0.00620 -0.00316 -0.00318 (0.00607) (0.00453) (0.00314) (0.00335) (0.00692) (0.00736) (0.00650) (0.00552) seclink 0.186 -0.0398 -0.0823 0.0592 0.0105 -0.198 0.0323 0.0660 (0.121) (0.0623) (0.0629) (0.0639) (0.127) (0.135) (0.137) (0.103) Constant 0.709*** 0.222*** 0.479*** 0.227*** 0.862*** 0.956*** 0.836*** 0.722*** (0.0722) (0.0437) (0.0200) (0.0457) (0.0627) (0.0354) (0.0778) (0.0575) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 6 Table S2.E continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy 0.120 0.0127 -0.122 -0.113 -0.144 -0.181 0.229 -0.00534 (0.218) (0.177) (0.156) (0.302) (0.135) (0.162) (0.632) (0.437) samesup -0.0396 -0.0535 -0.0543 -0.154 -0.0628 0.0612 -0.140 -0.119 (0.109) (0.124) (0.112) (0.155) (0.0912) (0.122) (0.238) (0.203) samecus 0.00614 0.0838 0.0223 -0.221 -0.0598 0.0908 -0.0925 -0.0228 (0.105) (0.118) (0.125) (0.160) (0.113) (0.146) (0.206) (0.179) same_sec -0.000247 -0.00125 0.0131 -0.153* 0.0164 0.0199 0.0360 0.0258 (0.0289) (0.0255) (0.0398) (0.0793) (0.0337) (0.0251) (0.0798) (0.0637) ldist 0.0117** -0.0262*** -0.0135* -0.0290** 0.0147 0.0196* 0.0233 0.0263* (0.00568) (0.00539) (0.00787) (0.0139) (0.0106) (0.0113) (0.0163) (0.0149) ad_fmage 0.000652 6.06e-05 -0.000896 -0.00314 0.00268 -0.00258*** -0.00563*** -0.00396*** (0.000792) (0.00149) (0.00168) (0.00201) (0.00347) (0.000727) (0.00161) (0.00145) ad_meduc -0.00515 -0.00804 -0.00486 -0.00711 0.00370 -0.00932 0.00860 -0.0167 (0.00883) (0.0112) (0.0133) (0.0280) (0.0117) (0.0116) (0.0311) (0.0174) ad_mexper 0.00114 0.00286 0.00291 0.000676 1.61e-05 -0.00271 -0.00168 -0.00409 (0.00151) (0.00312) (0.00340) (0.00321) (0.00226) (0.00204) (0.00453) (0.00283) ad_mfemale 0.0184 0.0541 0.00686 0.122 0.103 -0.0810** -0.220*** -0.109* (0.0299) (0.0597) (0.0613) (0.0932) (0.0670) (0.0389) (0.0687) (0.0582) ad_lemp 0.0151 0.0102 0.00309 0.0789** 0.00791 -0.0257** -0.0377 -0.0235 (0.0133) (0.0190) (0.0164) (0.0310) (0.0242) (0.0118) (0.0361) (0.0223) secdist -0.00441 -0.00225 -0.00607 0.0153 -0.00289 -0.00608 -0.0142 -0.00992 (0.00539) (0.00537) (0.00730) (0.0140) (0.00425) (0.00623) (0.0158) (0.0141) seclink 0.00281 0.00342 0.0772 0.0856 0.103 0.239 0.182 0.309 (0.129) (0.140) (0.129) (0.153) (0.104) (0.162) (0.278) (0.272) Constant 0.880*** 0.669*** 1.114*** 1.434*** 0.373*** 0.618*** 1.277*** 1.085*** (0.0528) (0.0639) (0.0828) (0.126) (0.0738) (0.0746) (0.173) (0.128) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 7 Table S2.E continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0754 -0.149 -0.0249 -0.00735 -0.0531 0.0488 -0.0320 -0.128 (0.118) (0.127) (0.127) (0.115) (0.100) (0.136) (0.0982) (0.202) samesup -0.0418 -0.0256 0.0347 0.00691 -0.0872* -0.0561 -0.0248 -0.0936 (0.0695) (0.0685) (0.0716) (0.0636) (0.0445) (0.0577) (0.0770) (0.131) samecus 0.139* -0.0812 0.0689 -0.00805 -0.0890 -0.0168 0.0341 0.0173 (0.0771) (0.0786) (0.0530) (0.0706) (0.0570) (0.0602) (0.0807) (0.157) same_sec -0.0432 -0.00901 -0.0229 -0.00346 0.0196 -0.000503 0.00499 0.00575 (0.0286) (0.0191) (0.0269) (0.0190) (0.0168) (0.0181) (0.0177) (0.0400) ldist -0.0253*** 0.00828 0.000736 -0.0106*** -0.00919* -0.000414 -0.0158*** -0.0306*** (0.00306) (0.00681) (0.00238) (0.00353) (0.00492) (0.00262) (0.00454) (0.00964) ad_fmage -0.000783 -0.00109** -6.72e-05 -0.000840 -0.00204*** 0.000622 -0.000481 -0.00268 (0.00110) (0.000533) (0.000214) (0.000913) (0.000637) (0.000540) (0.000841) (0.00177) ad_meduc -0.00389 0.0683*** -0.00358 -0.00499 -0.0128* -0.000465 -0.00838 -0.0173 (0.0138) (0.0200) (0.00224) (0.00638) (0.00772) (0.00544) (0.00635) (0.0139) ad_mexper -0.000458 -0.000422 -0.000162 -0.000941 0.00174 0.000808 0.00176 0.00239 (0.00122) (0.000639) (0.000389) (0.00160) (0.00177) (0.00114) (0.00209) (0.00426) ad_mfemale 0.0465 0.0439** -0.00111 0.00930 0.0263 0.00949 0.0201 0.0209 (0.0332) (0.0203) (0.00735) (0.0316) (0.0361) (0.0183) (0.0357) (0.0728) ad_lemp 0.0766*** 0.113*** 0.00249 -0.00226 0.00720 0.00812 0.00802 0.0109 (0.0194) (0.0150) (0.00405) (0.00950) (0.0104) (0.00891) (0.0114) (0.0214) secdist 0.000798 -0.00181 0.00421 -0.000384 -0.00289 -0.00140 0.000588 -0.00154 (0.00389) (0.00312) (0.00425) (0.00361) (0.00306) (0.00294) (0.00360) (0.00824) seclink -0.0355 0.0402 -0.0672 0.0231 0.0442 0.0375 0.00327 0.0368 (0.0852) (0.0762) (0.0742) (0.0747) (0.0564) (0.0701) (0.0909) (0.165) Constant 0.266*** 0.164*** 0.502*** 0.407*** 0.280*** 0.444*** 0.247*** 1.133*** (0.0399) (0.0426) (0.0148) (0.0406) (0.0421) (0.0280) (0.0433) (0.0997) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 8 Table S3.E: Results for the baseline specification with interaction terms (similar firm size x network variables) added (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES newprod invdum anyrd it pcf_tech suprat htrain ftrain supbuy 0.0486 -0.0468 0.0945 -0.0731 0.0770 -0.0218 0.0795 -0.0455 (0.126) (0.117) (0.108) (0.187) (0.240) (0.0333) (0.107) (0.134) samesup -0.0199 0.000180 0.0346 -0.0869 -0.0424 0.0150 0.0342 0.00815 (0.0409) (0.0464) (0.0413) (0.0672) (0.0671) (0.0150) (0.0451) (0.0455) samecus 0.0653 -0.0249 0.00500 0.0476 -0.0762 0.0168 0.0344 0.125** (0.0588) (0.0704) (0.0696) (0.0943) (0.0955) (0.0238) (0.0748) (0.0622) same_sec -0.0311* -0.0566** -0.0213 -0.0545* -0.0997*** -0.00225 -0.0164 -0.0301* (0.0170) (0.0259) (0.0140) (0.0279) (0.0345) (0.00484) (0.0124) (0.0177) ldist -0.00347 -0.00188 -0.00988* -0.0148** -0.0143* -0.00174 -0.0157*** -0.0149*** (0.00245) (0.00179) (0.00543) (0.00705) (0.00845) (0.00158) (0.00301) (0.00294) ad_fmage -0.000569 -0.000803*** -0.000984 -0.00179 -0.00319* -0.000143 -0.000171 0.00182 (0.000570) (0.000277) (0.00105) (0.00130) (0.00174) (0.000335) (0.000716) (0.00113) ad_meduc 0.00855 0.0142 -0.0100 0.0643** 0.0474** -0.00393* 0.0111 0.00641 (0.00907) (0.00958) (0.00830) (0.0285) (0.0226) (0.00237) (0.0184) (0.0160) ad_mexper -0.000794 0.000124 -0.000590 -0.000939 -0.00315* 0.000208 -0.00170** -0.000976 (0.000889) (0.000387) (0.00131) (0.00150) (0.00170) (0.000479) (0.000799) (0.000832) ad_mfemale -0.00325 0.00138 -0.00627 0.0871* 0.00782 0.00740 0.0166 -0.0125 (0.0184) (0.00605) (0.0284) (0.0450) (0.0479) (0.0112) (0.0256) (0.0178) ad_lemp 0.00199 0.0238** 0.0287* 0.182*** 0.118*** 0.0104** 0.0840*** 0.0975*** (0.0103) (0.00934) (0.0148) (0.0281) (0.0323) (0.00521) (0.0163) (0.0152) sizdist -0.0115* 0.00516 -0.00483 -0.0218** -0.00896 -0.00169 -0.00983* -0.00352 (0.00616) (0.00363) (0.00481) (0.00857) (0.0108) (0.00170) (0.00516) (0.00539) sizlink 0.000278 -0.00512 0.0801 0.149 0.156 -0.0129 0.0593 0.0190 (0.0934) (0.0844) (0.0965) (0.121) (0.150) (0.0262) (0.101) (0.0925) Constant 0.490*** 0.473*** 0.267*** 0.444*** 1.010*** 0.173*** 0.328*** 0.277*** (0.0195) (0.0222) (0.0434) (0.0617) (0.0742) (0.0135) (0.0371) (0.0375) Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 9 Table S3.E continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 supbuy -0.0539 0.0234 -0.0116 0.110 0.113 -0.187 0.0314 -0.0527 (0.221) (0.117) (0.105) (0.119) (0.253) (0.206) (0.178) (0.166) samesup 0.0625 -0.0281 0.00710 0.0795** -0.128* -0.0101 0.0224 -0.0831 (0.0845) (0.0514) (0.0430) (0.0396) (0.0776) (0.0915) (0.0877) (0.0671) samecus 0.203 0.0788 -0.00951 0.0786 0.178 0.0510 0.0241 -0.0704 (0.124) (0.0575) (0.0534) (0.0706) (0.125) (0.0807) (0.129) (0.101) same_sec -0.0467* -0.0391** -0.0169 -0.00163 -0.0365 -0.00872 -0.0140 0.00337 (0.0260) (0.0160) (0.0154) (0.00901) (0.0285) (0.0273) (0.0254) (0.0183) Ldist -0.0235*** -0.0115** -9.70e-05 0.00559 -0.00865 0.0106* -0.0173*** -0.0151*** (0.00836) (0.00586) (0.00149) (0.00666) (0.00815) (0.00569) (0.00559) (0.00468) ad_fmage 0.00230 -0.000401 0.000485 -0.00206*** -0.00416*** -0.000196 -0.000677 -0.00159 (0.00209) (0.000800) (0.000554) (0.000637) (0.00144) (0.000401) (0.00181) (0.00109) ad_meduc 0.0326 0.0299 0.000673 -0.0205*** 0.0372 -0.00399 -0.00806 -0.00947 (0.0351) (0.0199) (0.00498) (0.00749) (0.0279) (0.00757) (0.0117) (0.0110) ad_mexper -0.000872 -0.00154** 0.000481 -0.00178 0.00456** 0.000151 -0.00136 0.00305 (0.00165) (0.000779) (0.000648) (0.00120) (0.00210) (0.000934) (0.00301) (0.00262) ad_mfemale -0.00747 0.0448 0.00364 0.0231 0.0397 -0.00722 -0.00939 0.0575 (0.0402) (0.0297) (0.00972) (0.0305) (0.0422) (0.0131) (0.0550) (0.0547) ad_lemp 0.224*** 0.129*** 0.00198 0.0141 0.137*** 0.00377 -0.0120 0.00277 (0.0328) (0.0162) (0.00453) (0.0140) (0.0316) (0.00689) (0.0182) (0.0153) sizdist -0.0142 -0.00622 -0.00144 -0.00252 -0.00342 0.00411 -0.0213* -0.0172 (0.0103) (0.00555) (0.00432) (0.00496) (0.00931) (0.00468) (0.0116) (0.0109) sizlink 0.155 0.0634 -0.0661 -0.0341 0.0826 -0.0378 0.0416 0.142 (0.172) (0.0791) (0.110) (0.0773) (0.155) (0.180) (0.196) (0.136) Constant 0.707*** 0.221*** 0.486*** 0.228*** 0.859*** 0.946*** 0.852*** 0.736*** (0.0728) (0.0451) (0.0207) (0.0463) (0.0636) (0.0327) (0.0767) (0.0571) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 10 Table S3.E continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy 0.125 0.0163 -0.0765 -0.0814 -0.0843 -0.0435 0.335 0.159 (0.207) (0.159) (0.148) (0.283) (0.130) (0.122) (0.656) (0.424) samesup -0.0223 -0.0342 0.00905 -0.0898 0.0246 0.259** -0.0282 0.123 (0.0921) (0.0760) (0.0823) (0.114) (0.0699) (0.114) (0.131) (0.145) samecus 0.0134 0.0930 0.0574 -0.175 -0.00862 0.200 -0.0299 0.118 (0.0939) (0.112) (0.117) (0.136) (0.0939) (0.149) (0.203) (0.194) same_sec -0.0199 -0.00913 -0.0106 -0.0791* 0.00583 -0.000574 -0.0220 -0.0109 (0.0254) (0.0184) (0.0253) (0.0448) (0.0245) (0.0180) (0.0507) (0.0342) ldist 0.0105* -0.0246*** -0.0125 -0.0218* 0.0146 0.0214* 0.0231 0.0267* (0.00548) (0.00540) (0.00785) (0.0118) (0.0114) (0.0109) (0.0173) (0.0154) ad_fmage 0.000651 0.000131 -0.000819 -0.00304 0.00270 -0.00248*** -0.00553*** -0.00387*** (0.000795) (0.00149) (0.00168) (0.00199) (0.00349) (0.000720) (0.00161) (0.00144) ad_meduc -0.00510 -0.00813 -0.00494 -0.00722 0.00377 -0.00944 0.00855 -0.0164 (0.00883) (0.0112) (0.0133) (0.0280) (0.0117) (0.0115) (0.0312) (0.0174) ad_mexper 0.00114 0.00281 0.00284 0.000590 -9.34e-06 -0.00281 -0.00181 -0.00421 (0.00150) (0.00311) (0.00338) (0.00322) (0.00228) (0.00204) (0.00453) (0.00285) ad_mfemale 0.0182 0.0509 0.00313 0.118 0.102 -0.0855** -0.224*** -0.113* (0.0297) (0.0594) (0.0608) (0.0933) (0.0676) (0.0389) (0.0691) (0.0585) ad_lemp 0.0147 0.00275 -0.00544 0.0686** 0.00557 -0.0369*** -0.0492 -0.0340 (0.0128) (0.0203) (0.0170) (0.0335) (0.0270) (0.0135) (0.0371) (0.0253) sizdist -0.000854 -0.0197** -0.0232* -0.0256** -0.00605 -0.0290*** -0.0321 -0.0276* (0.00575) (0.00950) (0.0121) (0.0130) (0.00997) (0.00875) (0.0214) (0.0158) sizlink -0.0867 -0.142 -0.0390 -0.148 -0.0786 -0.135 0.197 -0.134 (0.179) (0.167) (0.188) (0.193) (0.124) (0.185) (0.338) (0.232) Constant 0.887*** 0.683*** 1.135*** 1.431*** 0.380*** 0.643*** 1.313*** 1.114*** (0.0516) (0.0635) (0.0801) (0.125) (0.0736) (0.0797) (0.183) (0.138) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 11 Table S3.E continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0539 -0.127 -0.0647 0.00506 -0.0270 0.0714 -0.0301 -0.107 (0.113) (0.117) (0.121) (0.0965) (0.0941) (0.130) (0.0835) (0.183) samesup -0.0787* -0.00827 -0.0169 0.0203 -0.0559 -0.0175 -0.0165 -0.0625 (0.0410) (0.0541) (0.0488) (0.0460) (0.0418) (0.0498) (0.0444) (0.0980) samecus 0.118* -0.0706 0.0396 0.000112 -0.0711 0.00507 0.0389 0.0351 (0.0670) (0.0731) (0.0409) (0.0685) (0.0574) (0.0507) (0.0745) (0.148) same_sec -0.0394* -0.0152 -0.00524 -0.00364 0.00734 -0.00610 0.00835 0.000865 (0.0205) (0.0145) (0.0141) (0.0120) (0.0102) (0.0131) (0.00888) (0.0242) ldist -0.0241*** 0.00934 0.00166 -0.00945*** -0.00983** -0.000753 -0.0150*** -0.0295*** (0.00291) (0.00671) (0.00203) (0.00308) (0.00466) (0.00218) (0.00441) (0.00943) ad_fmage -0.000751 -0.00104** -6.92e-05 -0.000800 -0.00204*** 0.000621 -0.000460 -0.00263 (0.00110) (0.000523) (0.000210) (0.000916) (0.000639) (0.000540) (0.000844) (0.00178) ad_meduc -0.00398 0.0682*** -0.00362 -0.00506 -0.0127* -0.000426 -0.00841 -0.0174 (0.0138) (0.0199) (0.00223) (0.00639) (0.00770) (0.00543) (0.00635) (0.0139) ad_mexper -0.000477 -0.000462 -0.000151 -0.000973 0.00173 0.000804 0.00174 0.00235 (0.00122) (0.000626) (0.000390) (0.00159) (0.00176) (0.00114) (0.00208) (0.00424) ad_mfemale 0.0453 0.0418** -0.000679 0.00752 0.0261 0.00929 0.0192 0.0187 (0.0330) (0.0199) (0.00740) (0.0312) (0.0359) (0.0183) (0.0355) (0.0725) ad_lemp 0.0736*** 0.108*** 0.00324 -0.00640 0.00672 0.00783 0.00581 0.00580 (0.0206) (0.0154) (0.00419) (0.00980) (0.0110) (0.00893) (0.0125) (0.0228) sizdist -0.00825* -0.0136** 0.00222 -0.0113* -0.00150 -0.000566 -0.00575 -0.0137 (0.00498) (0.00541) (0.00236) (0.00607) (0.00556) (0.00315) (0.00544) (0.0133) sizlink 0.0618 0.0871 -0.0124 0.0146 0.0293 -0.0730 -0.0575 -0.0383 (0.0728) (0.0923) (0.0945) (0.101) (0.0770) (0.0916) (0.0792) (0.197) Constant 0.270*** 0.174*** 0.495*** 0.414*** 0.284*** 0.446*** 0.250*** 1.143*** (0.0423) (0.0419) (0.0134) (0.0397) (0.0424) (0.0278) (0.0432) (0.0981) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 12 Table S4.E: Results for the baseline specification with a dummy variable for location in the same town added. (1) (2) (3) (4) (5) VARIABLES newprod invdum anyrd it pcf_tech supbuy 0.0515 -0.0460 0.103 -0.0751 0.0903 (0.128) (0.117) (0.108) (0.192) (0.239) samesup -0.0169 -0.00119 0.0481 -0.0644 -0.0170 (0.0401) (0.0457) (0.0404) (0.0662) (0.0685) samecus 0.0657 -0.0263 0.00823 0.0621 -0.0689 (0.0585) (0.0707) (0.0691) (0.0946) (0.0964) same_sec -0.0322* -0.0560** -0.0213 -0.0570** -0.0998*** (0.0171) (0.0260) (0.0141) (0.0282) (0.0345) Ldist 0.00649 0.00428 0.0242 -0.0411*** 0.0393 (0.0113) (0.00638) (0.0149) (0.0132) (0.0248) ad_fmage -0.000612 -0.000785*** -0.00100 -0.00187 -0.00322* (0.000572) (0.000273) (0.00105) (0.00131) (0.00172) ad_meduc 0.00905 0.0143 -0.00866 0.0636** 0.0495** (0.00917) (0.00966) (0.00823) (0.0286) (0.0230) ad_mexper -0.000757 0.000113 -0.000558 -0.000894 -0.00309* (0.000895) (0.000390) (0.00131) (0.00150) (0.00169) ad_mfemale -0.00129 0.000675 -0.00503 0.0900** 0.00996 (0.0186) (0.00596) (0.0284) (0.0453) (0.0480) ad_lemp 0.00665 0.0221** 0.0316** 0.189*** 0.123*** (0.00941) (0.00905) (0.0138) (0.0265) (0.0299) sametown 0.0697 0.0344 0.214*** -0.147* 0.338*** (0.0628) (0.0384) (0.0687) (0.0874) (0.129) Constant 0.416*** 0.443*** 0.0594 0.572*** 0.682*** (0.0659) (0.0460) (0.0821) (0.101) (0.152) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 13 Table S4.E continued (6) (7) (8) (9) VARIABLES suprat htrain ftrain pcf_hk supbuy -0.0227 0.0811 -0.0429 -0.0489 (0.0340) (0.107) (0.133) (0.221) samesup 0.0134 0.0440 0.0120 0.0864 (0.0141) (0.0473) (0.0414) (0.0801) samecus 0.0165 0.0394 0.126** 0.214* (0.0237) (0.0743) (0.0631) (0.124) same_sec -0.00250 -0.0173 -0.0303* -0.0479* (0.00490) (0.0123) (0.0177) (0.0263) ldist -0.00638 -0.0147 -0.00460 -0.0129 (0.00486) (0.0103) (0.0116) (0.0215) ad_fmage -0.000149 -0.000207 0.00180 0.00225 (0.000335) (0.000721) (0.00114) (0.00210) ad_meduc -0.00409* 0.0113 0.00684 0.0332 (0.00240) (0.0184) (0.0161) (0.0352) ad_mexper 0.000210 -0.00167** -0.000960 -0.000827 (0.000476) (0.000804) (0.000843) (0.00164) ad_mfemale 0.00760 0.0182 -0.0118 -0.00519 (0.0112) (0.0258) (0.0177) (0.0401) ad_lemp 0.0108** 0.0877*** 0.0991*** 0.230*** (0.00494) (0.0155) (0.0142) (0.0308) sametown -0.0275 0.0131 0.0664 0.0752 (0.0304) (0.0610) (0.0659) (0.130) Constant 0.198*** 0.310*** 0.211*** 0.627*** (0.0323) (0.0703) (0.0708) (0.133) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 14 Table S4.E continued (10) (11) (12) (13) VARIABLES directimport sellcredit subcont pcf_cont supbuy 0.0205 -0.0121 0.121 0.114 (0.116) (0.104) (0.120) (0.252) samesup -0.0197 -0.00127 0.0787** -0.116 (0.0502) (0.0403) (0.0367) (0.0746) samecus 0.0854 -0.0139 0.0720 0.184 (0.0566) (0.0524) (0.0691) (0.124) same_sec -0.0400** -0.0171 -0.00111 -0.0367 (0.0160) (0.0155) (0.00881) (0.0284) ldist -0.0318*** 0.000235 0.0613*** -0.00688 (0.0100) (0.00439) (0.0178) (0.0280) ad_fmage -0.000423 0.000479 -0.00207*** -0.00417*** (0.000805) (0.000556) (0.000637) (0.00145) ad_meduc 0.0292 0.000686 -0.0183** 0.0374 (0.0199) (0.00500) (0.00710) (0.0281) ad_mexper -0.00153** 0.000485 -0.00174 0.00457** (0.000780) (0.000645) (0.00120) (0.00209) ad_mfemale 0.0454 0.00392 0.0245 0.0402 (0.0297) (0.00950) (0.0305) (0.0420) ad_lemp 0.131*** 0.00261 0.0171 0.138*** (0.0150) (0.00473) (0.0133) (0.0313) sametown -0.121 0.00302 0.347*** 0.0135 (0.0787) (0.0270) (0.0787) (0.179) Constant 0.333*** 0.482*** -0.105 0.845*** (0.0753) (0.0337) (0.0835) (0.175) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 15 Table S4.E continued (14) (15) (16) (17) (18) VARIABLES reput_1 reput_2 reput_3 reput_4 reput_5 supbuy -0.186 0.0255 -0.0587 0.120 0.00681 (0.205) (0.179) (0.172) (0.206) (0.158) samesup -0.0154 0.0299 -0.0635 -0.0347 -0.0516 (0.0912) (0.0897) (0.0670) (0.0909) (0.0753) samecus 0.0473 0.0323 -0.0554 0.00917 0.0889 (0.0795) (0.129) (0.0985) (0.0918) (0.111) same_sec -0.00822 -0.0167 0.00115 -0.0204 -0.0119 (0.0272) (0.0256) (0.0186) (0.0255) (0.0184) ldist 0.0192 -0.0567** -0.0586** -0.0101 -0.0713*** (0.0198) (0.0246) (0.0242) (0.00907) (0.0179) ad_fmage -0.000181 -0.000753 -0.00165 0.000649 6.14e-05 (0.000403) (0.00180) (0.00109) (0.000792) (0.00149) ad_meduc -0.00372 -0.00927 -0.0109 -0.00587 -0.00967 (0.00762) (0.0116) (0.0112) (0.00881) (0.0114) ad_mexper 0.000145 -0.00133 0.00307 0.00113 0.00284 (0.000939) (0.00304) (0.00264) (0.00151) (0.00311) ad_mfemale -0.00770 -0.00670 0.0594 0.0180 0.0534 (0.0130) (0.0559) (0.0548) (0.0298) (0.0598) ad_lemp 0.00260 -0.00552 0.00741 0.0144 0.00859 (0.00647) (0.0178) (0.0147) (0.0132) (0.0189) sametown 0.0508 -0.229 -0.257* -0.127* -0.275*** (0.105) (0.148) (0.140) (0.0699) (0.106) Constant 0.900*** 1.058*** 0.971*** 1.007*** 0.934*** (0.114) (0.163) (0.140) (0.0652) (0.120) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 16 Table S4.E continued (19) (20) (21) (22) (23) (24) VARIABLES pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy -0.0811 -0.0781 -0.0726 -0.0254 0.295 0.170 (0.149) (0.286) (0.131) (0.116) (0.600) (0.415) samesup 0.00697 -0.107 0.0169 0.251** -0.000477 0.110 (0.0845) (0.110) (0.0665) (0.115) (0.130) (0.144) samecus 0.0594 -0.181 -0.0163 0.190 -0.00429 0.112 (0.118) (0.137) (0.0915) (0.148) (0.204) (0.192) same_sec -0.0134 -0.0820* 0.00514 -0.00266 -0.0246 -0.0129 (0.0259) (0.0446) (0.0246) (0.0182) (0.0515) (0.0350) ldist -0.0419* -0.0344 0.0581*** 0.0867** -0.122 0.0359 (0.0236) (0.0412) (0.0208) (0.0378) (0.105) (0.0661) ad_fmage -0.000902 -0.00314 0.00267 -0.00259*** -0.00563*** -0.00397*** (0.00167) (0.00200) (0.00347) (0.000726) (0.00158) (0.00145) ad_meduc -0.00577 -0.00753 0.00578 -0.00652 0.00239 -0.0159 (0.0133) (0.0277) (0.0114) (0.0112) (0.0327) (0.0175) ad_mexper 0.00289 0.000665 1.22e-05 -0.00271 -0.00175 -0.00409 (0.00341) (0.00321) (0.00225) (0.00204) (0.00449) (0.00284) ad_mfemale 0.00633 0.121 0.104 -0.0790** -0.223*** -0.109* (0.0613) (0.0935) (0.0671) (0.0388) (0.0694) (0.0580) ad_lemp 0.00219 0.0783** 0.00920 -0.0238** -0.0415 -0.0232 (0.0163) (0.0315) (0.0242) (0.0116) (0.0341) (0.0217) sametown -0.165 -0.0597 0.272*** 0.424** -0.870 0.0765 (0.138) (0.234) (0.0926) (0.188) (0.651) (0.368) Constant 1.279*** 1.472*** 0.115 0.219 2.129*** 1.023*** (0.158) (0.277) (0.119) (0.153) (0.654) (0.343) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 17 Table S4.E continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0528 -0.128 -0.0658 0.000924 -0.0319 0.0701 -0.0345 -0.115 (0.114) (0.119) (0.121) (0.0973) (0.0970) (0.130) (0.0826) (0.183) samesup -0.0695* 0.00511 -0.0192 0.0230 -0.0533 -0.0271 -0.0240 -0.0672 (0.0392) (0.0523) (0.0482) (0.0465) (0.0441) (0.0486) (0.0454) (0.102) samecus 0.124* -0.0620 0.0388 0.00431 -0.0669 0.000324 0.0372 0.0373 (0.0676) (0.0737) (0.0405) (0.0687) (0.0557) (0.0499) (0.0739) (0.147) same_sec -0.0403** -0.0167 -0.00506 -0.00512 0.00688 -0.00625 0.00744 -0.00113 (0.0206) (0.0147) (0.0141) (0.0121) (0.0102) (0.0131) (0.00895) (0.0247) ldist -0.0360** -0.00510 -0.00200 -0.0349*** -0.0361* -0.00344 -0.0358*** -0.0719** (0.0143) (0.0111) (0.00861) (0.0132) (0.0206) (0.00791) (0.0136) (0.0331) ad_fmage -0.000780 -0.00109** -6.08e-05 -0.000840 -0.00204*** 0.000619 -0.000480 -0.00268 (0.00110) (0.000533) (0.000213) (0.000911) (0.000635) (0.000539) (0.000838) (0.00177) ad_meduc -0.00431 0.0679*** -0.00379* -0.00587 -0.0137* -0.000537 -0.00913 -0.0188 (0.0138) (0.0200) (0.00228) (0.00625) (0.00782) (0.00544) (0.00647) (0.0140) ad_mexper -0.000461 -0.000433 -0.000159 -0.000957 0.00172 0.000804 0.00174 0.00237 (0.00121) (0.000641) (0.000389) (0.00160) (0.00176) (0.00114) (0.00208) (0.00426) ad_mfemale 0.0464 0.0437** -0.00108 0.00888 0.0259 0.00939 0.0198 0.0202 (0.0332) (0.0202) (0.00732) (0.0317) (0.0360) (0.0183) (0.0358) (0.0729) ad_lemp 0.0762*** 0.112*** 0.00229 -0.00314 0.00629 0.00803 0.00727 0.00940 (0.0194) (0.0150) (0.00407) (0.00938) (0.0102) (0.00892) (0.0113) (0.0211) sametown -0.0676 -0.0796 -0.0242 -0.149* -0.161 -0.0163 -0.125 -0.253 (0.0873) (0.0667) (0.0471) (0.0789) (0.117) (0.0459) (0.0813) (0.196) Constant 0.329*** 0.243*** 0.520*** 0.550*** 0.437*** 0.461*** 0.365*** 1.376*** (0.0946) (0.0680) (0.0489) (0.0876) (0.120) (0.0529) (0.0909) (0.218) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 18 Table S5.E: Results for an alternative specification in which a single dummy variable for connectedness (anylink) replaces the three trading relationship dummies in the baseline specification (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES newprod invdum anyrd it pcf_tech suprat htrain ftrain Anylink 0.0179 -0.0111 0.0406 -0.0210 -0.0428 0.00868 0.0502 0.0445 (0.0346) (0.0404) (0.0369) (0.0506) (0.0618) (0.0132) (0.0425) (0.0383) same_sec -0.0327* -0.0561** -0.0216 -0.0573** -0.1000*** -0.00228 -0.0174 -0.0306* (0.0171) (0.0260) (0.0141) (0.0280) (0.0346) (0.00493) (0.0123) (0.0178) Ldist -0.00477** -0.00128 -0.0105* -0.0172** -0.0154* -0.00194 -0.0168*** -0.0153*** (0.00238) (0.00175) (0.00543) (0.00694) (0.00833) (0.00159) (0.00298) (0.00292) ad_fmage -0.000607 -0.000786*** -0.00100 -0.00186 -0.00322* -0.000149 -0.000208 0.00181 (0.000574) (0.000274) (0.00105) (0.00131) (0.00173) (0.000335) (0.000718) (0.00114) ad_meduc 0.00859 0.0142 -0.00998 0.0644** 0.0474** -0.00392* 0.0112 0.00643 (0.00905) (0.00956) (0.00833) (0.0286) (0.0226) (0.00237) (0.0184) (0.0160) ad_mexper -0.000768 0.000113 -0.000576 -0.000886 -0.00313* 0.000211 -0.00167** -0.000977 (0.000897) (0.000387) (0.00132) (0.00151) (0.00170) (0.000477) (0.000804) (0.000837) ad_mfemale -0.00150 0.000617 -0.00558 0.0904** 0.00906 0.00764 0.0181 -0.0121 (0.0186) (0.00597) (0.0285) (0.0454) (0.0482) (0.0111) (0.0258) (0.0177) ad_lemp 0.00628 0.0219** 0.0304** 0.189*** 0.121*** 0.0110** 0.0876*** 0.0987*** (0.00943) (0.00902) (0.0137) (0.0264) (0.0298) (0.00490) (0.0155) (0.0143) Constant 0.483*** 0.475*** 0.264*** 0.431*** 1.005*** 0.172*** 0.323*** 0.275*** (0.0191) (0.0216) (0.0421) (0.0598) (0.0714) (0.0132) (0.0363) (0.0363) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 19 Table S5.E continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 anylink 0.126* 0.0158 -0.0111 0.0910*** 0.00262 -0.0115 0.0288 -0.0624 (0.0732) (0.0383) (0.0333) (0.0348) (0.0720) (0.0662) (0.0766) (0.0605) same_sec -0.0482* -0.0400** -0.0169 -0.00209 -0.0381 -0.00833 -0.0161 0.00153 (0.0261) (0.0160) (0.0155) (0.00909) (0.0286) (0.0272) (0.0257) (0.0187) ldist -0.0251*** -0.0123** -0.000268 0.00531 -0.00900 0.0111** -0.0198*** -0.0171*** (0.00825) (0.00572) (0.00143) (0.00670) (0.00797) (0.00562) (0.00578) (0.00487) ad_fmage 0.00225 -0.000419 0.000479 -0.00207*** -0.00415*** -0.000180 -0.000755 -0.00165 (0.00210) (0.000805) (0.000556) (0.000635) (0.00144) (0.000403) (0.00180) (0.00109) ad_meduc 0.0328 0.0299 0.000664 -0.0205*** 0.0371 -0.00396 -0.00792 -0.00931 (0.0351) (0.0199) (0.00501) (0.00751) (0.0279) (0.00760) (0.0117) (0.0110) ad_mexper -0.000847 -0.00153** 0.000484 -0.00178 0.00455** 0.000134 -0.00131 0.00309 (0.00165) (0.000778) (0.000643) (0.00120) (0.00209) (0.000930) (0.00304) (0.00264) ad_mfemale -0.00553 0.0457 0.00389 0.0235 0.0401 -0.00790 -0.00615 0.0601 (0.0402) (0.0297) (0.00948) (0.0306) (0.0420) (0.0130) (0.0558) (0.0550) ad_lemp 0.229*** 0.131*** 0.00257 0.0151 0.139*** 0.00220 -0.00418 0.00894 (0.0309) (0.0150) (0.00472) (0.0132) (0.0312) (0.00663) (0.0180) (0.0150) Constant 0.700*** 0.218*** 0.485*** 0.226*** 0.858*** 0.949*** 0.840*** 0.726*** (0.0716) (0.0440) (0.0201) (0.0456) (0.0625) (0.0329) (0.0775) (0.0575) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 20 Table S5.E continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf anylink -0.00813 0.0132 0.0229 -0.163** 0.000561 0.202** 0.0211 0.110 (0.0637) (0.0651) (0.0766) (0.0825) (0.0543) (0.101) (0.126) (0.132) same_sec -0.0204 -0.0123 -0.0133 -0.0811* 0.00513 -0.00213 -0.0242 -0.0124 (0.0253) (0.0186) (0.0258) (0.0446) (0.0245) (0.0182) (0.0514) (0.0348) ldist 0.0104* -0.0268*** -0.0151** -0.0248** 0.0139 0.0178* 0.0193 0.0234 (0.00544) (0.00528) (0.00735) (0.0122) (0.0106) (0.0108) (0.0168) (0.0149) ad_fmage 0.000651 6.48e-05 -0.000903 -0.00314 0.00267 -0.00260*** -0.00564*** -0.00397*** (0.000791) (0.00149) (0.00168) (0.00200) (0.00347) (0.000734) (0.00161) (0.00146) ad_meduc -0.00519 -0.00804 -0.00474 -0.00728 0.00388 -0.00916 0.00862 -0.0166 (0.00884) (0.0112) (0.0133) (0.0280) (0.0117) (0.0116) (0.0313) (0.0174) ad_mexper 0.00114 0.00286 0.00290 0.000673 9.82e-06 -0.00274 -0.00170 -0.00411 (0.00151) (0.00312) (0.00340) (0.00321) (0.00226) (0.00205) (0.00455) (0.00285) ad_mfemale 0.0184 0.0541 0.00673 0.121 0.103 -0.0815** -0.220*** -0.110* (0.0299) (0.0598) (0.0613) (0.0932) (0.0670) (0.0389) (0.0686) (0.0583) ad_lemp 0.0152 0.0103 0.00313 0.0786** 0.00786 -0.0259** -0.0373 -0.0235 (0.0132) (0.0190) (0.0164) (0.0311) (0.0242) (0.0118) (0.0363) (0.0224) Constant 0.886*** 0.671*** 1.121*** 1.416*** 0.376*** 0.626*** 1.293*** 1.097*** (0.0527) (0.0637) (0.0818) (0.120) (0.0731) (0.0777) (0.179) (0.133) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 21 Table S5.E continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD anylink 0.0100 -0.0310 -0.00607 0.0137 -0.0567 -0.0119 0.00523 -0.0296 (0.0363) (0.0371) (0.0347) (0.0400) (0.0383) (0.0354) (0.0407) (0.0912) same_sec -0.0411** -0.0161 -0.00511 -0.00460 0.00709 -0.00638 0.00716 -0.00126 (0.0205) (0.0146) (0.0141) (0.0122) (0.0103) (0.0131) (0.00899) (0.0246) ldist -0.0250*** 0.00777 0.00193 -0.0108*** -0.00999** -0.000812 -0.0156*** -0.0310*** (0.00300) (0.00657) (0.00201) (0.00316) (0.00469) (0.00225) (0.00431) (0.00889) ad_fmage -0.000771 -0.00109** -5.91e-05 -0.000843 -0.00204*** 0.000621 -0.000479 -0.00268 (0.00109) (0.000534) (0.000213) (0.000912) (0.000636) (0.000538) (0.000842) (0.00177) ad_meduc -0.00398 0.0684*** -0.00363 -0.00498 -0.0127* -0.000488 -0.00836 -0.0173 (0.0138) (0.0200) (0.00226) (0.00639) (0.00773) (0.00544) (0.00638) (0.0140) ad_mexper -0.000464 -0.000422 -0.000162 -0.000943 0.00174 0.000805 0.00176 0.00239 (0.00122) (0.000641) (0.000390) (0.00160) (0.00177) (0.00114) (0.00209) (0.00426) ad_mfemale 0.0465 0.0439** -0.00106 0.00925 0.0263 0.00943 0.0201 0.0209 (0.0333) (0.0202) (0.00734) (0.0316) (0.0361) (0.0183) (0.0357) (0.0728) ad_lemp 0.0767*** 0.113*** 0.00241 -0.00228 0.00726 0.00818 0.00802 0.0109 (0.0194) (0.0150) (0.00408) (0.00951) (0.0104) (0.00889) (0.0114) (0.0214) Constant 0.265*** 0.167*** 0.497*** 0.408*** 0.283*** 0.446*** 0.246*** 1.135*** (0.0404) (0.0421) (0.0133) (0.0401) (0.0425) (0.0280) (0.0430) (0.0995) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 22 Table S6.E: Results for the baseline specification with the geographical distance variable omitted (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES newprod invdum anyrd it pcf_tech suprat htrain ftrain supbuy 0.0567 -0.0451 0.112 -0.0429 0.104 -0.0188 0.107 -0.0208 (0.127) (0.117) (0.110) (0.187) (0.242) (0.0332) (0.103) (0.129) samesup -0.0119 1.25e-05 0.0585 -0.0425 -0.00192 0.0159 0.0639 0.0296 (0.0394) (0.0452) (0.0417) (0.0647) (0.0681) (0.0139) (0.0456) (0.0404) samecus 0.0702 -0.0249 0.0189 0.0732 -0.0530 0.0175 0.0522 0.139** (0.0586) (0.0706) (0.0697) (0.0962) (0.0980) (0.0240) (0.0736) (0.0608) same_sec -0.0328* -0.0562** -0.0227 -0.0583** -0.102*** -0.00262 -0.0189 -0.0318* (0.0171) (0.0260) (0.0141) (0.0277) (0.0345) (0.00494) (0.0125) (0.0176) ad_fmage -0.000527 -0.000762*** -0.000820 -0.00157 -0.00295* -0.000115 8.61e-05 0.00207* (0.000573) (0.000273) (0.00107) (0.00131) (0.00174) (0.000326) (0.000705) (0.00113) ad_meduc 0.00723 0.0138 -0.0130 0.0594** 0.0430* -0.00449* 0.00625 0.00194 (0.00897) (0.00949) (0.00844) (0.0281) (0.0227) (0.00233) (0.0185) (0.0159) ad_mexper -0.000720 0.000122 -0.000484 -0.000725 -0.00299* 0.000230 -0.00152* -0.000829 (0.000899) (0.000378) (0.00133) (0.00153) (0.00171) (0.000482) (0.000843) (0.000860) ad_mfemale 0.000367 0.00108 -0.00157 0.0970** 0.0150 0.00841 0.0246 -0.00612 (0.0187) (0.00591) (0.0287) (0.0459) (0.0487) (0.0116) (0.0259) (0.0177) ad_lemp 0.00597 0.0218** 0.0298** 0.189*** 0.120*** 0.0109** 0.0866*** 0.0979*** (0.00947) (0.00904) (0.0137) (0.0264) (0.0298) (0.00490) (0.0152) (0.0142) Constant 0.461*** 0.470*** 0.217*** 0.354*** 0.936*** 0.163*** 0.247*** 0.206*** (0.0215) (0.0191) (0.0393) (0.0585) (0.0678) (0.0129) (0.0403) (0.0382) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 23 Table S6.E continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 supbuy -0.0117 0.0439 -0.0118 0.101 0.128 -0.205 0.0643 -0.0232 (0.211) (0.117) (0.105) (0.118) (0.252) (0.204) (0.176) (0.164) samesup 0.115 -0.00401 -0.000994 0.0692* -0.106 -0.0290 0.0556 -0.0408 (0.0780) (0.0490) (0.0404) (0.0356) (0.0754) (0.0920) (0.0893) (0.0662) samecus 0.234** 0.0930 -0.0137 0.0726 0.191 0.0397 0.0441 -0.0459 (0.119) (0.0593) (0.0524) (0.0716) (0.125) (0.0807) (0.131) (0.0995) same_sec -0.0503* -0.0409** -0.0172 -0.00149 -0.0375 -0.00732 -0.0180 0.000124 (0.0262) (0.0162) (0.0155) (0.00911) (0.0286) (0.0273) (0.0256) (0.0188) ad_fmage 0.00269 -0.000210 0.000484 -0.00216*** -0.00401*** -0.000373 -0.000410 -0.00136 (0.00209) (0.000823) (0.000548) (0.000651) (0.00144) (0.000409) (0.00181) (0.00108) ad_meduc 0.0254 0.0263 0.000593 -0.0188*** 0.0345 -0.000792 -0.0137 -0.0143 (0.0348) (0.0198) (0.00493) (0.00718) (0.0277) (0.00740) (0.0117) (0.0111) ad_mexper -0.000609 -0.00141* 0.000487 -0.00183 0.00465** 4.15e-05 -0.00113 0.00325 (0.00169) (0.000775) (0.000640) (0.00120) (0.00208) (0.000930) (0.00305) (0.00265) ad_mfemale 0.00424 0.0504* 0.00401 0.0215 0.0437 -0.0121 0.00148 0.0666 (0.0404) (0.0291) (0.00943) (0.0318) (0.0424) (0.0135) (0.0561) (0.0555) ad_lemp 0.228*** 0.131*** 0.00258 0.0153 0.138*** 0.00292 -0.00527 0.00797 (0.0304) (0.0150) (0.00473) (0.0133) (0.0310) (0.00671) (0.0179) (0.0150) Constant 0.586*** 0.162*** 0.484*** 0.250*** 0.817*** 0.998*** 0.751*** 0.649*** (0.0711) (0.0380) (0.0178) (0.0440) (0.0570) (0.0199) (0.0754) (0.0571) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 24 Table S6.E continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy 0.108 0.0583 -0.0517 -0.0359 -0.106 -0.0709 0.312 0.142 (0.206) (0.156) (0.146) (0.285) (0.137) (0.127) (0.654) (0.426) samesup -0.0459 -0.0171 0.0265 -0.0764 -0.00414 0.225** -0.0204 0.0781 (0.0936) (0.0723) (0.0825) (0.102) (0.0610) (0.109) (0.127) (0.135) samecus -0.000330 0.105 0.0686 -0.163 -0.0209 0.181 -0.0232 0.0982 (0.0920) (0.113) (0.117) (0.135) (0.0934) (0.149) (0.200) (0.192) same_sec -0.0192 -0.0138 -0.0144 -0.0842* 0.00663 -0.00176 -0.0221 -0.0102 (0.0254) (0.0184) (0.0255) (0.0448) (0.0250) (0.0185) (0.0514) (0.0349) ad_fmage 0.000466 0.000527 -0.000639 -0.00273 0.00244 -0.00290*** -0.00598*** -0.00434*** (0.000790) (0.00149) (0.00166) (0.00202) (0.00339) (0.000776) (0.00168) (0.00152) ad_meduc -0.00205 -0.0159 -0.00923 -0.0141 0.00832 -0.00339 0.0153 -0.00830 (0.00862) (0.0112) (0.0134) (0.0279) (0.0119) (0.0103) (0.0306) (0.0156) ad_mexper 0.00104 0.00310 0.00304 0.000874 -0.000134 -0.00288 -0.00199 -0.00440 (0.00153) (0.00314) (0.00340) (0.00322) (0.00233) (0.00209) (0.00446) (0.00286) ad_mfemale 0.0144 0.0644 0.0126 0.131 0.0967 -0.0901** -0.229*** -0.121* (0.0298) (0.0604) (0.0612) (0.0906) (0.0692) (0.0414) (0.0687) (0.0620) ad_lemp 0.0157 0.00871 0.00232 0.0773** 0.00862 -0.0249** -0.0368 -0.0229 (0.0133) (0.0189) (0.0164) (0.0313) (0.0236) (0.0118) (0.0359) (0.0219) Constant 0.933*** 0.551*** 1.053*** 1.303*** 0.438*** 0.706*** 1.380*** 1.202*** (0.0482) (0.0645) (0.0707) (0.0940) (0.0651) (0.0995) (0.159) (0.152) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 25 Table S6.E continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0947 -0.138 -0.0680 0.0228 -0.0108 0.0719 -0.00565 -0.0577 (0.110) (0.120) (0.121) (0.0946) (0.0933) (0.130) (0.0827) (0.179) samesup -0.0391 -0.00338 -0.0213 0.0372 -0.0399 -0.0260 -0.00425 -0.0280 (0.0405) (0.0530) (0.0487) (0.0461) (0.0430) (0.0491) (0.0442) (0.0990) samecus 0.142** -0.0689 0.0371 0.0104 -0.0615 0.000717 0.0473 0.0573 (0.0702) (0.0757) (0.0405) (0.0694) (0.0555) (0.0497) (0.0752) (0.146) same_sec -0.0424** -0.0158 -0.00484 -0.00579 0.00631 -0.00629 0.00628 -0.00341 (0.0202) (0.0147) (0.0141) (0.0121) (0.0105) (0.0130) (0.00905) (0.0243) ad_fmage -0.000344 -0.00122** -9.45e-05 -0.000654 -0.00187*** 0.000633 -0.000208 -0.00214 (0.00109) (0.000568) (0.000213) (0.000912) (0.000632) (0.000548) (0.000842) (0.00177) ad_meduc -0.0113 0.0706*** -0.00308 -0.00813 -0.0157* -0.000676 -0.0130** -0.0264* (0.0133) (0.0203) (0.00215) (0.00636) (0.00804) (0.00538) (0.00656) (0.0149) ad_mexper -0.000230 -0.000495 -0.000174 -0.000846 0.00183 0.000813 0.00190 0.00267 (0.00127) (0.000597) (0.000386) (0.00161) (0.00177) (0.00114) (0.00209) (0.00424) ad_mfemale 0.0562* 0.0409** -0.00175 0.0134 0.0301 0.00974 0.0261 0.0328 (0.0340) (0.0196) (0.00740) (0.0317) (0.0367) (0.0183) (0.0362) (0.0733) ad_lemp 0.0752*** 0.113*** 0.00254 -0.00286 0.00668 0.00808 0.00713 0.00916 (0.0193) (0.0149) (0.00408) (0.00952) (0.0104) (0.00882) (0.0113) (0.0213) Constant 0.152*** 0.202*** 0.505*** 0.359*** 0.238*** 0.442*** 0.175*** 0.995*** (0.0396) (0.0328) (0.0104) (0.0378) (0.0369) (0.0326) (0.0374) (0.0852) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 26 Part II: Results for the Sudan 27 Table S1.S: Results for specifications with dummy variables for web, email and perceived consequences of disputes (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES Web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD Supbuy 0.0201 0.422 -0.461** -0.220 0.307 0.282 0.357 0.403 (0.330) (0.362) (0.186) (0.331) (0.261) (0.343) (0.365) (0.654) samesup -0.116* -0.133** -0.0748 -0.0456 -0.0410 -0.0600 -0.0665 -0.100 (0.0618) (0.0639) (0.0882) (0.0822) (0.0811) (0.0743) (0.0839) (0.140) samecus 0.135*** 0.119*** 0.0352 0.00193 -0.0673 -0.0878 0.0786** 0.0995 (0.0374) (0.0300) (0.0360) (0.0450) (0.0696) (0.0651) (0.0379) (0.101) same_sec -0.0232 -0.0628*** -0.0122 -0.0151 0.00143 -0.00288 0.0198** 0.0151 (0.0151) (0.0192) (0.0139) (0.0154) (0.0118) (0.0113) (0.00982) (0.0231) Ldist -0.00970*** -0.0153*** -0.00603*** -0.00417** -0.000201 -0.00280 -0.00233 0.00565 (0.00332) (0.00236) (0.00171) (0.00191) (0.00302) (0.00220) (0.00387) (0.00673) ad_fmage -0.00224*** -0.000539 -0.000665 0.000387 0.000438 -0.000245 0.00110 -0.000492 (0.000511) (0.000915) (0.000547) (0.000692) (0.000803) (0.000622) (0.00115) (0.00121) ad_meduc -0.00629 0.0166 -0.00185 0.00253 0.00206 0.000247 0.00507 -0.00423 (0.00611) (0.0108) (0.00285) (0.00439) (0.00528) (0.00489) (0.00615) (0.00882) ad_mexper 0.000328 -0.00140** 0.000439 0.000105 -0.000964 -0.000796 -0.000885 -0.000556 (0.00129) (0.000569) (0.000601) (0.000560) (0.000597) (0.000688) (0.00107) (0.00186) ad_mfemale 0.239*** 0.130*** 0.0277 0.00221 -0.00538 0.0114 0.0190 0.0656 (0.0527) (0.0431) (0.0244) (0.0204) (0.0228) (0.0258) (0.0293) (0.0552) ad_lemp 0.0819*** 0.0809*** 0.0121 -0.00904 -0.00439 -0.0177*** -0.0186** -0.0166 (0.0218) (0.0227) (0.0101) (0.00658) (0.00852) (0.00583) (0.00930) (0.0154) Constant 0.213*** 0.321*** 0.468*** 0.484*** 0.450*** 0.492*** 0.396*** 1.105*** (0.0363) (0.0353) (0.0214) (0.0201) (0.0258) (0.0201) (0.0383) (0.0531) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 28 Table S2.S: Results for the baseline specification with interaction terms (same sector x network variables) added (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES newprod invdum anyrd it pcf_tech suprat htrain ftrain supbuy 0.0838 -0.210 -0.302** 0.418 -0.258 -0.101 -0.124 0.378 (0.363) (0.328) (0.134) (0.586) (0.231) (0.0894) (0.265) (0.346) samesup -0.0210 -0.160* -0.150* -0.303** -0.274* -0.0141 0.0298 -0.0122 (0.0842) (0.0820) (0.0804) (0.131) (0.142) (0.0468) (0.0921) (0.0586) samecus 0.0415 0.00296 0.132*** 0.220*** 0.187* -0.0443 0.0826** 0.0219 (0.0263) (0.0355) (0.0363) (0.0623) (0.0988) (0.0293) (0.0396) (0.0704) same_sec -0.00501 -0.0138 -0.0238 -0.109** -0.0577 -0.00136 -0.0280 -0.0251 (0.0199) (0.0269) (0.0188) (0.0486) (0.0388) (0.0117) (0.0245) (0.0218) ldist 4.00e-05 -0.000943 -0.0126*** -0.0224*** -0.0184*** 0.00312 -0.00786** -0.000860 (0.00172) (0.00240) (0.00302) (0.00545) (0.00614) (0.00201) (0.00308) (0.00438) ad_fmage 0.000122 0.000256 0.000360 -0.00318*** -0.00102 -0.000523* -0.000422 0.000879 (0.000283) (0.000368) (0.00103) (0.00120) (0.00133) (0.000299) (0.000662) (0.000992) ad_meduc 0.00570 0.0409*** 0.00758 0.0189 0.0664*** 0.00381 -0.000344 -0.0144*** (0.00616) (0.0144) (0.00926) (0.0167) (0.0236) (0.00339) (0.00577) (0.00474) ad_mexper -1.72e-06 -0.000301 -0.000409 -2.83e-05 -0.000121 0.000258 -0.00155*** 0.000637 (0.000287) (0.000265) (0.00116) (0.00218) (0.00240) (0.000598) (0.000501) (0.00108) ad_mfemale 0.00121 0.0355* 0.0978** 0.435*** 0.241*** 0.0298* 0.0446 0.0848 (0.00629) (0.0187) (0.0396) (0.0909) (0.0751) (0.0176) (0.0375) (0.0542) ad_lemp 0.0231** 0.0524*** 0.0394** 0.167*** 0.197*** 0.0182*** 0.0567*** 0.0780*** (0.00911) (0.0143) (0.0175) (0.0421) (0.0457) (0.00569) (0.0170) (0.0205) secdist 0.00264 0.00283 0.00262 0.00285 0.00657 0.000701 -0.00142 0.00385 (0.00423) (0.00470) (0.00397) (0.00836) (0.00808) (0.00224) (0.00439) (0.00423) seclink -0.130 0.0200 -0.0662 0.0773 -0.0660 0.0322 -0.000292 -0.00623 (0.0960) (0.0816) (0.0833) (0.120) (0.130) (0.0428) (0.0753) (0.0703) Constant 0.461*** 0.375*** 0.332*** 0.507*** 0.812*** 0.292*** 0.377*** 0.0977** (0.0172) (0.0308) (0.0360) (0.0654) (0.0686) (0.0160) (0.0346) (0.0381) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 29 Table S2.S continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 Supbuy -0.209 -0.418** -0.459** 0.0418 -0.822** -0.752* -0.229 0.534 (0.344) (0.175) (0.192) (0.332) (0.392) (0.390) (0.636) (0.520) Samesup 0.0648 -0.145* -0.00667 0.0414 0.0172 -0.197 -0.104 -0.0594 (0.157) (0.0878) (0.0788) (0.0743) (0.153) (0.177) (0.166) (0.158) samecus 0.171 -0.0495 -0.0143 -0.0332 -0.0460 0.0466 0.0242 -0.112 (0.141) (0.0429) (0.0451) (0.0685) (0.106) (0.0544) (0.0665) (0.116) same_sec -0.0778* -0.0809** -0.0152 -0.0109 -0.0969* -0.000360 -0.0586 -0.0196 (0.0470) (0.0337) (0.0262) (0.0183) (0.0568) (0.0337) (0.0471) (0.0329) ldist -0.00648 0.00673** 0.00892** -0.00880** 0.00715 -0.00654** -0.00712** -0.000818 (0.00844) (0.00325) (0.00440) (0.00423) (0.00962) (0.00321) (0.00321) (0.00330) ad_fmage 0.000209 9.67e-05 -8.44e-05 -0.000669 0.000294 -0.000854 0.000376 0.000526 (0.00183) (0.000238) (0.000581) (0.000562) (0.00106) (0.000725) (0.00106) (0.000943) ad_meduc -0.0147 0.0215* 0.00583 0.00165 0.0279 0.000870 0.00329 0.00230 (0.0131) (0.0112) (0.00657) (0.00477) (0.0173) (0.00727) (0.00754) (0.00795) ad_mexper -0.000751 -9.70e-05 -0.000213 -0.000397 -0.000630 0.000670 0.000130 -0.00114 (0.00186) (0.000304) (0.000727) (0.000632) (0.00120) (0.000804) (0.000918) (0.000702) ad_mfemale 0.204* 0.00290 -0.0164 0.0110 -0.00177 0.0297 0.00214 -0.0102 (0.114) (0.00675) (0.0126) (0.0382) (0.0392) (0.0332) (0.0341) (0.0252) ad_lemp 0.207*** 0.0658*** 0.00494 -0.00398 0.0977*** 0.0220 -0.0130 -0.00427 (0.0499) (0.0145) (0.00781) (0.0123) (0.0264) (0.0157) (0.0113) (0.0115) secdist 0.00306 0.0117** 0.00195 0.00115 0.0117 -0.00229 0.00842 0.00366 (0.00872) (0.00572) (0.00557) (0.00395) (0.0104) (0.00729) (0.00806) (0.00605) seclink 0.0288 0.0323 0.0537 0.0182 0.0618 0.0901 0.0705 0.0602 (0.168) (0.0868) (0.0779) (0.0708) (0.153) (0.140) (0.137) (0.141) Constant 0.740*** 0.364*** 0.419*** 0.223*** 0.907*** 0.948*** 0.985*** 0.955*** (0.0815) (0.0299) (0.0313) (0.0434) (0.0587) (0.0309) (0.0326) (0.0309) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 30 Table S2.S continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy 0.287 0.392 0.281 0 -0.541* -0.551 1.573 0 (0.646) (0.645) (0.798) (0) (0.277) (0.785) (2.537) (0) samesup -0.0995 -0.0905 -0.0443 -1.434** -0.284 -1.043*** -0.802** -0.629** (0.162) (0.178) (0.156) (0.559) (0.185) (0.318) (0.394) (0.284) samecus -0.105 0.118** 0.0414 -0.946 0.0868 -0.624*** 0.284 -0.0639 (0.112) (0.0493) (0.0863) (0.684) (0.139) (0.209) (0.451) (0.219) same_sec -0.0118 0.0208 -0.00343 -0.320 -0.0704 -0.0915 -0.0903 -0.239 (0.0340) (0.0316) (0.0342) (0.291) (0.0618) (0.114) (0.212) (0.167) ldist -0.00328 -0.00496 0.00566 -0.0662* -0.0177 -0.0526** -0.0522 -0.0146 (0.00301) (0.00487) (0.00701) (0.0399) (0.0113) (0.0251) (0.0351) (0.0200) ad_fmage -0.000696 0.00154 -0.00117 -0.00498 0.00111 0.00861 0.00458 0.00600 (0.000812) (0.00156) (0.000985) (0.00781) (0.00216) (0.00650) (0.00745) (0.00460) ad_meduc -0.00143 0.00724 0.00139 0.0468 -0.00464 -0.0332 -0.0201 0.0120 (0.00704) (0.00944) (0.00863) (0.0477) (0.0116) (0.0229) (0.0267) (0.0251) ad_mexper -0.000772 -0.00118 0.000485 -0.00728 0.000767 0.00304 -0.00385 -0.00181 (0.000947) (0.00144) (0.00169) (0.00579) (0.00132) (0.00695) (0.00580) (0.00314) ad_mfemale 0.0157 0.0262 0.0370 0.376 -0.00715 -0.161 0.0561 -0.138 (0.0387) (0.0461) (0.0481) (0.582) (0.0613) (0.134) (0.274) (0.0959) ad_lemp -0.0267*** -0.0311** -0.0111 0.00746 0.0654** 0.0906 0.102 0.0611 (0.00876) (0.0133) (0.0136) (0.110) (0.0258) (0.0709) (0.114) (0.0599) secdist 0.00247 0.00173 0.00412 0.0382 0.00321 0.00876 0.00547 0.0307 (0.00628) (0.00628) (0.00674) (0.0511) (0.00845) (0.0259) (0.0315) (0.0217) seclink -0.00581 -0.0456 0.133 -0.244 0.0255 0.315 0.183 0.373 (0.171) (0.142) (0.118) (0.646) (0.238) (0.247) (0.451) (0.275) Constant 1.004*** 0.894*** 1.126*** 3.775*** 0.716*** 1.523*** 2.060*** 0.994*** (0.0242) (0.0489) (0.0506) (0.379) (0.0995) (0.290) (0.337) (0.222) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 31 Table S2.S continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES Web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.00413 0.397 -0.476** -0.236 0.285 0.287 0.384 0.377 (0.339) (0.366) (0.196) (0.346) (0.268) (0.361) (0.368) (0.674) samesup -0.131* -0.156** -0.0874 -0.0620 -0.0617 -0.0554 -0.0431 -0.126 (0.0718) (0.0742) (0.0938) (0.0857) (0.0897) (0.0844) (0.0981) (0.153) samecus 0.131*** 0.111*** 0.0326 -0.00390 -0.0737 -0.0866 0.0850** 0.0915 (0.0392) (0.0314) (0.0356) (0.0449) (0.0706) (0.0661) (0.0390) (0.0982) same_sec -0.0280 -0.0729*** -0.000199 -0.0330 -0.00775 -0.00419 0.0176 0.00165 (0.0233) (0.0280) (0.0187) (0.0240) (0.0218) (0.0180) (0.0190) (0.0383) ldist -0.00982*** -0.0156*** -0.00569*** -0.00465** -0.000431 -0.00284 -0.00242 0.00531 (0.00334) (0.00248) (0.00177) (0.00191) (0.00309) (0.00223) (0.00392) (0.00675) ad_fmage -0.00224*** -0.000541 -0.000663 0.000383 0.000436 -0.000245 0.00110 -0.000495 (0.000511) (0.000914) (0.000546) (0.000691) (0.000803) (0.000623) (0.00115) (0.00121) ad_meduc -0.00629 0.0166 -0.00187 0.00254 0.00206 0.000250 0.00508 -0.00423 (0.00612) (0.0108) (0.00285) (0.00439) (0.00528) (0.00489) (0.00615) (0.00882) ad_mexper 0.000328 -0.00140** 0.000438 0.000104 -0.000965 -0.000796 -0.000884 -0.000556 (0.00129) (0.000569) (0.000601) (0.000560) (0.000596) (0.000688) (0.00107) (0.00186) ad_mfemale 0.239*** 0.130*** 0.0276 0.00233 -0.00531 0.0115 0.0190 0.0657 (0.0527) (0.0431) (0.0244) (0.0204) (0.0228) (0.0258) (0.0293) (0.0552) ad_lemp 0.0819*** 0.0809*** 0.0121 -0.00905 -0.00439 -0.0177*** -0.0186** -0.0166 (0.0218) (0.0227) (0.0101) (0.00659) (0.00852) (0.00583) (0.00931) (0.0154) secdist 0.000999 0.00216 -0.00286 0.00399 0.00195 0.000335 0.000700 0.00289 (0.00406) (0.00460) (0.00383) (0.00428) (0.00380) (0.00340) (0.00358) (0.00712) seclink 0.0359 0.0587 0.0207 0.0473 0.0514 -0.00955 -0.0507 0.0643 (0.0691) (0.0755) (0.0753) (0.0788) (0.0805) (0.0832) (0.0840) (0.135) Constant 0.214*** 0.323*** 0.466*** 0.486*** 0.451*** 0.492*** 0.397*** 1.107*** (0.0360) (0.0350) (0.0217) (0.0202) (0.0259) (0.0198) (0.0384) (0.0537) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 32 Table S3.S: Results for the baseline specification with interaction terms (similar firm size x network variables) added (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES newprod invdum anyrd it pcf_tech suprat htrain ftrain Supbuy 0.0165 -0.203 -0.342** 0.448 -0.298 -0.0830 -0.122 0.369 (0.351) (0.327) (0.141) (0.571) (0.237) (0.0875) (0.263) (0.340) Samesup -0.100 -0.177** -0.175** -0.277** -0.327** -0.00326 0.0235 -0.0235 (0.0773) (0.0826) (0.0720) (0.118) (0.136) (0.0428) (0.0845) (0.0621) Samecus 0.0168 -0.00395 0.127*** 0.227*** 0.172* -0.0416 0.0793** 0.0194 (0.0228) (0.0341) (0.0370) (0.0631) (0.100) (0.0295) (0.0379) (0.0729) same_sec 0.00455 -0.00138 -0.0133 -0.0955*** -0.0299 0.00208 -0.0341* -0.00870 (0.0101) (0.0134) (0.0185) (0.0301) (0.0309) (0.00619) (0.0189) (0.0159) Ldist -2.85e-05 -0.00107 -0.0116*** -0.0213*** -0.0187*** 0.00251 -0.00783*** -0.000565 (0.00176) (0.00231) (0.00291) (0.00513) (0.00602) (0.00191) (0.00299) (0.00436) ad_fmage 0.000119 0.000256 0.000364 -0.00318*** -0.00102 -0.000526* -0.000422 0.000881 (0.000283) (0.000368) (0.00103) (0.00120) (0.00134) (0.000298) (0.000663) (0.000991) ad_meduc 0.00575 0.0410*** 0.00742 0.0187 0.0666*** 0.00394 -0.000371 -0.0144*** (0.00615) (0.0143) (0.00921) (0.0167) (0.0235) (0.00338) (0.00572) (0.00472) ad_mexper -1.79e-06 -0.000297 -0.000414 -3.09e-05 -0.000118 0.000262 -0.00155*** 0.000638 (0.000289) (0.000267) (0.00115) (0.00218) (0.00240) (0.000591) (0.000501) (0.00108) ad_mfemale 0.00116 0.0355* 0.0977** 0.435*** 0.241*** 0.0299* 0.0446 0.0847 (0.00633) (0.0187) (0.0396) (0.0909) (0.0750) (0.0175) (0.0375) (0.0541) ad_lemp 0.0252*** 0.0549*** 0.0364** 0.163*** 0.203*** 0.0216*** 0.0558*** 0.0788*** (0.00956) (0.0141) (0.0181) (0.0449) (0.0482) (0.00575) (0.0179) (0.0218) sizdist 0.00314 0.00374 -0.00467 -0.00608 0.00831 0.00545*** -0.00156 0.00135 (0.00234) (0.00350) (0.00441) (0.00880) (0.00907) (0.00150) (0.00396) (0.00390) sizlink 0.0812 0.0957 -0.0390 0.0123 0.0712 0.0212 0.0288 0.0180 (0.0713) (0.0813) (0.0815) (0.142) (0.138) (0.0459) (0.0966) (0.0641) Constant 0.457*** 0.370*** 0.334*** 0.510*** 0.802*** 0.287*** 0.379*** 0.0946** (0.0177) (0.0296) (0.0362) (0.0689) (0.0703) (0.0162) (0.0356) (0.0391) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 33 Table S3.S continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 supbuy -0.199 -0.421** -0.434** 0.0469 -0.809** -0.703* -0.209 0.558 (0.336) (0.173) (0.176) (0.330) (0.379) (0.372) (0.622) (0.515) samesup 0.0508 -0.158** 0.00810 0.0160 -8.64e-05 -0.111 -0.0894 -0.0473 (0.155) (0.0803) (0.0787) (0.0720) (0.140) (0.166) (0.171) (0.147) samecus 0.164 -0.0511 -0.0104 -0.0442 -0.0514 0.0757 0.0301 -0.109 (0.146) (0.0425) (0.0427) (0.0633) (0.0978) (0.0568) (0.0705) (0.116) same_sec -0.0644* -0.0299* -0.00606 -0.00578 -0.0460* -0.00877 -0.0213 -0.00305 (0.0375) (0.0157) (0.0137) (0.0132) (0.0279) (0.0213) (0.0281) (0.0193) ldist -0.00652 0.00761** 0.00869** -0.00813** 0.00779 -0.00713** -0.00615** -0.000492 (0.00833) (0.00319) (0.00430) (0.00403) (0.00945) (0.00315) (0.00304) (0.00327) ad_fmage 0.000210 0.000102 -8.47e-05 -0.000665 0.000300 -0.000857 0.000383 0.000530 (0.00183) (0.000240) (0.000582) (0.000563) (0.00107) (0.000726) (0.00106) (0.000943) ad_meduc -0.0147 0.0216* 0.00593 0.00153 0.0280 0.000944 0.00326 0.00232 (0.0130) (0.0112) (0.00658) (0.00477) (0.0173) (0.00721) (0.00756) (0.00795) ad_mexper -0.000748 -9.45e-05 -0.000210 -0.000397 -0.000626 0.000671 0.000131 -0.00113 (0.00185) (0.000305) (0.000727) (0.000633) (0.00121) (0.000803) (0.000917) (0.000702) ad_mfemale 0.204* 0.00266 -0.0164 0.0109 -0.00202 0.0297 0.00191 -0.0103 (0.114) (0.00685) (0.0126) (0.0382) (0.0391) (0.0333) (0.0342) (0.0252) ad_lemp 0.209*** 0.0685*** 0.00721 -0.00630 0.102*** 0.0230 -0.0128 -0.00365 (0.0527) (0.0146) (0.00814) (0.0134) (0.0270) (0.0161) (0.0120) (0.0125) sizdist 0.00333 0.00426 0.00365* -0.00403 0.00587 0.00220 0.000319 0.000895 (0.00868) (0.00367) (0.00217) (0.00348) (0.00592) (0.00551) (0.00450) (0.00373) sizlink 0.108 0.0586 0.0326 0.120 0.128 -0.167 0.0267 0.0428 (0.177) (0.0846) (0.0916) (0.0752) (0.142) (0.167) (0.198) (0.144) Constant 0.736*** 0.354*** 0.414*** 0.225*** 0.896*** 0.948*** 0.980*** 0.952*** (0.0851) (0.0299) (0.0305) (0.0444) (0.0585) (0.0307) (0.0328) (0.0316) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 34 Table S3.S continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy 0.281 0.369 0.342 0 -0.536** -0.427 1.637 0 (0.614) (0.637) (0.784) (0) (0.222) (0.674) (2.479) (0) samesup -0.104 -0.103 -0.00129 -1.579*** -0.245** -0.871** -0.745* -0.486** (0.142) (0.152) (0.158) (0.486) (0.120) (0.351) (0.401) (0.212) samecus -0.105 0.116** 0.0534 -0.967 0.102 -0.572*** 0.295 -0.0104 (0.113) (0.0515) (0.0881) (0.682) (0.146) (0.201) (0.445) (0.228) same_sec -0.00128 0.0275* 0.0162 -0.127 -0.0553 -0.0454 -0.0611 -0.0760 (0.0186) (0.0164) (0.0189) (0.157) (0.0393) (0.0681) (0.132) (0.0863) ldist -0.00318 -0.00529 0.00539 -0.0536 -0.0171 -0.0493** -0.0491 -0.0106 (0.00282) (0.00473) (0.00705) (0.0374) (0.0109) (0.0232) (0.0334) (0.0180) ad_fmage -0.000696 0.00154 -0.00117 -0.00484 0.00112 0.00862 0.00459 0.00602 (0.000812) (0.00156) (0.000984) (0.00780) (0.00216) (0.00651) (0.00746) (0.00458) ad_meduc -0.00140 0.00733 0.00155 0.0423 -0.00472 -0.0337 -0.0207 0.0115 (0.00705) (0.00940) (0.00863) (0.0469) (0.0117) (0.0228) (0.0268) (0.0252) ad_mexper -0.000771 -0.00118 0.000492 -0.00734 0.000762 0.00303 -0.00384 -0.00186 (0.000949) (0.00144) (0.00169) (0.00577) (0.00132) (0.00696) (0.00580) (0.00313) ad_mfemale 0.0157 0.0262 0.0370 0.369 -0.00735 -0.161 0.0564 -0.137 (0.0387) (0.0462) (0.0481) (0.581) (0.0615) (0.134) (0.275) (0.0962) ad_lemp -0.0257*** -0.0286** -0.00741 -0.0289 0.0637** 0.0797 0.0906 0.0626 (0.00946) (0.0142) (0.0141) (0.110) (0.0262) (0.0739) (0.119) (0.0624) sizdist 0.00162 0.00423 0.00595 -0.0566** -0.00242 -0.0165 -0.0188 0.00228 (0.00398) (0.00378) (0.00518) (0.0287) (0.00573) (0.0157) (0.0236) (0.0148) sizlink -0.00144 -0.0345 0.0535 -0.260 -0.118 -0.138 0.0733 0.0163 (0.155) (0.151) (0.139) (0.525) (0.140) (0.272) (0.539) (0.239) Constant 1.002*** 0.890*** 1.119*** 3.797*** 0.717*** 1.531*** 2.070*** 0.971*** (0.0257) (0.0492) (0.0500) (0.386) (0.0992) (0.295) (0.337) (0.220) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 35 Table S3.S continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0194 0.419 -0.460** -0.219 0.308 0.282 0.359 0.407 (0.329) (0.361) (0.185) (0.331) (0.262) (0.343) (0.365) (0.656) samesup -0.123* -0.137** -0.0510 -0.0526 -0.0479 -0.0570 -0.0680 -0.109 (0.0675) (0.0653) (0.0884) (0.0870) (0.0815) (0.0769) (0.0804) (0.146) samecus 0.132*** 0.117*** 0.0451 -0.000960 -0.0701 -0.0865 0.0780** 0.0961 (0.0390) (0.0337) (0.0360) (0.0464) (0.0689) (0.0655) (0.0394) (0.101) same_sec -0.0232 -0.0628*** -0.0122 -0.0152 0.00139 -0.00288 0.0198** 0.0150 (0.0150) (0.0192) (0.0139) (0.0154) (0.0118) (0.0113) (0.00980) (0.0231) ldist -0.00952*** -0.0147*** -0.00613*** -0.00419** -0.000521 -0.00287 -0.00275 0.00472 (0.00327) (0.00227) (0.00169) (0.00190) (0.00301) (0.00217) (0.00386) (0.00674) ad_fmage -0.00224*** -0.000535 -0.000666 0.000387 0.000437 -0.000245 0.00109 -0.000497 (0.000509) (0.000914) (0.000546) (0.000691) (0.000804) (0.000623) (0.00115) (0.00121) ad_meduc -0.00632 0.0165 -0.00184 0.00253 0.00212 0.000260 0.00515 -0.00405 (0.00612) (0.0108) (0.00283) (0.00439) (0.00528) (0.00489) (0.00614) (0.00882) ad_mexper 0.000328 -0.00141** 0.000438 0.000105 -0.000962 -0.000796 -0.000883 -0.000550 (0.00129) (0.000567) (0.000600) (0.000560) (0.000596) (0.000689) (0.00108) (0.00186) ad_mfemale 0.239*** 0.130*** 0.0277 0.00223 -0.00533 0.0115 0.0190 0.0658 (0.0528) (0.0431) (0.0244) (0.0204) (0.0228) (0.0258) (0.0293) (0.0552) ad_lemp 0.0811*** 0.0779*** 0.0123 -0.00887 -0.00280 -0.0174*** -0.0166* -0.0121 (0.0234) (0.0241) (0.0104) (0.00704) (0.00915) (0.00637) (0.00991) (0.0163) sizdist -0.00140 -0.00502 0.000652 0.000215 0.00254 0.000548 0.00332 0.00731 (0.00434) (0.00500) (0.00306) (0.00224) (0.00229) (0.00233) (0.00230) (0.00547) sizlink 0.0295 0.0101 -0.0930 0.0281 0.0310 -0.0114 0.0108 0.0441 (0.0756) (0.0932) (0.0893) (0.107) (0.0843) (0.0876) (0.0831) (0.152) Constant 0.214*** 0.325*** 0.468*** 0.483*** 0.448*** 0.492*** 0.394*** 1.100*** (0.0379) (0.0368) (0.0219) (0.0204) (0.0264) (0.0209) (0.0387) (0.0532) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 36 Table S4.S: Results for the baseline specification with a dummy variable for location in the same town added. (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES newprod invdum anyrd it pcf_tech suprat htrain ftrain supbuy 0.0148 -0.206 -0.341** 0.450 -0.304 -0.0872 -0.122 0.365 (0.351) (0.325) (0.141) (0.571) (0.230) (0.0886) (0.263) (0.349) samesup -0.0800 -0.154* -0.182*** -0.271** -0.308** 0.000282 0.0316 -0.0175 (0.0740) (0.0790) (0.0645) (0.109) (0.130) (0.0439) (0.0844) (0.0543) samecus 0.0243 0.00623 0.127*** 0.231*** 0.183* -0.0375 0.0837** 0.0282 (0.0221) (0.0336) (0.0347) (0.0608) (0.101) (0.0286) (0.0378) (0.0704) same_sec 0.00459 -0.00127 -0.0132 -0.0955*** -0.0298 0.00218 -0.0341* -0.00860 (0.0101) (0.0134) (0.0185) (0.0303) (0.0311) (0.00622) (0.0190) (0.0159) ldist 0.00220 -0.00557 -0.0333*** -0.0303 -0.0465* -0.0132 -0.0143 -0.0422** (0.00489) (0.00775) (0.0107) (0.0237) (0.0254) (0.00810) (0.0121) (0.0176) ad_fmage 0.000123 0.000252 0.000333 -0.00319*** -0.00105 -0.000544* -0.000431 0.000825 (0.000284) (0.000366) (0.00102) (0.00120) (0.00133) (0.000294) (0.000661) (0.000970) ad_meduc 0.00555 0.0413*** 0.00900 0.0195 0.0683*** 0.00494 0.000108 -0.0115*** (0.00617) (0.0146) (0.00923) (0.0168) (0.0236) (0.00343) (0.00594) (0.00447) ad_mexper 2.37e-06 -0.000322 -0.000497 -6.26e-05 -0.000238 0.000188 -0.00158*** 0.000459 (0.000288) (0.000271) (0.00120) (0.00223) (0.00248) (0.000587) (0.000493) (0.00104) ad_mfemale 0.00137 0.0348* 0.0950** 0.434*** 0.237*** 0.0277 0.0438 0.0795 (0.00625) (0.0189) (0.0396) (0.0916) (0.0764) (0.0178) (0.0375) (0.0539) ad_lemp 0.0234*** 0.0519*** 0.0369** 0.166*** 0.194*** 0.0163*** 0.0560*** 0.0731*** (0.00905) (0.0146) (0.0174) (0.0422) (0.0455) (0.00580) (0.0172) (0.0208) sametown 0.0118 -0.0321 -0.136** -0.0534 -0.187 -0.106** -0.0405 -0.270** (0.0297) (0.0543) (0.0674) (0.143) (0.158) (0.0499) (0.0729) (0.121) Constant 0.448*** 0.405*** 0.464*** 0.557*** 0.992*** 0.396*** 0.418*** 0.361*** (0.0318) (0.0568) (0.0815) (0.164) (0.175) (0.0529) (0.0834) (0.118) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 37 Table S4.S continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 supbuy -0.205 -0.425** -0.435** 0.0485 -0.817** -0.704* -0.208 0.559 (0.351) (0.175) (0.176) (0.329) (0.372) (0.373) (0.626) (0.518) samesup 0.0768 -0.144* 0.0147 0.0482 0.0351 -0.154 -0.0836 -0.0376 (0.147) (0.0760) (0.0788) (0.0810) (0.134) (0.166) (0.164) (0.145) samecus 0.185 -0.0417 -0.00873 -0.0302 -0.0261 0.0581 0.0300 -0.107 (0.141) (0.0419) (0.0437) (0.0649) (0.101) (0.0558) (0.0662) (0.117) same_sec -0.0642* -0.0298* -0.00604 -0.00569 -0.0455 -0.00888 -0.0213 -0.00307 (0.0377) (0.0157) (0.0136) (0.0131) (0.0284) (0.0213) (0.0281) (0.0194) ldist -0.0669** -0.0152 0.0158 -0.0130 -0.0608* -0.00770 0.0105 0.0157 (0.0320) (0.00972) (0.0166) (0.0108) (0.0325) (0.0124) (0.0190) (0.0186) ad_fmage 0.000134 7.33e-05 -7.32e-05 -0.000673 0.000226 -0.000855 0.000405 0.000552 (0.00181) (0.000250) (0.000582) (0.000561) (0.00105) (0.000721) (0.00106) (0.000952) ad_meduc -0.0105 0.0231** 0.00537 0.00195 0.0328* 0.000958 0.00210 0.00117 (0.0130) (0.0111) (0.00635) (0.00493) (0.0174) (0.00761) (0.00757) (0.00781) ad_mexper -0.00101 -0.000198 -0.000184 -0.000414 -0.000908 0.000669 0.000201 -0.00107 (0.00181) (0.000295) (0.000723) (0.000627) (0.00122) (0.000817) (0.000901) (0.000698) ad_mfemale 0.196* -0.000363 -0.0156 0.0104 -0.0108 0.0297 0.00399 -0.00830 (0.115) (0.00730) (0.0127) (0.0382) (0.0384) (0.0332) (0.0337) (0.0244) ad_lemp 0.200*** 0.0631*** 0.00573 -0.00447 0.0899*** 0.0218 -0.0111 -0.00237 (0.0499) (0.0147) (0.00790) (0.0127) (0.0261) (0.0158) (0.0113) (0.0116) sametown -0.392* -0.151** 0.0432 -0.0278 -0.448** -0.00566 0.107 0.104 (0.211) (0.0747) (0.101) (0.0632) (0.218) (0.0797) (0.117) (0.109) Constant 1.123*** 0.506*** 0.375*** 0.249*** 1.340*** 0.955*** 0.875*** 0.851*** (0.220) (0.0714) (0.105) (0.0780) (0.214) (0.0852) (0.130) (0.124) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 38 Table S4.S continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy 0.280 0.368 0.336 0 -0.533** -0.468 1.627 0 (0.614) (0.638) (0.774) (0) (0.221) (0.651) (2.462) (0) samesup -0.105 -0.114 0.0117 -1.654*** -0.278** -0.891*** -0.703** -0.467** (0.141) (0.164) (0.151) (0.452) (0.122) (0.333) (0.349) (0.200) samecus -0.107 0.109** 0.0640 -0.990 0.0883 -0.562*** 0.315 -0.00189 (0.112) (0.0478) (0.0885) (0.659) (0.140) (0.203) (0.445) (0.217) same_sec -0.00130 0.0275* 0.0164 -0.135 -0.0557 -0.0462 -0.0615 -0.0737 (0.0186) (0.0164) (0.0187) (0.156) (0.0392) (0.0677) (0.132) (0.0862) ldist 0.00532 0.0141 -0.0296 -0.0376 -0.00877 -0.176** -0.108 -0.0595 (0.0183) (0.0229) (0.0281) (0.104) (0.0453) (0.0766) (0.0973) (0.0577) ad_fmage -0.000684 0.00157 -0.00121 -0.00491 0.00113 0.00856 0.00454 0.00600 (0.000807) (0.00157) (0.000976) (0.00780) (0.00215) (0.00657) (0.00742) (0.00455) ad_meduc -0.00202 0.00592 0.00387 0.0442 -0.00531 -0.0235 -0.0153 0.0158 (0.00686) (0.00920) (0.00847) (0.0482) (0.0111) (0.0238) (0.0269) (0.0260) ad_mexper -0.000737 -0.00110 0.000334 -0.00721 0.000805 0.00248 -0.00410 -0.00205 (0.000935) (0.00142) (0.00174) (0.00587) (0.00131) (0.00707) (0.00564) (0.00304) ad_mfemale 0.0167 0.0285 0.0324 0.379 -0.00632 -0.178 0.0487 -0.147 (0.0385) (0.0459) (0.0481) (0.587) (0.0617) (0.132) (0.275) (0.0995) ad_lemp -0.0257*** -0.0289** -0.0153 0.0104 0.0662** 0.0747 0.0944 0.0556 (0.00862) (0.0133) (0.0135) (0.115) (0.0272) (0.0705) (0.116) (0.0599) sametown 0.0537 0.122 -0.231 0.158 0.0559 -0.818* -0.369 -0.333 (0.109) (0.129) (0.169) (0.615) (0.257) (0.428) (0.556) (0.301) Constant 0.950*** 0.773*** 1.351*** 3.600*** 0.660** 2.313*** 2.414*** 1.289*** (0.120) (0.148) (0.178) (0.725) (0.281) (0.566) (0.620) (0.394) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 39 Table S4.S continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0199 0.421 -0.460** -0.219 0.308 0.283 0.359 0.403 (0.330) (0.361) (0.185) (0.334) (0.265) (0.343) (0.366) (0.652) samesup -0.116* -0.132** -0.0749 -0.0461 -0.0419 -0.0604 -0.0675 -0.100 (0.0619) (0.0637) (0.0882) (0.0822) (0.0810) (0.0741) (0.0841) (0.140) samecus 0.135*** 0.120*** 0.0348 0.000112 -0.0704 -0.0892 0.0753** 0.100 (0.0374) (0.0298) (0.0362) (0.0447) (0.0704) (0.0654) (0.0382) (0.102) same_sec -0.0232 -0.0628*** -0.0122 -0.0152 0.00136 -0.00291 0.0197** 0.0151 (0.0151) (0.0192) (0.0139) (0.0154) (0.0118) (0.0113) (0.00988) (0.0230) ldist -0.0114 -0.0223** -0.00318 0.00653 0.0180 0.00552 0.0171 0.00181 (0.0161) (0.0101) (0.00891) (0.0114) (0.0149) (0.0123) (0.0163) (0.0294) ad_fmage -0.00225*** -0.000548 -0.000661 0.000401 0.000463 -0.000234 0.00112 -0.000497 (0.000512) (0.000914) (0.000544) (0.000693) (0.000807) (0.000621) (0.00115) (0.00121) ad_meduc -0.00617 0.0171 -0.00205 0.00178 0.000791 -0.000331 0.00372 -0.00397 (0.00618) (0.0110) (0.00309) (0.00439) (0.00522) (0.00478) (0.00602) (0.00851) ad_mexper 0.000321 -0.00143** 0.000451 0.000151 -0.000886 -0.000760 -0.000802 -0.000572 (0.00131) (0.000584) (0.000604) (0.000546) (0.000598) (0.000676) (0.00105) (0.00187) ad_mfemale 0.239*** 0.129*** 0.0281 0.00356 -0.00308 0.0125 0.0214 0.0652 (0.0532) (0.0433) (0.0244) (0.0202) (0.0223) (0.0258) (0.0292) (0.0554) ad_lemp 0.0817*** 0.0801*** 0.0124 -0.00778 -0.00225 -0.0167*** -0.0163* -0.0171 (0.0218) (0.0228) (0.0101) (0.00647) (0.00844) (0.00577) (0.00939) (0.0154) sametown -0.0112 -0.0448 0.0184 0.0691 0.118 0.0538 0.126 -0.0248 (0.0987) (0.0597) (0.0539) (0.0678) (0.0856) (0.0701) (0.0918) (0.172) Constant 0.225** 0.365*** 0.450*** 0.416*** 0.335*** 0.439*** 0.273*** 1.130*** (0.110) (0.0696) (0.0613) (0.0760) (0.0955) (0.0813) (0.106) (0.186) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 40 Table S5.S: Results for an alternative specification in which a single dummy variable for connectedness (anylink) replaces the three trading relationship dummies in the baseline specification (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES Newprod invdum anyrd it pcf_tech suprat htrain ftrain anylink 0.00393 -0.0278 0.0580 0.130** 0.0795 -0.0325 0.0719** 0.0147 (0.0227) (0.0295) (0.0382) (0.0597) (0.0914) (0.0254) (0.0318) (0.0611) same_sec 0.00407 -0.00218 -0.0151 -0.0982*** -0.0325 0.00232 -0.0344* -0.00885 (0.0102) (0.0133) (0.0186) (0.0302) (0.0312) (0.00624) (0.0189) (0.0160) ldist 0.000432 -0.000509 -0.0120*** -0.0218*** -0.0173*** 0.00318 -0.00799** -0.000379 (0.00165) (0.00221) (0.00301) (0.00530) (0.00599) (0.00195) (0.00314) (0.00434) ad_fmage 0.000123 0.000263 0.000370 -0.00317*** -0.00100 -0.000523* -0.000422 0.000881 (0.000283) (0.000368) (0.00104) (0.00120) (0.00134) (0.000299) (0.000663) (0.000992) ad_meduc 0.00565 0.0409*** 0.00750 0.0188 0.0663*** 0.00382 -0.000349 -0.0145*** (0.00616) (0.0144) (0.00929) (0.0167) (0.0236) (0.00339) (0.00577) (0.00474) ad_mexper -4.12e-06 -0.000299 -0.000408 -2.31e-05 -0.000120 0.000258 -0.00155*** 0.000637 (0.000287) (0.000265) (0.00116) (0.00218) (0.00240) (0.000598) (0.000500) (0.00108) ad_mfemale 0.00137 0.0358* 0.0985** 0.436*** 0.242*** 0.0297* 0.0447 0.0847 (0.00630) (0.0187) (0.0396) (0.0909) (0.0752) (0.0176) (0.0374) (0.0541) ad_lemp 0.0232** 0.0525*** 0.0396** 0.167*** 0.198*** 0.0182*** 0.0567*** 0.0780*** (0.00912) (0.0143) (0.0175) (0.0421) (0.0457) (0.00569) (0.0170) (0.0205) Constant 0.460*** 0.373*** 0.329*** 0.503*** 0.807*** 0.292*** 0.378*** 0.0956** (0.0174) (0.0302) (0.0357) (0.0660) (0.0687) (0.0158) (0.0345) (0.0381) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 41 Table S5.S continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 anylink 0.154 -0.0676* -0.00467 -0.0138 -0.0276 0.0122 0.00939 -0.0862 (0.120) (0.0378) (0.0365) (0.0540) (0.0810) (0.0566) (0.0629) (0.0940) same_sec -0.0650* -0.0305* -0.00597 -0.00526 -0.0456 -0.0102 -0.0219 -0.00255 (0.0374) (0.0157) (0.0137) (0.0130) (0.0279) (0.0212) (0.0282) (0.0192) ldist -0.00606 0.00820*** 0.00915** -0.00872** 0.00852 -0.00670** -0.00604** -0.000429 (0.00848) (0.00314) (0.00422) (0.00415) (0.00923) (0.00305) (0.00308) (0.00322) ad_fmage 0.000215 0.000108 -8.13e-05 -0.000670 0.000304 -0.000847 0.000386 0.000526 (0.00183) (0.000239) (0.000582) (0.000563) (0.00107) (0.000722) (0.00106) (0.000943) ad_meduc -0.0148 0.0214* 0.00583 0.00166 0.0278 0.000845 0.00323 0.00232 (0.0131) (0.0112) (0.00657) (0.00477) (0.0173) (0.00728) (0.00755) (0.00795) ad_mexper -0.000749 -9.66e-05 -0.000212 -0.000397 -0.000630 0.000675 0.000132 -0.00114 (0.00185) (0.000305) (0.000727) (0.000633) (0.00121) (0.000804) (0.000917) (0.000702) ad_mfemale 0.204* 0.00284 -0.0165 0.0107 -0.00219 0.0303 0.00219 -0.0105 (0.114) (0.00684) (0.0126) (0.0382) (0.0391) (0.0333) (0.0341) (0.0252) ad_lemp 0.207*** 0.0659*** 0.00494 -0.00403 0.0977*** 0.0221 -0.0129 -0.00431 (0.0499) (0.0145) (0.00782) (0.0123) (0.0264) (0.0157) (0.0113) (0.0115) Constant 0.738*** 0.357*** 0.417*** 0.223*** 0.901*** 0.948*** 0.980*** 0.953*** (0.0821) (0.0297) (0.0308) (0.0433) (0.0582) (0.0304) (0.0324) (0.0306) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 42 Table S5.S continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf anylink -0.102 0.0691 0.0513 -1.235** 0.00643 -0.660*** -0.000732 -0.202 (0.0897) (0.0494) (0.0742) (0.496) (0.108) (0.196) (0.332) (0.167) same_sec -0.00121 0.0264 0.0160 -0.141 -0.0578 -0.0484 -0.0709 -0.0800 (0.0184) (0.0166) (0.0188) (0.155) (0.0399) (0.0683) (0.134) (0.0878) ldist -0.00298 -0.00462 0.00617 -0.0604 -0.0171 -0.0512** -0.0505 -0.00942 (0.00283) (0.00473) (0.00705) (0.0386) (0.0110) (0.0240) (0.0344) (0.0185) ad_fmage -0.000696 0.00155 -0.00117 -0.00489 0.00113 0.00862 0.00462 0.00605 (0.000812) (0.00156) (0.000985) (0.00781) (0.00216) (0.00651) (0.00746) (0.00458) ad_meduc -0.00144 0.00719 0.00139 0.0458 -0.00477 -0.0333 -0.0203 0.0111 (0.00709) (0.00945) (0.00864) (0.0473) (0.0116) (0.0228) (0.0267) (0.0250) ad_mexper -0.000773 -0.00118 0.000487 -0.00731 0.000772 0.00303 -0.00385 -0.00185 (0.000949) (0.00144) (0.00169) (0.00577) (0.00132) (0.00695) (0.00580) (0.00313) ad_mfemale 0.0156 0.0266 0.0369 0.378 -0.00612 -0.161 0.0592 -0.134 (0.0388) (0.0461) (0.0481) (0.581) (0.0612) (0.134) (0.274) (0.0954) ad_lemp -0.0267*** -0.0310** -0.0111 0.00792 0.0656** 0.0909 0.103 0.0619 (0.00873) (0.0132) (0.0136) (0.110) (0.0259) (0.0709) (0.114) (0.0600) Constant 1.003*** 0.893*** 1.123*** 3.747*** 0.713*** 1.516*** 2.050*** 0.967*** (0.0243) (0.0487) (0.0500) (0.377) (0.0972) (0.288) (0.328) (0.212) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 43 Table S5.S continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD anylink 0.0864** 0.0722** 0.0110 -0.00828 -0.0597 -0.0799 0.0512 0.0621 (0.0349) (0.0303) (0.0340) (0.0396) (0.0560) (0.0536) (0.0344) (0.0842) same_sec -0.0245 -0.0640*** -0.0129 -0.0154 0.00163 -0.00268 0.0191* 0.0140 (0.0151) (0.0192) (0.0138) (0.0155) (0.0117) (0.0113) (0.00991) (0.0231) ldist -0.00957*** -0.0152*** -0.00596*** -0.00414** -0.000221 -0.00282 -0.00225 0.00576 (0.00332) (0.00235) (0.00170) (0.00190) (0.00301) (0.00220) (0.00387) (0.00672) ad_fmage -0.00224*** -0.000535 -0.000661 0.000388 0.000437 -0.000247 0.00110 -0.000489 (0.000512) (0.000917) (0.000545) (0.000691) (0.000803) (0.000622) (0.00115) (0.00121) ad_meduc -0.00634 0.0165 -0.00188 0.00251 0.00207 0.000260 0.00505 -0.00426 (0.00613) (0.0108) (0.00286) (0.00439) (0.00528) (0.00491) (0.00615) (0.00881) ad_mexper 0.000331 -0.00140** 0.000441 0.000105 -0.000965 -0.000796 -0.000884 -0.000554 (0.00129) (0.000569) (0.000601) (0.000560) (0.000597) (0.000688) (0.00107) (0.00186) ad_mfemale 0.240*** 0.130*** 0.0280 0.00235 -0.00547 0.0113 0.0193 0.0661 (0.0528) (0.0431) (0.0244) (0.0203) (0.0229) (0.0259) (0.0292) (0.0552) ad_lemp 0.0820*** 0.0810*** 0.0122 -0.00901 -0.00441 -0.0177*** -0.0185** -0.0165 (0.0218) (0.0227) (0.0101) (0.00658) (0.00852) (0.00582) (0.00929) (0.0154) Constant 0.213*** 0.321*** 0.467*** 0.483*** 0.450*** 0.492*** 0.396*** 1.105*** (0.0364) (0.0353) (0.0214) (0.0201) (0.0258) (0.0201) (0.0383) (0.0531) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 44 Table S6.S: Results for the baseline specification with the geographical distance variable omitted (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES newprod invdum anyrd it pcf_tech suprat htrain ftrain supbuy 0.0135 -0.203 -0.301** 0.520 -0.247 -0.0958 -0.0965 0.370 (0.350) (0.326) (0.122) (0.543) (0.198) (0.0877) (0.263) (0.342) samesup -0.0810 -0.153* -0.145** -0.206* -0.256** -0.00965 0.0552 -0.0183 (0.0731) (0.0784) (0.0617) (0.106) (0.127) (0.0439) (0.0839) (0.0521) samecus 0.0244 0.00572 0.130*** 0.241*** 0.188* -0.0420 0.0868** 0.0212 (0.0224) (0.0343) (0.0347) (0.0601) (0.101) (0.0288) (0.0366) (0.0709) same_sec 0.00462 -0.00131 -0.0138 -0.0963*** -0.0307 0.00222 -0.0344* -0.00870 (0.0101) (0.0134) (0.0185) (0.0299) (0.0311) (0.00627) (0.0189) (0.0159) ad_fmage 0.000122 0.000257 0.000346 -0.00323*** -0.00104 -0.000515* -0.000440 0.000881 (0.000281) (0.000370) (0.00101) (0.00123) (0.00132) (0.000303) (0.000640) (0.000998) ad_meduc 0.00582 0.0406*** 0.00240 0.00948 0.0592** 0.00516 -0.00374 -0.0146*** (0.00632) (0.0144) (0.00895) (0.0167) (0.0230) (0.00339) (0.00561) (0.00464) ad_mexper -2.91e-06 -0.000304 -0.000468 -0.000150 -0.000218 0.000276 -0.00160*** 0.000634 (0.000287) (0.000266) (0.00112) (0.00213) (0.00234) (0.000598) (0.000508) (0.00108) ad_mfemale 0.000903 0.0358* 0.106*** 0.449*** 0.252*** 0.0277 0.0497 0.0849 (0.00631) (0.0190) (0.0394) (0.0901) (0.0758) (0.0176) (0.0375) (0.0543) ad_lemp 0.0231** 0.0525*** 0.0404** 0.169*** 0.199*** 0.0179*** 0.0574*** 0.0780*** (0.00914) (0.0143) (0.0176) (0.0424) (0.0457) (0.00572) (0.0170) (0.0205) Constant 0.462*** 0.371*** 0.281*** 0.418*** 0.738*** 0.304*** 0.346*** 0.0941*** (0.0170) (0.0292) (0.0351) (0.0647) (0.0620) (0.0148) (0.0336) (0.0316) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 45 Table S6.S continued (9) (10) (11) (12) (13) (14) (15) (16) VARIABLES pcf_hk directimport sellcredit subcont pcf_cont reput_1 reput_2 reput_3 supbuy -0.182 -0.448** -0.464** 0.0753 -0.839** -0.683* -0.190 0.559 (0.337) (0.183) (0.188) (0.316) (0.400) (0.367) (0.612) (0.515) samesup 0.0918 -0.169** -0.0125 0.0743 0.00438 -0.134 -0.0645 -0.0357 (0.145) (0.0784) (0.0790) (0.0807) (0.138) (0.165) (0.162) (0.145) samecus 0.177 -0.0501 -0.0126 -0.0268 -0.0431 0.0617 0.0362 -0.104 (0.140) (0.0434) (0.0451) (0.0653) (0.101) (0.0559) (0.0673) (0.116) same_sec -0.0646* -0.0296* -0.00571 -0.00577 -0.0458* -0.00911 -0.0215 -0.00302 (0.0374) (0.0159) (0.0136) (0.0133) (0.0278) (0.0215) (0.0283) (0.0193) ad_fmage 0.000198 0.000121 -6.38e-05 -0.000693 0.000328 -0.000868 0.000371 0.000529 (0.00183) (0.000239) (0.000587) (0.000560) (0.00104) (0.000735) (0.00105) (0.000944) ad_meduc -0.0173 0.0249** 0.00973 -0.00224 0.0317* -0.00201 0.000648 0.00213 (0.0127) (0.0110) (0.00714) (0.00506) (0.0169) (0.00750) (0.00736) (0.00817) ad_mexper -0.000791 -5.26e-05 -0.000161 -0.000457 -0.000571 0.000634 9.60e-05 -0.00114 (0.00186) (0.000320) (0.000731) (0.000637) (0.00118) (0.000804) (0.000920) (0.000702) ad_mfemale 0.208* -0.00264 -0.0223* 0.0163 -0.00738 0.0341 0.00581 -0.0101 (0.115) (0.00679) (0.0131) (0.0374) (0.0386) (0.0335) (0.0338) (0.0253) ad_lemp 0.208*** 0.0651*** 0.00411 -0.00352 0.0972*** 0.0226 -0.0125 -0.00422 (0.0497) (0.0143) (0.00780) (0.0125) (0.0262) (0.0157) (0.0113) (0.0114) Constant 0.715*** 0.390*** 0.454*** 0.189*** 0.934*** 0.922*** 0.956*** 0.951*** (0.0681) (0.0253) (0.0235) (0.0354) (0.0586) (0.0351) (0.0367) (0.0284) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 46 Table S6.S continued (17) (18) (19) (20) (21) (22) (23) (24) VARIABLES reput_4 reput_5 pcf_rep lvademp dlemp growth1 growth3 pcf_perf supbuy 0.289 0.381 0.320 0 -0.477** -0.283 1.783 0 (0.610) (0.633) (0.789) (0) (0.199) (0.542) (2.384) (0) samesup -0.0961 -0.0986 -0.00838 -1.389*** -0.219** -0.723** -0.506 -0.431** (0.141) (0.160) (0.152) (0.399) (0.108) (0.316) (0.323) (0.174) samecus -0.104 0.115** 0.0546 -0.944 0.101 -0.556*** 0.356 0.00138 (0.111) (0.0481) (0.0868) (0.652) (0.147) (0.200) (0.439) (0.218) same_sec -0.00138 0.0274* 0.0165 -0.142 -0.0572 -0.0459 -0.0631 -0.0760 (0.0186) (0.0164) (0.0188) (0.155) (0.0403) (0.0665) (0.133) (0.0872) ad_fmage -0.000701 0.00153 -0.00115 -0.00429 0.00112 0.00889 0.00487 0.00611 (0.000817) (0.00157) (0.000985) (0.00791) (0.00218) (0.00643) (0.00738) (0.00457) ad_meduc -0.00271 0.00521 0.00401 0.0171 -0.0123 -0.0552** -0.0433 0.00651 (0.00692) (0.00983) (0.00894) (0.0516) (0.0148) (0.0231) (0.0271) (0.0237) ad_mexper -0.000789 -0.00120 0.000521 -0.00784 0.000674 0.00244 -0.00454 -0.00195 (0.000943) (0.00143) (0.00171) (0.00579) (0.00128) (0.00701) (0.00586) (0.00312) ad_mfemale 0.0176 0.0292 0.0329 0.411 0.00397 -0.133 0.0827 -0.132 (0.0385) (0.0468) (0.0486) (0.586) (0.0592) (0.132) (0.278) (0.0929) ad_lemp -0.0264*** -0.0307** -0.0117 0.00968 0.0667*** 0.0957 0.107 0.0619 (0.00871) (0.0132) (0.0137) (0.112) (0.0258) (0.0717) (0.114) (0.0600) Constant 0.991*** 0.875*** 1.148*** 3.477*** 0.644*** 1.297*** 1.838*** 0.924*** (0.0237) (0.0448) (0.0443) (0.256) (0.0655) (0.225) (0.261) (0.163) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 47 Table S6.S continued (25) (26) (27) (28) (29) (30) (31) (32) VARIABLES web email reput_D1 reputD_2 reputD_3 reputD_4 reputD_5 pcf_repD supbuy 0.0499 0.469 -0.442** -0.207 0.307 0.291 0.365 0.386 (0.314) (0.354) (0.178) (0.323) (0.262) (0.341) (0.366) (0.657) samesup -0.0880 -0.0882 -0.0567 -0.0331 -0.0404 -0.0517 -0.0596 -0.117 (0.0614) (0.0620) (0.0877) (0.0810) (0.0816) (0.0740) (0.0815) (0.139) samecus 0.140*** 0.127*** 0.0385 0.00420 -0.0672 -0.0862 0.0799** 0.0964 (0.0377) (0.0289) (0.0361) (0.0453) (0.0698) (0.0654) (0.0382) (0.100) same_sec -0.0236 -0.0634*** -0.0124 -0.0153 0.00142 -0.00298 0.0197** 0.0153 (0.0149) (0.0193) (0.0140) (0.0155) (0.0118) (0.0114) (0.00984) (0.0231) ad_fmage -0.00227*** -0.000573 -0.000677 0.000378 0.000438 -0.000250 0.00109 -0.000480 (0.000523) (0.000930) (0.000557) (0.000685) (0.000804) (0.000627) (0.00115) (0.00121) ad_meduc -0.0103* 0.0102 -0.00442 0.000750 0.00197 -0.000945 0.00408 -0.00183 (0.00592) (0.0108) (0.00292) (0.00431) (0.00559) (0.00490) (0.00650) (0.00938) ad_mexper 0.000270 -0.00149** 0.000405 8.14e-05 -0.000965 -0.000811 -0.000898 -0.000524 (0.00127) (0.000584) (0.000603) (0.000562) (0.000597) (0.000681) (0.00107) (0.00188) ad_mfemale 0.245*** 0.140*** 0.0316 0.00488 -0.00525 0.0132 0.0205 0.0620 (0.0523) (0.0431) (0.0245) (0.0202) (0.0230) (0.0257) (0.0299) (0.0557) ad_lemp 0.0827*** 0.0822*** 0.0126 -0.00866 -0.00437 -0.0174*** -0.0184** -0.0172 (0.0220) (0.0230) (0.0101) (0.00660) (0.00849) (0.00583) (0.00925) (0.0155) Constant 0.175*** 0.261*** 0.444*** 0.467*** 0.450*** 0.481*** 0.387*** 1.128*** (0.0331) (0.0361) (0.0241) (0.0224) (0.0230) (0.0185) (0.0337) (0.0481) Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 48