UNDERSTANDING THE FORMATION OF RECIPROCAL HYPERLINKS BETWEEN SELLERS IN AN E-MARKETPLACE

Zhaoran Xu1, Youwei Wang1, Yulin Fang2, Bernard Tan3, Hai Sun1

（1. Department of Information Management and Information Systems, Fudan University; 2. Department of Information Systems, City University of Hong Kong;

3. Department of Information Systems, National University of Singapore）Abstract

Online sellers in the e-marketplace cooperate with each other to increase resources and reduce

transaction costs, both of which are crucial to the success of small businesses. A commonly used IT-

enabled strategy is to ally with other online sellers by exchanging hyperlinks. This paper provides

theoretical guidance to sellers on how to choose partners to improve reciprocity rates in hyperlink

formation. Using the resource-based view and transaction-cost rationale, we examine the effects of

market conditions and seller reputation on reciprocity link formation, using real transaction data from

the largest online marketplace in China. The findings indicate that partners are less likely to exchange

hyperlinks if the two sellers sharing a link are in highly overlapping markets and are geographically

distant from one another, but the two factors weaken each other’s negative effects. The study also

explores the moderating effect of seller reputation, and finds that the negative effect of market

commonality is weakened by seller reputation. The results of this study can be extended to other types

of small business cooperation and are also useful to platform operators for designing mechanisms to

encourage cooperation among online sellers.

Keywords: online seller, hyperlink exchange, market commonality, geographical distance, seller

reputation

1 INTRODUCTION

Over the past decade, more and more sellers have conducted their business on e-marketplace

platforms, such as eBay, Amazon and Taobao (an e-marketplace owned by Chinese e-commerce giant

Alibaba). The large number of sellers leads to intensified competition in the e-marketplace. For

example, by the end of 2013, more than eight million active sellers were competing on Taobao. The

1

majority of these are small business owners [1-3] with limited resources and market presence, and are

vulnerable to environmental forces [4]. According to a recent study [5], in 2010, 38.4% of the sellers

in apparel shut down their businesses within six months of establishing their stores on Taobao. To

survive the competition, online sellers cooperate with each other and undertake the same alliance

activities as traditional bricks-and-mortar firms, such as co-branding and co-marketing. In addition,

they deploy distinct strategies which rely on Internet infrastructures in an e-marketplaces [6]. The

strategies involve forming alliances with other online sellers by exchanging website hyperlinks.

Hyperlink is an important feature of World Wide Web (WWW) [7]. It enables visitors to jump

from one webpage to another by clicking on links embedded in the hypertexts. Because there are

numerous webpages with similar information contents or services, these webpages have to attract

visitors’ attention by unique service offerings and techniques. Exchanging hyperlinks is a technique

commonly utilized by Internet web services (more broadly those techniques fall into the category of

search engine optimization). It is well known that exchanging hyperlinks can lift the rankings of the

website in major search engines such as Google [8, 9]. This is because websites with large number of

hyperlinks will be given higher priority in search engines’ ranking algorithm. In e-marketplaces, each

seller manages an online store and the key motivation of exchanging hyperlinks is simply to attract

and exchange web browsing traffic because this can lead to more online purchases [10, 11]. Taking

Taobao as an example, a focal seller links to another seller (link seller) by putting a hyperlink on its

store front (typically on the left sidebar). This is an outgoing link for the focal seller. When customers

browse the focal seller’s shop, they may visit the link seller by clicking on the hyperlink. The link

seller can establish a hyperlink back the focal seller, which would be an incoming link or reciprocal

link for the focal seller. When this happens, the two sellers exchange hyperlinks successfully.

Prior studies have shown that exchanging hyperlinks can help online sellers to achieve success

by growing the customer base [12], enhancing customer trust [13, 14], and eventually improving

sellers’ competitive advantage [15]. Hyperlinks can improve the seller’s revenue and profit but the

effects only lies in the incoming links [16]. Incoming links make it easier for customers to discover

the seller whereas outgoing links can reduce customer traffic and undermine the performance of the

2

seller [17]. Thus, online sellers with more incoming hyperlinks and fewer outgoing links tend to

perform better [10]. If hyperlink exchange fails and the target seller does not reciprocate by placing a

hyperlink back to the focal seller on its online store, the focal seller may suffer the loss of customers

from the outgoing links initiated.

Although it is obvious that having more incoming links and less outgoing links are often good

for focal sellers [10, 17], it is not clear how such goals can be achieved. This is because each seller

can only control its outgoing links but not the incoming links. What usually happens is that sellers

initiate hyperlinks to other sellers in the hope that some of them will reciprocate by linking back.

However, no prior study has examined how sellers may get more reciprocal links in e-marketplaces.

To fill this important gap in knowledge, this paper investigates the following research question: What

types of partner sellers are more likely to exchange reciprocal hyperlinks in an online marketplace?

To address this question, we build on the literature of alliance formation in the field of strategy

by conceptualizing hyperlink exchanges as exchanges of customer resources. In the strategy literature,

there are two kinds of antecedents in alliance formation. One is the individual characteristics of the

partner firm, such as status or age [18]; and the other is the dyad characteristics between the alliance

partners, such as market commonality [19]. These factors can be better understood using theoretical

perspectives like resource-based view and transaction-cost rationale [20].

Utilizing a dataset collected from Taobao, this study proposes and empirically validates the main

effects and the interaction of market commonality and geographical distance (two market conditions

between the partner sellers), and the moderating effects of reputation (a key characteristic of online

sellers). Market commonality is the extent of overlap in the market segments of alliance partners.

Geographical distance is the physical distance between alliance partners. Specifically, this study finds

that sellers are less likely to exchange hyperlinks if they operate in highly overlapping markets and

are geographically distant from each other, but the two factors weaken each other’s negative effects.

In addition, the negative effects of market commonality are ameliorated by seller reputation. These

findings have important theoretical and empirical implications. Theoretically, these findings extend

the generalizability of related theories from traditional bricks-and-mortar industries to e-marketplaces.

3

Practically, these findings help online sellers to make better decisions about how to cooperate with

other sellers and help platform operators make better decisions about how to enhance collaborations

in e-marketplaces.

2 THEORETICAL BACKGROUND

2.1 Resource-based view and transaction-cost rationale on alliances in e-marketplaces

The strategy literature offers theoretical guidance as to why firms enter into alliances. The

resource-based view argues that firms should possess resources that are rare, valuable, imperfectly

mobile, and non-substitunntable to achieve competitive advantage [21]. Thus, developing and

leveraging resources is a key driver for alliance formation [22]. The transaction-cost rationale

recommends that firms form alliances to minimize their fixed and continual transaction costs [23].

The literature suggests that firms tend to put resource concerns ahead of cost concerns when deciding

whether or not to engage in alliances [24] but considerations of economic costs can influence inter-

firm relationships [25]. Therefore, this study supplements the resource-based view with the

transaction-cost rationale in discuss alliance formation in e-marketplaces.

Alliance formation may differ depending on industry structure and competitive situation [26].

Given that this study is about e-marketplaces, the characteristics of the online environment have to be

considered when examining resource maximization and cost minimization in alliance formation.

Online sellers in e-marketplaces are small businesses which lack resources [27]. The resource-based

view suggests that, to survive, they need access to resources of alliance partners, especially customer

resources [6]. Indeed, it is critical for these small business to expand their customer base [28]. But it is

difficult for these small business to retain customers because online retailing allows customers to

transact with many different sellers [29]. Thus, in e-marketplaces, customer resources are vital but

easy to come and go. Sellers engage in alliances to cooperate and compete for customer resources at

the same time. The cooperation level (or competition level) depends on market conditions, such as

market commonality and geographical distance [30].

Another important resource for online sellers is their reputation. Customers attach considerable

importance to seller reputation in e-marketplaces [14] and seller reputation positively affects revenue 4

and total sales [31, 32]. Online sellers should consider the reputation of prospective partners when

forming alliances [33]. A good reputation improves the bbenefits of forming alliances [34]. Thus, this

study also examines the role of seller reputation in the exchange of hyperlinks.

Unlike production costs, transaction costs are incurred in organizing information, coordinating

behaviour, monitoring transactions, and safeguarding interests [35]. Online sellers use e-marketplaces

to select alliance partners and execute transactions, which lower search costs compared to traditional

approaches [36]. However, coordination costs are higher in e-marketplaces for two reasons. First, the

online environment is complex in the sense that partners can leverage on environmental uncertainty

and information asymmetry to be more opportunistic [37]. Second, competition tends to be fiercer in

e-marketplace and this creates conflicts in alliances, which increase coordination costs [38]. Thus, e-

marketplace alliances incur lower search costs and higher coordination costs compared to alliances in

bricks-and-mortar industries.

2.2 Market conditions and alliance formation in e-marketplaces

This study examines two market conditions: market commonality and geographical distance.

Market commonality is commonly known as “the degree of presence that a competitor manifests in

the markets it overlaps with the focal firm” [39]. When firms operate in overlapping markets, they

have higher market commonality and more collective strength compared to partner firms operating in

distinct industries [19].

Research has shown that firms with high market commonality are more likely to form alliances

because they share similar resources. This makes it easier for them to achieve economy of scale by

aggregating similar resources through alliances [19, 40]. Although there may be higher search costs

associated with screening prospective alliance partners [41], firms in the same markets are usually

quite familiar with each other and share common market knowledge [42]. This familiarity decreases

the costs of searching for prospective partners and makes the process less time-consuming [43]. Thus,

firms with high market commonality are likely to form alliances.

5

But research has also suggested that firms with low market commonality tend to cooperate [18,

22, 43]. Firms that are in different markets can also form alliances to develop new and complementary

resources [44]. In this situation, firms tend to cooperate when their partners have strengths that can

make up for their weaknesses [18]. They also cooperate to exploit business opportunities from their

partners in different markets [40]. Besides, firms with low market commonality are less likely to find

themselves competing in the same market [45]. Overall, past studies have indicated that market

commonality is important to alliance formation. However, given that the business environment in e-

marketplaces is different, it is important to re-examine this issue of alliance formation in the context

of e-marketplaces.

Geographical distance is commonly known as the spatial or physical distance between economic

actors, such as alliance partners [42]. Geographical distance is well documented in the international

business [46, 47] and industry cluster literature [48, 49]. In alliance formation, remote partners can

help firms reach out to different, diverse, and non-redundant resources that co-located firms cannot

[47]. Because it can be difficult for firms to reach customers in distant markets, remote partners can

provide access to these new markets, thus making alliance formation more likely [6, 12].

But having co-located partners can lowers cost associated with distance (such as communication

costs [50]) and search costs for identifying useful competences [51]. In addition, geographically

distant partners are more likely to behave opportunistically, which can undermine trust in the alliance

[42]. Thus, geographically distant firms tend to prefer deeper inter-partner relationships (e.g ., joint

venture or merger and acquisition) over alliances [52].

Studies have also reported that geographic distance is not particularly relevant to alliances under

heterogeneous industry characteristics [49]. Furthermore, advances in communications technologies

may weaken the role of geographical distance [53]. Therefore, it is useful to re-examine the role of

geographical location in an online environment. The effects of the two market conditions may be

interdependent. For example, partners in different markets (i.e., low market commonality) can help

each other access new markets [40] but geographical distance can make it more difficult and costly to

6

form alliances with partners [52]. Therefore, this study explores the interaction effects of market

commonality and geographical distance.

2.3 Firm reputation and alliance formation in e-marketplaces

Compared to bricks-and-mortar shops, customers tend to perceive more risks in transacting with

sellers in e-marketplaces. Thus, the reputation of online sellers plays a significant role in attracting

customers [54]. The concept of reputation is different from that of status but prior studies have often

confused the two [55]. Status represents the order or rank of the firm in a market whereas reputation

indicates the quality of the firm as determined by its previous actions and this is a market signal when

information is asymmetric [56].

The resource-based view suggests that reputation is a valuable resource to the firm [18] and can

affect the exchange of resources because high reputation enhances trust in inter-firm relationships

[57]. The transaction-cost rationale suggests that opportunism is a major concern in alliances and so a

good reputation helps to reduce cooperation costs by enhancing trust. For example, negotiation costs

are reduced when reputation improves trust between partner firms [58]. Thus, reputation can enhance

the competitive advantage of partners and has a positive effect on alliance formation [59]. This study

focuses on the moderating role of reputation on the effects of market conditions. For example, the

good reputation of the potential partner may reduce the opportunistic risk arising from geographical

distance.

3 HYPOTHESES DEVELOPMENT

3.1 Main effects of market conditions

Market commonality is commonly measured by sellers’ product categories [18, 22, 43]. A single

online seller can have multiple product categories. Overlapping categories between two partner sellers

increase market commonality.

Based on the resource-based view, a major benefit of forming alliance among online sellers is

that this increases the customer base for all sellers in the alliance [12]. As discussed above, research

has shown that both high and low market commonality between alliance partners can improve their

7

access to customers [22, 40, 43]. Traditionally, firms in overlapping markets form alliances to share

similar resources and compete with outsiders [19, 60]. But in e-marketplaces, there can be numerous

sellers in each product category that alliances are often loose and informal [28]. Moreover, sellers

with high market commonality are competitors [61] because it is easy for customers to move from one

seller to another. Thus, plenty of online traffic flows through hyperlinks connecting sellers. These

links provide opportunities for customers to leave [10], undermining the long-term interests of sellers

[29]. In this situation, partner sellers with many overlapping products (i.e., high market commonality)

are at greater risk of losing customers to their partners. On the contrary, partner sellers with few

overlapping products (i.e., low market commonality) can help each other by facilitating customer

purchase from both sellers through their hyperlinks.

The transaction-cost rationale suggests that search costs associated with finding the right partner

are lower when partners have high market commonality in bricks-and-mortar industries [40].

However, this cost advantage is not as salient in e-marketplaces because search costs are low,

regardless of market commonality [37]. Thus, the lower search costs that come with high market

commonality are not applicable in e-marketplaces. On the contrary, competition arising from high

market commonality brings more costs, such as the need to offer extra services to retain customers

[4]. Therefore, market commonality has a negative effect on hyperlink exchange (i.e., alliance

formation) in e-marketplaces:

H1: Partner sellers are more likely to exchange hyperlinks if market commonality among the

two sellers (in terms of product categories) is lower.

Every online seller has a physical location (their inventory site). In this study, geographical

distance is measured by absolute distance between a pair of partner sellers based on their respective

physical location [62]. Taking the resource-based view, prior research suggests that long-distance

alliances can provide access to new customers in remote markets and such access would be difficult

without the alliances [47, 48]. However, in e-marketplaces, market access is not restricted by location

because sellers can easily overcome barriers of geographical distance to reach far-flung customers.

Prior research also suggests that having nearby partners can help sellers increase their customer base

8

because customers prefer the convenience of visiting retail clusters [63]. Having nearby partners bring

in other resources, such as access to common specialised suppliers and skilled labour pools [48]. In e-

marketplaces, closely-located partnerships can still increase customer base because customers may

prefer to make multiple purchases from nearby sellers for reasons such as lower shipping costs and

shorter delivery schedule for these purchases collectively [63, 64]. Thus, lower geographical distance

can still contribute to alliance formation.

Past research suggests that it can be more costly to monitor and coordinate geographically distant

partnerships because partners are more likely to behave opportunistically [42, 46]. This opportunism

associated with environmental uncertainty [65] tends to be more pronounced in e-marketplaces due to

the volatility [37]. For example, a seller may pursue an unfair competitive strategy (e.g., offering deep

customer discounts) that undercuts its partner seller. Geographical distance can also undermine trust

between partner sellers. This is because sellers tend to be less familiar with the service and product

quality of distant partners and risk associating with disreputable distant partners [41]. Therefore,

geographical distance has a negative effect on hyperlink exchange (i.e., alliance formation) in e-

marketplaces:

H2: Partner sellers are more likely to exchange hyperlinks if geographical distance among the

two sellers is shorter.

3.2 Interaction effects of market conditions

The negative effects of market commonality on alliance formation lie in the loss of customer

resources when partner sellers’ markets overlap [40, 61]. However, geographical distance can reduce

this loss of customer resources because, if the partner sellers are far apart, there is a lower chance that

customers will move from one seller to its partner seller through hyperlinks, given that customers still

prefer nearby sellers for reasons such as lower shipping costs and shorter delivery schedule [64]. In

this way, geographic distance between partner sellers can lessen their competition in overlapping

markets and mitigate the associated costs. Therefore, geographical distance can weaken the negative

effects of market commonality and so sellers far away from each other are more likely to form

alliances even if they are in highly overlapping markets.

9

Based on the transaction-cost rationale, sellers tend to avoid long-distance partnerships because

these partnerships can incur higher monitoring and coordination costs as well as increase the risk of

associating with low-quality partners [46]. To avoid engaging with an opportunistic partner, sellers

need more market information to assess a prospective partner’s likely future behaviour [66], and

sellers with high market commonality tend to be more familiar with each other . This can reduce the

coordination costs of alliances that arises from geographical distance [42]. Thus, market commonality

can weaken the negative effects of geographical distance and so sellers in highly overlapping markets

are more likely to form alliances even if they are further away from each other.

H3: Market commonality and geographical distance have interaction effects such that they

weaken each other’s negative effects on the likelihood of exchanging hyperlinks.

3.3 Moderating effects of reputation

Reputation plays a very important role in e-marketplaces because of the uncertainty of the online

environment [54]. Sellers can improve their reputation and thereby attract more customers by forming

alliances with partners with good reputation [34]. The benefits of seller reputation in e-marketplaces

are well-known [59]. Going beyond past studies, this study examines the moderating effects of seller

reputation on market conditions (market commonality and geographical distance).

Partners with high market commonality tend to have intensive competition that causes a loss of

customers [40, 61] so high market commonality has negative effects on alliance formation. This

situation is aggravated when a partner seller has a better reputation that enables it to attract more

customers [18, 22]. Seller reputation is easily accessible in e-marketplaces due to ready availability of

customer feedback. Therefore, high reputation of a partner seller strengthens the negative effects of

market commonality on alliance formation.

H4: Higher reputation of a partner seller strengthens the negative effects of market commonality

on the likelihood of exchanging hyperlinks.

Geographical distance increases the monitoring and coordination costs as well as the risk of

associating with opportunistic partners [46]. To avoid engaging in less productive alliances, sellers

can assess a prospective partner’s likely future behaviour using its track record of past behaviour [66]. 10

This practice may be particularly important for partners that are far away from each other given that

they tend to have less information about each other [41]. However, such costs and risk arising from

geographical distance is alleviated when a partner seller has a better reputation because such partner

sellers tend to collaborate with other sellers so as to improve the overall customer purchase experience

[59]. Therefore, high reputation of a partner seller weakens the negative effects of geographical

distance on alliance formation.

H5: Higher reputation of a partner seller weakens the negative effects of geographical distance

on the likelihood of exchanging hyperlinks.

4 METHODOLOGY

4.1 Dataset

Our dataset was obtained from Taobao, the largest e-marketplace in China. Taobao labels

hyperlink exchanges as “friendship links”. Apparel sellers were examined in this study because

apparel has been the best-selling product on Taobao. In 2011, there were 375,000 apparel sellers

accounting for about 20% of Taobao’s 68 seller categories. Competition was very intense in such a

market and therefore sellers would be motivated to collaborate in various ways, including exchanging

friendship links.

Our data was collected in January 2011. The data was analysed at the dyad level, with the link

itself as the unit of analysis. Each link involved two sellers: a focal seller and a link seller. There were

two scenarios: (1) a focal seller sent an outgoing link to a link seller and the link seller decided

whether to send a reciprocal link back to the focal seller; and (2) a focal seller received an incoming

link from a link seller and the focal seller decided whether to send a reciprocal link back to the link

seller. Information on both scenarios was included in our dataset.

A total of 1,000 apparel sellers were randomly chosen as our focal sellers. Information was

collected for friendship links of focal sellers (including outgoing hyperlinks and incoming hyperlinks)

which met the following requirements: (1) the hyperlinks should be voluntary; (2) the hyperlinks

should connect sellers with different identities (seller IDs); (3) two sellers connected by the hyperlinks

should not share the same mailing address; and (4) the reciprocal hyperlinks were created within 20

11

days of the original hyperlinks (because the reciprocal hyperlinks might not be triggered by the

original hyperlinks if the time interval was too long). Our dataset comprised 924 outgoing hyperlinks

and 701 incoming hyperlinks for 323 focal sellers in the apparel industry. Results of t-tests revealed

that the difference between stock levels of sellers in our dataset and sellers in the population was not

significant.

4.2 Measurement

4.2.1 Dependent variable

When sellers received incoming hyperlinks as a consequence of sending outgoing hyperlinks to

other sellers, these would be deemed reciprocal hyperlinks. Our analyses examined whether each

hyperlinks was reciprocated and measured this as a binary variable, consistent with other studies on

alliance partner selection [6, 22, 67]. Each outgoing hyperlink sent by a focal seller would be coded

“1” if the focal seller received a reciprocal hyperlink after and “0” otherwise. Each incoming

hyperlink received by a focal seller would be coded “1” if the focal seller sent a reciprocal link after

and “0” otherwise.

4.2.2 Independent variables

Market commonality. Product category was used to measure the market commonality of sellers

[18, 22, 43]. Taobao classifies the products on its platform into 68 categories and codes sellers into

these categories based on major products they sell (using an algorithm that considers recently sold

products). Each seller would have one or two major product categories. Market commonality was

computed as follows:

Market commonality=¿ Focalselle r ' scategories ∩ Link selle r ' scategories∨ ¿¿ Focal seller ' scategories∪ Link selle r ' scategories∨¿¿

¿

For example, if a focal seller traded apparel and bags and a link seller traded apparel and

cosmetics, then they would have one common category (apparel) out of three categories (apparel,

bags, and cosmetics) and their market commonality would be 0.33. The lowest possible market

commonality of two sellers would be 0 (if they had no common categories) and the highest possible

market commonality of two sellers would be 1 (if all their categories were identical).

12

Geographical distance. Geographical distance was computed based on absolute distance between

the registered cities of the focal seller and the link seller [62]. This was measured in units of 1,000

kilometres to enlarge the coefficient estimates.

Reputation. After completing a purchase, customers in Taobao would rate the product quality of

the online seller. Seller reputation was measured by the ratio of positive ratings.

4.2.3 Control variables

Factors that might influence the likelihood of obtaining reciprocity hyperlinks were included as

control variables. A control variable was the seller’s business tenure (measured by number of months

since the seller’s online store was established). Established sellers might have more legitimacy in the

e-marketplace [68-70]. Thus, it might be beneficial for a new seller to collaborate with an established

seller but not the other way round. Another control variable was the numbers of existing friendship

links. This reflected the inclination of sellers to exchange hyperlinks and might affect their decisions.

Another control variable was whether sellers participated in the consumer rights safeguarding plan

(CRSP) tend to attract more customers.1 Sellers might prefer to exchange hyperlinks with partners that

participated in CRSP. Table 1 shows the measurements of all the variables.

Table 1. Variable measurements

Variables Measurements1 Reciprocal link 1 if focal seller or link seller received a reciprocal link, 0 otherwise

2 Market commonality ¿ Focal seller ' s categories∩ Link seller ' scategories∨ ¿¿ Focal selle r ' s categories∪Link selle r ' scategories∨¿¿

¿

3 Geographical distance Absolute distance between the registered cities of the focal seller and the link seller (in 1,000 kilometres)

4 Seller reputation Online seller’s ratio of positive ratings5 Seller tenure Online seller’s store age (in months).6 Friendship links Online seller’s number of existing friendship links7 Assurance mechanism 1 if online seller participated in CRSP, 0 otherwise

4.3 Estimation method

The dependent variable was binary in nature. Correspondingly, a logit regression was used for

estimation. The outcome variable p was the probability of receiving reciprocal hyperlinks.

p=Pr ( y ij=1)

1 Sellers participating in CRSP placed a deposit into their platform accounts. This deposit would be used to reimburse customers in instances where sellers were found to be at fault in transaction disputes. Participating sellers would be marked with a conspicuous icon on their online storefronts so customers could easily identify such sellers.

13

y ij={1, if reciprocal hyperlink recived0 , otherwise .

Because each seller could create multiple friendship links, the dataset was an unbalanced panel

data (with the seller as the panel variable and the hyperlink as the time variable). Random effects (RE)

model was used because RE parameter estimates could be applied to a random sample of the entire

population. Fixed effects (FE) model could not be used because FE parameter estimation required

variances in the dependent variable for the same seller. Therefore, if FE model was used, 225 sellers

out of the 327 sellers would have to be dropped and the sample size would then be too small. Thus,

we used RE model and the logit model specification as follows:

logit ( p )=ln ( p1−p

)=α i+x ij β+εij , α i IID ( α , σ2 ) ,

where x ij was the vector of independent variables and αi was a random variable distributed

independently of the regressors. The model was estimated through a maximum likelihood procedure

[71].

5 DATA ANALYSES

We analysed two situations in this study. In the first situation, focal sellers initiated hyperlink to

link sellers and it would be up to link sellers to decide whether to follow up with reciprocal hyperlinks

(focal seller → link seller). In the second situation, focal sellers received hyperlink invitations from

link sellers and it would be up to focal sellers to decide whether to follow up with reciprocal

hyperlinks (link seller → focal seller). Tables 2 and 3 summarize the descriptive statistics for the two

situations and show the Pearson correlation coefficients between the variables. Market commonality,

geographical distance, and seller reputation were significantly correlated to the dependent variable.

Multicollinearity was not an issue because the maximum variance inflation factor was 1.06 and 1.02

for apparel sellers initiating and receiving hyperlinks respectively, far below the threshold of 10 [72].

A hierarchical moderated regression was used to test the hypotheses. The variables were mean-

centered [73]. Models 1 to 5 in Table 4 present the regression results for the situation where focal

sellers first sent hyperlink invitations to link sellers. Market commonality and geographical distance

14

had negative main effects on the likelihood of obtaining reciprocal hyperlinks (see Model 1), thus

supporting H1 and H2. Market commonality and geographical distance had a positive interaction on

the likelihood of obtaining reciprocal hyperlinks (Model 2). The results suggested that market

commonality and geographical distance weakened the negative effects of each other, thus supporting

H3. Market commonality and seller reputation had a negative interaction on the likelihood of

obtaining reciprocal hyperlinks (Model 3). The results suggested that seller reputation strengthened

the negative effects of market commonality, thus supporting H4. Geographical distance and seller

reputation had no interaction on the likelihood of obtaining reciprocal hyperlinks (Model 4). The

results suggested that seller reputation did not change the negative effects of geographical distance,

thus rejecting H5. These results remained robust when all the interaction terms were examined

together (Model 5). Models 6 to 10 in Table 4 present the regression results for the situation where

focal sellers first received hyperlink invitations from link sellers. The results were consistent with

those in Models 1 to 5, thereby providing further evidence of robustness of these results.

15

Table 2. Descriptive statistics and Pearson correlation matrix for apparel industry (focal seller → link seller)Variable Mean S.D. 1 2 3 4 5 6 7

1 Reciprocal link 0.26 0.44 1.002 Market commonality 0.17 0.27 -0.12*** 1.003 Geographical distance 1.26 1.31 -0.09*** 0.03 1.004 Focal seller reputation 0.99 0.01 0.10*** 0.08** 0.07** 1.005 Focal seller tenure 13.54 15.68 0.04 -0.06* -0.09*** 0.09*** 1.006 Focal seller friendship links 13.08 14.00 0.29*** -0.09*** -0.01 0.04 0.03 1.007 Focal seller assurance mechanism 0.71 0.45 0.09*** -0.14*** 0.02 -0.10*** 0.21*** 0.02 1.00N = 924; * p < 0.1, ** p < 0.05, *** p < 0.01

Table 3. Descriptive statistics and Pearson correlation matrix for apparel industry (link seller → focal seller)Variable Mean S.D. 1 2 3 4 5 6 7

1 Reciprocal link 0.31 0.46 1.002 Market commonality 0.13 0.25 -0.08** 1.003 Geographical distance 1.07 1.09 -0.12*** -0.04 1.004 Link seller reputation 0.99 0.01 0.07** -0.03 -0.03 1.005 Link seller tenure 14.36 15.45 0.02 -0.07* -0.01 -0.01 1.006 Link seller friendship links 19.27 23.38 -0.11*** -0.02 -0.11*** 0.03 -0.11*** 1.007 Link seller assurance mechanism 0.62 0.49 0.08* -0.10** -0.02 -0.01 0.07** 0.01 1.00

N = 701; * p < 0.1, ** p < 0.05, *** p < 0.01

16

Table 4. Regression results

Focal seller → Link seller Link seller → Focal seller(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Market commonality -0.966** -0.890** -0.844** -0.965** -0.767* -1.210** -0.990* -1.290** -1.206** -1.116**

(-2.31) (-2.10) (-2.00) (-2.30) (-1.79) (-2.39) (-1.94) (-2.49) (-2.38) (-2.15)

Geographical distance -0.183** -0.183** -0.183** -0.203** -0.201** -0.457*** -0.428*** -0.455*** -0.453*** -0.428***

(-2.11) (-2.11) (-2.13) (-2.18) (-2.17) (-4.02) (-3.71) (-4.02) (-3.94) (-3.71)

Seller reputation28.37** 28.66** 26.38** 31.29** 28.92** 36.79* 36.32* 35.07 37.13* 36.38*

(2.22) (2.22) (2.09) (2.26) (2.09) (1.72) (1.69) (1.61) (1.74) (1.66)Market commonality* Geographical distance

0.684** 0.666** 0.912* 0.888*

(2.17) (2.12) (1.81) (1.76)Market commonality*Seller reputation

-77.83* -76.50* -174.3** -165.0**

(-1.87) (-1.79) (-2.08) (-2.03)Geographical distance*Seller reputation

6.716 5.843 5.135 0.0752(0.60) (0.52) (-0.24) (0.00)

Seller tenure -0.0104 -0.0112 -0.0105 -0.0104 -0.0113 -0.0020 -0.0021 -0.0018 -0.00200 -0.00192(-1.28) (-1.37) (-1.31) (-1.28) (-1.40) (-0.28) (-0.29) (-0.25) (-0.28) (-0.27)

Seller friendship links 0.0500*** 0.0503*** 0.0498*** 0.0501*** 0.0502*** -0.0028 -0.0021 -0.0031 -0.00274 -0.00255(7.13) (7.14) (7.14) (7.14) (7.15) (-0.28) (-0.21) (-0.31) (-0.27) (-0.25)

Seller assurancemechanism

0.699** 0.704** 0.720** 0.704** 0.726** 0.288 0.304 0.296 0.287 0.309(2.32) (2.32) (2.42) (2.34) (2.43) (1.24) (1.30) (1.28) (1.23) (1.32)

Intercept -2.265*** -2.274*** -2.267*** -2.283*** -2.287*** -0.720*** -0.721*** -0.740*** -0.720** -0.738**

(-7.63) (-7.62) (-7.73) (-7.64) (-7.70) (-2.63) (-2.62) (-2.71) (-2.63) (-2.68)N 924 924 924 924 924 701 701 701 701 701Log likelihood -454.84 -452.56 -453.16 -454.65 -450.86 -381.12 -379.57 -379.11 -381.09 -377.60

t statistics in parentheses;* p < 0.1, ** p < 0.05, *** p < 0.01

17

6 DISCUSSION AND IMPLICATIONS

6.1 Discussion of results

This study examines the effects of two market conditions (market commonality and geographic

distance) as well as the moderating effects of seller reputation on hyperlink exchange between sellers

in an e-marketplace. The results that support H1 suggest that two sellers are more likely to exchange

hyperlinks if they have low market overlap. In e-marketplaces, customers can move among sellers

with low switching costs [29]. Thus, market commonality creates substantial competitive tension for

alliance partners and online sellers who rely on the same customer resources are likely to become

direct competitors. However, the results that support H3 and H4 delineate the theoretical boundaries

of the effects of market commonality. In particular, the results that support H3 show that market

commonality has a positive effect on two sellers exchanging hyperlinks if they are located far away

from each other (see Figure 1), and the results that support H4 show that market commonality has a

positive effect on two sellers exchanging hyperlinks if the seller initiating the hyperlink has poor

reputation (see Figure 2). Consistent with past findings on firm alliances [19, 40], these results

demonstrate that long distance between partners or poor partner reputation can reduce competition in

high commonality situations because sellers have a better chance of attracting the customers of their

partners.

Figure 1. Geographic distance moderating the effects of market commonality

18

Figure 2. Seller reputation moderating the effects of market commonality

The results that support H2 suggest that two sellers are more likely to exchange hyperlinks if

they are geographically close to each other. In e-marketplaces, collaborating with nearby partners

allow sellers to offer customers lower shipping costs and shorter delivery schedule for their purchases

from partner sellers collectively [63, 64]. However, the results that support H3 show the theoretical

boundaries of the effects of geographic distance. Specifically, geographic distance has a positive

effect on two sellers exchanging hyperlinks if they have high market overlap (see Figure 3). The

results that reject H5 show that the effects of geographic distance are not moderated by seller

reputation. Even though high reputation of prospective partners can reduce the costs and risk of sellers

collaborating with these prospective partners, such reputation information may have a weak influence

[74] considering that the e-marketplace allows sellers to gather reasonably accurate information about

prospective partners located anywhere (nearby or far away).

19

Figure 3. Market commonality moderating the effects of geographic distance

6.2 Theoretical implications

Prior studies have explained the outcomes of hyperlink exchanges among sellers [10, 14, 17] but

these studies have not provided theoretically-grounded explanations for the factors (or combination of

factors) that lead to hyperlink exchanges. Because hyperlink exchanges are critical for the sustenance

of the community of sellers in e-marketplaces [10], it is important to understand the factors that

promote or hinder such alliances among sellers. In this regard, this study goes beyond past studies in

determining the antecedents of hyperlink exchanges among sellers in e-marketplaces.

This study investigates alliance formation outside the traditional bricks-and-mortar context in the

e-marketplace context. The results re-assess the generalizability of theories (such as resource-based

view and transaction-cost rationale) in the new context. Extending past findings about the effects of

market commonality on alliance formation in the traditional brick-and-mortar context [22, 40, 43],

this study re-examines these effects of market commonality in the e-marketplace context and show

that the effects arising from market commonality can be moderated by geographic distance and seller

reputation. While past studies have discussed the paradox due to geographic distance [46, 47, 48, 49]

and suggest that geographic distance may be less important in e-marketplaces [53], this study shows

that geographic distance still affects the decision of sellers in e-marketplaces in terms of whether they

should exchange hyperlinks with prospective partners. Past studies on strategy argue that geographic

distance undermines collaboration [40, 43] but this study demonstrates that geographic distance can

facilitate collaboration in the context of e-marketplaces.20

The moderating effects uncovered in this study enrich our theoretical understanding about how

sellers make decisions on the exchange of hyperlinks in the e-marketplace context. The negative

effects of market commonality can become positive effects when partners are located far away from

each other or when the seller initiating the hyperlink has poor reputation. In these conditions, the

drawbacks of collaboration (i.e., mutual competition) are lessened while the benefits of collaboration

(i.e., increased market) are heightened. Similarly, the negative effects of geographical distance can

become positive effects when partners have high market overlap. In this situation, collaboration

barriers arising from coordination and monitoring costs are reduced.

6.3 Managerial implications

The results of this study have managerial implications for sellers and platform operators in e-

marketplaces. In spite of the ubiquity of hyperlinks in e-marketplaces, sellers have little guidance as to

how they may benefit from hyperlink exchanges in the past. This study guides sellers in their

decisions about when to exchange hyperlinks. Specifically, when sellers are considering prospective

partners with high market overlap, they are likely to be better off exchanging hyperlinks with those

that are located far away or those with poor reputation. But when sellers are considering prospective

partners located far away, they are likely to be better off exchanging hyperlinks with those that they

have high market overlap. For sellers with weak reputation (e.g., new sellers) that have a greater need

for customer resources [18], they are more likely to succeed in exchanging hyperlinks with partners

with high market overlap.

Sellers in e-marketplaces have more difficulty differentiating their products [75] but they also

have more opportunities to cooperate in the dynamic environment [45]. Because these sellers must

often collaborate for mutual survival, this study offers guidance to sellers for their decisions on when

to collaborate. Specifically, market conditions (e.g., market commonality and geographical distance)

as well as partner characteristics (e.g., reputation) should factor into such decisions. The results of this

study offer insights into the trade-off that occurs when different combinations of market conditions

and partner characteristics come into the decision process. The payoff in making the right trade-off

can be significant for the numerous sellers in e-marketplaces.

21

Platform operators can also leverage on the results of this study to improve collaboration among

numerous sellers. For example, there are many seller associations on Taobao and some are organized

by geography (e.g., Association of Shanghai Sellers). At present, the key activities of such seller

associations are limited to sharing experience among members. Given that it can be beneficial for

sellers located near each other to collaborate under some conditions, the platform operator can

encourage collaboration among relevant members of such seller associations based on the results of

this study. Other seller associations on Taobao are organized by industry (e.g., Association of Apparel

Sellers). Again, the key activities of such seller associations are limited to sharing experience among

members. Given that it can be beneficial for sellers with high market overlap to collaborate under

some conditions, the platform operator can encourage collaboration among relevant members of such

seller associations based on the results of this study. Mutually productive collaboration is instrumental

for the survival of sellers in e-marketplaces.

6.4 Limitations and future research

This study has several limitations. First, the measurement of market commonality is limited in

accuracy. Market commonality is based on product type as classified by Taobao. The dataset only has

the first two major product categories of each seller. In practice, sellers may be involved in more than

two major product categories. Also, information on the percentage of sales for each seller in each

product category is not available. Future research leveraging on a richer dataset can shed light on

additional factors that may affect the decisions of sellers in exchanging hyperlinks.

Second, future studies can investigate other forms of alliance which incur higher switching costs

than exchanging hyperlinks. Such forms of alliance include sellers involved in jointly sourcing and

sellers sharing inventory. Also, sellers in e-marketplaces can customize and personalize their product

and service offerings, thereby increasing switching costs for customers [6]. Higher switching costs

may then affect the decisions of sellers on hyperlink exchange and alliance formation [76]. More

research is needed on these topics.

Third, e-marketplace alliances are loose and easy to break if all partners do not benefit equally

[28]. This study is confined to alliances which offer positive reciprocity (i.e., how sellers enter into 22

collaboration). Future research can examine when and how sellers may end their collaboration (e.g.,

removing existing hyperlinks with partners).

7 CONCLUSION

Many sellers in e-marketplaces are small businesses [1-3]. The selection of alliance partners is a

strategic decision for these sellers because they often lack resources and struggle to survive in a

competitive environment [4]. Hyperlink exchange is an IT-enabled means of alliance formation.

Although hyperlinks are ubiquitous on e-marketplaces, past studies have rarely offered guidance to

sellers about when they should exchange hyperlinks so as to raise their performance in e-marketplaces

[10, 14, 17]. Going beyond past studies, this study extends our theoretical understanding in this area

and offer guidance to sellers in terms of when they should exchange hyperlinks so as to increase their

performance in e-marketplaces. As e-marketplaces continue to offer more and better IT capabilities,

sellers would need guidance about how to leverage these IT capabilities to improve their performance.

Research in the direction of this study can help to address such issues and benefit the numerous sellers

in e-marketplaces.

References[1] S. Bandyopadhyay and S. Bandyopadhyay, Estimating Time Required to Reach Bid Levels in

Online Auctions, Journal of Management Information Systems, 26 (3) (2010) 275-301.

[2] Z. Qu, Y. Wang, S. Wang, and Y. Zhang, Implications of online social activities for e ‐tailers'

business performance, European Journal of Marketing, 47 (8) (2013) 1190-1212.

[3] X. Wu, R. Ma, and Y. Shi, How do Latecomer Firms Capture Value From Disruptive

Technologies? A Secondary Business-Model Innovation Perspective, IEEE Transactions on

Engineering Management, 57 (1) (2010) 51-62.

[4] M. H. Morris, A. Ko ak, and A. zer, Coopetition as a small business strategy: Implications for

performance, Journal of Small Business Strategy, 18 (1) (2007) 35-55.

[5] Y. Wang, S. Wang, Y. Fang, and P. Y. K. Chau, Store survival in online marketplace: An

empirical investigation, Decision Support Systems, 56 (1) (2013) 482-493.

23

[6] N. K. Park, J. M. Mezias, and J. Song, A resource-based view of strategic alliances and firm

value in the electronic marketplace, Journal of Management, 30 (1) (2004) 7-27.

[7] Z. Katona and M. Sarvary, Network formation and the structure of the commercial world

wide web, Marketing Science, 27 (5) (2008) 764-778.

[8] C. Dellarocas, Z. Katona, and W. M. Rand, Media, Aggregators and the Link Economy:

Strategic Hyperlink Formation in Content Networks, Management Science, 59 (10) (2010)

2360-2379.

[9] H. W. Park and M. Thelwall, Hyperlink Analyses of the World Wide Web: A Review, Journal

of Computer-Mediated Communication, 8 (4) (2003) 0.

[10] A. T. Stephen and O. Toubia, Deriving value from social commerce networks, Journal of

Marketing Research, 47 (2) (2010) 215-228.

[11] L. Vaughan, Web hyperlinks reflect business performance: A study of US and Chinese IT

companies, Canadian Journal of Information and Library Science, 28 (1) (2004) 17-31.

[12] P. Evans and T. S. Wurster, Getting real about virtual commerce, Harvard Business Review,

77 (1999) 84-98.

[13] P. B. Lowry, A. Vance, G. Moody, B. Beckman, and A. Read, Explaining and predicting the

impact of branding alliances and web site quality on initial consumer trust of e-commerce

web sites, Journal of Management Information Systems, 24 (4) (2008) 199-224.

[14] K. J. Stewart, How hypertext links influence consumer perceptions to build and degrade trust

online, Journal of Management Information Systems, 23 (1) (2006) 183-210.

[15] R. K. Srivastava, L. Fahey, and H. K. Christensen, The resource-based view and marketing:

The role of market-based assets in gaining competitive advantage, Journal of Management, 27

(6) (2001) 777-802.

[16] M. Thelwall and T. A. Sheldon, Exploring the link structure of the Web with network

diagrams, Journal of Information Science, 27 (6) (2001) 393-401.

[17] D. Mayzlin and H. Yoganarasimhan, Link to success: How blogs build an audience by

promoting rivals, Management Science, 58 (9) (2012) 1651-1668.

[18] Z. J. Lin, H. Yang, and B. Arya, Alliance partners and firm performance: resource 24

complementarity and status association, Strategic Management Journal, 30 (9) (2009) 921-

940.

[19] T. Das and B. S. Teng, Partner analysis and alliance performance, Scandinavian Journal of

Management, 19 (3) (2003) 279-308.

[20] T. K. Das and B. S. Teng, A resource-based theory of strategic alliances, Journal of

Management, 26 (1) (2000) 31-61.

[21] J. Barney, Firm resources and sustained competitive advantage, Journal of Management, 17

(1) (1991) 99-120.

[22] F. T. Rothaermel, W. Boeker, Old technology meets new technology: complementarities,

similarities, and alliance formation, Strategic Management Journal, 29 (2008) 47-77.

[23] O. E. Williamson, Comparative economic organization: The analysis of discrete structural

alternatives, Administrative Science Quarterly, 1991) 269-296.

[24] J. G. Combs and Ketchen, David J, Explaining interfirm cooperation and performance:

toward a reconciliation of predictions from the resource‐based view and organizational

economics, Strategic Management Journal, 20 (9) (1999) 867-888.

[25] K. M. Eisenhardt and C. B. Schoonhoven, Resource-based view of strategic alliance

formation: Strategic and social effects in entrepreneurial firms, Organization science, 7 (2)

(1996) 136-150.

[26] H. Yasuda, Formation of strategic alliances in high-technology industries: comparative study

of the resource-based theory and the transaction-cost theory, Technovation, 25 (7) (2005) 763-

770.

[27] T. Boyles, Small business and Web 2.0: hope or hype?, Entrepreneurialexecutive, 16 (2011)

[28] P. Chatterjee, Interfirm alliances in online retailing, Journal of Business Research, 57 (7)

(2004) 714-723.

[29] S. Burt and L. Sparks, E-commerce and the retail process: a review, Journal of Retailing and

Consumer Services, 10 (5) (2003) 275-286.

[30] Y. Luo, Coopetition in international business. Copenhagen: Copenhagen Business School

Press, 2004.25

[31] D. Houser and J. Wooders, Reputation in auctions: Theory, and evidence from eBay, Journal

of Economics & Management Strategy, 15 (2) (2006) 353-369.

[32] M. I. Melnik and J. Alm, Does a seller’s ecommerce reputation matter? evidence from ebay

auctions, Journal of Industrial Economics, 50 (3) (2002) 337-349.

[33] A. V. Shipilov, S. X. Li, and H. R. Greve, The prince and the pauper: search and brokerage in

the initiation of status-heterophilous ties, Organization Science, 22 (6) (2011) 1418-1434.

[34] G. U. Qian and L. U. Xiaohui, Unraveling the mechanisms of reputation and alliance

formation: A study of venture capital syndication in China, Strategic Management Journal, 35

(5) (2014) 739–750.

[35] B. A. Aubert, S. Rivard, and M. Patry, A transaction cost approach to outsourcing behavior:

some empirical evidence, Information and Management, 30 (2) (1996) 51-64.

[36] Y. Benslimane, M. Plaisent, and P. Bernard, Investigating search costs and coordination costs

in electronic markets: A transaction costs economics perspective, Electronic Markets, 15 (3)

(2005) 213-224.

[37] S. Standing, C. Standing, and P. E. D. Love, A review of research on e-marketplaces 1997–

2008, Decision Support Systems, 49 (1) (2010) 41-51.

[38] E. W. Tsang, Motives for strategic alliance: A resource-based perspective, Scandinavian

Journal of Management, 14 (3) (1998) 207-221.

[39] M.-J. Chen, Competitor analysis and interfirm rivalry: Toward a theoretical integration,

Academy of Management Review, 21 (1) (1996) 100-134.

[40] H. Mitsuhashi and H. R. Greve, A matching theory of alliance formation and organizational

success: Complementarity and compatibility, Academy of Management Journal, 52 (5) (2009)

975-995.

[41] T. E. Stuart, Interorganizational alliances and the performance of firms: A study of growth and

innovation rates in a high-technology industry, Strategic Management Journal, 21 (8) (2000)

791-811.

[42] A. Capaldo and A. M. Petruzzelli, Partner Geographic and Organizational Proximity and the

Innovative Performance of Knowledge‐Creating Alliances, European Management Review, 26

11 (1) (2014) 63-84.

[43] S. A. Chung, H. Singh, and K. Lee, Complementarity, status similarity and social capital as

drivers of alliance formation, Strategic Management Journal, 21 (1) (2000) 1-22.

[44] R. D. Ireland, M. A. Hitt, and D. Vaidyanath, Alliance management as a source of competitive

advantage, Journal of Management, 28 (3) (2002) 413-446.

[45] L. Wang and E. J. Zajac, Alliance or acquisition? A dyadic perspective on interfirm resource

combinations, Strategic Management Journal, 28 (13) (2007) 1291-1317.

[46] A. P. R. Ragozzino, The effects of geographic distance on the foreign acquisition activity of

US firms, Management International Review, 49 (4) (2009) 509-535.

[47] A. Zaheer and E. Hernandez, The geographic scope of the MNC and its alliance portfolio:

resolving the paradox of distance, Global Strategy Journal, 1 (1‐2) (2011) 109-126.

[48] C. Felzensztein, L. Huemer, and E. Gimmon, The effects of co-location on marketing

externalities in the salmon-farming industry, Journal of Business and Industrial Marketing, 25

(1) (2009) 73-82.

[49] C. Geldes, C. Felzensztein, E. Turkina, and A. Durand, How does proximity affect interfirm

marketing cooperation? A study of an agribusiness cluster, Journal of Business Research, 68

(2) (2015) 263-272.

[50] P. M. Rosenzweig and J. V. Singh, Organizational environments and the multinational

enterprise, Academy of Management Review, 16 (2) (1991) 340-361.

[51] O. Sorenson and T. E. Stuart, Syndication networks and the spatial distribution of venture

capital investments1, American Journal of Sociology, 106 (6) (2001) 1546-1588.

[52] H. Kranenburg, J. Hagedoorn, and S. Lorenz‐Orlean, Distance Costs and The Degree of Inter‐

Partner Involvement in International Relational‐Based Technology Alliances, Global Strategy

Journal, 4 (4) (2014) 280-291.

[53] F. Cairncross, The death of distance: How the communications revolution is changing our

lives: Harvard Business Press, 2001.

[54] S. S. Standifird, Reputation and e-commerce: eBay auctions and the asymmetrical impact of

positive and negative ratings, Journal of Management, 27 (3) (2001) 279-295.27

[55] M. Washington and E. J. Zajac, Status evolution and competition: theory and evidence,

Academy of Management Journal, 48 (2) (2005) 282-296.

[56] I. Stern, J. M. Dukerich, and E. Zajac, Unmixed signals: How reputation and status affect

alliance formation, Strategic Management Journal, 35 (4) (2014) 512–531.

[57] S. M. Wagner, L. S. Coley, and E. Lindemann, Effects of Suppliers' Reputation on the Future

Of Buyer - Supplier Relationships: The Mediating Roles Of Outcome Fairness And Trust,

Journal of Supply Chain Management, 47 (2) (2011) 29-48.

[58] A. Zaheer, B. Mcevily, and V. Perrone, Does Trust Matter? Exploring the Effectsof

Interorganizational and Interpersonaltrust on Performance, Organization Science, 9 (2) (1998)

141-159.

[59] Y. Bakos and C. Dellarocas, Cooperation Without Enforcement? A Comparative Analysis of

Litigation and Online Reputation as Quality Assurance Mechanisms, Social Science

Electronic Publishing, 57 (11) (2002) 127-142.

[60] P. Ingram and L. Q. Yue, Structure, Affect and Identity as Bases of Organizational

Competition and Cooperation, Academy of Management Annals, 2 (1) (2008) 275-303.

[61] T. K. Das and B. S. Teng, The dynamics of alliance conditions in the alliance development

process, Journal of Management Studies, 39 (5) (2002) 725-746.

[62] J. Knoben and L. A. Oerlemans, Proximity and inter‐organizational collaboration: A literature

review, International Journal of Management Reviews, 8 (2) (2006) 71-89.

[63] M. J. Eppli and J. D. Benjamin, The evolution of shopping center research: a review and

analysis, Journal of Real Estate Research, 9 (1) (1994) 5-32.

[64] A. Hortacsu, F. A. Martinez-Jerez, and J. Douglas, The Geography of Trade on eBay and

MercadoLibre, NET Institute Working Paper No. 06-09. Available at SSRN:

http://ssrn.com/abstract=939327 (2006)

[65] D. Mukherjee, A. S. Gaur, S. S. Gaur, and F. Schmid, External and internal influences on

R&D alliance formation: Evidence from German SMEs, Journal of Business Research, 66

(11) (2013) 2178-2185.

[66] A. Parkhe, Strategic alliance structuring: A game theoretic and transaction cost examination 28

of interfirm cooperation, Academy of Management Journal, 36 (4) (1993) 794-829.

[67] G. Ahuja, The duality of collaboration: Inducements and opportunities in the formation of

interfirm linkages, Strategic Management Journal, 21 (3) (2000) 317-343.

[68] C. E. Eesley and E. B. Roberts, Are You Experienced or Are You Talented?: When Does

Innate Talent versus Experience Explain Entrepreneurial Performance?, Strategic

Entrepreneurship Journal, 6 (3) (2012) 207-219.

[69] J. W. Lu and D. Xu, Growth and survival of international joint ventures: An external-internal

legitimacy perspective, Journal of Management, 32 (3) (2006) 426-448.

[70] S. Shane and M.-D. Foo, New firm survival: Institutional explanations for new franchisor

mortality, Management Science, 45 (2) (1999) 142-159.

[71] A. C. Cameron and P. K. Trivedi, Microeconometrics: methods and applications. Cambridge

Cambridge University Press, 2005.

[72] D. A. Belsley, E. Kuh, and R. E. Welsch, Regression diagnostics. New York: Wiley-

Interscience, 1980.

[73] L. S. Aiken, S. G. West, and R. R. Reno, Multiple regression: Testing and interpreting

interactions: Sage Publications Inc, 1991.

[74] M. A. Fuller, M. A. Serva, and J. S. Benamati, Seeing Is Believing: The Transitory Influence

of Reputation Information on E-Commerce Trust and Decision Making, Decision Sciences,

38 (4) (2007) 675-699.

[75] K. Uhlenbruck, M. A. Hitt, and M. Semadeni, Market value effects of acquisitions involving

Internet firms: A resource‐based analysis, Strategic Management Journal, 27 (10) (2006) 899-

913.

[76] J. Christoffersen, A Review of Antecedents of International Strategic Alliance Performance:

Synthesized Evidence and New Directions for Core Constructs, International Journal of

Management Reviews, 15 (1) (2013) 66-85.

29