CONFIDENCE IN LEARNING: INTER- AND … · CONFIDENCE IN LEARNING: INTER- AND INTRAORGANIZATIONAL LEARNING IN FOREIGN MARKET ENTRY DECISIONS JIATAO LI,1* CUILI QIAN,2 and FIONA K

Strategic Management JournalStrat. Mgmt. J., 36: 918–929 (2015)

Published online EarlyView 1 April 2014 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2250Received 6 February 2010; Final revision received 29 January 2014

CONFIDENCE IN LEARNING: INTER- ANDINTRAORGANIZATIONAL LEARNING IN FOREIGNMARKET ENTRY DECISIONS

JIATAO LI,1* CUILI QIAN,2 and FIONA K. YAO3

1 Department of Management, Hong Kong University of Science & Technology,Kowloon, Hong Kong2 Department of Management, City University of Hong Kong, Kowloon, Hong Kong3 Department of Business Administration, College of Business, University of Illinoisat Urbana-Champaign, Champaign, Illinois, U.S.A.

From an organizational learning perspective, we argue that the information signaled by thedistribution attributes of foreign investors already operating in a location will influence the entrydecisions of later arrivals by affecting their level of confidence in imitating. In the context of foreigninvestment decisions, the proportion of experienced firms in a location was shown to first increasea follower firm’s confidence about imitating them, but then to decrease it, due to anticipatedcompetition. The impact of learning from target organizations also varies with the experienceof the learning organization. Data on the location choices of 7,478 manufacturing ventures inChina by U.S. firms supported the hypotheses. The results provide a more integrated and nuancedunderstanding of learning in foreign direct investment. Copyright © 2014 John Wiley & Sons,Ltd.

INTRODUCTION

Organizational learning theorists have long con-tended that, when facing environmental uncertainty,organizations learn vicariously from their refer-ence groups, imitating or avoiding specific actionsbased on their perceived impact on others (Baumand Ingram, 1998; Baum, Li, and Usher, 2000;Cyert and March, 1963; Henisz and Delios, 2001;Kim and Miner, 2007). This line of research hasmainly focused on learning from a particular groupof reference firms. Apart from copying practicesthat are frequently used by similar others (Baumet al., 2000; Baum and Ingram, 1998), organiza-tions tend to imitate the practices of firms with cer-tain traits such as large size, high status, or good

Keywords: confidence; organizational learning; imitation;foreign investment; China*Correspondence to: Jiatao Li, Department of Management,HKUST, Clear Water Bay, Hong Kong. E-mail: [email protected]

Copyright © 2014 John Wiley & Sons, Ltd.

performance that give them confidence in imitation(Haunschild and Miner, 1997; Rhee, Kim, and Han,2006). However, these prior studies have severallimitations. First, they have tended to neglect thevariations within a reference group, which affect afollower firm’s confidence in learning through imi-tation (Rhee et al., 2006). For instance, a lack ofunanimity within a reference group about adoptinga certain practice should decrease the confidence ofa copying firm about adopting the same practice.Then, prior research on interorganizational learninghas focused on imitating other firms in adopting aparticular practice (Baum and Ingram, 1998; Baumet al., 2000; Kim and Miner, 2007; Schwab, 2007)without paying sufficient attention to the possiblecrowding and increased competition that might beanticipated as a result. When the number of otherfirms adopting a particular practice reaches a cer-tain level, it might decrease the confidence of alearning firm in adopting that practice. Finally, priorlearning studies have generally emphasized a firm’s

Inter- and Intraorganizational Learning in Foreign Market Entry Decisions 919

aggregate international experience such as the totalnumber of its foreign investments (e.g., Henisz andDelios, 2001)1 in analyzing its learning practices,but such studies have not further unpacked theeffects of the different types of experience.

This study was designed to address thosedeficiencies by examining how the informationsignaled by the distribution of attributes within areference group will affect a learning firm’s levelof confidence in learning. This takes into accountthat interorganizational influence is potentiallymore complicated than the simple linear corre-lation shown in previous studies. Based on theinterorganizational learning framework (Ingram,2002), we explore how the nature of the targetorganizations affects the confidence of a copyingorganization in learning, and how the nature ofthe learning organization’s experience interactswith its learning from target firms. The empiricalcontext was location choice by U.S. manufacturingfirms investing in China, a decision of strategicimportance to any multinational firm (Bastos andGreve, 2003; Henisz and Delios, 2001).

Specifically, the study was designed to addresstwo research questions. First, to what extent doesthe proportion of the previous entries made by expe-rienced investors affect the level of confidence oflater entrants in imitating them? When an organiza-tion tries to learn from a practice used by a referencegroup, it is possible that the members of the refer-ence group have used the practice in different waysand to different extents (Rhee et al., 2006). The dis-tribution of attributes among the reference groupwill signal to subsequent imitators and should affecttheir confidence in learning through imitation.2 Thishas until now received only limited attention frommanagement scholars. This study examined the pro-portion of experienced investors among the refer-ence group as a potentially influential measure ofthe distribution of experience. It tested the proposi-tion that the proportion of prior entries in a locationby experienced target firms should have an invertedU-shaped relationship with subsequent entrants’confidence in imitation and thus their probability ofimitating. The theoretical argument for the invertedU-shaped relationship goes beyond the conven-tional interorganizational learning and imitation

1 Thanks to an anonymous reviewer for this insight.2 A group led by Rhee has examined the influence of differencesin niche width and size among the reference groups on confidencein imitation (Rhee et al., 2006).

approaches. Rather, it emphasizes the role of sig-naling and the importance of considering the crowd-ing effect of anticipated competition—a neglectedtopic in previous studies. In particular, we argue thata high proportion of experienced foreign investorsin a location can initially signal a high probabil-ity of success for firms following their example anda munificent local environment. This increases afirm’s confidence in imitating. It will feel more con-fident in learning from experienced firms than inimitating inexperienced ones. Rather than simplycopying the most frequent location selections, aninvesting firm is more likely to follow closely thelocation choices of the experienced investors. Lateron, however, when a location is crowded with expe-rienced investors, this can signal serious competi-tion from the incumbents, which may decrease theconfidence of a copying firm in selecting that loca-tion. Organizations do not blindly copy the practicesof any reference group, after all. This will be par-ticularly salient if the location is crowded with alarge proportion of experienced firms. Bringing inthe competition argument in this way suggests thatlearning from the distribution of attributes among areference group may actually discourage a learningorganization from selecting a location that has beenselected by a large number of apparently suitablefirms.

This study’s second research question was aboutthe nature of the learning organization. Organiza-tions vary in their capacity for learning from others(Cohen and Levinthal, 1990). This study thus fur-ther examined how learning from target firms varieswith the learning firm’s experience in the host coun-try. Previous learning studies in international strat-egy have generally relied on aggregate measures ofexperience (e.g., Henisz and Delios, 2001),3 but thisstudy decomposed the investing experience of thelearning organization into experience in the particu-lar location in the host country and experience else-where in the country. It then investigated which ofthese more fine-grained specifications would havea stronger relationship with a firm’s foreign directinvestment (FDI) location decisions. The resultsprovide a more nuanced understanding of what sortof experience matters most and thus contribute tothe debate about whether and how intraorganiza-tional learning may interact with vicarious learningfrom others (Schwab, 2007).

3 Thanks to an anonymous reviewer for this insight.

Copyright © 2014 John Wiley & Sons, Ltd. Strat. Mgmt. J., 36: 918–929 (2015)DOI: 10.1002/smj

920 J. Li, C. Qian, and F. K. Yao

HYPOTHESES

Confidence in learning

A firm can learn from observing the choices ofother firms to decrease uncertainty (Haunschildand Miner, 1997; Henisz and Delios, 2001; Levittand March, 1988). After observing the patterns ofothers’ behavior, a firm can set out to learn thetechnologies, organizational forms, and managerialpractices that worked for them (Baum et al., 2000;Levinthal and March, 1993; Schwab, 2007).

In FDI, a firm must decide whether to followothers into the same location or to invest elsewhereand explore a new potential market. Prior entrants’experience can be an important information sourcefor firms evaluating whether a given location isworth pursuing (Abrahamson and Rosenkopf,1993; Feinberg and Gupta, 2004). This is ofparticular relevance for entry into an emergingeconomy where the uncertainty is likely to begreat. Observing and imitating other firms’ locationchoices can help reduce it (Baum et al., 2000;Henisz and Delios, 2001). Considerable evidencesupports the idea that organizations imitate thepractices previously adopted by relevant othersand that a firm will choose a location that haspreviously seen frequent successful entries (Baumet al., 2000; Haunschild and Miner, 1997; Palmer,Jennings, and Zhou, 1993).

Beyond simply imitating the actions or practicesthat have most frequently been used, firms may lookfurther into differences among the firms in their ref-erence group. Variations among those predecessorsmay add uncertainty to a decision to imitate. Whenthe potential model firms differ, a copying firm mustestimate how much confidence it has in imitating aparticular practice or behavior (Rhee et al., 2006).This should certainly be the case in the contextof selecting a location for FDI. Some of the firmsoperating in a location could be first-time foreigninvestments while others could be one of sequen-tial investments. Johanson and Vahlne (1977) havesuggested that expanding operations in a host coun-try involves the gradual acquisition and accumu-lation of experience over time by operating there,incrementally increasing the commitment to a for-eign market. Sequential investments made by expe-rienced firms may therefore have something extrato offer as models for imitation. So, a higher per-centage of sequential entries into a location by for-eign investors should tend to decrease uncertainty in

imitation and increase the probability of an imitat-ing firm entering that location. The higher percent-age of sequential entries may send a stronger signalabout a location’s attractiveness from experiencedentrants who have already gained some knowledgeabout location choices from their prior investments.It can also signal a munificent environment, enhanc-ing the confidence of a follower firm in imitating thelocation choice.

That confidence may, however, decrease lateron when the proportion of experienced investorsin a market becomes high. Experienced foreigninvestors operating in a location are potential com-petitors. Rivals play an important role in affectinglearning and its outcomes (Levitt and March, 1988).Strategic similarity among rivals may generatemore intense competition for resources, which mustalso be considered in learning through observation(McKendrick, 2001). Competition leads firms todifferentiate their strategies from those of similarothers (Gimeno et al., 2005; Li, 2008), as Li hasshown using data on foreign banks entering the U.S.market (Li, 2008). He showed that the presence ofdomestic banks in a particular location deterred theentry of foreign banks. Gimeno’s group has alsodemonstrated that the structure of domestic com-petition affects the behavior of new entrants in aforeign market (Gimeno et al., 2005).

If a potential investment location has becomecrowded with prior entries, the expectation of com-petition should discourage further entries into thatlocation. Beyond the deterrent effect of competitionfrom all prior entries, a high proportion of sequen-tial entries by firms that had already establishedthemselves in the host country should be a partic-ularly strong deterrent. Such experienced investorsmay have better knowledge of the business andpolitical environment, and following their lead willput a new entrant in direct competition with expe-rienced rivals. The presence of a high proportionof experienced rivals in a market may thus sig-nal intense competition and less favorable marketopportunities, decreasing the confidence of an imi-tating firm in choosing that location for its FDI.

In summary, management scholars have previ-ously examined the influence of the number of priorentries in a location on the choices of subsequententrants (Baum and Ingram, 1998; Kim and Miner,2007; Schwab, 2007), but the argument here is that,through signaling effects, the characteristics of theprior entrants (especially the proportion of expe-rienced investors among them) may also affect a



firm’s FDI location choice. Up to a point, a highpercentage of experienced prior entrants will signala munificent environment with promising marketopportunities. This will decrease a follower firm’suncertainty and enhance its confidence in imitation,thereby encouraging its entry into that location. Butbeyond a certain level, the prospect of intense com-petition may decrease a follower’s confidence inentering that location (Baum and Mezias, 1992).This suggests that an inverted U-shaped relationshipshould be observed.

Hypothesis 1 (H1): The percentage of sequen-tial investments by other firms in a location hasan inverted U-shaped relationship with the prob-ability that a firm investing subsequently willchoose that location.

The moderating role of experience

As Ingram (2002) has pointed out, learning is influ-enced not only by the nature of the model organi-zations (the distribution of their attributes in thiscontext), but also by the characteristics of the firmdoing the learning. Specifically, a learning firm’sown FDI experience in the host country should be animportant factor moderating the influence of modelorganizations in determining its FDI location choice(Chang, 1995; Henisz and Delios, 2001).

The first hypothesis predicts that, when there arefew experienced referents in a location, imitationwill tend to dominate in a follower firm’s deci-sion making. This positive relationship is likely tobe stronger for firms with less experience. Com-peting in a foreign market requires a firm to learnabout its culture, business practices, consumer pref-erences, policies and regulations, and possible pol-icy changes (Henisz and Delios, 2001). A firm canlearn from several information sources, especiallyfrom its own experience, but also vicariously fromreferent firms (Baum et al., 2000; Cyert and March,1963; Henisz and Delios, 2001; March and Simon,1958; Schwab, 2007). An inexperienced firm hasless information available from its own experience,and thus should pay more attention to signals fromothers as a main information source. When the ref-erence group signals a munificent environment, thiswill particularly encourage inexperienced firms. Incontrast, a firm with abundant experience in the geo-graphic market can look internally for information,relying less on external referents.

H1 further predicts that, when a location hostsa high percentage of experienced firms, the antic-ipated competition will tend to dominate and newfirms will be less likely to invest in that loca-tion. That negative relationship is also likely to bestronger for a firm with less experience. An experi-enced firm will be less sensitive to external cues andtherefore less susceptible to such discouragement.

Hypothesis 2 (H2): Greater experience inthe host country will weaken the relationshipbetween the percentage of sequential invest-ments by other firms in a location and theprobability that a subsequent entrant will choosethat location.

Relevant and transferable experience

The host country experience discussed in H2 canbe experience in the specific location under con-sideration, or experience in other locations in thehost country. In either case, the utility of the priorexperience will depend on whether it is perceived asrelevant and transferable (Levitt and March, 1988;Shrivastava, 1983; Yelle, 1979). By definition, onlyexperience considered relevant is likely to be per-ceived as valuable. Experience gained in other con-texts is more likely to be considered relevant to anew setting if the contexts can be viewed as shar-ing certain similarities and common principles (Sin-gley and Anderson, 1989). For instance, Delios andHenisz (2003) found that only international expe-rience gained in countries perceived as having asimilar culture is effective in minimizing the deter-ring effect on FDI of an uncertain policy environ-ment. Beyond that, only experience perceived astransferable is likely to be considered helpful inlearning about a new location. Experience gainedin other locations may not be easily transferredbecause of obstacles to knowledge transfer (Huber,1982, 1991).

To investigate what experience is most rele-vant and transferable, it is helpful to decompose afirm’s experience in the host country into experi-ence in the specific location and experience else-where in the country. Experience is most likelyto be considered relevant and transferable whenit was gained in the specific location where aninvestment is being contemplated. More experiencein a location implies more knowledge about it,which should make a firm less sensitive to others’actions. If a firm’s prior experience was acquired



in other locations, it may not be entirely relevantand/or easily transferable. Entering a geographicmarket requires location-specific knowledge, andlocal connections are often helpful (Henisz andDelios, 2001). Prior experience in other locationsmay provide little of either. As a result, even a firmwith abundant experience acquired elsewhere mayfeel obliged to rely on learning from local referents.Comparing the two experience profiles suggests that

Hypothesis 3 (H3): The moderating effect of afirm’s experience as predicted by H2 will bestronger if that experience was acquired in thespecific location rather than elsewhere.

RESEARCH METHODS

Sample

Data describing U.S. manufacturing investmentsin China from 1979 to 1995 were applied totest these hypotheses. Three data sources wereexploited. The first was data on direct investment inChina by U.S. firms from the research institute ofChina’s Ministry of Foreign Trade and EconomicCooperation (MOFTEC) in Beijing. Their databasecontains a brief profile of each foreign firm thathas operated in the country, providing data oninvestment location and industry, the identity ofthe foreign parent, total investment, number ofemployees, and the number of expatriate managers.The second data source, American Business inChina (1995), was used to confirm the informationin the raw data set, particularly the identities ofthe parent firms. The third data source, the ChinaStatistical Yearbook, provided data on location andindustry-level control variables.

During the study period, U.S. multinationalslaunched a total of 7,478 manufacturing sub-sidiaries in 29 manufacturing industries (two-digitSIC level), entering 29 of China’s 30 provinces,regions, and autonomous municipalities.4 Amongthe 7,478 investments, 6,382 were first-time FDIsin China, and 1,096 were identified as sequentialFDI entries. These data were formed into three

4 The study focused on foreign entries in manufacturing indus-tries, excluding those in service sectors where the governmenttended to enforce more restrictions on foreign ownership (Fu,2000). Focusing on manufacturing industries also allowed moreextensive comparison with the results of other FDI studies (Changand Park, 2005; Guillén, 2002).

different samples to test the hypotheses: the overallsample, which consisted of all initial and sequentialinvestments (7,478 investments); the first-time FDIsample, which included only initial entries (6,382investments); and the sequential FDI sample,which included only sequential investments (1,096investments).

Variables and measures

The dependent variable, FDI location choice, wasa dummy variable coded as “1” if a foreign invest-ment was located in a particular province and as “0”otherwise.

In order to capture learning from experienced ref-erence firms, the independent variable, percentageof sequential FDI entries in the same industry in thesame province, was computed by dividing the num-ber of sequential investments by the total numberof investments by all other U.S. firms in the sameindustry in the province or municipality. An invest-ing firm’s experience in China was represented bya count of the number of previous investments bythe parent firm in China (firm’s prior investmentsin China). This was further disaggregated into thefirm’s experience in the particular province, quanti-fied using a count of its previous investments there(firm’s prior investments in the province), and itsexperience in other Chinese provinces (firm’s priorinvestments in other provinces).

Most prior research on FDI location decisionshas taken an economic perspective, emphasizinglocation factors such as market potential and eco-nomic growth (Broadman and Sun, 1997; Chengand Kwan, 2000). In this study, GDP per capita,GDP growth, trade growth, and railway density foreach province in each year were included in theanalysis. Political factors were controlled for as wellby including a dummy variable, policy incentive,to indicate provinces that had one or more SpecialEconomic Zones (SEZs) and/or open coastal cities(OCCs) (Zhou, Delios, and Yang, 2002). Expatriatemanagers, quantified as the number of expatriates ina firm’s top management team, was included to con-trol for any potential effects of top management’sinternational experience (Carpenter, Sanders, andGregersen, 2001; Reuber and Fischer, 1997). Inaddition, total number of prior investments fromthe same home country in the same industry andprovince was a count of all other U.S.-invested man-ufacturing ventures in a particular province, andtotal number of prior investments from other home



countries in the same industry and province wasa count of all manufacturing ventures by non-U.S.firms in a particular province. These latter two vari-ables and their squared terms were both included inthe models to control for any effects of populationdensity on market entry probability (Hannan andFreeman, 1989). All of the independent and controlvariables were lagged one year for predicting entryin the following year.

Modeling

Nested logit models were formulated to investigatethe location decisions. Nested logit models areparticularly applicable when independence fromirrelevant alternatives (IIA) cannot be assumed,5

and Hausman-McFadden tests indeed indicated thatIIA could not be assumed for these data (Hausmanand McFadden, 1984). The hierarchical structureof the nested logit models in this study was basedon the spatial variation of China’s preferentialpolicies encouraging FDI. China has given firstpriority to the eastern coastal region and extendedpreference to the central and western regions inphases (Fu, 2000). Therefore, three regional clus-ters were included in the models: eastern, central,and western.

RESULTS

Table 1 presents descriptive statistics and corre-lation coefficients summarizing the 7,478 entrydecisions by U.S. multinationals during the studyperiod.

Table 2 reports the computed coefficients ofnine nested logit regression models representingthe FDI location choices, incrementally addingthe variables of interest. Model 1 included onlycontrol variables. In Models 2 and 3, the linearand squared terms representing vicarious learningfrom other firms were added. The percentage ofsequential FDI entries by other firms demonstrateda significant positive relationship (p≤ 0.01) withthe dependent variable while its squared term hada negative coefficient (p≤ 0.05), supporting H1.

5 Conditional logit modeling is a powerful tool for investigatinglocation decisions, as it directly measures how alternatives arecompared among the choice set (Greve, 2000; Shaver and Flyer,2000), but independence from irrelevant alternatives is essential ifconditional logit models are to be applied (Cheng, 2007; Hansen,1987).

Models 4–9 added the interaction terms betweenlearning from others and firm-level FDI experience(both centered, see Ryan, 1997). In Models 4 and 5,the number of a firm’s prior investments in China asa whole was included as a measure of FDI experi-ence. The coefficient of the interaction term relatingFDI experience in China and learning from oth-ers was negative and significant (p≤ 0.01), and theinteraction between FDI experience and the squaredterm representing learning from others was positiveand significant (p≤ 0.05), lending support to H2. InModels 6–9, FDI experience in China as a wholewas further decomposed into experience in the sameprovince and experience in other provinces. In Mod-els 6 and 7, the number of prior investments in theprovince was used. In Models 8 and 9, the number ofprior investments in other provinces was included.The interaction terms relating FDI experience inother provinces and vicarious learning from oth-ers were not significant, but the interaction termsrelating FDI experience in the same province andvicarious learning from others were significant, sup-porting H3. This suggests that prior experience inthe province provided valuable and relevant knowl-edge about the location; firms with such experiencedisplayed less reliance on others for their informa-tion. However, prior experience in other provincesdid not seem to offer knowledge about the locationthat the firms considered valuable. Firms with suchexperience were still sensitive to the actions of priorinvestors.

The first-time FDI sample and the sequentialsample were also modeled to test the robustnessof these conclusions. The nested logit regressionresults using the two subsamples remained consis-tent with the hypotheses.6 In addition, an alterna-tive measure of the distribution of attributes amongthe learning targets was constructed. The ratio ofsequential to first-time investments was computedby expressing the number of sequential investmentsas a proportion of the number of first-time invest-ments by all other U.S. firms in the same indus-try in the province/municipality. This alternativeratio measure constituted a direct comparison oflearning from experienced and inexperienced firms.The results using the two measures were againconsistent.

6 Due to space limitations, only the tables for the overall sampleare reported, but the detailed results using the two subsamples areavailable on request.



Tabl

e1.

Des

crip

tive

stat

istic

san

dco

rrel

atio

nsam

ong

the

vari

able

s

Var

iabl

eM

ean

S.D

.1

23

45

67

89

1011

12

1.FD

Ilo

catio

nch

oice

0.03

0.18

2.G

DP

per

capi

ta0.

290.

230.

083.

GD

Pgr

owth

0.23

0.11

0.04

0.46

4.T

rade

grow

th0.

270.

410.

01−

0.08

−0.

115.

Polic

yin

cent

ives

0.38

0.49

0.14

0.29

0.31

−0.

016.

Rai

lway

dens

ity0.

010.

010.

080.

48−

0.06

−0.

090.

167.

Exp

atri

ate

man

ager

s0.

964.

370.

010.

010.

01−

0.01

−0.

010.

018.

Firm

’spr

ior

inve

stm

ents

inC

hina

0.31

1.12

0.01

0.08

0.12

−0.

010.

020.

01−

0.01

9.Fi

rm’s

prio

rin

vest

men

tsin

the

prov

ince

0.01

0.16

0.13

0.04

0.05

0.01

0.05

0.02

−0.

010.

24

10.F

irm

’spr

ior

inve

stm

ents

inot

her

prov

ince

s0.

300.

09−

0.02

0.08

0.11

−0.

01−

0.01

−0.

01−

0.01

0.89

0.10

11.T

otal

num

ber

ofpr

ior

inve

stm

ents

from

the

sam

eho

me

coun

try

inth

esa

me

indu

stry

inth

esa

me

prov

ince

5.13

12.15

0.17

0.34

0.35

0.02

0.33

0.17

−0.

010.

030.

090.

02

12.T

otal

num

ber

ofpr

ior

inve

stm

ents

from

othe

rho

me

coun

trie

sin

the

sam

ein

dust

ryin

the

sam

epr

ovin

ce

76.37

248.

920.

080.

180.

260.

020.

30−

0.04

−0.

010.

020.

040.

010.

57

13.P

erce

ntag

eof

sequ

entia

lFD

Ien

trie

sin

the

sam

ein

dust

ryin

the

sam

epr

ovin

ce

0.02

0.08

0.05

0.12

0.12

0.01

0.11

0.07

−0.

010.

020.

030.

020.

190.

15

N=

216,

862.

All

corr

elat

ion

coef

ficie

nts

with

am

agni

tude

grea

ter

than

0.01

are

sign

ifica

ntat

the

0.05

leve

l.



Tabl

e2.

Coe

ffici

ents

ofne

sted

logi

treg

ress

ion

mod

els

ofFD

Ilo

catio

nde

cisi

ons

Mod

els

Var

iabl

es1

23

45

67

89

GD

Ppe

rca

pita

−0.

12−

0.13

−0.

11−

0.12

−0.

12−

0.12

−0.

12−

0.12

−0.

12(0.08)

(0.08)

(0.08)

(0.08)

(0.08)

(0.08)

(0.08)

(0.08)

(0.08)

GD

Pgr

owth

−0.

59−

0.59

−0.

57−

0.59

−0.

59−

0.59

−0 .

61−

0.58

−0.

57(0.58)

(0.58)

(0.58)

(0.59)

(0.59)

(0.59)

(0.59)

(0.58)

(0.59)

Tra

degr

owth

0.12

**0.

12**

0.12

**0.

12**

0.12

**0.

12**

0.12

**0.

12**

0.12

**(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Polic

yin

cent

ives

0.46

**0.

47**

0.47

**0.

48**

0.48

**0.

47**

0.47

**0.

48**

0.48

**(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

Rai

lway

dens

ity0.

10**

0.10

**0.

10**

0.10

**0.

10**

0.10

**0.

10**

0.10

**0.

10**

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01

)(0.01)

(0.01)

(0.01)

Exp

atri

ate

man

ager

s0.

040.

040.

040.

040.

040.

040.

040.

040.

04(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Firm

’spr

ior

inve

stm

ents

inC

hina

0.03

0.02

0.02

0.02

0.02

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Firm

’spr

ior

inve

stm

ents

inth

epr

ovin

ce1.

14**

1.23

**0.

96**

0.96

**(0.08)

(0.08)

(0.08)

(0.08)

Firm

’spr

ior

inve

stm

ents

inot

her

prov

ince

s−

0.30

**−

0.30

*−

0.32

**−

0.32

**(0.07)

(0.06)

(0.07)

(0.07)

Tota

lnum

ber

ofpr

ior

inve

stm

ents

from

the

sam

eho

me

coun

try

inth

esa

me

indu

stry

inth

epr

ovin

ce0.

06**

0.07

**0.

06**

0.06

**0.

06**

0.06

**0.

06**

0.06

**0.

06**

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(Tot

alnu

mbe

rof

prio

rin

vest

men

tsfr

omth

esa

me

hom

eco

untr

yin

the

sam

ein

dust

ryin

the

prov

ince

)2/1

00−

0.03

**−

0.03

**−

0.03

**−

0.03

**−

0.03

**−

0.03

**−

0.03

**−

0.03

**−

0.03

**(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

Tota

lnum

ber

ofpr

ior

inve

stm

ents

from

othe

rho

me

coun

trie

sin

the

sam

ein

dust

ryin

the

prov

ince

−0.

01−

0.01

−0.

01−

0.01

−0.

01−

0.01

−0.

01−

0.01

−0.

01(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(Tot

alnu

mbe

rof

prio

rin

vest

men

tsfr

omot

her

hom

eco

untr

ies

inth

esa

me

indu

stry

inth

epr

ovin

ce)2

/100

−0.

01−

0.01

−0.

01−

0.01

−0.

01−

0.01

−0.

01−

0.01

−0.

01(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

Incl

usiv

eva

lue

coef

ficie

nts

Eas

tern

0.74

**0.

73**

0.73

**0.

73**

0.74

**0.

74**

0.74

**0.

74**

0.74

**(0.03)

(0.02)

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Cen

tral

0.33

**0.

33**

0.33

**0.

34**

0.34

**0.

34**

0.34

**0.

33**

0.33

**(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Wes

tern

0.01

0.01

0.01

0.01

0.01

0.01

0.01

0 .01

0.01

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

Lea

rnin

gfr

omex

peri

ence

dta

rget

sin

the

refe

renc

egr

oup

Perc

enta

geof

sequ

entia

lFD

Ien

trie

sin

the

sam

ein

dust

ryin

the

sam

epr

ovin

ce0.

54**

1.21

**1.

23**

1.16

**1.

46**

1.50

**1.

20**

1.29

**(0.17)

(0.31)

(0.31)

(0.31)

(0.31)

(0.31)

(0.31)

(0.33)



Tabl

e2.

cont

inue

d

Mod

els

Var

iabl

es1

23

45

67

89

(Per

cent

age

ofse

quen

tialF

DI

entr

ies

insa

me

indu

stry

inth

esa

me

prov

ince

)2−

10.41

*−

1.61

*−

1.32

*−

1.69

*−

1.71

**−

1.47

**−

1.83

*(0.57)

(0.57)

(0.59)

(0.57)

(0.59)

(0.58)

(0.75)

Fir

m’s

expe

rien

cein

Chi

naas

am

oder

ator

(Fir

m’s

prio

rin

vest

men

tsin

Chi

na)×

(per

cent

age

ofse

quen

tialF

DI

entr

ies

inth

esa

me

indu

stry

inth

esa

me

prov

ince

)−

0.68

*−

1.08

**(0.21)

(0.34)

(Fir

m’s

prio

rin

vest

men

tsin

Chi

na)×

(per

cent

age

ofse

quen

tialF

DI

entr

ies

inth

esa

me

indu

stry

inth

esa

me

prov

ince

)21.

26*

(0.63)

Fir

m’s

expe

rien

cein

the

prov

ince

asa

mod

erat

or(F

irm

’spr

ior

inve

stm

ents

inth

epr

ovin

ce)×

(per

cent

age

ofse

quen

tial

FDI

entr

ies

inth

esa

me

indu

stry

inth

esa

me

prov

ince

)−

3.36

**−

8.05

**(0.61)

(1.04)

(Fir

m’s

prio

rin

vest

men

tsin

the

prov

ince

)×(p

erce

ntag

eof

sequ

entia

lFD

Ien

trie

sin

the

sam

ein

dust

ryin

the

sam

epr

ovin

ce)2

16.77

**(3.24)

Fir

m’s

expe

rien

cein

othe

rpr

ovin

ces

asa

mod

erat

or(F

irm

’sof

prio

rin

vest

men

tsin

othe

rpr

ovin

ces)×

(per

cent

age

ofse

quen

tialF

DI

entr

ies

inth

esa

me

indu

stry

inth

esa

me

prov

ince

)−

0.34

0.03

(0.26)

(0.54)

(Fir

m’s

prio

rin

vest

men

tsin

othe

rpr

ovin

ces)×

(per

cent

age

ofse

quen

tialF

DI

entr

ies

inth

esa

me

indu

stry

inth

esa

me

prov

ince

)2−

1.37

(1.77)

Log

likel

ihoo

d−

20,7

41−

20,7

36−

20,7

32−

20,7

27−

20,7

25−

20,7

13−

20,7

00−

20,7

31−

20,7

31L

ikel

ihoo

dra

tio8,

880

8,88

98,

896

8,90

88,

911

8,93

48,

961

8,89

88,

899

McF

adde

n’s

likel

ihoo

dra

tioin

dex

0.17

60.

177

0.17

70.

177

0.17

70.

177

0.17

80.

177

0.17

7

N=

216,

862.

Stan

dard

devi

atio

nsar

ein

pare

nthe

ses.

Eff

ects

ofin

dust

ryan

dye

ardu

mm

ies

are

nots

how

n.*p

≤0.

05;*

*p≤

0.01

leve

lof

confi

denc

e.



Moreover, in addition to the proportion ofsequential entries, the number of prior sequentialentries could also affect the location choice deci-sions. In the robustness tests, we first ran regressionanalyses with and without the number of the totalprior entries and found consistent results. Then,we included interaction terms relating the linearand squared terms for the proportion of sequentialentries and the number of prior sequential entries.Our predicted effect as in H1 is still strongly sup-ported. The number of prior sequential entries alsohas a positive main effect as expected. Although wedid not hypothesize about it, the number of priorsequential entries negatively moderated the effectof the proportion, also as expected.

DISCUSSION

These findings tend to confirm that a firm’s locationdecisions are indeed influenced by learning from areference group. However, the firm does not nec-essarily copy the most frequently used practices,but learns from decisions made by the more expe-rienced investors in the reference group. Learningfrom experienced firms seems particularly influen-tial in building confidence in imitation. The curvi-linear relationship demonstrated between learningfrom experienced firms and behavior suggests thata firm can learn from the variation in the practicesof its predecessors.

These results go beyond the conventional imi-tation studies by bringing in competition consid-erations. They show that, when a firm observesa high proportion of sequential entries made byexperienced firms, its confidence in imitating themand choosing that location decreases rather thanincreases. It is more inclined to choose another loca-tion, presumably to avoid competition. In addition,when considered in tandem with a firm’s own expe-rience, the curvilinear impact of learning from oth-ers’ experiences becomes weaker for firms withexperience in the host country. Compared withexperience in other host country locations, expe-rience in the specific location is apparently moresalient and transferable, which has a stronger weak-ening effect on learning from others.

These results contribute to our understandingof interorganizational learning by highlightingthe importance of how learning from experiencedtargets affects confidence in imitating. The resultsalso highlight the competitive dimension, which

does not seem to have been considered togetherwith imitation/learning in previous research.Another important finding of the study concernsthe interplay between vicarious learning and afirm’s own experience. The results show that firmswith greater levels of specific types of experi-ence are less sensitive to vicarious learning fromothers. The results thus reinforce the importanceof testing experience effects in a way that alignswith the precise capabilities that a firm needs todevelop.

This study responds to numerous calls for furtherwork on firm-level differences in learning from oth-ers (Baum and Ingram, 1998; Grant, 1996; Ingram,2002; Miner and Anderson, 1999). The resultssuggest some interesting implications for futureresearch. Previous research in the organizationallearning area has mainly focused on firm-levelfactors influencing search behavior, such as patentcitations (Rosenkopf and Nerkar, 2001; Stuart andPodolny, 1996), alliance formation (Beckman,Haunschild, and Philips, 2004; Rothaermel andDeeds, 2004), and interlocking networks (Beckmanet al., 2004). FDI location choice is a new settingfor the study of this phenomenon at the interorgani-zational level, and it would be enlightening if theseresults could be generalized to other empiricalsettings.

The reference groups in this study consisted offirms from the same country (Guillén, 2002; Li,Yang, and Yue, 2007; Yiu and Makino, 2002).Future research might profitably look at how firmslearn from those from different home countries.Future studies might also apply dynamic model-ing techniques such as event history analysis toexamine the competing risks associated with loca-tion choices in foreign entries. The findings hereimply that firms learn both from the practices oftheir predecessors and from their own experience,but further studies are needed to clarify the under-lying processes and the performance implications.In terms of firm experience, future research may gobeyond a firm’s experience in the host country orlocation and use the experience of its top managersto see whether the different experience measureswill yield similar results.

ACKNOWLEDGEMENTS

We would like to thank the Action Editor, TomiLaamanen, and two reviewers for their insights



and guidance throughout the review process. Theauthors contributed equally to the paper.

REFERENCES

Abrahamson E, Rosenkopf L. 1993. Institutional and com-petitive bandwagons: using mathematical modeling asa tool to explore innovation diffusion. Academy of Man-agement Review 18: 487–517.

Chang DK. 1995, 1988. American Business in China.Caravel, Inc: Torrance, CA.

Bastos PV, Greve HR. 2003. Interorganizational learn-ing and the location of manufacturing subsidiaries: isChain migration also a corporate behavior? Advancesin Strategic Management 20: 159–191.

Baum JAC, Ingram P. 1998. Survival-enhancing learningin the Manhattan hotel industry, 1898–1980. Organi-zation Science 44: 996–1016.

Baum JAC, Li SX, Usher JM. 2000. Making the nextmove: how experiential and vicarious learning shapethe locations of chain’s acquisitions. AdministrativeScience Quarterly 45: 766–801.

Baum JAC, Mezias SL. 1992. Localized competition andorganizational failure in the Manhattan hotel industry.Administrative Science Quarterly 37: 580–604.

Beckman CM, Haunschild PR, Philips DJ. 2004. Friendsor strangers? Firm-specific uncertainty, marker uncer-tainty and network partner selection. Organization Sci-ence 15: 259–275.

Broadman H, Sun X. 1997. The distribution of foreigndirect investment in China. World Economists 20:339–361.

Carpenter M, Sanders WG, Gregersen HB. 2001. Bundlinghuman capital with organizational context: the impactof international assignment experience on multinationalfirm performance and CEO pay. Academy of Manage-ment Journal 44: 493–512.

Chang S. 1995. International expansion strategy ofJapanese firms: capability building through sequentialentry. Academy of Management Journal 38: 383–407.

Chang S, Park S. 2005. Types of firms generating networkexternalities and MNCs’ co-location decisions. Strate-gic Management Journal 26: 595–614.

Cheng S. 2007. Structure of firm location choices: anexamination of Japanese greenfield investment inChina. Asian Economic Journal 24: 47–73.

Cheng LK, Kwan YK. 2000. What are the determinants ofthe location of foreign direct investment? The Chineseexperience. Journal of International Economics 51:379–400.

National Bureau of Statistics of China. 1980–1995. ChinaStatistical Yearbook. China Statistical PublishingHouse: Beijing.

Cohen WM, Levinthal DA. 1990. Absorptive capacity: anew perspective on learning and innovation. Adminis-trative Science Quarterly 35: 128–152.

Cyert RM, March JG. 1963. A Behavioral Theory of theFirm. Prentice-Hall: Englewood Cliffs, NJ.

Delios A, Henisz WJ. 2003. Political hazards, experi-ence, and sequential entry strategies: the international

expansion of Japanese firms, 1980–1998. StrategicManagement Journal 24: 1153–1164.

Feinberg SE, Gupta AK. 2004. Knowledge spillovers andthe assignment of R&D responsibilities to foreign sub-sidiaries. Strategic Management Journal 25: 823–845.

Fu J. 2000. Institutions and Investment. The University ofMichigan Press: Ann Arbor, MI.

Gimeno J, Hoskinsson RE, Beal BD, Van WP. 2005.Explaining the clustering of international expansionmoves: a critical test in the U.S. telecommunica-tions industry. Academy of Management Journal 48:297–319.

Grant RM. 1996. Toward a knowledge-based theory of thefirm. Strategic Management Journal 17: 109–122.

Greve HR. 2000. Market niche entry decisions: com-petition, learning, and strategy in Tokyo banking,1894–1936. Academy of Management Journal 43:816–836.

Guillén MF. 2002. Structural inertia, imitation, and foreignexpansion: South Korean firms, and business groups inChina, 1987–1995. Academy of Management Journal45: 509–525.

Hannan MT, Freeman J. 1989. Organizational Ecology.Harvard Business School Press: Cambridge, MA.

Hansen ER. 1987. Industrial location choice in Sao Paulo,Brazil: a nested logit model. Regional Science andUrban Economics 17: 89–108.

Haunschild PR, Miner AS. 1997. Modes of interorga-nizational imitation: the effects of outcome salienceand uncertainty. Administrative Science Quarterly 42:472–550.

Hausman J, McFadden D. 1984. Specification testsfor the multinomial logit model. Econometrica 52:1219–1240.

Henisz WJ, Delios A. 2001. Uncertainty, imitation andplant location: Japanese multinational corporations,1990–1996. Administrative Science Quarterly 46:443–477.

Huber GP. 1982. Organizational information systems:determinants of their performance and behavior. Man-agement Science 28: 135–155.

Huber GP. 1991. Organizational learning: the contributingprocess and the literature. Organizaiton Science 2:88–115.

Ingram P. 2002. Inter-organizational learning. In Com-panion to Organizations, Baum JA (ed). Malden, MA:Blackwell; 642–663.

Johanson J, Vahlne JE. 1977. The internationalizationprocess of the firm: a model of knowledge developmentand increasing foreign market commitments. Journal ofInternational Business Studies 8(1): 23–32.

Kim J, Miner AS. 2007. Vicarious learning from thefailures and near-failures of others: evidence fromthe U.S. commercial banking industry. Academy ofManagement Journal 50(2): 687–714.

Levinthal DA, March JG. 1993. The myopia of learning.Strategic Management Journal 14: 95–112.

Levitt B, March JG. 1988. Organizational learning. AnnualReview of Sociology 14: 319–340.

Li JT. 2008. Asymmetric interactions between foreignand domestic banks: effects on market entry. StrategicManagement Journal 29: 873–893.



Li JT, Yang JY, Yue DR. 2007. Identity, community, andaudience: how wholly owned foreign subsidiaries gainlegitimacy in China. Academy of Management Journal50: 175–190.

March JG, Simon H. 1958. Organizations. John Wiley &Sons: New York.

McKendrick DG. 2001. Global strategy andpopulation-level learning: the case of hard diskdrives. Strategic Management Journal 22: 307–334.

Miner AS, Anderson S. 1999. Industry and population-level learning: organizational, interorganizational, andcollective learning process. Advances in Strategic Man-agement 16: 1–30.

Palmer DA, Jennings PD, Zhou X. 1993. Late adoptionof the multidivisional form by large U.S. corporations:institutional, political and economic accounts. Admin-istrative Science Quarterly 38: 100–131.

Reuber AR, Fischer E. 1997. The influence of the manage-ment team’s international experience on the internation-alization behaviors of SMEs. Journal of InternationalBusiness Studies 28(4): 807–825.

Rhee M, Kim Y, Han J. 2006. Confidence in imitation:Niche-width strategy in the UK automobile industry.Management Science 52(4): 501–513.

Rosenkopf L, Nerkar A. 2001. Beyond local search:boundary-spanning, exploration, and impact in the opti-cal disk industry. Strategic Management Journal 22:287–306.

Rothaermel F, Deeds DL. 2004. Exploration and exploita-tion alliances in biotechnology: a system of newproduct development. Strategic Management Journal25: 201–221.

Ryan TY. 1997. Modern Regression Analysis. John Wiley& Sons: New York.

Schwab A. 2007. Incremental organizational learning frommultilevel information sources: evidence for cross-levelinteractions. Organization Science 18(2): 233–251.

Shaver MJ, Flyer F. 2000. Agglomeration economies,firm heterogeneity, and foreign direct investment inthe United States. Strategic Management Journal 21:1175–1193.

Shrivastava P. 1983. A typology of organizational learningsystems. Journal of Management Studies 20: 7–29.

Singley MK, Anderson JR. 1989. The Transfer of Cogni-tive Skill. Harvard University Press: Cambridge, MA.

Stuart TE, Podolny JM. 1996. Local search and the evo-lution of technological capabilities. Strategic Manage-ment Journal, Summer Special Issue 17: 21–38.

Yelle LE. 1979. The learning curve: historical review andcomprehensive survey. Decision Sciences 10: 302–308.

Yiu D, Makino S. 2002. The choice between joint ventureand wholly owned subsidiary: an institutional perspec-tive. Organization Science 13: 667–683.

Zhou C, Delios A, Yang YJ. 2002. Locational determinantsof Japanese foreign direct investment in China. AsiaPacific Journal of Management 19: 63–86.


Documents

CONFIDENCE IN LEARNING: INTER- AND … · CONFIDENCE IN LEARNING: INTER- AND INTRAORGANIZATIONAL LEARNING IN FOREIGN MARKET ENTRY DECISIONS JIATAO LI,1* CUILI QIAN,2 and FIONA K