Entering New Markets: The Effect of Performance Feedback ... and...(3) performance above aspiration...

Strategic Management JournalStrat. Mgmt. J., 38: 1416–1434 (2017)

Published online EarlyView 25 October 2016 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2561Received 2 September 2014; Final revision received 30 June 2016

ENTERING NEW MARKETS: THE EFFECT OFPERFORMANCE FEEDBACK NEAR ASPIRATION ANDWELL BELOW AND ABOVE ITOHAD REF1,* and ZUR SHAPIRA21 Faculty of Business Administration, Ono Academic College, Kiryat Ono, Israel2 Department of Management, Stern School of Business, New York University, NewYork, New York, U.S.A.

Research summary: This study draws on the resource-based view and the behavioral theoryof the firm to gain new insights about the effect of performance relative to aspiration level (i.e.,performance feedback) on the decision to enter new markets. Results show an inverted U-shapedrelationship between performance both below and above aspiration level, and the probability offirms to enter new markets. That is, when firms are well below or well above their aspiration level,they significantly change their behavior. This article develops a theoretical framework to clarifyand organize these findings.

Managerial summary: This study examines the effect of performance feedback, and particularly,large discrepancies between firm performance and aspiration level on the decision to enter newmarkets. It provides support to the role of performance feedback in affecting the decision to enternew markets, a factor that has received relatively little attention in the extensive literature that hasexamined the inducements of such moves. Results show that, as performance falls below or risesabove aspiration, a firm’s probability of entering new markets increases up to a certain point afterwhich this relationship decreases. This shows that the tendency to enter new markets is differentfor firms that are in the neighborhood of aspiration level compared to those that are well belowor above it. Copyright © 2016 John Wiley & Sons, Ltd.

INTRODUCTION

The strategy literature has given considerable atten-tion to the various inducements for entering newproduct markets and new geographic markets (i.e.,new countries), such as exploiting underutilizedexcess resources, spreading risk, increasing man-agers’ benefits, and accumulating market power(e.g., Chatterjee, 1990; Delios and Beamish, 1999;Grant, Jammine, and Thomas, 1988; Helfat andEisenhardt, 2004; Hitt et al., 2006; Montgomery,

Keywords: entering new markets; resource-based view;behavioral theory of the firm; performance feedback; wellbelow or above aspiration level*Correspondence to: Ohad Ref, Ono Academic College, 104Tzahal Street, Kiryat Ono, 55000, Israel. E-mail: [email protected].

Copyright © 2016 John Wiley & Sons, Ltd.

1994; Penrose, 1959; Rumelt, 1974). Relatively fewof these studies, however, have explored the effectof a firm’s performance as a factor in the deci-sion to enter new markets. These few studies havemainly focused on the effect of poor performance onthe propensity of firms to enter new markets. Theyhave argued that poor performance, which is oftena result of failure to cope with competition in cur-rent markets, or the result of a decline in demand fora firm’s existing products/services, may motivatefirms to search for new markets (new alternatives)for their existing resources (e.g., Campa and Kedia,2002; Matsusaka, 2001; Miller, 2004). Rumelt(1974) called this situation “the escape paradigm.”Poor performance may also motivate firms to enternew markets for other reasons, such as gainingaccess to resources unavailable in their existing

Entering New Markets: the Effect of Performance Feedback 1417

markets that may help them compete in a better way.Studies that have examined the effect of firm

performance on the probability of entering newmarkets have usually operationalized firm perfor-mance using its actual values. That is, they considerabsolute performance. New insights can be drawn,however, from the behavioral perspective, andspecifically, from the behavioral theory of the firm(BTOF) (Cyert and March, 1963). Key to thisperspective is the insight that absolute levels ofperformance are less meaningful than performancerelative to some reference point that actors (man-agers) aspire to meet or exceed (e.g., Cyert andMarch, 1963; Fiegenbaum, Hart, and Schendel,1996; Greve, 2003b; March and Shapira, 1992;Mishina et al., 2010). According to this perspective,using actual values as a measure of firm perfor-mance does not provide a clear delineation of goodor poor performance (Greve, 2003b).1 Instead, man-agers look to benchmarks to form their decisions,such as firm performance relative to historical orpeer performance. They evaluate firm performanceas a positive or negative deviation from target orgoal (aspiration), and this deviation affects theirbehavior.2 We believe that it is this relative perfor-mance that drives the decision to enter new markets.

As implied above, relatively little is known aboutthe effect of performance relative to aspiration level(i.e., performance feedback) on the decision to enternew markets. To the best of our knowledge, onlytwo studies have empirically examined this issue.Chang (1996) has studied entry into new productmarkets, but has not found support for his hypothe-sis that the poorer the performance relative to aspi-ration level, the more likely it is that firms will enternew product markets.3 Lin (2014) has examinedentry into new countries (i.e., enlarging geographicscope) and has found support for her hypothesisthat the more an organization’s performance fallsbelow its aspiration, the greater the probability thatit will enter new geographic markets. With regard

1 For example, for a firm that has historically been achieving a15 percent return on assets, a 5 percent return may mark the lowestpoint in its performance record. In contrast, for a firm with a poorperformance history, a five percent return on assets may be a greatachievement.2 The difference between actual performance and aspiration levelis called the attainment discrepancy, which is considered negativeif firm performance falls short of the level of aspiration, andpositive, if it rises above the level of aspiration (Lewin et al.,1944).3 It should be noted that Chang (1996) has not used a splinespecification to capture performance below and above aspiration.

to performance above aspiration levels, Lin has notfound conclusive results. In short, the effect of per-formance below aspiration level, and particularly,above it on the decision to enter new markets isstill not clear. Furthermore, some scholars have sug-gested that when decision makers, or their firms, arewell below or well above their aspiration level, theymay change their search behavior (e.g., Audia andGreve, 2006; Harris and Bromiley, 2007; Hu, Blet-tner, and Bettis, 2011; March and Shapira, 1992;Shimizu, 2007). However, we are unaware of anyresearch that has studied the effects of large dis-crepancies between firm performance and aspira-tion level (i.e., well below or well above aspirationlevel) on a firm’s propensity to enter new markets.4

Thus, the following questions remain unanswered:(1) Does firm performance relative to aspirationlevel, either below or above it but still in the neigh-borhood of aspiration, affect entry into new markets,and if so, in what way? (2) What are the effects oflarge discrepancies between firm performance andaspiration level (i.e., well below or well above aspi-ration level) on a firm’s propensity to enter newmarkets?

To answer the above questions, we also draw onthe resource-based view (Penrose, 1959; Peteraf,1993; Wernerfelt, 1984). The resource-based view(RBV) suggests three main motivations for enteringnew markets: (1) searching for better alternativesfor existing resources (e.g., Helfat and Eisen-hardt, 2004; Matsusaka, 2001; Penrose, 1960);(2) searching for a way to utilize excess resources(e.g., Chatterjee, 1990; Morck and Yeung, 1991;Penrose, 1959; Teece, 1982); and/or (3) searchingfor new resources to help cope with challengesin existing markets (e.g., Hitt et al., 2006). TheRBV complements the BTOF by explaining, basedon the above three motivations, specifically whythe propensity to enter new markets may eitherincrease or decrease in both underperforming firms(in the neighborhood of aspiration and well belowit) and outperforming firms (in the neighborhoodof aspiration and well above it).

The present study makes two contributions. Itsprimary contribution is to the literature on entryinto new markets. It provides support to the role ofperformance feedback in affecting the decision to

4 Note that the studies that have used absolute performance toexamine the effect of prior firm performance on the decision toenter new markets have not distinguished between poor and “verypoor” and good and “very good” performance.

Copyright © 2016 John Wiley & Sons, Ltd. Strat. Mgmt. J., 38: 1416–1434 (2017)DOI: 10.1002/smj

1418 O. Ref and Z. Shapira

enter new markets, and demonstrates that the dif-ference between performance and aspiration has anonlinear influence on firm behavior. More specif-ically, it shows an inverted U-shaped relationshipbetween performance below the aspiration level andthe probability of firms to enter new markets anda similar relationship between performance aboveaspiration level and the probability of firms to enternew markets. That is, when firms are well below orwell above their aspiration level, they significantlychange their behavior.

The current study also contributes to the BTOFliterature in more thoroughly delineating the effectof performance relative to aspiration level on man-agerial decisions. Most studies that have examinedthe effect of performance relative to aspiration levelon various managerial decisions have not distin-guished between situations that are in the neigh-borhood of aspiration level and those that are wellbelow or well above it. Accordingly, they have usu-ally modeled the effect as a modified spline func-tion (linear above and below aspiration level, butwith potentially different slopes). Only a handful ofstudies have empirically distinguished between sit-uations in the neighborhood of aspiration level andsituations that are well below or well above it, andconsequently, have used nonlinear functions (e.g.,Lehman et al., 2011; Shimizu, 2007). To the best ofour knowledge, they have done this solely for per-formance below aspiration. The present article joinsthese studies and demonstrates that there is a differ-ence between the effect of performance above aspi-ration, while in the neighborhood of aspiration, andthe effect of performance well above aspiration onthe behavior of decision makers. Hence, it extendsthe BTOF to apply not only to situations that arewell below aspiration level, but also to situationsthat are well above it.

THEORY AND HYPOTHESES

As mentioned above, we are interested not only insituations that are in the neighborhood of aspirationlevel (below and above), but also in situationsthat are well below or well above aspiration level.Accordingly, we divide the theory section into fourparts: (1) performance below aspiration still in theneighborhood of aspiration level; (2) performancebelow aspiration while well below aspiration level;(3) performance above aspiration still in theneighborhood of aspiration level; and (4)

performance above aspiration while well aboveaspiration level. Below, we discuss each of thesesituations in turn.

Performance below aspiration still in theneighborhood of aspiration level

According to the BTOF, failure to reach an aspira-tion level leads to a search for action to help improvefirm performance (e.g., Cyert and March, 1963;Greve, 2003b; Lant, Milliken, and Batra, 1992).The BTOF calls this kind of search “problemisticsearch.” It is an effort to identify alternatives tocurrent activities that resolve performance short-falls. Moreover, scholars adhering to the BTOF haveargued that performance below aspiration level notonly makes decision makers search for solutions,but also makes them more likely to accept riskysolutions (Baum et al., 2005; Greve, 2003a), such asentering new markets. The literature has shown thatas firm performance falls below aspiration, there is aprobable increase in organizational change (Greve,1998), strategic reorientation (Lant et al., 1992), andrisk taking (e.g., Bromiley, 1991).

Although the BTOF suggests that failure to reachaspiration level leads to a search for action to helpraise firm performance and accept seemingly riskysolutions, it does not explain specifically why entryinto new markets takes place. An explanation forthat can be found in the RBV, which suggeststhat entry into new markets can help firms addressweaknesses and unfavorable movements in demandconditions (Anand and Singh, 1997; Penrose, 1959;Peteraf, 1993; Wernerfelt, 1984). More specifically,the RBV suggests that poor performance, often theresult of failure to cope with competition in currentmarkets or the result of a decline in demand fora firm’s existing products/services, may motivatefirms to search in new markets for: (1) betteralternatives for their existing resources (e.g., Helfatand Eisenhardt, 2004; Matsusaka, 2001; Penrose,1960); (2) a way to utilize their excess resources(e.g., Chatterjee, 1990; Morck and Yeung, 1991;Penrose, 1959); and/or (3) new resources that canhelp them cope with the challenges in their existingmarkets (e.g., Hitt et al., 2006).

The first motive, searching in new markets forbetter alternatives for existing resources, is basedon the assumption that firms have an incentive touse their existing resources as profitably as possi-ble (Mahoney and Pandian, 1992; Penrose, 1959).Most resources, such as production capacity and



managerial time and attention, must normally beallocated among alternative markets. Hence, the useof these resources in one market usually precludestheir use in other settings, making them subject toopportunity costs. Levinthal and Wu (2010) namedthem “non-scale free resources/capabilities.”5 Itshould be noted that as long as “non-scale freeresources” are not being fully used in current mar-kets, that is, they are in excess, allocating them tonew markets is not subject to opportunity costs.6

Opportunity costs depend on the relative value ofthe resources in question between the existing usesand the new ones. When firm performance is poor,opportunity costs to transfer resources to new mar-kets are relatively low (Wu, 2013). In other words,when the existing resource-market position is lesslikely to be the most profitable option, redeployingfirm resources into new markets may help raise firmperformance (Helfat and Eisenhardt, 2004; Mat-susaka, 2001; Penrose, 1959; Wernerfelt, 1984).7

The second motive, searching in new markets fora way to utilize excess resources, is also based onthe assumption that firms have an incentive to usetheir existing resources as profitably as possible.These excess resources, as mentioned above, arenot subject to opportunity costs. Excess resourcesrepresent the difference between the resources cur-rently possessed by a firm and the resource demandof the current ongoing business (Mishina, Pollock,and Porac, 2004; Penrose, 1959). Failure to copewith competition in current markets and/or a declinein demand for a firm’s existing products/servicesmay often result in a decline in firm performanceand the creation of excess resources, such as excessproduction capacity and excess distribution capac-ity. Excess resources motivate firms to find a wayof using them more fully, including by expansioninto new markets (Chatterjee, 1990; Mahoney andPandian, 1992; Penrose, 1959, 1960; Teece, 1982).

The third motive for poorly performing firms toenter new markets is searching for new resourcesthat can help these firms cope with the challenges in

5 Resources and capabilities are used interchangeably in thisarticle.6 Some resources, such as brand name or patent, may have apublic good-like quality. For them, issues of opportunity costs andresource allocation are inconsequential since they can always have“excess capacity.” Levinthal and Wu (2010) named them “scalefree resources/ capabilities”.7 Redeployment of resources means withdrawing resources fromone economic activity (e.g., one product market) and transferringthem to another (e.g., Helfat and Eisenhardt, 2004; Sakhartov andFolta, 2014).

their existing markets. Examples might be gainingaccess to resources unavailable in the firm’s exist-ing markets like natural resources or raw materi-als, obtaining access to lower-cost resources suchas unskilled labor, and gaining access to new knowl-edge or competences.

Based on the BTOF, specifically on the prob-lemistic search reasoning and the argument thatnegative attainment discrepancy makes decisionmakers more likely to accept risky choices, and onthe RBV reasoning described above, we proposethat as performance falls below aspiration, in theneighborhood of aspiration level, a firm’s probabil-ity of entering new markets increases. However, webelieve that as performance falls farther and fartherbelow aspiration, at a certain point, the increase in afirm’s probability of entering new markets reachesan apex after which it decreases. In other words,we posit that there is an inverted U-shape relation-ship between negative attainment discrepancy andthe probability of entering new markets. The follow-ing subsection presents the theoretical argumentsfor the decrease in a firm’s probability of enteringnew markets well below aspiration level.

Performance below aspiration while well belowaspiration level

Large negative attainment discrepancy indicates avery low “match” between a firm’s resources andthe markets in which it operates. That is, it sig-nals low productivity of resources. It is reasonableto assume that firms in this situation would havedifficulty finding relevant new markets (new oppor-tunities) in which to redeploy their resources, orexploit their underutilized excess resources. In addi-tion, customers in new markets and owners of com-plementary resources in these markets have littleincentive to do business with poorly performingfirms whose future might be uncertain. Poor pro-ductivity of resources also greatly limits the searchadvisability in new markets for resources that canhelp poorly performing firms cope with challengesin their existing markets. Ceteris paribus, the lowerthe value of a firm’s resources in its current mar-kets, the higher the costs of acquiring the additionalresources needed to cope with the challenges inthese markets. In this situation, the appropriate strat-egy may often be liquidation (Helfat and Peteraf,2003; Matsusaka, 2001).

Another explanation for the decrease in the prob-ability of entering new markets where there is



a large negative attainment discrepancy could befound in the BTOF literature. Since falling fartheraway from a target and coming closer to extinc-tion are generally positively correlated, the effectof failure on the propensity to enter new marketsappears to depend on whether the decision makers’attention is focused on their hopes or on their fears(March and Shapira, 1992). That is, it depends onwhether decision makers perceive a negative attain-ment discrepancy as a repairable gap or as a threatto survival (Audia and Greve, 2006). Firms that arein danger of not surviving are more likely to tightentheir controls, conserve resources, repeat previousactions, reduce willingness to take risks, and avoidnew activities, such as entering new markets (Chenand Miller, 2007; D’Aveni, 1989; Staw et al., 1981).March and Shapira (1987, 1992) proposed that deci-sion makers do not direct their attention to a sin-gle reference point. Rather, they switch their focusbetween aspiration and survival levels, and then endup selecting one of them to focus on and guide theirdecision. That is, in the current context, a decreasein performance below aspiration leads to an increasein the propensity to enter new markets if the focus ison the aspiration level, but leads to a decrease in thetendency to enter new markets if the focus is on thesurvival point. March and Shapira (1987, 1992) fur-ther suggested that decision makers are more likelyto focus on the reference point most proximate totheir current position. Hence, as a firm falls fartherand farther below its target, its managers will tendto focus on survival, and therefore, the probabilityof entering new markets will fall.

Furthermore, entry into new markets generallyinvolves a high initial investment, which is at leastpartially specific to the market in question, andtherefore, involves sunk costs (Chatterjee, 1990;Johanson and Vahlne, 1977; Lu and Beamish,2004). Usually, this up-front investment can berecouped only after a certain period of activity inthe new market (Grant et al., 1988). That is, invest-ments in new markets usually bear fruit only in thelong run. Firms that focus on survival tend to payparticular attention to short-run performance issues(D’Aveni, 1989), and therefore, are less likely toenter new markets.

In conclusion, based on the arguments that asperformance falls below aspiration, in the neigh-borhood of aspiration level, a firm’s probability ofentering new markets increases, and the argumentsthat after a certain point, as performance falls fartherand farther below aspiration, a firm’s probability

of entering new markets decreases, we hypothesizethat:

Hypothesis 1 (H1): There is a curvilinear rela-tionship between negative attainment discrep-ancy and the probability of entering new markets.Specifically, as performance falls below aspira-tion, a firm’s probability of entering new marketsincreases up to a certain point after which thisrelationship decreases (i.e., there is an invertedU-shape relationship).

Figure 1 illustrates the curvilinear relationshipbetween negative attainment discrepancy and theprobability of entering new markets. In the neigh-borhood of aspiration level, as performance fallsbelow aspiration, the probability of entering newmarkets increases. After a certain point, as perfor-mance falls farther and farther below aspiration, theprobability of entering new markets decreases.

Performance above aspiration still in theneighborhood of aspiration level

Once a firm achieves its performance targets, themotivation for problemistic search goes down, andmanagers are more likely to become risk averse(Greve, 1998, 2003b; Iyer and Miller, 2008). Thislogic has received support in several empiricalstudies. Scholars have found, for example, that as afirm’s performance rises above aspiration, there is adecrease in the probability of organizational change(e.g., Audia, Locke, and Smith, 2000; Greve, 1998),risk taking (e.g., Wiseman and Bromiley, 1996), andR&D search intensity (e.g., Greve, 2003a).

The RBV complements the above argument andexplains specifically why the probability to enternew markets may go down as performance exceedsaspiration level. The explanation consists of threeparts. First, as performance rises above aspiration,the opportunity costs to transfer resources to newmarkets usually increases. In other words, whenfirm performance is relatively good, which indicatesa satisfactory “match” between a firm’s resourcesand the markets in which it operates, redeployingresources into new markets may be less reward-ing than using them in current markets (Matsusaka,2001). Second, performance above aspiration level,which often results from successfully competing incurrent markets and/or an increase in demand for afirm’s existing products/services, usually involves a



Positive attainment discrepancy Negative attainment discrepancy

Performance above

aspiration still in the

neighborhood of

aspiration level

Performance

above aspiration

while well above

aspiration level

Aspiration

Performance

below aspiration

while well below

aspiration level

Performance below

aspiration still in the

neighborhood of

aspiration level

A f

irm

’s p

rob

abili

ty o

f en

teri

ng

new

mar

kets

Figure 1. Hypothesized effects of negative and positive attainment discrepancy on a firm’s probability of entering newmarkets

highly efficient use of “non-scale free resources,”such as production and distribution capacity. Thatis, there are fewer non-scale free excess resources.Hence, it is reasonable to assume that as perfor-mance exceeds aspiration, the motivation to searchin new markets for a way to utilize these excessresources goes down. Third, it is also reasonableto assume that as performance exceeds aspirationlevel, a firm’s motivation to search in new marketsfor new resources to help it cope with the challengesin its existing markets goes down.

Notwithstanding the above, according to theRBV, there is another compelling motive that mayinduce outperforming firms to enter new mar-kets. As noted above, performance above aspi-ration points to a relatively good position for afirm’s resources. That is, a good “match” betweena firm’s resources and the markets in which it oper-ates. In many cases, the resources that underliethis good “match” have a public good-like qual-ity. For these resources, issues of opportunity costsand resource allocation are inconsequential sincethey can always have “excess capacity.” Levinthaland Wu (2010), as mentioned previously, namedthem “scale free resources/capabilities.” Examplesof such resources are brand name, patents, or rela-tionships with customers. The existence of a pool ofunderutilized resources may motivate firms to finda way of using them more fully by entering intonew markets (e.g., Chatterjee, 1990; Mahoney andPandian, 1992; Morck and Yeung, 1991; Penrose,1959, 1960; Teece, 1982). However, the existenceof underutilized resources is not a sufficient condi-tion for entering new markets. These resources mustalso be valuable and endow the firm with a positive

marginal return from entering a new market (Helfatand Lieberman, 2002). As implied above, the posi-tive attainment discrepancy indicates that this mightbe the case, that is, that the resources might be valu-able. The larger the positive discrepancy betweenfirm performance and aspiration level, the strongerthe indication.

A similar reasoning can be found in the BTOF. Itis based on slack search. Slack has been defined as“the difference between the potential performanceof an organization and the performance actuallyachieved” (Levinthal and March, 1981: 309). Slacksearch occurs when potential performance exceedsactual performance, that is, when a firm’s resourcesare not fully used. According to the BTOF, slackincreases the probability of experimentation, whichcan result in identifying and pursuing new opportu-nities (Iyer and Miller, 2008; Levinthal and March,1981). Moreover, the presence of excess resourcesalso appears to affect risk taking. Where organiza-tional slack is plentiful, it tends to lead to relaxationof control and increased experimentation, and thus,to relatively high levels of risk taking (March andShapira, 1992; Singh, 1986). That is, initiativespreviously deemed too risky, including entries intonew markets, are likely to be approved. Becauseslack is normally present in relatively successfulfirms, the BTOF has proposed that these firms, andnot just relatively unsuccessful ones, are expectedto invest in innovative activities (Cyert and March,1963: 189; Levinthal and March, 1981: 314; Pitelis,2007: 480).

While the BTOF explains why slack may enableexperimentation among outperforming firms, suchas entering into new markets, it does not explain



why slack may actually induce experimentation(Pitelis, 2007: 480). Such an explanation is pro-vided by the RBV. More specifically, as mentionedabove, the RBV claims that the existence of apool of underutilized resources, primarily, scale freeresources, motivates firms to find a way of usingthem more fully, including by entering new mar-kets. This is because they can provide services atnear-zero or relatively low marginal costs since theyhave already been hired to provide services for aspecific amount of time (Pitelis, 2007). This argu-ment is based on the assumption that, in general,firms pursue maximum feasible long-term profits(Penrose, 1959).

The above arguments are based on the utilizationof scale free/slack resources. However, even if a newmarket entry is based on exploitation of scale freeresources, it normally requires the use of additionalresources that are subject to opportunity costs, suchas managerial time and attention, to complementthe scale free resources. Recent empirical work hasprovided suggestive evidence that this is usuallythe case. It has shown that, ordinarily, there areopportunity costs when entering into new markets.Roberts and McEvily (2005), for example, havefound that entering a new pharmaceutical productmarket reduces a firm’s performance in its currentmarkets. Similarly, Wu (2013) has shown that prod-uct diversification is associated with a performancedecrease in current markets. Yet, in the neighbor-hood of aspiration level, these costs are relativelylow, which means that it is possible to find a bet-ter alternative (i.e., market) for resources that aresubject to opportunity costs. In other words, in theneighborhood of aspiration level, the relatively lowopportunity costs may allow expansion into newmarkets that is based on the use of valuable scalefree resources.

Based on the above arguments, we propose thatas performance rises above aspiration, in the neigh-borhood of aspiration level, a firm’s probabilityof entering new markets increases. However, webelieve that as performance rises farther and fartherabove aspiration, at a certain point, the increase in afirm’s probability of entering new markets reachesan apex after which it decreases. In other words,we posit that there is an inverted U-shape relation-ship between positive attainment discrepancy andthe probability of entering new markets. The follow-ing subsection presents the theoretical argumentsfor the decrease in a firm’s probability of enteringnew markets well above aspiration level.

Performance above aspiration while well aboveaspiration level

As mentioned above, even if a new market entryis based on the exploitation of scale free resources,such entry usually requires the use of resources thatare subject to opportunity costs to complement thescale free resources. Large positive attainment dis-crepancy indicates a very good “match” between afirm’s resources and the markets in which it oper-ates. It is reasonable to assume that a firm withsuch a relatively good match will have difficultyfinding relevant new opportunities (i.e., alternativemarkets) for resources that are subject to opportu-nity costs, such as managerial time and attention. Inother words, any alternative resource-market posi-tion is more likely to be inferior to the existing one,since the opportunity costs to use these resourcesin other markets are very high. In fact, for firmsthat are well above aspiration level, these resourcesare likely to constitute a “bottleneck” constrainingexpansion into new markets. It is therefore reason-able to assume that entering into new markets mayharm the performance of at least some of the firmswhose performance is well above their aspirationlevel.

In conclusion, based on the arguments that asperformance rises above aspiration, in the neigh-borhood of aspiration level, a firm’s probability ofentering new markets increases, and the argumentsthat after a certain point, as performance rises far-ther and farther above aspiration, a firm’s probabil-ity of entering new markets decreases, we hypothe-size that:

Hypothesis 2 (H2): There is a curvilinear rela-tionship between positive attainment discrep-ancy and the probability of entering new markets.Specifically, as performance rises above aspira-tion, a firm’s probability of entering new marketsincreases up to a certain point after which thisrelationship decreases (i.e., there is an invertedU-shape relationship).

Figure 1 also illustrates the curvilinear relation-ship between positive attainment discrepancy andthe probability of entering new markets. In sum,we propose that there is a curvilinear relationshipbetween both negative and positive attainment dis-crepancy and a firm’s probability of entering newmarkets.



DATA, MEASURES, AND METHODS

Data and time frame

Our primary sources of data are the annual Com-pustat database, the Compustat segment (markets)database, and firms’ annual financial statements.As mentioned above, our main objective for thisstudy is to examine the effect of performance feed-back, and particularly, large discrepancies betweenfirm performance and aspiration level on the prob-ability of entering new markets. It usually takesa fair amount of time, generally several years,for firms to respond to such feedback. We usea two-year window to observe entry into newmarkets because firms’ annual segment reports,under the Statement of Financial Accounting Stan-dards (SFAS 131) defined in 1997, present com-parisons of sales in each specific segment onlytwo years back. That is, we are restricted in ourability to analyze longer periods of time such asthree or four years. To enlarge our sample and toensure that our results will not be influenced byspecific-year effects, we use five time windows:1999–2001, 2001–2003, 2003–2005, 2005–2007,and 2007–2009, that is, 1999, 2001, 2003, 2005,and 2007 are the five base years. We retain firmsthat are incorporated in the U.S.A. We focus onAmerican firms to ease interpretation of the resultssince the sample is derived from a single home mar-ket. We select manufacturing firms (North Ameri-can Industrial Classification System [NAICS] code311111–339999) with base-year sales greater than$15 million in order to avoid the problem of com-paring very small firms with relatively small seg-ments to much larger firms with larger segments.We eliminate observations for which the total ofeither the product or the geographic segment salesis not within one percent of the total reported firmsales for that year. This requirement ensures thatall firm sales are allocated to individual productand geographic segments. Eventually, we are leftwith 3,926 observations (1,413 different firms),of which 362 are instances of entry into newmarkets (9.2%).

Variables

Dependent variable

The dependent variable is Entry into new marketst(product and geographic). It is a binary indicatorequal to one if the firm adds at least one new market

during a time window, and zero otherwise.8 Toidentify whether a firm has added a new market, wemanually compared the list of segments of each firmin its annual report of the last year in a given timeperiod to its list of segments in the base year of theperiod.

Independent variable

As mentioned above, the BTOF asserts that man-agers evaluate firm performance as a positive ornegative deviation from target (aspiration), and thisdeviation affects their behavior (Cyert and March,1963; Greve, 2003b). The BTOF stresses two typesof comparisons. The first is a comparison with thefirm’s own performance history (Cyert and March,1963; Levinthal and March, 1981). A firm’s recentperformance history is a benchmark against whichthe firm evaluates its current performance. Pastperformance is an indicator of how well a firmcan perform and can easily become a standard forhow well a firm should perform (Greve, 2003b).The second is a social comparison with the per-formance of comparable others (Cyert and March,1963; Festinger, 1954). In this case, the recentperformance of comparable firms is the benchmarkagainst which a firm evaluates its current level ofperformance. This raises the question: What is thereference group to which the focal firm comparesitself? For example, firms might choose to comparethemselves to subgroups in their industry basedon size, technology, or strategy, or may even com-pare themselves to corresponding firms in otherindustries. In general, external observers (suchas researchers) may have difficulty ascertainingthe appropriate social referent point (Washburnand Bromiley, 2012). This difficulty is probablyone of the main reasons that studies using socialcomparison have reported results that are consid-erably less consistent compared to studies basedon past performance (historical) aspiration (Boyleand Shapira, 2012). Following this observationabout social aspiration, and similar, for example,to Vidal and Mitchell (2015), and Moliterno andWiersema (2007), we focus in this study on com-paring firm performance with its own performancehistory.

8 Because the average number of new markets in a given two-yearwindow is relatively low (1.32), we have coded one or more entriesinto markets during a period as an event. For robustness test, weused the number of new markets as our dependent variable andre-estimated the models. The main results did not change.



The difference between a firm’s actual perfor-mance and its own performance history, that is, afirm’s attainment discrepancy, is our independentvariable. Firm performance was measured in periodt–1 (the base year), and the aspiration level wasobtained from the period (t–2) (one year prior tobase year). The decision to enter new markets isusually taken by corporate managers. Thus, over-all performance measures, such as return on assets(ROA) and return on sales (ROS), may be relevant tothese managers as they assess performance relativeto aspiration (Gavetti et al., 2012). We have cho-sen ROA (calculated as net income divided by totalassets) as our preferred performance proxy, becauseROA gives comparability across studies since it isthe most common measure in studies of perfor-mance effects on change and risk taking (Greve,2010), including in the two studies that have exam-ined the effect of performance feedback on enteringnew markets (Chang, 1996; Lin, 2014).9

In our hypotheses, we conjecture that when firmsare well below or well above aspiration level, theychange their search behavior. That is, we hypothe-size that the relationship between both negative andpositive attainment discrepancy and a firm’s prob-ability of entering new markets is curvilinear. Toexamine these propositions, we include a quadraticof firms’ negative/positive attainment discrepancyin our models.

Control variables

We control for firm size, firm age, firm’s degree ofproduct and geographic search experience, industrygrowth rate, and industry concentration. Similarto, for example, Mishina et al. (2010), we opera-tionalize firm size (Sizet–1, in logarithmic form) asthe number of employees in a firm (in thousands).Previous studies have shown that a firm’s size mayinfluence its willingness to enter new markets (e.g.,Chang, 1996) and also decision makers’ risk toler-ance (Audia and Greve, 2006). Controlling for age(Aget–1, in logarithmic form) is important becauseolder firms have more experience, and the oppor-tunity to experiment with different levels and typesof resources. A firm’s degree of product and geo-graphic search experience may affect its propensityto enter new markets. As a firm’s scope in a specific

9 For robustness test, we used ROS (calculated as net incomedivided by sales) and EBITDA (calculated as EBITDA dividedby sales) instead of ROA. The main results did not change.

domain increases, it is reasonable to assume that itaccumulates more domain-specific search routines(Ref, 2015). Our proxies for these two variablesare the number of product and geographic marketsthat each firm served in the base year, denotedas Product search experiencet–1 and Geographicsearch experiencet–1, respectively. Previous studieshave also argued that firms operating in industriescharacterized by few growth opportunities tend toexpand by entering new markets (e.g., Delios andBeamish, 1999). In other words, in industries withslow growth rates, entry into new markets may bethe only prospect for improving the profitabilityof the firm. Industry growth rates (Growtht–1) areobtained from the Annual Survey of Manufacturersby the U.S. Census Bureau and are allowed tobe negative. Exploitation of market power mayconstitute another motivation for expansion intonew markets (Montgomery, 1994). Therefore, wealso control for industry concentration. Marketconcentrations were obtained from the EconomicCensus by the U.S. Census Bureau, and reportthe Herfindahl-Hirschman indexes (HHIt–1) forthe 50 largest companies in each industry. Wenormalize the data to range between zero and one,where zero represents perfect competition and onerepresents monopoly. As mentioned above, we usefive periods of time; therefore, we include fourdummy variables to take into account any timetrends.

Similar to previous studies (e.g., Mishina et al.,2004; Ref, 2015), we investigate the entry into newmarkets of a sample of firms over a particular timeperiod, as a function of their slack/excess resourcesat the beginning of that period. Slack/excessresources represent the difference between theresources currently possessed by a firm and theresource demand of the current ongoing business(Greve, 2003a; Nohria and Gulati, 1996; Penrose,1959). Because resource demands of current ongo-ing business are unobserved, following previousstudies (e.g., Mishina et al., 2004), we use themedian value of the firm’s industry as a proxy forresource demands of current ongoing business.10

10 Industries are defined at the six-digit industry level in theNAICS, conditional on a minimum of 15 firms in the six-digitcode. If there are less than 15 firms, then we define an industrybased on the five-digit industry codes. If there are less than 15firms under this definition, then we move to four-digit codes andso forth. The 15 firm minimum is required in order to have a fairlyaccurate measure of the median industry values. For robustnesstest, we changed the minimum condition of 15 firms to 10 and



The value of a firm’s slack/excess resources wascalculated as the difference between resourcesat the firm’s level and the median core industryvalues. Positive values indicate the existence ofslack/excess resources and negative values indicateshortages in resources. Moreover, in using industrymedians, we also take industry-level differences inresource requirements into account. We use severalproxies for slack/excess resources:

First, we use two variables to capture slack/excess financial resources. The first proxy is afirm’s Available financial resourcest–1. It is basedon a firm’s current ratio and calculated as a firm’scurrent assets divided by current liabilities minusthe median core industry value. Current ratio mea-sures whether or not a firm has enough resources topay its debts over the following months. The secondproxy is Potential financial resourcest–1. It is basedon a firm’s leverage and is calculated as total liabili-ties divided by shareholder equity minus the mediancore industry value. Leverage captures the firm’sborrowing ability. A firm with a high debt-to-equityratio has a relatively low ability to obtain addi-tional funds through incurring debt, and thus, haslittle slack/excess. Second, for production excesscapacity (CAPt–1), similar to Kumar (2009), ourproxy is a firm’s net property plan and equipmentnormalized by sales, relative to its core industrynet property plan and equipment normalized bycore industry sales. Third, for excess technologicalcompetencies (R&Dt–1), our proxy is the differencebetween a firm’s R&D intensity (R&D expensesdivided by sales) and the industry median value.

Models

As mentioned above, the dependent variable is abinary indicator equal to one if a firm adds at leastone new market during a time window, and zerootherwise. Hence, we use a probit model to predictthe probability of entering new markets in a giventwo-year window.11,12 As estimated coefficients are

re-estimated the models. The main results did not change. Inanalyses not reported here, we used the mean of the firm’s industryinstead of the median as a proxy for the resource demand andre-estimated the models. The results did not change.11 For robustness test, we used the logit model instead of theprobit. The results did not change.12 As mentioned above, our database includes five time windows.Taking the full data set about 16 percent of the firms appearin all five windows, about 18 percent appear in four, about20 percent appear in three, another 20 percent appear in two, and

not obvious to interpret in the probit model (otherthan the sign and statistical significance), we presentaverage marginal effects (in Table 2, and Figures 2and 3) that reflect the impact of changing theindependent variable by one unit on the probabil-ity that the firm chooses to enter new markets.13

We estimate a model allowing for different attain-ment discrepancy marginal effects depending onwhether firms were performing below or above theiraspiration level while all other marginal effects wereconstrained to be equal across all firms (a splineapproach). We also use the split approach (separateregressions for underperforming and outperformingfirms), which is a more flexible approach that allowsfor all the parameters to vary between under- andoutperforming firms. In addition, we note that stan-dard errors are clustered by firm.

RESULTS

Table 1 presents descriptive statistics and correla-tions for the sample. On average, the firms in thesample are 38.1 years old (Aget–1), have 7.16 thou-sand employees (Sizet–1), and have 1.97 productmarkets (Product search experiencet–1) and 3.29geographic segments (Geographic search experi-encet–1) in the base year. Although not presentedhere, conditional on adding a new market, the aver-age number of new markets in a given time periodis 1.32, and the maximum is 5.

Table 2 presents the average marginal effectsfrom the probit regressions for underperformingand outperforming firms. As mentioned above,we use both the spline approach and the splitapproach. Model 1 presents the results for thespline approach, and Models 2 and 3 present the

the remaining 26 percent appear only once. However, we do notmake use of panel data techniques in our analysis for a numberof reasons. First, because we need at least two observations perfirm, we lose a significant number of observations when movingto panel data techniques. Moreover, because we further divide ourdata set into subcategories based on our independent variables (thesplit approach), the necessity to have at least two observationswithin each subcategory becomes particularly stringent.13 The average marginal effect for a variable x is generated by firstcalculating the marginal effect for each firm using their specificvalue of x and then averaging over the entire sample. This is incontrast to evaluating the marginal effect at the mean value of x.The advantage of the former methodology is that we do not haveto take a stand on whether the mean value of x is representative ofa typical firm, that is, the mean value of x may not actually existin the sample (Hoetker, 2007). To the extent that x approximatesa normal distribution, the two methodologies give very similaranswers.



Tabl

e1.

Des

crip

tive

stat

istic

san

dco

rrel

atio

ns

Var

iabl

eM

ean

S.D

12

34

56

78

910

1112

13

1.E

ntry

into

new

mar

kets

t0.

090.

281

2.(P

erfo

rman

ce-a

spir

atio

nle

vel)

t–1,<

0−

0.05

0.15

−0.

021

(0.1

88)

3.(P

erfo

rman

ce-a

spir

atio

nle

vel)

t–1,>

00.

050.

120.

050.

131

(0.0

02)(0.0

00)

4.Si

zet–

1a

7.16

19.9

6−

0.05

0.04

−0.

081

(0.0

02)(0.0

05)(0.0

00)

5.A

get–

1a

38.1

532

.83

−0.

070.

10−

0.14

0.34

1(0.0

00)(0.0

00)(0.0

00)(0.0

00)

6.G

row

tht–

1−

0.02

0.20

−0.

030.

06−

0.01

0.01

−0.

041

(0.0

36)(0.0

00)(0.6

50)(0.3

53)(0.0

07)

7.H

HI t

–1

0.26

0.20

0.05

−0.

060.

040.

12−

0.02

−0.

071

(0.0

01)(0.0

00)(0.0

22)(0.0

00)(0.3

37)(0.0

00)

8.A

vaila

ble

finan

cial

reso

urce

s t–

1

0.65

2.95

0.04

0.03

−0.

01−

0.14

−0.

150.

020.

031

(0.0

26)(0.0

68)(0.5

69)(0.0

00)(0.0

00)(0.1

06)(0.0

61)

9.Po

tent

ialfi

nanc

ial

reso

urce

s t–

1

1.69

74.8

40.

050.

010.

040.

000.

010.

010.

04−

0.02

1(0.0

02)(0.7

36)(0.0

18)(0.9

67)(0.4

77)(0.6

48)(0.0

08)(0.2

89)

10.

CA

P t–

10.

040.

340.

01−

0.01

−0.

030.

01−

0.00

0.02

0.00

0.02

−0.

021

(0.4

57)(0.4

16)(0.0

46)(0.7

36)(0.8

52)(0.1

65)(0.8

83)(0.1

47)(0.2

33)

11.

R&

Dt–

10.

030.

290.

00−

0.09

0.04

−0.

04−

0.10

0.05

−0.

010.

11−

0.01

0.66

1(0.9

53)(0.0

00)(0.0

06)(0.0

06)(0.0

00)(0.0

01)(0.3

25)(0.0

00)(0.3

10)(0.0

00)

12.

R&

Dt–

11.

971.

26−

0.01

0.07

−0.

110.

430.

32−

0.02(

0.01

−0.

180.

01−

0.01

−0.

091

(0.4

66)(0.0

00)(0.0

00)(0.0

00)(0.0

00)

0.28

1)(0.4

03)(0.0

00)(0.3

77)(0.3

46)(0.0

00)

13.

Geo

grap

hic

sear

chex

peri

ence

t–1

3.29

2.02

0.09

−0.

00−

0.02

0.16

0.12

−0.

020.

06x

−0.

070.

050.

02−

0.05

0.20

1(0.0

00)(0.7

93)(0.1

26)(0.0

00)(0.0

00)(0.1

07)(0.1

07)(0.0

00)(0.0

00)(0.1

29)(0.0

02)(0.0

00)

aT

heco

rrel

atio

nsar

esh

own

for

the

vari

able

sin

thei

rlo

gari

thm

icfo

rm.N

=3,

926.

P-v

alue

sin

pare

nthe

ses.



Table 2. Average marginal effects from Probit regressions for underperforming and outperforming firms.a The dependentvariable is entry into new marketsb

Model 1 Model 2 Model 3Spline

approachSplit approach—

underperforming firmsSplit approach—

outperforming firms

1. (Performance-aspiration level)t–1 < 0

−0.324 (0.000)c −0.302 (0.000)c

2. (Performance-aspiration level)t–1 > 0

0.207 (0.002)c 0.163 (0.010)c

3. Sizet–1 (in logarithmic form) −0.014 (0.000) −0.011 (0.024) −0.017 (0.001)4. Aget–1 (in logarithmic form) −0.016 (0.012) −0.017 (0.042) −0.013 (0.124)5. Growtht–1 −0.047 (0.045) −0.060 (0.062) −0.043 (0.180)6. HHIt–1 0.047 (0.034) 0.083 (0.003) 0.007 (0.811)7. Available financial resourcest–1 0.002 (0.065) 0.005 (0.003) −0.003 (0.302)8. Potential financial resourcest–1 0.000 (0.270) 0.000 (0.386) 0.000 (0.394)9. CAPt–1 0.043 (0.030) 0.050 (0.053) 0.037 (0.274)10. R&Dt–1 −0.053 (0.046) −0.090 (0.046) 0.015 (0.639)11. Product search experiencet–1 0.010 (0.023) 0.006 (0.345) 0.014 (0.015)12. Geographic search

experiencet–10.012 (0.000) 0.012 (0.000) 0.012 (0.000)

Number of observations 3926 1968 1958Wald chi2 105.23 (0.000) 77.50 (0.000) 52.34 (0.000)

a P-values in parentheses. Period dummies are not shown.b Entry into new markets is a binary indicator equal to one if the firm adds at least one new market during a time window, zero otherwise.c Note that there is only one marginal effect for attainment discrepancy, not separate effects for attainment discrepancy and its quadraticterm. See Footnote 14.

results for the split approach. Model 2 showsthe results for underperforming firms, and Model3, for outperforming firms. As mentioned in theindependent variable subsection, we hypothesizethat the relationship between both negative andpositive attainment discrepancy and a firm’s prob-ability of entering new markets is curvilinear. Toexamine these propositions, we include a quadraticof firms’ negative/positive attainment discrepancyin our models. The average marginal effect ofattainment discrepancy and its quadratic termare shown in Table 2. It is noteworthy at thispoint that there is only one marginal effect forattainment discrepancy, and not separate marginaleffects for attainment discrepancy and its quadraticterm.

For underperforming firms, the attainment dis-crepancy average marginal effects (Row 1) are bothnegatively and significantly related to entry intonew markets (the p-values in both Models 1 and 2are 0.000). For outperforming firms, the attainmentdiscrepancy average marginal effects (Row 2)are both positively and statistically associatedwith expansion into new markets (the p-values inModels 1 and 3 are 0.002 and 0.010, respectively).However, the average marginal effect presents only

one point of interest and does not tell us whetherthe relationship between the independent variableand the dependent variable is curvilinear. That is,we hypothesize that certain values of attainmentdiscrepancy positively affect the probability ofentering new markets, while other values have anegative impact. This behooves us to evaluate themarginal effect of attainment discrepancy and itsquadratic term at a number of different values ofattainment discrepancy in order to measure howthe effect differs based on its magnitude. In otherwords, we calculate the average marginal effectof attainment discrepancy and its quadratic termat different points along its range.14 Because eachvalue of attainment discrepancy provides a differentmarginal effect, the most concise and aesthetic way

14 The average marginal effect with respect to x1, for example, isequal to

1∕n

(n∑1

𝜑 (x′𝛽)(𝛽x1 + 2𝛽x21 x1

)),

where 𝜑 is the pdf of the normal distribution in the case of probit.The marginal effect is evaluated for each observation and then theaverage of these marginal effects is what we report in Table 2.In the cases where we evaluate the marginal effect for specificvalues of attainment discrepancy, we replace the observation’sactual value of attainment discrepancy with the value of interest.



-.5

0.5

Ave

rage

Mar

gina

l Effe

cts

-.89 -.78 -.67 -.56 -.45 -.34 -.23 -.12 -.01

Negative Attainment Discrepancy

50%75%90%95%99%

Figure 2. Average marginal effects of firms’ negativeattainment discrepancy for firms that have entered new

markets with 90 percent confidence intervals15

-.4

-.2

0.2

.4

Ave

rage

Mar

gina

l Effe

cts

.01 .135 .26 .385 .51 .635 .76 .885 1.01

Positive Attainment Discrepancy

50% 75% 90% 95% 99%

Figure 3. Average marginal effects of firms’ positiveattainment discrepancy for firms that have entered new

markets with 90 percent confidence intervals

to present these results is with a graph (Hoetker,2007; Wiersema and Bowen, 2009).16 Figures 2 and3 illustrate these average marginal effects for under-performing and outperforming firms, respectively.17

Hypothesis 1 (H1) predicts that as performancefalls below aspiration, a firm’s probability of enter-ing new markets increases up to a certain point afterwhich this relationship decreases (i.e., there is aninverted U-shape relationship). Figure 2 presents

15 The average marginal effect in Figures 2 and 3 are for theattainment discrepancy and its quadratic term at different pointsalong the range. In all figures, the percentages at the bottom of thefigures indicate the cumulative percentage of observations.16 For brevity and aesthetic reasons, we present the results graph-ically, but numeric tables (regression results) are available onrequest.17 Figures 2 and 3 are based on the split approach. For robustnesstest, we drew the results based on the spline approach. The figuresobtained from the two approaches are nearly identical.

.05

.1.1

5

Pre

dict

ed P

roba

bilit

y

-1 -.8 -.6 -.4 -.2 0

Negative Attainment Discrepancy

50% 75%90%95%99%

Figure 4. The effect of negative attainment discrepancyon the probability of underperforming firms to enter new

markets

the average marginal effects of firms’ negativeattainment discrepancy. In the neighborhood ofaspiration (the values on the right side of the graphnear zero), the average marginal effects are nega-tive and statistically significant (at -0.01 the p-valueis 0.000, at -0.12 the p-value is 0.001, at -0.23 thep-value is 0.004, and at -0.34 the p-value is 0.087).The average marginal effects gradually increase asthe negative attainment discrepancy increases, andafter a certain point, they become positive and sta-tistically significant (at -0.67 the p-value is 0.025, at-0.78 the p-value is 0.001, and at -0.89 the p-valueis 0.000). That is, when the negative attainmentdiscrepancy is very large, underperforming firmsreduce their propensity to enter new markets. Theseresults strongly support H1.

Figure 4 depicts the effect of negative attain-ment discrepancy on the probability of underper-forming firms to enter new markets. The presenta-tion method in this figure is slightly different fromthe one used in Figure 2, but essentially it showsthe same information. Generally speaking, in closeproximity to the aspiration level (near zero), a firm’spredicted probability to enter into a new marketis about six percent. As performance falls belowaspiration, the predicted probability increases. Thepredicted probability reaches its peak at about16 percent. Then it falls to a lower level than its levelin close proximity to aspiration level (near zero).

Hypothesis 2 (H2) predicts that as performancerises above aspiration, a firm’s probability ofentering new markets increases up to a certain pointafter which this relationship decreases (i.e., there isan inverted U-shape relationship). Figure 3 showsthe average marginal effects of firms’ positiveattainment discrepancy. In the neighborhood of



.06

.08

.1.1

2.1

4

Pre

dict

ed P

roba

bilit

y

0 .2 .4 .6 .8 1

Positive Attainment Discrepancy

50% 75% 90% 95% 99%

Figure 5. The effect of positive attainment discrepancyon the probability of outperforming firms to enter new

markets

aspiration (the values on the left side of the graphnear zero), the average marginal effects are positiveand statistically significant (at 0.01 the p-value is0.006, at 0.135 the p-value is 0.020, and at 0.26the p-value is 0.044). The average marginal effectsgradually decrease as the positive attainment dis-crepancy increases, and after a certain point, theybecome negative and statistically significant (at0.885 the p-value is 0.101, and at 1.01 the p-valueis 0.018). That is, when the positive attainmentdiscrepancy is very large, outperforming firmsreduce their propensity to enter new markets. Theseresults strongly support 2.18

Figure 5 depicts the effect of positive attain-ment discrepancy on the probability of outperform-ing firms to enter new markets. As in the case ofFigure 4, the presentation method in this figure isslightly different from the one used in Figure 3, butessentially it shows the same information. Gener-ally, in close proximity to the aspiration level (nearzero), a firm’s predicted probability to enter intoa new market is about seven percent. As perfor-mance rises above aspiration, the predicted proba-bility increases. The predicted probability reachesits peak at about 13 percent. Then it falls to a levelroughly similar to its level in close proximity toaspiration level (near zero).

Regarding the control variables (Table 2), asmentioned above, Model 1 presents the results for

18 Hu et al. (2011) have used an S-shaped function in their simu-lation for risk taking among outperforming firms. For robustnesstest, we included cubic terms (X3) of firms’ negative and specifi-cally positive attainment discrepancy in our models to test whetherthe relationships are perhaps S-shaped rather than those that wehave reported. The results remained the same as those presentedabove, thus providing stronger support for our hypotheses.

the spline approach in which the average marginaleffects are equal across all firms, and Models 2 and3 present the results for the split approach, whichis a more flexible method that allows for all theparameters to vary between under- and outperform-ing firms. Thus, if firms above and below aspira-tion differ in observable ways, the split approachallows us to better control for these differences.Firm size (Sizet–1), firm age (Aget–1), and indus-try growth rate (Growtht–1) are negatively related toentry into new markets (for underperforming firmsthe p-values are 0.024, 0.042, and 0.062, respec-tively. For outperforming firms the p-values are0.001, 0.124, and 0.18, respectively). Exploitationof market power (HHIt–1: Herfindahl-HirschmanIndex) is positively related to expansion into newmarkets; however, there is a significant difference inthe p-value between underperforming firms (0.003)and outperforming firms (0.811). This is an inter-esting result because scholars have usually arguedthat firms from more concentrated industries mayuse their power to enter new markets (e.g., Mont-gomery, 1994). Our results suggest that mainly“unsuccessful” firms use this power, but not allfirms in an industry. Available financial resourcesare positively related to expansion into new marketsfor underperforming firms (the p-value is 0.003).Excess in production capacity (CAPt–1) is positivelyrelated to entry into new markets (for underperform-ing firms the p-value is 0.053, and for outperformingfirms the p-value is 0.274). Excess in technologi-cal competencies (R&Dt–1) is negatively related toexpansion into new markets for underperformingfirms (the p-value is 0.046), and positively related toexpansion into new markets for outperforming firms(however, the p-value is 0.639). This is also an inter-esting result because scholars have usually arguedthat high technological capabilities may inducefirms to enter into new markets (e.g., Delios andBeamish, 1999; Silverman, 1999). Our results maysuggest that, for underperforming firms, increasingresearch and development efforts substitutes enter-ing into new markets. Finally, a firm’s previousproduct and geographic search experience is gen-erally positively related to entry into new markets.

DISCUSSION AND CONCLUSION

Our main objective for this study is to examine theeffect of performance feedback, and particularly,large discrepancies between firm performance and



aspiration level on the decision to enter new mar-kets. We provide support to the role of performancefeedback in affecting the decision to enter newmarkets, a factor that has received relatively littleattention in the extensive literature that has exam-ined the inducements of such moves (e.g., Chatter-jee, 1990; Delios and Beamish, 1999; Hitt et al.,2006; Matsusaka, 2001; Montgomery, 1994; Pen-rose, 1959). The results show that, as performancefalls below/rises above aspiration, a firm’s probabil-ity of entering new markets increases up to a cer-tain point after which this relationship decreases. Inother words, there is an inverted U-shape relation-ship between performance both below and aboveaspiration level and the probability of firms to enternew markets. This shows that the tendency to enternew markets is different for firms that are in theneighborhood of aspiration level compared to thosethat are well below or above it.

Despite the similarity in the effect of negativeattainment discrepancy and positive attainmentdiscrepancy on the probability of entering into newmarkets, it seems that there are some differencesin firm behavior in the two situations. Specifically,while in close proximity to aspiration level, thepredicted probability of under- and outperformancefirms to enter new markets is similar (see Figures 4and 5), it seems that the inflection point for under-performing firms is at a higher level compared tothe inflection point for outperforming firms. That is,there is a steeper relationship between performancerelative to aspiration level and the probability offirms to enter new markets for underperformingfirms. Furthermore, it seems that extreme negativeperformance relative to aspiration level has amore discouraging impact on the propensity offirms to enter new markets than extreme positiveperformance relative to aspiration level.19 Theseresults are consistent with the argument thatorganizations’ decision makers react more stronglyto threats than to opportunities (Kahneman andTversky, 1979). Hence, they tend to be more opento change when performance is below aspirationbut still in the neighborhood of aspiration levelthan when performance is above aspiration butstill in the neighborhood of aspiration level (Baum

19 Because Figures 4 and 5 are derived from separate regres-sions using different dependent variables, we cannot directly testwhether or not the graphs are statistically different. They are there-fore meant to be suggestive of very interesting and different pat-terns in the probability of entering new markets as a function ofattainment discrepancy.

and Dahlin, 2007). They also tend to be morediscouraged by a threat to survival (performancewell below aspiration level) than by difficulties infinding relevant new opportunities for non-scalefree resources that are subject to opportunity costs(performance well above aspiration level).

This study integrates key ideas from both theBTOF and the RBV to gain new insights into theeffect of performance relative to aspiration levelon the decision to enter new markets. Althoughthe two theories are different, they do have somesimilarities, and there is substantial scope forcross-fertilization that has largely gone unnoticedin the literature (Pitelis, 2007). We believe thatintegrating these two theories can be a fruitfulpath for future research, in particular, to helpus better understand organizational growth andchange.

As mentioned above, a handful of studies(Chang, 1996; Lin, 2014) have examined the effectof performance relative to aspiration level on thedecision to enter new markets. However, theirresults were generally not conclusive. One possibleexplanation for the difference between the signifi-cant results of this study and the results of the abovestudies may be related to the distinction we havemade between the neighborhood of aspiration leveland the areas that are far below and above it, andaccordingly, to the nonlinear relationships that wehave postulated between both negative and positiveattainment discrepancy, and a firm’s probability ofentering new markets. To corroborate the aboveexplanation we have re-estimated all of our modelsusing a linear-only version of attainment discrep-ancy.20 Generally speaking, where attainmentdiscrepancy enters only linearly, we have foundan insignificant relationship between this variableand the probability of underperforming and outper-forming firms to enter new markets. However, asmentioned above, when we have used a nonlinearfunction, that is, when we have also included aquadratic of firms’ negative/positive attainmentdiscrepancy in our models, we have found asignificant inverted U-shape relationship. That is,an incompatible assumption about the shape of therelationship may result in inconclusive results.

The use of a nonlinear relationship has broaderimplications for the behavioral theory of the firm.Most studies that have examined the effect of per-formance feedback on managerial decisions have

20 The results are available on request.



not distinguished between situations that are inthe neighborhood of aspiration level (below andabove) and situations that are well below or above it.Accordingly, they have usually modeled the effectas a modified spline function (linear above andbelow aspiration level, but with potentially dif-ferent slopes) (Harris and Bromiley, 2007). Thismay explain why different studies that have exam-ined the same phenomenon have found conflict-ing results. For example, Greve (2003a) has foundthat R&D intensity falls as performance increasesabove aspiration, while Chen and Miller (2007)have found that R&D intensity rises as perfor-mance increases above aspiration. Similarly, thismay explain why some scholars have not found sup-port for the BTOF logic and sometimes have evenfound results that have contradicted their hypothe-ses. For example, Iyer and Miller (2008) have foundthat contrary to their hypothesis, as performancefalls below aspiration, the probability of an acquisi-tion decreases. In all of these cases, the results maystem from the fact that researchers have postulatedan incompatible assumption about the shape of therelationship. Future research could replicate thesestudies using nonlinear relations.

As implied above, only a handful of studies haveempirically distinguished between situations in theneighborhood of aspiration level and situations thatare well below or above it, and consequently, haveused nonlinear functions (e.g., Lehman et al., 2011;Shimizu, 2007). To the best of our knowledge,these studies have done this only to examine theeffect of performance below aspiration level onmanagerial decisions. The present article joins thesestudies and demonstrates that there is a differencebetween the effect of performance above aspirationstill in the neighborhood of aspiration and theeffect of performance well above aspiration onmanagerial decisions. Hence, it extends the BTOFto apply not only to situations that are well belowaspiration level, but also to situations that are wellabove it.

While this article represents a step forward instudying the factors affecting the decision to enternew markets, our results also suggest several futureresearch opportunities. First, firms could use differ-ent vehicles/modes of entry into new markets, suchas acquisition and internal development (e.g., Chat-terjee, 1990; Martin and Salomon, 2003). Firms per-forming below aspiration may use different vehiclesfrom outperforming firms. Moreover, firms in theneighborhood of aspiration, either below or above,

may use different vehicles than firms that are wellbelow or above aspiration. Future research couldinvestigate these topics.

Second, following, for example, Chang (1996);Delios and Beamish (1999), and Matsusaka (2001),we consider entry into new markets as a searchprocess by which firms seek, for example, marketsthat are good matches for their resources (marketseeking), or access to inputs unavailable in theirexisting markets (asset seeking). Future researchcan examine whether the effect of performancerelative to aspiration level is different when it comesto different motivations for entry into new markets,such as those mentioned above.

Third, entering new markets is part of corporatereconfiguration (e.g., Helfat and Eisenhardt, 2004;Karim, 2006). In other words, it is part of the activ-ities in which firms engage when adding, redeploy-ing, recombining, or divesting resources or busi-ness units in order to improve their profitability.Future research could examine how the relationshipbetween both negative and positive attainment dis-crepancy, and the probability of entering new mar-kets is contingent on other reconfiguration activi-ties, such as divestment of business units.

Fourth, entering new product markets and newcountries are both sources of risk and complexity, inparticular, when such markets have little similarityto the markets in which a firm is already active. Theless related the new product market is to existingproduct markets and the more distant the new coun-try is from existing geographic markets, the greaterthe challenge of managing such a move (Johansonand Vahlne, 1977; Rumelt, 1974). Future researchcould examine, for example, whether negativeand positive attainment discrepancy also affect thedegree of relatedness/distance between the marketsin which a firm is already active and the newmarkets it chooses to enter, that is, on the amountof risk it is willing to take and how it will affect thisdecision.

Finally, as mentioned above, this study focuseson the effect of firm performance relative to its per-formance history (historical aspiration) on the deci-sion to enter new markets. Future research couldexamine the effect of firm performance relative tothe performance of comparable others (i.e., socialaspiration) on the decision to enter new markets. Inthis case, researchers are required to determine theparticular reference group to which the focal firmcompares itself.



ACKNOWLEDGEMENTS

We are most grateful to SMJ Editor ConstanceHelfat and two thoughtful reviewers for providingexcellent feedback. We also thank Elizabeth Boyle,J. P. Eggers, Naomi Feldman, Niron Hashai, DeepakHegde, Gabriel Natividad, Robert Salomon, andseminar participants at New York University forinsightful comments on earlier versions of thisarticle. All errors remain our own.

REFERENCES

Anand J, Singh H. 1997. Asset redeployment, acquisitionsand corporate strategy in declining industries. StrategicManagement Journal 18(S1): 99–118.

Audia PG, Greve HR. 2006. Less likely to fail: lowperformance, firm size, and factory expansion in theshipbuilding industry. Management Science 52(1):83–94.

Audia PG, Locke EA, Smith KG. 2000. The paradox ofsuccess: an archival and a laboratory study of strategicpersistence following radical environmental change.Academy of Management Journal 43(5): 837–853.

Baum JAC, Dahlin KB. 2007. Aspiration performance andrailroads’ patterns of learning from train wrecks andcrashes. Organization Science 18(3): 368–385.

Baum JAC, Rowley TJ, Shipilov AV, Chuang YT. 2005.Dancing with strangers: aspiration performance and thesearch for underwriting syndicate partners. Administra-tive Science Quarterly 50(4): 536–575.

Boyle E, Shapira Z. 2012. The liability of leading: bat-tling aspiration and survival goals in the Jeopardy!Tournament of champions. Organization Science 23(4):1100–1113.

Bromiley P. 1991. Testing a causal model of corporaterisk taking and performance. Academy of ManagementJournal 34: 37–59.

Campa JM, Kedia S. 2002. Explaining the diversificationdiscount. Journal of Finance 57: 1731–1762.

Chang SJ. 1996. An evolutionary perspective on diversifi-cation and corporate restructuring: entry, exit, and eco-nomic performance during 1981–1989. Strategic Man-agement Journal 17: 587–611.

Chatterjee S. 1990. Excess resources, utilization costs, andmode of entry. Academy of Management Journal 33(4):780–800.

Chen WR, Miller KD. 2007. Situational and institutionaldeterminants of firms’ R&D search intensity. StrategicManagement Journal 28: 369–381.

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

D’Aveni RA. 1989. The aftermath of organizationaldecline: a longitudinal study of the strategic andmanagerial characteristics of declining firms. Academyof Management Journal 32(3): 577–605.

Delios A, Beamish PW. 1999. Geographic scope, prod-uct diversification, and the corporate performance of

Japanese firms. Strategic Management Journal 20:711–727.

Festinger L. 1954. A theory of social comparison pro-cesses. Human Relations 7(2): 117–140.

Fiegenbaum A, Hart S, Schendel D. 1996. Strategic ref-erence point theory. Strategic Management Journal17(3): 219–235.

Gavetti G, Greve HR, Levinthal DA, Ocasio W. 2012. Thebehavioral theory of the firm: assessment and prospects.Academy of Management Annals 6(1): 1–40.

Grant RM, Jammine AP, Thomas H. 1988. Diversity, diver-sification, and profitability among British manufactur-ing companies, 1972–1984. Academy of ManagementJournal 31(4): 771–801.

Greve HR. 1998. Performance, aspirations, and risky orga-nizational change. Administrative Science Quarterly43: 58–86.

Greve HR. 2003a. A behavioral theory of R&D expen-ditures and innovations: evidence from shipbuilding.Academy of Management Journal 46: 685–702.

Greve HR. 2003b. Organizational Learning from Perfor-mance Feedback: A Behavioral Perspective on Innova-tion and Change. Cambridge University Press: Cam-bridge, U.K.

Greve HR. 2010. Positional rigidity: low performance andresource acquisition in large and small firms. StrategicManagement Journal 32: 103–114.

Harris J, Bromiley P. 2007. Incentives to cheat: the influ-ence of executive compensation and firm performanceon financial misrepresentation. Organization Science18(3): 350–367.

Helfat CE, Eisenhardt KM. 2004. Inter-temporaleconomies of scope, organizational modularity, andthe dynamics of diversification. Strategic ManagementJournal 25(13): 1217–1232.

Helfat CE, Lieberman MB. 2002. The birth of capabilities:market entry and the importance of pre-history. Indus-trial and Corporate Change 11(4): 725–760.

Helfat CE, Peteraf MA. 2003. The dynamic resource-basedview: capability lifecycles. Strategic ManagementJournal 24(10): 997–1010.

Hitt MA, Tihanyi L, Miller T, Connelly B. 2006. Interna-tional diversification: antecedents, outcomes, and mod-erators. Journal of Management 32(6): 831–867.

Hoetker G. 2007. The use of logit and probit models instrategic management research: critical issues. StrategicManagement Journal 28: 331–343.

Hu S, Blettner D, Bettis RA. 2011. Adaptive aspira-tion: performance consequences of risk preferences atextremes and alternative reference groups. StrategicManagement Journal 32: 1426–1436.

Iyer DN, Miller KD. 2008. Performance feedback, slack,and the timing of acquisitions. Academy of Manage-ment Journal 51: 808–822.

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

Kahneman DA, Tversky A. 1979. Prospect theory: analysisof decision under risk. Econometrica 47: 263–291.

Karim S. 2006. Modularity in organizational structure: thereconfiguration of internally developed and acquired



business units. Strategic Management Journal 27(9):799–823.

Kumar MVS. 2009. The relationship between product andinternational diversification: the effects of short–runconstraints and endogeneity. Strategic ManagementJournal 30: 99–116.

Lant TK, Milliken FJ, Batra B. 1992. The role of manage-ment learning and interpretation in strategic persistenceand reorientation: an empirical exploration. StrategicManagement Journal 13: 585–608.

Lehman DW, Hahn J, Ramanujam R, Alge BJ. 2011. Thedynamics of the performance-risk relationship withina performance period: the moderating role of deadlineproximity. Organization Science 22(6): 1613–1630.

Levinthal D, March JG. 1981. A model of adaptive orga-nizational search. Journal of Economic Behavior andOrganization 2: 307–333.

Levinthal D, Wu B. 2010. Opportunity costs and non-scalefree capabilities: profit maximization, corporate scope,and profit margins. Strategic Management Journal31(7): 780–801.

Lewin K, Dembo T, Festinger L, Sears PS. 1944. Level ofaspiration. In Personality and the Behavior Disorders,Hunt JMV (ed). Ronald Press Company: New York;333–378.

Lin WT. 2014. How do managers decide on internation-alization processes? The role of organizational slackand performance feedback. Journal of World Business49(3): 396–408.

Lu JW, Beamish PW. 2004. International diversifica-tion and firm performance: the S-curve hypothesis.Academy of Management Journal 47(4): 598–609.

Mahoney JT, Pandian JR. 1992. The resource-based viewwithin the conversation of strategic management.Strategic Management Journal 13(5): 363–380.

March JG, Shapira Z. 1987. Managerial perspectives onrisk and risk taking. Management Science 33(11):1404–1418.

March JG, Shapira Z. 1992. Variable risk preferences andthe focus of attention. Psychological Review 99(1):172–183.

Martin X, Salomon R. 2003. Knowledge transfer capacityand its implications for the theory of the multinationalcorporation. Journal of International Business Studies34(4): 356–373.

Matsusaka JG. 2001. Corporate diversification, value max-imization, and organizational capabilities. Journal ofBusiness 74(3): 409–431.

Miller DJ. 2004. Firms’ technological resources andthe performance effects of diversification: a longitu-dinal study. Strategic Management Journal 25(11):1097–1119.

Mishina Y, Dykes BJ, Block ES, Pollock TG. 2010.Why “good” firms do bad things: the effects of highaspirations, high expectations, and prominence on theincidence of corporate illegality. Academy of Manage-ment Journal 53(4): 701–722.

Mishina Y, Pollock TG, Porac JF. 2004. Are moreresources always better for growth? Resource stickinessin market and product expansion. Strategic Manage-ment Journal 25: 1179–1197.

Moliterno TP, Wiersema MF. 2007. Firm perfor-mance, rent appropriation, and the strategic resourcedivestment capability. Strategic Management Journal28(11): 1065–1087.

Montgomery CA. 1994. Corporate diversification. Journalof Economic Perspective 8(3): 163–178.

Morck R, Yeung B. 1991. Why investors value multina-tionality. Journal of Business 64(2): 165–187.

Nohria N, Gulati R. 1996. Is slack good or bad forinnovation? Academy of Management Journal 39(5):1245–1264.

Penrose ET. 1959. The Theory of the Growth of the Firm.John Wiley: New York.

Penrose ET. 1960. The growth of the firm – a casestudy: the Hercules Powder Company. Business HistoryReview 34(1): 1–23.

Peteraf M. 1993. The cornerstones of competitive advan-tage: A resource-based view. Strategic ManagementJournal 14(3): 179–191.

Pitelis CN. 2007. A behavioral resource-based view ofthe firm: the synergy of Cyert and March (1963) andPenrose (1959). Organization Science 18(3): 478–490.

Ref O. 2015. The relationship between product and geo-graphic diversification: a fine-grained analysis of its dif-ferent patterns. Journal of International Management21(2): 83–99.

Roberts PW, McEvily S. 2005. Product-line expansionand resource cannibalization. Journal of EconomicBehavior and Organization 57(1): 49–70.

Rumelt RP. 1974. Strategy, Structure, and Economic Per-formance. Harvard Business Review Press: Cambridge,MA.

Sakhartov AV, Folta TB. 2014. Resource relatedness,redeployability, and firm value. Strategic ManagementJournal 35(12): 1781–1797.

Shimizu K. 2007. Prospect theory, behavioral theory,and the threat-rigidity thesis: combinative effects onorganizational decision to divest formerly acquiredunits. Academy of Management Journal 50(6):1495–1514.

Silverman BS. 1999. Technological resources and thedirection of corporate diversification: toward an inte-gration of the resource-based view and transaction costeconomics. Management Science 45(8): 1109–1124.

Singh JV. 1986. Performance, slack, and risk taking inorganizational decision making. Academy of Manage-ment Journal 29: 562–585.

Staw BM, Sandelands LE, Dutton JE. 1981. Threat rigidityeffects in organizational behavior: A multilevel analy-sis. Administrative science quarterly 1: 501–524.

Teece DJ. 1982. Towards an economic theory of themultiproduct firm. Journal of Economic Behavior andOrganization 3: 39–63.

Vidal E, Mitchell W. 2015. Adding by subtracting: the rela-tionship between performance feedback and resourcereconfiguration through divestitures. Organization Sci-ence 26(4): 1101–1118.

Washburn M, Bromiley P. 2012. Comparing aspirationmodels: the role of selective attention. Journal ofManagement Studies 49(5): 896–917.

Wernerfelt B. 1984. A resource based view of the firm.Strategic management journal 5(2): 171–180.



Wiersema MF, Bowen HP. 2009. The use of limiteddependent variable techniques in strategy research:issues and methods. Strategic Management Journal30(6): 679–692.

Wiseman RM, Bromiley P. 1996. Toward a model of riskin declining organizations: an empirical examination

of risk, performance and decline. Organization Science7(5): 524–543.

Wu B. 2013. Opportunity costs, industry dynamics, andcorporate diversification: evidence from the cardiovas-cular medical device industry 1976–2004. StrategicManagement Journal 34(11): 1265–1287.


Entering New Markets: The Effect of Performance Feedback ... and...(3) performance above aspiration...

Documents

Transcript of Entering New Markets: The Effect of Performance Feedback ... and...(3) performance above aspiration...