2009 aff loss

8/2/2019 2009 Aff Loss

1/22

Copyright 2009 Strategic Management Society

Strategic Entrepreneurship JournalStrat. Entrepreneurship J ., 3: 105126 (2009)

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/sej.66

AFFORDABLE LOSS: BEHAVIORAL ECONOMICASPECTS OF THE PLUNGE DECISION

NICHOLAS DEW, 1 SARAS SARASATHY, 2 STUART READ, 3 and ROBERT WILTBANK 4*1Naval Postgraduate School, Monterey, California, U.S.A.2Darden Graduate School of Business Administration, University of Virginia,Charlottesville, Virginia, U.S.A.3IMD, Lausanne, Switzerland 4Atkinson Graduate School of Management, Willamette University, Salem,Oregon, U.S.A.

Affordable loss involves decision makers estimating what they might be able to put at risk and determining what they are willing to lose in order to follow a course of action. Using theentrepreneurs new venture plunge decision, this article combines insights from behavioraleconomics to develop a detailed analysis of the affordable loss heuristic. Specically, wedevelop propositions to explain how individuals: (1) decide what they can afford to lose; and (2) what they are willing to lose in order to plunge into entrepreneurship. The article alsodiscusses the implications of affordable loss for the economics of strategic entrepreneurship.Copyright 2009 Strategic Management Society.

INTRODUCTION

Several theories exist to provide guidance for theindividual facing the plunge decision the choice

faced by a potential entrepreneur to make an initialcommitment to a de novo venture. 1 Classic risk-return analysis is often prescribed as the way to helpmake this decision. The decision criteria used insuch analyses usually urge would-be entrepreneursto calculate the net present value (NPV) of futurerisk-adjusted returns while taking into account theiropportunity costs in terms of job market value (Benz,2006; Eisenhauer, 1995; Hamilton, 2000). For

example, Campbell (1992: 12) states that an indi-viduals decision whether to become an entrepreneurwill be based upon a comparison of the expectedreward to entrepreneurship and the reward to the

best alternative use of his [or her] time. Amit,Muller, and Cockburn (1995) found empiricalsupport for the hypothesis that the lower the op-portunity costs of individuals, the more likely theyare to undertake entrepreneurial activity.

Recently, an alternative approach based on realoptions has been suggested, particularly at the rmlevelfor example when a rm is considering takingthe plunge into new technology positioning projects(McGrath, 1997). Real options analysis enablesdecision makers to more accurately value invest-ment opportunities in instances where investments

can be incurred in stages (Dixit and Pindyck, 1994).In arguing for the value of viewing entrepreneurialinvestment decisions through a real options lens,McGrath (1999: 14) states that if investments arestaged so that expenditures end under poor condi-tions, losses can be contained.

Keywords: affordable loss; entrepreneurship; behavioral eco-

nomics; effectuation*Correspondence to: Robert Wiltbank, Atkinson GraduateSchool of Management, Willamette University, Salem, OR97301, U.S.A. E-mail: [email protected]

1 In this article, we use the term plunge decision interchange-ably with the terms entry into entrepreneurship and the self-employment decision .

8/2/2019 2009 Aff Loss

2/22

106 N. Dew et al.

Copyright 2009 Strategic Management Society Strat. Entrepreneurship J. , 3: 105126 (2009)

DOI: 10.1002/sej

A third approach based on the affordable loss principle has been outlined by Sarasvathy (2001a).This heuristic was induced from empirical studies of entrepreneurial expertise (Sarasvathy, 2001b), asexperts (Chase and Simon, 1973) exhibit high per-formance in their domains (Ericsson and Lehmann,

1996). Affordable loss is one component of effectua-tion, a set of heuristics for making decisions underuncertainty. While there is a large body of theoreti-cal work on the nancial and behavioral economicbases for neoclassical investment theory (NCIT) andreal options (Dixit and Pindyk, 1994; McMullen andShepherd, 2006; Wennberg, Folta, and Delmar,2006; Lee, Peng, and Barney, 2007), there is verylittle detail on how affordable loss works, littleclarity about the behavioral assumptions upon whichit is based, and few particulars about how thisconcept relates to these other approaches. Mostimportantly, as it currently stands, affordable loss islittle more than an observed heuristic induced fromstudies of entrepreneurial expertise. It is not yettheoretically situated in either nancial or behavioraleconomics. This article attempts to address this gapby explicitly relating the heuristic to existing resultsfrom behavioral economics. 2 In other words, to takepreliminary steps to address the research question of how a behavioral economic perspective theoreticallycan deepen our understanding of the empiricallyinduced decision heuristic of affordable loss. The

plunge decision of the entrepreneur provides auniquely appropriate context for deriving thesetheoretically meaningful relationships.

This article makes two key contributions to schol-ars of strategic entrepreneurship. First, we hope toexpand the theoretically informed and practicallyuseful toolbox available to decision makers underuncertainty. Second, we hope to contribute to theexciting new conversation about a more creative view of entrepreneurship and the market process,as fostered by this journal.

After a brief literature review on the making andnding of entrepreneurial opportunities and the risk-taking rationalities they entail, we briey summarizethe empirical basis for the use of affordable loss byexpert entrepreneurs. We begin the section entitled

Behavioral Aspects of Affordable Loss with an

outline of the key features of the affordable lossheuristic in comparison with NCIT and real options,and then delve into the behavioral aspects of theplunge decision using affordable loss. Thereafter,we discuss the implications of affordable loss for thefrequency of start-up activity, the cost of failed

starts, and the efciency of new ventures thatgrow.

MAKING AS WELL AS FINDINGENTREPRENEURIAL OPPORTUNITIES

Going beyond a discovery viewof entrepreneurship

It is perhaps not an accident that the very rst articlein the special launch issue of the Strategic Entrepre-neurship Journal outlines the alternative possibili-ties of studying entrepreneurship as an engine of making, and not merely one of discovery (Alvarezand Barney, 2007). Several other articles pick up onthis theme in a variety of ways as well. Even tradi-tional sociological approaches were pushed beyondthe deterministic inuences of existing social net-works to the formation of new networks (Aldrichand Kim, 2007). And Baron (2007) emphasized theactive element in new venture creation, even whileemphasizing the role of well-trodden relationshipsbetween automated cognitive processes resulting inrecognition of opportunities already fully formedand out there in the environment.

Of particular note is Millers (2007) exposition of risk and rationality that offers a contingent perspec-tive on risk and rationality. His starting point is priordescriptions of the entrepreneurial process as a func-tion of a set of three possibilitiesopportunityrecognition, opportunity discovery, and opportunitycreation (Littlechild, 1986; Buchanan and Vanberg,1991; Sarasvathy et al. , 2003). Miller argues thatthese three descriptions imply conceptions of risk and rationality that are process contingent becausethe different descriptions involve unique sourcesof risk that, in turn, require different rationalresponses.

This framework leads Miller to argue that conven-tional interpretations of risk-taking behaviors (e.g.,as maximizing expected utility) may be unique, his-torically situated frames or paradigms that may bestiing broader thinking about risk and rationality.Instead, there may be alternative ways of understanding entrepreneurship that call for other

2 We leave explication of links to nancial economics to futureendeavors. For a bare-bones beginning in this direction, seeSarasvathy (2008).

8/2/2019 2009 Aff Loss

3/22

Affordable Loss 107


DOI: 10.1002/sej

perspectives on risk and rationality (Miller, 2007:60). Entrepreneurs may not be strictly bound to asingle form of rationality: instead, they may displaya practical, situational rationality that involvesswitching cognitive gears to adapt their decision-making style to the exigencies of their situation.

Three views of the entrepreneurial process

These situations can be differentiated according tothree conceptions of the entrepreneurial process thatcontain within them different assumptions about theknowledge (ignorance) of decision makers withregard to the future.

Recognition

The view of entrepreneurship as an opportunity rec-ognition process involves matching sources of supply and demand that pre-exist and bringing themtogether through an existing rm or a new rm(Sarasvathy et al., 2003). The conception of risk inthis process is based on unpredictability: possiblefuture states are, in principle, knowable, but in theabsence of complete knowledge, individuals areforced to rely on their own limited information, onwhich they form subjective probability estimates

(Miller, 2007). Rationality, herein, consists of maximizing the subjective expected utility of theentrepreneur.

Discovery

Entrepreneurship as a process of opportunity discov-ery involves a different conceptualization of risk andrationality. In this view, either demand or supplyexists, but not both. Therefore, entrepreneurialopportunities involve the search for and discoveryof the nonexistent side of a market transaction(Sarasvathy et al., 2003). In this view of entrepre-neurship, risk arises because of the unknowablecharacter of search processes, which raises the pos-sibility that the entrepreneur may be truly surprisedby what he/she nds (Miller, 2007). Rationality,herein, consists of managing the search process in asatisfactory fashion, i.e., setting appropriate aspira-tion levels, exploring efciently, and learning fromexperience.

Creative

Entrepreneurship as a process of opportunity cre-ation supposes that neither demand nor supply existsin an obvious fashion and that both, therefore, mustbe created by entrepreneurial interventions in themarketplace (Sarasvathy et al., 2003:). A distin-guishing feature of this view of opportunity is thatentrepreneurs have a causal role in establishingopportunities. In this conceptualization, risk is aproduct of uncontrollability: it is the freedom of other agents to act creatively in the marketplace thatexposes the entrepreneur to the risk of downsidelosses. According to Miller (2007: 58), entrepre-neurship as a process of opportunity creation raisessome questions that challenge the mainstream con-ceptualizations of risk and rationality. Rationaldecision making in the context of such risks may

involve limiting entrepreneurial investments toaffordable losses (Miller, 2007).

Risk and rationality in the creative view

In every context of uncertainty, paying attention todownside possibilities is essential to making gooddecisions. Even in the case of high-potential oppor-tunitiessuch as those involving defensible patentsin healthcare and technologythere is always achance things will not work out. Hence, we deductour investment in the venture (which equals the costof failure, should failure occur) from our calcula-tions of expected return. Moreover, we might try tolimit the downside by spreading investment overseveral projects (portfolio diversication) or bystaging the actual deployment of funds (real optionslogic).

However, in the case of the creative process, thevery existence of the upside may be in doubt. Takethe case of the absurdly unlikely venture 1800-AUTOPSY. Until 1988, the world got along withoutthe services of a company providing autopsies ondemand. With the growing success and increasingdemand for the companys services over the last twodecades, one could argue ex post that there was latentdemand that simply went unnoticed until VidalHerrera recognized the opportunity with the unerringeye of the attentive entrepreneur. But what would hiselevator pitch have been in 1988? Or for that matter,that of Starbucks in 1980, when according to reliablehistorical accounts, coffee consumption in the U.S.had been steadily declining for 20 years (Koehn,2001). Common sense suggests that while we might

8/2/2019 2009 Aff Loss

4/22

108 N. Dew et al.


DOI: 10.1002/sej

be able to calculate what we would lose in suchventuresnamely all that one chooses to invest, asKnight (1921) arguedwe cannot normatively pre-scribe what we ought to invest because the upside isvirtually unknowable. And if we make the decisionin comparison with other opportunities that offer

more predictable upsides, the creative project willalways be discarded under any rational metric.

It is within this context of entrepreneurship as acreative process that we begin to consider how indi-viduals decide what they are willing to lose (i.e.,their affordable loss) in order to take the plunge intoentrepreneurship. The fundamental asymmetrybetween the calculability of losses and the unpre-dictability of gains both fuels the creative processand is an outcome of it. According to literaturefocused on this problem, human imagination andfreedom of action are the fundamental features of creative market processes (Buchanan and Vanberg,1991; Littlechild, 1986; Shackle, 1979). This cre-ative initiative makes the future indeterminate and,therefore, suspends the logic of consequential rea-soning (March, 1994). Expectations about the future,though not beyond conjecture, are frequently awednot only because historical data either do not existin this space or tend to point in multiple directionsin equivocal fashion, but also because of the limitsof anticipating how ones own actions will interactwith those of other actors in the marketplace. In suchcases, how then do we characterize risk and itsappropriate (rational) response? Clearly, the stan-dard calculus of optimizing risk/return has signi-cant drawbacks. Modied versions of risk/returnthat involve min-max reasoning or the application of real options is also of limited applicability owing tothe meaninglessness of estimated payoffs. Instead,the central concern of the entrepreneur is with thehazard of downside loss, i.e., the possibility of losses and the decision makers aversion to loss(Kahneman and Tversky, 1979; March and Shapira,1987; Miller and Leiblein, 1996; Miller and Reuer,1996; Sortino and Satchell, 2001; Sortino and vander Meer, 1991; Thaler et al. , 1997).

Millers suggested bases for managing risk inthe creative space

Miller (2007) suggests three possible solution spacesfor how entrepreneurs may deal with risk in theopportunity creative space, each of which is relevantto our exposition of affordable loss in the nextsection.

Identity

In the less than fully specied creative context,looking inwards to ones own identity (rather thanoutwards to the environment) may provide an impor-tant guide for entrepreneurial action. As Sarasvathyand Dew (2005a) showed, entrepreneurs oftenexplain their actions and decisions in terms of theiridentities, rather than their preferences or interests.It serves them well to have a strong sense of identity(who we are rather than what we want) and of process (how to make decisions rather than whatdecisions to make) when outcomes are highly unpre-dictable. This is a case of procedural rather thansubstantive rationality (Simon, 1976).

Values and preferences

Eminent scholars such as Sen (2003) have arguedthat rationality includes critical reection on onesown values and preferences, not just maximizingchoices based on them. Here again, who the entre-preneur is plays an important role by allowing him/ her to manage preference conicts, experiment withnewly acquired preferences and even construct newones. Rationality is, thus, a dynamic outcome of preference processing by individuals.

Emotions

As Miller points out, noncognitive aspects of risk taking have been largely neglected in the literatureon risk perception. Yet, practitioners often remark that the emotional aspects of risky decisionshowthey feel about the risksis highly inuential intheir decision processes. Moreover, recent empiricalevidence suggests that emotional responses to risk are better predictors of behavior than cognitiveassessments (Loewenstein et al., 2001).

In sum, Millers overriding claim is that entrepre-neurs who exhibit skillful performances may do soby operating according to more than one approachto risk and rationality. They may operate accordingto plural rationalitiesby applying decision-makingtechniques contingent on their perceptions of theirsituation at hand. Their choices about which deci-sion technologies to apply are probably not arbitrary,but are acquired through practical experience andare largely tacit, i.e., invoked automatically based onpattern recognition (Miller, 2007). This implies thatcognitive studies of expert entrepreneurs may helpreveal this practical, tacit rationality and help

8/2/2019 2009 Aff Loss

5/22

Affordable Loss 109


DOI: 10.1002/sej

understand how it works (Baron and Ensley, 2006;Sarasvathy, 2008).

The empirical basis for affordable loss

Evidence is beginning to accumulate on the use of

effectual logicincluding the affordable loss heu-risticand an outline of its impact on performanceis beginning to emerge. The original evidence col-lected by Sarasvathy (1998 and 2001a) using a rep-resentative sample of expert entrepreneurs has sincebeen replicated and compared with novices andexperienced managers, and also studied in thecontext of private equity investing and through ameta-analysis of previously published work onentrepreneurial performance. Dew et al. (2009)delineate the differences between novices andexperts in the use of the logic overall, and Readet al. (2009) investigate applications to marketingdecisions with additional data from experiencedmanagers. Results indicate that expert entrepreneurswere signicantly more likely to pay attention todownside risk and use affordable loss heuristics thaneither the novices or the managers. Read, Song, andSmit (2009) could nd only four studies that hadmeasures they could relate to affordable loss, so theyfailed to nd a signicant relationship with newventure performance. However, their meta-analyticstudy of 35 articles totaling investigations of 9,897ventures did support a positive relationship betweenthree of the effectuation principles and new ventureperformance. Another study using a scenario surveymethod examined angel investors in their use of prediction-oriented (as opposed to control-oriented)strategies, the latter including affordable loss strate-gies (Wiltbank et al. , 2009). Results showed thatprediction-oriented angels made signicantly largerventure investments, while those who emphasizednonpredictive strategies, such as affordable loss,experienced a signicantly lower number of invest-ment failures without a reduction in the number of successes overall.

Given that expert entrepreneurs have a demon-strated preference for nonpredictive strategies, suchas affordable loss, and because this preference hasbeen acquired as part of their expertise-developmentprocess, it may be signicantly related to positivenew venture performance. Therefore, it would beuseful to understand in more depth how all entrepre-neurs can use affordable loss as part of their newventure decision-making toolbox and what that mayimply for entrepreneurial performance at meso- and

macrolevels of analysis. That is the task we take upin the rest of this article. We address the formerthrough the discipline of behavioral economics andthe latter as a derived model of performance.

BEHAVIORAL ASPECTS OFAFFORDABLE LOSS

Before we consider the behavioral aspects of afford-able loss, it may be useful to outline its main featuresand compare them with those relevant to more famil-iar decision tools, such as NPV and real options.

Overlaps and differences between NPV,real options, and affordable loss

We would like to begin by noting that there exist

both overlaps and differences between affordableloss and the other two approaches. 3 As Miller (2007)argued so well, the key to decision making in thecreative setting is that entrepreneurs can use multi-ple rationalities contingent upon the particularitiesof their identity and venture ideas. Moreover, theycan (and should) draw from an extended toolbox of strategies that include everything from NPV, min-max, and real options to affordable loss, integrativenegotiation, leveraging slack, and even gut feel andintuition. Yet the differences are worth emphasizingsimply because they make a differenceboth inhow entrepreneurs perceive problems and in howthey tackle them. And differences in their choicesalso lead to differences in outcomes, whether at theirown or more macrolevels.

The most fundamental difference, of course, isthat affordable loss is rmly grounded in behavioraltheory (bounded cognition and psychology) abouthuman reasoning, whereas neoclassical investmenttheory (expected returns) and real options theory arebased on the expected utility model that behavioraleconomists continually inveigh against. This meansthe theories are substantially different in terms of their description of the reasoning process itself. Italso means these differences, and the consequencesimplied by them, are empirically testable usingstandard behavioral economic methods such asexperiments.

One could investigate the descriptive accuracy of the affordable loss model in comparison with real

3 Again, we thank our alert, though anonymous, reviewers forpressing us on this point.

8/2/2019 2009 Aff Loss

6/22

110 N. Dew et al.


DOI: 10.1002/sej

options reasoning. In the previous section, we pro-vided relevant empirical evidence that shows theprevalence of affordable loss in expert entrepreneur-ial decision making. With regard to the use of realoptions in strategic management in large corpora-tions, a recent survey of accumulated evidence con-

cludes that even if real options has succeeded as away of thinking, the extent of acceptance and appli-cation of real options today has probably not livedup to expectations created in the mid- to late-1990s(Triantis, 2005: 8). While there are several publishedtheoretical papers concerning real options thinkingin entrepreneurship, empirical evidence is rathersparse, and what does exist appears to be unsup-portive of any actual use of real options by entrepre-neursespecially in terms of the upside potential of opportunities. A recent study that used data from alarge longitudinal study of entrepreneurship in theknowledge-intensive sector between 1989 and 2002, 4 concludes:

Our study informs real options theory becausewhile the theory emphasizes the dynamic natureof nancial investment decisions, actually veryfew studies fully test this assumption on individ-ual human decision makers. However, our studywas not able to prove that entrepreneurs indepen-dent of the entry choice make complicated judg-ments taking into account dueling option andmixed effects of irreversibility and uncertainty.One explanation is that our irreversibility mea-sures are highly imperfect. If not so, the main goalfor entrepreneurs in this study seems to be tominimize possible lossesusing the option todeferbut they are not considering growthoptions. This is in line with previous work madeon nascent entrepreneurs in Sweden and in othercountries that shows the same pattern: most entre-preneurs do not at all consider growth as an optionearly in the new venture formation process(Delmar and Davidsson, 2000). They are toofocused to get the venture operational and togather information about the basic viability of their opportunity (Wennberg, Folta, andDelmar, 2006).

One explanation for this is that authentic realoptions analyses are performed rather infrequently

by entrepreneurs. Perhaps this is because the infor-mation requirements of the theory are very high forall but the most simple of problems, the formal cal-culations required are substantially more complexthan the heuristic version of the theory, and duelingoptions frequently compete with one another in deci-

sion problems (Folta and OBrien, 2004). The theory,therefore, runs into problems of nancial literacyand data constraints. By comparison, affordable lossis information light and computationally simple.

The affordable loss heuristic involves decisionmakers estimating what they might be able to put atrisk and examining what they are willing to lose inorder to follow a particular course of action. In prin-ciple, affordable loss might be used at all levels of analysisindividual, rm, economy, etc.and in awide variety of contexts, such as new product devel-opment, new policy initiatives, the building of newinstitutions and, of course, new venture start-updecisions. However, the plunge decision of the indi-vidual entrepreneur provides a quintessential illus-tration of the affordable loss principle and is thefocus of the exposition in this article.

Take the case of an entrepreneur who is consider-ing quitting employment in order to start a rm. 5 Classic risk-return analysis suggests some marketresearch and competitive analysis should be done toestimate the potential risk and return to the venturebefore deciding whether or not to take the plunge.The entrepreneurs musings might go as follows: Iestimate that I need $2 million to start this venture,and I hope to break even in two years. I can put in$250,000, so I need to raise $1.75 million before Ican take the plungeeven without taking intoaccount the opportunity costs of forgoing two yearssalary.

Considered this way, taking the plunge is a matterof predicting parameters as accurately as possible inorder to make a good decision.

In contrast, affordable loss suggests that entrepre-neurs set an upper bound on what they are willingto lose in order to start the venture. This entrepreneurmight think I have always wanted to be my ownboss. I think I can afford to take two years and investmy $250,000 to try this out. In the worst case

4 The study was provided by Statistics Sweden and coveredmore than 3,300,000 individuals, representing more than 70percent of the active Swedish active labor market.

5 Some empirical data indicates that most new rms are startedon a part-time basis (Wennberg, Folta, and Delmar, 2006).Thus, the plunge decision may, in fact, occur in stages. Plung-ing in stages does not materially alter the analysis we presenthere and, therefore, for expository convenience, we focus on aone-time plunge decision.

8/2/2019 2009 Aff Loss

7/22

Affordable Loss 111


DOI: 10.1002/sej

scenario, I will lose the money and be back on the job market in two years.

In this case, taking the plunge involves designinga venture using what the entrepreneur has and whatothers may eventually provide. This may or may notinclude additional funding of $1.75 million.

This approach to the plunge decision entails indi-viduals judging what they are willing to lose in orderto take the plunge into entrepreneurship. Thisinvolves assessing what means are available to themand precommitting to how much they are willing tolose. Sarasvathy (2001a: 250) uses the concept of affordable loss as a criterion for selecting betweeneffects in effectuation, referring to a predeterminedlevel of affordable loss or acceptable risk. In turn,these insights build on Sarasvathy, Simon, and Lave(1998), where the decision processes of entrepre-neurs and bankers were examined. In that study,entrepreneurs were found to pick a level of risk theyfelt comfortable with and then focus on manipulat-ing returns. The economist George Shackle (1966)also refers to the term in an early paper, where hepostulates that the entrepreneur might characterizeeach venture opportunity according to the possiblegains and losses, and suggests affordable loss is usedin the evaluation of which venture opportunity anentrepreneur might pursue.

It is practical and reasonable to regard the focus-loss, in absolute terms, as depending on the natureand scale of the enterprise concerned. Thus, bychoice of an appropriate kind, or an appropriatesize, of plant or enterprise, he can adjust the great-est amount he stands to lose, that is, his focus loss,to the amount which, given the size and characterof his assets, he can afford to lose (Shackle, 1966:765).

Our starting point for analyzing the use of afford-able loss by potential entrepreneurs is the observa-tion that information about the potential downsideof a venture is more salient to the decision makerthan information about the potential upside of theventure. Salience refers to the distinctiveness andprominence of information (Mehta, Starmer, andSugden, 1994). Information that is more salientgrabs the attention of decision makers. The salienceof information may be the result of a number of dif-ferent cognitive factors that lead particular informa-tion to be perceived as standing out, suggestingitself, or just seeming obvious or natural to notice(Schelling, 1960; Mitchell et al., 2004). While

(normative) expected returns reasoning is agnosticabout the salience of upsides and downsides (and,therefore, weights both upside and downside infor-mation equally in computing a choice), affordableloss reasoning involves decision makers attendingunequally to the downside information about the

decision because it is more salient as a decision cri-terion. Downside information is, therefore, over-weighted as a choice criterion by comparison to the(normative) expected returns model.

Why is information about the downside moresalient than information about the upside when itcomes to launching a venture? The difference mayoccur because of perceived differences in the natureand source of the information used in such calcula-tions. To calculate the upside case for a venture, theentrepreneur has to estimate future revenues, costs,and possible risks that inuence the cost of capitalfor a venture. This involves looking outward tocollect information about the environmentcustomer preferences, supplier costs, competitoractivities, nancing alternatives, etc. Almost all of the information required for such calculations isexogenousabout things that are for the momentoutside the decision makers controlalmostentirely dependent on the effect to be created, andlargely reliant upon predictive information, such asestimates and expectations. Typically, this informa-tion is translated into net present value/discountedcash ow models.

Entrepreneurs may have good reasons for under-weighting this information in the plunge decision.While the upside potential of a venture is critical inmotivating the plunge decision, entrepreneurs maystill underweight upside potential as a salientdecision criterion for two reasons. First, from aninformation processing perspective (Simon, 1978),underweighting may occur because exogenous infor-mation is regarded as too fuzzy, noisy, and unreli-able to drive the choice process. Some empiricalevidence supports this assertion. Studies of ventureinvestors (such as venture capitalists) focus on man-agement quality and deemphasize business plans,indicating that professional investors tend to treatnancial estimations rather skeptically (Gomperset al., 2006). A survey of Inc. 500 founders askedwhether they had written formal business plansbefore they launched their companies and foundonly 40 percent said yes. Of those, 65 percent saidthey had strayed signicantly from their originalconception, adapting their plans as they went along.In a similar vein, only 12 percent of this years Inc.

8/2/2019 2009 Aff Loss

8/22

112 N. Dew et al.


DOI: 10.1002/sej

500 group said theyd done formal market researchbefore starting their companies (Bartlett, 2002: 63).Other studies have documented that expert entrepre-neurs clearly reject predictive data on market oppor-tunities, such as market research on new productideas (Dew et al. , 2009).

A second reason why entrepreneurs may under-weight upside potential is overcondence and over-optimism (Camerer and Lovallo, 1999; Cassar,2008; Casson, 2005). In this case, an entrepreneursoptimism that their venture will be a homerun success negates the discriminating value of carefullycalculating expected returns, since all upside sce-narios are assumed to dwarf the initial investmentsneeded to start the venture. Again, from an informa-tion processing perspective this actually reduces thesalience of upside data in the plunge decision.

By contrast, information about the potential down-side to launching a venture is often rather concreteand highly salient to potential entrepreneurs. To cal-culate affordable loss, all of the information entre-preneur needs to know is endogenoustheir currentnancial condition and a psychological estimate of their commitment in terms of the worst case sce-nario. Instead of looking outward for information inorder to decide how much money to commit to anew venture, entrepreneurs looks inward to assessthe means available for starting the proposed ventureand to estimate how much they are willing to lose.The estimate of affordable loss does not depend onthe venture, but varies from entrepreneur to entre-preneur and even across his/her life stages and cir-cumstances. Because this information is about theentrepreneurs own life, current commitments, andaspirations, it involves trade-offs between subjectiverisks and values over which the entrepreneur canassert some control. Owing to its relative concrete-ness, controllability, and the specter of loss, poten-tial entrepreneurs may use the worst case scenarioas a focal point for the plunge decision and pay agreat deal of attention to it as a discriminating deci-sion criterion (Sarasvathy, 1998).

Thus, consistent with bounded rationality, afford-able loss involves using a smaller information setthan is required in (normative) expected returns rea-soning. By allowing estimates of affordable loss todrive their decisions about which venture they start,entrepreneurs focus on information that is moresalient in determining their nal choice, and they putaside less salient information that does not deter-mine the decision. Again, this does not negate themotivating effect of the upside potential of a venture:

our intention is not to minimize the importance of this factor (nancial or otherwise, articulated or not).We merely stress that upside data is usually not dis-criminating and reliable enough to be the key deci-sion criterion that triggers an entrepreneur to takethe plunge.

We summarize these differences in Figure 1below:

Expected return

Expected return

Expected return

Investment

Investment

Investment

NCIT: EXO

ROR: PartiallyENDO

NCIT&

ROR

FOCUS

NCIT: EXOALFOCUS

ROR:

ROR: EXO

AL: Mostly ENDO

AL: Mostly ENDO

a: Overall space

Time

$

ROR: Partially ENDO

b: Real options

Time

$

ALFOCUS

AL: Mostly ENDO

AL: Mostly ENDO

c: Affordable loss

Time

$

RORFOCUS

Figure 1. Differences in the theoretical models guidinginvestment in new ventures

8/2/2019 2009 Aff Loss

9/22

Affordable Loss 113


DOI: 10.1002/sej

Figure 1a is a simplication of the overall argu-ment, made more nuanced through Figures 1b and1c. While both neoclassical investment theory(NCIT) and real options reasoning (ROR) treatexpected returns and investments as exogenous tothe decision maker with the plunge decision deter-

mined by the difference between the two (comparedto opportunity costs), affordable loss reasoningfocuses on the (endogenous) investment amount,with the plunge decision determined by theentrepreneurs willingness to lose this amount.

The necessity to take environmental endogeneitymore seriously has been emphasized by Adner andLevinthal (2004) in their recent critique of ROR.They argue that such endogeneity is precisely whatcharacterizes strategically interesting settings,where, having made an initial investment, rms canactively engage in follow-on activities that can inu-ence outcomes and identify new possible actions andgoals (Adner and Levinthal, 2004: 120). They dis-tinguish this situation from the exogenous opportu-nity set typically posited in real options and expectedvalue reasoning where the assumption is that thenature and quality of options are independent of therms interim activities. The implicit imagery bothin NCIT and ROR is of a rm buying a ticket toengage in some prespecied opportunity set (Adnerand Levinthal, 2004: 120). This ignores the roleof agency in shaping and molding initiatives andpossibilities.

Interestingly, in an empirical study that found evi-dence that supports predictions from a real optionsperspective, OBrien, Folta, and Johnson (2003:526) concluded that furthermore, whether or notthey are versed in the formality of real optionstheory, it appears that most entrepreneurs astutelyevaluate their concerns over uncertainty with respectto the degree of irreversibility associated with theirinvestment. As noted in the introduction, the realoptions literature has been lacking in empirical dem-onstrations of the theoretical interaction betweenuncertainty and irreversibility. Our results are uniquein that they indicate that the degree of irreversibilityassociated with a new venture can be inuenced bythe nature of the industry being entered, the locationselected, and even the characteristics of theentrepreneur.

This result is consistent with an effectual useof the affordable loss heuristic that does not pre-clude the possibility that entrepreneurs can mold,shape, transform, and reconstitute current realitiesincluding their own limited resourcesinto new

opportunities. Both the upside and downside of aventuring opportunity are taken to be endogenous.On the downside, entrepreneurs using affordableloss reasoning may attempt numerous ways of low-ering their resource investment in a new venture. Atthe limit, some ventures may be launched with zero

resources. Entrepreneurs are motivated to do thisboth by a combination of risk acceptance and lossavoidance, i.e., accepting risk as inevitable and thenstriving to minimize their downside loss. They mayalso be motivated by their skepticism about theinformation needed to make an upside case for theventure, which they may treat as endogenous to theirown efforts. So, instead of making a calculated beton an exogenously given upside, they seek out asmany ways as possible to increase the potentialreturns of the venture by actively trying to make thescenario better. We present a simple illustration of these arguments in Figure 1c, with a comparison toreal optionsper Adner and Levinthals (2004)argumentsin Figure 1b.

We now turn to building a theoretical basis foraffordable loss rooted in insights from behavioraleconomics. We begin by breaking up the decisionspace into three parts: (1) the preference for takingthe plunge; (2) the ability to take the plunge; and(3) the depth of the plunge.

Behavioral aspects of the preference for taking

the plungeOne of the interesting issues regarding the plungedecision has always been whether the motivation tobecome an entrepreneur is largely psychological orsubject to real inuence by nancial incentives.Motivation may have to do with any number of things including upside potential, psychologicalreasons (such as the desire for independence) andsocioeconomic factors (such as downsizing, power-distance, being an immigrant, and so on) (Swedberg,2000). The likelihood of actually acting upon any of these motivations, however, would have to take intoaccount things like the degree and intensity of moti-vation (willingness to lose any given sum) and reso-lution of conicts in nancial and nonnancialmotivation (risking independence versus riskingsecurity for example), where reducing the level of one below a threshold might make the conict dis-appear and make the plunge more affordable.

Most developmental economists and policymak-ers appear to assume that the motivation to plungedepends upon societal and economic incentives to

8/2/2019 2009 Aff Loss

10/22

114 N. Dew et al.


DOI: 10.1002/sej

do so. Hence, the endless streams of seed capitalprograms and culture-related incentives (such aspeer lending) increases the number of people start-ing new ventures. Research on psychological factorsof motivation is also interestingarguing for avariety of nonnancial motivators (Baum, Locke,

and Smith, 2001; Gimeno et al ., 1997) that are pre-sumably intrinsic to the entrepreneur and not easilyamenable to change through external incentives. Yetconsiderable recent evidence from psychology andbehavioral economics suggests that preferences dochange over time and that they may even be con-structed at times, for example, through the veryprocess of researchers trying to elicit them in thecourse of their investigations. As Paul Slovic (1995:365) stated in his address to the American Psycho-logical Association entitled The construction of

preference :

The meaning of preference and the status of value may be illuminated by this well-knownexchange between three baseball umpires. I callthem as I see them, said the rst. I call them asthey are, claimed the second. The third disagreed.They aint nothing till I call them, argues thethird.

Affordable loss tends to call them in the sense thatit disconnects objective (exogenously given) perfor-mance probabilities and resource requirements fromthe actual act of plunging. Thus, it allows potentialentrepreneurs to construct their preference for takingthe plunge even when their motivations are ambigu-ous and so-called rational decision criteria argueagainst taking the plunge. The statistics of newventure success and failure argue that any rationalcalculation based on expected return ought to biasthe decision against plunging, simply because of thelarge failure rate. In fact, it would take either a verylarge potential for gain or a high level of risk-tolerance to overcome the failure rate. However,affordable loss lessens the impact of possible failurebecause it makes failure clearly survivable by con-straining the loss to something that the entrepreneurregards as affordable and is willing to lose in orderto pursue the venture (the venture is consideredworth doing even if the invested amount is lost). Thisincreases the likelihood of plunging irrespective of the motivation to enter into entrepreneurship. Thereare at least four ways that the use of affordable lossas a decision heuristic increases the probability of entry into entrepreneurship: (1) it reduces the

threshold of nancial risk taking required; (2) itallows potential entrepreneurs to focus on thingswithin their control and proceed in spite of thingsoutside their control, thereby increasing both con-dence and creativity; (3) it makes explicit the factthat the upside is at least partly, maybe even largely,

endogenous to their own actions and those of theirstakeholders; and (4) it enables potential entrepre-neurs to choose projects that matter to them in waysbeyond the economic upside. Thus, even if the nan-cial upside is what decides the particular set of venture ideas they are considering (i.e., elements of the choice set), factors beyond the nancial upside(it is worth it even if I lose my investment in it) shapethe actual decision of which venture ideas to act on(i.e., provide choice criteria).

By reducing the nancial constraints, affordableloss increases the set of potential entrepreneurs whocan afford to take the plunge. And if a person isalready highly motivated to become an entrepreneur,by endogenizing the upside, affordable loss increasesthe probability he/she nds something worth plung-ing into. The former are provided with more reasonsfor saying yes and the latter with more reasons forsaying no to taking the plunge.

Thus, affordable loss reasoning is a biased mecha-nism for taking the plunge. It increases an individu-als probability of entering into entrepreneurshipeven if the failure rate is high and irrespectiveof exact motivations, nancial or otherwise (ascompared to expected returns and real optionsreasoning). Stated as a proposition:

Proposition 1: An entrepreneur using affordableloss reasoning will be more likely to take the

plunge than one using either expected returns or real options reasoning.

Behavioral aspects of the ability totake the plunge

Behavioral economics offers insights not only aboutthe willingness of people to take the plunge, but alsoabout their ability to do so, given that they do wantto become entrepreneurs. For example, how dopeople decide which resources are framed as afford-able to lose or not? On the one hand, it might bepossible to draw up an objective estimate of thedecision makers current nancial condition, i.e., apersonal balance sheet. On the other hand, we needto understand why some things get mentallyaccounted for or categorized as losable and other

8/2/2019 2009 Aff Loss

11/22

Affordable Loss 115


DOI: 10.1002/sej

things do not. This problem looks like a classicmental accounting problem applied to the plungedecision of the entrepreneur (Thaler, 1999). Let usexamine what the literature on mental accountingcan tell us about moving toward a rigorous under-standing of this particular aspect of the affordable

loss principle.The notion of mental accounting was rst devel-

oped in a paper by Thaler (1985) and later summa-rized by him (Thaler, 1999). Mental accountingemerges fairly straightforwardly from boundedrationality: creatures with limited cognitive process-ing capabilities require ways of keeping track of their money with limited memory space. Thayertheorized that people categorize resources in orderto keep track of them, much like accountants do inrms. For example, they create separate mentalcompartments for long-term savings (such as that forretirement and childrens education) and others forshort-term expenses (such as entertainment andleisure activities).

A key implication of mental accounting is theviolation of the fungibility premise of economics,i.e., that resources are automatically arbitragedacross different accounts (Thaler, 1999). A simpleway to think about this is that for Homo Economicus,money by any other name is still money, but formost Homo Sapiens, money in one mental accountis just simply not the same as money in anotheraccount. Because of this nonfungibility characteris-tic, mental accounting suggests that consumers mayborrow at high interest rates in some accounts evenwhile they save at much lower interest rates inothers. Similarly, some resources may be mentallyaccounted for in accounts that the entrepreneur willnot put at risk, whereas other resources are accountedfor in accounts that are available for risky investingin entrepreneurship. Just as the accounting of spend-ing behavior affects how consumers spend (Prelecand Loewenstein, 1998), the accounting of resourcesby entrepreneurs may affect how entrepreneurs makethe plunge decision.

Take, for example, the impact of windfallssuchas inheritanceson the plunge decision. A famousexample of this is Fred Smith investing his $2 millioninheritance (as well as his sisters $2 million) to startFedex. Prior research has found that individuals whoreceive an inheritance are signicantly more likelyto enter entrepreneurship than individuals who donot receive an inheritance (Holtz-Eakin, Joulfaian,and Rosen, 1994). However, the explanation for thisnding is debatable. From a cognitive perspective,

inheritances should make no difference to how aperson reasons through the decision to become anentrepreneur. This has led some researchers to con-clude that would-be entrepreneurs must be liquidityconstrained, but inheritances lift this constraint and,hence, enable entry into entrepreneurship. However,

Cressy (1996: 1253) subsequently showed this wasnot the case, pointing out that a reason why othershave seemingly identied start-up debt-gaps may bethe failure to test a sufciently rich empiricalmodel.

Affordable loss provides an alternative explana-tion for these empirical results: windfalls changewhat the potential entrepreneur accounts for aslosable ; they increase the entrepreneurs mentalbudget of affordable loss. This is because inheri-tances are more likely to be accounted for as housemoney and, therefore, they are more freely availablefor betting , i.e., that inheritances are mentallyaccounted for as funds available for risky investingin entrepreneurship (Thaler and Johnson, 1990;Weber and Zuchel, 2003). This suggests that wind-falls will have a larger impact on the likelihood of plunging than the same amount of money accumu-lated through savings, for instance. Thus, windfallshave a positive impact on start-up activity becauseof the effect they have on the entrepreneurs calcula-tion of affordable loss, not directly because they liftliquidity constraints. 6 Other examples of such wind-falls are stock options, lottery winnings (Lindh andOhlsson, 1996), and unexpected increases in assetprices (property prices, for example).

On the other side of the mental ledger are resourcesthat are accounted as being unavailable for spendingon entrepreneurship. Thaler (1990) suggests thatagents may use prudential heuristics, i.e., rules of thumb that preclude borrowing against or spendingcertain resources. For example, individuals mayhave rules that preclude borrowing against certainaccounts that are mentally accounted for as belong-ing to other parts of their life (for example, funds setaside for retirement, such as 401K, pension, etc.) ormentally accounted for as belonging to others (such

6 This conjecture could be tested through a market entry game-theoretic experiment where two randomly selected groups of subjects are asked to decide how much they would invest onentry. Both would be given similar levels of resources, withone group attaining it through a windfall and the other grouphaving saved it from their accumulated earnings. Since bothhave the same liquidity overall, the cognitive hypothesis wouldbe that subjects will place similar bets on market entry; thebehavior hypothesis would be that they will bet differently.

8/2/2019 2009 Aff Loss

12/22

116 N. Dew et al.


DOI: 10.1002/sej

as childrens college funds). In some instances, thehome may be accounted for in a not to be put at risk account because dependents rely on the home.

We hypothesize that two aspects of mentalaccounting are generally applicable to the plungedecision in terms of affordable loss. The rst is cou-

pling. Prelec and Loewenstein (1998) have proposedthat forms of payment that more closely couplepayment and consumption are avoided if possiblebecause they are more likely to call forth thoughtsof payment that undermine the pleasure of consump-tion. Prelec and Loewenstein (1998: 4) found thatcoupling . . . refers to the degree to which consump-tion calls to mind thoughts of payment, and viceversa. Some nancing methods, such as credit cards,tend to weaken coupling, whereas others, such ascash payment, produce tight coupling.In entrepreneurship, there are several well-knownstories of entrepreneurs who started their businesseson their credit cards (e.g., EDS by Ross Perot andthe founding of Home Depot). Because credit cardsweaken the coupling between decision and payment,we expect that this will affect the way potentialentrepreneurs evaluate the costs of taking theplunge. Other examples of weak coupling mayinclude loans from family members that have exi-ble or unspecied payback terms (sometimes casu-ally referred to as spending somebody elses money).Research on family business, for example, refers tothe relatives money as patient capital .7 The use of this money allows an entrepreneur (or familymember) to continue operating in business withoutthe heavy pressure of deadlines to repay or earn aspecic return in a very short period of time (forexample, see the work of Sirmon and Hitt, 2003).

Since individuals will seek the hedonic benets of decoupling where possible, those involved in entre-preneurship will prefer to use mental accounts thatmore weakly couple the experience of riskingresources with the decisions being made. Theseaccounts are less painful to lose than accounts thatare more strongly coupled with the plunge decision.As a result:

Proposition 2: Weakly-coupled forms of payment will raise a potential entrepreneurs level of affordable loss and, therefore, increase both thelikelihood of taking the plunge and the ability totake it.

The second aspect of mental accounting thatdirectly impacts the affordable loss levels leading tothe plunge decision has to do with the effects of accounting in different units, i.e., time versus money(Okada and Hoch, 2004; Soman, 2002). 8 Becausenancial losses are painful, Thaler (1999: 188) states

that we should expect to see that some of the discre-tion inherent in any accounting system will be usedto avoid having to experience losses.

Two qualities differentiate time and money. First,the value of time is more ambiguous than money.Second, time is perishable; it cannot be stored ininventory or saved for later use the way money canbe. Third, people tend to calculate returns on timeinvested using nonnancial metrics. For example,entrepreneurs such as Scott Cook, founder of Intuit,talk about celebrating failure, for they value learningfrom failure. These differences suggest that lossespaid for in time may be experienced as more afford-able than losses paid for in money because theirambiguity means they can be accounted for moreexibly.

Therefore, we might expect that the currency, inwhich mental accounting occurs, matters for raisinglevels of affordable loss for potential entrepreneursand, thereby, increases the probability that they willactually take the plunge. 9 To our knowledge, theeffects of mentally accounting in time has not so farbeen studied in an entrepreneurial setting. However,anecdotal evidence suggests that entrepreneurs mayexibly substitute time for money in the new venturesetting. This is sometimes referred to as sweat equity .Thus, sweat equity may be a factor in the plungedecision and some proportion of entrepreneurs maysubsequently sweat it out over longer periods of time. Because of the ambiguity and perishability of time, ventures where the investment can be easilyconverted into time inputs are perceived to be moreaffordable than ventures that require cash invest-ments. Stated as a proposition:

7 We thank the editor for turning our attention to this importantand relevant stream of research.

8 Our analysis focuses on money and time, but we recognizethat individuals may have other important resourcessuch asreputationthat may warrant investigation from a behavioralperspective in future work.9 It is intriguing to also consider currency other than money andtime. A nonexhaustive list could include reputation and rolodex.While we speculate that accounting in these nonnancial cur-rencies will also likely increase the ability to take the plunge,current research in behavioral economics does not yet addressthese. A deeper empirical understanding of the entrepreneurialplunge decision, we believe, can bring original insights tobehavioral economics on some of these unstudied units of mental accounting.

8/2/2019 2009 Aff Loss

13/22

Affordable Loss 117


DOI: 10.1002/sej

Proposition 3: When entrepreneurs account intime (versus money), they will have higher levelsof affordable loss and, therefore, will be morelikely to take the plunge.

Behavioral aspects of the depth of the plunge

The concept of affordable loss urges individuals toenter into entrepreneurship based on a loss that, eventhough it is only possible, is known to be affordableand that they have come to terms with before takingthe plunge. Given that preferences may change overtime (Ainslie, 2001; March, 1978) raises the impor-tant theoretical question of how the potential entre-preneur secures the plunge decision. In other words,how do entrepreneurs make the plunge work overrelevant periods of time during which they may besorely tempted to second guess themselves and quit,while at the same time ensuring their ability to quitwhen it would be prudent for them to do so?

When making the plunge decision effectually, themechanism that both supports the willingness to losean affordable set of resources and enables the quit-ting of the new venture at the point those resourcesare exhausted is the precommitment that one iswilling to lose a select and nite set of resourcesover which one has control. For instance, if theentrepreneur says My affordable loss is $200,000and two years of my life in this venture, then thischoice rests not only on a commitment that the entre-preneur is willing to lose two years and $200,000,but also on a commitment that after this point, theentrepreneur will indeed quit the venture, i.e., this isall that the entrepreneur is willing to lose. 10 More-over, the commitment is based on entrepreneurspreferences over their means, not on information inthe environment that might change over time andlead them to change their minds. Thus, the commit-ment aspect of affordable loss is double edged: itinvolves the resolve that one is willing to lose certainresources, as well as a constraint that this is all oneis willing to lose. Both aspects rely on some kind of commitment mechanism.

What enables human beings to make such com-mitments to themselves? One explanation is thatemotions play an important role in serving as com-mitment devices that enable people to behave

consistently over time (Frank, 1988; Nesse, 2003).For a vivid description of the emotional state thatsometimes accompanies the plunge decision, con-sider the following example drawn from Tom Fatjosautobiography (Fatjo and Miller, 1981). Fatjo wasan accountant in Houston when a meeting with the

people living in his subdivision challenged him totake up the garbage collection problem the commu-nity was facing. In 1970, he borrowed $7,000 for hisrst truck. Every day, Fatjo woke up at 4 a.m. tocollect garbage for two hours before changing intoa suit to go to work in his accounting ofce. He didthis for over a year before he quit his day job tofound the waste management giant Browning Ferris.Of course, when he made the decision to take theentrepreneurial plunge, he did not know he wouldend up building a $1 billion enterprise. Here is howhe describes his moment of decision:

Within a week, I was almost frantic. My foodwouldnt seem to digest and I had a big knot inmy chest. When I was doing one thing, I thoughtof two others which had to be done that sameday.

The pressure just kept building. Even though itwas cold, my body was damp from continuousperspiration. Since so much of what I was doingin the accounting rm had to be done by the endof the tax year and involved important decisionswith key clients, I needed to spend time thinkingthrough problems and consulting with them asthey made decisions. I was caught in a triangle of pressing demands, and I felt my throat constrict-ing as if there were wires around my neck.

That night I was exhausted, but I couldnt sleep.As I stared at the ceiling, I fantasized all ourtrucks breaking down at the same time. I wastrying to push each of them myself in order to getthem going. My heart began beating faster in thedarkness and my body was chilled. The horriblethought that we might fail almost paralyzed me.

I wanted to quit and run away. I was scared todeath, very lonely, sick of the whole deal. As hardas I tried to think about my life and what wasimportant to me, my mind was just a confusedmass of muddled images . . . I remembered com-mitting myself to make it in the garbage businesswhatever it takes ! I lay back on my pillow and felta deep sigh within myself Good Lord, so this iswhat it takes , I thought, then rolled over and gotsome restless sleep. (Fatjo and Miller, 1981:32)

10 Note that the argument would be the same even if he/she doesnot make an immutable decision to quit, but makes only a rmcommitment to seriously consider the decision to quit.

8/2/2019 2009 Aff Loss

14/22

118 N. Dew et al.


DOI: 10.1002/sej

Fatjos decision embodies the principle of affordableloss. In his narrative of the plunge decision, Fatjoappears to be grappling with the worst case scenarioand striving to come to terms with it through a com-mitment that explicitly divorces his decision fromthe possible consequences.

Yet Fatjos narrative also illustrates the powerfulemotional component of such decisionsand thedangers that accompany those emotions. It showshow the decision had the kind of felt, embodiedcomponent that Miller (2007) has urged us to recog-nize. Fatjo was damp from continuous perspiration,felt (his) throat constricting, and exhausted butcouldnt sleep. Eventually, the way he silenced themuddled images was by making a commitment dowhatever it takes . Such emotional intensity, whileperhaps necessary to overcome the opportunity costsassociated with giving up a highly paid white-collar

job, may also induce blind fervor of the kind thatleads to nancial ruin. Here a precommitment toaffordable loss levels can put the brakes on sucha rush to ruin.

Among the emotions that act as bulwarks againstbreaking commitments to ourselves, guilt and shameare prominent. People readily use these feelingsagainst themselves, i.e., people feel badly if theydont keep their promises to themselves. In effect,these emotional predispositions provide incentivesto act in particular ways or serve as a tax on notbehaving in certain ways (Elster, 2000). Thus, froman economic perspective, these emotions are valu-able commitment-rendering devices that are directedat the selfat securing a decision you made to your-self so that you do not renege on it in the future. 11 These emotion-based commitments have to work with whatever stop-loss mechanisms and proceduresthey employ to limit their losses.

Therefore, the affordable loss heuristic providesthe potential entrepreneur with the resolve to takethe plunge and also the emotional back up requiredto quit when the time comes. This heuristic hasan important side effect: it serves to protect the

entrepreneur from the well-known susceptibility toescalation of commitment bias (Staw, 1976). In con-ventional analyses of investment decisions, research-ers have found that investors often throw goodmoney after bad when they nd projects underper-forming in terms of their expectations of return

(Staw, 1981). At these points of disappointment,investors begin to perceive the costs thus far incurredas sunk and so arrive at irrational decisions to investmore money in bad projects. Affordable loss pro-vides a safeguard against this by shifting the empha-sis to the downside at all times. In fact, even beforeone begins, affordable loss insists on a precommit-ment to quit when the affordable loss amount isactually lost. Unlike standard cases of escalation of commitment where the upside still reigns supremein the decision-making processand, therefore, thecosts incurred thus far come to be seen as sunkthe focus in the case of affordable loss is always onthe downside and the precommitment to quit ensuresthat the decision to continue is not about unendinghopes of the upside, but once more thinking throughwhether any new investments are worth losinginterms of nonnancial upsides that the entrepreneurreally values enough to make the new investmentworthwhile. We speculate, therefore, that:

Proposition 4: Entrepreneurs who make the plunge decision using the affordable loss heuristicwill be less susceptible to escalation of commit-ment than those who use calculations of expected returns. 12

DISCUSSION

In this section, we discuss several implications of the affordable loss heuristic that are relevant forresearch and public policy in entrepreneurship.These are, in turn, the frequency of venture start-up,the costs of failure, and the efciency implicationsof underinvestment in entrepreneurship.

Frequency of venture start-upA long-standing and important research puzzle inthe economics of entrepreneurship is the issue of

11 This type of self-commitment can be traced to Homersparable about Ulysses, who had himself bound to the mast of his ship in order to resist the temptatious song of the Sirens(Elster, 2000). Essentially, in a moment when his thoughts wereclear and passions were in a state of balance (a moment of reective equilibrium/considered judgment), Ulysses decidedto take precautions against future changes in his preferences.Commitment devices, then, are normally thought of as ways toprotect oneself against glitches in ones judgments; the idea,then, is that commitments keep the decision maker rational overtime.

12 Here we have deliberately limited the scope of our analysisof the emotional aspects of decision making to the issue of commitment. However, we note that a variety of propositionsrelated to emotion could be imagined (for example, recent work on grief [Shepherd, 2003]), and we, therefore, return to thisissue in our concluding remarks. We are grateful to an anony-mous reviewer for drawing our attention to this point.

8/2/2019 2009 Aff Loss

15/22

Affordable Loss 119


DOI: 10.1002/sej

excess entry and suboptimal returns observed inempirical studies of entrepreneurship (Geroski,1996; Caves, 1998). The statistics of new venturesuccess and failure argue that any rational calcula-tion based on expected return ought to bias decisionmakers against plunging, simply because of the

large failure rate. But the fact is that this does notdeter entrepreneurs from (over)entering. Whilestudies have suggested a wide range of individualcharacteristics that might help explain these data,some key explanations have recently been contestedin the literature. For instance, Moore and Cain (2007)signicantly moderate the conclusions of Camererand Lovallo (1999) regarding overcondence/refer-ence group neglect; and Miner and Rajus (2004)results contest Stewart and Roths (2001) regardingthe long-standing hypothesis of preference for risk.

In our view, entry into entrepreneurship is a choicethat may be supported by a variety of reasoningprocesses that are contingent on the exact nature of the decision problem as well as on the characteristicsand circumstances of the decision maker. Someentrepreneurs may indeed be more risk loving thanothers. Others may exhibit over- or undercondenceand yet others may simply be ignorant of failurerates. But at least someif not mostexpert entre-preneurs use affordable loss. We have shown in thisarticle that affordable loss is not only empiricallyinduced, but also theoretically consistent with whatwe know about human information processing ingeneral and recent ndings in behavioral economicsin particular.

By contrast, more formal economic models of entrepreneurial entry start from the assumption thatindividuals enter into entrepreneurship only when itpays to do soi.e., only when the expected value oroption value of the plunge is positive (based on dataabout payoffs, failure rates, and probabilities). Com-paring affordable loss to heuristics that begin withexogenous expected return reveals that when indi-viduals use the affordable loss heuristic, they maytake the plunge even if the failure rate is high andirrespective of potential gain, i.e., they may take theplunge when the expected value of entering entre-preneurship is negative. Therefore, affordable losssuggests a higher entry rate than expected returns.

This prediction is important because it suggestsan alternative behavioral explanation for the excessentry/poor returns puzzle. Importantly, one virtue of this explanation is that it does not depend on anycognitive mechanisms or biases that are specic tothe entrepreneurial population, i.e., it does not

depend on an empirical difference between entrepre-neurs and nonentrepreneurs. 13 Moreover, affordableloss is a heuristic for taking the plunge at any givenlevel of liquidity, so it applies to the whole popula-tion of potential entrepreneurs regardless of hetero-geneity in the initial distribution of wealth/resources.

But perhaps most importantly, the affordable lossheuristic is teachable and learnable with the addedbenet of low cognitive burden. Put simply, thefollowing four steps capture the heuristic in theclassroom:

Think through what you can afford to loseamounts set aside in weakly-coupled mentalaccounts, sudden windfalls, savings you havebeen setting aside for an entrepreneurial debut,etc.

Think through how much you are willing to losefor the particular project steps you are actuallyplanning to takehalf of the above amount, forexample, so you can try two projects instead of one, in case the rst one fails.

Take those steps at those levels of investment if you feel comfortable that those steps are worthtaking even if you lose all your investmenti.e.,think through nonmonetary benets.

Think creatively about how you can reduce actualcash outows on this investmentand continuallystrive to drive it close to zero.

This cognitive simplicity suggests that we can sup-plement nancial incentives for increasing the fre-quency of entrepreneurial start-ups with a pedagogicalone, surely a claim worth investigating both froma normative policy perspective as well as from adescriptive scientic one.

Costs of failure

The fallibility and error-prone nature of entrepre-neurial efforts have been well argued in the literature(Christensen and Knudsen, 2004). This is why expe-rienced and self-aware entrepreneurs have failurermly in mind when they take the plunge based onaffordable loss reasoning. These entrepreneurs andtheir stakeholders explicitly consider the costs of assembling and disassembling new ventures. The

13 Though differences will emerge from differential learningopportunities, i.e., there will be differences between how adeptexpert entrepreneurs are at using affordable loss compared tonovice entrepreneurs.

8/2/2019 2009 Aff Loss

16/22

120 N. Dew et al.


DOI: 10.1002/sej

affordable loss principle works to keep potentiallosses per stakeholder down (with each stakeholderindividually assessing their own affordable loss)while keeping the venture open to unexpected newpossibilities on the upside that may come from avariety of sourcesinternal, organic, market driven,exogenous, or entirely ad hoc and unpredictable.

Figure 2 presents graphically the overall argumentrelating the use of affordable loss as opposed toexpected return to the performance of the rm. Thehorizontal axis on Figure 2 is time. The vertical axismeasures nancial investment at each point in time.Note that this is exactly the same as the amount of money lost at each point in time, should the venturefail at that time. Assume now that for any givenventure that survives and grows over time, thereexists an ideal level of investment required. For thepurposes of this argument, we can limit this assump-tion to the ex post actual investment level. Figure 2posits a generic S-curve to capture the cumulativeshape of these investments over the life of the sur-viving venture. The S-curve is widely acknowledgedto adequately capture the diffusion process in a newmarket (Rogers, 1995). The only leap this assump-tion makes is that investment required by the newventure (if it survives and grows) will follow thegrowth pattern of the market. Note that for our argu-ment to hold, several other types of cumulativeinvestment curves would work just as well.

When entrepreneurs take the plunge based onexpected returns, they can make one of two types of errors: they can overinvest or underinvest. And their

investment performance in the given venture heavilydepends on the accuracy of their predictions.However, when entrepreneurs plunge based onaffordable loss, their investments grow as a functionof survival (with incremental investments beingmade based on affordable loss reasoning). One con-sequence of this is that they would almost alwaysunderinvest in relation to the ex post actual invest-ment curve. However, should an unpredictable orexternal shock occur, entrepreneurs using affordableloss are almost always likely to lose less than predic-tion-oriented entrepreneurs. It is in this sense thataffordable loss reduces the cost of failure, irrespec-tive of the probability of failure (Sarasvathy,2001a).

In summary, this implication, when taken togetherwith the implication that affordable loss leads tomore frequent venture start-up, means that afford-able loss results in more entry into entrepreneurship,but when failures occur, the losses are smaller. Incontrast, reasoning from an expected returns (NPV)basis results in fewer entries and larger losses whenfailures occur. From a policy perspective, althoughexcess entry by the wrong types of entrepreneursmay be costly, the lower costs of failure are a benet.Overall, which alternative is normatively most desir-able may depend largely on factors such as the pre-vailing technology regime and institutional regime(Winter, 1984; Lee, Peng, and Barney, 2007). Webelieve that sorting these considerations out bothfrom micro- and macroperspectives would provideseveral exciting projects for future research.

Investment based on

Affordable loss

Low

High

Investment levels / failure costs

Timeline

Control gap:Use ofeffectual logic

External shock Investment based on

Expected return

Prediction gap:Investmentsin accuracy

Investment based onExpected return

Actual investmentrequired (Ex-post)

Figure 2. Firm performance: affordable loss and expected returns compared

8/2/2019 2009 Aff Loss

17/22

Affordable Loss 121


DOI: 10.1002/sej

Efciency implications of underinvestment

Several other implications follow from the hypoth-esis that entrepreneurs frequently (and, in a sense,deliberately) undercapitalize their ventures. First,this issue has important policy implications because

liquidity constraints have often been used as anexplanation for the high failure rate of new rms,i.e., rms fail because they are undercapitalized and,therefore, run into cash ow problems (Cressy,1996). While undercapitalization might increase therisk of ruin for a rm (Baxter, 1967) it may neverthe-less lower the risk of ruin of the entrepreneur (whosurvives to start another venture) and lower the costsof failure per venture (Sarasvathy, 2001a). Nonethe-less, since rm failures are visible and measurable,many government initiatives across the globe seek to supplement the resource bases of new ventures bytax breaks or preferential nancing arrangements.Second, underinvestment might also imply that rmsfounded by entrepreneurs who use affordable lossreasoning are more likely to miss homeruns, i.e., areless likely to capture their full upside potential(regardless of whether this is a homerun or mediocreopportunity) in markets with explosive growth andhigh rates of return that require large amounts of nancing quickly. Here the argument is one of missed opportunity rather than outright failure.

In a recent empirical paper, Wiltbank et al. (2009)tests these predictions using data from a sample of angel investors (wealthy individuals who act asinformal venture capitalists by placing their ownmoney directly into early stage new ventures). Inter-estingly, the empirical results do not support theabove predictions about failure and missed opportu-nities. The results of the Wiltbank et al. study showthat angels who emphasize effectual strategies (of which affordable loss is one component) actuallyexperienced a reduction in investment failures (notan increase) without a reduction in their number of homeruns. One possible explanation for these coun-terintuitive ndings is that the affordable loss heu-ristic may tend to be used in combination with othertactics. Though nancial resources are clearly veryimportant in new ventures, they are not the onlyresources that are important for eventual success orfailure: ventures may survive and thrive because thefounding entrepreneur/s managed to nd ways of supplementing the nancial resources of the venture.The literature on how entrepreneurs may use socialco-opting strategies to establish legitimacy andsecure access to underutilized resources appears

very relevant here (Starr and McMillan, 1990; Bakerand Nelson, 2005). Or entrepreneurs may use effec-tual strategies such as bringing on board a variety of self-selected stakeholders that help shape and growthe market organically rather than through nancialinvestments (Sarasvathy and Dew, 2005b). Or they

may attempt to substitute sweat equity for nancialresources, i.e., invest large amounts of their ownlabor into their venture.

Finally, undercapitalization also has implicationsfor the plurality of new venture investor types.Undercapitalization speaks to an obvious gap inmost theories of the rmproperty rights, resource-based view, behavioral theory, contracting, transac-tion cost, etc. But these theories do not explain howthe rm was put together in the rst place and, there-fore, do not take into account the implications of thestart-up situation for the efciency of the subsequentbundle of assets or contracts that constitutes theventure (Hellmann, 2000). One hypothesis is that theappropriate reasoning approach of venture investorsis contingent on the life stage of the venture. Whereasreasoning based on affordable loss may increase thelikelihood of entrepreneurs taking the plunge, theiruse of the heuristic may limit the ventures growthpotential down the road. This suggests a theoreticalreason why predictive stakeholderssuch as venturecapitalistsmay be necessary to the survival andgrowth of high-potential ventures. It also leads to aninteresting paradox that good entrepreneurs may,under some circumstances, make bad investors fornew ventures.

CONCLUSION

We began this article with the objective of develop-ing a deeper understanding of the affordable lossheuristic as a part of the toolbox available to indi-viduals contemplating the entrepreneurship plungedecision. In doing so, our aim was to contribute tothe exciting new conversation emerging in our eldon a more creative view of entrepreneurship.

Interestingly, the downside focus that a behavioraleconomic view of affordable loss brings to theplunge decision may be used with all three views of entrepreneurial opportunities. For example, even inthe case of opening a franchise for a well-establishedcompany such as McDonalds, potential franchiseescan evaluate their plunge using an affordable lossheuristic. They can ask themselves not only howthey can raise the initial investment required to open

8/2/2019 2009 Aff Loss

18/22

122 N. Dew et al.


DOI: 10.1002/sej

the franchise, but also come to grips with worst casescenarios on what could go wrong and come up withcreative ways to reduce that potential loss. Our prop-ositions will all hold even in cases of such mundaneor more readily recognizable opportunities. But incases where opportunities are made , rather than

found, affordable loss is particularly useful becausethose are precisely the cases where the upside ismost unpredictable. In cases where entrepreneursare choosing between ventures with highly unreli-able but promising upsides, or in cases where theyare driven largely by non-nancial motives, afford-able loss gives them reasoned and even systematicdecision criteria without the necessity of spendingtime, money, and effort on estimating upsidesprobably an exercise in ction rather than fact orforecast in any case (Goodman, 1955).

We conclude with three departing thoughts. First,there are some important questions regarding theplunge decision. What decision-making tools shouldbe taught to potential entrepreneurs? Should weteach them only models based on NCIT and ROR inan entrepreneurial setting, or should we also teachthem how to use the affordable loss heuristic? 14 When is it more or less appropriate to teach each of these different decision-making tools? As Miller(2007) has suggested, with regard to risk taking, thebroader issue at stake is the whole notion of what isdesirable as rational behavior. This is a widelydebated issue among philosophers, psychologists,and economists. Are decisions rational if the proce-dure is logical (as highlighted by Simon, 1978) oronly if they express substantive rationality (in thesense of conforming to the expected utility model)?Are they rational if the outcomes are good or onlyif the decision input is substantively correct? In arecent article, Haselton and Nettle (2006: 63) explainthat many of the simple heuristics that people actu-ally use perform just as well as complex normativemodels under real-world conditions of partial knowl-edge (Gigerenzer and Todd, 1999). There are evencircumstances in which they perform better thannormative modelsthe so-called less-is-more effect.The less-is-more effect occurs because simple heu-ristics can exploit structural features of the decision-making environments that are noisy and uncertainand contain multiple cues.

Thus, human minds appear to work using a set of simple heuristic procedures and perform best when

decision problems are presented in ways that lever-age natural capability by putting them in ecologi-cally valid formats. One conjecture is that theaffordable loss heuristic is possibly another exampleof the less-is-more and biased-is-better effects(Haselton and Nettle, 2006). It uses less information

and it is biased against external information. And itmay produce better results in a specic environmen-tal context, i.e., one that is noisy, uncertain, andcontains multiple cues. The implication of this argu-ment is that we should teach students decision tech-nologies that are adaptive (and, therefore, appropriate)across a spectrum of circumstances. This means thatwe might usefully teach potential entrepreneursabout both the affordable loss heuristic and the EUmodel as part of a package of (contingently applied)decision-making tools.

Second, we believe there is a signicant opportu-nity to enrich research on entrepreneurial cognitionwith psychological research on how individuals feelabout decisions, actions, and thoughts. We notemuch excitement among entrepreneurship research-ers and some pioneering work developing on thetopic of entrepreneurial cognition (Mitchell et al. ,2004). Researchers have also been developing ideasabout the role of affect in entrepreneurship and asso-ciated areas such as creativity and innovation (Adlerand Obstfeld, 2007; Goss, 2005; Shepherd, 2003).Our view is that entrepreneurship involves morethan cold cognitive processes. Key entrepreneurialdecisions (such as the plunge decision) are deeplypersonal choices that are frequently viewed as sig-nicant life choices; therefore, we should expect theentrepreneurs feelings about these decisions to playan important role in such choices. There remainssignicant scope for further research on this topicabove and beyond the commitment issues we havehighlighted in this articlethat could help us betterunderstand entrepreneurial behaviors, while atthe same time, usefully informing pedagogy andpractice.

Finally, if affordable loss plays a role in the plungedecisions of entrepreneurs, the individuals objec-tive function may not be directed at prot maximiza-tion. Selecting a decision strategy rooted in affordableloss fundamentally prioritizes control of downsideloss above the maximization of potential upside.This is not to say that affordable loss will alwaysresult in a suboptimal result from a societal perspec-tive or that expected return will always entail assum-ing more risk than affordable loss. But what it doessuggest is that existing research puzzles about

14 We are grateful for the comments of an anonymous reviewerwho prompted us to think about this issue in more depth.

8/2/2019 2009 Aff Loss

19/22

Affordable Loss 123


DOI: 10.1002/sej

entrepreneurial wealth creation may, in part, be anartifact of theoretical perspectives that ignore crucialbehavioral factors that shape the decision to start anew venture. If researchers assume prot maximiza-tion is the priority of entrepreneurs, entrepreneurialoutcomes are correctly measured according to ROI,

IRR, and perhaps sales revenue and sales volume.But if the entrepreneur looks to manage risk throughaffordable loss, the focus may be on differentandperhaps conictingdependent variables. There-fore, to the extent that the theory expressed in thisarticle is empirically signicant, it raises fundamen-tal questions about the implicitly assumed risk-taking practices of entrepreneurs.

For instance, we do not know of any historicalstudy that specically examines the risk-taking heu-ristics used by well-known entrepreneurs such as theWedgwoods, Hersheys, Edisons, Watsons, Dells,and Schultzes of the world. But we are fascinated bythe possibility of what we might nd were we toexamine accounts of their decision processesespecially in terms of contingent relationshipsbetween their use (or lack of use) of affordable loss,risk taking, and eventual outcomes over a career of multiple entrepreneurial ventures. Edison, forexample, had been on the brink of bankruptcy, andHershey and Heinz had been through more than one.Even Wedgwood bet his entire net worth at leastonce in his career. Which of these were strategic (asin the case of North American Phonograph Companythat allowed Edison to buy back the rights to hisinvention), exogenous (as in the case of Hersheysearlier ventures), and/or avoidable through the use of precommitments to affordable loss levels (as in thecase of Edisons Portland Cement Company)? Thus,rethinking entrepreneurial outcomes in the contextof the behavioral processes we have described in thisstudy (such as contemplating preferences for becom-ing and being an entrepreneur, and exiting venturesto meet self-imposed precommitments rather thanbecause the venture failed) has potential for signi-

cantly enrichingif somewhat complicatingourunderstanding of entrepreneurial wealth creation.

REFERENCES

Adler PS, Obstfeld D. 2007. The role of affect in creativeprojects and exploratory search. Industrial and Corpo-rate Change 16(1): 1950.

Adner R, Levinthal D. 2004. Real options and real trade-offs. Academy of Management Review 29(1): 120126.

Ainslie G. 2001. Breakdown of Will . Cambridge UniversityPress: Cambridge, U.K.

Aldrich H, Kim P. 2007. Small worlds, innite possibili-ties? How social networks affe

2009 aff loss

Documents