Industrial and Organizational Psychology, 2 (2009), 130–143.Copyright c© 2009 Society for Industrial and Organizational Psychology. 1754-9426/09

FOCAL ARTICLE

Executive Selection—What’sRight . . . and What’s Wrong

GEORGE P. HOLLENBECKHollenbeck Associates

AbstractAlthough recent reviews of executive selection have catalogued much that we as industrial–organizational (I–O)psychologists are doing right in our research and practice, we are confronted with the facts that executiveselection decisions are often, if not usually, wrong and that I–O psychologists seldom have a place at the tablewhen these decisions are made. This article suggests that in our thinking we have failed to differentiate executiveselection from selection at lower levels and that we have applied the wrong models. Our hope for the futurelies not in job analyses, developing new tests, meta-analyses, or seeking psychometric validity, but in viewingexecutive selection as a judgment and decision-making problem. With the right focus, applying our considerablemethodological skills should enable us to contribute toward making better judgments. When we have a bettermousetrap, organizations (if not the world) will beat a path to our door.

Introduction

How we select executives is a hot topicnot only in the popular press but also inindustrial-organizational (I–O) psychology,which seems to have a resurgence of interestin employee selection as a whole andin executive selection in particular (seereviews of executive selection by Howard,2001; Thornton, Hollenbeck, & Johnson, inpress; and of employee selection in Farr &Tippins, in press; see Highhouse [2008] inthe September 2008 issue of this journal fora discussion of related issues). This articleis an attempt to place executive selectionin the context of employee selection as awhole. It considers in passing the manythings I–O psychology is doing well in thearea but examines in more detail where wehave failed and the implications of those

Correspondence concerning this article should beaddressed to George P. Hollenbeck.E-mail: geoholl@livingston.net

Address: Hollenbeck Associates, 109 Whippoor-will Drive, Livingston, TX77351

failures, followed by suggestions to enhanceour contributions.

Who is an executive? Silzer’s (2002)definition follows; he includes generalmanagers, corporate officers, and heads ofmajor organizational functions and businessunits. ‘‘High potentials’’ are those deemedto have the potential to become executives.Our discussion focuses on executives ratherthan high potentials, although some ofwhat we say will apply to high potentialselection as well, especially ‘‘what we aredoing right.’’ And, although there may bewide variations in the selection processdepending on whether the candidates arefrom inside or outside an organization, andselecting CEOs has its particular issues,these differences will often be glossed overin the interest of brevity.

The flexible format of Perspectives allowsa presentation more or less in the formof a ‘‘last lecture’’—reflections over along career—based on the lessons of theliterature, years of practice as an I–Opsychologist and as an executive, and

130

Executive selection 131

counsel with many wise colleagues. Aneffort will be made to forgo the useof citations in expressing ‘‘the commonwisdom,’’ reserving those for references thatmight be useful as our younger and brightercolleagues pursue the questions raised.

Before listing our rights and wrongs, weneed to make explicit some key assump-tions, based perhaps more on philosophyand experience than on data:

• Executive performance matters—alot. Some have argued that executiveperformance doesn’t matter much inan organization’s success or failure.I don’t believe that—I believe thatexecutives and executive selectionare absolute keys to organizationsuccess. If you doubt that assumption,read Finkelstein (2003) or Akst’s(2008) review of Carroll and Mui’sBillion-Dollar Lessons.

• Executive success is predictable wehear voices saying that most (70%) ofthe variance in executive success isunpredictable. There is indeed a longand tenuous link between what anexecutive does and whether or not heor she ‘‘succeeds,’’ but in this articleI hold to the notion that the worldis not random, and that if we kneweverything, our selections would beperfect. Everything is not knowable tous, of course, prior to hiring, despiteour talent for identifying (with 20/20hindsight) every cause of failure.

• People don’t change—much. Peo-ple grow, learn, and develop newrepertoires of behavior, but thereis a ‘‘there there.’’ After becom-ing adults, if not before, peoplehave underlying characteristics thatserve as bases for their behaviorover the years. The Greeks knewthis and called it character. True,sometimes small changes in behav-ior can make significant changesin results, but patterns emerge atseveral levels—other things beingequal, extraverts remain extraverts,turnaround executives tend to stay

that way. The implication, of course,is that selection is important. Mostexecutives believe this; selection peo-ple believe this; most developmentpeople, not surprisingly, see the worldfrom the other side!

• Past performance predicts future per-formance—sometimes. Again, no-body would argue that people can’tget better or worse at what they do,but knowing what people have donein similar situations in the past gives usa window on what they are most likelyto do in the future. Given dramaticallydifferent performance demands, pre-dicting from past performance mayget dicey, and we may need to predictfrom more fundamental perspectives.

• The people make the place. We canthank Ben Schneider for eloquentlybringing to our attention in his 1985SIOP Presidential Address the impor-tance of the people in determiningwhat an organization is like. Althoughmany factors influence ‘‘the place’’(e.g., industry, company, compensa-tion structure), those of us who haveworked in different organizations willattest to how different places canbe—IBM is not Merrill Lynch; Mer-rill Lynch is not Fidelity Investments.And, executives are disproportion-ately important in what the organi-zation is like, starting with the CEO.The old saying goes, ‘‘The fish stinksfrom the head down.’’ The impli-cation again: executive selection isimportant.

Given that context, let’s look first at thegoods.

What’s Right

Our knowledge of and sophistication inthe technology of selection in general hasgrown exponentially in the past half cen-tury. Although the basic process and princi-ples may be the same (see Campbell, Dun-nette, Lawler, & Weick, 1970; Dunnette,1966; Guion, 1965), we have a much better

132 G.P. Hollenbeck

understanding of many of the issues. Con-sider for example the following list, in noparticular order and certainly not exhaus-tive. There are so many ‘‘rights’’ that eachis given short shrift, largely citation free:

• Better interviews—if interviews arestill the most widely used selectiontechnique, they may also be themost studied technique. They maystill be subject to many of the age-old criticisms, but our understandingof the process has in many casesbeen translated into more effectiveinterview processes and interviewertraining.

• Better tests—testing theory and tech-nology have developed to the pointwhere if a thing exists, we can mea-sure it reliably, often in half a dozenor fewer well-chosen test items. Well-developed tests today exist for not justcognitive abilities but for personality,interest, derailment factors, and so on,not to mention the technology to testcandidates around the world.

• Assessment centers—whether thoughtof as a testing technology or assess-ment methodology, assessment cen-ters have been a major focus ofI–O psychology and have stimulateda wealth of research, thinking, andapplications.

• 360 feedback—although 360 feed-back existed in a nascent form inthe 1960s at Exxon and at IBM inthe form of a ‘‘Rate Your Manager’’process, 360 feedback has becomestandard practice for executive devel-opment and is widely used for mea-suring performance as well as execu-tive potential, especially with internalcandidates.

• Fit—seldom is an executive selectiondone today without considerationof the strategic fit of the executiveand the position. Strategy had notordinarily been the province of I–Opsychologists when it hit our screenin the early 1980s, and it has becomea regular feature of our efforts.

• Session Cs—wise practitioners haveknown for years that ‘‘the conver-sation’’ where executives discussedthe capabilities and possibilities ofother executives was more importantthan performance appraisals or rec-ommendations. Honed to a fine art byGeneral Electric as ‘‘Session Cs,’’ thisroll up of executive selection infor-mation has become widespread incorporations, both large and small.

• Range of candidates—diversity, in allits forms, has become the order of theday in the executive suite, whethergender, geography, culture, age . . .

you name it.• Hiring strategies—talent manage-

ment is another ‘‘order of the day,’’and a big part of that is strategic hiring,not just for ‘‘fit’’ of individual execu-tives but for optimizing the use of anorganization’s resources (e.g., whereto seek ‘‘acceptable’’ and where toseek ‘‘outstanding’’ employees).

• Board involvement—CEOs oncechose their own replacements. Nolonger. Boards of directors are notonly involved in the selection of theCEO and other senior executives,good board practice now calls for aregular review of the executive talentpools throughout the organization.

• Behavioral competency models—abenchmarking study by the AmericanProductivity and Quality Council in2004 found that every one of their bestpractice organizations had developeda behavioral competency model,designed to guide their selection anddevelopment efforts. It is difficultto find any organization today thatdoesn’t have its competency model.

In addition, other factors are beingconsidered in the selection process. A lookback at executive selection over the yearsis surprising in the narrow range of factorsthat were typically included in our modelsand even in the historical disdain for someof the variables routinely considered today:

Executive selection 133

• Personality is back—after ignoring,if not denying the importance ofpersonality variables, personality is anintegral part of our thinking. At the toplevels, the range in cognitive ability isoften thought to be so narrow thatmuch of the performance variationlies in these noncognitive factors.

• Leadership is in—in a world wherethe terms ‘‘leader’’ and ‘‘executive’’are largely interchangeable, it may behard to believe that ‘‘leadership’’ perse was not a dimension in the orig-inal AT & T Assessment Center. The1970s and 1980s saw an emphasis onleadership to the point where we neednot ask, ‘‘Is leadership important?’’but instead, ‘‘Is leadership the ONLYthing that is important?’’ Assessing‘‘leadership’’ is one of our major pri-orities in executive selection.

• Relationships. Whether called emo-tional intelligence, social intelligence,or simply interpersonal relationships,leadership and executive positions aretoday seen as relational. The inabilityto manage relationships and the ‘‘topteam’’ is often seen as the primarycause of executive failure.

• Derailers—we seek not only ‘‘successfactors’’ in our executives but are alertto derailment factors that may causethem to fail. Books are written, testsare available, and many companiesdevelop their own lists of derailers.

• Global capabilities—the world maybe ‘‘flat’’ or ‘‘curved,’’ but organiza-tions have become global and manyexecutives must perform on a globalstage. Organizations include globalsuccess factors in their selection aswell as development activities.

By any standard, this is an impressivelist. My apologies for omissions; readershopefully will add their own favorite‘‘rights.’’ Clearly there is no right ‘‘rights’’list! Even with an incomplete list, however,we must conclude that I–O psychology asa field is doing a great many things better.

We have much to be proud of. There is,however, another side.

What’s Wrong

Despite all we have learned and theprogress made, we are faced with thefollowing two very uncomfortable truths:(a) the results of executive selection are nobetter than ever—and that is dismal and(b) I–O psychologists—with few excep-tions—don’t have a place at the table whenexecutive selections are made.

Executive failure is common. First, the con-sistently dismal results. To my knowledgethere are few, if any, actual data on therate of executive failure over the years. Thecase here is made from the events of the dayand from the judgments of our distinguishedcolleagues.

High profile executive failures are inthe news with distressing regularity, fromEnron’s executive team to the rash of recentfirings in financial services (e.g., O’Nealat Merrill Lynch, Prince at Citicorp, Muddat Fannie Mae, Cayne at Bear Stearns).The finance industry may be the onemost in today’s news, but on the insidepages we find that no industry is immune(automobile, advertising, fashion, computerhardware and services, to name a fewreported recently) nor is the phenomenonan ‘‘American’’ disease.

Are these very visible executive failuresexceptions to the rule or is failure morewidespread? Our esteemed colleagues thinkthe problem is much broader. To quote afew over the years:

Drucker, writing 20+ years ago (1985),was quite critical of organizations’ successat executive selection: ‘‘their batting aver-age is no better than. 333. At most one-thirdof such decisions turn out right; one thirdare minimally effective; and one-third areoutright failures. In no other area of manage-ment would we put up with such miserableperformance’’ (p. 22). Sorcher (1985) writ-ing about the same time reached a similarconclusion.

134 G.P. Hollenbeck

Continuing the theme in the 1990s,Nadler and Nadler (1998) wrote: ‘‘In theirmost introspective moments, most execu-tives readily acknowledge that selecting theright people for the right jobs constitutes oneof their most important responsibilities. Fewdecisions they make will have as direct animpact on every facet of the organization.Yet, few other decisions are made in suchan illogical, slipshod manner’’ (p. 229).

Hogan confidently proclaimed in 2003‘‘we know for a fact . . . that perhaps two-thirds of the people currently in leadershippositions in the Western world will fail; theywill be fired, demoted, or kicked upstairs’’(p. xv).

More recently Hogan, Hogan, andKaiser (in press) gathered a dozen ofthese estimates of executive failure andconcluded ‘‘published estimates of the baserate of managerial failure . . . range from 30to 67 percent, with an average of about50 percent. . . . Based on the data, wesuggest that two thirds of existing managersare insufferable and that half will eventuallyfail’’ (p. 3).

Executives themselves are perhaps moresanguine about their ability to selectexecutives, but even the best of themconfess to a less than perfect record. Jackand Suzy Welch (2007) recently wrote:‘‘Look, hiring great people is brutally hard.New managers are lucky to get it right halfthe time. And even executives with decadesof experience will tell you that they makethe right calls 75% of the time at best’’(p. 102).

Whether widespread executive failure isa ‘‘fact’’ or not may be debatable, butsuch is the widespread ‘‘common wisdom.’’And, the high failure rate reflects selectionfailure. That argument is supported by myown experience that behind most businessfailures is a predictable executive failureas well as by the more extensive andsystematic studies of business and executivefailure (e.g., Finkelstein, 2003).

Arguable, perhaps, executive failuresare NOT the result of bad selections butinstead the inevitable workings of bad

luck, inadequate development, conspira-cies of organizational life, or ‘‘black swan’’changes in the external environment. High-house (2008) even argues that the larger partof employee (and presumably executive)success is simply largely unpredictable andthat to expect more is unrealistic. Othersmay hold that the high failure rate reflectsthe fact that executive jobs have become socomplex that success is hardly an option.We would counter that had the right peo-ple been chosen, the executives and theirorganizations would not have failed.

I–O psychologists have limited involve-ment in executive selection. Our seconduncomfortable truth is that when executiveselections are made, I–O psychologists arelargely absent from the table. Especially atthe senior levels, the task of selecting exec-utives is reserved for executive recruitersand more senior executives. This perhapsobvious conclusion was reached indepen-dently in two recent reviews (Howard,2001; Thornton, et al., in press).

Individual psychological assessment isperhaps our largest inroad into the area,but we have always had an uncomfortablerelationship with individual assessment (seethe historical review by Highhouse, 2002),and assessments are used with only asmall fraction of the executives selected.What many would consider our bestassessment tool, the assessment center,is used even more sparingly and moreoften for development than selection, andmore often at the high potential thanthe executive level. In general, assessmentcenters don’t fit the realities of executiveselection described by executive recruiters(Fernandez-Araoz, 2005).

Instead, executive search firms driveselection at senior levels and seem at leastto write most of the books and articles(e.g., Fernandez-Araoz, 2007; Wackerle,2001). Some search firms use assessmentby psychologists as input to their search,but more typical is the view expressedby Fernandez-Araoz (2007), himself anexecutive of Egon Zehnder International,a large international search firm: ‘‘after

Executive selection 135

conducting my first comprehensive reviewof all the relevant research that was outthere, I came to the reluctant conclusionthat most academics were missing thepoint. They were falling into the trapof producing statistically significant butmanagerially irrelevant findings’’ (p. 210).

We take as further evidence of our lackof influence that most books written byI–O psychologists that specifically addressexecutive selection don’t seem to gaintraction. How many of us are awareof Sorcher (1985)? Hollenbeck (1994)quickly went out of print. Sessa andTaylor (2000), one of the few researchstudies of how executive selection takesplace, was recently numbered 883,617on Amazon.com’s best seller list, withused copies selling for $.01! These lasttwo were part of an effort by the Centerfor Creative Leadership (CCL) to studyexecutive selection, but, other than asimulation used in the Leadership at thePeak program, the effort apparently didnot survive the death of its author, RichardJ. Campbell.

The net result is that a critical areaof organization success proceeds largelywithout us. This conclusion, if not correct,is certainly arguable and should stimulateinterest. It begs for the answers to thequestions: What’s wrong? Why do we haveso little influence and get no respect?

Getting it backwards. What’s wrong?What’s wrong is that we are doing itwrong—we are using the wrong models.Before explaining that conclusion, here is aword on how we arrived at it.

Sometimes something not exactly ‘‘ontopic’’ is the stimulus for crystallizing one’sthoughts on a long-simmering issue. Inthis case, the stimulus turned out to beJim Collins’ Good to Great (2001). I hadset out to explore ‘‘Level V’’ leadership,thinking that I might be the only personalive who had not read Good to Great —itturns out that I wasn’t—in September 2008,7 years after publication, Good to Greatis #38 on Amazon.com’s list of ‘‘BestSellers in Books,’’ and #1 in their Strategy

and Competition, Systems and Planning,and Leadership areas. For the uninitiated,Collins and his team selected on the basis offinancial criteria companies who had madethe transition from ‘‘good’’ to ‘‘great.’’ Theymatched those companies with those whohad not made the transition, or who hadmade it but didn’t sustain it. Then theysought what differentiated the sets. Level Vleaders were some of the differentiators. Ifyou are in the talent management world, Ihighly recommend the first three chaptersof this book.

One of the findings that surprised Collinswas that getting their strategies in place wasNOT what good-to-great companies didfirst; FIRST, they got the right people. Withthe right people ‘‘on the bus’’ (to use Collins’metaphor), you can, if needed, changedirections a mile down the road withoutchanging people; with the right people, theissues of motivation, compensation, andso on, take care of themselves, and, ifyou have the wrong people your strategywon’t matter—they can’t get you thereanyway. Collins’ arguments are simple andconvincing.

The question then becomes ‘‘who arethe right people?’’ and ‘‘what is ‘right?’ ’’Collins answers it in this way: ‘‘thegood-to-great companies placed greaterweight on character attributes than onspecific educational background, practicalskills, specialized knowledge, or workexperience.’’ It’s not that these otherattributes aren’t important but that Level Vleaders had recognized the key importanceof placing character first.

It was at this point that my thoughtscame together—eureka! We are doing itwrong! We are doing it backwards! Here isour well-accepted selection model; we baseour selection, in order of importance, on:

1. Competencies . . . behaviors . . . whata person does.

2. Competence . . . results . . . what aperson gets done.

3. Character . . . character-istics . . . whoa person is.

136 G.P. Hollenbeck

Here is what our model should be:

1. Character

2. Competence

3. Competencies.

These three Cs deserve some discussion.Competencies have become the dom-

inant HR model for selection and devel-opment systems; rare is the company thatdoesn’t have a behavioral competencymodel, either unique to the company ora generic one.

Defining ‘‘competencies’’ is not sosimple. If ever there was a Humpty Dumptyword, meaning exactly what one wants itto mean, it is ‘‘competencies.’’ Wikipediadefines competency as ‘‘a standardizedrequirement for an individual to properlyperform a specific job.’’ But lest onethinks that settles the issue, the Wikipedia‘‘discussion’’ begins with the comment‘‘This entry is totally off beam.’’

A history of how we got where weare today in the world of competenciesis outside the scope of this article. Clearlyimportant influences were the 1960s trendstoward behavioral objectives and dimen-sions as well as David McClelland’s workon competency modeling (e.g., McClelland,1973). Howard (2001) describes differentcompetency modeling approaches: In addi-tion to behavioral and clinical job-drivencompetency models and strategy-drivenapproaches, Howard describes ‘‘a role-based approach . . . (that) provides morebreadth than job-driven analyses withoutnecessarily sacrificing rigor. . . . Each rolelinks to competencies that specify how toeffectively enact that role’’ (italics ours).

For most in the I–O and human resourcesarena, the behavioral objectives movementwon out over the clinical-driven or traitmodels. Our reading of the key elementsof competencies today is that they are‘‘behavioral,’’ objective, and that they serveas the basis for entire human resourcesystems, including selection from the shopfloor to the executive suite. Thornton andByham’s (1982) definition of dimension

seems to fit competencies as well: ‘‘[W]emean a cluster of behaviors that are specific,observable, and verifiable, and that can bereliably and logically classified together’’(p. 117).

Competence is quite simply what itsays—what one can do, demonstrated bywhat one has done. It is exemplified bysimple questions: ‘‘Can she run a marketingorganization?’’ ‘‘Can he handle a P andL?’’ ‘‘Can she motivate a sales force?’’Combined with character-istics, this is howexecutives are typically evaluated, judged,and selected by other executives. Executivesanswer these questions by examining,however ineffectively, what a person hasdone. Only after affirming that an executiveactually gets done what needs to get done,do they ask ‘‘How does she go aboutit?’’ McCall and Hollenbeck (2007) givespecific examples of the language andconcepts that executives actually use fromtheir observation of executives making thesedecisions and also contrast competence andcompetencies in detail.

Character turns out to be a considerablylarger can of worms, with a long history, alarge literature, and a wide range of defi-nitions. A recent book on character (Klann,2007) states ‘‘Traditionally character hasbeen defined as ‘the combination of emo-tional, intellectual, and moral qualities thatdistinguishes a person,’ and that the word isderived from the Greek Kharassein meaningto engrave or inscribe.’’ Guinness (1999),in discussing the origin of the word char-acter, also makes the Greek reference—‘‘ahall mark or distinguishing sign’’ that distin-guishes one person from another; the innerform that makes a person or thing what it is(pp. 12–13). Without giving an explicit def-inition, Collins goes on to write that Level Vleaders use dimensions such as ‘‘character,work ethic, basic intelligence, dedicationto fulfilling commitments, and values (that). . . are more ingrained.’’

Collins was hardly the first to emphasizecharacter in selecting leaders (we couldgo back to Plato). Much to our surprise,Thornton and Byham’s (1982) description ofthe 1920–1930s German officer assessment

Executive selection 137

process includes character as one of thecharacteristics assessed and defines it as‘‘integrity, selflessness, idealism’’ (p. 24).And, Abraham Zaleznik (HBS AlumniBulletin, 2008) describes his primary careercontribution as emphasizing character inleadership: ‘‘It’s character that counts . . . Idon’t think it’s useful for people to worryabout the rules of leadership or command.I think the central concept is character. . .’’. This is a theme echoed by Tichy andBennis (2007) in the exercise of judgmentthat character means doing the right thing,putting the greater good first, honesty,integrity, having values, a moral compass. . . knowing what you will do and whatyou won’t.

Definitions of character turn out tohave both nonevaluative elements (e.g.,judgment, work ethic) as well as moralisticones. Guinness (1999) points out that theGreek usage focused on the ‘‘stamped’’quality—as in stamping coins with theirlasting value—and did not carry the‘‘moral’’ connotation; the Hebrews addedthe moral element, the ‘‘righteous’’ tone.Today’s usage almost always includesthe moral element and sometimes thenonevaluative element.

As a point of interest, the psychoana-lytically inclined seem to have been morecomfortable with the notion of executivecharacter, appearing at least to producethe lion’s share of writing and thinking(e.g., see Levinson, 1994; Sperry, 1999).Sperry writes: ‘‘My view is that charac-ter is the dimension of personality thatdescribes how individuals conduct them-selves in interpersonal and organizationalsituations. . . . Generally speaking an indi-vidual with a good character or reputationis considered to be responsible, trustworthy,respectful, fair, caring, and a good citizen,meaning being cooperative and playing bythe rules’’(p. 212).

For the purpose of this article, I will gowith Collins’ view that defines characteras those more enduring (if not ingrained)qualities that enable a person to contributeto organizations. This view embraces boththe nonmoral or Greek elements (judgment,

work ethic, etc.) as well as the moralor Hebrew elements of honesty, integrity,sacrificing self interest for the greater good,and treating people with respect.

Many people writing and thinking aboutcharacter are unlikely to jump on thisparticular definitional bandwagon, but weshould be able to arrive at a workingdefinition that will at least serve as a startingpoint. A review of character from thisperspective would be a real contribution.

So, what’s wrong with doing it backwards?Why doesn’t it work? My quarrel here is notwith competencies or with competences;they are obviously useful starting pointsfor much training and development, andI do not advocate substituting character-istics for either of them. I argue with theimportance we assign to them. However, asCollins quite eloquently makes the case, inselecting their executives, great companiesstart with ‘‘who.’’

Why doesn’t starting with behavioralcompetencies work in selecting executives?Because executive jobs, especially themega jobs of today in complex, globalorganizations, are very very different thanjobs at the bottom of the hierarchy. And withthose differences, we fail on two counts.

First, as we move up the hierarchy, wego from jobs in which behavior and resultsare tightly coupled to jobs where there isa very loose coupling between behaviorand results. In lower level jobs, there oftenis a ‘‘right way’’ to do the job—scientificmanagement worked, and it worked for areason. There may well be a ‘‘right’’ wayto operate the widget machine (and even aright way to land an airplane!). The ‘‘right’’behaviors are closely tied to the results onegets.

At the top, however, and in complexorganizations, we cannot ‘‘specify howto effectively enact that role’’ (to giveHoward’s definition of competency adifferent emphasis). The coupling changes,becoming increasingly loose, so that atthe top levels executive job behaviors arevery loosely coupled to results. There is noprescription of behaviors that will succeed

138 G.P. Hollenbeck

or fail —there are as many combinationsand nuances as there are people; successfulexecutives come in all shapes and sizes andbehave (within limits of tolerable deviance)in as many ways as there are people. This iswhat equifinality is about—the same resultmay come from many different behaviors.Executives (like golf swings) come in asmany variations and do their work in asmany ways as there are executives.

To make matters worse, there is alsoa loose coupling at the executive levelbetween results and organization perfor-mance. This, by the way, makes itself feltin our much-discussed ‘‘criterion problem’’and is a source of Highhouse’s (2008) frus-tration that performance at any level (but,we might add, especially at the executivelevel) is difficult to predict.

Intriguingly, this intuitively obvious factwas as evident in the 1960s as it is today.In a world that we suspect was muchless loosely coupled, Campbell et al. (1970)wrote ‘‘The means for accomplishing man-agerial job objectives are rarely specifiedin detail. A particular manager may changethe way a job is done in order to capi-talize on his own propensities, strengths,weaknesses, and stylized patterns of behav-ing’’ (p. 13). Despite their at least beginningawareness of the individuality of execu-tive jobs, Campbell et al. continued to seek‘‘those kinds of job behaviors resulting ineffective managing.’’

Perhaps, the nail in the coffin of equifi-nality was the classic paper by Wernimontand Campbell (1968). Beautifully writtenand argued, the paper’s point is that ourselections will be more successful if weuse behaviors, rather than traits, to pre-dict behaviors. Of course, they were rightat lower levels; however, at the executivelevels, we are trying to predict behaviorsonly after character and competence, and,even if we put behaviors first, the practical-ities of executive selection preclude muchbehavior-to-behavior prediction that maybe best accomplished through simulation.

As an afternote, in our increasinglyknowledge-based workforce, this loose/tightcoupling distinction between executive/

managerial and lower level jobs is diminish-ing, arguing for a who/what/how throughoutthe organization.

And using the same old model. In additionto doing it backwards, our field hasstuck with our classical personnel selectionmodel, seeking to correlate predictors withcriteria, hoping for large samples. We dothis even though it continues to disappointus in terms of research, results, or respect.Our model is so thoroughly ingrained in allof us that we keep hoping that the right setof predictors and criteria and a large enoughN will (finally) produce the results we seek.

Intriguingly, here also it wasn’t supposedto be that way. Campbell et al. (1970)wrote: ‘‘much research and many com-pany practices imply the stripping awayof managerial individuality . . . statisticalequations are being used for ‘managers ingeneral.’ In contrast, decisions about man-agers and managers, jobs are individualdecisions. . . . All this points to the needfor a truly individualized approach to thetask of managing managerial effectiveness’’(pp. 3–4, italics theirs).

Going on to declare that this individual-ized approach is ‘‘our fundamental point ofview,’’ they then proceed to write the classicreference in manager/executive selectionusing our traditional psychometric, statis-tical, large sample point of view. Theirhope seemed to be that the future woulddefine individuals, organizations, and envi-ronments so thoroughly that statistical pre-dictions for individuals would be possible.In fact, their solution to the problem thatmanagerial jobs differ so much from eachother is to include a larger range of jobs inour job analysis!

This ‘‘by-and-by-we-will-understand’’optimism must be understood in thecontext of the times. Many of us in graduateschool in the 1950s and 1960s can recallwhen psychology was one field, AmericanPsychological Association (APA) had oneconference, and professors professed thatsomeday there would be a unifying theory

Executive selection 139

of human behavior bringing together every-thing from eye-blink conditioning to testsand measurements.

Guion (1965) in his classic PersonnelTesting, written about the same time, rec-ognized the same issues, even though hissolution was slightly different and his wasa longer-term optimism: ‘‘Statistical valida-tions of test batteries for higher managerialpersonnel will, because of the uniquenessof managerial jobs, have to use a sys-tem of synthetic validities. Such a systemwill not be available until (1) job situa-tion variables, (2) organizational variables,and (3) performance variables are clearlyidentified and measurable’’ (p. 474). Littledid Guion know how right he was whenhe went on to say that solutions to theseresearch problems were ‘‘decades away.’’

The point is that historically I–O hasbeen well aware that our standard psycho-metric models won’t serve us in makingindividual executive selection decisions.And yet, we persist in criticizing the onlyindividualized approaches available (High-house, 2002).

Is one model enough? We have pro-ceeded as if ‘‘executive selection’’ isa single process, one set of principles,applied to a like sample of candidates.Although the argument here took as itsstarting point Collins’ notion of good-to-great, obviously every organization couldnot and should not aspire to greatness.Many executive selections won’t followthe Character/Competence/Competenciesmodel. After all, one of the things we aredoing right is attending to a variety of hiringstrategies—different organization demandswill require different approaches.

Organizations may have different hiringstrategies as they progress through differentstages and also in different parts ofthe organization. Fidelity Investments, forexample, historically selected for characterin its core competence—mutual fundmanagement; it hired at the entry level witha rigorous program of selection followedby learning to ‘‘run money.’’ For itsother executive positions, Fidelity was well

known to select for competence (what aperson could do in the short run), payingwhatever it took to get the right person; itmanaged with a philosophy of ‘‘churn ’emand burn ’em,’’ keeping the few who turnedout to have Fidelity defined ‘‘character’’ asproven in what one recruiter described asa ‘‘Darwinian’’ culture. Fidelity, as a result,has excelled at producing fund managersbut from all appearances has never beenable to develop a corps of outstandingexecutives (the who rather than the what)to carry it past its patriarch Ned Johnson’snot so distant (at age 78 years) retirement.

Companies such as J&J, General Electric(GE), and Procter & Gamble almost neverhire executives from outside. As good-to-great organizations, their selections at everylevel place heavy emphasis on character,with an increasing emphasis upon com-petence and competencies as executivesmature. Organizations focused on short-term turnarounds with intense immediatepressures for results are likely, and rightfullyso, to focus on competence—can the exec-utive do the job required now. Charactercan wait until later.

These examples illustrate that no singleapproach to executive selection is uni-versally appropriate, and they invite usto explore different solutions for differ-ent situations. My argument is that ouradherence to doing it backwards (compe-tencies/competency/character) has kept uson the sidelines and prevented our learn-ing from what we need to learn to be TierI players.

Not having a place at the table, perhapsit is not surprising that I–O as a field iswoefully uninformed about how executiveselection actually takes place. Certainly,there are those few colleagues who are verywell informed and participate in the pro-cess, but they are few and far between. Withso little knowledge about how selectionsactually take place, although maintainingthat our statistical validation models apply,it is again not surprising that we do notmake the contribution we feel we deserve.Fernandez-Araoz (2005), cited previously,offers a very realistic and very readable

140 G.P. Hollenbeck

description by a clearly knowledgeableinsider of the constraints surrounding exec-utive selections.

‘‘So Now What?’’? Getting Back inthe Game

That we get no respect and don’t have aseat at the table are symptoms, not thecause of our problems. Nobody is buyingbecause we don’t have much to sell. Thisis not a marketing problem; our traditionalpsychometric, selection model has led usdown the wrong paths.

Can we get back on the right track? Ibrim with optimism—the analytic abilityand problem focus of I–O psychologistsmake us very desirable commodities inboth industry and academia. Here are someroads to relevance:1. Change our mindset. We must viewselecting executives not as a selectionproblem but as a problem in decisionmaking and judgment. This is our client’spoint of view, let’s adopt it. Taking this viewtakes us out of our old thinking boxes andfrees us to examine the extensive literatureon how judgments are made, to examinethat in ourselves, and to find what wecan offer to executives who make thosejudgments. Cronbach and Gleser (1957)started us in that direction 50 years ago,but somehow we abandoned their line.Colarelli and Thompson (2008) hint at thatview in suggesting why we continue touse our intuition in the face of constantcriticism. As they point out, we don’t useour selection models for making decisionsthat are important to us personally—whyshould executives?

Tichy and Bennis (2007), neither notablefor their published work in selection, havebegun exploring judgment as the foundationof good leadership, dividing judgment intopeople judgments, strategy judgments, andcrisis judgments. Their discussion of ‘‘howto’’ and ‘‘how do’’ executives make peoplejudgments (yes, based on character andcompetence) may raise as many questionsas it answers, but it is a start toward whatwe need to do.

2. Reframing the problem from selectionto judgment leads us in very differentdirections. We begin to ask ‘‘what do weneed to know to make this judgment andhow should we make it?’’ Here are some ofthe questions:

What do we need to know about the job?

What do we need to know about theorganization? The external environment?

What do we need to know about theperson?

How can we find out what we need toknow?

How do we arrive at a decision, once wehave the information?

When viewed from the decision maker’spoint of view, we begin to ask also ‘‘whatdoes the executive making this judgmentwant to know?’’ (This in itself would be aninteresting study. I was tempted to contacta few executives for an answer but decidedthat fell outside the scope of this article.)

Given the importance assigned to char-acter, assessing it deserves at least someconsideration. I have taken a wide angleapproach to character, defining it to includeboth the nonevaluative and evaluative(moral) aspects. Viewed this way, charac-ter is a market basket of relatively stableabilities and characteristics. Some of thesewe should be able to assess using alreadyavailable tools; some may require new toolsor sharpening the ones we have.

Our problem may be more that our focushas been in the wrong order and on thewrong things rather than that I–O is bereftof useful tools. Mahler and Wrightnour(1973) provided a systematic alternative (adescendant of which is now used at GE)to the interview self-narrative for assessingcompetence that could equally as well beused for character. Sorcher (1985) provideda framework for finding and using whatan organization knows about an executivecandidate that is useful in all the three Cs.Tichy and Bennis advocate examining anexecutive’s past judgments as windows on

Executive selection 141

character without offering specific tools.An underlying theme in each of theseis searching a candidate’s background fortelltale events that provide the evidencerequired to make judgments.

Our task is almost certainly easier forinternal than external candidates, but a formof this background analysis (references)is a primary tool of search professionals,whether done poorly or well. We havea history in biodata and in using workinventories for selection, but especiallyat the executive level we haven’t begunto refine the techniques. On the processside, Sessa and Taylor (2000) observe thatteams, if constituted properly, make betterjudgments than individuals.

At the high potential level, we may havemore latitude in the tools that can be usedand that in fact companies are alreadyusing. High potentials are usually internal,at lower levels and eager to demonstratetheir talents however possible via tests,simulations, assessment centers, whatever.We may have greater opportunities (andgreater success) free from organizationconstraints at the high-potential level, butthat does not mean we should despair at theexecutive level.

We may indeed know more than wethink we know and have more tools thanwe think.3. Begin to address the topic. Some mightsay ‘‘readdress’’ the topic, but at least froma research and analysis standpoint, wehave ignored executive selection. Thorntonet al. (in press) lament the lack of availableresearch to review. In their recent AnnualReview chapter on personnel selection,Sackett and Lievens (2008) do not eveninclude the topic. When queried about this,Sackett (personal communication) replied‘‘Our assigned task was to select and citethe best of the empirical literature sincethe last Annual Review chapter. There’snothing on executive selection because weweren’t aware of any recent empirical workon the topic to cite.’’

Clearly, there is much to be done,whether or not it would meet ‘‘A-level’’publication criteria. Fifteen or so years

ago, the CCL sponsored a conferenceon executive selection as a kickoff toan executive selection research program.Given CCL’s mission as a developmentorganization, perhaps it is not surprisingthat selection as an area of research didnot survive that effort. In addition to Sessaand Taylor’s (2000) book, one of theproducts that did survive was an executiveselection simulation that has been usedin its Leadership at the Peak program.Although CCL’s data from that simulationare available to researchers, the verycarefully developed simulation could offera ready-made research tool for exploringthe range of executive selection questionsaddressed here.

One study that begs to be done wouldaddress why executive selections fail. Inter-estingly enough, studies of why executiveshave failed abound, but only one studysurfaced that examined why the selectionfailed; that study by an executive searchconsultant, Fernandez-Araoz (2007) lookedinternally at their search firm results anddid not, to my mind, provide the breadthor rigor that we need. The causes of execu-tive failure are no doubt many, but studyingthe selection failure would be more fruit-ful than one more study of why executivesfail. If nothing else, we will learn a greatdeal about what works and doesn’t workand gain a perspective on the magnitudeof our task and promising directions forresearch. A recent example illustrates thesepossibilities.

Wackerle’s (2001) book on executiveselection is an excellent resource fromthe viewpoint of an executive searchprofessional as to how executive selectionshould be done and the issues involved.In Chapter 3, Wackerle describes his‘‘blue chip search’’ (pp. 23–40) in whicheverything went right—the right person waschosen by the right people at the right time,as he puts it. The person chosen to beCEO of MassMutual was Bob O’Connell,then at AIG. Reading this book in 2008,I was curious as to how O’Connell hadfared as the CEO at MassMutual. Googling‘‘MassMutual’’ produced this June 23,

142 G.P. Hollenbeck

2005, press release: ‘‘Massachusetts MutualLife Insurance Terminates Employment ofRobert O’Connell for Cause.’’ The pressrelease reports a list of ‘‘engaging ina systematic and pervasive pattern ofwillful abuse of authority, violations of theCompany’s code of conduct, and other actsof willful gross misconduct.’’ Intriguingly,all the seven specifics listed were actionstaken by O’Connell to enhance his personalfinances or to hide his actions. (Definitelya character issue here!) And intriguingly,the selection also apparently did involve anI–O psychologist.

Other than Wackerle’s book and thepress release, I have no knowledge of theMassMutual selection, but these facts cryout for an answer to the question, ‘‘Howdid they miss this?’’ The recent bailoutof AIG and a passing knowledge of theAIG executive culture both suggest thatO’Connell was an executive failure and aselection failure.

Can research find out where selectionfailures go wrong? Certainly there are obsta-cles to this research: There won’t always beclear lines between executive failure andselection failure; the succession process,especially at the CEO level, is fraught withsecrecy and political issues. Vancil (1987)with his remarkable access to executivesuites found that in any CEO selection thereis an elephant problem—there are as manyviews as there are viewers. However, afteraction reviews of selection failures shouldbe less subject to organizational pressuresthan entire succession process, and if not,the different viewpoints would be informa-tive.4. Critically examine the executive job.Every company and every consulting firmhas its competencies, but there is surpris-ingly little integrative research about execu-tive jobs today. ‘‘Research’’ for I–O shouldbe a big tent, from qualitative analyses toquantitative ones, from analyses such aspolitical scientists have done on leader-ship, to the meta-analyses of job analyses.One such qualitative study is Renshon’s(1998) study of how the job of Presidentof the United States has changed and the

attendant changes in the requirements ofthe job, and the needs for assessment isthe type of analysis we need of executivejobs in general. Renshon’s psychoanalyticapproach to assessing candidates may notbe our cup of tea, but his approach isinsightful.

Other unanswered questions include thefollowing: How indeed do jobs changeas we ascend the organization ladder? Atwhat point do results become more persondependent than job dependent? At whatpoint do our models lose their validity? Areexecutive jobs more alike or more differentand at what levels? Can we indeed developthe data to inform judgments with syntheticvalidity? These are some of the questionswe might address.5. Develop an integrative explanatory andpredictive model of the executive. Manyif not most of us in I–O would beuncomfortable using the psychoanalyticmodel as our primary model of theexecutive. However, as Renshon’s studyillustrates, that model does explain therelationship among variables and supportspredictions and judgments. It is time thatwe started putting together a model ofthe executive that brings together all welearn and enables us to make judgmentsand predictions across a wide range ofsituations. Admittedly we may be decadesaway from a useful approach, but we havemany reasons to be optimists that it can bedone.

Carrying out these suggestions facesmany obstacles. For practitioners familiarwith executive jobs (and we have many),there may be few rewards for research-ing the topic or for publishing integra-tive thinking. For academics, many mayknow little about executives and somework in an environment that does notreward highly applied research. Despite theobstacles, so important an area for orga-nization success will not go long withoutattention. There is indeed cause for opti-mism. I–O psychology has the character,competence, and competencies to con-tribute.

Executive selection 143

ReferencesAkst, D. (2008). Expensive mistakes. Wall Street

Journal, September 24, A27.Campbell, J. P., Dunnette, M. D., Lawler, E. E., III., &

Weick, K. E., Jr. (1970). Managerial behavior, per-formance, and effectiveness. New York: McGraw-Hill Inc.

Colarelli, S. M., & Thompson, M. (2008). Stubbornreliance on intuition and subjectivity in employeeselection: Statistical decision aids are evolutionarilynovel. Industrial and Organizational Psychology:Perspectives on Science and Practice, 1, 347–351.

Collins, J. (2001). Good to great. New York: Harper-Collins Publishers, Inc.

Cronbach, L. J., & Gleser, G. C. (1957). Psychologicaltests and personnel decisions. Urbana, IL: Univer-sity of Illinois Press.

Drucker, P. E. (1985). Getting things done: How tomake people decisions. Harvard Business Review,July-August, 22–26.

Dunnette, M. D. (1966). Personnel selection andplacement. Belmont, CA: Brooks Cole PublishingCo.

Farr, J., & Tippins, N. (Eds.) (in press). Handbook ofemployee selection. Clifton, NJ: Psychology Press.

Fernandez-Araoz, C. (2005). Getting the right peopleat the top. Sloan Management Review, A46(4),67–75.

Fernandez-Araoz, C. (2007). Great people decisions.Hoboken, NJ: John Wiley & Sons.

Finkelstein, S. (2003). Why smart executives fail. NewYork: Penguin Group.

Guinness, O. (Ed.). (1999). Introduction. Charactercounts. Grand Rapids, MI: Baker Books.

Guion, R. M. (1965). Personnel testing. New York:McGraw-Hill.

HBS Alumni Bulletin (March 2008). Where are theynow? Abraham Zaleznik. p. 4.

Highhouse, S. (2002). Assessing the candidate as awhole: A historical and critical analysis of individ-ual psychological assessment for personnel deci-sion making. Personnel Psychology, 55, 363–396.

Highhouse, S. (2008). Stubborn reliance on intuitionand subjectivity in employee selection. Industrialand Organizational Psychology: Perspectives onScience and Practice, 1, 333–342.

Hogan, R. (2003). Foreword. In D. L. Dotlich., & P. C.Cairo (Eds.), Why CEOs fail. San Francisco: Jossey-Bass.

Hogan, J., Hogan, R., & Kaiser, R. B. (in press). Man-agement derailment: Personality assessment andmitigation. In S. Zedeck (Ed.), (in press). AmericanPsychological Association handbook of industrialand organizational psychology. Washington, DC:American Psychological Association.

Hollenbeck, G. P. (1994). CEO selection: A street-smart review. Greensboro, NC: Center for CreativeLeadership.

Howard, A. (2001). Identify, assessing, and selectingsenior leaders. In S. J. Zaccaro, & R. Klimoski

(Eds.), The nature of organizational leadership. SanFrancisco: Jossey-Bass.

Klann, G. (2007). Building character. San Francisco:Jossey-Bass.

Levinson, H. (1994). Why behemoths fell: The psy-chological roots of corporate failure. AmericanPsychologist, 49, 428–436.

Mahler, W., & Wrightnour, W. (1973). Executivecontinuity: How to build and retain and effectivemanagement team. Homewood, IL: Dow Jones-Irwin.

McCall, M. W., Jr., & Hollenbeck, G. P. (2007). Get-ting leader development right: Competence notcompetencies (chap. 4). In J. A. Conger, & R. E.Riggio (Eds.), The practice of leadership. San Fran-cisco: Jossey-Bass.

McClelland, D. C. (1973). Testing for competencerather than for ‘‘intelligence.’’ American Psycholo-gist, 28, 1–14.

Nadler, D. A., & Nadler, M. B. (1998). Championsof change: How CEOs and their companiesare mastering the skills of radical change. SanFrancisco: Jossey-Bass.

Renshon, S. A. (1998). The psychological assessmentof presidential candidates. New York: Routledge.

Sessa, V. I., & Taylor, J. J. (2000). Executive selection:Strategies for success. San Francisco: Jossey-Bass.

Silzer, R. (2002). Preface. In R. Silzer (Ed.), The 21stcentury executive: Innovative practices for buildingleadership at the top. San Francisco: Jossey-Bass.

Sorcher, M. (1985). Predicting executive success. NewYork: John Wiley & Sons.

Sperry, L. (1999). The 1999 Harry Levinson Lecture.Leadership dynamics: Character assessment in theexecutive selection process. Consulting PsychologyJournal Practice and Research, 51, 211–217.

Thornton, G. T., III, & Byham, W. (1982). Assessmentcenters and managerial performance. New York:Academic Press.

Thornton, G. C., III, Hollenbeck, G. P., & Johnson,S. K. (in press). Selecting leaders: Executives andhigh potentials. In J. Farr & N. Tippins (Eds.), (inpress) Handbook of employee selection (chap. 38).Clifton, NJ: Psychology Press.

Tichy, N. M., & Bennis, W. G. (2007). Judgment: Howwinning leaders make great calls. New York: ThePenguin Group.

Vancil, R. F. (1987). Passing the baton: Managingthe process of CEO selection. Boston: HarvardBusiness School Press.

Wackerle, F. W. (2001). The right CEO: Straight talkabout making tough CEO selection decisions. SanFrancisco: Jossey-Bass.

Welch, J., & Welch, S. (2007). Hiring wrong—andright. Business Week, January 29, 102.

Wernimont, P. F., & Campbell, J. P. (1968). Signs,samples and criteria. Journal of AppliedPsychology, 52, 372–376.