Categories
Journal of Management

Celebrating the “essential:” the impact of performance on the functional favoritism of CEOs in two contexts

Celebrating the “essential:” the impact of performance on the functional favoritism of CEOs in two contexts – includes appendix

Danny Miller

This paper argues that an organization’s performance can influence the extent to which CEOs favor the importance and competence of one functional department above all others. Specifically, financial success induces such favoritism in CEOs, not in stable but in uncertain environments where there is a good deal of scope and motivation for attributional opportunism and superstitious learning. Financial weakness also induces functional favoritism on the part of CEOs, not in uncertain but in stable environments where conditions are fight for escalation of commitment and threat-rigidity responses. These findings were confirmed using subgroup regressions, moderated regressions and two-group LISREL analyses.

An organization’s performance may affect profoundly a CEO’s evaluations of his or her policies and departments (Miller & Porter, 1988; Milliken & Lant, 1991). Good performance may cause a leader to attribute special merit to focal tasks and skills, often those tied to the executive’s reputation, objectives or political interests (Miller, 1993). This attribution may also cause CEOs to elevate the relative status of the functional department within which such tasks and skills are said to reside (Miller, 1990). Such “favoritism” draws resources, attention and prestige towards one function and away from all others.

Certainly the concentration on a special skill has been much applauded by recent resource-based views of strategy (Wernerfelt, 1984; Barney, 1991). And often a company’s most valuable skill is to be found within a single department (Pascale, 1989). Scholars of organizational decline, however, have pointed to the long-run hazards of such concentration. Finns’ attentions and skills may become too obsessively focused to address important market contingencies (Miller, 1990, 1994; Wilensky, 1971; Walsh, 1995). Given its potential implications, therefore, it is unfortunate that we know so little about the sources of functional favoritism. This study will explore one of those sources–the past performance of an organization.

Preliminary Definitions

Functional Favoritism may be defined as the attribution, in this case by CEOs, of exceptional relative importance and competence to a functional department–marketing or R&D, for example. We will address two aspects of functional favoritism: (1) the relative importance ascribed to the tasks of one department vis-a-vis all others; and (2) the relative competence ascribed to a department, again vs. others in the firm, in carrying out those tasks. Our arguments and hypotheses are expected to apply to both aspects of favoritism but will be tested separately.

Overview: Performance and Favoritism in Two Environments

Miller (1990, 1993, 1994) suggests that high performance can induce functional favoritism in CEOs. He believes that powerful parties in thriving organizations tend to attribute overriding importance and virtuosity to only one or two functions–those that implement their vision, help them politically, or excel in achieving a prized objective. Miller (1990, 1994) suggests that this is done opportunistically by top executives to affirm their visions of the organization and to enhance their reputations. Favoritism also may be elicited by superstitious learning in which successes are used to amplify pre-existing biases (Miller & Chen, 1993, 1994, 1996; Milliken & Lant, 1991).

The literature on how organizations react to problems suggests, however, that failure too may cause favoritism. But now the driving force is more defensive than opportunistic. Staw (1981) and his followers maintain that failure causes managers to escalate their commitment to a course of action–to a prized task or activity (Brockner, 1992; Staw & Ross, 1978). In the attempt to vindicate their past policies and protect their reputations, some leaders react to failure by attending more assiduously and defending more avidly their favorite activities and functions (Halberstam, 1986, pp. 377-390).

The thesis of this research is that both success AND failure can induce favoritism, the former in uncertain contexts, the latter in stable and predictable ones. Specifically, the learning and attributional processes that link success to favoritism will be argued to occur mostly in uncertain rather than stable contexts (Lant & Montgomery, 1987; Miller, 1990; 1993; Milliken & Lant, 1991). By contrast, the defensive escalation and threat-rigidity responses that link failure to favoritism will be argued to occur mostly in stable, rather than uncertain, contexts (Brockner, 1992; Staw, 1981; Staw, Sandelands & Dutton, 1981).

Favoritism under Success–The Uncertain Context

In successful companies, opportunism and superstitious learning may induce top managers to attribute overriding importance and competence to particular functional areas and their activities. We will argue that this is most likely to take place in uncertain environments where attributional ambiguities and substantial managerial discretion provide both the scope and the motive for success-based favoritism.

Opportunistic Attributions

Success presents top managers with an opportunity to enhance their reputations by taking credit for good performance (Salancik & Meindl, 1984: Staw, McKechnie & Puffer, 1983). One way that CEOs can call attention to the value of their contributions is by allying or identifying themselves with, and then praising, particular functional skills and units (Halberstam, 1986; Miller, 1990; Pascale, 1989). For example, if the CEO has glorified marketing. bet his reputation on that department, or hails from there, he may boost his image by ascribing credit to the marketing function (Dearborn & Simon, 1958). In contexts of average performance, executives will gain less by making such attributions (Miller & Porter, 1988; Salancik & Meindl, 1984; Staw et al., 1993).

We expect that such opportunism born of success will be more likely in uncertain than in stable environments. Environmental uncertainty may be defined as the degree of dynamism and unpredictability of competitor and customer behavior and operating methods (Duncan, 1972; Khandwalla, 1977; Dess & Beard, 1984). In order for managers to be motivated to take credit for positive outcomes, they must have had some real or perceived discretion to act (Hambrick & Finkelstein, 1987; Kiesler & Sproull, 1982; Staw et al., 1983). In stable environments where products, markets and methods are uniform and unchanging, it is often difficult for top executives to take or to feel much credit for good results. By contrast, in changing, uncertain environments, managers must work harder to perform well. Here the top echelon can more credibly celebrate success by highligting the unique contributions of the function that implements their vision (Hambrick & Finkelstein, 1987; Miller, 1990).

Opportunistic or self-flattering interpretations also work best when there is ambiguity in the relationship between actions and outcomes (Walsh, 1995). When it is not entirely obvious just who or what is responsible for results, self-serving attributions that celebrate a function disproportionately are more credible. And, as we have argued, in uncertain contexts the sources of success may well be ambiguous and multifaceted (Walsh, 1995). Here there is scope for opportunistic interpretations and thus a strong link between past success and functional favoritism. In more stable and predictable contexts, however, there is less ambiguity regarding sources of success, and hence the scope for opportunistic attributions is more limited (Miller & Porter, 1988).

Superstitious Learning

Success reinforces the notion of many CEOs that their firms possess a special skill or advantage vis-a-vis their competitors (March, 1991; Miller. 1990, 1994). Leaders, often wrongly, view good performance as a confirmation of their recipes (Milliken & Lant, 1991), recipes that tend to feature the contributions and skills of a prized department (Wright, 1979). Levitt & March (1988) call such inferences “superstitious learning.” The enhanced confidence that leaders have in their recipes increases the importance and merit they ascribe to their favorite functions, and it decreases their comparative regard for other functions (Miller, 1993). Leaders of firms with poorer performance, by contrast, are less likely to be so confident.

We expect that superstitious learning–and again, therefore, success-based favoritism–will occur in uncertain rather than stable settings. Uncertain environments demand attention, present equivocal cues, and display intertemporal, variety (Levitt & March, 1988; Weick, 1979). They generate a complex stream of challenges that are ever-changing and tough to decipher. Such complexity and ambiguity promote superstitious interpretations that play up the importance and achievements of prized functions (Miller, 1993). Since uncertainty makes it impossible for anyone to know exactly the sources of success, administrators have the latitude to ascribe felicitous flux and happy coincidence to the excellence of their pet departments (Levitt & March, 1988). The causes of success in stable contexts, on the other hand, are unlikely to be so ambiguous, and hence superstitious attributions will be that much less credible (see Figure 1).

[Figure 1 ILLUSTRATION OMITTED]

H1: Functional favoritism will follow success in uncertain

environments.

Favoritism under Failure–The Stable Context

Since success was argued to promote favoritism, it will now appear paradoxical that failure can have a similar result. But we expect the latter relationship to hold in stable, not uncertain environments, and for different reasons. Companies operating in stable contexts are enured to constancy and stability. In such environments, failure may provoke defensive “escalation of commitment” and threat-rigidity reactions as executives, fearful of change, strive to protect their self-images, their positions, and their reputations (Brockner, 1992: Salancik & Meindl, 1984; Staw, 1981). These reactions are less apt to occur, however, in dynamic, uncertain contexts, where critical re-evaluations are common and failure motivates search and adaptation (Burgelman, 1994; Eisenhardt, 1989).

Escalation of Commitment

By challenging the competence of top executives, corporate failure can be damaging politically and threatening psychologically. Thus, to vindicate themselves leaders may escalate their commitment to an established competence or course of action (Brockner, 1992, Staw, 1976; 1981; Staw & Ross, 1987). They may also try to salvage reputations and cherished programs by championing pet competences and departments (Wright, 1979). The escalation argument suggests that any CEO who has openly touted or lavishly funded a particular department and its activities would especially champion that team or effort when the firm Is doing poorly. Indeed, executives may celebrate what they view to be the organization’s most essential actors and skills, in part to deny failure, and in part to dispel pessimism and discourage rebellion (Meyer & Starbuck, 1991; Milliken & Lant, 1991; Salancik & Meindl, 1984). By emphasizing the relative importance and achievements of special units, CEOs are able to shore up myths of organizational competency and to reaffirm the soundness of current priorities (Miller, 1990).

The conditions for such commitment responses are most likely to be found in stable, not uncertain contexts. Staw (1981) has argued that escalation of commitment in the face of failure is most common when there are powerful norms of consistency–where too much change on the part of managers is believed to signify weakness and vacillation. Such norms are likely to be stronger in stable environments than in uncertain settings where change and adaptation are viewed as both normal and necessary (Bums & Stalker, 1961; Nystrom & Starbuck, 1984; Staw & Ross, 1987). Indeed, in uncertain environments, failure will induce many managers to re-examine organizational priorities and skills, and call into question established practices. CEOs operating in such eternally changing settings will be less afraid to re-evaluate and broaden their priorities and skills in the face of failure. This, in fact, is a very common way of coping with challenges in a turbulent environment (Burgelman, 1994; Eisenhardt, 1989).

Staw (1981) also contends that managers are especially apt to escalate commitment when they are confident that they can predict the outcome. This again is most likely to hold in stable settings where priorities and policies are “tried and true,” and where the context is believed to be well understood. What worked, for so long in the past, is expected to work in the future. And under the shadow of poor performance, top managers in stable settings will be especially decisive in defending their priorities. Uncertain environments, by contrast, discourage the notion that the future will be like the past (Burgelman, 1994).

Threat-Rigidity Effects

“Threat-rigidity effects” too may cause managers to react to failure by favoring a familiar function and its activities and by de-emphasizing the importance of less familiar and less central tasks and issues (Staw, Sandelands & Dutton, 1981). Under threat, managers tend to resort more exclusively to their dominant responses and beliefs. They resist change, eschew refinement. nuance and novelty, and concentrate on the essential and familiar. This is a common reaction to threat in many walks of life (D’Aveni & MacMillan, 1990: Dutton & Jackson, 1987; Fox & Staw, 1979; Holsti, 1969).

Such threat-rigidity effects are likely to occur in stable, not uncertain environments. Stable contexts foster longstanding, inertial organizational priorities and competency myths (Starbuck & Milliken, 1988). The more established these myths, the more they encourage threat-rigidity effects. Venerable priorities and entrenched beliefs are defended zealously in the face of failure because they are so deeply and widely held and because they are embedded in the political fabric of the organization (Pfeffer, 1983). The resources and prestige of top managers come to be tied to these myths, which thus become avidly celebrated when reputations are threatened by failure (Kiesler & Sproull, 1982: Meyer & Starbuck, 1991). Functional favoritism is one form of such celebration. In dynamic environments, on the other hand, change will be more common, and so competency myths will be less entrenched and less politically useful (Burgelman, 1994; Eisenhardt, 1989).

Stable environments and the complacency and inertia they foster also prevent managers from learning about their mistakes and their markets. Hence. administrators try to grapple with problems by concentrating on the familiar (Starbuck, Greve & Hedberg, 1978). The insularity produced by stable environments may even promote wishful thinking–a belief that problems are temporary or that more resources applied in the same direction will achieve objectives (Hambrick & D’Aveni, 1988). By contrast, the shocks administered by dynamic environments generally discourage such hyper-optimism (Burgelman, 1994, see Figure 1).

H2: Functional favoritism will follow failure in stable environments.

Given that H1 proposes that favoritism will follow success in uncertain environments, and given that H2 proposes that favoritism will follow failure in stable environments, one can deduce that:

H3: Environmental uncertainty will moderate the relationship

between performance and favoritism.

Method

Sample

The study included 45 independent firms and 20 strategic business units (SBUs) in the office and residential furniture industry (SIC codes 2511, 2512, 2521, 2522). All our SBUs were independent profit centers, all were functionally organized, and all had their own top executives (here we call them CEOs) who responded to our survey. All units had sales exceeding $10 million and were single business units with structures organized by function. This industry was deemed suitable for testing our hypotheses as it includes firms facing uncertain and dynamic niches as well as those competing in very stable ones. Firms were approached by sending a covering letter requesting a personal interview, along with a copy of our questionnaire. We followed up with a minimum of three calls to each firm during which we attempted to set up an appointment to conduct a telephone interview with the Chief Executive Officer (CEO).

In order to ensure the independence of the responses, we attempted never to sample more than one SBU from the same parent, nor did we ever include an SBU if the parent had been polled. Because we insisted on having interviews of at least one hour with busy CEOs of significant sized companies, our response rate was only 20%, after three follow-ups. The final sample consisted of 65 firms, with mean sales of $122 million (standard deviation of $279). The mean number of employees was 1,364 (s.d. 2,713).

Measurement

Functional Favoritism: Two kinds of indexes can be used to assess favoritism: the variance across functional areas in attributions of importance and competence, and the extent to which the favorite function dominates all others (Miller & Chen 1993). For purposes of this study, these functional departments are manufacturing, marketing/sales, engineering, product design and development, and distribution/logistics. Interviews with experts in the furniture industry revealed these to be the departments most likely to create a distinctive competence.

To avoid defensive, polite or “politically correct” responses, we refrained from asking CEOs directly to rank their five functional departments. Instead, we first polled CEOs about the importance to–and the achievements of–their firms for a set of 31 concrete and explicit competences, (see Appendix). Only then did we ask them to allocate functional responsibility to each of the competences. Competences were identified as being critical to competition in the furniture industry based on an extensive review of the literature in marketing, new product management and manufacturing, and on pre-test interviews with six CEOs in the industry.

Responding CEOs were first asked to rate the importance of each of the 31 competitive competences to implementing a firm’s strategy. A 7-point scale was used with anchors of “least important” and “most important.” All items were defined for the executives. CEOs then were asked to rate the firm’s performance on each of the 31 items on a 7-point scale ranging from “poor” to “excellent. Finally, for each item, CEOs were asked to allocate in percentages the relative degree to which each of the five functional areas was responsible for each competency (the sum of the responsibility ratings did not always equal 100%–but it averaged about 95%).

We measured favoritism in the assignments of both importance and competence. Our importance indexes are based on the importance scores attributed to each functional area. For each of the 31 items, the importance rating was multiplied by the percentage responsibility rating for each functional area to obtain the weighted importance score. These weighted scores were summed across all 31 items to get the total weighted importance’ score for each functional area. For each firm there were five such total scores, one for each functional area.

The Concentration of Importance index is the standard deviation for each firm across the five total weighted importance scores. It reflects the extent to which CEOs assign the most importance to a few functions. The Dominance of Importance index is the highest total weighted importance score of the five functional areas, and it shows the extent to which the function deemed most important dominates all others. Both of these importance indexes were divided by the mean total weighted importance score to adjust for differences in the mean scoring tendencies of CEOs. This reduces rater bias and removes a key source of common method variance.

Two more indexes assess attributions of competence. They are based on the products of three scores for each of the 31 competence items: the importance score for the item (coded 1 to 7), the performance relative to competitors on that item (recoded -3 to +3), and the percentage functional responsibility score. These product scores were summed across all 31 items by functional area to derive a total weighted competence score for each functional area. The total weighted competence scores were used to derive two competence indexes of favoritism.

The Concentration of Competence index is the standard deviation among the five total weighted competence scores. This index shows the extent to which CEOs assign the most competence to a few functions. The Dominance of Competence index is the highest total weighted competence score among the five, and it shows the extent to which the function perceived to be the most competent is believed to surpass all others. Again, to reduce common method variance, both indexes were divided by the mean total weighted competence score.

The correlation coefficients of Table 1 show that all four indexes are highly intercorrelated, especially those within the importance and competence categories. It appears that ascriptions of competence sometimes go hand-in-hand with ascriptions of importance. This may be due in part to a tendency towards self-flattery–to imply that what one is good at is indeed important, and vice versa.

[TABULAR DATA 1 NOT REPRODUCIBLE IN ASCII]

Performance: Financial performance was evaluated along the following dimensions: return on total assets (ROA), return on investment (ROI), and growth in return on investment (ROI growth). CEOs were asked to compare the performance of their firm relative to its major competitors on a 7-point scale, where I represented “worst in the industry” and 7 represented “best in the industry.” Given that this is a study of how CEOs react to performance, their perceptions of performance would be more germane to our predictions than would objective performance. We did, however, want to establish the degree to which the subjective performance ratings were borne out by the actual financial results supplied by one-third of the firms. The correlations of subjective versus actual performance were as follows: for ROA, 0.55, for ROI, 0.49 and for ROI growth, 0.65 (all statistically significant at p [is less than] 0.01). Thus, the subjective ratings of performance do relate to the actual financial results.

Environmental Uncertainty: Uncertainty was measured using the five-item, 7-point scales of Miller & Droge (1986), which were based on the work of Khandwalla (1977). The mean of the five scales was taken as an indicator of overall uncertainty for each firm. The Cronbach alpha reliability for the overall scale was 0.55. According to Van de Ven & Ferry (1979), this was an acceptable level of reliability for a multi-attribute dimension such as this.

Analyses

Descriptive statistics and product-moment correlations are presented in Table 1. The first part of the table shows the means and standard deviations for the entire sample and for the subsamples split on uncertainty. The second part of the table shows the correlation matrix for the entire sample. And the third part of Table 1 presents the correlation matrices for the subsamples.

The research hypotheses were tested in three ways. First, we used the separate multiple regression analyses of Tables 2 and 3; the former presents the results for the uncertain environment subsample in order to test Ell; the latter presents the results for the stable environment subsample to test H2. These tables show the top and bottom thirds of the sample when it was divided along the uncertainty dimension. In order to verify the robustness of this split, we also bifurcated the sample along the median for uncertainty. The results for those subsamples very closely 2 replicated the findings on Tables 2 and 3.(2)

Table 2. (Top 1/3) Multiple Regression Analyses: High Uncertainty Subsample Standardized Betas

(P-Values in Parentheses)(*)

A. Concentration of Importance

1. Log Employees -.209(.183) -.044(.429)

2. Uncertainty .354(.069) .252(.155)

3. Performance:

ROA .490(.023)

R01 .280(.131)

R01 Growth

4. [R.sup.2] .442 .142

5. F(p-value) 2.923(.039) 0.830(.249)

B. Dominance of Importance

1. Log Employees -.026(.461) .066(.397)

2. Uncertainty .279(.146) .245(.169)

3. Performance:

ROA .395(.070)

R01 .155(.269)

R01 Growth

4. [R.sup.2] .255 .089

5. F(p-value) 1.367(.149) 0.488(.348)

C. Concentration of Competence

1. Log Employees -.247(.175) -.202(.195)

2. Uncertainty -.022(.466) -.069(.382)

3. Performance:

ROA .431(.056)

R01 .451(.033)

R01 Growth

4. [R.sup.2] .250 .229

5. F(p-value) 1.330(.155) 1.486(.129)

D. Dominance of Competence

1. Log Employees -.172(.262) -.134(.289)

2. Uncertainty -.023(.466) -.059(-401)

3. Performance:

ROA .406(.078)

ROI .400(.056)

ROI Growth

4. [R.sup.2] .196 .170

5. F(p-value) 0.973(.219) 1.023(.205)

A. Concentration of Importance

1 . Log Employees .023(.465)

2. Uncertainty .229(.187)

3, Performance:

ROA

R01

R01 Growth .493(.035)

4. [R.sup.2] .277

5. F(p-value) 1.530(.129)

B. Dominance of Importance

1. Log Employees .069 (.399)

2. Uncertainty .233 (.194)

3. Performance:

ROA

R01

R01 Growth .379(.087)

4. [R.sup.2] .168

5. F(p-value) .919(.231)

C. Concentration of Competence

1. Log Employees -.230(.203)

2. Uncertainty -.002(.497)

3. Performance:

ROA

R01

R01 Growth .303(.139)

4. [R.sup.2] .160

5. F(p-value) 0.772(.265)

D. Dominance of Competence

1. Log Employees -.219(.221)

2. Uncertainty .016(.477)

3. Performance:

ROA

ROI

ROI Growth .227(.212)

4. [R.sup.2] .111

5. F(p-value) 0.498(.345)

Note: One-tailed p-values are reported.

Table 3. (Bottom 1/3). Multiple Regression Analyses: Low Uncertainty Subsample Standardized Betas

(P-Values in Parentheses)(*)

A. Concentration of Importance

1. Log Employees .402 (.058) -.307 (.119)

2. Uncertainty .183 (.269) .199 (.216)

3. Performance

ROA -.500 (.029)

ROI -.353 (.091)

ROI Growth

4. [R.sup.2] .338 .235

5. F(p-value) 2.214 (.067) 1.333 (.153)

B. Dominance of Importance

1. Log Employees .385 (.054) .255 (.158)

2. Uncertainty .120 (.292) .144 (.281)

3. Performance:

ROA -.627 (.008)

ROI -.440 (.049)

ROI Growth

4. [R.sup.2] .419 .254

5. F(p-value) 3.131 (.031) 1.473 (.134)

C. Concentration of Competence

1. Log Employees .635 (.005) .495 (.024)

2. Uncertainty .013 (.473) .008 (.486)

3. Performance:

ROA -.578 (.008)

ROI -.449 (.036)

ROI Growth

4. [R.sup.2] .501 .360 .335

5. F(p-value) 4.353 (.012) 2.441 (.055)

D. Dominance of Competence

1. Log Employees .619 (.006) .474 (.029)

2. Uncertainty .011 (.479) .005 (.489)

3. Performance:

ROA -.581 (.009)

ROI -.445 (.038)

ROI Growth

4. [R.sup.2] .489 .341

5. F(p-value) 4.153 (.014) 2.246 (.066)

A. Concentration of Importance

1. Log Employees .264 (.153)

2. Uncertainty .176 (.249)

3. Performance

ROA

ROI

ROI Growth -.212 .(204)

4. [R.sup.2] .143

5. F(p-value) 0.780 .(262)

B. Dominance of Importance

1. Log Employees .203 (.207)

2. Uncertainty .148 (.275)

3. Performance:

ROA

ROI

ROI Growth -.355 (.083)

4. [R.sup.2] .188

5. F(p-value) 1.080 (.195)

C. Concentration of Competence

1. Log Employees .436 (.033)

2. Uncertainty .031 (.445)

3. Performance:

ROA

ROI

ROI Growth -.406 (.043)

4. [R.sup.2]

5. F(p-value) 2.352 (.058)

D. Dominance of Competence

1. Log Employees .417 (.039)

2. Uncertainty .028 (.449)

3. Performance:

ROA

ROI

ROI Growth -.420 (.038)

4. [R.sup.2] .330

5. F(p-value) 2.301 (.061)

Note: One-tailed p-values are reported.

Because we employed four indexes of favoritism and three indexes of performance, there are 12 separate regressions for each of the subsamples. The dependent variables are the various indexes of favoritism and the independent variables are size, uncertainty, and performance. Size, assessed as the log number of employees, was a control variable that was included in case small firms tended to rely more upon a single function than larger ones.

To test H3 we ran a set of moderated regression analyses for the total sample (Table 4). An interaction term was used to establish the differential impact of performance on favoritism for different levels of uncertainty. Each regression takes one of the favoritism indexes as the dependent variable and, size, uncertainty, performance and a product of uncertainty and performance (the moderator term) as independent variables. In order to avoid multicollinearity with the main effects, we standardized the components of the interaction (moderator) term before multiplying them (Smith & Sasaki, 1979). Our p-values assess the significance of the incremental variance explained by the interaction terms after they were added to the full models.

Table 4. Moderated Regression Analyses: Total Sample Standardized Betas

(P-Values in Parentheses)(*)

A. Concentration of Importance

1. Log Employees .073 (.299) .070 (.304)

2. Uncertainty -.037 (.269) -.061 (.327)

3. Performance:

ROA .032 (.408)

ROI -.044 (.375)

ROI Growth

4. Interaction .395 (.003) .258 (.030)

5. [R.sup.2] .144 .074

6. F(p-value) 2.107 (.047) 1.041 (.198)

B. Dominance of Importance

1. Log Employees .107 (.217) .085 (.267)

2. Uncertainty -.026 (.421) -.029 (.413)

3. Performance:

ROA -.056 (.342)

ROI -.139 (.155)

ROI Growth

4. Interaction .396 (.003) .239 (.040)

5. [R.sup.2] .161 .084

6. F(p-value) 2.406 (.031) 1.196 (.162)

C. Concentration of Competence

1. Log Employees .294 (.015) .220 (.045)

2. Uncertainty -.203 (.059) -.164 (.099)

3. Performance:

ROA -.171 (.099)

ROI -.152 (.119)

ROI Growth

4. Interaction .361 (.004) .330 (.005)

5. [R.sup.2] .223 .187

6. F(p-value) 3.577 (.006) 2.984 (.014)

D. Dominance of Competence

1. Log Employees .286 (.017) .211 (.053)

2. Uncertainty -.199 (.062) -.156 (.111)

3. Performance:

ROA -.183 (.084)

ROI -.159 (.110)

ROI Growth

4. Interaction .360 (.004) .322 (.007)

5. [R.sup.2] .223 .119

6. F(p-value) 3.587 (.006) 2.835 (.017)

A. Concentration of Importance

1. Log Employees .024 (.429)

2. Uncertainty .015 (.455)

3. Performance:

ROA

ROI

ROI Growth -.035 (.397)

4. Interaction .367 (.004)

5. [R.sup.2] .137

6. F(p-value) 1.947 (.059)

B. Dominance of Importance

1. Log Employees .005 (.486)

2. Uncertainty .017 (.450)

3. Performance:

ROA

ROI

ROI Growth -.083 (.271)

4. Interaction .352 (.006)

5. [R.sup.2] .128

6. F(p-value) 1.797 (.073)

C. Concentration of Competence

1. Log Employees .198 (.066)

2. Uncertainty -.143 (.139)

3. Performance:

ROA

ROI

ROI Growth -.196 (.068)

4. Interaction .320 (.008)

5. [R.sup.2] .183

6. F(p-value) 2.747 (.019)

D. Dominance of Competence

1. Log Employees .181 (.084)

2. Uncertainty -.141 (.141)

3. Performance:

ROA

ROI

ROI Growth -.223 (.045)

4. Interaction .317 (.009)

5. [R.sup.2] .185

6. F(p-value) 2.786 (.019)

Note: (*) One-tailed p-values are reported.

For our third set of analyses, we ran two two-group LISREL models (Joreskog & Sorbom, 1989), one for importance and one for competence. The model for importance is presented in Figure 2. This figure presents half of the entire LISREL model, the other half having the same constructs and paths. The two halves or groups represent the high and low uncertainty subsamples (top and bottom thirds) respectively. The two group LISREL model estimates the measurement model and the structural paths in both subsamples simultaneously.

[Figure 2 ILLUSTRATION OMITTED]

The measurement model consists of one endogenous construct (i.e., favoritism) and three exogenous constructs (uncertainty, size and performance) in each group. Favoritism has two indicators–dominance and concentration–each of which is measured with error.(1) Thus the basic measurement model estimated for favoritism consists of two factor loadings and two error terms in each group, and these were specified to be equal across groups. Uncertainty and size were measured without error in each group (i.e., the factor loadings were set to 1 and the error terms to 0). Performance was measured, with errors, by ROA, ROI and ROI growth, and again the measurement model (i.e., factor loadings and errors) was specified to be equal across the two groups. Finally, the following pattern was specified for phi, the covariance matrix of the exogenous constructs:

1. all diagonal elements were specified to be equal across groups:

2. the covariance between size, the control variable, and performance was estimated but constrained to be equal across groups; and

3. all other off-diagonal elements were set to zero.

The structural part of the model consists of three paths in each group: from uncertainty, size and the performance construct, respectively, to the favoritism construct. These paths, of course, were not constrained to be equal across groups as we wished to assess their differences between the subsamples. We were most interested in a cross-group comparison of the estimates of the performance-favoritism relationship when the constrained measurement models and the other structural paths are estimated in both groups simultaneously. The paths of interest here are between constructs and not between variables, as was the case in the regression models.

Findings

Results of the Regression and Moderated Regression Analyses

Table 2 shows that in the high uncertainty subsample, performance is positively and significantly related to the various importance indexes of favoritism in four of the six cases, two of these coefficients at beyond the 0.05 level. The findings are similar for the competence indexes of favoritism with four of the six coefficients reaching significance, again two at beyond the 0.05 level. Overall, then, H1 received moderate support.

Table 3 indicates that in the low uncertainty subsample, performance is negatively and significantly related to the importance indexes of favoritism in five of the six cases, three at beyond the 0.05 level. For the competence indexes of favoritism, six of the six coefficients attain significance, all at beyond the 0.05 level. All in all, H2 received strong support. The results of Tables 2 and 3 were confirmed when we used a median split on uncertainty to bifurcate the sample.2

The consistency of these results becomes even more apparent in Table 4 when we assess H3. We ran moderated regression analyses on the total sample and tested for the significance of a performance-uncertainty multiplicative interaction term. The results establish that the relationship between performance and favoritism is indeed conditioned by environmental uncertainty. The interaction term was positive and explained significant additional variance after all other terms had been included in the regressions in all 12 out of 12 analyses (all coefficients were significant at beyond the 0.05 level, 10 at beyond the 0.01 level).

Results of the LISREL Analyses

The results of the two group LISREL analyses can be classified along three dimensions: (1) Overall fit of the specified models across both groups simultaneously; (2) Indicators of fit within each group of the specified model; and (3) Analyses of the significance of individual paths in each group.

Overall Fit: In well-fitting models, the p value of the chi-square statistic will be greater than 0.05. For the importance scores, the chi-square was 45.12 (d.f. = 36, p = 0. 142), while for the competence scores, it was 88.14 (d.f. = 37, p = 0.00). Clearly, the former model fits very well overall, while the latter does not.

It is important to note that we did not engage in a priori model pruning, that is, selectively eliminating paths or variables that the regression analyses had already indicated would be nonsignificant (uncertainty, for example, was never a significant regressor at p [is less than] 0.05). One way to improve the overall fit of the LISREL models, and of course the regression models, would have been to drop uncertainty. Our goal, however, was NOT to obtain the best fitting LISREL model, but rather to test our theory completely by examining the performance-favoritism paths across groups with all variables in the model (Joreskog & Sorbom, 1989).

Group-Specific Fit: Although Tables 5 and 6 show some lack of fit in each of the models, particularly for competence scores, there were no serious indications of misfit. For example, there were no correlations greater than 1.0 nor any negative multiple correlations. The squared multiple correlations for the structural equations, in fact, ranged from 0.256 to 0.488, and for our models these numbers are equal to the total coefficients of determination for all structural paths jointly.

Table 5. Two-Group LISREL Results for Importance Scores

A. Group Specific Indicators of Fit

Top 1/3 Bottom 1/3

1. Goodness of fit index 0.755 0.865

2. Modification indices over 5 0 0

3. Root mean square residual 0.311 0.305

4. Squared multiple correlations 0.350 0.351

B. LISREL Measurement Model Estimates

(Note: Measurement models were specified equal across groups)

Unstandard Standard

Estimate Error

1. Favoritism loadings:

Dominance Set to 1.0 0.000

Concentration 0.549(***) 0.066

2. Performance loadings:

ROA Set to 1.0 0.000

ROI 1.964(***) 0.087

ROI Growth 0.756(***) 0.108

Common Metric

Standardized Estimate

1. Favoritism loadings:

Dominance 0.433

Concentration 0.237

2. Performance loadings:

ROA 1.673

ROI 1.612

ROI Growth 1.264

C. LISREL Structural Path Estimates

Top 1/3

Structural path to Unstd. Est. Common Metric

Favoritism from: (Std. Error) Std. Est.

1. Size 0.012 (0.050)(*) 0.040

2. Uncertainty 0.276 (0.160) 0.281

3. Performance 0.132 (0.044)(***) 0.511

Bottom 1/3

Structural path to Unstd. Est. Common Metric

Favoritism from: (Std. Error) Std. Est.

1. Size 0.107 (0.051)(**) 0.353

2. Uncertainty 0.090 (0.161) 0.091

3. Performance 0.139 (0.044)(***) 0.537

Notes: (*), (**), (***) Significantly different from zero at 0.10, 0.05 and 0.01, respectively.

Table 6. Two-Group LISREL Results for Competence Scores

A. Group-Specific Indicators of Fit

Top 1/3 Bottom 1/3

1. Goodness of fit index 0.768 0.791

2. Modification indices over 5 3 4

3. Root mean square residual 0.305 0.299

4. Squared multiple correlations 0.256 0.488

B. LISREL Measurement Model Estimates

(Note: Measurement models were specified equal across group)

Unstandard Standard

Estimate Error

1. Favoritism loadings:

Dominance Set to 1.0 0.000

Concentration 0.636(***) 0.014

2. Performance loadings:

ROA Set to 1.0 0.000

ROI 1.019(***) 0.066

ROI Growth 0.798(***) 0.099

Common Metric

Standardized Estimate

1. Favoritism loadings:

Dominance 2.989

Concentration 1.901

2. Performance loadings:

ROA 1.627

ROI 1.658

ROI Growth 1.298

C. LISREL Structural Path Estimates

Top 1/3

Structural path to Unstd. Est. Common Metric

Favoritism from: (Std. Error) Std. Est.

1. Size -0.012 (0.190) -0.056

2. Uncertainty -0.282 (0.604) -0.040

3. Performance 0.487 (0.165)(***) 0.265

Bottom 1/3

Structural path to Unstd. Est. Common Metric

Favoritism from: (Std. Error) Std. Est.

1. Size 1.547 (0.399)(***) 0.720

2. Uncertainty -0.502 (1.269) -0.072

3. Performance -1.324 (0.347)(***) -0.721

Notes (*), (**), (***) Significantly different from zero at 0.10, 0.05 and 0.01, respectively.

Path Estimates: LISREL estimates both a measurement model (parts B, Tables 5 and 6) and a structural path model (parts Q. For each path, we present the unstandardized estimate and the standard error as well as the common metric standardized solution. The measurement models were specified equal across groups, but the structural paths were estimated separately for each group. In specifying the measurement models we followed the normal convention of setting to I the lambda estimate of one variable per construct. Finally, the common metric standardized solution is presented rather than the within-group solution because only it maintains equality constraints and permits comparison of parameters across subgroups (Joreskog & Sorbom, 1989, pp. 238-242).

The results show first that the measurement model lambdas estimated are all significant at less than 0.001 for both models of favoritism. In both models, the performance factor loadings of ROA, ROI and ROI growth show the same pattern of results.

The structural paths indicate that the relationship between performance and favoritism is significantly different from zero at the 0.01 level or less in each subgroup in both models: in the high uncertainty subgroup it is always positive, and in the low uncertainty subgroup it is always negative. Size had a variable impact on favoritism, and uncertainty never attained significance as a predictor in any of the subgroups. In conclusion, when measurement is modelled, the results are stronger and more significant than those reported in the individual regression analyses. Both of the hypotheses were supported.

Discussion and Conclusions

Patterns in the Findings

Three different kinds of analyses, each with its own distinctive strengths and weaknesses, showed that the relationships between performance and functional favoritism were opposite in the high vs. low uncertainty subsamples. Most of the regressions and each of the LISREL analyses revealed these opposite relationships to be statistically significant. In short, there was much convergence among the findings from the subgroup regression analyses, the moderated regression analyses, and the two-group LISREL analyses. And the path results improved when we accounted for measurement variability by using LISREL. Clearly, the overall connections between the constructs were very consistent and in line with our hypotheses.

It is also interesting that in the subgroup regression analyses the importance indexes of favoritism do not support either of the hypotheses as strongly as the competence indexes of favoritism. The political arguments we made in support of our hypotheses maintained that CEOs will try wherever possible to exploit their association with a particular function. If so, they are apt to celebrate the competence of that function even more than the importance of the issues it deals with (Halberstam, 1986; Miller, 1990).

Contexts for Favoritism

There is a growing literature on the impact of performance on managerial perceptions (Milliken & Lant, 1991). But there appears to be a basic schism in this research. Some authors studying organizational learning and attribution processes have stressed that success causes conceptual bias and narrowness (March, 1991; Miller, 1993; Milliken & Lant, 1991). Other researchers studying escalation of commitment, prospect theory, and threat-rigidity reactions have argued that failure produces a similar result (Brockner, 1992; Kahneman & Tversky, 1979; Staw, 1981, Staw et al., 1981). This research on functional favoritism, which can be considered one type of bias or narrowness, suggests that there may be merit to both views–but each in its own particular context.

We found that functional favoritism is NOT simply a product of good or bad performance. The environment is all-important. In reacting to success or failure, leaders must have the psychological and political incentives to favor particular functions. Under success, these incentives flourish in the environment for superstitious learning and opportunism provided by an uncertain context. And under failure, such incentives are born of the escalation and threat-rigidity effects fostered by a stable context.(3)

Favoritism and Simplicity

Miller (1990, 1993) has argued that successful organizations become more simple over time–their upper echelons come to concentrate on a single objective, their strategic repertoires encompass a narrowing array of skills, and their attentions come to be focused on fewer aspects of the environment. Although such simplicity can help firms develop special competences, in the long run it impedes adaptation and erodes organizational resilience. Our research suggests that these tendencies and biases will be more prevalent in uncertain than in stable environments. Here, favoritism based on success may narrow managerial worldviews and therefore, perhaps reduce an organization’s resilience and adaptability.

The literature on escalation of commitment and threat-rigidity effects argues that perceptual narrowness and departmental or strategic obsessions may well be intensified by failure (Brockner, 1992; Staw, 1981). These reactions and biases too can exacerbate problems and impede much needed adaptation. But our results indicate that they will be most likely to occur in stable environments.

Further Research

Our discussion has been speculative. Longitudinal research will be needed to evaluate the relative merits of explanations using economic or ecological rationality versus those based on myth creation, superstitious learning, and opportunism. Unfortunately, our cross-sectional study cannot help us here. Another limitation of this research is that we did not directly measure any learning, political or attribution processes. These can best be studied by in-depth analyses of a few organizations. It would be useful, we think, for subsequent researchers to begin to look in finer detail at the processes behind managerial favoritism. The challenges are great, but so may be the rewards.

One of the major limitations of this study is that we have looked only at a single industry. Perhaps CEOs in more turbulent and uncertain industries such as software and semiconductors may be less given to favoritism. Instead, the many challenges they face might lead in the normal course to greater eclecticism in evaluating their functions. This eclecticism may be reversed mainly when poor performance indicates a need to change (Miller & Chen, 1993, 1996; Walsh, 1995).

Notes

(1.) For the competence measurement model, one error term was set to zero since the dominance and concentration indexes are so highly correlated.

(2.) To establish the robustness of our findings we re-ran all analyses by splitting the sample into two instead of three equal parts, again based on the level of environmental uncertainty. We then tested Hypotheses 1 and 2 on the top and bottom halves of the sample respectively. These results also show support for both Hypotheses. Hypothesis 1, which predicted a positive relationship between performance and functional favoritism in the uncertain subsample, was supported in 7 of the 12 regressions (6 at beyond the 0.05 level). Hypothesis 2, which predicted a negative relationship between performance and favoritism in the stable subsample, was supported in 11 of the 12 regressions (9 at beyond the 0.05 level).

(3.) Some might argue that functional favoritism causes extremes of performance. Our research. however, shows that it is past performance that is followed by favoritism. Moreover, the results indicate that environment moderates the relationship between performance and favoritism. It would be hard to argue that favoritism would lead to good performance in uncertain environments — settings that demand much parallel effort from and intensive collaboration among numerous functional departments (Lawrence & Losrch, 1967: Galbraith, 1973). Similarly, it is difficult to see why favoritism would be especially likely to lead to poor performance in stable settings, where some firms may actually benefit from the efficiencies wrought by specialization and focus (Burns & Stalker, 1961: Thompson, 1967).

References

Aguilar, F.J. (1967). Scanning the business Environment. NY: Macmillan.

Aldrich, H. (1979). Organizations and environments. Englewood Cliffs. NJ: Prentice Hall.

Amburgey, T.L. & Miner, A.S. (1992). Strategic momentum: The effects of repetitive, positional and contextual momentum on merger activity. Strategic Management Journal, 13: 335-348.

Argyris, Chris & Schon, Donald. (1978). Organizational learning: A theory of action perspective, Reading. MA: Addison Wesley.

Barney, J. (1991). Firm resources and sustained competitive advantage Journal of Management, 17: 99-120.

Blau, P.M. & Schoenherr, Richard. (1971). The structure of organizations. NY: Basic Books.

Burgelman, R.A. (1994). Fading memories: A process theory of strategic business exit in dynamic environments. Administrative Science Quarterly, 39: 24-56.

Burns, T. & Stalker, G.M. (1961). The management of innovation. London: Tavistock.

Brockner, J. (1992). The escalation of commitment to a failing course of action: Toward theoretical progress. Academy of Management Review, 17: 39-61.

Cohen, W.M. & Levinthal, D. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35: 128-152.

Cyert, R.M. & March, J.G. (1963). A behavioral theory of the firm. Englewood Cliffs. NJ: Prentice Hall.

D’Aveni, R. & MacMillan, I. (1990). Crisis and the content of managerial communications. Administrative Science Quarterly, 35: 634-657.

Dearborn, D.C. & Simon, H.A. (1958). Selective perception: A note on the departmental identification of executives. Sociometry, 21: 140-144.

Dess, G. & Beard, D. (1984). Dimensions of organizational task environments-Administrative Science Quarterly, 29: 52-73.

Duncan, R.B. (1972). Characteristics of organizational environments and perceived environmental uncertainty. Administrative Science Quarterly, 17: 313-327.

Dutton, Jane & Jackson, S. (1987). Categorizing strategic issues: Links to organizational action. A, Academy of Management Review, 12: 76-90.

Eisenhardt, Kathleen M. (1989). Making fast strategic decisions in high volocity environments. Academy of Management Journal, 32: 543-576.

Finklestein, S. & Hambrick, D. (1990). Top-management-team tenure and organizational outcomes: The moderating role of managerial discretion. Administrative Science Quarterly, 35: 484-503.

Fox, F.V. & Staw, B.M. (1979). The trapped administrator: Effects of job insecurity and policy resistance upon commitment to a course of action. Administrative Science Quarterly, 24: 449-471.

Galbraith, Jay. (1973). Designing complex organizations. Reading. MA: Addison Wesley.

Halberstam, D. (1986). The Reckoning, New York: Avon.

Hall, R. (1992). The strategic analysis of intangible resources. Strategic Management Journal, 13: 135-144.

Hambrick, Donald C. & D’Aveni, Richard A. (1988). Large corporate failures as downward spirals. Administrative Science Quarterly, 33: 1-23.

Hambrick, D.C. & Finkelstein, S. (1987). Managerial discretion: A bridge between polar views of organizational outcomes. Pp. 369-406 in B. Staw & L. Cummings (eds.). Research in organizational behavior. Vol. 9. Greenwich, CT: JAI Press.

Hannan, M.T. & Freeman, J. (1977). The population ecology of organizations-American Journal of Sociology, 82: 929-964.

Holsti, O. (1969). Crisis, escalation, war. Montreal: McGill-Queens University Press.

Huber, G.P. (1991). Organizational learning., Organization Science, 2: 88-115

Janis, Irving. (1972). Victims of group think. Boston: Houghton Mifflin.

Joreskog, K.G. & Sorbom, D. (1989). LISREL 7: A guide to the program and applications. Chicago, IL: SPSS Inc.

Kahneman, D. & Tversky, A. (1979). Prospect theory: An analysis of decisions under risk. Econometrica. 47: 263-291.

Kets de Vries, Manfred & Miller, D. (1994). The neurotic organization. San Francisco, CA: Jossey Bass.

Khandwalla, Pradip N. (1977). The design of organizations. New York: Harcourt Brace Jovanovich.

Kiesler, Sara & Sproull, L. (1982). Managerial response to changing environments: Perspectives on problem sensing from social cognition. Administrative Science Quarterly, 27: 548-570.

Lant, T.K. (1992). Aspiration level updating: An empirical exploration. Management Science, 38: 623-644.

Lant, T.K., Milliken, F.J. & Batra, B. (1992). The role of managerial learning and interpretation in strategic persistence and reorientation. Strategic Management Journal, 13: 585-608.

Lant, T.K. & Montgomery, D.B. (1987). Learning from strategic success and failure. Journal of Business Research, 15: 503-518.

Lawrence, P.R. & Lorsch, J. (1967). Organization and environment. Boston: Harvard.

Levitt B. & March, J.G. (1988). Organizational learning. Annual Review of Sociology, 14: 319-340.

Lippman, S. & Rumelt, R. (1982). Uncertain imitability. Bell Journal of Economics, 13: 418-438.

March, J.G. (1991). Exploration and exploitation in organizational learning. Organization Science, 2: 71-87.

March, James G. & Simon, Herbert A. (1958). Organizations. New York: Wiley.

Meyer, A.D. (1982). Adapting to organizational jolts. Administrative Science Quarterly, 27: 515-537.

Meyer, A.D. & Starbuck, W.H. (1991). Organizations and industries in flux: The interplay of rationality and ideology. Working paper, University of Oregon.

Meyer, Marshall & Zucker, L. (1989). Permanently failing organizations. Newbury Park. CA: Sage.

Miller, Danny. (1990). The Icarus paradox: How exceptional companies bring about their own downfall. New York: Harper Collins.

–. (1993). The architecture of simplicity. Academy of Management Review, 18: 116-138.

–. (1994). What happens after success: The perils of excellence. Journal of Management Studies, 31: 325-358.

–. (1991). Stale in the saddle: CEO tenure and the match between organization and environment. Management Science, 37: 34-52.

Miller, D. & Chen, M.J. (1993). The simplicity of strategic repertoires. Best Paper Proceedings of the Academy of Management Meeting: 9-14.

–. (1994). Sources and consequences of competitive inertia. Administrative Science Quarterly, 39: 1-24.

–. (1996). The simplicity of competitive repertoires. Strategic Management Journal, 17: 419-439.

Miller, D. & Droge, C. (1986). Psychological and traditional determinants of structure. Administrative Science Quarterly, 31: 539-560.

Miller, D. & Friesen, P.H. (1984). Organizations: A quantum view. Englewood Cliffs, NJ: Prentice Hall.

Miller, D.T. & Porter, C. (1988). Errors and biases in the attribution process. Pp. 3-29 in L.Y. Abramson (ed.). Social cognition and clinical psychology. New York: Guilford Press.

Milliken, F.J. & Lant, T.K. (1991). The impact of an organization’s recent performance history on strategic persistence and change. Pp. 129-156 in P. Shrivastava, A. Huff & J. Dutton, (eds.), Advances in strategic management, Vol. 7. Greenwich, CT: JAI Press.

Mintzberg, Henry. (1983). Power in and around organizations. Englewood Cliffs. NJ: Prentice Hall.

Nystrom, P.C. & Starbuck. W.H. (1984). To avoid crises, unlearn. Organizational Dynamics: 55-64.

Pascale, R.T. (1989). Managing on the edge. New York: Simon & Shuster,

Pfeffer, J. (1981). Power in organizations. New York: Pitman.

Prahalad, C.K. & Hamel. G. (199). The core competence of the corporation. Harvard Business Review. (May-June): 79-91.

Salancik, G. & Meindl, J. (1984). Corporate attributions as strategic illusions of management control. Administrative Science Quarterly, 29: 238-254.

Smith, K.W. & Sasaki, M. (1979). Decreasing multicollinearity: A method for models with multiplicative functions. Sociological Methods and Research. 8(1): 35-56.

Starbuck, W.H. (1965). Organizational growth and development. In J.G. March (ed.) Handbook of organizations. Chicago: Rand McNally.

Starbuck, W.H., Greve, A. & Hedberg. B.L.T. (1978). Responding to crises. Journal of Business Administration. 9: 111-137.

Starbuck, W.H. & Milliken. F.J. (1988). Challenger Fine-tuning the odds until something breaks. Journal of Management Studies. 25: 319-340.

Staw, B.M. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action. Organizational Behavior and Human Performance. 16: 27-44.

–. (1981). The escalation of commitment to a course of action. Academy of Management Review. 6: 577-587.

Staw, B.M., McKechnie, P. & Puffer. S. (1983). The justification of organizational performance. Administrative Science Quarterly, 28: 582-600.

Staw, B.M. & Ross. J. (1978). Commitment to a policy decision: A multitheoretical perspective. Administrative Science Quarterly, 23: 40-64.

–. (1987). Understanding escalation situations: Antecedents, prototypes and solutions. Pp. 39-78 in B.M. Staw & L.L Cummings (eds.), Research in organizational behavior. Vol. 9. Greenwich CT: JAI Press.

Staw, B.M., Sandelands, L. & Dutton, J. (1981). Threat-rigidity effects in organizational behavior. A multilevel analysis. Administrative Science Quarterly, 26: 501-524.

Van de Ven, A. & Ferry, D. (1979). Measuring and assessing organizations. New York: Wiley, 1979.

Walsh, J.P. (1995). Managerial or organizational cognition. Organization Science, 6: 280-321

Weick, K.E. (1979). The social psychology of organizing. Reading. MA: Addison Wesley.

Wernerfelt, B. (1984). A resource-based view of the firm. Strategic Management Journal, 5:171-180.

Whyte, G. (1986). Escalating commitment to a course of action: A reinterpretation. Academy of Management Review, 11: 311-321.

Wilensky. H. (1971). Organizational intelligence. New York: Free Press.

Williamson, O. (1975). Markets and hierarchies. New York: Free Press.

Wright, P. (1979). On a clear day you can see General Motors. New York: Avon

Appendix 1

The Thirty-one Competencies

1. Product Flexibility (Customization): The ability to handle difficult, nonstandard orders, to meet special customer specifications and to produce products characterized by numerous features, options, sizes and/or colors.

2. Volume Flexibility: The ability to rapidly adjust capacity so as to accelerate or decelerate production in response to changes in customer demand.

3. Process Flexibility: The ability to produce low quantities of product cost efficiently so that product mix changes are easily accommodated.

4. Low Production Cost: The ability to minimize the total cost of production (inclusive of labor, materials, and operating costs) through efficient operations, process technology and/or scale economies.

5. New Product Introduction: The ability to rapidly introduce large numbers of product improvements/variations or completely new products.

6. Delivery Speed: The ability to reduce the time between order taking and customer delivery to as close to zero as possible.

7. Delivery Dependability: The ability to exactly meet quoted or anticipated delivery dates and quantities.

8. Production Lead Time: The ability to reduce the time it takes to manufacture products.

9. Product Reliability: The ability to maximize the time to product failure or malfunction.

10. Product Durability: The ability to maximize the time to product replacement.

11. Quality (Conformance to Specifications): The ability to manufacture a product whose operating characteristics meet established performance standards.

12. Design Quality (Design Innovation): The ability to provide a product with capabilities, features, styling, and/or operating characteristics that are either superior to those of competing products or unavailable with competing products.

13. Product Development Cycle Time: The ability to minimize the time it takes to develop new products.

14. Product Technological Innovation: The ability to engage in new product development involving major advances in product technology.

15. Product Improvement/Refinement: The ability to further develop and refine existing products.

16. New Product Development: The ability to develop new products for existing markets.

17. Original Product Development: The ability to develop original (i.e., “new-to-the-world”) products that create entirely new markets.

18. Brand Image: The ability to create a positive or favorable image in the customer’s mind when he/she hears the product’s brand name.

19. Competitive Pricing: The ability to offer a lower product price than direct competitors.

20. Low Price: The ability to offer one of the lowest or the lowest available product price.

21. Advertising and Promotion: The ability to create effective advertising and/or promotional campaigns.

22. Target Market(s) Identification and Selection: The ability to identify promising target markets and select the best ones for consideration.

23. Responsiveness to Target Market(s): The ability to respond to the needs and wants of the firm’s target market(s).

24. Pre-Sale Customer Service: The ability to service the customer during the purchase decision process (i.e., before the customer buys the product).

25. Post-Sale Customer Service: The ability to service the customer after the sale of the product to ensure continuing customer satisfaction.

26. Broad Product Line: The ability to provide a comprehensive set of related items within a given product line offering.

27. Widespread Distribution Coverage: The ability to effectively provide widespread and/or intensive distribution coverage.

28. Low Cost Distribution: The ability to minimize the total cost of distribution.

29. Selective Distribution Coverage: The ability to effectively target selective or exclusive distribution outlets.

30. Personal Selling Proficiency: The ability to successfully move products through personal selling activities.

31. Company Reputation: The ability to create a positive or favorable image in the customer’s mind when he/she hears the company’s name.

Direct all Correspondence to: Danny Miller, Ecole des Hautes Etudes Commerciales and Columbia University. 4642 Melrose Avenue, Montreal, P.Q., Canada H4A 2S9.

COPYRIGHT 1997 JAI Press, Inc.

COPYRIGHT 2004 Gale Group