Planning - Learning SMJ Article

8/13/2019 Planning - Learning SMJ Article

1/25

Strategic Management JournalStrat. Mgmt. J., 20: 889913 (1999)

LEARNING TO PLAN AND PLANNING TO LEARN:

RESOLVING THE PLANNING SCHOOL/LEARNING

SCHOOL DEBATEPETER J. BREWS1* and MICHELLE R. HUNT2

1The Fuqua School of Business, Duke University, Durham, North Carolina, U.S.A.2The Kenan-Flagler Business School, The University of North Carolina, Chapel Hill,North Carolina, U.S.A.

This paper resolves the long-standing debate between the two dominant process schools instrategy. Analysis of the planning practices of 656 firms shows that formal planning andincrementalism both form part of good strategic planning, especially in unstable environments.

Environment neither moderates the need for formal planning nor the direction of the

planning/performance relationship, but does moderate firm planning capabilities and planningflexibility. In unstable environments planning capabilities are far better developed and formalplans more amenable to change. The planning/ performance relationship is, however, moderatedby planning duration: at least four years of formal planning are required before external

performance associations are noted. Copyright 1999 John Wiley & Sons, Ltd.

INTRODUCTION

A recent bitter debate between two prominentstrategy academicians considers a question vitalto the theory and practice of strategy: what types

of planning should firms utilize in their strategyformation behaviors? Ansoff, flying the PlanningSchool flag, contends that formal planning isbeneficial in both stable and unstable environ-ments (Ansoff, 1991, 1994) while Mintzberg, art-iculating the Learning School view, favors logicalincrementalism, especially in unstable environ-ments (Mintzberg, 1991; 1994a, 1994b). Thispaper presents a resolution to the debate, andreports on a study investigating whether environ-mental conditions moderate the type of planningfirms employ in their strategy formation activities.

The impact of the length of time a planning

Key words: formal strategic planning, logicalincrementalism, environmental stability, firm perform-ance* Correspondence to: Professor Peter J. Brews, The FuquaSchool of Business, Duke University, Box 90120, Durham,NC 27708-0120, U.S.A.

CCC 08869383/99/10088925 $17.50 Received 10 June 1998Copyright 1999 John Wiley & Sons, Ltd. Final revision received 21 April 1999

regime has been employed at a firm on theplanning/performance relationship was alsoexplored, together with other Learning Schoolcritiques of the Planning School. After reviewingprior research, the key constructs utilized in the

study are presented and discussed. Then, hypoth-eses are presented, followed by the study method-ology and a report of the findings. Once the studylimitations are acknowledged, a discussion of thekey implications of the findings, for research andpractice, concludes the paper.

THEORY DEVELOPMENT

Prior research

Few issues have attracted more attention in strat-

egy research than the relationship between themode of strategic planning adopted by the firmand the economic performance of the firm.Regrettably, decades of planning/performanceresearch have yielded inconsistent findings. Areview of 18 empirical studies testing the effectof formal strategic planning on economic per-formance concluded the link was tenuous


2/25

890 P. J. Brews and M. R. Hunt

(Pearce, Freeman and Robinson, 1987). A meta-analysis of 21 studies found that the formal stra-tegic planning/performance link was weak, witha correlation of 0.1507 (Boyd, 1991). Morerecently, a meta-analysis of 26 studies concludedthat strategic planning positively influenced firm

performance (Miller and Cardinal, 1994), whilea similar analysis of 14 studies investigating theeffects of planning on small firm financial per-formance concluded that the relationship, thoughsmall, was significant and positive (Schwenk andSchrader, 1993). The inconsistencies in findings,and the weak planning/performance relationshipsobserved have been key in the rejection of formalplanning as the one best way to plan(Mintzberg, 1994d).

One methodological explanation for the incon-sistencies and perhaps the most serious indictment

of early planning/performance research stemsfrom the poor conceptualizations and measure-ment protocols utilized to operationalize the plan-ning construct (Boyd, 1991). Crude dichotomousor trichotomous classifications of planningbehaviors were employed: comparing formal, longrange planners with non-formal, long range plan-ners (Thune and House, 1970) or comparing non-planners with incomplete planners and completeplanners (Kudla, 1980). Following the inconsis-tent and often counterintuitive findings emergingfrom the first two waves of planning/performance

research (Pearce et al., 1987) more sophisticatedGuttman scaling techniques (Guttman, 1944)were employed to measure the planning constructin the so-called third wave of planning/performance research (Pearce et al.,1987). These more sophisticated methodologies(see for example Wood and LaForge, 1979, 1981;Fredrickson and Mitchell, 1984) have in generalproduced stronger planning/performance relation-ships than the earlier work (Priem, Rasheed andKotulic, 1995).

A second substantive explanation for the incon-sistencies relates to the impact of environment onthe type of planning employed by firms. Somestudies (for example Eisenhardt, 1989; Goll andRasheed, 1997; Hart and Banbury, 1994; Millerand Cardinal, 1994; Miller and Friesen, 1983;Priem et al., 1995) found that formal strategymaking processes or planning are positivelyassociated with firm performance in unstable, tur-bulent or dynamic environments. Other studiesconcluded formal strategic planning is best suited

Copyright 1999 John Wiley & Sons, Ltd. Strat. Mgmt. J., 20: 889913 (1999)

to stable environments (Fredrickson, 1984; Fred-rickson and Mitchell, 1984; Fredrickson andIaquinto, 1989; Mintzberg, 1973) but recommendincrementalism for unstable, complex, dynamicenvironments facing high uncertainty, disconti-nuity and/or rapid change. The impact of environ-

ment on the type of planning firms should employin particular, and the environments effect on theplanning/performance relationship in general thusremains unclear.

This study utilized a measurement protocolconsistent with the third wave of planning/ per-formance research, and controlled for environ-mental stability in its design. However, unlikeFredrickson (1984), Fredrickson and Mitchell(1984) and Fredrickson and Iaquinto (1989), stra-tegic decision making activities or processesbased on hypothetical decision scenarios did not

measure firm planning activities. Rather, a widelyaccepted conceptualization of strategy articulatedin terms of ends and means formed the basis ofthe planning construct. The specificity of endsand means in firm level strategic planning proc-esses was the independent variable of interest.Based on this conceptualization, a research instru-ment measuring the range of planning behaviorsrecommended by Planners and Learners alike wasdeveloped, suitable for use across different indus-tries and firms.

Though this studys independent variable sca-

ling was based on the planning/learning polarity,strategy formation is more complex than thesetwo extremes represent. Other modes of strategyformation have been suggested (see for exampleChaffee, 1985; Bourgeios and Brodwin, 1984;Shrivastava and Grant, 1985; Mintzberg and Wat-ers, 1982), summarized and integrated by Hart(1992) into five distinct strategy making postures.Mintzbergs narration of ten schools of thoughteach describing different and discrete parts of thestrategy formation puzzle further underlines thiscomplexity (Mintzberg, 1990a). The planningversus learning dichotomy is reflected in HartsRational and Transactive modes and MintzbergsPlanning and Learning Schools of Thoughtrespectively.

Strategic ends and means

Both Chandler (1962) and Andrews (1971)emphasize ends and means in their conceptsof strategy. Hofer and Schendel (1978: 1819)


3/25

Learning to Plan and Planning to Learn 891

compared thirteen authors conceptualizations ofstrategy and strategy formulation, and allincluded ends and means, though defined in anumber of fashions and under a variety ofnames. Clearly, ends and means are distinctivethough related concepts widely used in the strat-

egy literature, worthy of separate treatment inboth research and practice (Dess, 1987; Schendeland Hofer, 1979). Ends relate to whatan organi-zation desires to achieve, while means relate tohow an organization intends achieving theseends. Ends describe specified desired futurestates organizations aspire to, while means, incontrast, are actionoriented (MacCrimmon,1988), typically defining the steps an organi-zation intends to take to achieve or reach itsdesired objectives or goals. Synonyms for endsinclude mission, purposes, goals or objectives,

while synonyms for means range from strategiesor policies, to alternatives, programs or actionplans, and even resource allocation activities ordecisions made by firms.

Ends with varying specificity are found in thestrategy formation process (Tosi and Carroll,1968), implemented at different levels withinorganizations (Anthony, 1965). Ends may be hier-archically ranked from broad, higher level ends,i.e., the grand design (Granger, 1964), mission,or strategic intent (Hamel and Prahalad, 1989)of an organization, to lower level, more limited

and specific operational objectives or goals(Granger, 1964; King and Cleland, 1987). Lowerlevel operational goals or objectives wereexpressly excluded from this study, as firmbehavior at the strategic, rather than the oper-ational level was the area of focus. In keepingwith the approach adopted by, inter alia, Chaffee(1985), Narayanan and Fahey (1982), and Mitroffand Emshoff (1979), strategic was presumed torelate to unstructured, non-programmable, non-routine problems, for which no predeterminedexplicit set of ordered responses existed in theorganization. At the strategic or higher level,strategic ends and means can include the defi-nition of business purpose or mission, formalizingthe firms major objectives or goals, determiningstrategies to achieve these objectives, and definingthe broad resource allocation commitments relatedto the strategies (Camillus, 1986; MacCrimmon,1988). Congruent with these conceptualizations,ends and means in the study were defined forrespondents as follows:


Ends: are the major, higher level purposes, mis-sion, goals or objectives set by organizations,each of which (should there be more than one)significantly influences the overall directionand viability of the firm concerned.

Means: are the patterns of action whichmarshal/allocate organizational resources intopostures that, once implemented, increase theprobability of attaining organizational ends.

To ensure respondent understanding, synonymsfor ends and means were included on the frontof the questionnaire as follows: Synonyms forends include a firms mission, purpose, goals, orstrategic objectives, while synonyms for meansinclude strategies, policies, alternatives, programsor action plans. Respondents were also cau-

tioned: Since the research focuses on strategicends or means, only major or important ends ormeans should be considered part of the study.

Ends and means specificity, firm

performance, and environments

Conceptualizations of strategic planning rangefrom the formalized processes of the SynopticModel (reflective of the Planning School) to thedisjointed processes of the Incremental Model(favored by the Learning School). According to

the Synoptic Model planning is a deliberate,rational, linear process (Chaffee, 1985) whereends are specified first, followed by means (Dess,1987; Fredrickson and Mitchell, 1984; and Schen-del and Hofer, 1979). Deliberate means emergefrom the strategy formation process fully speci-fied, ripe for implementation through detailedattention to objectives, programs, and operationalplans of ever increasing specificity (Mintzberg,1990b). Synoptic Formalism is considered bestsuited to predictable, stable contexts where uncer-tainty is low (Fredrickson, 1984; Mintzberg,1973, 1990b), as fully specified plans emergingfrom synoptic processes promote conception andthinking rather than learning, a vital process inunstable conditions (Mintzberg, 1990b). In stable,predictable contexts multiple, quantified, enduringends can be developed (Quinn, 1980) facilitatingcontrol in a tightly coupled system (Chaffee,1985).

In contrast, strategy formation according to theIncremental Model is an adaptive, incremental,


4/25


complex learning process (Lindblom, 1959;Mintzberg, 1978; Mintzberg, 1990a; Quinn, 1980)where ends and means are either specified si-multaneously, or are intertwined (Fredrickson andMitchell, 1984). Unlike the Synoptic approach,ends are rarely announced or recorded in a formal

planning document, and when they areannounced, they remain broad, general, and non-quantified (Quinn, 1980). Means, rather thanemerging from the planning process fully formedand ripe for implementation, develop and evolveover time as organizations learn from environ-mental interaction (Mintzberg, 1990a). In contrastto Synoptic Formalism, Incrementalism is rec-ommended for unstable, complex, dynamic con-texts with high uncertainty, or environments fac-ing discontinuity or change (Fredrickson andIaquinto, 1989; Fredrickson and Mitchell, 1984;

Mintzberg, 1990a). The increased uncertainty ofunstable environments requires less formalizationand more flexible, organic structures (Burns andStalker, 1961; Lawrence and Lorsch, 1967). Thearticulation of strategy in a formal plan isunsuited to unstable environments as such activi-ties implicitly assume conditions of stability orpredictability (Mintzberg, 1990b: 184), suggest-ing the following four hypotheses:

Hypothesis 1: There will be a positive

relationship between the specificity of ends

and performance among firms operating instable environments.

Hypothesis 2: There will be a negative


and performance among firms operating in

unstable environments.


relationship between the specificity of means


stable environments.




unstable environments.

Planning specificity and planning flexibility

Critics assert formal planning introduces rigidityand inflexibility into organizations (Bresser and


Bishop, 1983; Mintzberg, 1990a; 1990b; 1994a).Described as Plannings change pitfall(Mintzberg, 1994a) planning is feared to setorganizations on pre-determined, inflexible paths(Mintzberg, 1990b) thereby creating resistanceto change (Mintzberg, 1990a). If this rigidity

hypothesis (Hart and Banbury, 1994) holds,more specific plans should display greaterinflexibility, suggesting the following two hypoth-eses:


relationship between the specificity of ends and

planning flexibility among firms, regardless of

environmental stability.



and planning flexibility among firms, regard-less of environmental stability.

Since no prior work linked planning flexibilitywith environment, environment was not expectedto moderate the planning specificity/planningflexibility relationship.

Planning duration and the

planning/performance relationship

A lapse of time is required before the perform-

ance effects of a strategy or new strategy regimeare noted. Boyd (1991: 366) asserts that strategicdecision making today may not show up in thebalance sheet for several years, while Robinsonand Pearce (1983) contend strategic planningtypically improves economic performance by theend of a three- to five-year period. Bracker andPearson (1986) in one of the few studiesexplicitly controlling for planning duration, di-vided the planning history of the firms in theirsample into two periods: less than 5 years, and5 years or more. Significant performance differ-ences between the two periods were noted, withlong planning history firms outperforming shortplanning history firms by an average of 125%.Fulmer and Rue (1974) after finding no positiveplanning/performance relationship concluded thatas 50 percent of their sample had implementedplanning systems within two years of the study,insufficient time had passed for the benefits ofplanning to be enjoyed. In contrast, Gup andWhitehead (1989) found no statistically signifi-


5/25


cant relationship between the length of time bankshad engaged in planning, and their financial per-formance. Given the uncertainty surrounding theimpact of planning duration on theplanning/performance relationship, the followingtwo hypotheses were investigated:



and firm performance, regardless of plan-

ning duration.



and firm performance, regardless of plan-

ning duration.

RESEARCH METHODS

Measurement of variables

Ends and means specificity

A review of the literature revealed a need todevelop items and scales for the independent vari-ables of interest in the study. Though ends andmeans consensus has been measured in pastresearch (West and Schwenk, 1996; Bourgeois,1980) no investigation of ends and means speci-

ficity was discovered. Accordingly, five closed-ended Guttman type scales measuring ends speci-ficity, and four measuring means specificity weredeveloped. Statements ranging in choices fromunspecified to very specific were presented, and inevery scale but one respondents indicated whichone statement/item best described their firm (seeAppendix for scales and the scoring protocolsemployed). Individual scale scores were summedto obtain the overall ends and means specificityscores. Scale statements were constructed to capturethe differing properties of ends and means as

characterized by the Synoptic and IncrementalModels respectively, with low specificity scoresindicating ends and means typical of the Incremen-tal Model, and high specificity scores denoting endsand means characteristic of the Synoptic Model.

Accordingly, organizations with very specificends would possess many, precisely quantified,formally documented, time-limited ends, rangingfrom a statement of firm mission to statements ofspecific market share/sales growth targets and


other key result areas. Very specific means wouldbe reflected in plans that set out exact plans and/orprograms for implementation, describing in detailthe actions and steps required for implementation.These specific means would be used to direct firmaction and behavior and measure timely perform-

ance against plan. They would also be formallydocumented and distributed among firm members.Conversely, few broad ends that change and evolveas conditions dictate would characterize less speci-fic ends, while unspecific means would be broadand unstructured, evolving as circumstances war-rant and acting as loose guides only. Such unspeci-fic ends and means would rarely be announced,and if so, in broad terms.1

Environmental stability

Prior research has employed two principalmethods to categorize environmental stability.Lawrence and Lorsch (1967) used subjectiverespondent ratings to categorize environmentalvolatility, while Tosi, Aldag, and Storey (1973)developed three objective measures of environ-mental volatility (market, technological, and earn-ings volatility). Other objective measures of vola-tility include Dess (1980), who used multipleobjective measures such as the degree of changein industry sales, cost/price margins, and valueadded as criteria to measure the

stability/instability of industries. Hart and Ban-bury (1994), following Dess and Beard ( 1984),characterized firm competitive environmentsacross four dimensions: change, unpredictability,complexity, and munificence. Because of themulti-industry, multinational nature of the sample,objective indicators of environmental stabilitywere inappropriate. Respondents self-selected oneof four industry/business environments that most

1

Though we measure a narrower construct than the strategicplanning construct validated by Boyd and Reuning-Elliot(1998) we include five of the seven key planning indicatorsof strategic planning identified by Boyd and Reuning-Elliot(1998). These five planning indicators are mission statement,long term goals, annual goals, short term action plans, andongoing evaluation. Unlike the Boyd and Reuning-Elliotmeasurement model, the utility of ours is not restricted tomeasuring formal planning only. We measure both formaland incremental planning. Boyd and Reuning-Elliot doacknowledge this limitation in their model. As recommendedby Boyd and Reuning-Elliot, our model is parsimonious, usesindicators which are well grounded in theory, and representsdifferent aspects of the strategic planning process (ibid.: 188).


6/25


Table 1. Respondent self-categorization of industry/business environment stability (n = 426)

Industry/Business Environment Number Group ISselecting score meancategory

Group 1: A mature, stable industry or context, where change is relatively 104 33.72

little/slow, the technology is mostly stable and known, and in general trendscan be forecasted. Though some uncertainties exist, stability and predictabilityare more present than change. New entrants into the industry are infrequent,as are exits of current competitors. Both the rules of competition, and thepositions of current competitors are reasonably well established, and stable.(Choice 1)

Group 2: A mature industry or context which in the past was stable, but is 273 44.58currently experiencing considerable change and realignment, due to factors suchas deregulation, technological change, or new entrants of powerful competitors.Industry rationalization and realignment are now occurring, under conditions ofchange, uncertainty, and discontinuity. The rules of competition, and the positionof current competitors, are changing considerably. (Choice 2)

Group 3: An industry or context where technological change is rapid, and/or 36 53.53

product obsolescence is quick, and/or competition to innovate is very high.Trends are difficult to forecast, and uncertainty is high. Most decisions arecomplex, made with imperfect information, under conditions of rapid,continual change. Though some established competitors exist, remainingcompetition is challenging, and competitive rivalry is intense. (Choice 3)

Group 4: A young industry in which the rules of competition are currently 13 49.77emerging and being established. Though industry leaders exist, technologyand the rules of competition are still developing, and uncertainty is high.Trends are difficult to forecast, and uncertainty is high. Most decisions arecomplex, made with imperfect information, under conditions of rapid,continual change. (Choice 4)

Note: Scores for Groups 1, 2 and 3 differed at the 0.05 level. Scores for Groups 2 and 4 and Groups 3 and 4 did not differat the 0.05 level.

closely described their firms industry/businessenvironment. Each choice described a differentenvironment, ranging from mature and stable toyoung and highly unstable. Table 1 presents thefour choices, and the number of respondents se-lecting each.

Since the categorization of environmental sta-bility was vital to the study, additional data weregathered to validate that the four choices rep-resented environments with clearly different sta-bility conditions. In another section of the ques-tionnaire respondents rated the stability of theirfirms primary industry across ten industry sta-bility factors, using a seven point rating scale.These industry stability factors are reported inthe Appendix. To complete the validation ofenvironmental stability, respondents were alsoasked to indicate which word of two in threepairs best described the business environment orindustry of their firm: slow or rapid change;


stable or unstable; and predictable or unpredict-able. Where slow change, stable, or predictablewas chosen, a score of 1 was given. Where rapidchange, unstable, or unpredictable was chosen, ascore of 1 was given. A third, not applicablechoice was also included for each pair, scored 0when checked.

The industry stability ratings and the respective1, 0, 1 scores received for the three pairs ofwords were summed to obtain Industry Stabilityscores (IS Scores) for each firm in the sample.2

The IS Score mean of firms in each environmentcategory was then calculated (reported in column3 of Table 1) and compared to the meansobtained for the other three choices. Each meanwas expected to significantly differ from the other

2 Regressing the sum of the three word pair scores ( 1, 0,1) on the IS Scores excluding the sum of the word pairscores produced a strong correlation with r = 0.55, p 0.001.


7/25


three, in the following fashion: lowest for theMature/Stable Context (Choice 1), and increasingsequentially for the next three choices.

Firm performance

Similar to the measurement of environmentalstability, objective and subjective measures havecalibrated firm performance in planning/performance research. Objective criteria includesales growth, in either gross or net terms(Fredrickson and Mitchell, 1984; Leontiades andTezel, 1980; Pearce et al., 1987), growth in netincome (Wood and LaForge, 1979), return onassets/sales (Fredrickson and Mitchell, 1984;Leontiades and Tezel, 1980; Pearce et al., 1987),abnormal stockholder returns (Kudla, 1980), stockprice performance (Ansoffet al., 1970), dividend

per share or earnings per share (Ansoff et al.,1970), or total return to investors (Rhyne, 1987).Subjective criteria include respondent ratings ofplanning system effectiveness (Chakravarthy,1987), respondent rating in comparison to thefirms overall industry (Pearce et al., 1987), orrespondent perceptions of their firms currentprofitability, growth/share, future positioning,quality, and social responsiveness (Hart and Ban-bury, 1994). No one criterion or set of criteriadominates, with the outcome variables selectedmostly reflecting the preferences of the

researchers involved.Because of the multi-industry, multinational

characteristic of the sample, and the inability toreliably control for industry effects and/or othercompetitive groupings, objective measures of fi-nancial or business performance were unsuitable.Following and based on Pearce et al. (1987) andChakravarthy (1987) two subjective perceptualmeasures of performance were developed: OverallFirm Performance (OFP), and Planning Perform-ance (PP). OFP captured overall firm performancedata relative to competitive peers and firm marketshare change over the past 510 years (an exter-nally anchored performance measure) while PPcaptured the planning performance of respondentfirms (an internally anchored performance meas-ure evaluating planning capabilities andeffectiveness). The OFP and PP scales aredetailed in the Appendix. In the OFP scales, aNot Applicable box was included for firms thatdid not have listed securities, or fordivisions/business units that did not have any


relevant stock performance to rate. When thisbox was checked, the average of the other twofactors was used as a proxy for stock priceperformance. OFP scores were obtained by sum-ming the three industry peer comparison scoreswith the market share change score. PP scores

were obtained by summing the six individualitem ratings provided by respondents.

Planning flexibility and planning duration

No prior research measuring planning flexibilitywas located. Accordingly, and similar to thescales measuring Ends and Means Specificity, aGuttman type scale was developed to measurethe construct (see Appendix). To permit facevalidation of the flexibility scales respondents alsoindicated the flexibility of their firms mission or

business purpose using the same scale statements.Since firm mission or business purpose is suppos-edly the most stable or enduring of strategic ends(Hamel and Prahalad, 1989; Granger, 1964), andsince means should be more flexible than ends(Hamel and Prahalad, 1989), the average flexi-bility scores of the three sets of flexibility datawere expected to rank accordingly.3

Since no particular method was favored in pastresearch to measure planning duration, a morefine-grained categorization of planning durationthan previously employed was utilized. Respon-

dents reported the period of time their firms hadbeen conducting the planning as indicated in thequestionnaire in one of five choices: Less thanone year; 13 years; 47 years; 815 years; andmore than 15 years. Comparing data obtained onfirms with a planning duration of three years orfewer (choices 1 and 2) with firms reporting aplanning duration of 4 years or more (choices 3to 5) tested the impact of planning duration.

3 We must recognize this journals anonymous reviewer whopointed out that though we label our construct PlanningFlexibility we are really measuring the change in or alterationof plans. Since words such as flexibility or rigidity arepervasive in the literature, we adopted Planning Flexibility.Furthermore, and pursuant to the same reviewers comments,we must acknowledge that we obtain only average measuresof flexibility. As illustrated in the three flexibility scoresgathered in the study, flexibility does differ depending onthe type and level of the ends or means involved. Since lowerlevel ends or means were excluded from the study, averagemeasures of ends and means flexibility were sufficient thoughadmittedly coarse-grained.


8/25


Hypothesis testing

The study investigated three contingent relation-ships: that the form and direction of theplanning/performance relationship is contingentupon environmental stability; that the

planning/performance relationship is contingentupon planning duration; and that the planningspecificity/planning flexibility relationship is notcontingent upon environmental conditions. Con-tingency relationships may be tested by two tech-niques: moderated regression, and sub-groupanalysis (Li and Simerly, 1998). Since the formand direction of the relationships were underinvestigation, moderated regression with dummyvariables was utilized (Aiken and West, 1991).Sub-group analysis is more appropriate when thestrength of the relationship is at issue

(Venkatraman, 1989). However, where appropri-ate, t-tests or one way ANOVAs were alsoperformed, with the moderator as the independent(grouping) variable and the five key study vari-ables (Ends Specificity, Means Specificity, OFP,PP, and IS Scores) as dependent variables.Implicitly, the null hypothesis that the inde-pendent (grouping) sample means were all equalwas tested against the alternative hypothesis thatthe means were not all equal. When the ANOVAprovided a basis to reject the overall null hypoth-esis and accept the alternative hypothesis, a post-

hoc multiple comparison of means wasperformed. Following Winer (1971) and Brackerand Pearson (1986) the Scheffe test was selectedto conduct the post-hoc multiple comparisons.The lack of any significant difference in thedependent variable means was considered indi-cation of the lack of moderator influence on thevariables involved.

Sampling procedures and sample

characteristics

Senior and mid-level executives attending 39 edu-cational programs offered at three businessschools and senior executives affiliated with twoother private organizations provided data for thestudy.4 While not as fine grained as interpretive

4 Details on the breakdown of number of responses obtainedper program are available from the first author. No singleprogram contributed more than 9% of the sample and onlytwo contributed more than 5%. One of the private organi-zations was an industry body for the hydraulic pump industry,


methods (Eisenhardt, 1989) or scenario basedmethods (Fredrickson, 1984) the survey methodwas employed to obtain a large sample, oftenable to complement or clarify more interpretivemethods (Lee, 1991). Though a convenience sam-ple, the sample size and diversity add confidence

and credibility to the study findings.Respondents were either mailed questionnairesbefore their program and asked to bring the ques-tionnaire with them, or handed questionnairesonce in attendance and verbally requested to com-plete and hand back the instrument by the endof the program, always at least one week ahead.At distribution, three points were stressed: thatparticipation was entirely voluntary; that strategyformation at the business unit, divisional, or SBUlevel was the focus of the study, and ideal respon-dents were thus members of a firm/business

unit/division who were familiar with the strategicplans and strategy formation processes utilizedby their firm/business unit/division; and that therewere no correct or incorrect answers to any ofthe questionsthe desired response to each ques-tion was the one which best described the prac-tices or situation in the respondentsfirm/business unit/division.

Additional data were gathered to control forfirm age and size, respondent employment tenure,respondent country of residence, and respondentmanagerial level according to business title. Firm

age and size were employed to quantify the sam-ple characteristics, while respondent tenure wasused to exclude responses from unreliable respon-dents. Grouping by respondent managerial levelwas used to test whether managerial level impac-ted the study findings, following work that hasfound middle management involvement in strat-egy formation improves organizational perform-ance (Wooldridge and Floyd, 1990; Floyd andWooldridge, 1992). Country of residence was

who provided access to its address list, and the other privateorganization was The Executive Council (TEC), an inter-national organization of CEOs of companies from differentindustries who meet periodically to discuss business issues.24 cases in the sample are from these two sources. Thehydraulic pump industry was approached to gather responsesfrom a single industry, and TEC was included to increase thenumber of small and medium size enterprises in the sample.Exclusion of the 24 cases from the sample has no noticeableeffect on the study findings. The response rate from thehydraulic pump industry was insufficient to permit any sta-tistical testing of the research instrument or findings, for eitherinstrument validation purposes, or for investigation of anyspecific industry effects.


9/25


used to validate the study findings across differ-ent countries.

In total, 668 responses were collected. Ques-tionnaires from respondents at their firms for ayear or less were excluded, reducing the sampleto 656. Apart from these twelve cases, respondent

employment tenure was impressive: 89 percenthad been at their firm four years or longer, and73 percent indicated tenures of longer than sevenyears. Fifty percent reported more than 10 yearsof service. The exclusion of firms with a planningduration of three years or less (choice 1 or 2 inthe planning duration categorization) reduced thefinal sample size by 230 to 426 cases.

Firms from 20 countries are represented,though 47 percent are United States based, and44 percent South African in origin. Though firmsize and age varies, the findings are most relevant

to large, mature organizations in business for along period of time73 percent of firmsemployed more than 500 and 59 percentemployed more than 1000. 95 percent were morethan seven years old, with 92 percent being olderthan ten years. A wide range of industries is alsorepresented, indicated in Table 2 below.

The IS Score mean per industry category isalso included in Table 2. The relative values and

Table 2. Sample industry representation (n = 425)

Industry Categorization Count IS Score Percent of mean all Cases*

Wholesaling and/or retailing 34 41.29 8Mining and/or metals extraction 20 37.75 5Heavy manufacturing 63 43.08 15Light manufacturing 71 42.13 17Financial services: banking 22 46.91 5Financial services: insurance 27 41.85 6Financial services: other 12 43.75 3Engineering 18 43.89 4Consulting services 8 49.13 2Other services 27 42.11 6Commercial/industrial construction 11 41.27 3Residential construction 5 47.00 1Agriculture or farming 4 36.75 1Travel and tourism 5 34.80 1Pump manufacturing 8 36.75 2Telecommunications 19 50.53 5Railroad Transportation 7 39.29 2Utilities 8 37.13 2Other 56 44.43 13

One respondent failed to report industry data.*Totals more than 100, due to rounding.


range of these means adds face validity to the ISScore methodology: telecommunications has thehighest (50.53, n = 19) while more stable indus-tries such as mining or metals extraction, tourism,or utilities have the lowest means. Owing to therelatively few respondents who selected environ-

ment Choices 3 or 4 as indicated in Table 1 (n= 36 and 13 respectively) these were merged intoa single, high instability category. The similarityin IS Score mean of Choices 3 and 4 madecombining these two choices into a single VeryUnstable category reasonable. The IS Score meanof the combined Very Unstable group was 52.53(see Table 5), significantly different from the ISScore mean of the other two groups. A trifurcatedcategorization of environmental stability was thusutilized: Mature/Stable; Mature/Unstable; andVery Unstable.

The data on managerial level permitted divisionof the sample into three groups: Top Management(n = 65) which included Chairman, CEO, COO,President, CFO, and Managing Director/DeputyManaging Director (mainly for South Africa,where Managing Director is equivalent toPresident); Senior Management (n = 137) whichincluded EVP/SVP, SBU Director, DivisionalManager, Partner, Director, Principal, and General


10/25


Manager/Assistant GM (mainly for SouthAfrica); and Middle Management (n = 191)which included Vice President, Sales Manager,Manager, and Assistant Vice President. Broadlyspeaking, almost half of the sample wereTop/Senior management (n = 202). Thirty-three

respondent titles could not be easily categorizedinto the three management groups. Moderatedregression and one way ANOVA with man-agement level as the dummy/grouping variableindicated management level did not moderate orimpact any of the study findings.5

Construct Validity

Cronbachs Alpha was calculated for each vari-able measured by multi-item scales. The com-puted alpha coefficients (reported in Table 3)

range from 0.77 to 0.93, indicating adequatereliability and internal consistency. The threeplanning flexibility score means formission/business purpose, ends in general, andmeans in general ranked as expected (MissionFlexibility Score mean = 4.22; Ends FlexibilityScore mean = 3.42; and Means Flexibility Scoremean = 2.66). Each of these differed from theother two at p 0.001.6

RESULTS

Descriptive statistics and Pearson correlations arepresented in Table 3. Four base models reflectingthe planning specificity/firm performance relation-ships for the entire sample net of moderatoreffects are included for comparative purposes. Allfour models are significant, and each planningspecificity/performance beta coefficient is sig-nificant and positive.7

5 Details of this analysis are available from the study authors.6

Some firms did not possess a specific missionstatement/business purpose, reducing the sub-sample used totest whether the three flexibility scores differed to n = 367.This explains why the average scores for Ends and MeansFlexibility reported here differs slightly from the data reportedin Table 3, where n = 426.7 The models differ systematically across beta strengths andexplanatory power, contingent upon the dependent variable.That the planning specificity/OFP models have weaker betasand less impressive R2s than the planning specificity/PP mod-els is unsurprising, given that many factors other than planningspecificity influence external firm performance. Since planningspecificity captures many dimensions of firm planning,stronger betas and greater explanatory power are to be


Table 4 reports the regression results testingHypotheses 14, first with Ends Specificity asthe independent variable, followed by MeansSpecificity.

All four models are significant, and in eachcase 3 ( the planning specificity/performance

coefficient for Very Unstable environments) issignificant and positive, ranging from 0.38 to0.43. Particularly remarkable are the strong plan-ning specificity/OFP relationships noted in Mod-els I and III, both of which are stronger thanthe planning specificity/PP relationships noted inModels II and IV, and the average relationshipsreported in the planning specificity/OFP BaseModels in Table 3. In addition, no significantnegative relationships are noted in the fourth andfifth terms of any model, which test the moderat-ing effects of environment on the

planning/performance relationships. Accordingly,Hypotheses 1 and 3 are accepted, and Hypotheses2 and 4 are rejected. However, in Models IIand IV significant 4s indicate that the planningspecificity/PP relationships are moderated byenvironmental stability, though not in the case ofMature/Unstable environments. Though remain-ing positive throughout, the planningspecificity/PP associations are strongest inMature/Stable Environments, and not significantin Mature/Unstable environments.8

One way ANOVA with environment as the

independent (grouping) variable indicated that theES Score, MS Score, PP Score and IS Scoremeans were statistically different. The result ofthis analysis is reported in Table 5.

Table 5 indicates that the ES Score mean forfirms operating in Mature/Stable environments isthe lowest, and differs significantly from the meanof firms operating in Mature/Unstable environ-ments. A similar pattern is displayed across MSScore means. In addition, the PP Score meansgrow with environmental instability, and the high

expected. That the performance variable specification somaterially effects the planning/performance relationshipstrength and model explanatory power emphasizes the impor-tance of outcome variable selection in this line of research.Using external or internal measures of performance aloneunderrates the complexity of the planning/performancerelationship. Similar patterns are noted in the data after con-trolling for the country of origin of respondents (UnitedStates, n = 202; South Africa, n = 182).8 The pattern of findings reflected in Table 4 are replicatedin the data after controlling for country of origin (UnitedStates, n = 202; South Africa, n = 182).


11/25


Table 3. Descriptive statistics, Pearson correlations, and base models ( n = 426)

Variables Mean SD CA* 1 2 3 4 5 6 71. Ends Specificity 16.83 5.06 0.77 1.002. Means Specificity 14.81 4.11 0.83 0.60 1.003. Overall Firm Performance 13.77 3.43 0.81 0.33 0.36 1.004. Planning Performance 28.00 7.05 0.93 0.56 0.66 0.52 1.00

5. Environmental Stability 42.84 1 0.10 0.80 0.20 0.16 0.01n 0.17 1.006. Ends Flexibility 3.33 1.26 0.01n 0.06n 0.01n 0.01n 0.01n 1.007. Means Flexibility 2.58 1.30 0.01n 0.18 0.01n 0.08n 0.00n 0.36 1.00

Correlations are significant at p 0.05 unless accompanied by an n, indicating not significant.*Cronbachs Alpha.

Base models: Planning specificity/firm performance regressions (n = 426)

Dependent VariablesOverall Firm Performance Planning Performance

Independent Variable:Ends Specificity t t

(Constant) 18.03*** 15.12***Ends Specificity 0.33*** 7.29*** 0.56*** 13.85***R2 0.11 0.31Adjusted R2 0.11 0.31F(1,424) 53.14*** 191.71***

Base models: planning specificity/firm performance regressions (n = 426)

Dependent VariablesOverall Firm Performance Planning Performance

Independent Variable:Means Specificity t t

(Constant) 15.82*** 11.66***Means Specificity 0.36*** 7.94*** 0.66*** 18.18***R2 0.13 0.44Adjusted R2 0.13 0.44F(1,424) 63.11*** 331.13***

Beta coefficients are standardized.*p = 0.05; **p = 0.01; ***p = 0.001

PP Score mean noted among firms in VeryUnstable environments differs significantly fromthe PP Score means of firms operating in the

other two environments. As expected, the ISScore means differ.9

9 Controlling for country of origin, the USA IS Score meanwas higher than the South African mean, and the SouthAfrican OFP mean was higher than the USA mean. Thehigher USA IS Score mean is understandable, given that theUSA is probably the most competitive business environmentin the world today, while South Africa is emerging from thedecades of isolation and siege during apartheid, which causeda decline in competition in general in the country. Thedifference in OFP means is difficult to explain.


Table 6 reports the regression results testingHypotheses 5 and 6.

Model V is significant, and provides some

interesting findings. The coefficient for EndsSpecificity surprisingly is negative (3 = 0.44,t = 2.94, p 0.01) indicating that in VeryUnstable environments specific ends areassociated with greater flexibility.10 Moreover, thesignificant coefficients for 4 and 5 indicate that

10 Planning Flexibility is scaled so higher scores indicategreater inflexibilitya negative coefficient means high speci-ficity with high flexibility.


12/25


Table 4. Planning specificity/ firm performance relationships, moderated by environmental stability

Dependent Variables

Overall Firm Performance Planning Performance(Model I) (Model II)

Independent Variable:

Ends Specificity t t

(Constant) 5.52*** 6.92***Mature/Stable: Dummy1 0.32 1.28 0.59** 2.70**Mature/Unstable: Dummy2 0.11 0.41 0.42 1.83Ends Specificity 0.43** 3.06** 0.38** 3.10**Ends Specificity by Mature/Stable 0.33 1.40 0.44* 2.10*Ends Specificity by Mature/Unstable 0.06 0.20 0.26 1.07R2 0.14 0.33Adjusted R2 0.13 0.33F(5,420) 14.03*** 42.10***

Dependent Variables

Overall Firm Performance Planning Performance(Model III) (Model IV)


(Constant) 4.58*** 6.23***Mature/Stable: Dummy1 0.23 0.84 0.71** 3.16**Mature/Unstable: Dummy2 0.14 0.49 0.63** 2.66**Means Specificity 0.43** 2.80** 0.41*** 3.30***Means Specificity by Mature/Stable 0.26 0.94 0.55* 2.51*Means Specificity by Mature/Unstable 0.03 0.09 0.47 1.86

R2

0.16 0.46Adjusted R2 0.15 0.46F(5,420) 15.96*** 72.75***


environment moderates the ends specificity/endsflexibility relationship. In both mature environ-ments (Mature/Stable and Mature/Unstable) endsspecificity is positively correlated with endsinflexibility, though less strongly so inMature/Unstable than in Mature/Stableenvironmentsas environmental instabilitygrows, so does ends flexibility. Model VI has nosignificant regression coefficients indicating thatMeans Specificity is neither related to MeansFlexibility nor moderated by environment.Hypotheses 5 and 6 are accordingly rejected. Thedata clearly refute that planning specificity byitself is associated with planning inflexibility orrigidity. Rather, planning specificity is more


associated with planning inflexibility in stableenvironments, with this association reducing inboth strength and direction (for Ends Specificity)as environmental instability grows.11

Table 7 reports the regression results testingHypotheses 7 and 8.

11 Calculating Planning Specificity/Planning Flexibility Pear-son correlations and controlling for environment display asimilar picture. The strongest positive specificity/flexibilitycorrelations are noted among firms operating in Mature/Stableenvironments (ES/EF r = 0.17, p 0.1; MS/MF r = 0.30,

p 0.001, n = 104) and the weakest/most negative corre-lations are noted in the Very Unstable environments (ES/EFr = 0.42, p 0.05; MS/MF r = 0.00, not significant, n =49). The Mature/Unstable correlations lie in between (ES/EFr= 0.01, not significant; MS/MF r= 0.15, p 0.05,n = 243).


13/25


Table 5. Key study variables and environmental stability (n = 426)

Summary Table of Means

Group ES Score MS Score PP OFP IS Score

Group 1: Mature/Stable 15.65 13.91 26.83 14.18 33.72

Group 2: Mature/Unstable 17.18 15.12 27.96 13.46 44.58Group 3: Very Unstable 17.36 14.96 30.73 14.63 52.53All Groups 16.83 14.81 28.00 13.77 42.84

Analysis of Variance

Variable F pES Score 3.79 0.02MS Score 3.37 0.04PP 5.24 0.01OFP 3.44 0.03IS Score 101.61 0.001

Using the Scheffe Test, the Group 1 ES Score mean differed significantly from the Group 2 ES Score mean, and the Group1 MS Score mean differed significantly from the Group 2 MS Score mean. The Group 3 PP Score mean differed significantlyfrom the other two PP Score means.All IS Score means differed significantly.

All four models in Table 7 are significant, andexamination of the regression coefficients in eachequation clearly illustrates the impact of PlanningDuration. In Models VII and IX, 2 is insignifi-

cant, showing that when planning had beenunderway for three years or less, Overall FirmPerformance is not associated with Ends or MeansSpecificity. The strong positive coefficients forEnds and Means Specificity by Planning Durationof 4 years or more (Model VII: 3 = 0.40, t =2.58, p 0.01; Model IX: 3 = 0.41, t =2.37, p 0.05) indicates that OFP is positivelyassociated with planning specificity once planninghas been underway for at least four years. ModelsVIII and X display a different pattern. The sig-nificance of 2 in both indicates Planning Per-

formance is associated with Ends or Means Speci-ficity, and the insignificance of 3 in bothindicates planning duration does not moderate the

In the case of Model VI, though the Means Specificity/MeansFlexibility correlations are significant for both Mature/Stableand Mature/Unstable environments when calculated individu-ally, combining these with the lack of correlation noted inVery Unstable environments (r = R2 = 0) swamp the signifi-cant individual correlations. Similar to Table 4, the patternof findings reflected in Table 6 are replicated in the data aftercontrolling for country of origin.


planning specificity/PP relationship. Accordingly,Hypotheses 7 and 8 are partially rejected. Plan-ning duration clearly moderates the planning/OFPperformance relationship, and from an external

performance perspective excluding firms with aplanning duration of three years or less is valid.Planning duration does not significantly impactthe planning specificity/planning performancerelationship.12

One way ANOVA using Planning Duration asthe independent (grouping) variable indicated thatthe means of two of the five variables of interest(OFP and PP) differed significantly. The ScheffeTest confirmed that the Group 1 PP Score meandiffered significantly from the Group 5 PP Scoremean, while the Group 1 OFP Score mean dif-

fered significantly from all other Group OFPScore means. These results are reported inTable 8.

12 The data testing planning durations impact on the relation-ships and variables of interest in the study include the 230cases initially excluded from the 426 case main sample.


14/25


Table 6. Planning specificity/planning flexibilityrelationship, moderated by environmental stability

Dependent Variable

Ends Flexibility(Model V)

Independent Variable:Ends Specificity t

(Constant) 7.79***Mature/Stable: Dummy1 0.90*** 3.40***Mature/Unstable: Dummy2 0.71* 2.56*Ends Specificity 0.44** 2.94**Ends Specificity by 0.86*** 3.43***

Mature/StableEnds Specificity by 0.81** 2.77**

Mature/UnstableR2 0.03Adjusted R2 0.02F(5,420) 2.61*

Dependent Variable

Means Flexibility(Model VI)

Independent Variable:Means Specificity t

(Constant) 3.12**Mature/Stable: Dummy1 0.38 1.28Mature/Unstable: Dummy2 0.25 0.78Means Specificity 0.00 0.00

Means Specificity by 0.44 1.51Mature/StableMeans Specificity by 0.30 0.88

Mature/UnstableR2 0.04Adjusted R2 0.03F(5,420) 3.32

Beta coefficients shown are standardized.*p = 0.05; **p = 0.01; ***p = 0.001

LIMITATIONS OF THE STUDY

Though the study has some interesting and valu-able findings, some limitations must be acknowl-edged. First, the study is based upon a singleresponse per firm/division/business. Though thescales are valid and reliable between firms, multi-ple respondents from the same firm would permitwithin firm validation of the study scales. Second,subjective perceptual data are used to measurethe variables of interest in the study. This is ofmost concern in the measurement of the inde-


pendent variable, firm level planning. Previousresearch has confirmed subjective measures ofperformance typically correlate strongly with sec-ondary objective measures (Dess and Beard,1984; Dess, 1987; Hart and Banbury, 1994; Ven-katraman and Ramanujam, 1987) and subjective

measures of industry volatility have been foundto correlate with other objective measures (Snyderand Glueck, 1982). Furthermore, the significantlydifferent IS Score means per environmental cate-gory (Mature/Stable; Mature/Unstable; and VeryUnstable) and the range of industry based ISScore means reflected in Table 2 provideadditional confirmation that industry/environ-mental stability has been appropriately measured.In addition, the significant and as predicted differ-ences in the three measurements of planningflexibility validate the flexibility measurement.

Regarding the subjective measurement of Endsand Means Specificity, content analysis of actualplanning documents may have provided a moreaccurate picture of the specificity of plans emerg-ing from respondent firms. However, such a meth-odology is impractical over a large sample, andexcludes firms with no plans at all. The thirdlimitation relates to the limited number of smallfirms represented in the sample. To confirm theapplicability of the findings to smaller firms, moresmall companies must be included in the sample.Finally, the use of cross sectional data to study

an essentially longitudinal construct (the planningprocesses inherent in the notion of planningduration) must be acknowledged as a limitation.

DISCUSSION OF FINDINGS

Regarding environments,

planning/performance, and the planning

school/ learning school debate

This study concludes environment does not mod-erate the type of planning firms pursue. Externalfirm performance (OFP) and internal planningperformance (PP) are clearly associated with for-mal, specific planning, regardless of environment.Furthermore, the key planning prescriptions ofthe Incremental Model received no support. Inunstable environments, relying on a few broad,unspecified, unannounced ends, and leavingmeans to emerge as an organization interacts withits environment is inadvisable. Multiple specificends, and specific, deliberate means are prefer-


15/25


Table 7. Planning specificity/ firm performance relationships, moderated by planning duration

Dependent Variables

Overall Firm Performance Planning Performance(Model VII) (Model VIII)

Independent Variable:

Ends Specificity t t

(Constant) 12.73*** 10.86***Plandum: Dummy1 0.24 1.69 0.10 0.77Ends Specificity 0.09 1.32 0.45*** 7.30***Ends Specificity by Duration 4 yrs 0.40** 2.58** 0.22 1.60R2 0.08 0.29Adjusted R2 0.08 0.29F(3,652) 18.88*** 88.93***

Dependent Variables

Overall Firm Performance Planning Performance(Model IX) (Model X)


(Constant) 11.02*** 8.83***Plandum: Dummy1 0.26 1.66 0.19 1.00Means Specificity 0.10 1.70 0.55*** 9.34***Means Specificity by Duration 4 yrs 0.41* 2.37* 0.22 1.59R2 0.09 0.40Adjusted R2 0.09 0.40F(3,652) 22.10*** 147.48***


able. The findings do not, however, preclude theuse of incrementalism in strategy formation. Clearevidence of the co-existence of formal, specific,andflexible planning was noted in Very Unstableenvironments: plans can be both specific andflexible. In fact, sound specific planning before-hand may limit the amount of incrementalism(and learning) a firm faces later, while firmsoperating without any specific planning mayspend more time in experimentation and trial anderror, thus affecting performance. Instead of beingthe antithesis of incrementalism, formal specificplanning may be a necessary precursor to success-ful incrementalism. Stated in the terms of a strongsupporter of the Incremental Model (Mintzberg,1987) specific plans may represent the intendedstrategy while the inevitable incremental changesthat follow as intentions become reality represent


the emergent, or realized part of the firmsdeliberate strategy. Both are necessary andneither is sufficient. Such integration has beensuggested before, both within the field of strategy(Camillus, 1982; Goold, 1992) and in other fields(Michael, 1973).

The strong planning/performance relationshipsnoted in this study support recent studies thatreport a positive relationship between formalplanning processes and performance (Eisenhardt,1989; Goll and Rasheed, 1997; Hart and Banbury,1994; Priem et al., 1995). With the planningconstruct adequately specified and measured,strong and consistent planning/performancerelationships are noted. However, the positiverelationships noted in Very Unstable environ-ments are inconsistent with the findings of Fred-rickson and his colleagues (Fredrickson, 1984;


16/25


Table 8. Key study variables and planning duration (n = 647)

Summary Table of Means

Group* ES Score MS Score PP OFP IS Score

Group 1: 1 year (n = 26) 14.65 13.50 24.19 10.60 41.00

Group 2: 13 yrs (n = 204) 17.54 14.83 27.22 13.29 44.20Group 3: 47 yrs (n = 243) 17.09 15.13 28.17 13.65 43.08Group 4: 815 yrs (n = 114) 16.91 14.79 28.09 13.91 42.88Group 5: 15 yrs (n = 60) 16.63 15.47 29.18 14.30 42.73All Groups 17.06 14.94 27.79 13.52 43.28

*9 respondents indicated 0 for Planning Duration as their firms possessed No specifically developed strategic plans of any

substance and thus were prompted to skip the remainder of Section 2 (measuring Strategic Means Specificity and Planning

Duration) and go onto Section 3 of the questionnaire. Since all these firms had been in business for longer than 4 years, they

were included in the main study sample ( n = 426) but excluded from the analysis reported in this table.

Analysis of Variance

Variable F p

ES Score 2.35 0.05MS Score 1.65 0.16PP 3.41 0.01OFP 5.87 0.00IS Score 0.87 0.48

Using the Scheffe Test, the Group 1 PP Score mean differed significantly from the Group 5 PP Score mean, and the Group1 OFP Score mean differed significantly from all other group means.

Fredrickson and Mitchell, 1984; Fredrickson andIaquinto, 1989). The inconsistency could be due

to many factors including different researchdesigns and conceptualizations of the planningconstruct, or different measurement ofindustry/environmental volatility. Alternatively,the findings of this study (and others contra-dicting the Fredrickson studies) may fail to cap-ture the true nature of the relationships underinvestigation. However, after almost fifteen years,the Fredrickson studies remain to be replicatedoutside the firms and industries of the originalstudies (Priem et al., 1995) and contrary resultshave been noted in a similar study of three otherindustries (Judge and Miller, 1991). In contrast,the Planning School prescriptions have now beenthe subject of considerable investigation overmany years, and are receiving growing supportfrom this and other recent work. The LearningSchool claims would benefit from similar inten-sive scrutiny.

The integration of synoptic formalism andincrementalism suggested in this paper questionsthe value of the recent Ansoff/Mintzberg debate,


and indeed the general condemnation of formalstrategic planning as a desirable firm behavior

(Mintzberg, 1991; Mintzberg, 1994c). If anything,the dissatisfaction with formal strategic planninghas surfaced the practices to be avoided in plan-ning, rather than providing support for the propo-sition that the remedy for bad planning is noplanning. For example, the first of Mintzbergsplanning fallacies (detachment) is avoided byplacing planning responsibility in the hands ofengaged line managers. The second fallacy (pre-determination) is avoided by acknowledging, asthis paper points out, that good planning is aboutboth synoptic formalism and incrementalism. Thethird fallacy of formalization (Mintzberg,1994a) may not be fallacious after all: as thisstudy illustrates, there is room for formal, specificplanning in the work of managers, and no supportfor firms which rely on incremental processesalone. This more inclusive, both/and conceptuali-zation of the strategy formation process is congru-ent with recent findings that suggest successfulstrategy-making requires multiple strategy forma-tion modes or capabilities (Hart and Banbury,


17/25


1994). Forming and reforming specific plans andmaking incremental adjustment as implementationproceeds are probably two of the capabilitiesrequired. The remedy for bad planning is goodplanning, which includes incrementalism withinits ambit.

Regarding planning duration and planning

capabilities

The significant role planning duration plays inmoderating the planning specificity/OFP relation-ship surfaces a variable that has received littleattention to date. The failure to control for plan-ning duration may explain some of the inconsis-tencies and brittle nature of past findings. Brackerand Pearsons (1986) findings are replicated,though 4 years, and not 5 years, appears the

critical cut-off point. Two factors probablyexplain planning durations impact on theplanning/performance relationship. First, and asalready pointed out above, a lapse of time isnecessary before the fruits of good planning showup in externally oriented performance data.Implementation takes time, and the planningspecificity/OFP relationship time lag providesclear indication of this. Second, the quality ofplanning probably improves with timethe lowerPP score mean for firms planning for a year orless, and the increase in PP score means over

time supports this notion. Anecdotal data fromcases (see for example Simon and Weston, 1990;Porter and Dougherty, 1976) provide more finegrained corroboration that firm planning capabili-ties can and do improve over time. This inter-pretation accords with resource based theory thatholds firms possess different combinations or lev-els of capabilities (Barney, 1991; Dierickx andCool, 1989; Hart and Banbury, 1994; Wernerfelt,1984). Planning duration may be a key factorcontributing to the accumulation (or lack ofaccumulation) of firm level planning capabilities.

As much as firms plan to learn (i.e., specificplanning a priori facilitates subsequentincrementalism) they probably in the early yearsof planning learn to plan. More research examin-ing how firm planning processes/capabilitiesdevelop and are learned over time is needed, inparticular specifying how this learning processunfolds. Research of this nature would accordwith recent calls for more evolutionary ordynamic perspectives on strategy (Schendel,


1996; Barnett and Burgelman, 1996) though theemphasis here is on the evolution of strategyprocess rather than content. Regrettably, thestrong and perhaps overstated condemnation offormal planning that has become the norm insome Learning School critiques may have unin-

tentionally devalued planning process research.This studys findings, hopefully, revive the needfor such research and highlight some avenues forfuture research to rectify the imbalance.

Planning durations impact on theplanning/performance relationship also providessome basis to counter the concern that correlationdoes not prove causation. An opposite causalspecification has been proposed (see for exampleHopkins and Hopkins, 1997 who assert that highfirm performance promotes good planning) ratherthan sound planning contributes to high perform-

ance. That time must pass before systematicplanning/external performance relationships arenoted supports the latter causal model: improvedplanning accompanied by implementation capa-bilities that also may improve over time, botheventually contributing to superior external firmperformance.

Regarding environments and planning

Rejecting environment as a moderator of theplanning/performance relationship is not the only

insight this study provides into theenvironment/planning nexus. The findings alsosupport an alternative theory on the environmentsimpact on planning. Rather than being amenableto classic formal strategic planning (as has beenargued by many to date) stable environments mayin fact require less planning. Having refined theroutines to operate in the stable environment,planning is not needed until the environmentchanges. Planning capabilities would thus remaindormant and/or underdeveloped. Three studyfindings support this interpretation: the lower endsand means specificity scores of firms inMature/Stable environments, the lower PP scoremean for firms operating in Mature/Stableenvironments; and the greater planning inflexi-bility exhibited by firms operating in stableenvironments. In short, stable environments maycontribute to firm planning capabilitydiscountsenvironment may be another factor(in addition to planning duration) that moderatesthe planning capabilities accumulated by firms.


18/25


By demanding more sophisticated planning,unstable environments may force the developmentof planning capabilities reflected in planningcapability premia. This line of argument alsoexplains Iaquinto and Fredricksons surprisingrecent finding that top management team agree-

ment about planning comprehensiveness wasnegatively correlated with industry stability(Iaquinto and Fredrickson, 1997). Better planners(operating in unstable environments) should dis-play closer agreement about comprehensivenessthan poorer planners (who populate stableenvironments) should. Similar to early organi-zation theorists (see for example Burns andStalker, 1961; Lawrence and Lorsch, 1967; Bour-geois, 1985) the suggested causal sequence isenvironment determines organization, whichdetermines effectiveness. Here the organization

component is firm level planning capabilities.That environment moderates planning flexi-

bility opens up an additional line of inquiry forplanning/performance research. In the case ofends, the findings clearly show that as environ-mental instability grows, so does flexibility. Inthe case of means, though displaying a similarunderlying structure in that the meansspecificity/inflexibility relationships weaken asenvironmental instability increases, no significantrelationships were noted. At the very least, plan-ning specificity is not systematically associated

with inflexibilitythe rigidity hypothesis (Hartand Banbury, 1994) is not supported. Neitheris the lack of any direct flexibility/performanceassociation concerning. To note performanceeffects, plans must be both flexible andspecificpoorly specified flexible plans will not enhanceperformance. However, since this is the first studyto discover that environment moderates planningcapabilities and flexibility, definitive conclusionsshould not be drawn. Further research is neededto confirm the findings presented here.

Implications for practice

The findings have four important implications forpractitioners. First, the importance and value offormal, specific strategic planning is underlined.Second, though plans should be specific, theymust also be flexible, especially in unstableenvironments. Once formed, firms must be pre-pared to rework and amend plans incrementallyas implementation proceeds. At times, even full-


scale abandonment may be necessary. Third, forfirms new to specific planning, the importanceof planning duration cannot be over-emphasized.Though the findings show that formal specificplanning is an important and integral part of themanagement task, and that significant perform-

ance effects can and do accompany such activi-ties, these results are usually only seen in time.Persistence in planning, as much as specificityand flexibility in planning is key to the realizationof any performance enhancements. Similar tomost activities or processes, the quality improvesthe more it is done. Firms new to planning mustexpect initial efforts to be somewhat off the mark,and procedures should be put in place to ensurethat learning to plan is accomplished asefficiently as is possible. Lastly, firms emergingfrom stable environments should be especially

cautious. Since one possible outcome of operatingin a stable environment is an underdevelopedplanning capability, the learning curve faced bysuch firms when more sophisticated planningbecomes necessary may be steep. This appliesfor example to firms facing deregulation, such asutilities, or firms facing increased competitionfrom a prior position of oligopoly or comfortablyregulated competition, such as banks or telecom-munications providers. However, notwithstandingthe disadvantages such firms may face, this studyshows that the time and effort devoted to planning

(a factor within managements control) ratherthan the volatility of the external environment (afactor mostly beyond managements control) isthe moderator of importance. When the goinggets tough, the tough go planning: formally, spe-cifically, yet with flexibility and with persistence.And once they have learned to plan, they planto learn.

ACKNOWLEDGMENTS

The authors would like to thank the many execu-tives who provided the data for this study. Theconstructive comments of the referees are alsogratefully acknowledged.

REFERENCES

Aiken, L. S. and S. G. West (1991). MultipleRegression: Testing and Interpreting Interactions.Sage, Newbury Park, CA.


19/25


Andrews, K. R. (1971). The Concept of CorporateStrategy. Dow Jones-Irwin, Homewood, IL.

Ansoff, H. I. (1991). Critique of Henry MintzbergsThe Design School: Reconsidering the basic prem-ises of strategic planning, Strategic Management

Journal, 12(6), pp. 449461.Ansoff, H. I. (1994). Comment on Henry Mintzbergs

Rethinking strategic planning, Long Range Plan-ning, 27(3), pp. 3132.

Ansoff, H. I., J. Avner, R. G. Brandenburg, F. E.Portner, and R. Radosevich (1970). Does planningpay off? The effects of planning on success ofacquisitions in American firms, Long Range Plan-ning, 3(2), pp. 27.

Anthony, R. N. (1965). Planning and Control Systems:A Framework for Analysis. Division of Research,Harvard Business School, Boston, MA.

Armstrong, J. S. (1982). The value of formal planningfor strategic decisions: Review of empiricalresearch, Strategic Management Journal, 3(3), pp.197211.

Barnett, W. P. and R. A. Burgelman (1996). Evolution-

ary perspectives on strategy, Strategic ManagementJournal, Summer Special Issue, 17, pp. 519.

Barney, J. (1991). Firm resources and sustained com-petitive advantage, Journal of Management, 17(1),pp. 99120.

Bourgeois, L. J. (1980). Performance and consensus,Strategic Management Journal, 1(3), pp. 227248.

Bourgeois, L. J. (1985). Strategic goals, perceiveduncertainty, and economic performance in volatileenvironments, Academy of Management Journal,28(3), pp. 548573.

Bourgeois, L. J. and D. R. Brodwin (1984). Strategicimplementation: Five approaches to an elusivephenomenon, Strategic Management Journal, 5(3),

pp. 241264.Boyd, B. K. (1991). Strategic planning and financial

performance: A meta-analytic review, Journal ofManagement Studies, 28(4), pp. 353374.

Boyd, B. K. and E. Reuning-Elliot (1998). A measure-ment model of strategic planning, Strategic Man-agement Journal, 19(2), pp. 181192.

Bracker, J. S. and J. N. Pearson (1986). Planningand financial performance of small, mature firms,Strategic Management Journal, 7(6), pp. 503522.

Bresser, R. K. and R. C. Bishop (1983). Dysfunctionaleffects of formal planning: Two theoretical expla-nations, Academy of Management Review, 8(4), pp.588599.

Burns, T. and G. M. Stalker (1961). The Managementof Innovation. Tavistock, London.

Camillus, J. C. (1982). Reconciling logical incremen-talism and synoptic formalism An integratedapproach to designing strategic planning processes,Strategic Management Journal, 3(3), pp. 277283.

Camillus, J. C. (1986). Strategic Planning and Manage-ment Control: Systems for Survival and Success.Lexington Books, Lexington, MA.

Chaffee, E. E. (1985). Three models of strategy,Academy of Management Review,10(1), pp. 8998.

Chakravarthy, B. S. (1987). On tailoring a strategic


planning system to its context: Some empirical evi-dence, Strategic Management Journal, 8(6), pp.517534.

Chandler, A. (1962). Strategy and Structure. MITPress, Cambridge, MA.

Dess, G. G. (1980). The relationship between objectiveand subjective measures of manufacturers com-petitive environments: Implications for firm per-formance, doctoral dissertation, University ofWashington.

Dess, G. G. (1987). Consensus on strategy formulationand organizational performance: Competitors in afragmented industry, Strategic Management Jour-nal, 8(3), pp. 259277.

Dess, G. G. and D. Beard (1984). Dimensions oforganizational task environments, AdministrativeScience Quarterly, 29(1), pp. 5273.

Dierickx, I and K. Cool (1989). Asset-stock accumu-lation and sustainability of competitive advantage,

Management Science, 35(12), pp. 15041511.Eisenhardt, K. M. (1989). Making fast strategic

decisions in high velocity environments, Academy

of Management Journal, 32(3), pp. 543576.Floyd, S. W. and B. Wooldridge (1992). Middle man-

agement involvement in strategy and its associationwith strategic type: A research note, Strategic Man-agement Journal, Summer Special Issue, 13, pp.153167.

Fredrickson, J. W. (1984). The comprehensiveness ofstrategic decision processes: Extension, observations,future directions, Academy of Management Journal,27(2), pp. 445466.

Fredrickson, J. W. and A. L. Iaquinto (1989). Inertiaand creeping rationality in strategic decision proc-esses, Academy of Management Journal, 32(3), pp.516542.

Fredrickson, J. W. and T. R. Mitchell (1984). Strategicdecision processes: Comprehensiveness and perform-ance in an industry with an unstable environment,

Academy of Management Journal, 27(3), pp. 399423.

Fulmer, R. M. and L. W. Rue (1974). The practiceand profitability of long range planning, ManagerialPlanning, 22, pp. 17.

Goll, I. and A. M. Rasheed (1997). Rational decision-making and firm performance: The moderating roleof environment, Strategic Management Journal,18(7), pp. 583591.

Goold, M. (1992). Design, learning and planning: Afurther observation of the Design School debate,

Strategic Management Journal, 13(2), pp. 169170.Granger, C. H. (1964). The hierarchy of objectives,

Harvard Business Review, 42(3), pp. 6374.Gup, B. E. and D. D. Whitehead (1989). Strategic

planning in banks: Does it pay?, Long Range Plan-ning, 22, pp. 124130.

Guttman, L. A. (1944). A basis for scaling qualitativedata, American Sociological Review, 9, pp. 139150.

Hamel, G. and C. K. Prahalad (1989). Strategicintent, Harvard Business Review, 67(3), pp. 6376.

Hart, S. (1992). An integrative framework for strategy-


20/25


making processes, Academy of ManagementReview, 17, pp. 327351.

Hart, S. and C. Banbury (1994). How strategy-makingprocesses can make a difference, Strategic Manage-ment Journal, 15(4), pp. 251269.

Hofer, C. and D. E. Schendel (1978). Strategy Formu-lation: Analytical Concepts. West Publishing Com-pany, St. Paul, MN.

Hopkins, W. E. and S. A. Hopkins (1997). Strategicplanning-financial performance relationships inbanks: A causal examination,Strategic Management

Journal, 18(8), pp. 635652.Iaquinto, A. L. and J. W. Fredrickson (1997). Top

management team agreement about the strategicdecision process: A test of some of its determinantsand consequences, Strategic Management Journal,18(1), pp. 6375.

Judge, W. Q. and A. Miller (1991). Antecedents andoutcomes of decision speed in different environmen-tal conditions, Academy of Management Journal,34(2), pp. 449463.

King, W. R. and D. I. Cleland (1987). The choice

elements of strategic management. In W.R. Kingand D.I. Cleland (eds.), Strategic Planning and

Management Handbook. Van Nostrand Reinhold,New York, pp. 6369.

Kudla, R. J. (1980). The effects of strategic planningon common stock returns, Academy of Management

Journal, 23(1), pp. 520.Lawrence, P. R. and J. W. Lorsch (1967). Organization

and Environment: Managing Differentiation andIntegration. Harvard Business School Press, Bos-ton, MA.

Lee, A. S. (1991). Integrating positivist and inter-pretive approaches to organizational research,Organization Science, 2, pp. 342365.

Leontiades, M. and A. Tezel (1980). Planning percep-tions and planning results, Strategic Management

Journal, 1(2), pp. 6576.Li, M. and R. L. Simerly (1998). The moderating

effects of environmental dynamism on the ownershipand performance relationship, Strategic Manage-ment Journal, 19(2), pp. 169179.

Lindblom, C. E. (1959). The science of muddlingthrough, Public Administration Review, 19(2), pp.7988.

Lindsay, W. M. and L. W. Rue (1980). Impact of theorganization environment on the long range planningprocess: A contingency view. Academy of Manage-ment Journal, 23(3), pp. 385404.

MacCrimmon, K. R. (1988).Essence of strategy: Ends,means and conditions. In J. H. Grant (ed.), Strategic

Management Frontiers. JAI Press, Greenwich, CT,pp. 4972.

Michael, D. N. (1973). On Learning to Plan andPlanning to Learn. Jossey-Bass, San Francisco, CA.

Miller, C. C. and L. B. Cardinal (1994). Strategicplanning and firm performance: A synthesis of morethan two decades of research, Academy of Manage-ment Journal, 37(6), pp. 16491665.

Miller, D. and P. H. Friesen (1983). Strategy makingand environment: The third link, Strategic Manage-


ment Journal, 4(3), pp. 221235.Mintzberg, H. (1973). Strategy making in three

modes, California Management Review, 16(2), pp.4453.

Mintzberg, H. (1978). Patterns of strategy formation,Management Science, 24(9), pp. 934948.

Mintzberg, H. (1987). Crafting strategy, HarvardBusiness Review, 65(4), pp. 6675.

Mintzberg, H. (1990a). Strategy formation: Schools ofthought. In J. W. Fredrickson (ed.), Perspectiveson Strategic Management. Harper Business, NewYork, pp. 105235.

Mintzberg, H. (1990b). The Design School: Recon-sidering the basic premises of strategic man-agement, Strategic Management Journal, 11(3), pp.171195.

Mintzberg, H. (1991). Learning 1, planning 0: Replyto Igor Ansoff, Strategic Management Journal,12(6), pp. 463466.

Mintzberg, H. (1994a). Rethinking strategic planning,Part I: Pitfalls and fallacies, Long Range Planning,27(3), pp. 1221.

Mintzberg, H. (1994b). Rethinking strategic planning,Part II: New roles for planners, Long Range Plan-ning, 27(3), pp. 2230.

Mintzberg, H. (1994c). The fall and rise of strategicplanning, Harvard Business Review, 72(1), pp.107114.

Mintzberg, H. (1994d). The Rise and Fall of StrategicPlanning. Free Press and Prentice Hall International,New York.

Mintzberg, H. and J. Waters (1982). Tracking strategyin an entrepreneurial firm, Academy of Management

Journal, 5, pp. 465499.Mitroff, I. I. and J. R. Emshoff (1979). On strategic

assumption-making: A dialectical approach to policy

and planning, Academy of Management Review,4(1), pp. 112.

Narayanan, V. K. and L. Fahey (1982). The micro-politics of strategy formation, Academy of Manage-ment Review, 7(1) pp. 2534.

Pearce, J. A., E. B. Freeman and R. B. Robinson(1987). The tenuous link between formal strategicplanning and performance, Academy of Manage-ment Review, 12(4), pp. 658675.

Porter, M. E. and J. B. Dougherty (1976). EG&G, Inc.(A) Condensed. Harvard Business School Case, 9377027.

Priem, R. L., A. M. A. Rasheed and A. G. Kotulic(1995). Rationality in strategic decision processes,

environmental dynamism and firm performance,Journal of Management, 21(5), pp. 913929.

Quinn, J. B. (1980). Strategies for Change: LogicalIncrementalism. Richard D. Irwin, Homewood, IL.

Rhyne, L. C. (1986). The relationship of strategicplanning to financial performance, Strategic Man-agement Journal, 7(5), pp. 423436.

Rhyne, L. C. (1987). Contrasting planning systems inhigh, medium, and low performance companies,

Journal of Management Studies,24(4), pp. 363385.Robinson, R. B. and J. A. Pearce II. (1983). The

impact of formalized strategic planning on financial


21/25


performance in small organizations, Strategic Man-agement Journal, 4(3), pp. 197207.

Schendel, D. E. (1996). Evolutionary perspectives onstrategy, Strategic Management Journal, SummerSpecial Issue, 17, pp. 14.

Schendel, D. E. and C. W. Hofer (eds.) (1979). Stra-tegic Management. Little, Brown, Boston, MA.

Schwenk, C. R. and C. B. Schrader (1993). Effectsof formal strategic planning on financial performancein small firms: A meta-analysis, EntrepreneurshipTheory and Practice, 17(3), pp. 5364.

Shrivastava, P. and J. Grant (1985). Empirically derivedmodels of strategic decision-making processes, Strategic

Management Journal, 6(2), pp. 97113.Simon, R. and H. Weston (1990). Central Maine

Power Company: Goals and objectives program(A), Harvard Business School Case, 9-190-065.

Snyder, N. H. and W. F. Glueck (1982). Can environ-mental volatility be measured objectively?, Acad-emy of Management Journal, 25 (1), pp. 185192.

Thune, S. S. and R. J. House (1970). Where long rangeplanning pays off, Business Horizons, 13(4), pp. 8187.

Tosi, H. L. and S. J. Carroll (1968). Managerialreaction to management by objectives, Academy of

Management Journal, 11, pp. 415426.Tosi, H. L., R. Aldag and R. Storey (1973). On the

measurement of the environment: An assessment of theLawrence and Lorsch environmental uncertainty subscale, Administrative Science Quarterly, 18, pp. 2736.

APPENDIX: Questionnaire Items(Bold numbers indicate scoring protocol)

Section I: Strategic Ends Specificity

1. Please indicate which one statement most closely describes your firm.

No ends have been developed for our firm in the strategy formation process. 1

A few (less than five) ends have been developed for our firm in the strategy formation process, but they 2remain undocumented and informal.

A few (less than five) ends have been developed for our firm and formally documented in the strategy 3formation process.

A number (greater than five) of ends have been developed for our firm in the strategy formation process, 4but they remain undocumented and informal.

A number (greater than five) of ends have been developed for our firm and formally documented in the 5strategy formation process.

Many ends have been developed for our firm and formally documented in the strategy formation proccess, 6including a statement of firm mission/purpose, and specification of strategic objectives/goals for differentareas of the firm.

If you selected the first statement above (no firm ends), please go to Section II of this questionnaire .


Venkatraman, N. (1989). The concept of fit in strategyresearch: Towards verbal and statistical correspon-dence, Academy of Management Review, 14, pp.423444.

Venkatraman, N. and V. Ramanujam (1987). Planningand performance: A new look at an old question,

Business Horizons, 30(3), pp. 1925.Wernerfelt, B. (1984). A resource based view of the

firm, Strategic Management Journal, 5(2), pp.171180.

West, C. T. and C. R. Schwenk (

Documents

Planning - Learning SMJ Article