Untangling the relationship between strategic planning and performance: The role of contingency factors
Powell, Thomas C
Strategy researchers have devoted considerable attention to the relationship between strategic planning and performance, with over 30 empirical studies appearing since the publication of Thune and House’s study in 1970. This literature has been reviewed extensively by Armstrong (1982), Pearce, Freeman, and Robinson (1987), and Shrader, Taylor, and Dalton (1984).
The reviewers agree that findings in this research stream are nearly impossible to reconcile. For example, Shrader et al. (1984, p. 154) concluded that “there is no clear systematic relationship between long-range planning and organizational performance,” and Pearce et al. (1987, p. 671) concluded that “empirical support for the normative suggestions by strategic planning advocates that all firms should engage in formal strategic planning has been inconsistent and often contradictory.”
Although the problems vary from study to study, the reviewers have focused on two methodological shortcomings, namely, poor measurement of strategic planning processes and the neglect of important contingency variables such as firm size, industry, and generic strategy. This study attempts to untangle the planning-performance relationship by introducing valid, reliable measures for key strategy-making constructs, and by investigating the role of contingency variables in moderating the planning- performance relationship. The following section provides a brief review of the planning-performance literature, and subsequently introduces the contingency variables and hypotheses. The empirical findings are then presented and discussed, along with their implications or researchers and practitioners.
STRATEGIC PLANNING AND FINANCIAL PERFORMANCE
Planning adherents have widely asserted that formal strategic planning provides important benefits to firms (Steiner, 1979; Thompson & Strickland, 1987). It generates information, ensures a thorough consideration of all feasible options, forces the firm to evaluate its environment, stimulates new ideas, increases motivation and commitment, enhances internal communications and interaction, and has symbolic value, reassuring stakeholders (such as creditors, potential customers, and the investment community) that the firm has set a confident, proactive course for the future. Since a firm would presumably be more profitable with these characteristics than without them, strategic planning is said to have financial performance consequences for the firm.
However, even planning adherents such as Steiner (1979) concede that formal planning is not a costless activity, and that these costs must be weighed against the benefits just mentioned. For example, a strategic planning program that entails a staff person or department devoted exclusively to formal planning incurs significant, direct human resources expenses. If planning is done by line managers, the expense is less direct, but may be greater overall, since it removes managers from operations to gather information, attend committee meetings, and produce reports. This expense becomes far greater as senior managers become involved in the planning process. Of course, consultants can be hired to produce strategic plans, but this also is expensive, and researchers and practitioners are unanimous in their contempt for strategic plans prepared without the involvement, and consent, of top management and relevant line managers (Andrews, 1980; Quinn, 1980).
Organizational scholars have also argued that formal strategic planning deludes managers into a false sense of control and security, and obscures outworn assumptions, producing the strategic “blind-spots” and complacency that inevitably lead to crises (Miller, 1990; Starbuck, Greve, & Hedberg, 1978). Moreover, Brunsson (1982) has argued that formal planning threatens organizational well-being by delineating environmental uncertainties, producing internal conflicts, creating a confusing array of alternatives that cannot be rationally compared (in part because their consequences are unknowable), and focusing explicitly on negative data (e.g., the organization’s weaknesses, its competitors’ strengths, and threats in the environment). He argues that effective actions result from internal motivation and commitment, and that formal planning is inimical to both.
The planning-performance studies seek to determine empirically whether strategic planning produces positive net financial performance. These studies have employed a variety of planning definitions and measures, but they share a common interest in exploring the financial performance consequences of the basic tools, techniques, and activities of formal strategic planning, e.g., systematic intelligence-gathering, market research, SWOT analysis, portfolio analysis, mathematical or computer modelling, formal planning meetings, and written long-range plans.
In reviewing these studies, Pearce, Robbins, and Robinson (1987) identified three successive waves of empirical research. The first wave, led by Ansoff et al. (1970) and Thune and House (1970), was characterized by planner/non-planner classification schemes, and generally positive planning-performance relationships. Thune and House, for example, classified 92 survey respondents either as formal or informal planners, and, using a sub-sample of 36 of these 92 respondents, compared the two planning types on four objective performance measures covering a seven-year period. The authors reported “a remarkable association between economic performance and long-range planning” (p. 83), although the results varied somewhat across six broad industry classifications. The authors could not fully account for the industry differences, but recommended that future studies attend more closely to industry as a contingency variable.
The second wave, which included studies by Fulmer and Rue (1974), Grinyer and Norburn (1975), Kallman and Shapiro (1978), and Kudla (1980), was characterized by more discriminating planning classification schemes, and occasional findings of non-significant or negative planning-performance relationships. Fulmer and Rue, for example, separated 386 survey respondents into four planning categories (which they later collapsed into two), and three broad industry classifications, and reported a positive planning-performance relationship in durables industries, and a negative (but non-significant) relationship in non-durables industries. They concluded that their industry classifications were probably overbroad, and that “there is no simple, across the board relationship between the financial success of the firm and the use of long-range planning. Obviously, such variables as timing, luck, and the immeasurable quality of ‘overall management competence’ have a more direct relationship on a firm’s performance than the formality of its long-range planning” (p. 7).
The third wave included studies by Bracker and Pearson (1986), Fredrickson (1984), Pearce, Robbins, and Robinson (1997), Rhyne (1986), and Robinson and Pearce (1983). This group was characterized by increasingly sophisticated planning and performance measures, and a wide variety of findings. Pearce et al. (1987) were among the first to argue that business strategy content and strategic planning may be interdependent, but did not use Porter’s generic strategy framework, and did not produce significant results. Rhyne factor-analyzed survey responses from 210 Fortune 1000 companies to derive an eight-item scale of “planning openness,” which was used to classify firms into five planning categories. However, the author did not classify firms by industry, as Thune and House (1970) had done. The author reported positive planning-performance relationships in three of the five years studied (using ten-year moving averages of ROI as the performance measure), and no significant relationships in the other two years, concluding, on a positive note, that “the results provide assurance that the prescriptions of strategic management theory are indeed valid” (p. 435). It will be shown that this may be a spurious result, given that all sample firms were large, and that firm size may be associated with both strategic planning and profitability.
The reviewers focused on two primary flaws in this line of research: poor measurement of strategic planning — few studies used reliable scales for measuring strategic planning (in some cases, planning was measured by a single item), and the scales were not comparable from one study to the next; and neglect of contingency variables — most of the studies examined simple, bivariate planning-performance relationships, ignoring potential contingency variables, particularly firm size, industry, and generic strategy. The neglect of contingency variables has proven particularly vexing, and may be responsible for apparent contradictions in the earlier research. Indeed, Pearce et al. (1987, p. 671) concluded that “the principal methodological concern is the lack of attention to contextual influences,” and Shrader et al. (1984, p. 165) stated that, “one theme drawn from this review . . . is that future research needs to deal with variables upon which effective strategic planning is contingent for specific industries and organizations.”
One result of neglecting contingency variables is increased risk of type I experimental error, i.e., reporting significant, but spurious, planning-performance correlations. As noted above, it has been shown in separate studies that firm size correlates both with formal strategic planning (Blau & Schoenherr, 1971) and with profitability (Schoeffler, Buzzell, & Heany, 1974; Shepherd, 1972). Therefore, one would expect planning and profitability to correlate positively merely because firms come in different sizes. Unfortunately, many planning-performance studies have failed to control for firm size effects, and some have rendered formal planning prescriptions based on this spurious correlation (e.g., Malik & Karger, 1975).
Moreover, neglect of contingency variables may also increase the risk of type II experimental error, i.e., reporting no significant correlations when correlations do exist within homogeneous subsamples. For example, non-significant planning-performance correlations have been reported in studies using large, heterogeneous samples of firms, notably firms competing in different industries, or using different strategies in the same industries (e.g., Kudla, 1980, and nearly all studies using Fortune 1000 or similar databases). If the planning-performance relationship is a contingent one — for example, if planning is more profitable in some industries than in others, or in some strategic groups more than in others — then the low overall correlation would not be a surprise. As a result, studies that ignore the basic contingency variables of strategic management research, i.e., industry membership and strategy, are likely to miss potentially significant planning-performance correlations that lay beneath the surface in their heterogeneous samples.
CONTINGENCY VARIABLES AND HYPOTHESES
The previous discussion suggests that, to reduce the risk of experimental error, future planning-performance studies should control for firm size effects, and for basic contingency variables in strategic management research, including industry membership and strategy. Other contingency variables might also be explored, including technology and age of the firm. However, based on their prominence in the organizational research, and on their ex ante likelihood of confounding results in planning-performance research, this research focuses on size, industry membership, and generic strategy. This section explains further why these variables are theoretically important, and presents hypotheses for empirical testing.
Firm size as a contingency variable. Perhaps the most frequently-studied context variable in the contingency research, and a critical organizational variable since Weber’s (1947) initial studies of bureaucracy, is organizational size. The Aston researchers found that size is closely associated with formalization, being a strong predictor of reliance on paperwork and the use of formal procedures (Pugh, Hickson, & Hinings, 1969). Blau (1970) and others have noted that larger organizations adopt more formal procedures — including formal strategic planning — in order to improve control, since increased size also brings greater organizational complexity and carries with it the need for coordination through formalization, standardization, and controls (Blau & Schoenherr, 1971; Child, 1972; Miller & Droge, 1986; Pugh et al., 1968, 1969).
Small firms, on the other hand, tend to forego formal strategic planning because their environments are comprehensible and their internal operations manageable by a single person or small team, without the need for systematic, formal scanning, extensive internal analysis, or detailed, written long-range plans (Lorange & Vancil, 1976: Mintzberg, 1979). In fact, comprehensive planning may actually harm small firms by diverting human and financial resources into an unnecessarily extensive planning process, and by exerting excessive formal control. For these reasons, Mintzberg (1973, 1979) hypothesized that the adoption of formal strategic planning methods was closely, and positively, linked to firm size, and subsequent empirical studies have, without exception, strongly supported this hypothesis (see, for example, Fredrickson, 1984, and Miller & Friesen, 1984).
The problem arises for planning-performance research because researchers have demonstrated connections not only between size and formal planning, but also between size and profitability. The PIMS researchers, for example, reported that the most statistically-significant proportion of ROA variance in their sample was attributable to market share, an intra-industry size measure (Buzzell, Gale, & Sultan, 1976; Schoeffler, Buzzell, & Heany, 1974), and this finding is consistent with profitability research in the industrial organization economics and strategic management literatures (e.g. Porter, 1976, 1980; Shepherd, 1972). These studies attributed the size-profitability connection to large firms’ capacities to gain sustainable cost advantages, i.e., advantages protected by entry and mobility barriers, such as economies of scale (in purchasing, research, marketing, or production), economies of scope, vertical integration, and dominance over suppliers. As suggested earlier, the mutual correlation of firm size with both profitability and strategic planning is problematic because it tends to produce positive and significant, but spurious, planning-performance correlations. These considerations give rise to the following hypotheses:
H1: Overall, strategic planning correlates significantly with financial performance (but only because both strategic planning and performance correlate with a third variable, firm size).
H2: When firm size effects are removed, strategic planning does not correlate significantly with financial performance.
H3: The correlation between strategic planning and performance is significantly greater among large firms than among small firms.
Industry as a contingency variable. Some industries consistently produce higher returns on invested capital than others, for reasons that have been extensively examined by industrial organization economists and strategy researchers (e.g., Porter, 1980). Moreover, organization theorists have suggested that planning may be more prevalent in some industries than in others, with instability and uncertainty emerging as the critical contingency variables (Lawrence & Lorsch, 1967). For example, Lindblom (1959) suggested that planning is only possible in stable, predictable environments, and Mintzberg (1973, 1979) has argued that, in order to plan comprehensively, an organization “must face an environment that is reasonably predictable and stable” (1979, p. 50). Similarly, in one of the few studies to control for industry membership, Fredrickson and Mitchell (1984, p. 464) concluded that “synoptic processes are appropriate for firms in stable environments and incremental processes are appropriate for firms in unstable environments.” In fact, this finding dramatizes the need for contingency-based research: Had the researchers not controlled for industry membership, they would have found no planning-performance correlations at all, since the positive correlation in the stable industry would merely have offset the negative correlation in the unstable industry.
These findings suggest that the planning-profitability correlation should be greater in stable than in unstable industries. However, there also exists evidence to the contrary. Miller and Friesen (1983, p. 225), for example, found that among a group of successful firms “there appears to be a modest tendency for increased dynamism to be met by increased analysis,” whereas, among unsuccessful firms, there existed, “a weak negative relationship between changes in analysis and dynamism.” Therefore, although researchers agree that environmental stability moderates the planning-performance relationship, the direction of its effects is open to dispute, and in need of reconciliation. Hypothesis 4 tentatively adopts the null hypothesis, as follows:
H4: The correlation between strategic planning and financial performance does not differ among firms in a stable industry from among firs in an unstable industry.
Generic strategy as a contingency variable. Gilbert and Strebel (1986) and Porter (1980) have suggested that the two fundamental bases of competitive advantage are cost leadership and differentiation. According to Porter, generic strategy is interdependent with a firm’s culture, structure, and strategic planning process. Cost leadership requires (p. 40) “tight cost control, frequent, detailed control reports, structured organization and responsibilities, and incentives based on meeting strict quantitative targets,” whereas differentiation requires “creative flair, and subjective measurement and incentives instead of quantitative measures.” As a result, comprehensive strategic planning — which involves specific goals, quantitative planning techniques, and explicit, written long-range plans — is, under the Porter approach, more consistent with the cost leadership generic strategy than with that of differentiation. Studies that reported near-zero planning-performance correlations, but neglected generic strategy differences, may have overlooked the offsetting effects of a positive correlation among cost leaders and a negative correlation among differentiators. This issue may also be connected to firm size, since true cost leadership is generally restricted to firms large enough to establish dominance over suppliers and take full advantage of available scale economies. Thus, as in the other hypotheses, firm size is partialled from the analysis of H5, which is as follows:
H5: The correlation between strategic planning and financial performance is significantly greater among cost leaders than among differentiators.
DATA AND MEASURES
Sample
In order to test these hypotheses, it was necessary to study firms both in stable and unstable industries. To accomplish this, the study was focused on two four-digit SIC-code industries that met the following criteria: (1) the industries contrasted as much as possible in environmental stability, but were otherwise similar (e.g., both manufactured consumer products); (2) the four-digit codes were narrowly-defined industries with natural competitors, rather than broad groups or miscellaneous industry categories; (3) the industries were sufficiently fragmented to generate a large sample size [i.e., each had at least 250 firms listed in the combined Dun and Bradstreet (1988) and Standard and Poor’s (1988) directories]; and (4) at least 80% of the firms in the industries were undiversified firms competing primarily in that industry according to Dun and Bradstreet (1988). [Note: this criterion was added because it is not generally valid to pool diversified and undiversified firms, as many previous studies have done (e.g., Kudla, 1980). A fundamental premise in strategic management is that single-business and diversified contexts give rise to entirely different strategic and structural concerns (Hofer & Schendel, 1978). This is a serious problem, since many of the contingency variables — such as industry membership and generic strategy — become either diluted or meaningless in diversified contexts].
A stable industry was defined as one that met the following criteria: A relatively low variance in average annual change in total shipments for all four-digit codes; an average annual change in total shipments near the median for all four-digit codes; a relatively low variance in average annual industry employment over an extended period; and anecdotal data from industry experts and participants supporting the objective data.
An extensive review of all four-digit SIC codes resulted in the selection of industry codes 2512 (wood upholstered furniture) as the stable industry, and 2335 (women’s dresses) as the unstable industry. Using the criteria for industry stability, the furniture industry ranked as one of the most stable of all the manufacturing industries reviewed. In a Department of Commerce (1987) study of 219 four-digit industries, it had the eighth-lowest variance of total shipments between 1973 and 1987, it ranked exactly at the median of all industries studied (110th of 219 industries studied) in average change in annual shipments between 1972 and 1987, and its variance in annual employment was low in comparison to other industries (a variance of 13,700 on a mean of 280,500). The women’s apparel industry, on the other hand, ranked in the upper 20% of industries in variance in total shipments, had average annual changes in shipments (-2.6%) far below the median, and had a high variance in annual employment (a variance of 90,250 on a mean of 119,200). Moreover, data obtained in interviews with industry participants, consultants, and industry analysts supported these objective assessments; anecdotal data suggested that the fashion-consciousness and entrepreneurial nature of the women’s apparel industry contrasted sharply with the product and market stability typical of the furniture industry.
Data Collection
The sampling design required the researchers to gather unpublished data from a large sample of firms. In order to collect these data, and to sample from both publicly- and privately-held firms (and firms located across a broad geographic area), the mail survey method was adopted. However, since mail surveys are vulnerable to a wide range of measurement and methodological pitfalls, special care was taken in designing and administering the instrument. These procedures are outlined in this section, and further details, along with the complete survey, are available from the authors.
The survey was designed in booklet form and administered according to the principles of the Total Design Method (Dillman, 1978). e survey was exhaustively pretested — through personal interviews with academics, industry participants, consultants, and industry experts — and was pilot-tested in a sample of 30 firms. The final survey was then mailed, along with a personal cover letter, to the CEOs of all firms listed in Dun’s Million Dollar Directory (1988) and Standard and Poor’s Register (1988) for SIC codes 2512 (furniture) and 2335 (apparel). A follow-up postcard was sent one week after the initial mailing, a second survey and cover letter were mailed two weeks after the initial mailing, and non-respondents were called beginning in the fourth week.
Of the 544 firms receiving the survey, 113 responded, for a response rate of 20.8%. This response rate is consistent with those of other studies using a similar methodology, and met the expectations for this research design, considering its requirement for direct CEO involvement, the sensitivity of much of the requested information, the large number of variables measured, and the high proportion of privately-held firms in the population. Furthermore, the two industry samples represented not 20% of a large, heterogeneous collection of firms (e.g., the Fortune 500, which resemble one another only in size), but 20% of two relatively homogeneous populations to which findings could be generalized legitimately. To establish further the external validity for these industry samples, the median firm size and profitability of sample firms were compared to known population parameters in each industry. These data (available from the author) showed very slight, statistically insignificant differences between the sample statistics and population parameters, strongly supporting the external validity of the sample data for both industries.
Although it will be possible to generalize findings from the obtained samples to the industry populations, generalization beyond these two industries can only proceed cautiously. Because the industries studied here are, by necessity, fragmented, median firm sizes are small; thus, it will not be possible to generalize with confidence to more concentrated industries or to populations of larger firms. Nonetheless, the range of firm sizes (from 14 to 11,000 employees) and annual sales (from $850,000 to $500 million) in the sample are broad, and the mean number of employees (742.5) and sales ($39 million) would not seem to bias the sample unduly toward small firms. To minimize any possible large-firm bias, firm size was measured as the natural log of the number of employees.
An attempt was made to establish interrater reliability among multiple respondents in a sub-sample of firms, but although the results suggested excellent interrater reliability, the response among this sub-sample was insufficient to establish conclusive results. Among the six firms for which two responses were obtained per firm, the mean of the six intrafirm correlations was r =.70, compared to a mean of r =.27 for the interfirm correlations. Furthermore, 81% of all intrafirm responses (306 of 378 items) fell within a single point of one another on the five- and six-point scales employed, compared to the 55% (208 of 378) that would be expected by chance. Although these data are suggestive of interrater reliability, they cannot be considered conclusive because of the small sub-sample of firms involved. However, since all respondents were CEOs, and most firms were relatively small and undiversified, it seems reasonable to believe that the respondents were well-informed about their firms, and that response or function bias was minimal.
Measurement
Strategic planning was measured using scales developed by Miller (1987). In a principal components analysis of a wide range of strategic planning variables, Miller (1987) found that four variables — future orientation, explicitness, analysis, and scanning — loaded on a single factor. This factor, which Miller labelled “rationality,” is a broad characterization of planning processes that encompasses nearly all of the operational definitions employed in the planning-performance literature. According to Miller (p. 20), the variables in this factor “all relate to the synoptic and planning modes and represent systematic, analytical decision-making. This approach contrasts with the purely spontaneous, intuitive modes found with severely bounded rationality.”
“Rationality” is not the only strategic planning construct that has received research attention; for example, Miller’s principal components analysis generated two additional constructs — which he labelled “assertiveness” and “interaction” — and Bourgeois (1980) has investigated the performance consequences of strategy-making consensus. However, Miller’s findings do suggest that strategic planners, synoptic planners, formal planners, sophisticated planners, rational planners, and comprehensive decision-makers share certain important behavioral characteristics, namely: (1)goal-setting — they set broad, long-range goals and specific, short-range objectives, and these goals are communicated to key organization members; (2) scanning — they search their environments systematically for competitive and market information; and (3) analysis — they hold planning meetings, discuss long-range strategic issues, and develop explicit, written plans, budgets, and forecasts.
Miller’s findings suggest that these three dimensions are highly intercorrelated and, therefore, can be measured both separately and as a single, meaningful strategic planning construct. Since this research seeks comparability with previous planning-performance studies, it employs these three dimensions, and observes their relationships with profitability both separately and as a single strategic planning construct. This research refers to the overall construct not as “rationality,” but as “strategic planning.”
For this research, Miller’s (1987b) strategy-making scales (goal-setting, scanning, analysis, and overall planning) were abridged, based on suggestions received during pre-testing (the scales are given in the Appendix). The generic strategy measures (product innovation, product quality differentiation, and production costs) were original to this research, but were based on attributes of strategy identified by Andrews (1980), Hofer and Schendel (1978), Porter (1980), and Steiner (1979). Firm size was measured by the natural logarithm of the number of full-time employees (Blau & Schoenherr, 1971), and industry stability was measured as a dichotomous variable representing industry affiliation (apparel industry = 0, furniture industry = 1).
Financial performance was measured by survey items concerning profitability and sales growth over a three-year period. Although the use of subjective performance measures is widespread in organizational research and has been justified elsewhere (e.g., Dess, 1987; Lawrence & Lorsch, 1967), their use does invite explanation. In this research, subjective measures were employed for the following reasons. Given differences in accounting conventions, especially concerning inventory valuation, depreciation, and officers’ salaries and expenses (particularly in smaller firms), it was not clear that financial statement data were more accurate measures, or were more comparable across firms, than subjective evaluations. Since all respondents were CEOs, it was assumed they were reasonably well informed of their firms’ financial positions. Many of the firms were privately-held, and did not provide confidential information from their financial statements as a matter of policy. Since no survey identification numbers were used, respondents were assured of complete confidentiality, and had little incentive to provide misleading subjective assessments. Irrespective of the convergent validity between objective and subjective performance measures, CEO perception of performance can be regarded as an important dependent variable in and of itself.
Despite these justifications, the convergent validity of the subjective measures was tested by obtaining objective performance measures from a subset of firms in the overall sample. These firms were asked to provide detailed information from their financial statements for the three fiscal years from 1985 through 1987, including total sales, total assets, and net income after taxes for each year. From this information, average ROA and sales growth were computed for each responding firm. Of the total of 113 respondents, a subset of 52 firms provided both the subjective and objective financial information, and the correlation between these two measures was computed as a test of the convergent validity of the subjective measures. For sales growth, the correlation between the subjective and objective measures was .69, and for profitability the correlation was .58 (each is significant at less than p
It was possible to establish the internal consistencies of the strategic planning scales using Cronbach’s alpha (Cronbach, 1951). Although no acceptable range has been established for this index, Van de Ven and Ferry (1979) have suggested that, for a scale of three items, alpha should fall between .70 and .90 for a narrow construct, between .55 and .70 for a moderately broad construct, and between .35 and .55 for a very broad construct. In order to establish Cronbach reliabilities for the scales, a pilot test was conducted in a sample of 30 firms in a variety of industries, alphas were computed, and the scales were fine-tuned as necessary. In the pilot study, the Cronbach alphas for the goal-setting, scanning, and analysis scales ranged from .74 to .83, and in the actual field study, from .75 to .79. Not only were all reliability coefficients acceptable in both the pilot test and the field study, but their similarities under the two different testing conditions suggests that the scales are robust with respect to changes in experimental settings.
Results and Discussion
The strategic planning-performance correlation. The first two hypotheses consider, respectively, the zero-order correlation between strategic planning and financial performance, and the partial correlation when firm size effects are removed [Note: although some hypotheses are directional, two-tailed tests were used in all testing to avoid overstating statistical significance, and in recognition of the possibility of contrary results — see Welkowitz, Ewen, & Cohen, 1976, pp. 131-133, for an extended discussion of this rationale.] These results are presented in Table 1, which shows the correlations for both the profitability and sales growth measures of financial performance. Table 1 shows that, for the profitability measure, H1 and H2 are supported by the data. (Table 1 omitted) As hypothesized, the zero-order correlation between strategic planning and profitability is positive and statistically significant (r=.20; p
However, Table 1 also shows that the large zero-order correlation between planning and sales growth (r=.49; p
Why should strategic planning correlate with sales growth, but not with profitability? One reason can be found in the traditional theory of strategic planning, namely that planning requires immediate outlays in exchange for expected longer-run benefits. As Steiner (1979) observes, “planning is expensive . . . The time of many people is occupied and costs are incurred for special studies and information” (p. 45). Planning expenses are similar to advertising expenses in that they are often designed to increase market share, but may suppress short-run profitability (in this sample, advertising correlated .02 with profitability and .17 with sales growth). The market share-profitability dilemma has long been recognized as central in strategic management thought (Buzzell, Gale, & Sultan, 1975), and there is no reason to believe that planning — as a forward-looking but costly activity should be exempt from its effects. The good news (for planners) is that planning and sales growth are associated, but without producing a net decrease in short-run profitability.
Firm size as a contingency variable. H3 considers whether the correlation between strategic planning and financial performance is greater among large firms than among small firms (the significance test for the difference between r coefficients uses the r to z transformation and normal curve test developed by R.A. Fisher). To test H3, firms were classified either as “small” (120 or fewer employess: n=42),”medium” (between 120 and 350 employees: n=31), or “large” (over 350 employees: n=40), and differences were tested between the small and large subsamples. These classifications were used in order to maximize the contrast between the largest and smallest firms in the overall sample, and to obtain roughly equal subsamples of large and small firms, while retaining reasonable statistical power. Table 2 presents the analysis, and shows that H3 is supported for both the profitability and sales growth measures of financial performance. (Table 3 omitted) For each performance measure, the strategic planning-performance correlation is highly significant, for large firms (r=.49 and r=.68 for profitability and sales growth, respectively, each significant at p
It is interesting to note that, among small firms, the correlations between strategic planning and financial performance were not significant in either direction. This suggests that, although strategic planning is more nearly associated with performance among large firms than among small firms, it does not appear to have harmful effects on smaller firms. In fact, the correlation between strategic planning and sales growth was positive and nearly significant for small firms. Later analysis (given in Table 4) will shed additional light on this finding, showing that, among small firms, the strategic planning-performance relationship varies with industry membership. (Table 4 omitted)
Industry stability as a contingency variable. H4 considers whether the strategic planning-performance correlation is greater among firms in the stable industry (furniture) than among firms in the unstable industry (apparel). As noted earlier, previous empirical research on this issue has produced conflicting findings, Fredrickson and Mitchell (1984) reporting a higher planning-performance correlation in a stable industry, Miller and Friesen (1983) reporting a higher planning-performance correlation in unstable industries.
In testing H4, all firms in the overall sample (n=113) were divided into their respective industry subsamples. The findings in this study support the idea that industry stability is an important contingency variable, and, as shown in Table 3, they tend to support Miller and Friesen’s view. (Table 3 omitted) They show that, in the unstable industry, the planning-profitability correlation is positive and significant (r=.36; p
Why is planning associated with profitability in the unstable industry but not in the stable one? To examine this issue further, it will be useful to combine the analyses of hypotheses 3 and 4, by examining subgroups formed both by industry membership and by firm size. Table 4 presents this analysis. This analysis shows that planning and profitability are highly correlated among large firms in both industries, but that significant differences emerge between small firms in the two industries. Specifically, the correlation between planning and profitability is .25 among small apparel firms and -.25 among small furniture firms, and the difference is even greater for the goal-setting dimension (r=.49 for small apparel firms and r=-.31 for small furniture firms).
The analysis in Table 4 sheds additional light on two of the key findings thus far. In Table 4, firms are classified both by industry and by firm size. In order to create similarly-sized subsamples that were also large enough to yield meaningful results, definitions for small and large firms varied slightly from those used in Table 2, as described in the table (this was considered a reasonable procedure, since Table 4 does not involve hypothesis testing, and may help to explain results in some of the other tables). First, Table 4 provides underlying data concerning the near-zero planning-profitability correlation among small firms in the overall sample; the near-zero correlation in the overall sample was concealing a negative correlation in the stable industry, offset by a positive correlation in the unstable industry. Second, the data in Table 4 help to explain why planning appears to be more profitable in the unstable industry than in the stable industry — as before, the answer lies in the small apparel firm subgroup, where planning and profitability are, contrary to expectation, highly correlated.
Why should planning and profitability be correlated among small firms in an unstable industry? Although this finding seems unusual, it is possible to sketch a provisional explanation that also helps to reconcile the earlier conflict between the findings of Fredrickson and Mitchell (1984) and Miller and Friesen (1983). This explanation will suggest that, under ordinary conditions, planning is indeed more profitable in stable than in unstable industries, but that certain types of instability — such as an external shock – may call for centralization, goal-setting, and a temporary reversion to formal strategic planning.
The domestic apparel industry has long been subject to market instabilities, competitive uncertainty, and the vagaries of a fashion-conscious society. However, if anecdotal evidence is correct, they have coped with, or “internalized,” this instability by centralizing authority in driven, impulsive entrepreneurs who evaluate market needs and determine styles and production quantities according to experience and intuition, rather than rational planning principles. Using Mintzberg’s (1973, 1979) typologies, they have adopted simple structures and adaptive strategy-making modes, rather than machine bureaucracies and “planning” strategy-making modes. The evidence in this study bears out such an assessment: Controlling for size differences, apparel firms were significantly more centralized and less formalized than furniture firms, and they planned less comprehensively (see the Appendix for formalization and centralization measures).
For many years, centralized authority and intuitive strategic planning were effective means for coping with the unstable apparel market, especially for smaller manufacturers. However, in recent years, low-priced, lookalike imports began to erode the competitive advantages of domestic manufacturers, creating a severe industry shakeout (Standard and Poor’s, 1988), particularly during the period covered by this study (1985-1987). This external shock differed from the ordinary instability with which apparel firms were familiar: Intuitively-derived, incremental changes in styles or quantities were no longer adequate responses, since neither product differentiation nor cost leadership provided defensible competitive advantages against the foreign threat.
This external shock constituted a crisis for many firms in the apparel industry, so that a successful response required a complete reevaluation of the firm’s environment, leadership, mission, goals, and strategy (Starbuck, Greve, & Hedberg, 1978). In other words, “second- order change” (Watzlawick, Weakland, & Fisch, 1974) or “double-loop learning” (Argyris, 1977) were needed, rather than simple, incremental change within the existing system. Such changes required a departure from these firms’ habitual, intuitive strategy-making modes, and temporarily rewarded a more comprehensive planning approach. It seems reasonable, therefore, that planning and profitability may have been correlated among apparel firms, both large and small, during this turbulent period.
Although this is a post hoc account of an unexpected empirical finding, it may provide an important insight into the apparently contradictory findings of Miller and Fredrickson. If this account is accurate, formal strategic planning may or may not be appropriate in an unstable environment, depending on the type of instability encountered. For ordinary, ongoing market instability, such as that in the sawmill industry studied by Fredrickson, planning is neither necessary nor desirable, especially for small firms, since the instability has become a functional fact of life, and can be managed through centralized authority, or an adaptive strategy-making mode, or both. However, for an external shock or industry crisis, the adaptive mode is no longer adequate, since the firm’s habitual behavior programs and routines become an important source of the problem itself. To cope successfully, firms may need to revert temporarily to a planning mode, scanning their environments and re-evaluating their fundamental strategic aims and competitive position. For this reason, planning may be effective under some types of industry instability, but not others; and, unless the nature of industry instability is specified, empirical studies — such as those of Miller and Fredrickson — might easily yield conflicting results.
Generic strategy as a contingency variable. H5 considers whether the correlation between planning and financial performance is greater among firms using cost leadership strategies than among those using differentiation strategies. Using a classification suggested by Miller (1987a), two kinds of differentiation strategies were considered, namely innovative differentiation and product novelty. Using responses on the generic strategy scales, firms were first divided into cost leaders (n-40), innovative differentiators (n=35), and stuck in the middle (n=28), and then into cost leaders (n=40), novelty differentiators (n=33), and stuck in the middle (n=30). The cutoff points were determined by considerations of sub sample size and statistical power, with stuck in the middle firms generally falling between 2.5 and 3.5 on the five-point scales.
Table 5 presents this analysis, and shows that the planning-profitability correlation is significantly greater among cost leaders (r=.47; p
CONCLUSIONS
The purpose of this study has been to advance the planning-performance research by examining the role of contingency variables — particularly firm size, industry, and strategy — in moderating the planning-performance relationship. In so doing, it was found that, although planning and profitability correlated directly with one another, this correlation was spurious, vanishing when firm size was held constant. Planning and sales growth, on the other hand, correlated significantly even after firm size effects were partialled. These results suggest that planning and profitability may not be generally associated, perhaps because of costs associated with the human and financial resources required to support a formal planning program. In this study, however, these resources did appear to be well-invested in the long run for firms seeking growth, as suggested by the significant planning-sales growth correlation. These results suggest, at a minimum, that future planning-performance studies should not be misled by direct planning-performance correlations, but should also consider the confounding effects of firm size.
Although planning and profitability were not generally associated, they were associated in subsamples formed according to key contingency factors. Specifically, planning was more nearly associated with profitability among large firms than among small ones, among firms in an unstable industry than among firms in a stable industry, and among cost leaders than among differentiators. As such, this research demonstrates how contingency factors can reveal critical dimensions of the planning-performance relationship, dimensions that remain hidden in heterogeneous samples, and in studies that do not account for contingency variables.
Concerning the impact of industry instability on the planning-performance relationship, the findings tended to support those of Miller and Friesen (1983). However, the conditions in the sample industries were such that it was possible to provide a provisional reconciliation of the earlier findings. In particular, it was possible to study the joint effects of size and industry stability to construct a reasonable reconciliation, namely that planning and performance may not ordinarily be associated in unstable industries (as Fredrickson found), but that certain kinds of instability — such as an industry-wide crisis resulting from foreign competition — may temporarily reward firms that plan comprehensively, especially those that scan their environments vigorously and systematically, analyze internal resources, and match their internal capabilities to the demands of the new environment. This explanation, though tentative, is testable and can be explored further in future research.
In the current study, the researcher instituted a number of controls to avoid pitfalls in previous planning-performance studies, and otherwise common to the survey methodology, including sampling from two homogeneous populations of undiversified firms, extensive pretesting and pilot testing, the use of proven, reliable scales, testing for interrater reliability, testing for the convergent validity of the performance measure, and comparing sample statistics with population parameters. To avoid misstating the significance of the correlations, only partial correlations were used, and infinite industry populations were assumed.
Nonetheless, the study is open to a number of fair criticisms. For example, although the researchers deliberately chose similar industries (i.e., mature, fragmented, consumer products manufacturing industries) in order to isolate the industry stability variable, it is still possible that another industry difference, and not the stability difference, might have produced the reported industry effects. Furthermore, the similarities between the industries implies that caution should be used in generalizing the findings to other types of industries, e.g., oligopolies, growth industries, or industries in the service, transportation, or retail sectors. The author is not aware of any confounding variable that might have produced the industry effects, nor of any reason why the planning-performance relationships described here would be invalid in other industry contexts.
Another problem, and one shared with other profitability studies, concerns causality and the cross-sectional design of the study. It has not been shown that any firm’s financial performance improved or declined as a result of adopting strategic planning – the study merely contrasts the performance of firms that employed varying planning practices during a fixed period of time. Furthermore, it is possible that, since profitable firms have more resources to devote to strategic planning, profitability may, in effect, cause strategic planning to arise (although, if this is true, it does not explain why profitable firms do not plan more than unprofitable firms in the furniture industry). Thus, the results in this study cannot strictly be construed as demonstrating a causal planning-performance relationship in the apparel industry, but merely an association between them, perhaps with bi-directional causation.
It should also be noted that this study, like previous planning-performance studies, is concerned only with financial measures of performance, and, specifically, with profitability and sales growth. The researchers acknowledge the existence of multiple other performance measures, and of multiple stakeholders, and do not argue that profitability and sales growth are more important than other measures — it is merely the measure with which the planning-performance studies have been principally concerned, and a measure readily amenable to analysis. Accounting data were employed rather than stock data because of the large proportion of privately-held firms in the population.
In taking a contingency approach, this study explored only those contingency hypotheses with strong ex ante theoretical justification (with a single exception – the joint industry-firm size effect — which was open to dispute in previous research). However, a less conservative approach might have yielded further results. For example, it would have been possible to subdivide the generic strategy subsamples to examine joint strategy-industry-size effects. Because the author had no ex ante theoretical expectations about such effects, they were not explored in this research. In any event, small subsample sizes would have precluded meaningful interpretation of the findings.
A final problem concerns a potential survivor bias in the data — if strategic planning caused organizations to fail completely, this study would not have detected it, since no non-surviving organizations were studied. This is a potential problem in the apparel industry, since the industry suffered a large number of Chapter 11 bankruptcies in the early 1980s and, of course, it was not possible to obtain planning data for these firms. The author knows of no reason why planning may have been an important cause of these failures (although non-planning may have been), which industry sources have widely attributed to low-cost foreign competition.
For future planning-performance research, several contributions would be appropriate. Current results are suggestive, but they do not entirely clarify the role of industry factors in shaping the planning-performance relationship. The explanation provided above is tentative, and certainly open to empirical testing in a wider range of industries. Furthermore, it may be worthwhile, using a contingency approach, to attempt a rigorous meta analysis of previous planning-performance research. Such a study would prove difficult at best, because of the widely differing methodologies and samples employed in previous studies. However, it may now be possible to begin reconciling this line of work. Finally, it may prove useful to examine the planning- performance relationship from the perspective of the resource theory of the firm (Barney, 1991; Wernerfelt, 1984). In order for strategic planning to contribute to firm performance, it must constitute a rent-producing strategic resource, i.e., it must be economically valuable, scarce, and difficult to imitate. Although resource theory has not yet proven its explanatory power, it may prove useful to derive empirical predictions based on this approach. It is hoped that the findings in this study can help to stimulate these further contributions.
APPENDIX MEASUREMENT SCALES
Strategic planning. Respondents were asked to indicate, on a scale of 0 to 5, the accuracy of twelve statements concerning their firms’ strategic planning processes. As a group, these items comprise the overall strategic planning scale. The items were anchored at either extreme with the words Very Accurate or Not at all Accurate. The statements, with the variable being measured in parentheses, were as follows:
1. We have broad, long-range goals known to all managers (GOALS)
2. We have specific, short-term goals known to all managers (GOALS).
3. Our firm’s actions are based more on formal plans than on intuition (ANALYS).
4. We have a manager or department devoted exclusively to formal planning (ANALYS).
5. We hold regular managers’ meetings to discuss overall strategy (ANALYS).
6. We use mathematical and computer models as planning aids (ANALYS).
7. We have a written plan for the next twelve months (ANALYS).
8. Our planning outlook is more long-term than short-term (ANALYS).
9. We search systematically for information about our competitors (SCANNG).
10. We use special market research studies (SCANNG).
11. We search systematically for new products, acquisitions, and investments (SCANNG).
12. Most of our decisions aim to solve immediate problems or crises (FUTUR).
Generic strategy. Respondents were asked to indicate, on a scale from 1 to 5, the accuracy of eight statements concerning their firms’ strategies. The scale was anchored at either extreme with the words Very Accurate or Not at all Accurate. The statements, with the variable being measured added in parentheses, were as follows:
1. We command a higher price than other firms by making a distinctive, high quality product (PRODIF).
2. Our prices are among the lowest in the industry (PRODIF).
3. We are often first to introduce innovative products (INNOV).
4. We spend more heavily on R&D than our competitors (INNOV).
5. We focus on a narrow, specific customer group (BREDTH).
6. Our production costs are among the lowest in the industry (PRCOST).
Financial Performance. Respondents were asked to indicate, on a scale of 1 to 5, how they perceived their firms’ profitability relative to other firms in their industry over the preceding three years. The scale was anchored at each point with the words Outstanding, Above average, Average, Below average, and Poor.
Formalization. Respondents were asked to indicate, on a scale of 0 to 5, the extent to which their firms use various formal organizational features. The scale was anchored at either extreme with the words Used to a great extent or Not used at all. These items were as follows:
1. Written employment contracts
2. An employee handbook (dealing with benefits, work rules, etc.)
3. A formal organizational chart
4. Written job descriptions
5. A preprinted form for employee performance appraisal
Centralization. Respondents were asked to consider six decisions, and to indicate the organizational level at which each decision would be made in their firm. The levels given were (1) Owner or CEO, (2) Upper management; (3) Middle management; (4) Lower management; and (5) Non-management. The six decisions were:
1. Setting delivery dates for orders
2. Choosing the type or brand of new computer equipment
3. Policy concerning overtime to be worked by shop workers
4. Accounting methods to be used
5. Suppliers to be used
6. Whether to introduce a new product
REFERENCES
Andrews, K. (1980). The concept of corporate strategy. Homewood, IL: Irwin.
Ansoff, H., Avner, J., Brandenburg, R., Porter, F., & Radosevich, R. (1970). Does planning pay? The effect of planning on success of acquisitions in American firms. Long Range Planning, 3, 2- 7.
Argyris, C. (1977). Double-loop learning in organizations. Harvard Business Review, 55, 115-125.
Armstrong, J. S. (1982). The value of formal planning for strategic decisions: Review of empirical research. Strategic Management Journal, 3, 197-211.
Blau, P., & Schoenherr, R. (1971). he structure of organizations. New York: Basic Books.
Bourgeois, L. (1980). Performance and consensus. Strategic Management Journal, 1, 227-248.
Bracker, J., & Pearson, J. (1986). Planning and financial performance of small, mature firms. Strategic Management Journal, 7, 503-522.
Brunsson, N. (1982). The irrationality of action and action rationality: Decisions, ideologies, and organizational actions. Journal of Management Studies, 19, 29-44.
Buzzell, R., Gale, B., & Sultan, R. (1975). Market share a key to profitability. Harvard Business Review, 53, 97-106.
Child, J. (1972). Organizational structure, environment, and performance: The role of strategic choice. Sociology, 6, 2-22.
Cohen, J., & Cohen, P. (1983). Applied multiple regression-correlation analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.
Cronbach, L.J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297-334.
Dess, G.(1987). Consensus on strategy formulation and organizational performance: Competitors in a fragmented industry. Strategic Management Journal, 8, 259-278.
Dillman, D. (1978). Mail and telephone surveys: The total design method. New York: Wiley.
Dun and Bradstreet (1987). Industry norms and key business ratios. New York: Author.
Dun and Bradstreet (1987). Million dollar directory Parsippany, NJ: Author.
Fredrickson, J., & Mitchell, T. (1984). Strategic decision processes: Performance in an industry with an unstable environment. Academy of Management Journal, 27, 399-423.
Fulmer, R., & Rue, L. (1974). The practice and profitability of long-range planning. Managerial Planning, 22, 1-7.
Gilbert, X., & Strebel, P. (1986). Developing competitive advantage. In H. Mintzberg, VB. Quiner, & R. James (Eds.), e strategy process (pp. 70-79). Englewood Cliffs, NJ: Prentice-Hall.
Grinyer, P., & Norburn, D. (1975). Planning for existing markets: Perceptions of executives and financial performance. Journal of the Royal Statistical Society A, 138, Part I, 70-97.
Hedberg, B. (1981). How organizations learn and unlearn. In. P. Nystrom & W. Starbuck (Eds.), Handbook of organizational design (pp. 3-27). New York: Oxford University Press.
Hickson, D, Pugh, D., & Pheysey, D., (1969). Operations technology and organization structure: An empirical reappraisal. Administrative Science Quarterly, 14, 378-397.
Hofer, C., & Schendel, D. (1978). Strategy formulation: Analytical concepts. St. Paul, MN: West.
Inkson, J., Pugh, D., & Hickson, D.(1970). Organization, context, and structure: An abbreviated replication. Administrative Science Quarterly, 15, 318-329.
Kallman, E., & Shapiro, H. (1978). The motor freight industry — a case against planning. Long Range Planning, 11, 81-86.
Kudla, R. (1980). The effects of strategic planning on common stock returns. Academy of Management Journal, 23, 5-20.
Lawrence, P., & Lorsch, J. (1967). Organization and environment. Boston, MA: Harvard.
Lindblom, C. (1959). The science of ‘muddling through’. Public Administration Review, 19, 79-88.
Lorange, P., & Vancil, R. (1977). Strategic planning systems. Englewood Cliffs, NJ: Prentice-Hall.
Malik, Z., & Karger, D. (1975). Does long-range planning improve company performance? Management Review, 64, 27-31.
Miller, D. (1987a). The structural and environmental correlates of business strategy. Strategic Management Journal, 8, 55-76.
Miller, D. (1987b). Strategy-making and structure: Analysis and implications for performance. Academy of Management Journal, 30, 7-32.
Miller, D. (1990). The Icarus paradox. New York: Harper & Row.
Miller, D., & Droge, C. (1986). Psychological and traditional determinants of structure. Administrative Science Quarterly, 31, 539-560.
Miller, D., & Friesen, P. (1983). Strategy-making and environment: The third link. Strategic Management Journal, 4, 221-235.
Miller, D., & Friesen, P. (1984). Organizations: A quantum view. Englewood Cliffs, NJ: Prentice-Hall.
Mintzberg, H. (1973). Strategy-making in three modes. California Management Review, 16, 44- 53.
Mintzberg, H. (1979). The structuring of organizations. Englewood Cliffs, NJ: Prentice-Hall.
Pearce, J., II, Freeman, E., & Robinson, R., Jr. (1987). The tenuous link between formal strategic planning and financial performance. Academy of Management Review, 12, 658-675.
Pearce, J., II, Robbins, D.K., & Robinson, R., Jr. (1987). The impact of grand strategy and planning formality on financial performance. Strategic Management Journal, 8, 125-134.
Porter, M. (1976). Interbrand choice, strategy, and bilateral market power. Cambridge, MA: Harvard University Press.
Porter, M. (1980). Competitive strategy. New York: Free Press.
Pugh, D., Hickson, D:, & Hinings, C. (1969). An empirical taxonomy of structures of work organizations. Administrative Science Quarterly, 14, 115-126.
Pugh, D., Hickson, D., Hinings, C., & Turner, C. (1968). Dimensions of organization structure. Administrative Science Quarterly, 13, 65-91.
Quinn, J. a (1980). Strategies for change: Logical incrementalism. Homewood, IL: Irwin.
Rhyne, L. (1986). The relationship of strategic planning to financial performance. Strategic Management Journal, 7, 423-436.
Robert Morris Associates (1987). Annual statement studies. Philadelphia, PA: Robert Morris.
Schoeffler, S., Buzzell, R., & Heany, D. (1974). The impact of strategic planning on profit performance. Harvard Business Review, 52, 137-145.
Shepherd, W. (1972). The elements of market structure. Review of Economics and Statistics, 54, 25-37.
Shrader, C., Taylor, L., & Dalton, D.(1984). Strategic planning and organizational performance: A critical appraisal. Journal of Management. 10. 149-171.
Standard and Poor’s Corporation (1987). Standard and Poor’s register. New York: Author.
Standard and Poor’s Corporation (1988). Industry surveys. New York: Author.
Starbuck, W., Greve, A., & Hedberg, B. (1978). Responding to crises. Journal of Business Administration, 9, 111-137.
Steiner, G. (1979). Strategic planning. New York: Free Press.
Thompson, A., & Strickland, A. (1987). Strategic management (4th ed.). Plano, TX: Business Publications.
Thune, S., & House, R. (1970). Where long-range planning pays off. Business Horizons, 13, 81-87.
U.S. Department of Commerce (1987). US. industrial outlook. Washington, DC: Government Printing Office.
U.S. Office of Management and Budget (1972). Standard industry classification manual. Washington, DC: Government Printing Office.
Van de Ven, A., & Ferry, D. (1979). Measuring and assessing organizations. New York: Wiley.
Watzlawick, P., Weakland, J., & Fisch, R. (1974). Change: Principles of problem formulation and problem resolution. New York: Norton.
Weber, M. (1947). The theory of social and economic organization. New York: Oxford University Press.
Welkowitz, Ewen, J.R., & Cohen, J. (1976). Introductory statistics for the behavioral sciences. New York: Academic Press.
Address all correspondence to Thomas C. Powell, Bryant College, 1150 Douglas Pike, Smithfield, RI 02917, U.S.A.
Copyright Administrative Sciences Association of Canada Jun 1994
Provided by ProQuest Information and Learning Company. All rights Reserved
