Shareholder value and investment strategy using the general portfolio model

Shareholder value and investment strategy using the general portfolio model

Stanley F. Slater

Resource Allocation Theory

A central question that diversified companies must answer is “How should resources be allocated across the business units in the corporate portfolio?” To create value for shareholders, resource allocation decisions should be consistent with the principle of modern financial theory, which states that only those investment opportunities that have a positive net present value should be funded. As Myers (1984: 128) explains,

“A strategic commitment of capital to a line of business is an investment project. If management does invest, they must believe the value of the firm increases by more than the amount of capital invested — otherwise they are throwing money away. In other words, there is an implicit estimate of net present value.”

Value-Based Planning

One approach that makes those estimates explicit is value-based planning (Rappaport, 1986). Value-based planning applies discounted cash flow techniques to analysis of strategic alternatives. However, there are numerous shortcomings associated with value-based planning (Day & Fahey, 1988) including (a) bias in forecasts of future activity, (b) invalid assumptions underlying the baseline scenario, (c) opportunistic behavior by managers, and (d) difficulty in estimating divisional cost of capital. Thus, though value-based planning is theoretically attractive, it is a complex methodology that is difficult to implement effectively.

Portfolio Planning

To cope with the many influences on the resource allocation decision, managers reduce the complexity of their decision-making environment through the processes of selective attention and simplification (Pfeffer & Salancik, 1978). Portfolio planning models, which were developed by consulting organizations such as the Boston Consulting Group and McKinsey and Company (Abell & Hammond, 1979), facilitate this process by focusing managerial attention on the broad constructs of competitive strength and market attractiveness.

Through the 1970’s and the 1980’s, portfolio planning techniques were widely adopted by U.S. corporations. In 1979 it was estimated that 36% of the Fortune “1000” and 45% of the Fortune “500” employed these techniques, with an additional 25 to 30 new firms introducing the approach every year (Haspeslagh, 1982). A 1984 survey indicated that Fortune “500” companies were maintaining or slightly increasing their use of planning techniques such as portfolio planning (Ramanujam & Venkatraman, 1987a & b). At a recent conference (Swartz, 1990), representatives from several major corporations such as GE, US West, and DuPont indicated that they continued to use portfolio management in the resource allocation process.

Portfolio planning techniques are attractive because they simplify the resource allocation problem:

“The strategic implications that are normally drawn from the matrix have, as a primary message, the assignment of investment priorities to the various businesses of the firm.” (Hax & Majluf, 1984: 174)

According to portfolio planning theory, investments should be made in business units according to their relative competitive position and the relative attractiveness of the industry in which the units compete. Resources should be concentrated in businesses that have a strong competitive position in an attractive industry and should be reallocated from weak units in unattractive industries, as depicted in Figure 1.

The presumption is that portfolio planning theory is consistent with modern finance theory. In other words, opportunities in competitively strong business units in attractive markets (Invest position) probably have a rate of return that exceeds their cost of capital. The rate of return on the majority opportunities in business units in the harvest section of the matrix is likely to be below the cost of capital.

However, the widespread use of portfolio planning suggests that a corporation might not be able to achieve a competitive advantage, and thus superior returns to shareholders, through its use because the majority of its competitors also use portfolio planning. This assumes, however, that different corporations implement portfolio planning similarly. Due to the variety of portfolio planning models (e.g., BCG, Life-Cycle, A.D. Little, McKinsey, etc.) and the variety of approaches for implementing a particular model (Wind, Mahajan, & Swire, 1983), there is substantial variability in the application of the theory by corporations. Furthermore, a resource-based theory of the firm (Barney, 1991; Wernerfelt, 1984) allows that some firms may be able to achieve above average returns, even if all firms use the same planning model, if they invest in unique, nonimitable resources. Therefore, realized excess returns may differ substantially among practitioners of portfolio planning.

Although portfolio matrices have been criticized for being overly simplistic representations of the complex influences on the resource allocation decision (e.g., Seeger, 1984; Wensley, 1981, 1982), Hax and Majluf (1991: 194) recently observed, “It is our experience that portfolio matrices can assist in bringing intelligent and appropriate communicational opportunities to the hard issue of portfolio management.” Therefore, it seems that though the consistency between modern financial theory and portfolio planning may be questionable, these techniques seem to be useful to and are used by managers.

However, portfolio models have been subjected to few empirical tests (Reiman, 1987: 48), none of which address whether adherence to the prescriptions of portfolio planning leads to superior corporate performance (Hamermesh, 1986:18). These criticisms and research deficiencies have stimulated calls for additional empirical analyses of portfolio planning techniques (Hambrick, et al., 1982: 529; Reiman, 1987: 48). Bettis and Hall (1981: 35) suggest that research on the corporate performance implications of portfolio planning, “should form an important agenda item for future research in business policy.” That these calls have gone unheeded is particularly troublesome in light of Camerer’s (1985: 3) criticism that many policy academicians are satisfied with, “an intuitive sense of correctness or plausibility” of their models.

This study tests whether an investment strategy that is consistent with prescriptions from the general portfolio model is positively associated with returns to shareholders. The study utilizes the Industry Attractiveness-Business Position matrix to examine the relationship between industry attractiveness, business position, investment strategy, and excess returns to shareholders. Results indicate that an investment strategy based on portfolio analysis may actually destroy shareholder value.

Research Design

Research on portfolio planning is complicated for several reasons. First, the positioning of business units in a corporation’s portfolio requires that numerous subjective judgements be made. Second, in a widely diversified corporation, portfolio management involves a large number of business units, which further complicates data access for researchers. Third, there are several potential portfolio models that a corporation could use, and these could result in different investment priorities being assigned to the same business unit. Finally, compilation of the necessary data from a primary source on each business unit in a large sample of corporations would be very difficult.

To address these problems we take a “realized” (Mintzberg, 1978) perspective on a corporation’s portfolio strategy. Regardless of whether the industry attractiveness – business position matrix is used, all corporations make investments that are either consistent or inconsistent with prescriptions from this matrix. This matrix is also the closest to being a “generic” model because the dimensions of all models represent, in some sense, industry or market attractiveness and competitive position. We operationalize all of the relevant constructs using secondary data from the Compustat database. Arguments for the validity of these measures are presented later. The database provides a sample of 129 corporations that have useable data. Although this research design may limit the ability to generalize our findings, we believe that this study is an important first step in the analysis of the corporate performance implications of portfolio planning.

Hypotheses

Specifically, this study tests two hypotheses. Hypothesis 1 is the central premise of portfolio planning, which is that corporations whose actual investment strategy is consistent with the prescriptions from the matrix should realize superior value creation. Value creation results from making investments in attractive opportunities, those that have a rate of return in excess of their cost of capital, and from reducing the size of investment in unattractive areas (Myers, 1984).

Rate of return is a function of market-level conditions and business-specific characteristics. Market conditions dictate the average profitability that can be expected in an industry. However, as Porter (1985) noted, industry structure does not dictate individual business profitability. Different businesses achieve rates of return that are above or below the industry average depending on whether they possess a sustainable competitive advantage. Because portfolio models use industry attractiveness and competitive position as their dimensions, they should predict a business unit’s likelihood of achieving a superior rate of return. Therefore,

Hypothesis 1: There is a positive association between an investment strategy that is consistent with prescriptions from the matrix and excess returns to shareholders.

Hypothesis 2 is derived from the central premise and from the value-based planning literature and is concerned with overinvestment (i.e., that is an excessive allocation of the corporation’s capital to a business unit that is at a competitive disadvantage or a unit in a relatively unattractive industry). In these business units it is difficult to find projects whose return on investment exceeds the cost of capital. Therefore, the rate of return for these business units should be relatively low. The value-based planning literature (e.g. Hax & Majluf, 1984: chapter 10; Rappaport, 1986; Reiman, 1987) predicts value destruction in this situation. Consequently,

Hypothesis 2: There is a negative association between overinvestment in relatively unattractive opportunities and excess returns to shareholders.

The other possible realized strategy is underinvestment, that is investment in relatively attractive opportunities but at a lower rate than is recommended by the matrix. Rates of return in these relatively attractive opportunities should be high and should probably exceed their cost of capital. Although depriving these businesses of capital that they could productively employ may not be a value-destroying strategy, it is also not a value-maximizing strategy. Also, underinvestment could result in a deterioration of competitive position that could destroy value in the long run. Due to the ambiguity of the relationship, underinvestment is not expected to be strongly associated with value creation or destruction.

Sample

The sample consists of 129 corporations reporting data on an average of 593 lines of business annually from 1983-1989 on PC-Compustat. The 7-year time frame is important as it allows the stock market time to react both to announcements of corporate investment plans and to announcements of the results from those plans. This permits us to examine the effect of investment strategy on excess returns to shareholders from the perspective of strategy as a pattern in a stream of decisions (Mintzberg, 1978) rather than from the event study perspective. Consequently, we are studying decisions and their outcomes instead of relying solely on investors’ expectations of the outcomes of those decisions.

Assignments of SIC codes to business groups have been criticized as somewhat subjective. However, it is Compustat (Standard & Poors), not the companies, that makes the assignments. Therefore, “(i)t is reasonable to assume that Compustat’s SIC code assignments are carefully and consistently executed,” and that this database is a reliable source of business-level information (Davis & Duhaime, 1989: 7).

To be selected, a firm had to meet the following requirements:

1. A primary SIC code between 2000 and 2799, or 3300 and 3799. This covers all manufacturing industries with the exception of chemicals, pharmaceuticals, and laboratory, photographic, test, and measurement equipment. These industries were excluded because firms in these industries tend to place greater emphasis on R&D expenditures than on capital spending in their resource allocations. Firms are not required to report line of business data on R&D expenditures.

2. No less than four lines of business identified by SIC code for the years 1983 through 1989.

3. Fiscal year ending in December to allow matching balance sheet data with common stock returns.

4. Price, share, dividend, and book value information available on PC-Compustat. Sales, operating profit, identifiable assets (those that can be uniquely attributed to the line of business), depreciation, capital expenditures, and SIC code for each line of business on PC-Compustat.

Variable Definitions

1. Unit of analysis: business group. It is common in strategic management to use portfolio analysis to determine missions for business units in the corporate portfolio. However, as Haspeslagh (1982: 70-71) pointed out, portfolio techniques are appropriate at multiple levels in the organization. The lowest level would be a portfolio of products within a business unit. The highest level of aggregation would be a portfolio of groups of related business units. An example of this is the sector system implemented at GE in 1977 to reduce cognitive demands on top management. A sector is a “macrobusiness or industry area” and the sector executive was responsible for integrating SBU strategies into a sector strategic plan (Bower, Bartlett, Christensen, Pearson, & Andrews 1991: 719) that was then presented to top management. Because 4-digit SIC (line of business, comparable to a business group) data is the only level available from Compustat, this study bases its analyses on the perspective of the corporation as a portfolio of business groups.

2. Matrix prescription. The position on the industry attractiveness-competitive position matrix is determined for each business group as follows.

Two dimensions of competitive position are used. These are the business group’s Return on Assets relative to industry (4-digit SIC) average ROA and its sales growth compared to the industry average. These indicators were used by GE in 1980 as two of three criteria for assessing competitive position (Bower et al., 1991: 729). The calculation of Industry ROA and sales growth for a given year includes all business segments (a corporation’s lines of business) identified by the 4-digit SIC code on PC-Compustat. This includes businesses beyond those in our sample.

We justify the use of relative ROA as an indicator of competitive position instead of attempting to assess the traditional subjective factors necessary for a business to achieve competitive advantage for the following reasons: (a) a business should not be able to earn an above average ROA in its industry if it does not possess a competitive advantage (Porter, 1985: xv); (b) Return on Investment (of which ROA is one form) is a business’s most important operational financial objective (Hill & Jones, 1989) because it is a primary driver of shareholder value creation (Varaiya, Kerin & Weeks, 1987); and (c) relative ROA commonly has been used to represent a business’s market power in both strategy and finance research (e.g., Barton, 1988; Bettis, 1981; Logue & Merville, 1972). We use business group sales growth also because of its widespread importance as a corporate objective and because of its traditional role in portfolio planning (e.g., the BCG matrix).

For a given year, all business groups within an SIC are ranked by ROA and by sales growth and rated as high, medium, or low on these dimensions. A strong competitive position is characterized by high or medium ROA and high sales growth, or by high or medium sales growth and high ROA, relative to all other business groups in the industry. A weak competitive position is characterized by medium or low ROA and low sales growth or by medium or low sales growth and low ROA, relative to all other business groups in the industry. A moderate competitive position is characterized as high-low, medium-medium, or low-high on ROA and sales growth, respectively. Thus, each business group is compared to all other groups in the same 4-digit SIC and placed in one of three equal-sized competitive groupings depending upon the group’s ROA and sales growth relative to its competitors.

We define industry attractiveness as industry (4-digit SIC) ROA relative to all-industry ROA and industry growth rate relative to all-industry growth. The industry attractiveness for each year was determined by ranking the ROA and sales growth rate for each industry and by forming three equal-sized groups using the same procedure as described for the determination of competitive position.

Again, industry growth is used because of its traditional role in defining industry attractiveness in portfolio models. In 1980, GE used market growth and profitability as two of six criteria for assessing market attractiveness (Bower et al., 1991: 729). Relative industry ROA is used instead of the traditional description of industry attractiveness as a composite of the strength of the forces of industry structure (e.g., buyer power, threats from substitutes, barriers to entry). Industry structure influences attractiveness because it affects “the prices, costs, and required investment of firms in an industry — the elements of return on investment” (Porter, 1985: 5). An industry return on investment in excess of the “free market” return will stimulate the inflow of new capital into an industry as it offers good potential for above average individual business rates of return (Porter, 1980: 5-6). In contrast, below market returns will cause capital to flee to more attractive opportunities. Although some of the forces that lead to high industry profitability are barriers to entry that make the industry unattractive to new entrants, all of the business groups in our sample are already in the industry. Consequently, relative industry profitability is a valid measure of industry attractiveness for this group of businesses. Current ROA is used instead of an expected figure because the business cannot know what future ROA will be. Jacobson and Aaker (1985) found that previous years’ ROI has substantial power for explaining current ROI. Thus, current ROA is a reasonable predictor for future ROA.

The matching of the competitive position and industry attractiveness determines a group’s position on the matrix. Business groups in cells labeled “Invest” are in very favorable positions on one dimension and are at least in a moderately favorable position on the other. Business groups in cells labelled “Maintain” are in moderately favorable positions on both dimensions or have a strength on one offset by a weakness on the other. Finally, groups in cells labelled “Harvest” are in relatively unfavorable positions on one or both dimensions, with no strengths to offset the weakness.

To determine whether our test is sensitive to the method of defining business group mission, we use four alternative business group ranking schemes. Consistency of results across the different models demonstrates robust findings. Model 1 uses the composite ROA-Sales Growth measures as described above. Model 2, the first alternative model, uses ROA only to rank the business groups.

Models 3 and 4 drop the assumption that there are equal numbers of business groups (3/9) in each strategy type. If opportunities are distributed somewhat normally, “invest” and “harvest” opportunities occupy the tails of the distribution, with “maintain” occupying the middle. Consequently, there are probably fewer “invest” and “harvest” opportunities than there are “maintain” opportunities. Therefore, Model 3 defines “Invest” business groups as those in very attractive industries (highest 1/3 based on average ROA) and with medium to strong competitive positions (top 2/3 based on relative ROA in the industry). Similarly, “Harvest” business groups are in very unattractive industries (lowest 1/3 on ROA) and medium to weak competitive positions. Model 4 defines “Invest” business groups as those in attractive industries (top 2/3 based on average ROA) and with strong competitive positions. “Harvest” business groups are in the bottom two-thirds of the industries but with weak competitive positions. Thus, under models 3 and 4, two-ninths of the sample are in the “invest” group, two-ninths are in the “harvest” group, and five-ninths fall in the “maintain” groups.

3. Business group’s actual investment strategy. The ratio of capital expenditures to total assets is an indicator of the business group’s actual investment profile for a particular year. Capital expenditure is only one type of resource allocation, albeit a very important one. Other resources that might be assigned on the basis of a portfolio position would include personnel, research and development funds, and advertising expenditures. However, corporations do not report data on these categories at the line of business level. As previously stated, firms from selected industries were excluded from this study due to their emphasis on R&D instead of capital investment in their resource allocations.

Consistent with our other operationalizations, we ranked the ratios of actual capital expenditure to total assets (CE/TA) for all business groups and assigned a “Growth” rating to those in the top third, a “Maintain” rating to the middle third, and a “Harvest” rating to the lower third.

4. Consistency of investment policies with matrix prescription. Consistency of actual investment strategy with matrix prescription is determined by matching the prescriptions of Figure 1 with the relative ranking of CE/TA for each year in the sample. Consistent investment occurs along the diagonal of the matrix in cells labeled 1, overinvestment occurs in groups located in cells labeled 2, and underinvestment occurs in groups in cells labeled 3.

Only capital expenditures attributed to a line of business were considered. The percentage of a corporation’s capital expenditures over the 7-year study period that are invested in business groups in cells labeled 1 is the value for consistency; the percentage invested in business groups in cells labeled 2 is overinvestment; and the percentage invested in groups in cells labeled 3 is underinvestment.(1)

Although the idea of investing in high ROA businesses in high ROA industries to create shareholder value may seem so intuitively obvious that it is not interesting, previous research has shown that the simple ROA-shareholder value linkage is not at all strong. Empirical analyses using accounting ROI as a surrogate for economic rate of return have been labeled as “totally misleading enterprises” (Fisher & McGowan, 1983: 91) and “of doubtful value” (Bentson, 1985: 64). In a study of the relationship between annual stock return and accounting ROI, Jacobson (1987) found an |R.sup.2~ of only 2%, statistically significant but not managerially meaningful. Consequently, demonstrating the basis for improved explanation of shareholder value creation would be a significant contribution.

5. Excess returns to shareholders. The objective of corporate strategy is to allocate resources so that they create value for the organization’s shareholders in excess of the risk-adjusted expected return to shareholders or, alternatively, in excess of the market return. Thus, the dependent variable is excess returns to shareholders over the study period. Annual common stock returns are calculated using PC-Compustat year-end closing prices and all dividend payments for each year. Excess returns for each year are calculated as the difference between the actual annual return to shareholders and an expected return based on the capital asset pricing model(2) (Fama et al., 1969). This figure provides a risk (represented by a systematic risk factor, |beta~) adjusted return to shareholders. The excess returns are compounded over the study period to determine geometric mean return of the annual excess returns over the 1983-1989 period.(3)

Analysis

The hypotheses are tested using the following cross-sectional regression model. By definition, the sum of the three independent variables is 1.0 or 100% of the investment for a given year. This calculation creates an exact collinearity with the normal regression intercept term and makes the regression matrix non-invertible. Therefore, tests of the three hypotheses are performed using regression through the origin, a no-intercept model. The difficulty with the no-intercept model is that the |R.sup.2~ is redefined and cannot be compared directly with |R.sup.2~s from standard regression models.(4)

|R.sub.i~ = ||beta~.sub.l~|C.sub.i~ + ||beta~.sub.2~|O.sub.i~ + ||beta~.sub.3~|U.sub.i~ + |epsilon~

where;

|R.sub.i~ – Average annual excess returns to shareholders over the 1983-1989 period for firm i.

|C.sub.i~ – % of investment consistent with matrix prescriptions from 1983-1989 for firm i.

|O.sub.i~ – % of investment where actual allocation is above prescribed from 1983-1989 for firm i.

|U.sub.i~ – % of investment where actual allocation is below prescribed from 1983-1989 for firm i.

In addition to the full model, we drop the underinvestment variable as it is the least theoretically interesting and because, as presented in Table 3, its coefficient is not significant in the full model. A traditional OLS regression model with intercept can then be calculated along with the standard |R.sup.2~.

To further assess the robustness of our test, we estimate generalized least squares models using a pooled, time-series, cross-sectional analysis of each year’s investment profile and excess returns to shareholders. The results of these tests show a reduction in the overall significance levels of the independent variables, but no change in signs. We also test whether firm size (measured as the market value of the firm’s equity) influences our results by including firm size as an additional independent variable. We find that firm size does not affect the relationships between the investment strategy variables and excess returns to shareholders. (Contact the second author for any unreported results from the analyses.)

Results

Descriptive statistics and intercorrelations are displayed in Tables 1 and 2. TABULAR DATA OMITTED Table 1 provides summary information at the firm and business group level. The mean level of firm sales over the 7-year period is $3.8 billion. Net income of the firms in the sample averages over $156 million. Total assets average roughly $3.5 billion and capital expenditures over the 7 years averages $223 million. Average firm size based on market value of the common stock is $2.1 billion. Table 2 contains descriptive statistics and intercorrelations for the independent and dependent variables.

The results of the regression analyses are displayed in Tables 3 and 4. Table 3 contains the results of the analysis when the data is combined over the entire study period. Table 4 has the GLS results using the pooled data. The top half of each table shows the results from the no-intercept models and the bottom half shows the results when the “underinvest” variable is dropped from the analysis.

The results reported for the no-intercept models do not support hypothesis 1. In fact the association between consistency and excess returns is negative and significant in all four of the models. Consistent with hypothesis 2, the coefficients for the overinvestment variable are negative and significant in all models, indicating that overinvestment is a value-destroying strategy. As expected, the coefficients for underinvestment are not significant in three of the four no-intercept models. TABULAR DATA OMITTED TABULAR DATA OMITTED Consequently, underinvestment is dropped from the analysis and regression with the intercept is conducted.

In the OLS models with underinvestment excluded, the equations range in significance from p|is less than~.01 to p|is less than~.001 and explain from 6% to 21% of the variance (adjusted |R.sup.2~) in excess return to shareholders. Relative to other studies that rely on accounting based measures to predict market returns (e.g., Jacobson, 1987; Jose, Nichols, & Stevens, 1986; Varaiya, Kerin, & Weeks, 1987; Woo, 1985), the overall explanatory power of these models is notable.

Contrary to the prediction of hypothesis 1, the coefficient of the consistency variable is negative (p |is less than~.01 in three out of four models analyzed).

Consistent with hypothesis 2, the coefficient for overinvestment is negative and statistically significant in all of the models. The empirical evidence based on the regression results supports the hypothesis that overinvestment in unattractive opportunities is associated with destruction of shareholder value.

The results from the GLS pooled models reported in Table 4 are qualitatively similar to the results reported in Table 3. There is a moderate change in the absolute value of the parameter estimates towards zero. However, the coefficients, in general remain negative, are statistically significant, and follow the pattern of the coefficients reported in Table 3.

TABULAR DATA OMITTED

Discussion and Suggestions for Future Research

Not only is investment that is consistent with prescriptions from the general portfolio model not positively associated with creation of shareholder value, it appears to be associated with value destruction. This is true whether a risk adjusted or market adjusted measure of return to shareholders is used, and it is true in four different formulations of the business groups’ matrix position. As noted earlier, portfolio planning techniques have been criticized on a number of points. Several of those points seem relevant to understanding why portfolio planning might lead to value destruction.

As Wensley (1981) suggests, the correspondence between portfolio planning techniques and discounted cash flow (DCF) techniques may not be very high. DCF techniques are forward looking and are based on estimates of future cash flows from projects or businesses. In contrast, definition of competitive position and industry attractiveness are often based on historical information that may be irrelevant for the future. This could result in the corporation investing in businesses where competitive position cannot be maintained for reasons that are not apparent in backward scanning (Seeger, 1984). Furthermore, the generic recommendations from the portfolio (Invest, Maintain, Harvest) have “limited managerial value and could be very myopic in terms of offering creative alternatives to top management” (Kerin, Mahajan, & Varadarajan, 1990: 94). In other words, the prescriptions offer little guidance for developing and sustaining competitive advantage, the true source of value creation.

The results of this study indicate that portfolio planning techniques should be used very cautiously. Reliance on historical information about businesses and markets is likely to lead top management to overestimate or to miss opportunities. In either case, the outcome is a misallocation of resources. Portfolio planning techniques can be conceptually useful devices for understanding a business’s prospects. However, those prospects must be based on a thorough analysis and realistic assessment of current and future customer needs and market structure and the unit’s requirements for and likelihood of developing and sustaining competitive advantage.

There are two important implications of this study for scholars. First, given the findings from this study, we must be very careful of how we teach portfolio planning techniques. It is imperative that we impress on our students the fact that naive or simplistic application of the models can be more destructive than productive. However, it is important to remember that it was a simplistic model that was tested in this study, not how corporate managers manage portfolio planning. The many conceptual and methodological criticisms about portfolio planning that have been raised over the years obviously remain as concerns.

Second, as use of portfolio planning remains high, an important area for future research is in the implementation of portfolio planning. Research by Wind, Mahajan, & Swire (1983) showed that assignment of business unit mission (invest, maintain, harvest) can vary depending on the selection of a particular portfolio model (i.e., BCG vs. GE-McKinsey vs. Life-cycle). Although this study examined four operationalizations of the general model, research comparing the effectiveness of different models would be very useful. Another important area for research is on the suitability of different control systems, management styles, reward systems, and other implementation characteristics for portfolio planning and management (see Hamermesh, 1986 for several case studies of portfolio planning implementation).

Although this study offers useful insights into the relationship between investment strategy based on the general portfolio model and creation of shareholder value, there are limitations to its external validity. One limitation is that this study evaluates the results of realized strategy, not intended strategy. Thus, it cannot be interpreted as an analysis of the way that portfolio planning is employed. Aside from the fact that previous surveys show that large companies are likely to use these techniques and that the companies in this sample are relatively large, there is no evidence that the companies in the sample actually use portfolio planning to make resource allocation decisions.

Another limitation of this study is the use of historical secondary data for measurement of competitive position, industry attractiveness, and resource allocations. Although business group ROA and sales growth relative to industry norms should be indicators of competitive position and industry ROA and sales growth relative to all-industry norms should be associated with industry attractiveness, it is accepted that competitive position and industry attractiveness are qualitative as well as quantitative constructs. Furthermore, these measures do not recognize the possibility that a business would characterize its competitive position or an industry’s attractiveness differently based on future prospects than it would based on historical performance. Finally, this study does not consider resource allocations besides capital investment. Consequently a corporation’s definition of position, attractiveness, and investment strategy could be different from that used in this study. These limitations notwithstanding, this study is a first step in the analysis of the central premise of portfolio planning.

1 The total of consistent, overinvestment, and underinvestment will not sum to 1.0 for all firms due to missing data. Missing data occur most often because industry (SIC) comparison data for the business group is unavailable.

2 Specifically, for each year we calculate the firm’s actual return based on capital gains and dividends. The excess return is calculated as E|R.sub.it~ = |R.sub.it~ – ||r.sub.Ft~ + ||beta~.sub.i~ (|R.sub.mt – |r.sub.Ft~)~; where |R.sub.it~ is the actual return, |r.sub.Ft~ is the one year treasury rate for the years 1983-1989, |beta~ is the beta for the firm reported by Compustat, and |R.sub.mt~ is the S&P 500 with dividend rate of return for the years 1983-1989.

3 Additionally, we calculated market adjusted returns using the S&P 500 to represent the market. This is a common methodology employed in the finance literature (Brown & Warner, 1985) and is given as: E(|r.sub.it~) = R|i.sub.lt~ – R|m.sub.lt~. R|i.sub.lt~ is the total return in the year t for firm i. R|m.sub.lt~ is the total return in year t for the S&P 500. There are no substantive differences in the results using this procedure to calculate excess returns. Thus, we report only the results based on the CAPM excess returns.

4 Specifically, in a normal regression model with an intercept the |R.sup.2~ is defined as the ratio of the sum of squares regression/corrected total sum of squares. In the no-intercept model, the denominator is not corrected for the mean and is simply |summation~|y.sup.2~. Thus, the |R.sup.2~ in the no-intercept model is given as the ratio of sum of squares regression/uncorrected total sum of squares.

References

Abell, D.F., & Hammond, J.S. 1979. Strategic market planning. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Alberts, W.W., & McTaggart, J.M. 1987. Value based strategic investment planning. Interfaces, 14 (1): 138-151.

Barney, J. 1991. Firm resources and sustained competitive advantage. Journal of Management, 17 (1): 99-120.

Barton, S.L. 1988. Diversification strategy and systematic risk: Another look. Academy of Management Journal, 31 (1): 166-175.

Bentson, G.J. 1985. The validity of profits-structure studies with particular reference to the FTC’s line of business data. American Economic Review, March: 37-67.

Bettis, R.A. 1981. Performance differences in related and unrelated diversified firms. Strategic Management Journal, 2: 379-393.

Bettis, R.A., & W.K. Hall. 1981. Strategic portfolio management in the multi-business firm. California Management Review, Fall: 23-38.

Bower, J.L., Bartlett, C.A., Christensen, C.R., Pearson, A.E., & Andrews, K.R. 1991. Business policy text and cases. Homewood, IL: Irwin.

Brown, S.J., & J.B. Warner. 1985. Using daily stock returns: the case of event studies. Journal of Financial Economics, 14: 3-31.

Camerer, C. 1985. Redirecting research in business policy and strategy. Strategic Management Journal, 6: 1-15.

Davis, R., & Duhaime, I.M. 1989. Business level data disclosed under FASB No. 14: Effective use in strategic management research. Proceedings of the National Academy of Management: 7-11.

Day, G., & Fahey, L. 1988. Valuing market strategies. Journal of Marketing, 52 (July): 45-57.

Fama, E., Fisher, L., Jensen, M., & Roll, R. 1969. The adjustment of prices to new information. International Economic Review, 10 (1): 1-21.

Fisher, F.M., & McGowan, J.J. 1983. On the misuse of accounting rates of return to infer monopoly profits. American Economic Review, March: 82-97.

Hambrick, D., MacMillan, I., & Day, D. 1982. Strategic attributes and performance in the BCG matrix — A PIMS-based analysis of industrial product businesses. Academy of Management Journal, 25 (3): 510-531.

Hamermesh, R.G. 1986. Making strategy work. New York: John Wiley.

Haspeslagh, P. 1982. Portfolio planning: Uses and limits. Harvard Business Review, January-February: 58-73.

Hax, A., & Majluf, N. 1984. Strategic management: An integrative perspective. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Hax, A., & Majluf, N. 1991. The strategy concept and process. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Hill, C., & Jones, G. 1989. Strategic management theory. Boston, MA: Houghton Mifflin.

Jacobson, R. 1987. The validity of ROI as a measure of business performance. American Economic Review, 77 (3): 470-478.

Jacobson, R., & Aaker, D. 1985. Is market share all that it’s cracked up to be? Journal of Marketing, 49 (Fall): 11-22.

Jose, M.L., Nichols, L.M., & Stevens, J.L. 1986. Contributions of diversification, promotion, and R&D to the value of multiproduct firms: a Tobin’s q approach. Financial Management, Winter: 33-42.

Kerin, R.A., Mahajan, V., & Varadarajan, P.R. 1990. Strategic market planning. Needham, Heights, MA: Allyn & Bacon.

Logue, D.E., & Merville, L.J. 1972. Financial policy and market expectations. Financial Management, 1 (2): 37-44.

Lorange, P., & Murphy, D. 1984. Considerations in implementing strategic control. Journal of Business Strategy, 4 (4): 27-35.

MacMillan, I.C., Hambrick, D.C., & Day, D.L. 1982. The product portfolio and profitability — A PIMS-based analysis of industrial-product businesses. Academy of Management Journal, 25 (4): 733-755.

Mintzberg, H. 1978. Patterns in strategy formulation. Management Science, 24: 934-948.

Myers, S.C. 1984. Finance theory and financial strategy. Interfaces, 14 (1): 126-137.

Pfeffer, J., & Salancik, G.R. 1978. The external control of organizations: A resource dependence perspective, New York: Harper & Row Publishers.

Porter, M. 1980. Competitive strategy, New York: The Free Press.

Porter, M. 1985. Competitive advantage, New York: The Free Press.

Porter, M. 1987. From competitive advantage to corporate strategy. Harvard Business Review, May-June: 43-59.

Ramanujam V., & Venkatraman, N. 1987a. Planning and performance: A new look at an old question. Business Horizons, May-June: 19-25.

Ramanujam V., & Venkatraman, N. 1987b. Planning system characteristics and planning effectiveness. Strategic Management Journal, 8: 453-468.

Rappaport, A. 1986. Creating shareholder value. New York: The Free Press.

Reiman, B. 1987. Managing for value. Oxford, OH: Planning Forum.

Seeger, J.A. 1984. Reversing the images of BCG’s Growth/Share Matrix. Strategic Management Journal, 5 (1): 93-97.

Snow, C., & Hambrick, D. 1980. Measuring organizational strategies: Some theoretical and methodological problems. Academy of Management Review, 5 (4): 527-538.

Swartz, G.S. 1990. Organizing to become market-drive. Marketing Science Institute Conference Summary, Report # 90-123.

Varaiya, N., Kerin, R., & Weeks, D. 1987. The relationship between growth, profitability, and firm value. Strategic Management Journal, 8 (5): 487-497.

Wensley, R. 1981. Strategic marketing: Betas, boxes, or basics. Journal of Marketing, 45 (Summer): 173-182.

Wensley, R. 1982. PIMS and BCG: New horizons or false dawn? Strategic Management Journal, 3 (2): 147-158.

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

Wind, Y., Mahajan, V., & Swire, D. 1983. An empirical comparison of standardized portfolio models. Journal of Marketing, 47: 89-99.

Woo, C. 1984. An empirical test of value-based planning models. Management Science, September: 1031-1050.

COPYRIGHT 1992 JAI Press, Inc.

COPYRIGHT 2004 Gale Group