MULTIPLE BUSINESS GOALS AND MULTfilLE BusINESS !NSlRUMENTS: THE APPliCATIoN OF TIlE THEoRY OF MACROECONOMIC POliCY TO BusINESS DECISIONS Beheruz N. Sethna Larry Allen Dee Wellan Lamar University Beaumont. Texas Introduction A manager is often called upon to make decisions regarding the allocation of scarce resources between two or more business, management, or marketing instruments in order to be able to achieve a certain goal. The analysis of such a problem is not new to business. However, when there is more than one goal to be achieved, each of which is dependent upon a set of business, management, or marketing instruments, the prob- lem is much more complex. The purposes of this paper are to: • Analyze the situation of multiple business instruments being used to achieve multiple business goals, using selected principles of macroeconomics, • To present recommendations for the use of such instruments that suggest that instruments with comparative advantage should be selected to achieve each goal, and • To illustrate how violation of the principles presented in the paper might lead to undesirable results. The theory of economic policy may effectively be used to understand and help re- solve such problems without restricting the form of the function as multivariate statis- tical methods usually require. An additional advantage is that it provides a dynamic analysis rather than a comparative-static one. The present work is essentially an ap- plication to business or management, of the theory of macroeconomic policy in a multiple goal, multiple instrument setting. Tmbergen [7], a pioneer of economic policy in a multiple goal and instrument con- text, developed the Tinbergen principle, which states that the number of policy instru- ments must be equal to the number of goals if all the goals are to be achieved. This applies in the context of a fIxed target value for each goal, as is common in business The fIrst author wishes to acknowledge the assistance of Dr. Sonmez Atesoglu, Pro- fessor of Economics at Oarkson Univesrity, in the early stages of the paper. Spring 1991 Sethna, Allen & Wellan: Multiple Business Goals 53 practice. If the decision-maker can make a trade"-off between goals. then the instru- ments are only required to be less than or equal to the number of goals. This is the case of variable targets. as discussed by TInbergen [6] and by Theil [5]. The variable targets approach requires the policy maker to be able to specify a continuous trade-off function Qike the utility function) between the two goals. Although some trade-off is possible between goals, such a "utility function" is difficult (at best) to specify in management practice. It is much more probable that a manager is required to meet specific fixed target values in particular target or goal areas, such as profit ($X per unit), market share (Y%), etc. Consequently, the fixed targets approach is more ap- plied even in actual economic policy issues. For the above reasons, it is the fixed targets situation that is analyzed in this paper. A major application of the fixed targets approach was presented by Mundell [3, pp. 70-77]. Mundell's principle of effective market classification states that policy instru- ments should be aimed at the goals on which they have the greatest influence, or in other words, the greatest comparative advantage. In his review of Mundell's analysis of this principle in the context of the use of monetary and fiscal policy (two instru- ments) to achieve internal and external economic equilibrium (two goals), Makin has observed that, while Mundell's basic principle seems reasonable enough, "it is not imme- diately obvious that its violation can lead to rather disastrous results [2, pp. 460-461]." The current work makes use of the fact that there are different effects on goals of the various business instruments controlled by a manager. Such effects were found by Stowsand and Wenzel [4, pp. 243-257] in eight different product groups. Also, Cronin and Skinner [1, pp. 9-22] found that several business instruments (ROA, li- quidity, leverage, etc.) had different effects on business outcomes (profitability, [man- cial structure, and marketing outcomes). In this paper, the above analysis of macroeconomic policy will be applied in a busi- ness context to assist in dynamic analysis of the problem of multiple business goals and multiple business instruments. The conclusions of the analysis will be shown to be consistent with good business practice. It will also be demonstrated how failure to conform to the derived guidelines may lead to "disastrous" results. Analysis Consider the case of a brand for which the manager must determine the appropriate business strategy. There are many instruments of strategy available to the manager, and any or all of the elements of the marketing mix may be used to achieve a par- ticular target or goal. In deciding the business plan for the brand, the manager may allocate resources between the instruments he or she controls, depending upon their relative effect on the brand. Such an analysis of effects, whether intuitively done or based on multivariate analysis, is not new to the field of business. However, when two or more (sometimes conOicting) targets have to be met, the problem is much more complex. It is this problem that is analyzed here. 54 Journal of Business Strategies For the purpose of this analysis, consider two instruments: 1. Unit Price (P) of the brand, 2. Advertising expenditure (A) per sales period. Vol. 8, No.1 Only two business instruments have been chosen for convenience of illustration. The mathematical analysis is generalizable and may be extended to any number of instru- ments. Any or all of the instruments that a manager controls, including the elements of the marketing mix, may be used. In this paper, the analysis is concerned with the meeting of two goals: 1. Per Unit Profit (C), 2. Share of Market (at the end of the sales period) (M). These two target areas have been chosen because of their importance and the fre- quency of their occurrence in business practice; and because they often require con- flicting inputs, which makes the problem more interesting. As in the case of the in- struments, the analysis is generalizable and may be extended to any number of tar- gets. Note that the design is consistent with the Tinbergen principle which states that where there are fixed target values for policy goals, the number of policy instruments must be equal to the number of goals if all the goals are to be obtained. Suppose that a target profit per unit for the brand is set at C* (dollars) and that the desired market share is targeted at M*. These targets may be determined at any level of the company. For example, the target profit, C* may be $1 per unit for a rela- tively low-price, high-frequency consumer purchase item; and the target market share M* may be 10%. These figures are only illustrations and would be vastly different for industrial goods or other products. Each of the two target areas are affected by the two business instruments mentioned above. These may be expressed in a completely general form as: C = f(p,A) (1) or, in other words, Per Unit Profit is some Qinear or non-linear) function of Unit Price and Advertising Expenditure and M = f(p,A) (2) or, in other words, Market Share is some (linear or non-linear) function of Unit Price and Advertising Expenditure. Spring 1991 where: Sethna, Allen & Wellan: Multiple Business Goals C = Per Unit Profit M = Market Share P = Unit Price A = Advertising Expenditure 55 Further, it appears reasonable to assume that, for many products; as price increases, with other variables being held constant (ceteris paribus): • Per unit profit increases, at least in the short term, because more money is available for the bottom line. (3) • Market share decreases, since customers buy less of that brand and more of competing brands or products. This is the case of the classic downward-sloping demand curve, often (though not always) found in practice. (4) As advertising expenditure increases with other variables being held constant: • Per unit profit decreases, since less money is available for the bottom line (5) • Market share increases (assuming that the advertising is any good) because more people are made aware of the brand, acquire more knowledge of the brand, and perhaps are persuaded to buy the brand. (6) Expressions (1) through (6) are quite acceptable intuitively and have been found to be substantiated by numerous academic and corporate findings. They are working as- sumptions only; violation of these will not invalidate the procedure (although the re- sults will be modified accordingly). We use the notation MA to denote the incremental change in market share (M) re- sulting from a unitary change in advertising, with all other variables remaining con- stant. Graphically, it may be thought of as the slope of the market share (M) line with respect to advertising expenditure (A). (For those readers more comfortable with calculus, MA is the partial derivative of M with respect to A, 6M/fiA) Similarly, we use the notation Mp to denote the incremental change in market share (M) resulting from a unitary change in price, with all other variables remaining constant). Graphi- cally, it may be thought of as the slope of the market share (M) line with respect to price (P). (Mp is the partial derivative of M with respect to P, 6M/6P.) In business practice, the usual way of determining the incremental effect of one variable upon another is to do an empirically-based analysis. For example, a regres- sion analysis may be done with Market Share (M) as the dependent variable, and Advertising Expenditure (A) and Price (P) as the independent variables. Then, the regression coefficients for A and P represent MA and M~ respectively, because they represent the incremental change in M resulting from incremental changes in A and P respectively, with the other variable being held constant. 56 Journal of Business Strategies Vol. 8, No.1 We can then rephrase Expressions (3) through (6), using the above notation for the sake of brevity: As price increases, profit increa~s, so As price increases, market share decreases, so As advertising increases, profit decreases, so As advertising increases, market share increases, so (7) (8) (9) (10) Given that both business instruments (price, P, and advertising, A) affect goals (per unit profit, C, and market share, M), we can find various combinations of P and A that can achieve together levels of C and M. For example, if our targeted market share is M* or 10%, it may be that a price of $5 per unit and advertising expenditure of $40,000 per year may achieve this target Now if we increase the price to $6, market share may be in danger of dropping be- low the desired 10% level. However, by increasing the advertising level to say, $45,000 per year, (perhaps singing the praises of an upscale brand) the manager may be able to avert any decline of market share, leaving M at the desired 10% level. So, we now have two combinations of P and A that lead to the desired market share, M* or 10%: these are (p=$5 and A=$40,OOO) and (p=$6 and A=$45,OOO). Similarly, several other combinations of P and A may be found; for example (p=$6.50 and A=$55,OOO), (p=$4 and A=$30,OOO), etc. may also result ia a market share of 10%. Note that, as we increase price, we must increase advertising (and vice versa) to remain at the desired level of market share. This is a result of the basic relationships discussed in Expres- sions (3) - (6); the numbers, however, are for illustrative purposes only. The verbal arguments in the preceding paragraph lead to the generalizable result that several combinations of P and A may be found, that lead to the targeted value, M*, of market share. These combinations of P and A may be found experimentally or math- ematically (by use of the calculus) once Expressions (1) and (2) are empirically deter- mined. For example, regression analysis on corporate data with Market Share as the dependent variable and Price and Advertising Expenditure as independent variables is one way of empirically determining Expression (2): M = f(p,A). In a similar vein, different combinations or trade-offs between P and A can lead to the desired per unit profit, C*. For example, if a price of $4.50 and advertising of $38,000 per year achieve the desired per unit profit of $1, we may be able to retain that $1 profit per unit by simultaneously increasing the price to $5.00 and increasing the advertising to $40,000. Note that, as in the analysis of market share, an increase of price must be combined with an increase in advertising (and vice versa) to pre- serve the targeted per unit profit C*. Figure 1 shows two lines: C=C* and represents the combination of P and A that lead to the targeted per unit profit of C*, and M=M* represent the combinations of P and A that yield the targeted market share of M*. Note that both lines slope upwardS; this is consis- tent with the conclusions of the precffiiDg paragraphs that showed that, to preserve the tar- geted levels of C* and M*, if we increase P we must increase A, and vice versa. Spring 1991 Sethna, Allen & Welian: Multiple Business Goals Figure 1 57 Case 1: Assumes Comparative Advantage for Advertising in Affecting Market Share and for Price in Affecting Profit M=M* Region ill cM* t Price C>C* MC* M>M* Region n Advertising --..,. Figure 1 is drawn such that the slope of the M=M* line is greater than the slope of the C=C· line. This is a representation of comparative advantage for advertising in affecting market share and for price in affecting profit. That is, market share is more sensitive to advertising expenditure changes than to price changes. The justifi- cation for the preceding statement is not immediately obvious and may be provided by a mathematical justification using differential calculus or by using graphical and verbal arguments. The latter approach is favored here and is described in the follow- ing paragraphs. If advertising expenditure has a relatively large effect on market share (M) and price has a small effect on M, we might expect the following to occur. As an example, let us start with one of the points of the M=M* line: let us say that the targeted market share of 10% is achieved when Price (P) = $5 per unit and Advertising (A) = $40,000 per year. Now, if we increase the price to $6 (a 20% increase), that would, ceteris paribus, lead to a loss of customers because of the downward-sloping demand curve. To prevent this from occurring, we increase advertising. If advertising has a large 58 Journal of Business Strategies Vol. 8, No.1 relative effect on market share, we may need to increase advertising only by, say, $1,000 (a 2.5% increase; this high impact of advertising is exaggerated for illustration). So to move from one point on the M=M* line to another point, we have a relatively large movement on the P (price) dimension for a relatively small movement on the A (Ad- vertising) dimension. This corresponds to the relatively steep slope of the M=M* line in Figure 1. Further, if price has a relatively large effect on per unit profit (C) and advertising a relatively small effect or C, we might expect the following scenario. As an illustra- tion, we consider one of the points on the C:C* line: price (P) = $4.50 and advertis- ing (A) = $35,000 per year. Now, if price has a relatively large effect on profit (be- cause of the product's cost structure), and we increase price to $5 (approximately an 11% increase), we may be able to remain at our desired profit level even in spite of a relatively large increase in advertising, say, by $7,000 (a 20% increase). So, to move from one point on the C:C* line to another, we have a small increase in P accompa- nied by a large increase in A This shows up as a low slope for the C:C* line in Figure l. Figure 1, then, represents a case where advertising has a high effect (relative to price) on market share, and price has a high effect (relative to advertising) on per unit profit. This is the case of comparative advantage for advertising in affecting market share and for price in affecting per unit profit. The reverse case will be discussed later (Figure 2). Figure 1 may be divided into four regions formed by the intersection of the M=M* and C:C* lines. These are labeled Regions I, II, III, and N. For any value of ad- vertising, a point above the C:C* line represents a higher price than the one that achieves the desired per unit profit; it therefore represents a profit higher than C* (since, as price increases, ceteris paribus, so does per unit profit). So all points above the line C:C* represent values of C which are greater than C*. Hence, Regions I and II are labelled as C>C*. Correspondingly, points below the line C:C* (Regions ill and IV) are labelled as CM*. Correspondingly, Regions I and IV are labelled as MC* or CM* or MC*), market share is less than the target (M 0, CA < 0, Mp < 0, MA > 0, as described Expressions (7) - (10», which implies that either instrument may be used to achieve either goal. However, this approach may lead to potentially disastrous results, as the following analysis will show. As mentioned earlier, CA is the notation used to denote the incremental change in profit from a unitary change in advertising, with all other variables remaining constant. It may be thought of as the slope of the per unit profit (C) line with respect to adver- tising expenditure (A). Alternatively, it is the partial derivative of C with respect to A Let us assume that, as stated in Expression (9), we find that CA < °based on an empirical study. This coefficient implies that, ceteris paribus, we expect a decrease in C if A increases, and vice versa. In Figure 1, consider Point W where C < C* and M < M*. If the principle of effective market classification is violated or ignored, very undesirable results may re- sult. In order to increase per unit profit, advertising is cut back to a level corresponding to V which, in fact, decreases market share. In order to increase market share, price is decreased (since the empirical study showed the commonly-found result of Mp < 0), moving the brand further away from the desired target. Here, it does not matter which comes first - the advertising cut-back or the price decrease; the result is not an expected or desirable one, in spite of the above (intrinsically sound) business strat- egies. 60 Journal ofBusiness Strategies Vol. 8, No.1 Let us now consider using Mundell's principle, which is concerned with the dynamic analysis-the method to be followed to achieve the targets. Since price has the com- parative advantage in affecting ~ Mundell's principle would suggest increasing price to that corresponding to X where C=C*. Since advertising has the comparative ad- vantage in influencing M, the principle would suggest that advertising be increased to that corresponding to Y (M=M*) so that the brand may be moved from X to Y In a similar fashion, by price and advertising increases, the brand can move towards Q and so achieve both targets. There is no theoretical significance to starting with a price increase. In a given business situation, it may be more appropriate to start with an increase in advertising input. It must be emphasized that these results occur because of the relative comparative advantages of price and advertising in affecting the two target areas (profit and mar- ket share) represented by the relative slopes of the two lines, C=C* and M=M*, in Figure 1. The opposite case would arise if the comparative advantages were reversed. This is discussed in Case 2, which follows. FJgUre 2 Case 2: Assumes Comparative Advantage for Advertising in Affecting Profit and for Price in Affecting Market Share C=C* t Pri c>c*Mc* M>M* Advertising --~ Region n Region ill Figure 2 is divided into four regions, bounded by the lines C=C* and M=M*, respec- tively. Each one of these regions (except for Region ~ in which C > C*, M > M*) Spring 1991 Sethna, Allen & Wellan: Multiple Business Goals 61 represents a strategy problem because at least one target is not met. Consider, for example, the dilemma faced by the manager in Region IT represented by C 0) and advertising (MA > 0), without considering comparative advantage, the advertising input may be increased (in an attempt to increase market share) to a level corresponding to \l, then price may be increased (in an attempt to increase profit), moving the brand further and further away from the targets. Business Implications and Potential Applications 1\vo cases were considered in the above analysis: One case in which advertising has the comparative advantage in affecting market share, and the other case in which its comparative advantage is in affecting profit. In fact, they may be taken as being instances of an entire spectrum of cases ranging from an extreme situation where the market share is inelastic to price, and dominated by advertising and other marketing inputs; to the other extreme in which price is the all important instrument in affecting market share. Case 1 assumed that market share was "dominated': (to a greater or lesser extent) by advertising effects and price was the less dominant influence. Such cases are very common in business practice. This is the case of the premium product (as seen from a product portfolio viewpoint), or the case of a quality product in a price-insensitive market (viewed as segmentation analysis), or the case of the image building task (at initial stage(s) in the product life=eycle), etc. The above examples (in parentheses) are intended only as instances of potential marketing applications. The theoretical analy- sis showed that the appropriate business strategy, if profit and market share targets are not met, is to increase advertising and price in the indicated manner. This is consis- tent with business judgment. Case 2 is representative of the situation in which price has the dominant effect on market share and advertising has relatively little influence. The business applications 62 Journal of Business Strategies Vol. 8, No.1 of this case are opposite to the ones mentioned above: the standardized product or market, or a product in the later stage(s) of the life-cycle. The theoretical analysis indicated that the appropriate strategy, if targets were not met, would be decrease price and advertising as shown in Figure 2. This, too, is consistent with business judgment. The analysis provides a rigorous theoretical base to some commonly held business beliefs. Although the analysis is a general one, it is not any less pragmatic for that reason. The only parameters required for analysis of a specific brand in a particular market are the partial derivatives, (Cp, CA, Mp, and M,J in the cases discussed above. These parameters may be empirically obtained in a given situation. One method for obtaining them is to use the regression coefficients of a regression analysis with Mar- ket Share as the dependent variable, for example, and Price and Advertising (lagged, if appropriate) as the independent variables. As shown by Stowsand and Wenzel [5, p. 254] in their study of eight different product groups, and Cronin and Skinner [1, pp. 14-17] in their study of the different effects of several business instruments (ROA, liquidity, leverage, etc.) on business outcomes (profitability, financial structure, and marketing outcomes), different instrument coefficients are obtained that show differing effects on business goals. The current paper provides a normative case for the appro- priate use of such analyses. Some conclusions of the analysis which would not have been immediately obvious without the benefit of this theory are described below: (a) The fact that the slopes or regression coefficients, Cp, CA, Mp, and MA , are found to be of the appropriate sign in an empirical study (Cp > 0, CA < 0, Mp < 0, MA > 0) is !lQ indication that either instrument should be used to achieve a particular target. For instance, the analysis indicates that, even if price increases have a positive effect on profit and a negative effect on market share, as expected, it should not be used as a business instrument to affect market share unless it has the comparative advantage. (b) Violation of the principle of effective market classification leads to disas- trous consequences. To continue with the example mentioned in (a) above, if price is used as an instrument to change market share, the brand will move further away from its targets, unless it has the comparative advantage. Based on the findings of this paper, it is recommended that a manager develop an understanding of the relative strength or comparative advantage of the instruments at his or her command. One method for doing this is through regression analysis, where the regression coefficients are indicators of the relative strength. Other quantitative or accounting data may also help determine such relationships and effects. IT no histori- cal data is available, the manager may consider a judgemental analysis based on other similar products inside or outside the company. It is further recommended that the manager use the instrument that has the comparative advantage to achieve each of the goals. Spring 1991 Sethna, Allen & Wellan: Multiple Business Goals Summary and Conclusions 63 The approach in this paper analyzes the problem of multiple goals and: multiple in- struments in the business context and provides a rigorous theoretical base for some business strategies; it presents recommendations for the use of such instruments that suggest that instruments with comparative advantage should be selected to achieve each goal; and it illustrates how violation of the principles presented in the paper might lead to undesirable results. The conclusions of the theoretical analysis are shown to be con- sistent with business judgment. Also shown are some conclusions of the theory which are not immediately obvious, and dangers of violation of the theoretical principles. The analysis presented in this paper uses a dynamic approach rather than a com- parative-static one. In this respect, it adds an additional dimension to the solution of such business problems. This analysis is made possible by incorporating theories of macroeconomic policy into business framework. The resulting expansion of the bound- aries of knowledge and integration of two fields (macroeconomics applied to the strat- egy of the firm) may well be contributions of this paper. The procedure described in this paper is a general one and may be applied to a number of corporate or social goals. References 1. Cronin, U., Jr. and Skinner, SJ. "Marketing Outcomes, Financial Conditions, and Re- tail Profit Performance." Journal of Retailing. Vol. 60, No.4 (1984) pp. 9-22. 2. Makin, J.H. Macroeconomics. Hinsdale, IL.: The Dryden Press (1975), pp. 460-461. 3. Mundell, R.A. "The Appropriate Use of Monetary and Fiscal Policy for Internal and External Stability." International Monetary Fund, Staff Papers, Vol. 55, No. 3 (March 1962), pp. 401-415. 4. Stowsand, H. and Wenzel, W. "Market Mechanics: A Study to Measure the Effect of Marketing Instruments on the Market Position of Fast-Moving Consumer Goods." Jour- nal of Business Research, Vol. 7, No.3 (1979), pp. 243-257. 5. Theil, H. Optimal Decision Rules for Government and Industry. Amsterdam: North Holland (1964). 6. Tinbergen, J. Centralization and Decentralization in Economic Policy. Amsterdam: North Holland (1954). 7. Tinbergen, J. On the Theory of Economic Policy. Amsterdam: North Holland (1951). Multiple Business Goals and Multiple Business Instruments: The Application of the Theory of Macroeconomic Policy to Business Decisions