Skip to main content Accessibility help

Management Accounting in an Early Multidivisional Organization: General Motors in the 1920s*

  • H. Thomas Johnson (a1)


Introduction of the multidivisional enterprise, which has generally replaced the functional or departmental form of organization in the twentieth century, required management accounting techniques that provided both divisional and top management with data with which to evaluate individual managers' performance, company-wide performance, and future company policy. Professor Johnson discusses the development of these controls at General Motors, the results obtained with them in practice, and their alleged shortcomings, especially in respect to division managers' attitudes towards constructive decisions that might tend to limit in the short run the rate of return on the investment entrusted to them. He concludes with some observations on the influence that organization has had on corporate goals.



Hide All

1 Williamson, Oliver E., Corporate Control and Business Behavior: An Inquiry Into the Effects of Organization Form on Enterprise Behavior (Englewood Cliffs, N.J., 1970), 175.Chandler's, Alfred D. classic thesis on the development of modern large-scale enterprise, including the multidivisional form of organization, is developed in Strategy and Structure: Chapters in the History of the Industrial Enterprise (Cambridge, Mass., 1962 and Garden City, N.Y., 1966). His study of multidivisional enterprise and of General Motors in particular is carried further in Giant Enterprise: Ford, General Motors and the Automobile Industry (New York, 1964) and Pierre S. Du Pont and the Making of the Modern Corporation, co-author Stephen Salsbury (New York, 1971). Additional observations on the importance of accounting information in multidivisional enterprise are found in Chandler, and Redlich, Fritz, “Recent Developments in American Business and their Conceptualization,” Business History Review, XXXV (Spring, 1961), 131.

2 The following remarks about Durant and the 1920 inventory crisis at GM are drawn from: Chandler and Salsbury, Pierre S. Du Pont, Chs. 16–18; Dale, Ernest, “Contributions to Administration by Alfred P. Sloan, Jr., and GM,” Administrative Science Quarterly 1 (19561957), 3062; Sloan, Alfred P. Jr., My Years with General Motors (New York, 1963), Ch. 2.

3 The sources of information on GM's post-1921 accounting practices that were used to prepare this paper are: Bradley, Albert, “Setting Up a Forecasting Program” Annual Convention Series: No. 41, American Management Association (New York, 1926), 320 (reprinted in Chandler, Giant Enterprise, 121–141); Brown, Donaldson, “Pricing Policy in Relation to Financial Control,” Management and Administration, 7 (February, 1924), 195198 and (March, 1924), 283–286, “Pricing Policy Applied to Financial Control,” Management and Administration, 7 (April, 1924), 417–422, “Centralized Control with Decentralized Responsibilities,” Annual Convention Series: No. 57, American Management Association (New York, 1927), 3–24, “Some Reminiscences of an Industrialist” (unpublished mss. at Eleutherian Mills Historical Library, Wilmington, Delaware, 1957); Chandler, Strategy and Structure, Ch. 3; Chandler and Salsbury, Pierre S. Du Pont, Chs. 17–21; Fordham, Thomas B. and Tingley, Edward H., “Control Through Organization and Budgets,” Management and Administration, 6 (December, 1923), 719724, 7 (January, 1924), 57–62, (February, 1924), 205–208 and (March, 1924), 291–294; Mark, R. C., “Internal Financial Reporting of General Motors,” Federal Accountant (November 12, 1952), 3141; Mott, C. S., “Organizing a Great Industrial,” Management and Administration 7 (May, 1924), 523527 (reprinted in Chandler, Giant Enterprise, 115–127); Sloan, Alfred P. Jr., My Years With General Motors (New York, 1963), Ch. 8; Swayne, Alfred H., “Mobilization of Cash Reserves,” Management and Administration 7 (January, 1924), 2123; U.S. Senate, Committee on the Judiciary, Subcommittee on Antitrust and Monopoly, Administered Prices, Report and Hearings, Vols. 6 and 7 (1958); Wennerlund, E. Karl, “Quantity Control of Inventories,” Management and Administration, 7 (June, 1924), 677682.

4 Sloan, My Years, 140. Brown, “Some Reminiscences of an Industrialist,” 61.

5 On the development of Du Pont's accounting system to World War I see Johnson, H. Thomas, “Management Accounting in an Early Integrated Industrial: E. I. du Pont de Nemours Powder Company, 1903–1912,” Business History Review XLIX (Summer 1975), 184204. Brief remarks on changes in Du Pont's accounting system from 1914 to about 1921 are in Chandler, Strategy and Structure, 81–82 and 132–133. On the transfer of Du Pont's methods to GM see Sloan, My Years, 116–117. Brown, “Centralized Control with Decentralized Responsibilities,” passim. Sloan, My Years, 140.

6 Brown, “Pricing Policy in Relation to Financial Control,” 195–196 and “Some Reminiscences”, 59–65.

7 Brown, “Centralized Control with Decentralized Responsibilities, 5; Sloan, My Years, 141. Other statements that attest to GM management's primary concern with profits are in Fordham and Tingley, “Control Through Organization and Budgets,” 719 (“Business is operated only to make a profit”) and Mark, “Internal Financial Reporting,” 33.

8 Brown, “Pricing Policy in Relation to Financial Control,” 197. Harlow H. Curtice once testified that “standard volume” was based on a daily operating rate of 2 shifts, or 16 hours per day, although he did not state the number of, days per year. He said that GM's average production over 30 years (1928 to 1957) was within 7 per cent of “standard volume.” See U.S. Senate, Administered Prices: Hearings (1958), Vol. 6, 2521 and 2522.

The 20 per cent return on investment rate applied, it would seem, to the Corporation's overall net earnings after taxes. Because corporate expenses per se and income taxes were not allocated to divisions, one assumes that the target return on investment for individual divisions, on which product pricing policy was based, must have been considerably higher than 20 per cent. As for what items the company included in its investment base, published accounts give an ambiguous picture. In public statements, investment is often defined as “net worth” or “stock-holders investment” (see testimony by Harlow Curtice cited above). Published examples of the company's internal reports (see Figure 1) suggest, however, that all assets were included in the investment base, at least for some purposes. For additional information see Mark, “Internal Financial Reporting,” 37–38.

9 An excellent discussion of the auto industry's interwar pricing policies is in Vanderblue, Homer B., “Pricing Policies in the Automobile Industry,” Harvard Business Review XVII (Summer, 1939), 385401.

10 Sloan, My Years, 129.

11 Bradley, “Setting Up a Forecasting Program,” 7.

12 Ibid., 3.

13 The detailed steps involved in preparing all segments of a divisional budget were described by two executives of Delco-Light Co. See Fordham and Tingley, “Control Through Organization and Budgets.” Limiting the sales of each division to a specific price range followed from GM's overall product policy, enunciated in late 1921, to minimize duplication and competition among divisions. See Sloan, My Years, 65. Although the company's basic product policy indicated a need for some top-level coordination of divisional sales plans, no such coordination existed before 1924. However, top management took a more serious interest in divisional sales forecasts after the overoptimism of certain general managers caused a serious overproduction crisis in that year. For more information on this famous episode in GM's history see Sloan, My Years, 129–134 and Chandler and Salsbury, Pierre S. Du Pont, 549–554. Sloan, My Years, 136–137.

14 Bradley, “Setting Up a Forecasting Program,” 7–8; Vanderblue, “Pricing Policies,” 397 n. 14, U.S. Senate, Administered Prices: Report, 106.

15 Bradley, “Setting up a Forecasting Program,” 7.

16 To derive Brown's standard price ratio (although in a somewhat different fashion than he follows in “Pricing Policy, …,” 283–286 and 417–422), assume a division that produces only one product with capital investment (variable and fixed) K that is comprised of two parts: one a function of sales, f(S) (e.g., the ratio of variable cash and receivables to sales), and the other a function of factory production costs, f(C) (e.g., the ratio of variable inventories to cost and fixed plant investment to cost at standard volume). If S is the division's annual dollar sales, C is the factory production costs (variable, semi-variable, and fixed) at standard volume, a is the “turnover” ratio f(S) ÷ S, b is the “turnover” ratio f(C) ÷ C, and r is the division's target return on investment at standard volume, then

But rK is also the division's total profit at standard volume, P; and by definition

If the division's commercial and selling expenses (variable, semi-variable and fixed) are not included in C and if they are stated as a ratio to sales, cS, then

Substituting (2 ) into (4) and simplifying, we get

Dividing through by the standard volume output, T,

Thus C/T is the unit factory cost at standard volume and S/T is the unit standard price that will return r percent on K at standard volume. Those readers who examine Brown's articles should note that he refers to standard price as “base price” and he generally uses the word “normal” where I have used “standard.”

I find it surprising that no economist or accountant has ever derived or analyzed Brown's target return on investment price formula. While economists often discuss GM's standard price policy, they never go into specifics and they seldom refer to Brown's 1924 articles. And accounting writers seem to have ignored Brown altogether. In fact, one group of accountants in 1959 published what they thought was an original return on investment pricing formula; however, they merely reinvented the wheel that Donaldson Brown had already discovered some 35 years earlier. See Return on Capital as a Guide to Managerial Decisions, N.A.A. Research Report No. 35 (December 1, 1959), 44.

17 Although this definition of the standard price implies that the division produces only one product, it is easily extended to a multiproduct (or multimodel) situation if one assumes either that costs and investment can be allocated among products or that the mix of products can be pre-determined. Published information does not indicate how GM actually calculates standard prices in practice.

18 Alfred Sloan remarked once that “an alternative approach to our standard-volume policy would have been to evaluate prices strictly in terms of actual unit costs at actual or anticipated production levels.” However, “the use of the actual unit-cost type of evaluation would have been socially and economically unsound,” he argued, because of the industry's cyclical demand and the company's high fixed costs. Naturally, “unit costs would drop in times of high volume and increase during periods of low production. Any attempt to raise prices during periods of low volume, even if competition permitted, to recover the higher unit costs could have deflated sales still further, with the result of still lower profits, less employment, and a generally depressive effect on the economy.” See Sloan, My Years, 147.

19 Brown, “Pricing Policy Applied to Financial Control,” 417–422. Sloan, My Years, 146–148. U.S. Senate, Administered Prices: Hearings, Vol. 6, 2519–2521. Cordtz, Dan, “Car Pricing, …,” Wall Street Journal (December 10, 1957), reprinted in Ibid., Vol. 7, 3496–3502. Vanderblue, “Pricing Policies,” 396–401. Mark, “Internal Financial Reporting,” 34. Homer Vanderblue once suggested that cost does not determine auto price; instead, it determines auto quality. Thus, if division managers regarded the coming year's price as more or less given, they probably coordinated proposed price with standard price by postponing, whenever possible, costly quality (or “style”) improvements. Vanderblue, “Pricing Policies,” 395.

20 Although GM's Price Study forecast procedure may have been inspired initially by similar forecasting tools that the DuPont Corporation had used for many years before 1920, it is notable that GM's forecasts by 1923 served a very different purpose than DuPont's forecasts had ever served up to that time. DuPont's forecasting procedure was originally designed to control cash, working capital, investment, and financing requirements; and the forecasts that Durant's successors brought to GM from DuPont in 1920–21 were intended primarily to cope with a severe liquidity crisis. By late 1922, however, when Alfred Sloan said that Price Study forecasts first appeared at GM (My Years, 129), the liquidity crisis was gone. Now GM faced the durable capital goods producer's endemic problem of how to control revenue and expense during short-run periods of sharp and unpredictable change in market demand; this problem apparently did not affect DuPont's management to the same degree at that time. Thus, GM modified DuPont's forecast procedures to suit a different set of circumstances. Indeed, most of Du Pont's procedures required modification before they could be used at GM; and one surmises that Donaldson Brown directed the major changes. See Sloan, My Years, 118.

21 Bradley, “Setting Up a Forecasting Program,” 13. Sloan, My Years, 138.

22 Bradley, “Setting Up a Forecasting Program,” 13.

23 Sloan, My Years, 138; Dale, Ernest, The Great Organizers (New York, 1960), 100.

24 Parker, Robert H., Management Accounting: An Historical Perspective (London, 1969), 6770; Solomons, David, “The Historical Development of Costing,” in Solomons, D., ed., Studies in Costing (London, 1952), 36; on Gillette see Raymond, R. H., “History of the Flexible Budget,” Management Accounting (August, 1966), 12; on Westinghouse see Dale, The Great Organizers, 151 and 165–166; Brown, “Pricing Policy … Financial Control.” “In these annual budgets, fixed and non-variable expenses are treated separately from variable so that the figures can be readily adjusted for changes in volume at any time during the year,” Sloan, My Years, 143. See also Mark, “Internal Financial Reporting,” 34.

25 Fordham and Tingley, “Control Through Organization and Budgets,” 722.

26 To preserve simplicity, the annual sales are forecast in Table 1 using the dollar equivalent of the standard price ratio. In practice, however, division managers prepared the standard volume forecast using the actual price that was proposed for the coming model year (see Bradley, “Setting Up a Forecasting Program,” 7); Brown, “Centralized Control with Decentralized Responsibilities,” 21, and “Pricing Policy in Relation to Financal Control,” 285.

27 Examples of some of these reports are in Mark, “Internal Financial Reporting,” 36. Albert Bradley once noted that “on the basis of our forecast we work up a balance sheet and income statement [and] during each month as it develops we make up a daily balance sheet and income account, so that we can tell you this afternoon where we stood at the close of business yesterday.” See the discussion that follows “Setting Up a Forecasting Program,” 19.

28 Brown, “Some Reminiscences of an Industrialist,” 48–49.

29 “In order to allow for freedom of action with respect to variations in raw material and labor prices, it is not desirable to establish the [standard price] in terms of dollars per unit of product, but rather in terms of percentage of factory cost [at standard volume].” Brown, “Pricing Policy in Relation to Financial Control,” 285, 286.

30 Bradley, “Setting Up a Forecasting Program,” 15–18.

31 Accountants have studied the effects of internal accounting systems on management behavior and economists have studied the effects of internal structure on company performance only in the past decade or two. Thus, GM's New Era management system anticipated management accountants' and economists' ideas about motivating profit-oriented behavior by almost fifty years. For some additional insight into how advanced GM's management accunting system was, see Fordham and Tingley, “Control Through Organization and Budgets,” 291–294.

32 The unique goal pursuit properties of the multidivisional structure are summarized in Williamson, Corporate Control, Ch. 8. (especially 133–134). A succinct statement of economists' views on corporate goal formation is in Scherer, F. M., Industrial Market Structure and Economic Performance (Chicago, 1970), 3133 and 282–283. A more elaborate discussion of economists' theories of corporate goal formation, including a thorough examination of the behavioral theory of the firm, is in Cohen, Kalman J. and Cyert, Richard M., Theory of the Firm: Resource Allocation in a Market Economy (Englewood Cliffs, N.J., 1965), Chs. 16–17.

33 Top management's dedication to “entrepreneurial” profit-making is theoretically no less in a centralized company where operating responsibility is delegated to functional department managers than it is in a decentralized company, like GM, where operating responsibility is delegated to division managers. In practice, however, it is often quite different. In the functionally departmentalized firm, top management often finds that it must sacrifice some of its commitment to company-wide profits in order to achieve smooth coordination of the operating departments' activities. The reason is that the performance of these departments, since they are cost centers or revenue centers, cannot easily be evaluated in terms of their contribution to company-wide net profit. Except where a department head's performance is egregiously bad, then, it is difficult to detect, especially in large organizations, if department heads are doing everything they can to enhance company-wide profits. Therefore, top managers in large-scale functionally departmentalized firms often find that the easiest way to achieve smoothly coordinated operations is to trade-off some company-wide net profit for a certain amount of departmental “slack” (e.g., overstaffmg, inefficiency, and so forth), to the extent that market conditions allow such a trade-off to be made. See Williamson, Oliver E., “Managerial Discretion, Organization Form, and the Multi-division Hypothesis,” in Marris, Robin and Wood, Adrian, eds., The Corporate Economy: Growth, Competition, and Innovative Potential (Cambridge, Mass., 1971), 358359.

34 Solomons, David, Divisional Performance: Measurement and Control (Homewood, Ill., 1965), 154.

35 The problem of distinguishing between the division's performance and the division manager's performance is succinctly analyzed in Horngren, Charles T., Cost Accounting: A Managerial Emphasis (Englewood Cliffs, N.J., 1977), 713. “To the extent that performance and subsequent resource assignments can be connected in rational ways, the resource allocation process can be made responsive to differential performance. Ceteris paribus, those parts of the organization that are realizing superior performance will increase in relative size and importance, with all of the beneficial local gains associated with expansion.” Williamson, “Managerial Discretion,” 362. Different return-on-investment targets existed for different products at least since 19–10 at the DuPont Corporation, whose methods provided so many of the basic ideas for GM's post-1921 accounting system. Johnson, “Management Accounting in an Early Integrated Industrial,” 198–199.

36 It is notable that division controllers at GM were the only company officials who had dual responsibility; they reported both to the division manager and to the corporate controller. Williamson, Corporate Control, Ch. 8. An independent functionally structured firm will tend to be less profitable than a semi-autonomous division of similar size largely because the chief executive of the independent company is under less pressure than the general manager of the division to suppress slack and inefficiency within his organization. Whereas the performance of the division manager is promptly and comprehensively scrutinized by profit-oriented top managers, the performance of the independent company's chief executive is evaluated infrequently and imperfectly by outsiders (e.g., participants in capital or product markets, stockholders, “raiders,” and so forth) whose incomplete knowledge about the company's operations precludes them from threatening the chief executive's position unless his performance is egregiously bad.

37 Horngren, Cost Accounting, 742–744.

38 Brown, “Centralized Control with Decentralized Responsibilities,” 8. Mark, “Internal Financial Reporting,” 32.

39 Sloan, My Years, Ch. 7.

40 32nd Annual Report of the General Motors Corporation, Year Ended December 31, 1941, 38–40 and 81. Sloan, My Years, 407–420. The Bonus Plan provided “the means to attract and to hold in the employ of the Corporation men of proven and potential ability. …” Brown, “Centralized Control with Decentralized Responsibilities,” 13. “The Bonus Plan established the concept of corporate profit in place of divisional profits. …” Sloan, My Years, 409.

* The research for this study was supported by funds from The Canada Council and The Eleutherian Mills-Hagley Foundation. The author is responsible for the views expressed in this article. He gratefully acknowledges the assistance of Elaine Bowe Johnson.

Management Accounting in an Early Multidivisional Organization: General Motors in the 1920s*

  • H. Thomas Johnson (a1)


Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed