business policy unit VI essay and DQ question

ECONOMIC CRISIS AND THE CENTRALIZATION OF CONTROL OVER THE MANAGERIAL PROCESS:

CORPORATE RESTRUCTURING AND NEO-FORDIST DECISION-MAKING*

HARLAND PRECHEL

Texas A&M University

I analyze the effects of the recent intensification of control over the managerial pro- cess in a large steel corporation in the United States. Corporate restructuring was an attempt to overcome constraints on capital accumulation and resolve the contra- dictions and inefficiencies embedded in the previous controls over the managerial process. Formal controls over the managerial process were intensified to standardize decisions, improve product quality, and reduce costs. Additional findings include: (1) discretion over many decisions was centralized in decision centers where conceptual activities were performed; (2) the new formal controls increased surveillance over production managers; (3) many decisions were eliminated, which reduced the need for some managers and eliminated four layers of the managerial hierarchy; and (4) the corporation became more tightly coupled and more flexible. These Neo-Fordist controls entailed a structure of capital accumulation that separated conception from execution while enhancing control over the managerial process.

The 1980s initiated a period of transfor-mation for manufacturing corporations in the United States. The back-to-back reces- sions of the early 1980s, escalatitig interest rates, declining profits, and foreign competi- tion demanded higher quality products and relentless cost cutting.

Current analyses of the effects of corporate restructuring focus on intra- and interorgan- izational relations or human resources (Lin- coln and Kalleberg 1985; Florida and Ken- tiey 1991), the effect of computer-directed technologies on blue-collar work (Shaiken 1984; Howard 1985; Kelley 1990; Zetka 1991), and the computerization of office management systems (Attewell 1987; Baxter 1989). Although these studies make impor-

' Direct all correspondence to Harland Prechel, Department of Sociology. Texas A&M Univer- sity, College Station. TX 77843-4351, This paper ha.s benefited substantially from comments from Pat Akard, Jon Alston, John Boies, Dan Clawson. Derek Gill. Joe Gorton, John Harms, Wolf Hyde- brand, Stephen Kalberg. William Rothstein, and several anonymous ASR reviewers. 1 thank Tag- gert Steel Corporation (a pseudonym) for access to its archives, personnel, and other organiza- tional resources necessary to this study. [Review- ers acknowledged by the author are Kenneth Ben- son, Jerald Hage. Michael Wallace, and James Zetka, —ED.\

tant contributions, researchers have given in- sufficient attention to the effect of the new conditions of capital accumulation' on the managerial process (i.e., the collective be- havior of managers and the incentive struc- tures that motivate and control their behav- ior). I suggest that, like the labor process dur- ing a previous historical period (Braverman 1974; Noble 1977; Edwards 1979; Clawson 1980; Gordon. Edwards and Reich 1982). control over the managerial process in the contemporary era is being centralized.

Researchers who have addressed changes in the managerial process often have dia- metrically opposing views. Some have sug- gested that corporations are establishing de- centralized cooperative work teams, extend- ing managerial freedom, weakening the boundaries between managers and the man- aged, emphasizing informal networks, and increasing autonomy and participation (Piore and Sabel 1984; Peters and Waterman 1984; Kanter 1989; Chandler 1990). Others have argued that the managerial process is becom- ing more centralized: Managers are further

' Capital accumulation is the mobilization, transformation, and exploitation of inputs to in- crease the total capital of the corporation (Marx [1887] 1977:711-61; Bowles and Edwards 1985: 86-87; Sweezy [1942] 1970:92-94).

American Sociological Review, 1994. Vol. 59 (October:723-745) 723

724 AMERICAN SOCIOLOGICAL REVIEW

removed from decision-making centers, au- tonomy is declining, and middle-managerial work is being degraded (Shaiken 1984; Car- ter 1985; Heydebrand 1985; Smitb 1990; Burris 1993). Some argue that the ideology of managerial decentralization obscures this degradation (Smith 1990).

After a decade of debate, little progress has been made toward resolving this debate because current theories fail to address the question: How does a corporation maintain control over complex manufacturing pro- cesses that involve hundreds of managers, while improving product quality and in- creasing the corporation's capacity to re- spond to the changing conditions of capital accumulation? There are two interrelated obstacles to answering tbis question. First, empirical studies of corporate restructuring lack sufficient depth to show where deci- sion-making infonnation is located and who has access to it, and what is being central- ized and decentralized. Second, the concep- tion and execution of managerial decisions are presented as being located in the same place in the corporation. Modifications in the formulation of these variables together with an investigation of the location of deci- sion-making Information are necessary to determine whether centralization or decen- tralization is occurring.

I analyze the response of one large U.S. integrated steel corporation to its competitive environment. My objectives are: (1) to iden- tify how the previous account-control sys- tem^ created an incentive structure that un- dermined corporate profits, (2) to identify the salient characteristics of the emerging con- trol system,^ and (3) to show how changing the loci of decision-making information and discretion produced a different set of author- ity relations within the managerial process.

account-control system is the historically specific mode of quantification designed to achieve organizational standards, targets, and goals. It includes a mechanism lo direct behav- ior, an evaluation procedure, and an apparatus of discipline and reward intended to elicit coopera- tion and compliance.

^ Organizational controls include mechanisms that regulate individual conduct and direct behav- ior toward achieving organizational goals. Con- trol is aimed at maintaining a predictable level and type of performance (Edwards 1979; Child 1984).

CURRENT PERSPECTIVES ON MANAGERIAL CENTRALIZATION AND DECENTRALIZATION

Contingency theory suggests that an organi- zation's structural arrangements are contin- gent upon several internal and external vari- ables.'' It stresses the importance of tech- nology as a determinant of organizational change. Technology is defined as the tech- niques used to transform inputs into outputs, including both physical technology and the knowledge and skills of personnel (Wood- ward 1965; Perrow 1967), Contingency theory also maintains that the greater the task uncertainty, the more information that must be processed by decision-makers to achieve a given level of performance (Galbraith 1977:36). Task uncertainty contributes to de- centralization of decisions and informal con- trol in several ways. First, if the relation be- tween means and ends is poorly understood, rules are not established to govern decision- making, and authority is delegated (Burns and Stalker 1961; Thompson 1967; Penow 1970; Galbraith 1973; Lawrence 1981). Sec- ond, if problems are nonroutine and unpre- dictable, solutions cannot be preprogrammed and decision-making discretion falls on per- sonnel at lower levels (Burns and Stalker I96I; Blau and Scott 1962; Woodward 1965; Hage and Aiken 1969; Child 1972). Third, when the environment is rapidly changing, decision-making authority is allocated be- cause rules cannot cover unanticipated con- ditions, and deferring decisions up the mana- gerial hierarchy reduces organizational re- sponsiveness and flexibility (Burns and Stal- ker 1961; Chandler 1962; Woodward 1965; Lawrence and Lorsch 1967; Thompson 1967; Dornbusch and Scott 1975).

While contingency theory focuses prima- rily on the contingencies (like uncertainty) that contribute to outcome (e.g., decentrali-

'' There are at least two analytically distinct ways to present contingency theory. Advocates argue that "the best way to organize depends on the demands of the environment" (Scott 1981:96). However, this view of "contingency theory has been heavily criticized as being tautological" (Hall 1991:283). For this reason. 1 present the specific claims of contingency theory and evalu- ate the theory's capacity to explain phenomena (also see Simpson 1985).

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 725

zation), decision-making theory focuses on the decision-making process to specify how decentralization is possible. Decision-mak- ing theory suggests three mechanisms that foster decentralization. First, reducing the numher of information sources available to decision-makers limits the possibility of con- tradictory information and differences of per- ception and cognition (March and Simon 1958:127; Ungson, Braunstein, and Hall 1981), Second, bureaucratic rules that be- come the basis of more precise control at the lower levels accompany decentralization (Child 1972). Third, premise controls are implemented to restrict behavior by limiting the content and flow of information, thereby limiting the search for alternatives. When successfully implemented, premise controls restrict the range of stimuli used in decision- making. The specification of information sources, rules, and premise controls combine to limit the search for alternatives and ensure predictable and consistent decision-making. However, premise controls are limited to nonroutine decisions made by professionals, and they operate at high levels of the organi- zational hierarchy (Simon 1957:xii. 79; Per- row 1986:128-29). Therefore, like contin- gency theory, decision-making theory sug- gests that increasing task complexity results in decentralization and that complex tasks, such as coordination of the manufacturing process, are achieved through informal com- munication and cooperation among manag- ers (March and Simon 1958;!60-61). While these theories advance our understanding of how task complexity and uncertainty affect the location of decision-making in an orga- nization, they ignore the effect of variations in the capital accumulation process (e.g., de- mand for higher quality products) on the managerial process.

More recently, post-Fordist arguments have suggested that information is critical to management in an increasingly competitive global economy. Post-Fordism maintains that access to and control over information and a capability for instant data analysis are now essential to decisions concerning internal corporate behavior, flexible manufacturing, the manufacture of specialized products, and the coordination of diverse corporate inter- ests (Harvey 1991). Whereas some post- Fordist arguments have suggested that infor-

mation fosters decentralization and au- tonomy at lower levels of the corporation (Piore and Sabel 1984), others suggest that access to information contributes to central- ization (Dohse, Jurgens, and Malsch 1985). Post-Fordist theory has not demonstrated whether centralization or decentralization is occurring because it has not examined how information is used in the decision-making process.

Although contingency theory, decision- making theory, and post-Fordism all implic- itly or explicitly acknowledge the role of in- formation in the decision-making process, each perspective conceptualizes information in highly abstract terms. None gives explicit attention to the kind of information used, its location in the organization, who has access to it, or how it is used in decision-making. To determine whether decision-making is tightly controlled or subject to wide discretion, re- search must analyze the design of informa- tion systems and the spatial dimensions of the managerial process—the organizational distance between the conception and execu- tion of decisions. Moreover, centralization and decentralization must be treated as theo- retical constructs that illuminate empirical processes rather than as empirical absolutes.

A THEORETICAL REFORMULATION: NEO-FORDIST DECISION-MAKING

I suggest that new methods of formal control over the managerial process are emerging to ensure standardization of decisions and prod- ucts. Neo-Fordist managerial controls entail a sharp break from the past, when concep- tion and execution of decisions tended to be located in the same place in the managerial process.^ The introduction of technically ad- vanced accounting and information-process- ing systems make it possible to centralize authority while decentralizing the responsi- bility to execute activities conceptualized at higher levels in the managerial hierarchy.^

^ Therefore, I use the term neo-Fordism. Also, the post-Fordist literature has focused primarily on the labor process and neglected the manage- rial process (Harvey 1991; Sayer and Walker 1992).

''Previous research has assumed that decision- making and authority are located in the same

726 AMERICAN SOCIOLOGICAL REVIEW

Therefore, decision-making and authority must be treated as separate variables, so that the location of authority apart from the loca- tion of decision-making can be considered.

Decentralization of authority exists, for example, when the organization's controls of one level over a lower level are so abstract as to leave decision-making discretion with the lower-level manager (Stinchcombe 1990: 114-15). Therefore, the key feature of decen- tralized authority is the abstractness of infor- mation flows and the specificity of how to u.^e that information. To address this issue, the subterranean processes of collecting, calcu- lating, and redistributing information must be analyzed.

As Weber ([19211 1978:224) argued, ra- tional calculation creates the possibility of establishing a single decision-making crite- rion and a unified system of formal control over distinct spheres of the corporation. These efforts to increase economic perfor- mance cannot be dissociated from authority because profitability "is a concept which is applicable to every discrete act which can be individually evaluated in terms of busi- ness accounting techniques" (Weber 1978: 96). Autbority is highly centralized in manufacturing because many complemen- tary processes must be coordinated under continuous common supervision (Weber 1978:1156). This coordination creates the possibility of subjecting social actors to stringent discipline and standardizing effort and product quality by establishing incen- tives and shared interests for eacb actor in the social structure (Weber 1978:137; Wa- ters 1990:198).

Rational calculation has a dual role in the corporation. First, it evaluates operating costs in relation to market opportunities to determine "the expected advantages of every projected course of economic action," in- cluding the success of profit-making activi- ties and the income-yielding power of the corporation (Weber 1978:81-92; Kalberg 1983). Second, it ensures control and coop- eration through "the rational conditioning and training of work performances" (Weber 1946:261). The application oi formal ratio- nality extends quantitative calculation to its

place in the organization (Van de Ven and Ferry 1980; Child 1984; Kanter 1989; Fombrun 1992).

technical limits to translate actions into nu- merically calculable terms (Weber 1978:85- 86; Sbaiken 1984).

Three formal controls are based on ratio- nally calculated data. First, budget control calculates the receipts and expenditures of each production unit at the end of an ac- counting period. Corporations use budget controls to determine capital allocations and establish expected profit levels. Budget con- trols provide information: (I) to determine tbe operating costs of each organizational unit, (2) to evaluate the degree to which man- agers remain within their budgets, and (3) to measure the optimum profitability of work- ers and the material means of production (Weber 1946:261, 1978:86-87; also see Chandler 1962, 1977)."' Second, bureaucratic coo/ro/establishes the possibility of carrying out organizational activities based on objec- tive considerations to ensure calculability of results. Business management emphasizes calculable rules because tbe success of the enterprise depends on precision, consistency, and above all, speed of operations (Weber 1978:74-75, 975). Third, technical control is a bundle of rationally calculated rules em- bedded in the physical technology of tbe en- terprise that dictates activity at the point of production and is "oriented economically to profit-making" (Weber 1978:65-67, 108; also see Braverman 1974; Edwards 1979). The selection of technology involves "con- sideration of costs" and the degree to which "these expenditures will pay off in terms of money obtained through the sale of tbe goods" (Weber 1978:66).

Moreover, rational calculation is histori- cally contingent and occurs slowly "in the absence of tbe objective need for it" (Weber 1978:106, 89; also see Marx [18931 1981a: 211-12, [1894] 1981b:312).s Researchers sensitive to historical conditions have shown.

' Economists use outpul control to denote mea- surements derived from market prices (Ouehi and Maguire 1975:568-69). They suggest that these controls ensure that operating units are governed in a quasi-market fashion (Williamson and Ouchi 1983:26). Budgei control is a type of output con- trol.

"Theorists from other perspectives agree that historical variation in the calculation of costs and prices is a critical dimension of managerial strat- egies .(Schumpeter 1950; Sombart 1953).

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 727

for example, that unexpected contingencies or crises between 1870 and 1910 resulted in a shift toward a more purposive control sys- tem that required new modes of calculation to deliver more precise information to deci- sion-makers (Chandler 1962; Garner 1976; Cbatfield 1977; Noble 1977; Levenstein 1991). Other researchers have argued that or- ganizations adapt to historically specific eco- nomic conditions (Hamilton and Biggart 1988; Precbel 1991b): New organizing tech- niques and revolutionary changes emerge from crises (Benson 1977; Prechel 1991b), and crises result in the formalization and tightening of organizational controls (Pfeffer and Leblebici 1973). Crises are critical points at wbich an organization must reori- ent and revitalize itself to survive (Weitzel and Jonsson 1989: 104).

My formulation of neo-Fordist decision- making theory suggests that top managers are restructuring the managerial process in response to historically specific constraints on capital accumulation. Rational calculation is more frequently used to make decisions about where costs can be cut and to ascertain how control over tbe managerial process can be increased. These data are centralized, and experts (e.g., accountants, engineers, metal- lurgists) analyze thetn and establish control over the tnanagerial process by defining the premise of decision-making, and distributing information to operating managers on a "need-to-know basis."^

Based on tbis theoretical reasoning and my case study, I suggest four propositions. First, rational calculation intensifies to restructure the account-control system and overcotne two historically specific constraints on capi- tal accumulation: high manufacturing costs and low-quality products. Second, decision- making information is centralized and dis- tributed on a "need-to-know basis" to limit managers' decision-making discretion at tbe point of production. Third, technologies that deliver information are itnplemented to speed up the transfer of information through- out tbe corporation, thereby increasing orga- nizational flexibility and responsiveness.

Fourth, neo-Fordist decision-making controls reduce the information-processing activities of production managers, permitting a reduc- tion in the number of managers and number of levels in tbe managerial hierarchy.

If contingency tbeory is correct and the environment is rapidly changing, decentrali- zation of decision-making and autbority will occur to increase organizational flexibility. This decentralization will be characterized by informal coordination and control of the manufacturing process. If decision-making tbeory is correct, decentralization of deci- sion-making and authority will be accompa- nied by elimination of information sources and an increase in formal rules and premise controls. However, decentralization will be restricted to upper levels of the hierarchy and, as is true of contingency theory, coordi- nation will be done informally. Post-Fordist arguments are supported if information be- comes more important to the decision-mak- ing process, and its use increases organiza- tional flexibility. The neo-Fordist decision- making perspective is supported if con- straints to capital accumulation result in the use of information to centralize control over more spheres of tbe managerial process, while increasing organizational fiexibility.

CASE STUDY AND METHODS

This case study presents evidence to evalu- ate the explanatory power of the existing theories and my formulation of neo-Fordist decision-making theory. I am not systemati- cally testing or falsifying theories. Rather, evaluating the explanatory power and scope of a tbeory entails an examination of tbe de- gree to wbich a theory directs attention to- ward certain phenomena and not toward oth- ers (Prechel I991a:698; Akard 1992:600).'^ I develop an alternative neo-Fordist decision- making theory because contingency theory, decision-making, and post-Fordist theory ne- glect critical dimensions of the managerial

''Braverman (1974:82) analyzed how control over the labor process centralized information and redistributed that information on a "need-to- know basis."

'" Moreover, disputes over competing theories in social science are rarely concerned with "truth" or "falsehood." but rather with the relative use- fulness of different levels of analysis or episte- mological assumptions that in turn reflect differ- ences over the central problem of interest (Alford and Friedland 1985: Akard 1992).

728 AMERICAN SOCIOLOGICAL REVIEW

process and historical conditions." They also fail to explain the relationship between the macro-level capital accumulation process and the micro-level managerial process.

I investigate changes in the managerial pro- cess at Taggert Steel Corporation (a pseud- onym), a large Nortb American integrated steel plant (i.e., a plant that produces steel from raw tuaterials). In the late 1970s, aver- age employment iji this plant was 24,000, wbicb included 6,000 managers. Historically, most of Taggert Steel's capital was invested in steel. As early as the 1930s, tbe corpora- tion emphasized flat products, which are at the high-cost, high-quality end of the market. Flat products Include sheet steel for the auto- mobile and appliance industries. In 1982, Taggert's largest market was the automobile industry, which purchased 30 percent of Taggert's total steel tonnage. Steel products made up more than 65 percent of corporate sales and Taggert manufactured more than "5,000 finished products and 50,000 interim products" (Middle Manager D, Account- ing).'^ Moreover, unlike some integrated steel corporations, Taggert did not retreat from steel manufacturing during the 1980s; tbe percentage of its revenue from steel sales in- creased in the early and mid-1980s.

Taggert Steel Corporation is an appropri- ate case to analyze for several reasons. First, the steel industry was affected severely by global competition. Therefore, insofar as cor- porate restructuring is a response to market competition, these processes should occur in this type of corporation. Second, increasing the quality of consumer products (e.g., auto- mobiles) was a fundamental challenge to U.S. industry in the 1980s.'^ Because Taggert depended on customers who also suffered constraints on capital accumulation, Taggert is an appropriate case through whicb to examine the effects of the resource-depen- dent relations among industrial sectors.

" Although organizational sociologists argue that corporations are profoundly historical phe- nomena and must be understood in their histori- cal contexts (Clegg 1981), few historical studies of corporate change exist.

'̂ An interim product is the output from each production unit in an integrated manufacturing process.

'̂ For example, improving the fuel efficiency of automobiles required lighter steel.

Third, the steel industry is a dominant seg- ment of the economy, and it spans an impor- tant period in tbe developtnent of U.S. capi- talism.'''Fourth, if these changes occurred in a steel corporation—especially during a pe- riod when increased interdependence among manufacturers was necessary to ensure prod- uct quality—sucb changes are probably oc- curring in other sectors.

Although a case study generates concerns about whetber the case is typical or atypical, tbe key is to demonstrate the ways in whicb the case resembles or differs from others, and why tbose similarities and differences are theoretically important. This case study dif- fers from previous research on organizational decision-making in a theoretically important way. Contingency theory and decision-mak- ing theory are based on studies conducted between the 1950s and 1970s, when corpo- rations were not confronting a crisis of capi- tal accumulation. In contrast, Taggert Steel was facing falling profits and increasing de- mands for quality products. That is, my study is unique theoretically; it examines tbe cor- porate response to an increasingly competi- tive economic environment. If this indepen- dent variable (constraints on capital accumu- lation) is important, this case study sbould generate new insights.

Two additional interrelated characteristics distinguish Taggert from other large manu- facturing corporafions. First, the steel indus- try historically has exhibited a bigb degree of oligopolistic behavior.'^ Second, tbe se- vere constraints on capital accumulation that emerged in tbe steel industry in the early 1980s resulted in an accelerated rate of cal- culation that may not bave occurred in other industries. However, tbis case study can still illustrate the general trends typical of corpo- rate restructuring. In fact, the rapid rate of cbange following tbe constraints on capital accumulation at Taggert made corporate re- structuring easier to observe. Changes that occur over a long period are more difficult to observe, and the time needed to complete such researcb often makes it infeasible.

'•* Manufacturing accounted for 23 percent of the GNP at the end of the 1980s, a slightly higher share than in \9^0 (Forlune 1991).

'̂ However, oligopolistic behavior also existed in other industries (e.g., automobile, chemical).

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 729

I conducted historical analysis in conjunc- tion with qualitative (and some quantitative) analysis to obtain detailed data on past and present changes. Because I conducted the re- search wben controls over the managerial process were being constructed, the people I interviewed had been thinking a great deal about tbe structuring of the system. They were therefore probably better informants than people who merely operate a system created by their predecessors (Stinchcombe 1990:76).

The qualitative data were obtained from corporate documents and from many hours of structured and unstructured interviews with informants in various departments (e.g., Ac- counting, Corporate, Industrial Relations, Manufacturing, Researcb, Sales, Systems). Moreover, I interviewed informants at all lev- els of the managerial hierarchy. In addition to interviewing a wide range of middle manag- ers, I conducted structured interviews with corporate senior managers (i.e., general man- agers and above). I conducted follow-up in- terviews witb several informants and in some cases interviewed informants a third time. I also conducted unstructured interviews with many other informants. The structured inter- views were recorded and transcribed.

I obtained the historical data from public and internal corporate documents. The pub- lic documents included readily available itetns, such as IO-K forms and quarterly and annual reports. Internal documents included archival data, such as strategic plans, corpo- rate histories, public presentations by corpo- rate executives, press releases, corporate memoranda, and communications to employ- ees. I verified the narrative presented here by cross-referencing interview data witb the public and archival data to resolve problems associated with the reliability of informants' memories. The quotations presented here emphasize points documented in this verifi- cation process. The identification of manag- ers' positions correspond to the managerial hierarchy presented in Table 1.'^

Table 1. Managerial Hierarchy: Taggert Corporation, 1960 to 1984

Management Level Organization Unit

""Gaining access to the corporation entailed an agreement not to reveal the identity of the corpo- ration or its managers. Hence quotations from in- terviews specify only whether they were top or middle managers and the area of organizational activity. When the quotation is from a document

Top Management

1. President

2. Vice-president

Middle Management

3. General Manager

4. Assistant General Manager

5. Manager

6. Superintendent

7. Assistant Superintendent

Lower Management

8. General Foreman

9. Assistant General Foreman

10. Foreman

Corporate

Product line: Steel manufacturing*

Product groups: Flat products, bar products.

Structural products'*

Production units: Hot rolling mills.

Blooming mill.';, cold rolling mills.

Operating responsibilities within production units

° Froduct-hne denotes products in an industrial sec- tor (e.g.. oil. steel, real estate, chemical).

'' Product group (e.g., gasoline, diesel fuel) refers to a type of product within a product line or industrial sec- tor (i.e., oil production).

THE NARRATIVE

Historical Context

From the early 1900s through tbe 1960s, cor- porations in the steel industry pursued two interrelated strategies. On the one hand, the steel industry limited cotnpetition by setting prices and closely coordinating production capability with market demand (Barnett and Schoi-scb 1983; Sobe! 1984; Prechel 1990). In this oligopolistic environment, Taggert's profit-making strategy maximized capability utilization rates (the percentage of produc- tion capability in use) by manufacturing a wide product line. This strategy, bowever, kept total operating costs high because fre- quent changes in the production units were labor intensive and because initial runs often produced low-quality steel that had to be

that specifies the manager, I also identify the type of document. All other quotations are referenced as "public" or "intemal" documents.

730 AMERICAN SOCIOLOGICAL REVIEW

rerolled or sold at a lower price. Yet, because capability utilization rates were high and re- duced the per-unit cost, tbis strategy success- fully maxitnized profits in this oligopolistic environment.

On tbe other hand, like many other corpo- rations (Lazonick 1992), Taggert maintained a bigh price-earnings (P/E) ratio and pro- vided stockholders witb high dividend yields. Wbile this strategy satisfied stock- holders, it reduced the amount of capital available for reinvestment. However, because tbe industry limited competition, the high P/ E ratio and low rate of capital reinvestment were successful strategies for capital accu- mulation in the oligopolistic era.

Until the 1940s, several dimensions of the control structure remained centralized. Typi- cal of large corporations, Taggert's top man- agement monitored a wide range of operat- ing and coordinating decisions and "met ev- ery morning to review incoming orders and schedule the production units" (Top Manager A, Corporate; see Chandler 1962). Within tbis structure, top management directly su- pervised cost inputs for the various produc- tion units. However, as Taggert's size and complexity increased in tbe postwar era as a result of producing a wider product line, its centralized managerial system became inad- equate. Bounded rationality (i.e., the cogni- tive limits of individual decision makers) re- stricted top managetnent's ability to exercise control over day-to-day operations (Simon 1957; March and Simon 1958; Chandler 1962. 1977). To ensure smooth operation of the manufacturing process, decision-making and autbority were delegated from top man- agement and the general or plant manager to the assistant general managers and managers, as shown in Table 1.

Once tbe loci of decision-making discre- tion were decentralized, these units operated semi-autonomously. Moreover, as in other large corporations, once top management re- linquished control over the operating units, little information was available for evaluat- ing their effectiveness (i.e., their use of re- sources to pursue organizational goals [Raff and Temin 1991; Prechel fortbcoming]).

By the 1960s, tbe size and complexity of the manufacturing facility had increased sig- nificantly; supervision of these organization units exceeded the cognitive limits of the

managers and assistant general managers. To overcome these constraints, decision- tnaking and autbority were delegated to the superintendents and assistant superinten- dents responsible for tbe individual manu- facturing units (Table I). Similar to tbe del- egation of authority from top management to upper-middle management, this reloca- tion of the loci of discretion resulted in a new structural arrangement in whicb control over the managerial process was further de- centralized. Middle managers obtained au- thority over their respective organizafional units (e.g., mills) and operated them semi- autonomously. Once top management relin- quished direct control, it relied on budgets to control operating costs and on bureau- cratic controls to specify middle managerial autbority.

The complexity of steelmaking. however, limited the degree to which bureaucratic con- trols could routinize the managerial process. On the one hand, formal rules could not gov- ern decisions regarding coordination because it was difficult to establish rules governing processes that require linking several produc- tion units. As in other expanding corpora- tions (Chandler 1977:413), personal coop- eration among middle managers became the means of coordinating production units and aligning the fiow of materials through the manufacturing process. On the other hand, the numerous variables in the steelmaking process, together with the variety in Taggert's product line, made it difficult to establish rules and regulations to govern pro- duction decisions. Establishing a separate set of rules for thousands of products with unique production specifications restricts manufacturing flexibility (Silver 1982; Shaiken 1984). Therefore, Taggert also de- centralized operating decisions.

In short, the complexity of steel manufac- turing restricted the degree to wbich top man- agement could formalize or centralize con- trol over two functional areas of the corpora- tion. Informal decentralized control remained the primary means to (1) govern many pro- duction decisions and (2) coordinate tbe link among the production units. Production man- agers based their decisions on a combination of formal reports, craft knowledge, and other informal information. As Middle Manager A in Manufacturing reported:

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 731

[I obtained information] from the morning re- ports, what the assistants would tell me, from walking around, watching the production sheets, and a general feel of what was happen- ing.

Under this managerial process, production managers retained control over the four con- ception steps and the final execution step of the decision-making process. Operating managers bad access to the information, pro- cessed tbat information, determined what was to be done, authorized what was to be done, and executed tbe decision.

Contingency theory and decision-making tbeory explain these historically specific de- velopments in the managerial process. Con- tingency tbeory suggests that wben task com- plexity increases, routinized control is diffi- cult to achieve and decision-making and au- tbority are decentralized. Decision-making tbeory predicts that more precise rules and limited information sources accompany de- centralization. Whereas rules specified the span of authority, budgetary control limited tbe premise of decisions. Moreover, as pre- dicted, premise controls operated at a high level in the hierarchy (i.e., organizational units).

Contradictions in the Account-Controt System. 1960s and 1970s

The account-control system shaped manage- rial interests and over time, certain dimen- sions of the organizational structure came to oppose eacb another. Opposing forces are considered contradictions—oppositions or incompatibilities between two crucial com- ponents of a social structure (Zeitz 1980; Lyng and Kurtz 1985; Prechel 1991b). The contradiction of concern bere is how for- mally rational controls shaped managers' in- terests such that their decisions opposed tbe corporation's substantive goals of profits and product quality.'•* Contradictions become cri-

" In contrast to formal rationality, substantive rationality Is not restricted to calculation of the most adequate available methods, but applies some criterion of ultimate ends. Substantive ra- tionality is an evaluative concept denoting the degree to which an activity provides the needs, furthers the ends, or accords with the values of a given social group, and applies "certain criteria of ultimate ends . . . and measures the result of

ses when they generate structural imbalances that hinder organizational adaptations neces- sary to attain its goals (Heydebrand 1977; Zeitz 1980; Prechel 1991b).

A contradiction in budget control—an un- intended consequence of expansion—gradu- ally emerged at Taggert. To take advantage of the demand created by rapid economic growth in the 1960s and early 1970s, Taggert increased its production capability by 24 per- cent. By the late 1960s, tbe lower and middle segments of the managerial hierarchy had expanded to eight levels (Table I). Whereas bureaucratic control defined managerial re- sponsibility and the span of control, budget control remained top management's primary means of control over the managerial pro- cess. Budget control provided top manage- ment with tbe operating cost of each produc- tion unit (e.g., blast furnace) and enabled management to determine whether middle managers were within their budgets. Most important, the ability to remain within their budgets became the criterion for evaluating middle managers' skills and to "ascertain if tbey merited a salary increase" or promotion (Middle Manager A, Accounting; also see Weber 1946; Chandler 1977; Johnson and Kaplan 1991).

Tbis incentive to keep operating costs within assigned budgets, however, sometimes resulted in decisions that undermined prod- uct quality—one of the corporation's substan- tive goals. For example, if mill superinten- dents stopped the manufacturing process to reroll defective steel, tbeir operating costs increased, which inturn refiected poorly on their tnanagerial abilities. As Middle Man- ager A in Industrial Relations reported,

[The manager ofj one mill would pass on steel that really didn't meet specifications, hoping that it would slip through. Nobody wanted to take the hit on poor quality. They regularly tried to pass on the defective material to the next mill and let them take the hit.

Decisions to keep defecfive products in the manufacturing process, however, increased capital investments, like labor and energy, in products that were sold at a lower price, scrapped, or rerolled. Although these deci-

economic action" (Weber 1978:85, 108-109; also see Antonio 1979).

732 AMERICAN SOCIOLOGICAL REVIEW

sions ensured that managers remained within tbeir budgets, the production of low-quality steel decreased revenues or increased manu- facturing costs.

In two critical ways, budget controls re- sulted in behaviors tbat contradicted corpo- rate goals. First, together witb the increased size and complexity of each organizational unit (some of which had more tban 500 em- ployees), budget controls encouraged middle managers to focus on tbe organizational ac- tivities within their formal span of control. Second, because production managers were evaluated by their ability to remain within their budgets, this account-control system created an incentive structure for managers to make decisions based on the effect on their particular organizational unit. As Middle Manager A in Industrial Relations put it:

You had superintendents of mills, little kings; that was their fief. All they were concerned about was their mill. And they passed their products on to the next guy, who was only con- cemed with his mill.

Moreover, as the organization expanded, up- per-level middle managers became spatially separated from tbe point of producfion. As a result, they did not understand the details of specific manufacturing processes and were less capable of evaluafing whether cost in- creases were justified.

Together with the increased size and com- plexity of the manufacturing process, this in- centive structure gradually reified budget controls into production-unit goals. Although budget controls made production managers accountable to top management, they encour- aged these managers to disregard the conse- quences of their decisions for downstream production units, which contradicted the budget's intended function of controlling costs. This contradiction was not detected because managers' actions were not judged against the profitability goals of tbe corpora- tion. Budget control alone did not guarantee tbe efficient'^ use of resources because for- mal rationality is indifferent to whether ac- counting follows substantively rational prin- ciples.'^

'" Efficiency is defined as the ratio of inputs like labor and raw materials to outputs (Chandler 1962:37; Pfeffer and Salancik 1978:11).

'^Because formal rationality focuses on means.

Equally important, when flawed products "slipped through," downstream managers be- came responsible for those products. This practice of passing cost on to tbe manager of the next production unit created deep-seated "divisions" among these managers and, in several cases, led to "rivalry and open hostil- ity" (Middle Manager A, Industrial Rela- tions). These confiicts in social relations un- dermined the persona! cooperation necessary to ensure the efficient linking of the manu- facturing process.̂ "^ The subsequent increase in coordinating costs remained undetected because budget controls could not identify the manager responsible for the costs associ- ated with linking technologies that require the joint effort of several managers to yield a single outcome (March and Simon 1958:22- 32).

There are two reasons why these internal constraints on capital accumulation were not detected. First, the account-control sys- tem was cost-effective enough to generate an acceptable rate of return in Taggert's oli- gopolistic environment in which the indus- try limited competition by setting prices and coordinating production capability witb market demand. Second, as long as manag- ers remained witbin their budgets, they met the technical criterion of success. In short, tbis contradiction remained undetected be- cause tbe account-control system did not evaluate each organization unit's perfor- mance in relationship to substantive corpo- rate goals.

I

Competition and Crisis: Restructuring the Account-Controt System

In tbe 1960s and 1970s, Taggert's mean ca- pability utilization rates were 89 and 94 per- cent, and its mean rates of return were 9.6 and 9.7 percent respectively (Moody's 1960- 1988; public documents). By 1980, bowever, global steelmaking capability exceeded de- mand, and high-quality steel from domestic mini-mills and foreign producers began to

the consequences need not be rational. Rather, they only have to be traceable to the mode of cal- culation (Weber 1978:96; McNeil 1978; Antonio 1979).

"̂ This incentive structure created the same problem in other corporations {Business Week 1993:81).

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 733

1985

Figure 1 Raie of Return on Equity (After Taxes): Tagger! Steel Corporation and All Manufacturing Corporations

Sources: Moody's (1960-1988); U.S. E>epartment of Commerce (1965-1987).

erode integrated steel's market share. As a result, the oligopolistic price-setting .struc- ture of tbe U.S. steel industry collapsed. Be- tween 1975 and 1981, manufacturing cost increases exceeded price increases by 3.3 percent per year, a cumulative difference of 23.1 percent (public documents). By 1980, Taggert's rate of return had fallen to 2.3 per- cent. In response, Taggert implemented a cost-cutting and quality improvement strat- egy. The subsequent cost reductions were in- adequate. During the 1981-1982 recession, Taggert failed to realize a profit for the first time since the Great Depression (see Figure 1). Moreover, on the basis of its .calculation of price and cost trends, Taggert projected that its break-even point (i.e., when costs equal revenues) would increase from 77 per- cent of the capability utilization rate in 1982 to 86 percent in 1983. Taggert's capacity for accumulating capital was further constrained by its customers, who demanded higher quality steel so tbey could lower their costs, improve their product quality, and compete in their markets. Top Manager B in Corpo- rate stated:

Every year, without exception, the quality pa- rameters became more exact. One reason why this happened, and will continue, is that steel customers realized they could lower their manufacturing costs by insisting on higher quality steel. (Internal docuniient)

In response to these constraints on capital accumulation, Taggert developed a second

cost-cutting program to reduce its manufac- turing costs by $82 million in 1982 and $200 million in 1983 (public document). Aitbough tbis strategy reduced operating costs, it did not restore profits or increase product qual- ity. The steel industry remained highly com- petitive, and for tbe third consecutive year Taggert did not realize a profit (Eigure 1). Under these conditions of capital accumula- tion, "quality became the key to better earn- ings" (Top Manager A, Manufacturing, inter- nal document). A core problem remained: The incentive structure created by the ac- count-control system encouraged production managers to make decisions without regard for their effects on downstream units, result- ing in conflict, "adversarial relationships, and lack of cooperation" (Top Manager C, Manufacturing, internal document).

To overcome internal and external con- straints to capital accumulation, Taggert ini- tiated a restructuring strategy. According to Middle Manager A in Sales,

When we went into the reorganization, we took the view that the marketplace had changed, our customers had changed, their needs had changed, the demands are changing, and as a result of that we have got to identify those things better and have a market driven steel company. Not have it driven by production, not have it driven by technology, but have it driven by the marketplace.

The cornerstone of restructuring the mana- gerial process entailed tbe intensification of rational calculation. In 1984, Taggert began

734 AMERICAN SOCIOLOGICAL REVIEW

to collect more precise data on tbe manufac- turing process to standardize decision-mak- ing criteria, introduce more precise cost con- trols, improve product quality, and hold man- agers accountable for their decisions.

The processes of defining and organizing ob- jectives are built on the need for standardiza- tion and conformance to customer specifica- tions. The basic concept behind this process is that once we establish a base, the operating de- partments are accountable for managing to that base. (Middle Manager B, Research, internal document; my emphasis)

Like many other corporations, Taggert had established a "standard cost system" in tbe 1950s. In the 1970s, Taggert established a financial concepts seminar to educate management in the use of "financial ac- counting and cost accounting information . . . to make better decisions" (Middle Man- ager A, Accounting). But, in the absence of an objective need for it, rational calculation evolved slowly throughout tbe oligopolistic era. For example, Taggert began to "look at product detail in tbe late 1960s" (Middle Manager D, Accounting), and by 1972 bad identified 1,200 cost points (Middle Man- ager A, Systems, public document). In con- trast, wben profits declined and competitive pressures intensified, Taggert accelerated the rate of calculation. By 1985, Taggert bad identified more than 50,000 cost points (Middle Manager D, Accounting).'^' More- over, the corporation used these data in sig- nificantly different ways: Whereas the ac- count-control system had collected cost data to generate budget controls for each produc- tion unit, the system was restructured to col- lect data on more cost inputs and to calcu- late more precisely "the actual cost that products accumulate" as they move through tbe manufacturing process (Middle Manager B, Accounting).

In our manufacturing processes there are a couple of thousand steel specifications or reci- pes. In the past, we had some [product] groups that didn't have specific specifications. There are vast differences in the cost alloys; tellurium

'̂ A cost point is a cost that is identified. Cost points include, for example, the energy cost to transform raw materials (e.g., limestone, iron ore, etc.) into molten steel.

versus some steel that has rare earth additives in it. The cooking time, heating practices, manufacturing process, and the raw steelmak- ing end of it also create variances in the cost. (Middle Manager D. Accounting)

Restructuring the account-control system also entailed disaggregating existing cost data.

What we did is pull out costs that were grouped, that were never identified. Some costs, for cer- tain variables, you didn't know because you just had cost on the whole [organization] unit. (Middle Manager D, Accounting)

In addition, Taggert collected market data and calculated tbe profit margin on eacb of its 5,000 products.

We measure the competitive gap by . . . identi- fying the gap between our cost and the Japa- nese product delivered in our market area. (Middle Manager D, Accounting)

Once Taggert made these calculations, it could identify and eliminate its less efficient production units.

Centralizing Control Over the Managerial Process

Once collected, Taggert entered these cost and market data into a central database. Technical experts (e.g., accountants and in- dustrial engineers) accessed this information to define cost centers at tbe point of produc- tion, establish cost controls for these centers, and determine if manufacturing costs ex- ceeded the pre-establisbed costs. These ac- tivities were centralized in a new organiza- tion unit that combined the quality control and technology units. This new unit, the manufacturing decision center, was elevated to the top of the manufacturing hierarchy, equivalent to the plant manager.

In contrast to recent analyses suggesting that control is restricted to evaluating entire product lines (Chandler 1990:232; Fligstein 1990:15), technical experts in the manufac- turing decision center used these data to cal- culate the most cost-efficient means of pro- ducing each product. These data were then used to determine whicb products to manu- facture and to establish more precise controls in their manufacture. According to Middle Manager A in Accounting:

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 735

Determining the profitability of each product and the cost effectiveness of the production units is fundamental to [our] conception of ac- counting and business management.

Taggert couid then "maximize profits . . . through selective marketing" (Middle Man- ager D, Accounting).

The more precise cost data were also used to intensify managerial accounting, the "di- mension of accounting that transforms stan- dardized cost data into decision-making in- formation" (Middle Manager A, Account- ing). This dimension of tbe account-control system provided tbe quantitative data which established tbe premise for making a deci- sion. By 1987, Taggert had computer moni- tors at each decision site to transmit tbe most recent accounting data from tbe decision center to the point of production.

Tbe identification of additional cost cen- ters also made it possible to intensify respon- sibility accounting—cost accounting by area of responsibility. By specifying costs more precisely, the account-control system simul- taneously identified where in the manufac- turing process cost inputs occurred. As Mid- dle Manager A in Accounting stated:

One section [of the accounting activity] is called product cost and their basic job is to take the cost information from the standard cost sys- tem to identify the area of responsibility.

The previous account-controls had identified production units (e.g., blast furnaces); the new controls established cost centers at the point of production to monitor specific deci- sions. As described by Middle Manager A in Accounting:

The cost center is a control function. The more cellular the cost-centering structure, the more a cell can be examined [to determine] if it is functioning reasonably. (My emphasis)

In addition, bureaucratic controls defining spans of control, lines of autbority, and areas of responsibility were specified in greater detail: "The limits of power or decision-mak- ing were escalated up the ladder to be more clearly delineated, reviewed, and approved" (Middle Manager A, Corporate). While re- sponsibility accounting identified the points at which specific cost inputs exceeded pre- determined costs, bureaucratic controls iden-

tified the manager responsible for those costs. Together these formal controls made it possible to identify the middle managers re- sponsible for the costs of coordinating the manufacturing process, and the lower-level managers responsible for specific production costs. These controls over the managerial process also enabled upper-level managers to evaluate the cost-efficiency of specific deci- sions and hold even the lowest-level manag- ers accountable for their decisions.

Production problems that could not be readily resolved by premise controls were al- located to management teams—groups of managers whose areas of responsibility were affected by particular (typically coordinat- ing) decisions. To overcome the conflictual social relationships among these managers and to better control tbe collective behavior of tbese teams, Taggert implemented a team leadership program. Top management bired outside consultants to establish a unifted set of values to encourage lateral communica- tion, promote participatory decision-making and problem-solving, encourage the delega- tion of decision-making from middle- to low- level managers and show how eacb manager's responsibilities and decisions con- tributed to costs.^^

The team process was an attempt to get all of these people singing the same song—coordi- nating, thinking the same, communicating the same values downward, having the same pri- orities. (Middle Manager B, Human Relations, my emphasis)

These teams also used rationally calculated data to make decisions. Witbin this structure, when a difference between standard costs and actual costs was identified, management teams investigated the source and collec- tively made a decision based on the cost data. The emphasis on team decision-making, to- gether with a predetermined rationally calcu- lated criterion, operated as an unobtrusive control—control of the cognitive premise underlying action (Perrow 1986:129). It also established a mechanism whereby managers supervised each other to ensure that they em-

retain the loyalty of managers whose au- thority was being eroded, these consultants also held seminars on how the restructured manage- rial process empowered managers.

736 AMERICAN SOCIOLOGICAL REVIEW

ployed rationally calculated data in the deci- sion-making process.

After specifying cost and bureaucratic con- trols, Taggert implemented technical controls under the guidance of engineers from a Japa- nese steel corporation. Tbis phase of restruc- turing placed coordination of the production process in the newly created "operations con- trol center." This control center cyberneti- cally scheduled orders with similar produc- tion specifications to reduce the downtime needed to adjust tbe manufacturing process to meet different product specifications. This new unit transformed steelmaking from batch production to continuous processing, which links "operations leading to a more efficient fiow . . . from initial steeimaking to delivery at the customer's plant" (internal docu- ment).^^ However, continuous processing is not merely a different way to organize pro- duction; it entails centralization of control over coordinating activities.

The second technical control was the in- corporation of statistical quality control (SQC) directly into the manufacturing pro- cess. Statistical quality control applies statis- tical principles and techniques to solve pro- duction problems and increase control over the manufacturing process, which in turn im- proves the cost economics of manufacturing a product (Deming 1969:367). By the early- 1980s, General Motors began to implement SQC in its manufacturing processes, and in 1984 it informed Taggert tbat it would no longer purchase Taggert's steel if it did not implement SQC (Middle Manager C, Indus- trial Relations). General Motors assumed that SQC would improve the quality of steel, and higher quality steel would lower the cost of manufacturing automobiles. Although Taggert had begun to implement SQC in 1983, this demand from its largest customer in its largest market sped up the implementa- tion of SQC.

The use of SQC was not limited to Tag- gert's customers in tbe automobile industry. As Middle Manager C in Systems reported:

Each of our customers is using statistical qual- ity programs to come up with specific measure-

^̂ Continuous processing lowers costs by re- ducing capital investments in unfinished invento- ries and by cutting expenditures on the energy used to reheat unfinished steel.

ments of what they feel the quality ought to be from their steel suppliers. And they are con- tinuing to narrow the window on those require- ments.

Supplier corporations like Taggert used SQC to improve quality and meet the statistical parameters established by their customers in two ways. First, Taggert used SQC to evalu- ate tbe tolerance capabilities of its manufac- turing facilities (i.e., to determine whether a production unit could manufacture a product to customers' specifications). Second. Tag- gert used statistical techniques to "monitor every production process to make sure that tbe right material goes in and the right mate- rial comes out" (Top Manager B, Sales, in- ternal documents).

We used to inspect quality. But we never had a statistical program in place that we could make sure that we were meeting the quality standard that the customer wanted. [That] dramatically changed the way we are doing our manufactur- ing process. {Middle Manager C, Systems)

As was true for cost control and bureau- cratic control, Taggert implemented techni- cal controls to overcome tbe contradiction between organization unit and corporate goals. SQC "is a means to an end. It defines quality as conformance to [preestablished] standards. That is our end" (interna] docu- ment). ,

Computerization: Further Standardization

By 1986, technical experts in the manufac- turing decision center had calculated "the standards for all significant product charac- teristics" (internal document). When Taggert received an order, sales personnel transferred the product specifications to the manufactur- ing decision center where technical experts programmed the production control comput- ers to those standards. Transmitting decisions directly to the point of production bypassed operating managers who had previously made those decisions.

Statistical controls and computers were also used to establish a reiterative control system—a system that repeatedly measures and monitors the product as it is being manu- factured to ensure conformance to predeter- mined standards. If the product exceeds the rationally calculated tolerance limits, the

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 737

production control computer automatically adjusts tbe technology during the manufac- turing process to conform to the "standards necessary to meet the customer's require- ments" (Middle Manager B, Systems).

Production units are cotnputer-monitored for conformance to the new standards to assure maximum control of manufacturing and the identification and more timely correction of quality problems during manufacturing. (Inter- nal document, my emphasis)

This is a combination of statistical process con- trol, which is a tool for monitoring the output of a process, and standardization, which is an approach to guarantee customers their specified product. The idea is to find the best way of do- ing something, then make sure it is done that way every time. (Top Manager A, Manufactur- ing, internal document; my emphasis)

Implementation of these controls involved a shift in the systems department from com- puter applications to database management. As Middle Manager C in Systems stated:

Since I came to the company [1976], systems has moved from a developer of [computer] ap- plications to being responsible for corporate data bases. We are the people who . . . make that data available to ihe people who need it to make good decisions.

Using these data, experts from systems and other technical areas (e.g., engineers) devel- oped computer models to identify the most cost-efficient blend of raw materials and pro- cess variables (e.g., heat, rolling pressure) required to meet product quality goals. These computer models not only eliminated many decisions, but processed the information in "less than two minutes." In tbe past, "it would have taken a couple of people three or four weeks to make those calculations" (Middle Manager A, Research, internal document).

Hyper-Centralization and Formalization of Control: The Hierarchiat Computer System

To process and transmit these data to the ap- propriate location in the corporation, Taggert restructured its computing activities by es- tablishing an interconnected three-level com- puter hierarchy. Tbe business computer at the top of the hierarchy accessed information to coordinate raw material inventory with sales and production schedules, traced each prod-

uct from tbe titne its raw materials entered the manufacturing process to its delivery date, and stored information on billing and projected delivery dates.

At the lowest level, Taggert used comput- ers to increase automated management—the application of computer information-pro- cessing systems to aid in decision-making and controlling the manufacturing process (McManus 1983; Pond and Harvey 1985). Computer-aided manufacturing (CAM) pro- grams identified and immediately corrected problems with individual manufacturing technologies. These programs corrected de- fects as the steel was being manufactured by "automatically measuring the sbape and ad- justing the rollers to increase the pressure on the stress point" (Middle Manager D, Sys- tems).

[These controls take] the humanistic element out of the decision-making process. The opera- tors used to have to eyeball the strip as it fed through the mill and manually adjust it to en- sure flatness. The new system automatically measures the shape and adjusts the rollers, cor- recting any shape defects as the steel is being rolled in order for us to make a good, consis- tent product. (Top Manager B, Manufacturing, internal document)

These applications of formal rationality rep- resent an extreme example of the spatial separation of conception and execution: Computers executed decisions conceived in the manufacturing decision center. For pro- duction problems that exceeded the capabili- ties of tbe technology, production managers consulted with tbe decision center or used managerial accounting techniques to solve the problem.

At the middle level, a computer-integrated manufacturing (CIM) system connected the individual CAM computers so that manufac- turing was a continuous process. Together tbe CAM and tbe CIM systems formalized dimensions of tbe managerial process tbat had previously escaped calculation.

By 1989, the implementation of tbis com- puter hierarchy, which "provides a linkage between order processing, planning, sched- uling, manufacturing, and customer service" (public document), was complete. These cen- tralized controls were set up to ensure the continuous use of production units and to standardize decision-making.

738 AMERICAN SOCIOLOGICAL REVIEW

The Tightly Coupted. Flexible Organization

This hierarchy of computers also tightened horizontal and vertical coupling. A new "plant planning and scheduling system," which was linked to the "operations control center," calculated the information pro- grammed into tbe CIM system. These cyber- netic controls tightened horizontal coupling of each production unit by centralizing and formalizing the coordinating functions. This dimension of tbe control system reduced (1) the downtime needed to realign the manufac- turing facilities, (2) the capital investment in semi-finished inventories, and (3) tbe energy used to reheat products manufactured in dis- crete stages. Taggert more tightly coupled the manufacturing process to achieve

. . . market-oriented goals and maintain a con- tinuous processing strategy, which requires that units be very tightly coupled together. Continu- ous processing cannot succeed unless functions are centralized and performed by an organiza- tion whose perspective is the entire production process. (Top Manager A, Manufacturing, in- ternal document; my emphasis)

Similarly, the hierarchial computer system vertically coupled tbe point of production di- rectly to the decision center. This vertical link enabled tbe decision center to control the computer-aided manufacturing (CAM) system and transmit premise controls to the computer terminals at the point of produc- tion. If customers requested changes in an order, this information was transmitted to the operators, who located and changed these speciftcations minutes before manufacturing would begin. These controls tightly coupled the organization, while increasing respon- siveness and flexibility by increasing the organization's capacity to transmit decision- making information "to operations, or sales, or general management so they can make tbe proper decision, wbich could vary under dif- ferent economic situations" (Middle Man- ager D, Accounting; my emphasis).

Whereas production managers under tbe previous control system were semi-autono- mous and controlled conception and execu- tion, restructuring centralized many concep- tion functions, decentralized execution, and held managers responsible for proper execu- tion. These new controls over the managerial

Table 2. Transformation of the Managerial Hierarchy: Taggert Corporation. 1960 to 1985

Old Managerial Hierarchy New Managerial Hierarchy

Top Management

1. President

2. Vice-President

President

Vice-President

Middle Management

3. General Manager General Manager

4. Assistant General Assistant General Manager Manager

5. Manager

6. Superintendent

7. Assistant Superintendent

Lower Management

8. General Foreman Section Manager

9. Assistant General Supervisor Foreman

10, Foreman

process reduced the information-processing activities of middle management. As Middle Manager C in Systems put it:

Middle managers used to be the people that provided the information. They acted as an in- formation source. With the increase in informa- tion technology, middle management through- out America is threatened because they are no longer the information sources. (My emphasis)

By reducing the information-processing activities of middle management, these con- trols reduced tbe number of decisions, deci- sion-makers, and layers of the managerial hi- erarchy. By tbe late 1980s, Taggert had re- duced the number of middle levels of man- agement from five to two. and lower levels of management from three to two (Table 2). These changes eliminated 28 percent of Taggert's managers. Restructuring of tbe rtianagerial process also entailed restructur- ing of skill. Managerial skills shifted from understanding how materials react to changes in the manufacturing process (e.g., tbe effects of pressure, beat) to an under- standing of the account-control *̂

-** Although contemporary manufacturing en- tails restructuring of skills, whether more or less skill is required is a separate issue (see Burris 1993).

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 739

SUMMARY

Taggert restructured its managerial process to overcome constraints on capital accumu- lation tbat resulted from contradictions em- bedded in its account-control system. These contradictions produced a crisis in the early 1980s when the conditions of capita! accu- mulation changed. Top management's deci- sion to establish more precise formal con- trols was also influenced by Taggert's oligopsonic relationship with steel consum- ers, wbo were under pressure to increase tbe quality of tbeir products so as to ensure prof- its in tbeir industrial sectors.

Tbis study of Taggert Steel Corporation produced several findings. First, hyper-quan- tification was the cornerstone of Taggert's strategy to more precisely regulate and con- trol the managerial process. The constraints on capital accumulation renewed an empha- sis on breaking down the manufacturing pro- cess into discrete activities, identifying tbe costs of those activities, and centralizing control over those activities.^^ Taggert inten- sified rational calculation to standardize pro- duction decisions, coordinate links among tbe production units, and guarantee tbat the goals of the production units advanced the goals of tbe corporation. Whereas managers previously were semi-autonomous, Taggert centralized autbority to unify "the entire managerial system into one hierarchical sys- tem" (Top Manager B, Manufacturing, inter- nal document). These findings support the first proposition: that Taggert intensified ra- tional calculation to overcome constraints on capital accumulation.'^^

Second, the corporation centralized market data and cost data. Technical experts used these data (1) to specify tbe means to manu- facture the highest quality product at the lowest cost, and (2) to establish controls over

-•'' Although technology was an important fac- tor in corporate restructuring, the technologies discussed had been available for several decades. Statistical quality control had been available since the 1950s, and the computer capacity to process and transmit information had been avail- able before the 1970s.

^̂ This finding is consistent with research on the labor process which concluded that greater control emerges from changes in the capital ac- cumulation process (Edwards 1979:viii),

the managerial process to ensure that the product was manufactured to specifications. Taggert also elaborated its bureaucratic con- trols so it could identify the managers re- sponsible for specific decisions. Together these controls detected when operating man- agers did not follow premise controls. More- over, team decision-making, in which team members monitored each otber to ensure that the formally rational criterion is used, oper- ated as an unobtrusive premise control. Managerial teams had autonomy in making some decisions, but this team "autonomy" was premised on the rationally calculated data available to them. These findings sup- port the second proposition: that decision- making information is centralized and tben distributed on a "need-to-know basis" to in- crease control over the managerial process.

Third, these controls tightened the coupling of two spheres of the corporation. On the one hand, the decision center was tightly coupled to the manufacturing process, which was transformed from batch production to con- tinuous processing. On the other hand, mar- kets were tightly coupled to tbe point of pro- duction. These formal controls enabled the organization to transmit product specifica- tions from the top of the manufacturing hier- archy directly to the point of production, thereby increasing tbe corporation's respon- siveness to changing market conditions. This finding supports the third proposition: that in- formation technologies are established to in- crease the rate at which information can be transferred throughout the corporation and hence increase corporate responsiveness and flexibility.

Fourth, these attempts to redefine tbe single "one best way" to manufacture each product separated conception from execution by placing production managers fartber away from the loci of decision-making. Whereas execution occurred at the point of produc- tion, Taggert relocated conception activities, like processing information and determining what was to be done, to the decision center where experts made decisions and stipulated the parameters governing other decisions.^^

^̂ In contrasi to arguments that see engineers as support staff operating outside the managerial hierarchy (Mintzberg 1979; Chandler 1990:43), experts are incorporated into the hierarchy and located closer to its apex.

740 AMERICAN SOCIOLOGICAL REVIEW

Tbis organizational control transmitted the premise of decisions througbout tbe manage- rial hierarchy and embedded many decisions directly in the manufacturing process. The new managerial process replaced many of the information-processing activities of manage- ment and controlled others more tightly, wbich tben eliminated four of the eight lay- ers of production management. This finding supports the fourth proposition: that the in- formation-processing activities of production managers declined, reducing the number of managers and levels of the managerial hier- archy needed.

An important question remains: Was the hyper-quantification and centralization of in- formation used to establish control over the collective behavior of management at Taggert an exception? Although the crisis in capital accumulation at Taggert may bave ac- celerated the rate of rational calculation to a degree tbat may not have been typical of other corporations, I suggest that the con- straints on capital accumulation and tbe re- source dependence among corporations are promoting rational calculation as a way to centralize and formalize control over the managerial process throughout the manufac- turing sector.

When Taggert's customers demanded that Taggert adopt rational controls, Taggert then pressured its suppliers to adopt similar con- trols. In the late 1980s, Taggert imple- mented a strategy of buying from suppliers based on whether they implemented statisti- cal quality controls. To enforce compliance, Taggert established inspection teams that used statistical techniques to evaluate sup- pliers' manufacturing processes to deter- mine tbeir capacity to meet Taggert's prod- uct specifications. Moreover, Taggert re- duced "tbe number of suppliers" (i.e., single-sourcing) based on their use of these rationally calculated controls (Top Manager A, Sales). Taggert became more tightly coupled to a small number of suppliers be- cause the reaction of standardized inputs during manufacturing could be more accu- rately predicted, thus improving the corporation's capacity to meet the quality standards imposed on it by its customers. Tbis suggests tbat tighter resource-depen- dent organizational networks are emerging, which pressure corporations to establish

more formal controls over their managerial processes.^^

DISCUSSION

This case study situates restructuring of the managerial process within the historically specific context of capital accumulation and clarifies how the corporation simultaneously centralized control and increased its flexibil- ity. Tbe analysis shows that transformation of tbe social structure is an inherently dialecti- cal process, tbat historically specific social structures provide incentives and motives for action (Prechel 1990), that contradiction is a central property of organizational change (Benson 1977; Heydebrand 1977; Prechel 1991b), and that macro-level constraints on capital accumulation produce meso-level (the corporation) transformations tbat inten- sify control over micro-level behaviors.

Contingency theory explains organiza- tional change during an oligopolistic era when rapid economic growth and industry's oligopolistic behavior ensured stable mar- kets. During this period, organizational com- plexity increased, and the corporation decen- tralized decision-making and authority. Flex- ible coordination and control of tbe manufac- turing process were achieved through infor- mal communication and cooperation among production managers. However, contingency theory claims (that centralization reduces flexibility and rapidly changing environ- ments produce decentralization) do not ex- plain contemporary restructuring of the managerial process. Clearly, Taggert simul- taneously centralized control and increased flexibility in response to the rapidly chang- ing conditions of capital accumulation.

Decision-making theory specifies two ways in which centralization occurs. First, the number of information sources available to decision-makers is reduced to limit differ- ing perceptions. Second, premise controls are established to restrict bebavior by limit- ing the content and flow of information. However, the tbeory incorrectly posits that

*̂ The rise in CAD-CAM industry shipments from less than $100 million in 1976 to $5.6 bil- lion in 1986 (Harris 1987) suggests that these technical controls are being implemented throughout industry.

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 741

premise controls are restricted to the higher levels of the managerial hierarchy and, like contingency theory, suggests that coordina- tion of the manufacturing process is achieved through informal communication and coop- eration among production managers.

My study supports a post-Fordist argu- ment, that in late capitalism information be- comes more important to the decision-mak- ing process, and information is used to in- crease organizational flexibility. However, post-Fordism presents no clear conceptual framework to explain how information is used to increase organizational flexibility, or whetber centralization or decentralization is occurring. Therefore, post-Fordist theory has a limited capacity to explain recent changes in the managerial process. Similar to contin- gency theory and decision-making tbeory, post-Fordist theory does not explain how the corporation formalizes and centralizes con- trol while increasing its capacity to respond to a rapidly changing environment.

In contrast, neo-Fordist decision-making theory can explain these transformations in the managerial process. Top management at Taggert implemented a strategy of hyper- quantification and centralization to increase its control over the collective bebavior of managers. Data were used to program pro- duction-control computers, establish premise controls over manufacturing decisions, iden- tify managers responsible for executing deci- sions, and monitor whether operating manag- ers followed the decision-making parameters established at higher levels in the managerial hierarchy. On the one hand, these transfor- mations in the managerial process increased surveillance over operating managers. On the other hand, tbese controls advanced standard- ization by limiting the search for alternatives at tbe point of production, which enhanced tbe probability of predictable outcomes.'^'

'̂' Although the concern here is with transfor- mations in the managerial process, a partial as- sessment of outcome is possible. Taggert became the first U.S. integrated steei company to obtain contracts with a Japanese automobile corporation. It also received product quality awards from Japa- nese and U.S. automobile corporations. In addi- tion, the number of managers per million tons of steel produced declined from 1,125 in 1982 to 737 in 1986, a 35 percent increase in output per manager.

This case study corroborates several other studies on organizational change that empha- size historical conditions and crises. Corpo- rate restructuring at Taggert represented the corporation's adaptation to specific eco- nomic conditions (Hamilton and Biggart 1988; Prechel 1991b). Moreover, new orga- nizing techniques and revolutionary changes emerged from crises (Benson 1977; Prechel 1991b), which in turn resulted in formaliza- tion and tightening of organizational controls (Pfeffer and Leblebici 1973). Centralization and formalization occurred because modern capitalism demands "calculability of results" (Weber 1978:975). These efforts to increase standardization are intertwined with in- creased control over participants.

This conclusion challenges decentraliza- tion arguments, which suggest that corporate restructuring extends managerial freedom, widens informal networks, and increases aw- tonomy—control over work activities and de- cisions, and freedom to be innovative (Freidson 1970; Larson 1977). I suggest that the decentralization argument is flawed be- cause it has not analyzed recent changes in the managerial process in sufficient detail, and has obscured what is being centralized and decentralized. Tbe argument fails to rec- ognize the spatial separation of conception from execution, and does not acknowledge tbat decentralization entails more precise controls at the point of production to ensure tbe concomitant standardization of social ac- tion and product quality.

Tbe way the centralization of control af- fected tbe managerial process resembles the way the transfer of information from labor to management in the early twentieth century affected the labor process.^'^ As was true of tbe labor process (Braverman 1974), once Taggert could process information and trans- mit it throughout the organization, general knowledge of the manufacturing process by production managers became unnecessary to tbe functioning of the system, and informa- tion was distributed on a "need-to-know ba-

•'" Just as the centralization of authority over the labor process was not absolute (Buraway 1979), centralization of authority over the mana- gerial process was not absolute. Rather, the his- torical trend is toward centralization of control over more spheres of the manufacturing process (Weber 1946).

742 AMERICAN SOCIOLOGICAL REVIEW

sis." In addition to using technical controls to exercise direct control over tbe manufac- turing process, experts in the decision cen- ters distributed information in such a way that it defined the premise of operating deci- sions. These controls over tbe managerial process separated conception from execution in order to increase tbe predictability of management's collective behavior.

These formal controls were based on the principles of scientific management: mea- surement, quantification, and the separation of conception from execution (Taylor [1911] 1967). However, tbe application of these principles in the contemporary era extends scientific management beyond tbe tall, hierarchial and rigid, rule-based managerial system tbat characterized Taylorism and Fordism in three ways. First, under neo- Fordist decision-making, controls rapidly process, change, and transmit decision-mak- ing information. Second, neo-Fordist con- trols have flatter hierarchies because tbey eliminate many of the information-process- ing activities of management. Tbird, the ca- pacity of neo-Fordist controls to transmit in- formation tightly couples the corporation, while increasing its responsiveness and flex- ibility.

HARLAND PRECHEL is an Associate Professor of Sociology at Texas A&M University. He is cur- rently engaged in a study of restructuring the cor- porate form in the 1980s and 1990s. This is part of a larger project on corporate change in the postwar political economy of the United States. His other interests include economic sociology, historical methods in sociology, and historically grounded theory.

REFERENCES

Antonio, Robert. 1979. "The Contradiction of Domination and Production in Bureaucracy: The Contribution of Organizational Efficiency to the Decline of the Roman Empire." Ameri- can Sociological Review 44:%95-9\2.

Akard, Patrick. 1992. "Corporate Mobilization and Political Power: The Transformation of U.S. Economic Policy in the 1970s." American Sociological Review 57:597-615

Alford, Robert and Roger Friedland. 1985. Pow- ers of Theory: Capitalism, the State, and De- mocracy. Cambridge, England: Cambridge University.

Attewell, Paul. 1987. "Big Brother and the Sweatshop: Computer Surveillance in the Au-

tomated Office." Sociological Theory 5:87-99. Barnett, Donald and Louis Schorsch. 1983. Steel

Upheaval in a Basic Industry. Cambridge, MA: Bal linger.

Baxter, Vern. 1989. "The Process of Change in Public Organizations." The Sociological Quar- terly 30:283-304.

Benson, Kenneth. 1977. "Organizations: A Dia- lectical View." Administrative Science Quar- terly 2 0 : 2 2 9 ^ 9 .

Blau, Peter and W. Richard Scott. 1962. Formal Organizations. San Francisco, CA: Chandler.

Bowles, Samuel and Richard Edwards. 1985. Un- derstanding Capitalism. New York: Harper and Row.

Braverman. Harry. 1974. Labor and Monopoly Capital: The Degradation of Work in the Twen- tieth Century. New York: Monthly Review.

Burawoy, Michael. 1979. Manufacturing Con- sent. Chicago, IL: University of Chicago.

Burns, Tom and George Stalker. 1961. The Man- agement of Innovation. London, England: Tavi stock.

Burris, Beverly. 1993. Technocracy at Work. Al- bany, NY: State University of New York.

Byrne. John. 1993. "Horizontal Corporation." Business Week, 20 D e c , pp. 76-81.

Carter, Bob. 1985. Capitalism. Class Conflict, and the New Middle Class. London, England: Routledge and Kegan Paul.

Chandler, Alfred. 1962. Strategy and Structure. Cambridge, MA: Massachusetts Institute of Technology.

. 1977. The Visible Hand. Cambridge, MA: Belknap-Harvard University.

1990. Scale and Scope. Cambridge, MA: Belknap-Harvard University.

Chatfield, Michael. 1977. A History of Account- ing Thought. Huntington, NY: R.E. Kreiger.

Child, John. 1972. "Organization Structure and Strategies of Control: A Replication of the Aston Study." Administrative Science Quar- terly 17:163-77.

. 1984. Organization. London, England: Harper and Row.

Clawson, Dan. 1980. Bureaucracy and the Labor Process: The Transformation of U.S. Industry, 1860-1920. New York: Monthly Review.

Ciegg, Stewart. 1981. "Organizational Control." Administrative Science Quarterly 26:545-62.

Deming, W. Edwards. 1969. "Statistical Control of Quality in Japan." Paper presented at the In- ternational Conference on Quality Control, 2 2 - 24 Oct., Tokyo, Japan.

Dohse. Knuth, Ulrich Jurgens, and Thomas Malsch. 1985. "From 'Fordism' to Toyotism'? The Social Organization of the Labor Process in the Japanese Automobile Industry." Politics and Society 14:115-46.

Dornbusch, Stanford and W. Richard Scott. 1975.

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 743

Evaluation and the Exercise of Authority, San Francisco, CA: Jossey-Bass.

Edwards, Richard. 1979. Contested Terrain: The Transformation of the Workplace in the Twen- tieth Century. New York: Harper.

Fligstein, Neil. 1990. The Transformation of Cor- porate Control. Cambridge, MA: Harvard Uni- versity.

Florida, Richard and Martin Kenney. 1991. "Transplanted Organizations: The Transfer of Japanese Industrial Organization to the U.S." American Sociological Review 56:381-98.

Fortune. Stewart, Thomas. 1991. "The New American Century: Where We Stand." Spring/ Summer, pp. 12-23.

Fombrun, Charles. 1992. Turning Points: Creat- ing Strategic Change in Corporations. New York: McGraw-Hill.

Freidson, Eliot. 1970. Professional Dominance. Chicago, IL: Aldine.

Galbraith, Jay. 1973. Designing Complex Orga- nizations. Reading, PA: Addison-Wesley.

. 1977. Organization Design. Reading, PA: Addison-Wesley.

Garner, S. Paul. 1976. Evolution of Cost Account- ing to 1925. Tuscaloosa, AL. University of Ala- bama.

Gordon, David, Richard Edwards, and Michael Reich. 1982. Segmented Work, Divided Work- ers: The Historical Transformation of Labor in the United States. Cambridge, England: Cam- bridge University.

Hage, Jerald and Michael Aiken. 1969. "Routine Technology, Social Structure, and Organiza- tional Goals." Administrative Science Quar- terly 14:366-76.

Hall, Richard. 1991. Organizations: Structures. Processes, and Outcomes. Englewood Cliffs, NJ: Prentice-Hall.

Hamilton, Gary and Nicole Woolsey Biggart. 1988. "Market. Culture, and Authority: A Comparative Analysis of Management and Or- ganization in the Far East." American Journal of Sociology 94(Supp.):S52-S94.

Harris, Deborah. 1987. "U.S. CAD/CAM Indus- try: Bright Future in Foreign Markets." Busi- ness America 10:2-5.

Harvey, David. 1991. The Condition of Post- modernity. Cambridge, MA: Basil Blackwell.

Heydebrand, Wolf. 1985. "Technarchy and Neo- Corporatism: Toward a Theory of Organiza- tional Change Under Advanced Capitalism and Early Socialism," Current Perspectives in So- cial Theory 6:1 \-\2%.

1977. "Organizational Contradictions in Public Bureaucracies." The Sociological Quar- terly 18:85-109.

Howard, Robert. 1985. Brave New Workplace. New York: Viking Penguin.

Johnson, H. Thomas and Robert Kaplan. 1991.

Relevance Lost: The Rise and Fall of Manage- ment Accounting. Boston, MA: Harvard Busi- ness School.

Kalberg, Stephen. 1983. "Max Weber's Univer- sal-Historical Architectonic of Economically- Oriented Action." Current Perspectives in So- cial Theory 4:253-88.

Kanter, Rosabeth Moss. 1989. When Giants Learn to Dance. New York: Simon and Schu- ster.

Kelley, Maryellen. 1990. "New Process Technol- ogy, Job Design, and Work Organization: A Contingency Model." American Sociological Review 55:\9\-20i.

Larson, Magali S. 1977. The Rise of Profession- alism. Berkeley, CA: University of California.

Lawrence, Paul. 1981. "The Harvard Organiza- tion and Environment Research Program." Pp. 3311-37 in Perspectives in Organizational De- sign and Behavior, edited by A. H. Van de Ven and W. F. Joyce. New York: Wiley Inter- science.

Lawrence, Paul and Jay Lorsch. 1967. Organiza- tions and Environment. Boston, MA: Harvard University Graduate School of Business Ad- ministration.

Lazonick. William. 1992. "Controlling the Mar- ket for Corporate Control." Pp 153-199 in En- trepreneurship, Technology. Innovation and Economic Growth, edited by F. Scherer and M. Perlman. Ann Arbor, MI: University of Michi- gan.

Levenstein, Margaret. 1991. "The Use of Cost Measures: The Dow Chemical Company, 1890-1914." Pp. 7I-116 in Inside the Business Enterprise: Historical Perspectives on the Use of Information, edited by P. Temin. Chicago, IL: University of Chicago.

Lincoln, James and Arne Kalieberg. 1985. "Work Organization and Workforce Commitment: A Study of Plants and Employees in the U.S. and Japan." American Sociological Review 50:738- 60.

Lyng, Stephen and Lester Kurtz. 1985. "Bureau- cratic Insurgency: Vatican and the Crisis of Modernism." Social Forces 63:901-22.

March, James and Herbert Simon. 1958. Organi- zations. New York: John Wiley and Sons.

Marx, Karl. 11887] 1977. Capital. Vol. I. New York: Vintage Books.

. [1893] 1981a. Capital. Vol. 2. New York: Vintage Books.

. [1894] 1981b. Capital. Vol. 3. New York: Vintage Books.

McManus, George. 1983. "Steel Takes an Orga- nized Look at Quality." Iron Age, 3 Oct., pp. MP9-MP22.

Mintzberg, Henry. 1979. The Structuring of Or- ganizations. Englewood Cliffs, CA: Prentice- Hall.

744 AMERICAN SOCIOLOGICAL REVIEW

M o o d y ' s . 1960-1988. Moody's Handbook of Common Stock. New York: Dun and Bradstreet.

McNeil, Kenneth. 1978, "Understanding Organi- zational Power: Building on the Weberian Legacy." Administrative Science Quarterly 23:65-90.

Noble. David. 1977. American by Design. New York: Alfred A, Knopf.

Oucbi, William and Mary Ann Mcguire. 1975. "Organizational Control: Two Functions." Ad- ministrative Science Quarterly 20:559-69.

Peters, Thomas and Robert Waterman, 1984. In Search of Excellence. New York: Harper and Row,

Perrow, Charles, 1967. "A Framework for the Comparative Analysis of Organizations," American Sociological Review 32:194-204.

. 1970, Organizational Analysis: A So- ciological View. Belmont, CA: Wadsworth,

~. 1986. Complex Organizations: A Criti- cal Analysis. New York: Random House.

Pfeffer, Jeffrey and Huseyin Lebtebici, 1973. "The Effect of Competition on Some Dimen- sions of Organizational Structure." Social Forces 52:268-79.

Pfeffer, Jeffrey and Gerald Salancik, 1978, The External Control of Organizations: A Resource Dependence Perspective. New York: Harper and Row.

Piore, Michael and Charles Sabel, 1984, The Sec- ond Industrial Divide. New York: Basic Books.

Pond, James and Robert Harvey, 1985, "Dream Factory Leap-Frog the Opposition." Iron Age, 15 Nov., pp. 26-36,

Prechel, Harland, 1990. "Steel and The State: In- dustry Politics and Business Policy Formation, 1940-1989." American Sociological Review 55:648-68,

. 1991a. "Conflict and Historical Varia- tion in Steel Capital-State Relations: The Emergence of State Structures and a More Prominent, Less Autonomous State." American Sociological Review 56:693-698,

-, 1991 b. "Irrationality and Contradiction in Organizational Change: Transformation in the Corporate Form of a U.S. Steel Corpora- lion," The Sociological Quarterly 32:423-55,

-. Forthcoming. "Corporate Restructur- ing of Information Gathering Systems." In Technology Policy and Social Innovation in the Firm, edited by J. Niosi. London, England: Pinter Press.

Raff, Daniel and Peter Temin, 1991, "Business History and Recent Economic Theory." Pp. 71-116 in Inside the Business Enterprise^ ed- ited by P, Temin. Chicago. IL: University of Chicago,

Sayer, Andrew and Richard Walker. 1992. The New Social Economy: Reworking the Division

of Labor. Cambridge. MA: Blackwell. Schumpeter, Joseph, 1950. Capitalism, Socialism

and Democracy. New York: Harper Torch- books,

Scott, Richard. 1981. Organizations: Rational, Natural, and Open Systems. Englewood Cliffs, NJ: Premice-Hall.

Shaiken, Harley, 1984. Work Transformed: Auto- mation and Labor in the Computer Age. New York: Holt, Reinhart, and Winston.

Silver, Marc L, 1982, "The Structure of Craft Work: The Construction Industry," Pp, 232-52 in Varieties of Work, edited by P, L. Stewart and M. G, Cantor, Beverly Hills, CA: Sage.

Simpson, Richard. 1985, "Social Control of Oc- cupations and Work," Annual Review of Soci- ology 11:415-36.

Simon, Herbert, 1957. Administrative Behavior. 2d ed. New York: Macmillian,

Smith, Vicki. 1990. Managing in the Corporate Interest: Control and Resistance in an Ameri- can Bank. Berkeley, CA: University of Califor- nia.

Sobel, Robert. 1984. The Age of Giant Corpora- tions: A Microeconomic History of American Business. 1914-1984. Westport, CT: Green,

Sombart, Werner, 1953, "Medieval and Modern Commercial Enterprise." Pp. 25-40 in Enter- prise and Secular Change, edited by F, C, Lane and J. Riemersma, Homewood, IL: Irwin,

Stinchcombe, Arthur, 1990, Information and Or- ganizations. Berkeley. CA: University of Cali- fornia.

Sweezy, Paul. [1942] 1970. The Theory of Capi- talist Development. New York: Monthly Re- view.

Taylor, Frederick. 11911] 1967, The Principles of Scientific Management. New York: W,W. Norton.

Thompson, James, 1967. Organizations in Action. New York: McGraw-Hill,

Ungson, Gerardo Rivera, Daniel Braustein, and Phillip Hall, 1981. "Managerial Information Processing." Administrative Science Quarterly 26:116-34.

U.S, Department of Commerce. 1965-1987. Quarterly Financial Report for Manufacturing, Mining, and Trade Corporations. Washington, DC: United States Bureau of the Census.

Van de Ven. Andrew and Diane Ferry, 1980. Measuring and Assessing Organizations. New York: John Wiley and Sons,

Waters, Malcolm, 1990. "interest and Procedural Norms in the Analysis of Collegiality." Ameri- can Journal of Sociology 96:192-99,

Weber, Max. 1946. From Max Weber: Essays in Sociology. Translated by H. Girth and C. W. Mills. New York: Oxford University.

, 11921] 1978, Economy and Society. Translated by G. Roth and C, Wittich. Berke-

CENTRALIZING CONTROL OVER THE MANAGERIAL PROCESS 745

ley. CA: University of California, Weitzel, William and Ellen Jonsson, 1989, "De-

cline in Organizations: A Literature Integration and Extension." Administrative Science Quar- terly 34.9 \-W9.

Williamson. Oliver and William Ouchi. 1983. "The Market Hierarchies Program of Re- search." Pp. 13-34 in Power, Efficiency and Institutions, edited by A, Francis, J. Turk, and P, Willman, London, England: Heinemann Educator Books.

Woodward, Joan, 1965. Industrial Organization: Theory and Practice. London, England: Oxford University,

Zeitz, Gerald. 1980. "Interorganizational Dialec- tics," Administrative Science Quarterly 25:72- 88,

Zetka. James R , Jr. 1991. "Automated Technolo- gies, institutional Environments, and Skilled Labor Processes: Toward an Institutional Theory of Automated Outcomes." The Socio- logical Quarterly 32:557-74,