Strategy and irreversibility in supplier relations: the case of the U.S. automobile industry.
Theoretically, a firm could act as a holding company for several such clusters, but this seems impossible in practice. First, there is no way for corporate management to commit itself not to intervene in the affairs of its subsidiaries, and it is hard to distinguish good from bad interventions a priori.(92) As the firm gets larger, top management resources are stretched thin, and the proportion of bad interventions rises. Second, as the firm grows, its parts become more insulated from outside pressure, and it is harder for Dramatic changes are under way in U.S. automakers' relationships with their suppliers.(1) Before these changes began in the 1980s, the automakers' dealings with outside suppliers had been characterized by short-term contracts (usually one-year), arms'-length relationships, and many (usually six to eight) suppliers per part. The fierce competition among outside suppliers for the automakers' continued patronage contrasted sharply with the comfortable position of the inside suppliers, who were usually guaranteed some or all of the business for an indefinite period. Nevertheless, there was little communication between the internal components divisions and either the central engineering groups who designed the parts produced by the divisions or the assembly plants they supplied.
Since 1980, however, the automakers have been moving (albeit in fits and starts) toward a very different supplier relations system. In the new system, only a few suppliers provide each type of part, and information is interchanged extensively between buyer and supplier. Contracts with outside suppliers are increasing in length (three- to five-year contracts are now common), and the automakers are reducing their commitment to their own components divisions, ending such practices as guaranteeing them business and in some cases divesting them completely.
These changes seem surprising on several counts. First, the automakers' willingness to give up monopsony power over their outside suppliers in favor of long-term, often sole-source contracts appears to defy economic logic. Second, a reduction in vertical integration at a time when both the technology and the market structure of the industry are in flux seems to contradict the predictions of organization theorists that vertical integration should rise with increasing uncertainty. Their reasoning is that vertical integration allows decisions to be made sequentially; as the state of nature is revealed, "internal adaptations can be made by fiat."(2) In contrast, dealing with a financially independent supplier requires either a contract that provides in advance for every contingency or costly post-contract haggling. As uncertainty rises, so does the number of contingencies and therefore also the expense of contracting.
Finally, the new supplier relations cannot be attributed to the discovery of some organizational form never before seen in the United States. In contrast to linear views of history, in which each change must represent a new, improved product, the recent changes move the U.S. auto industry back to a supplier relations system seen earlier in the twentieth century, when relationships with outside suppliers were close and vertical integration was rare.
Why are these changes in automakers' supplier relations occurring? More generally, what determines which type of supplier relations system is chosen by a firm? What are the consequences for the evolution of organizational and technical capability of buyers and suppliers in that industry?
In response to these questions, this article argues that a buyer firm with power in its final-product market can use that power to change the structure of its input markets.(3) In choosing its supplier relations strategy, however, the buyer faces a trade-off between promoting technical progress in the industry and maintaining buyer bargaining power.
Before the import challenge of the 1970s, three firms--General Motors (GM), Ford, and Chrysler--shared control of the U.S. auto industry; they were free to choose a supplier relations strategy that favored buyer bargaining power over technical change. With the entry of Japanese and European automakers producing high-quality, innovative products, however, U.S. automakers no longer were able to make profits by minimizing supplier bargaining power at the expense of quality and innovation. This analysis thus attributes the change in supplier relations not to the rectification of errors caused by past bad management, but rather to a change in the type of strategy that maximized profits.
In the article, I develop this interpretation using a conceptual framework based on Albert Hirschman's distinction between "exit" and "voice" methods of problem resolution.(4) The focus will be on explaining the first anomaly mentioned: why the automakers found it profitable to give up monopsony power over their outside suppliers. I will discuss the other two anomalies (the relationship between vertical integration and uncertainty, and the return to previous patterns of supplier relationships) insofar as they help to shed light on the first.
In the following section, I briefly describe the framework and discuss the implications of different supplier relations choices for 1) the buyer firm's bargaining power (measured by the percentage of buyer-supplier joint profits the buyer receives), and 2) the nature and rate of technical change in the industry.(5) The next section uses the framework to look at the evolution of supplier relations in the early days of the U.S. auto industry. I then present the results of interview-based research into the "exit-based" system of supplier relations prevalent in the postwar industry. Finally, I examine the situation in the early 1990s, in which some relationships are becoming "voice-based," whereas others are better characterized as a new form of exit.
Economic analysis of the vertical relationships between firms has typically focused on the "make-buy" problem.(6) It is assumed that a buyer firm has only two choices about how it obtains its components. It can make them in a wholly owned subsidiary, in which case the buyer is assumed to have perfect knowledge about and control over all aspects of the subsidiary's operation. Alternatively, it can "buy" them on an arms'-length market; in this case the customer knows only the price of the product and nothing about the firm or its production processes.
Contrary to this assumption, equity investment in a downstream process (that is, a decision to "make" rather than to "buy") is neither necessary nor sufficient for administrative coordination. For example, as Therese Flaherty points out, a financially independent buyer and supplier who have had long-term dealings may have closer administrative relations than would two divisions of the same holding company.(7) The key determinant of the nature of information exchanged by a buyer and supplier is not the existence of a financial connection, but rather the strength of the commitment between them. In what follows, I will briefly describe a new conceptual framework for supplier relations, one that is based on problem-solving mechanisms rather than on financial arrangements.(8)
Many problems can arise in a relationship between buyer and supplier. For example, suppose one party wants the other to undertake a specific action (lower its price, improve its quality), but the other party refuses, either because it lacks the capability (decision-makers in the firm do not know how to implement the proposed changes), or because it lacks the incentive (decision-makers have other, more profitable, courses of action open to them). Or, one party may perceive a problem (the buyer may find that its products are not selling as well as it would like) and may believe that changed behavior by the other party would contribute to a solution, but the aggrieved party is not able to propose a specific course of action.
In this article, I concentrate on responses to problems arising in a particular type of buyer-supplier relationship, one in which the buyer can make "take-it-or-leave-it" offers to suppliers. These offers include not only the price of the input, but also the length of the (implicit or explicit) contract and the nature of the task that the supplier is to perform (amount of design work or subassembly, for example). This assumption is reasonable for the U.S. auto industry; the automakers can use some of the power and profits derived from their favorable final-product market position to influence the structure of their input markets.
Borrowing Hirschman's terminology, we can identify two types of responses to problems arising in a buyer-supplier relationship: 1) exit, where the buyer's response to problems with a supplier is to find a new supplier, and 2) voice, where the buyer's response is to work with the original supplier until the problem is corrected.(9) Whereas the exit strategy secures compliance by use of the "stick" of threats to withdraw from the relationship, the voice strategy relies on the "carrot" of increased profits for both parties resulting from improved products.
The key to the exit strategy is making credible the buyer's threat to leave if its demands are not met. Therefore, the buyer must have access to many interchangeable suppliers or the ability to tool up quickly for in-house production. The key to the voice strategy is setting up a communication system that will allow the rich flow of information essential to the "let's work things out" approach. For example, a Stanford Business School case describes a contract between an automaker and Signetics (a division of N. V. Phillips) for high-quality semiconductors:
As part of the contract, the customer was prepared to supply capital for development and testing.... The arrangement would also involve both the customer and Signetics in identifying types of defects and their sources. This ... required a high level of teamwork: "our product problems" would be solved jointly.... The automotive parts manufacturer would have to participate in Signetics' decision-making process. |There would have to be~ open discussion of device characterization data, identification of critical parameters, and the operating environment of the device in the customer's system.(10)
Differences in supplier relations strategies can be summarized along two dimensions: the degree of administrative coordination and the nature of incentive systems used by the buyer and its suppliers. "Administrative coordination" describes the nature and amount of information that flows between producing units. At the lowest level of administrative coordination, only price information is exchanged, as in the "market" described in economics textbooks. At intermediate levels, other types of information flow across the boundaries of producing units, such as financial data or characteristics of plant and equipment. At the highest levels of administrative coordination, there is a continuous flow of information, and parties no longer act as order-giver and order-receiver; rather, both parties adjust as a result of the two-way transfer of information.
"Incentive systems" are the mechanisms (both explicit and implicit) used by a buyer and a supplier to elicit desired behavior from the other party. An important incentive is the granting or withholding of commitments to undertake particular actions in the future.(11) A supplier has a commitment from a buyer when the supplier knows that the buyer will continue to purchase its products for some length of time. This assurance can be provided in several ways, including equity investment or long-term loans, long-term contracts, and parties' concern for their reputations for fair dealing. Commitment can also be provided involuntarily, as when a buyer faces an oligopolistic supplier industry. If a firm can obtain an input from only a few vendors, the firm's ability to exit from relationships with them is very weak.
The voice and exit methods of problem resolution outlined here each requires a different combination of administrative coordination and commitment. Figure 1 shows four regions of supplier relations strategies.
To implement the voice-based strategy, the buyer needs a high degree of administrative coordination with its suppliers, which requires the buyer to make a strong commitment to only a few suppliers. This high administrative coordination leads to high costs of switching suppliers resulting from the high cost of maintaining extensive communication systems with more than one supplier, the learning effects as the parties get to know each other, and the need for trust when exchanging proprietary information.
In contrast, an exit-based strategy requires low commitment in order to maintain the credibility of the buyer's threat to leave. Therefore, administrative coordination must also be low. Since high administrative coordination requires high commitment, the high administrative coordination/low commitment area of Figure 1 is marked "infeasible." Conversely, the low administrative coordination/high commitment region is called "stagnant," because the buyer has few tools available for problem resolution; therefore, it has neither the threat of exit nor the channels through which to exercise voice.
The buyer's choice of method for problem resolution is an important one, because it affects both the buyer's and the supplier's relative bargaining power and their propensity to introduce new technologies of various types.
The exit strategy gives the buyer a great deal of bargaining power, because it requires little commitment to any one supplier. Conversely, the voice strategy reduces the buyer's bargaining power by increasing the cost of switching between suppliers.(12) In contrast, where significant investment is required, either to communicate about a technical problem or to implement a solution to that problem, voice is likely to be superior. Without the detailed information and long-term commitment characteristic of a voice relationship, a supplier's innovation may well be inapplicable to the customer's needs. Even if it is applicable, a supplier without some type of market power will face problems in raising enough capital to introduce the innovation and in appropriating rents from it.(13)
I have argued that the exit strategy maximizes buyer bargaining power, whereas the voice strategy maximizes the rates of most types of technical change.(14) Therefore, a trade-off exists between buyer bargaining power and industry technical change.(15) This application of exit/voice analysis allows us to integrate Marxist analyses (which usually focus on the zero-sum aspects of economic relationships) and management theory (which often emphasizes positive-sum outcomes). What determines which point on this trade-off is chosen at a particular juncture?
The driving forces in this model are the concepts of "strategy" and "irreversibility." Strategy enters because firms with power in their final-product market can use that power to change the structure of their input markets. Such firms can act not only to minimize the cost of inputs on a given cost curve, but also to affect the location of the cost curve itself by altering barriers to entry in their supplier industries.(16) By their choice of supplier relations strategy, buyer firms affect not only the prices they pay for inputs, but also the potential for technical progress by suppliers and by the downstream firms themselves. A buyer's strategy is irreversible in the short run, however, because it leads both supplier and buyer firms to develop particular capabilities and expectations, in a manner that tends to be self-reinforcing (see Figure 2 for a schematic summary of this argument). That is, history matters. Whereas the expectations and capabilities generated by a given supplier relations system tend to be self-reinforcing, the changes in final-product market structure resulting from that supplier relations system may be de-stabilizing. Under certain conditions, a cycle of supplier relations systems could be generated in which each system carried the seeds of its own destruction.(17)
First, a firm would gain final-product market power at least in part through its ability to manage a voice supplier relations system. However, because of suppliers high bargaining power in a voice-based system, the firm would find it profit-maximizing to switch to the exit system in order to appropriate the return to its favorable final-product market position. Although this exit system would increase barriers to entry for domestic firms by limiting their ability to purchase engineering and coordination services on the market, it would leave the incumbent vulnerable to challenge by an entrant able to establish a voice system (perhaps behind tariff walls in a foreign country). The incumbent would have either to switch back to voice or to exit from the industry. In either case, the new dominant firm(s) would practice voice-based supplier relations. These firms might then try to increase their bargaining power with suppliers by switching to exit, and the cycle would begin again.
A necessary condition for such a cycle is that either the incumbent's discount rate be high or the entrants' market penetration be slow. Either of these conditions ensures that the value of the incumbent's profits during the exit period exceeds the later losses caused by failure to blockade entry. The scenario of oscillation between exit and voice supplier relations seems to fit the history of the U.S. auto industry fairly well; the Big Three enjoyed fifty years of domestic oligopoly, and American managers in general have been castigated for their short time horizons.(18)
In the earliest days of U.S. auto production (before 1909), it was auto assemblers who were dependent on their suppliers. Since supplier firms had grown up primarily to serve older industries such as bicycles and carriages, they were more established than their automotive customers. These suppliers served as an important source of working capital for the fledgling automakers, because assembly and sale frequently required less time than the 30- to 90-day credit provided by the supplier. In 1903, Ford's entire contribution to the manufacture of the cars that bore the company's name was to install bodies, wheels, and tires on completed chassis made by the Dodge brothers, owners of a Detroit machine shop.(19)
Suppliers also were an important source of technical change in the industry. For example, in 1901, entirely as a result of greater precision in machining, Henry Leland was able to increase the horsepower of the Olds curved-dash runabout engine from 3.0 to 3.7. It is not clear how much information was exchanged between supplier and customer, or how much exchange was necessary to achieve improvements over existing practice in such a young industry. Some inventions were originally aimed at other industries. Hyatt invented the flexible roller bearing in an attempt to perfect a sugar cane-grinding machine; Timken invented the tapered roller bearing in 1899 to reduce the need for lubrication of carriages and wagons.(20)
In this early period, boundaries between supplier and assembler were fluid. For example, George Holley gained a "considerable reputation" with his one-cylinder car, the Holley Motorette, but he later came to specialize in carburetors. Conversely, when further improvements in his engine yielded 10.25 horsepower in 1902, and Olds refused to accept them because they would have required radical change in the car's design, Leland formed the Cadillac Automobile Company to assemble cars himself. A decade later, the Dodge brothers began planning to use their Ford dividends to finance entry into auto assembly.(21)
As automakers became more stable, they offered their suppliers both more commitment and more administrative coordination. Around 1903, the young Alfred Sloan, then president of Hyatt Roller Bearing Company, was summoned by Henry Leland, the W. Edwards Deming of his day, and castigated for not continually striving for the tightest tolerances possible. "You must grind your parts. Even though you make thousands, the first and the last should be precisely alike," Leland said. As a result of their conversation, Sloan gained "a genuine conception of what mass production should really mean.... I was determined to be as fanatical as he in obtaining precision in our work."(22)
At Ford during 1909-14, "the Company was not then averse to purchasing virtually all of its materials and parts from independent producers." The automaker shared its growing management expertise with suppliers:
The Ford Motor Company purchased materials for its componentsmakers, reorganized their manufacturing processes, supervised their larger policies, and, in some cases, aided them in financing production. The Company became so dependent upon the production of its specialized suppliers that its own operations were frequently within thirty minutes of suspension because of tardy deliveries of parts or materials.(23)
In a similar effort to assure supply, William C. Durant at General Motors persuaded parts makers such as Weston-Mott in Utica and Alfred Champion in Boston to move their operations close to his in Flint. Smaller assemblers were even more tightly linked with their suppliers; their designs were so specialized that the bankruptcy of one firm often meant the bankruptcy of the other as well.(24)
With the rapid rise to dominance of Ford and (GM, however, this voice arrangement with financially independent suppliers soon proved unstable. Parts makers found themselves holding large investments specific to one or two customers. For example, Sloan writes proudly in his autobiography of the significant investments Hyatt had made by 1916: an 80 percent expansion of floor area (to 750,000 square feet) in three years; three private railroad sidings; its own fire department; a staff of chemists and metallurgists, who saw to it that "every step in the development of raw material into antifriction bearings was checked by scientific methods."(25)
But this investment was highly vulnerable; as Sloan wrote, "One dismal fact was revealed by our accounting: More than half our business came from Ford, and our other big customer, General Motors, dwarfed the remainder. If either Ford or General Motors should start making their own bearings or use some other type of bearings, our company would be in a desperate situation." So, when William Durant invited Sloan to lunch to discuss a proposal to buy him out, Sloan was ready to listen.(26)
Suppliers who did not make such specific investments came to be high-cost producers. As Ford gained capital, he gradually began to make parts that previously he had bought; his company "always made them cheaper than the former makers," because Ford constructed machinery "to do just that one job, whereas the outside manufacturers had to consider other products with the same machine."(27)
For the automakers, the combination of fast growth and industry consolidation gave them both the desire and the wherewithal to use voice within their organizations while moving toward exit relationships with outside suppliers. Fast growth left both Ford and GM short of trained managers. One important source of managers was the parts industry, which was more established and better managed.(28) Sloan, Charles Kettering, the Fisher brothers, and S. L. Mott, who all became important to the success of GM, joined the firm when their parts-making businesses were bought out by Durant or the du Ponts.
At Ford, the purchase of the Keim Mills 1911 brought three key people into the company: William S. Knudsen, who directed the expansion of the assembly operation; W. H. Smith, the most important of several Keim engineers who made numerous advances in adapting machine tools for mass production; and John R. Lee, who became head of the Sociological Department. Before the purchase, "cooperation developed along a broad front," as Ford invested heavily in Keim machinery, and "Ford men spent much time in Buffalo" at the Keim plant.(29) Together the companies developed a process for pressing steel into such parts as axle housings and crankcases. After a strike in 1912, Ford moved the machinery and key managers to highland Park and integrated them into Ford's day-to-day operations, where their experience and interaction with Ford executives contributed greatly to the development of mass production.
This emphasis on vertical integration's role in facilitating administrative coordination by bringing skilled people into the company is somewhat different from other accounts. For example, one well-known discussion of GM's 1924 acquisition of Fisher Body focuses on the problem of asset specificity--Fisher's alleged refusal to build its stamping plants next to GM assembly plants. In contrast, Alfred Chandler and Stephen Salsbury cite GM's desire to gain access to the Fishers' management skills as the reason for GM's increase of its ownership of Fisher Body from 60 to 100 percent in 1924; they do not mention specificity of physical assets as a consideration. (It would seem that 60 percent ownership should have given GM sufficient power to force Fisher to locate its plants in places convenient to GM.)(30)
My focus is also somewhat different from that of Richard Langlois and Paul Robertson, who argue that vertical integration is important for process innovation because it lowers the cost of communication, and that the key benefit of Ford's acquisition of Keim was that "because much of the assembly activity was collected in one place, Ford engineers were able to perceive opportunities for grasping economies of scale that would not have been apparent to decentralized suppliers.... |Keim's stamping~ technology sparked the interest of Ford engineers, and metal-stamping was soon adapted to producing crankcases, axles, housings, and even bodies. "(31)
In focusing on the importance of vertical integration in providing opportunities for the extensive informal communication necessary for working out the logic of mass production, Langlois and Robertson have made an important contribution. In their account, however, vertical integration creates this opportunity by providing geographic proximity and new technology. I would argue, first, that vertical integration is neither necessary nor sufficient for proximity (after all, the Keim operation was not moved to Detroit immediately after Ford's purchase); and second, that "the modern machine tools from the Keim works were no more valuable than the inventive skills of the experienced leaders who came over from Buffalo."(32) That is, it was not so much that "technology)" sparked the interest of Ford engineers, but rather that integrating Keim personnel into Ford's management provided valuable opportunities for intellectual cross-fertilization. Finally, not all process innovation benefits from vertical integration. Although centralization was important for working out the logic of mass production, decentralization seems more conducive to recent process innovations such as statistical process control.
At first, vertical integration promoted the use of voice to solve problems in the auto industry. In the 1920s and 1930s, however, as Langlois and Robertson document, vertical integration proved less efficacious as a result of outside suppliers' excess capacity and of consumers' demand for product innovation. Financially independent suppliers developed many new products, such as the closed steel body, and companies with less vertical integration, such as Chrysler, were best able to incorporate these products into their cars.(33) In contrast, Ford in 1927 had to shut down its operations for nine months in order to change over from the Model T to the Model A. Whereas a systemic rearrangement of the production process benefits from vertical integration, product innovation may benefit from decentralization, which gives access to a multiplicity of ideas.(34) By the 1940s, vertical integration at Ford had reached levels that clearly reduced corporate efficiency (the company's far-flung empire included rubber plantations in Brazil and vast tracts of forests); Henry Ford's successors sold off many of these assets, leaving GM as the most integrated producer.
Besides vertical integration, the other piece of the automakers' supply strategy was the establishment of exit relationships with financially independent suppliers. Industry consolidation meant that Ford and GM were able to engage in supplier relations practices that improved their bargaining position even at some cost to efficiency. Over the next several decades, U.S. automakers created a fiercely competitive components industry. They did this by reducing barriers to entry by taking complex functions like engineering and R&D almost completely in-house and by dividing components into small, easy-to-produce pieces and hiring managers to coordinate the assembly of these parts centrally. They also employed several (six to eight) competing suppliers for each part, offered only short-term (one-year) contracts, and required suppliers to license major innovations.
The history of this marginalization of outside suppliers has yet to be written. It seems to have been a gradual process. In the 1920s, U.S. automakers refused to use aluminum in significant quantities because it meant doing business with Alcoa, a single-source supplier. This was true even though European manufacturers used a large number of aluminum parts to save weight, and though Henry Ford "was constantly trying to keep down weight" in the design of the Model A, according to one of his chief engineers.(35)
However, Kenneth Alexander reports sole-sourcing of products such as forgings and castings until the early 1950s, when a desire to avoid (among other things) work stoppages that shut down assembly plants brought a "more rigid adherence to the multisource practice."(36) Where capital intensity made multiple sourcing prohibitively expensive, either the supplier was asked to retain slack capacity or the part was vertically integrated. For example, Ford and Chrysler began in-house production of car bodies during this period. The automakers also stepped up price pressure; all of them disallowed increased prices to cover higher wages resulting from "annual improvement factors" negotiated into labor contracts, and GM began to visit suppliers to demand a share of the benefits of cost reductions made after the original bid.(37)
This dual system of vertical integration and exit-based relations with outside suppliers helped to protect U.S. automakers against successful entry by domestic producers, but it left them vulnerable to entry by the Japanese in the 1970s. By the late 1980s, the Big Three were trying to return to a more voice-based system. The next two sections explore this recent shift in supplier relations strategy.
The Exit-Based System of Supplier Relations
Earlier, "exit" and "voice" were discussed as pure strategies employed in isolation from each other. The reality, of course, is more complicated. In this section, I discuss the system of supplier relations used in the U.S. auto industry from roughly the 1950s to the 1970s.(38) This system was exit-based in the sense that most supplier relationships were located in the low administrative coordination/low commitment area of Figure 1. The relationships within this exit-based system fell into three main modes, which I will call 1) simple exit, 2) voice with cheating, and 3) financial integration.(39) (Figure 3 shows where each fits on the map presented in Figure 1.)
Simple Exit * In this mode, each part was subject to annual competitive bidding, on detailed blueprints provided by the automaker, in which the winner was in normal circumstances the low bidder. There was almost no administrative coordination and almost no use of feedback in the sense defined in the previous section.(40) In the 1970s, this mode accounted for about 99 percent by number of firms and about two-thirds by dollar volume of the automakers' outside sourcing.(41) Simple exit was used primarily in relations with suppliers of easy-to-make parts (such as small stampings, metal finishing, plastic knobs, clips, fasteners, and wire harnesses), where superior performance was not perceived to be a source of competitive advantage. In addition, the automakers obtained some of their requirements for more complex parts on a simple exit basis to keep down the prices of suppliers in the other two modes (who usually had a somewhat greater role in designing their components).(42)
Voice with Cheating * In this arrangement, suppliers who offered engineering excellence or some other form of superior service received implicit long-term commitments.(43) With some of these suppliers, mechanisms for administrative coordination developed as well. A Bendix executive described his firm's long-time relationship with Ford engineers:
We talk to those guys every day--keep each other up on new developments--it's all very informal. They'll say, "We're coming up with a new model, and we would like to put your Series 100 brake in it. However, we'd like you to work on it a little bit...." And we'll try to convince them that our Series 100--or something we can make on the same equipment, is really just what they need, so as to preserve our economies of scale. But, since they're the customer, we can't push too hard.(44)
Voice-based relationships were not explicitly nurtured by the official purchasing policies. Instead, these relationships were maintained because of personal trust between individuals at the two firms or because of a firm's status as a traditional supplier. Over the decades that these relationships endured, fairly complex norms of reciprocity (often with asymmetric rights and responsibilities reflecting differential bargaining power) arose, as buyers and suppliers attempted to forge continuing relationships in an atmosphere characterized by both intense conflict and mutual dependence.
However, "cheating," or reneging on these implicit contracts, was not unknown.(45) In my interviews, for example, I found widespread agreement by both supplier and automaker executives that if an automaker asked a supplier to build a plant to supply a particular input, the automaker had an obligation to provide enough business (either of that input or of a substitute) to keep the plant full. In one such instance, a supplier built a plant to make a highly automaker-specific part "on a handshake." In the early 1960s, however, a supplier built a plant to make radios for Ford. Within two years, Ford took the work in-house, since it had acquired a subsidiary (Philco) with that capability. Although Ford gave the supplier some substitute work, it was insufficient, and the supplier was still trying to figure out what to do with the plant in October 1984. In a more recent case, after the same supplier had spent a good deal of money and developed a major innovation on a product that it had supplied to Ford for years, Ford made use of the new design but awarded almost all of the business to a Brazilian company with lower labor costs.(46)
The voice-with-cheating mode accounted for about 1 percent by number of firms and one-third by volume of automaker purchases. This mode was more prevalent on technically demanding parts such as brakes and steering wheels; the firms involved tended to be large, technically capable suppliers such as Bendix, Rockwell, and TRW.(47)
Financial Integration * This mode was characterized by very high commitment, expressed through 100 percent equity ownership with very few divestitures or closures during the 1950-80 period. However, administrative coordination was not necessarily higher for internal components divisions than it was for outside suppliers in the voice-with-cheating mode. A Chrysler engineering executive said that the major determinant of trust was not whether a supplier was inside or outside, but rather how long Engineering had been working with a particular supplier on a particular product. In longer term relationships, confidence in the other party's abilities and trustworthiness had time to grow, and people learned whom to contact if there were difficulties. With both inside and outside suppliers, most of the engineering of a product was done centrally, "in almost splendid isolation" from the engineering of the process to be used to make that product. An additional step of manufacturing engineering then had to be performed in order to suit the product to the supplier's manufacturing facility.(48)
The level of commitment to inside suppliers was extremely high. In the early days of the industry, acquisition of components firms permitted the integration of skilled managers into the automaker's organizations, but by the 1950s key positions in the divisions were often staffed, as former GM executive John DeLorean put it, by "the logic of the unobvious choice. This means promoting someone who was not regarded as a contender for the post....|The practice~ earns for you his undying loyalty because he owes his corporate life to you."(49)
The result of the Big Three's insulation from competition in final-product markets was a stagnant relationship with in-house component suppliers. Among the most formidable of the in-house suppliers was GM's Fisher Body Division, which by the 1980s "had become so powerful that they could literally change the direction of a vehicle program at whim.(50) In the 1970s, the division would submit a single cost estimate for the entire body (including trim), refusing to break down the figure beyond general estimates for materials, overhead, and labor. In one incident, a Fisher representative resolved a design dispute by saying, "If you want this car to have some doors, you'll do it our way."(51)
At Chevrolet, the most experienced buyers were assigned to products purchased from allied divisions. Since these divisions usually refused to provide cost data, buyers had to establish their negotiating positions by comparing costs from previous years or by soliciting outside quotes. Only if the cost differential were large and the purchaser prepared to fight would the business be awarded to the outside supplier.(52)
The degree of financial integration varied a great deal across automakers. General Motors produced 60-70 percent of its value-added in-house in 1980; Chrysler produced 30-35 percent, and Ford about 50 percent.(53) Between the 1950s and the 1970s, all three automakers made all of the parts commonly defined as key to the essence of an automobile (engine, transmission, axles) for each of their models. One important motivation for vertical integration was to achieve an "assured supply" of key components.(54)
Under the exit system of supplier relations, U.S. automakers were quite successful in preventing their vendors from sharing in the rents from the makers' final-product market oligopoly. Using Internal Revenue Service and annual report data, Robert Crandall found that peacetime rates of return for 1936-61 for the seven major assemblers averaged 23.8 percent and were greater by at least four percentage points in every year than those of their suppliers, who averaged only 8.2 percent. Using data from the First National City Bank Monthly Letter for 1947-65, he showed that the average rate of return for auto and truck assemblers of 20.2 percent was significantly greater than the average 13.2 percent rate of return for all manufacturing, whereas the average 13.8 percent return for automotive equipment producers was not.(55) More recent data show that publicly traded suppliers to automotive assemblers averaged a 12 percent return on equity for 1972-82, well below (and far more cyclical than) the Standard and Poor 500's average of 15.5 percent.(56)
The assemblers' success appears all the more remarkable in view of certain conditions in the auto industry at the time. One might have predicted that factors such as the moderately high technological requirements for producing an automobile and the technical feasibility of achieving economies of scale through mechanization would have led to above-average supplier profits by creating barriers to entry. One might also have predicted that the combination of the automakers' high fixed costs and highly uncertain product specifications and the concomitant necessity for post-contractual engineering changes would have left the automakers vulnerable to threats by suppliers to refuse contract modifications unless they were paid handsomely for them.(57)
How did the automakers manage to overcome these conditions? The source of their success lay in the ability to maintain a credible threat of exit, which they did by using rents gained from power in their final-product market to transform the structure of their input markets. The automakers made huge investments in engineering and management staffs whose work simplified the task of being an auto supplier, thereby reducing the barriers to entry into supplier industries.(58) This task simplification took two main forms: 1) automaker-provided blueprints, obviating the need for design capability for firms in the simple exit mode; and 2) automaker coordination of subassembly, in which each supplier provided only a very small piece of the automobile, so that each firm had to procure, manage inventory for, and assemble only a few parts.
Both suppliers and buyers agreed that one purpose of the division of labor was to prevent suppliers from developing expertise, because "they |automakers~ don't want to lose control," as one supplier put it. A Ford manufacturing executive, in commenting on an earlier draft of this article, wrote that the statement had been "true in the past" and added that the automakers' "motivation |was~ more driven by profit ... than by tyrannical control."(59)
The automakers were most successful in maintaining a threat of exit with suppliers in the simple exit mode. Often six to eight firms would supply a single part, and dozens would supply a general class of component. For example, until the early 1980s, Ford had twenty-seven wiring-harness suppliers.(60)
David Teece has emphasized that one explanation for vertical integration is that financial control of complementary assets allows an entrepreneur to appropriate the benefits of innovation.(61) The automakers' success in protecting their oligopoly rents points to another way of ensuring that one appropriates the benefits of an innovation: by making sure that complementary assets are provided on a competitive market. Control can be exercised through markets as well as through hierarchies.
However, a supplier relations strategy utilizing the simple exit mode alone would have suffered from a severe lack of technical change. Therefore, since the leading U.S. automakers did not totally abandon technology-based competition, they found it advantageous to establish more voice-based relationships with some suppliers. This voice-with-cheating mode involved the automakers in a complex balancing act, since they wanted to maintain high bargaining power while still taking advantage of suppliers' expertise. Suppliers knew that the only way to make economic profits was to invest in new products and processes that the automakers could not monitor so well, yet they feared that the automakers would take the new technology in-house before the suppliers had had a chance to recoup their investment.
Financially integrated divisions served largely as a source of assured supply.(62) By the 1950s, it became difficult to compete without this assurance; it has been argued that the automakers' acquisition of independent suppliers was an important factor in the demise of smaller automakers such as Kaiser. One factor leading to Chrysler's near-bankruptcy in 1980 was its inability to make more than 350,000 of its popular Omni Horizons because the company had contracted with Volkswagen to supply that number of engines, and VW refused to supply any extras.(63)
The major drawback to this supplier relations system from the consumer's point of view was that it provided disincentives to innovation. These disincentives were especially severe for firms in the simple exit mode, where intense competition led to supplier industries populated by tiny firms lacking in both organizational and physical capital. Researchers have described the bidding process in the tool and die industry: "Many customers will get five quotations for a job. Invariably, at least one quotation is too low because the manager made errors in estimating the job, or doesn't understand his costs, and he gets the work." The result was firms so small that diffusion of capital-intensive technologies was sometimes blocked; for example, a 1967 study found that the cost of a numerical control die mill was several times the total invested capital of a typical body die manufacturer.(64) Similar undercapitalization has been reported in the metal finishing industry. One firm, for example, had only one chemist; when this person was on vacation, chemical costs regularly jumped 20-25 percent, and quality plummeted.(65)
Similar though less severe problems with automaker lack of commitment plagued the voice-with-cheating mode. The introduction of electronic fuel injection in the United States was delayed for over a decade because of Bendix's fears that the automakers would integrate into the technology before Bendix could recoup its investment in developing it.(66) Similarly, the necessity of competing with low-overhead firms in the simple exit mode who were low-cost producers of existing designs reduced the ability of other firms to invest in improving designs, especially given problems with keeping such designs proprietary.
In addition, lack of administrative coordination severely hindered quality control in all three types of exit-based relations. One supplier executive whom I interviewed made weatherstripping, a product whose fit with the automakers' sheet steel was critical for preventing leaks. Instead of investing in administrative coordination--that is, learning enough about the characteristics of each other's processes and products to enable them to work together to find a design in which the parts fit well and that each could produce with consistently high conformance to specification--the parties fought over who was to blame for quality problems and made ad hoc adjustments. These adjustments meant that the supplier was producing a slightly different product for each of the automaker's assembly plants to which it sold, thus adding to the supplier's cost; sometimes these adjustments (made at the behest of the automaker) sacrificed watertight fit for speed of installation.
The record of internal suppliers was mixed. For capital-intensive innovation, it was quite good. The automakers were leaders in factory automation, as practiced, for example, at the Cleveland engine plant.(67) The high degree of commitment from the automakers gave internal suppliers both access to resources and the ability to appropriate returns from their investments.
In part because of the automakers' final-product market power, however, commitment became excessive, and stagnation set in. Risk avoidance was also in part a profit-maximizing strategy for dominant firms. The tone-setting incident for the automakers' policy toward radical innovation was the failure of GM's attempt to introduce the copper-cooled engine in 1922. The effort failed largely because the product was forced on the Chevrolet division in violation of GM's policy of divisional autonomy. However, the lesson that Alfred Sloan drew from this incident was quite broad: "it was not necessary to lead in design or run the risk of untried experiment."(68) According to one commentator, "While the policy spawned by this technological failure helped to protect the operating divisions from the uncertainties of technological change, it also isolated mass-produced cars from the influence of advanced technology."(69)
The Voice System of Supplier Relations
About 1980, the automakers began a fundamental reexamination of their supplier relations systems. The main impetus was their rapid loss of market share to foreign automakers, stemmed only by voluntary restraint agreements; since the late 1970s, the Japanese market share in the United States had fluctuated between 25 and 30 percent. In the early 1980s, several widely publicized studies attributed 20-30 percent of the Japanese cost advantage of $1,500-$2,000 per car to the voice-based Japanese supplier relations system. In addition, these studies found, the Japanese system led to far higher levels of quality and faster new product introductions.(70) In part as a result of the nature of Japanese and European entry, and in part because of exogenous advances in industries such as electronics, technology became an important factor in competition. As a Ford manufacturing executive put it, "The thing which is driving us to new relations with our suppliers is the desire for more coordination which will lead to more innovation.... We must find ways to use the capabilities of our suppliers, which we have ignored in the past because of NIH |Not-Invented-Here syndrome~."(71)
As a result, the automakers' official policy toward suppliers has undergone a significant change, from the exit-based strategy described earlier toward voice-based strategy. The Ford executive quoted above went on to say: "We recognize that our suppliers are not interchangeable. It requires some commitment on our part for them to develop state-of-the-art manufacturing capability; they have to amortize their investments. We see suppliers as extensions of our company; once you're in bed with them, you're there."(72)
There has been a great deal of experimentation, but it seems fair to categorize supplier relationships in the new voice system into two modes: reputation-based voice and financially based voice. Although the rhetoric has shifted to emphasizing cooperation, the reality is more complicated; a new form of exit is also alive and well in the new system. This section will discuss the nature of the new voice system, the problems of transition from exit to voice, and the reasons for the persistence of exit.
Voice Relationships Reborn * In reputation-based voice, administrative coordination with outside suppliers was far higher than under the voice-with-cheating mode of the traditional system. Suppliers in this mode were involved in a project from the early stages of design, and continuing contact with the automaker was maintained to work out quality problems and to practice just-in-time inventory techniques.
At Chrysler under the "Integrated System Supplier" (ISS) program, for example, a supplier for an instrument panel was chosen before a single drawing was made. The ISS firm was to design the instrument panel and to negotiate contracts with firms supplying plastic moldings and the various gauges. At first, the subcontractors and some of their materials were specified by Chrysler, but in later years this control was to be relaxed. The ISS firm had to make sure the parts met quality requirements, assemble the parts, and deliver the completed instrument panel as one module to the Chrysler assembly plant.(73)
In a similar move, Ford has put TRW in charge of safety systems on Ford cars for the indefinite future. The rationale is that because safety devices represent a complex technology outside Ford's core expertise, it is best to leave responsibility for them to a specialist. However, since there is a great deal of interaction between safety systems and the rest of the car (for example, warning lights and air-bag activation devices must be integrated into the car's electrical system), Ford and TRW engineers must be in close communication.(74)
The high level of administrative coordination characteristic of this mode of supplier relationship is sustained by long-term contracts. Whereas before 1980, "if a division wanted to go beyond a year with a supplier contract, it practically had to get dispensation from the Pope," five years later longer agreements were quite common.(75) In 1984, 30 percent of GM's contracts with outside suppliers were multi-year; in 1985, 70 percent of Ford's were.(76)
Since these contracts are usually incomplete, the parties' concern for their reputations for fair dealing is also important.(77) There is a great deal of room for judgment calls. In the GM-Signetics contract, for example, major uncertainties about levels and allocations of costs remained, and Signetics management thought that one or two such issues would arise every year. According to a Signetics document written after the contract was signed, "we will request that the customer pay the full $2.4 million investment" in test equipment needed to carry out the contract, and "it is hoped that the customer would agree to pay a significant portion" of the $1 million first-year expenses for additional quality control personnel.(78)
The financially based voice mode is characterized by high administrative coordination among inside suppliers. Instead of the old sequential product development chain, the automakers have set up committees with representatives from different functional areas such as purchasing, manufacturing, quality control, finance, and engineering. One benefit of such communication is the reduced chance that engineers will design unmanufacturable products.
Whereas in the reputation-based voice mode, the automakers sought to increase administrative coordination by increasing commitment to their outside suppliers, in financially based voice in the mid-1980s they sought to increase coordination by decreasing commitment to their inside suppliers.
This decreased commitment took various forms. In many cases, the automakers dropped full capacity utilization as a goal and began to require in-house divisions to compete for business just as outside suppliers would. Chrysler went the farthest in this direction. In 1987, ending a transition period begun in 1985, Chrysler grouped all of its "non-core" parts plants into an independent subsidiary named Acustar, from which Chrysler has no obligation to buy and which is free to sell to outside customers. Chrysler then attempted to sell Acustar, but backed off because of United Auto Worker (UAW) objections; instead, the company negotiated a reduction in job classifications and work rules. Chrysler will, however, close three of Acustar's twenty-nine plants and sell one.(79)
General Motors closed several components plants between 1982 and 1987. In 1986, GM began its "red light, green light, yellow light" study, in which it did a competitive analysis of every product made by its components divisions. Operations that made money and had "world-class" quality were given a green light; unprofitable operations (amounting to 10 percent of sales) were put on "red light" status and set for closing or sale; and borderline operations were given a "yellow light" and a chance to improve.(80) As at Chrysler, implementation of these plans was postponed because of objections from component-division executives and because of Attachment C of the UAW's 1987 contracts with each of the three U.S. automakers, which gave captive parts plants until September 1990 to become competitive with their outside rivals.(81) Few components plants were actually closed until the GM board's revolt in 1992.
Another indicator of changed direction is Ford and Chrysler's divestiture in functional areas; each made 40 to 45 percent cuts in its white-collar work force in the early 1980s. Although many of these cuts simply reduced organizational slack, others reflected increased outsourcing of key functions. At Chrysler, for example, the purchase of outside engineering services grew threefold between mid-1984 and mid-1986 and was planned to increase by the same amount or more between 1986 and 1990.(82)
The effects of the new, voice-based supplier relations system on technical change can already be seen. The use of interfunctional committees, although plagued in some cases by committee members' lack of authority, has reduced product development lead times, thus increasing the speed with which new technologies can be incorporated into new products. Simply convincing the finance and engineering departments at GM's Chevrolet-Pontiac-Canada division to review designs at the same time, rather than sequentially, reduced lead time by fourteen weeks in 1988.(83)
At Chrysler, executives at Acustar said that the parent corporation's reduction in commitment had had a salutary impact on their incentives "to seek out new ways of doing things, because we know that Chrysler does have alternatives," as one of them put it.(84) One example of an initiative that would not have occurred under the old arrangement, he said, was a joint-venture negotiation then under way with a Canadian firm to use flame lamination to bond fabrics, an alternative to the traditional, labor-intensive process of sewing them together.
At a division owned by another automaker, the reduction in commitment has led managers to seek to use the opportunities for efficiency provided by a common career path. When the corporation dragged its feet on setting up mechanisms for administrative coordination, division managers saw to it that liaisons with experience in the division were stationed both at assembly plants and at the design center to improve communication in case of a problem. As a result of this and other measures, the division's quality dramatically improved in recent years; it has gone from being rated one of the worst suppliers to the corporation's joint venture with the Japanese to one of the best.(85)
Conversely, increased commitment to outside suppliers has led to increased adoption of just-in-time inventory techniques and to increased R&D spending. For example, seat manufacturer Lear Siegler agreed to build an R&D center in return for a five-year contract.(86)
Elsewhere, I have provided statistical evidence of the impact of voice relationships on innovation, based on survey research done in 1989.(87) Contract length and trust in customer emerge as two key determinants of supplier adoption of computer numerically controlled (CNC) machine tools, even after controlling for such factors as the skill of the supplier in making new process introductions. Similarly, suppliers who meet more frequently with their customers are able to carry less inventory.
It seems likely, however, that the increased commitment required for a voice relationship will ultimately reduce the automakers' bargaining power. In the short run, suppliers will invest heavily to be in a good position to win multi-year contracts; but in the long run, the automakers will face significant switching costs, because they are dismantling many of the structures that kept barriers to entry low in their supplier industries. First, they are drastically reducing their supplier base; Ford cut the number of its suppliers from 5,000 in 1982 to 2,300 in July 1987.(88) Second, suppliers are now expected to participate in the design of products and to coordinate their subassembly; automakers are allowing some of their engineering capability to "atrophy," as a Ford executive put it, in order to reduce fixed costs. In addition, suppliers are expected to adopt statistical process control and just-in-time inventory techniques, both of which are manager-intensive innovations. According to a 1983 report,
The increasing emphasis on quality expected |by the automakers~ is likely to foster further the process of industry consolidation ... |since it~ often necessitates supplier retooling at a time of overcapacity in the industry.... With the financial strength and management expertise to utilize advanced storage and retrieval systems, companies such as Dana, Federal-Mogul, and Eaton will be in a position to gain market share on the strength of their ability to time accurately the delivery of high-quality products.(89)
The preceding analysis has focused on the automakers' vulnerability to suppliers. In contrast, popular perceptions of the industry are that automakers are powerful and suppliers are weak. I do not argue that the popular perception is wrong. Rather, I argue that the automakers' power over their suppliers is due not to some exogenous characteristic of the industry, but rather to a particular strategy adopted by Detroit's Big Three. The auto companies, with their high fixed costs (resulting from their huge investments in plant, equipment, and organization) and moderate-to-high engineering requirements, are highly vulnerable to oligopolistic suppliers. The automakers' investment in keeping the supplier industry competitive has been the key to their success in maintaining bargaining power. As the automakers reduce this investment under pressure from foreign competition, I predict that their bargaining power will fall.
Given the difficulty of obtaining bargaining power and technical change with outside suppliers simultaneously, why are the automakers moving away from financially based voice, which would seem to offer the potential for great technical change (because of high commitment) with full capture of joint profits (resulting from a shared bottom line)? Although the automakers have perhaps been too quick to abandon their internal suppliers, given the legacy of mistrust that attempts to establish reputation-based voice relationships must overcome, vertical integration is certainly no panacea.
There are two reasons for divesting internal suppliers. First, the automakers believe it would be difficult to undo the effects of past over-commitment--more difficult than establishing new commitments with outside suppliers. The divisions' sheltered life has left some of them unable to compete (particularly with European or Japanese manufacturers) on quality, technology, or cost. The auto executives with whom I spoke (particularly those in finance) were especially concerned about the effects of past commitments on shop-floor labor costs and work rules. In 1985, UAW workers in components divisions earned on average $20 per hour including benefits; independent parts plants paid between $5 and $13 per hour for similar work.(90) However, as the example of Fisher Body shows, management problems were at least as important a source of high costs as were production workers.
Second, the nature of uncertainty faced by the automakers has changed from predominantly marginal to radical. Under marginal uncertainty, the possible states of nature are limited to situations, such as a supply interruption, that involve the allocation and pricing of specific assets. Under these conditions, it is helpful to have the ability to engage in "adaptive sequential decision-making" about the costs and benefits of proposed remedies (the cost of transporting the part by air freight versus the cost of a temporary shutdown of the downstream plant, for example). Since such emergencies arise infrequently, and the procedures for dealing with them are similar for a broad class of parts, vertical integration makes sense.
In radical uncertainty, in contrast, under some states of nature it may be optimal to make radical changes (or a series of incremental changes whose overall effect is radical) in the nature of the asset used. Decision making under this type of uncertainty is very information-intensive. Because of bounded rationality, however, an activity can have high administrative coordination with only a few other activities--that is, firms
|Printer friendly Cite/link Email Feedback|
|Publication:||Business History Review|
|Article Type:||Industry Overview|
|Date:||Dec 22, 1991|
|Previous Article:||The modernization of production in the French automobile industry between the wars: a photographic essay.|
|Next Article:||The corporation nobody knew: Bruce Barton, Alfred Sloan, and the founding of the General Motors "family."|