The Invisible Handshake: The Development of the Japanese Automotive Parts Industry

Michael J. Smitka 1
Washington and Lee University

The Japanese auto industry is not only less integrated into parts productionthan the U.S. Big Three, but it also organizespurchasing differently. In Japan,car makerstypicallycontractout subassembly and component manufacturing, while the Big Three primarilypurchase simple parts. ThuswhileChrysler maybuyfrom5,000suppliers (andGM 20,000), Japaneseauto companies buy from 200-300 firms--thoughthese direct suppliers subcontract simple parts production to numerous small firms. Likewise, procurementin Japan is based on long-term "strategic" partnerships rather than the short-termcontracting whichhas been typical of the U.S. untilthe pastfiveyears. One consequence is thatJapanese auto firms are smaller. In 1985Toyotaand Nissan together employed a scant 120,000, whilein NorthAmericaGM aloneemployed 419,000 in 1988. This patternis repeated across mostindustries; two-thirds of the entireJapanese labor force are found in small establishments, while two-thirds of U.S. workersare in largefirms. Because of this structure, managers at Japanese suppliers take over tasks which in Detroit are performed by the visible hand of middle management. In particular,Japanese managersat both suppliersand assemblers face the challengeof coordinatingactivities across firm boundaries.The Japanese auto industry developed innovative approaches to contracting to govern this system, whichinfluenced practice in muchof manufacturing. The parallelis obvious. While in the U.S. in the 1920sGM was one centerfor experimentation with the "visible hand"of internal management,in Japan in the 1950s Toyota developedan "invisible handshake" for managing strategic purchasing.The implications mirror those in the U.S.--having lowered interfirm transactions costs,Japanese companies on the margin resorted to purchasing rather than vertical integration. Management innovations in Japanled not to an increase in the scaleof firms,but (on the margin)to a decrease in scale.

1Susan Helper (School ofManagement, Case Western Reserve) and David Millon (Law
School, Washington and Lee) helpedme structure this paperand improve the title;the discussant,Helen Shapiro (Harvard BusinessSchool),gave a useful critique. For
Japanese language referencessee [10].

BUSINESS AND ECONOMIC

HISTORY, Second Series, VolumeNineteen, 1990.

Copyright (c) 1990by the Business HistoryConference.ISSN 0849-6825.

163

164

Here I try to reveal the outline of the invisiblehandshake and its evolution. SusanHelper [6] shedslight on how the visiblehand molded purchasing patternsin the U.S. My approachreflects three theoretical perspectives. First, I believethat organizational changeis inducedby the external environment; in general, inertiadominates--strategy is reactive, not

proactive.(Thisis a variation of the familiareconomic modelof induced technicalchange.) Second,history matters: the particular timing and
patterns of development in labor, capital and product markets have an enduring impact. (In Japan,for example, the lack of a major market downturn after 1949wascentral.) Third, it alsofollows that adaptations can and often will be novel,and may evenprovidesuperiortools for handling generic management problems. But "better" management will not automatically be adoptedelsewhere. Inferior practicespersist:without severepressures from the externalenvironment, "best" practicemay not diffuseto otherfirms,industries, or countries.This will be particularly true wherecompetition is imperfect and large sizemakesinstitutional change costly, as with the U.S. Big Three. Business historyis a potentially usefulundertaking--even if curiosity and fun is what actually motivates our efforts. We hope to understand why and how corporate management evolved, and henceto suggest its strengths and limitations. The visiblehand helpedgive birth to large organizations in the U.S., but in today's rapidlychanging environment thesenow appear to be a sourceof rigidityrather than a sourceof strength. By contrast, in Japan reliance on the invisiblehandshake--or what in the U.S. are now often called strategicalliances--provided strong incentivesfor technical changeand gave birth to smaller,more flexible organizations.Both are

central issues of the current competitiveness debate. (Seemy forthcoming book for more detail on theseissues IT0].) Two further points deservementioning. In her overviewof the evolution of U.S. contracting, Helper utilizesthe exit/voiceterminology of A.O. Hirschman. In contrast, I draw upon a transactions cost and principle-agent (optimal contracting) perspective. Despite differing terminology and nearly polar practices in the U.S. and Japanese auto industries, I believeour analyses of the relativestrengths and weaknesses of alternativecontracting paradigmsare similar. Furthermore,we argue elsewherethat current practice is converging [7, TT]. The automotive industry is increasingly a worldindustry.Japan is no longera high-growth, developing economy.Japanese and Americanproducers thusface similar challenges. That wasnot always the case.
Inducement to Change:The Adoption of a SubcontractingStrategy

Until the beginning of World War H--T937for Japan--Ford and GM dominated the Japanese automotive industry.At their peak,theyassembled over 30,000unitsannually in Japan. Most of their outputwas on a CKD (completely knocked down)basis, using imported parts. Therewere firms which manufactured commonreplacement items (tires, wheels,batteries, brakelinings, piston rings), butwhenNissan, Toyota, andIsuzuentered the

industryduring 1936-1937, they found few potentialdomestic parts

165

producers, and not all of them were interested in the automotive business. Existing steelproducers, for example, did not believe Toyotawasviable,and refusedto supplysteel of the requisitetypesand consistency neededfor large castings and forgings. The fledglingauto firms were thus forced to integrate vertically. Toyotamadeits ownglass, electrical components, and specialty steel for castings, as well as many of its own machinetools. Nissan Motorsturnedto sister firmsin the Nissan zaibatsu, including Tobata Casting andHitachi [5]. (Silvermakesthis argument for otherindustries, countries and eras [9].) During the war the auto firmswere forcedto turn out munitions, not vehicles. From August 1945 suchproductionceasedand the facilitiesof many firms were temporarilydesignated for reparations to Southeast Asia and placedunder seal. Nissan, Toyota,and Isuzu,the pre-warentrants, continued partialoperations. Alongwith turningout pots,pans,and sundry items, they and the major aircraft producersentered or reentered the automotive industry. They repairedU.S. jeeps,made four wheel trucksand motor scooters, and turned out heavy trucks and buses. Passenger car production resumedin significant volumein 1955 and surpassed truck production only in 1967. In response to this environment, the previousstrategyof vertical integration was reversed.The DodgeLine policies of April 1949provided the impetusand a channel throughwhichthe U.S. Occupation successfully quelledthe postwarinflation. The resultingrecession, however,led many large firms to reduce their work force and produced bitter labor confrontations.The three dominanttruck producers--Toyota, Nissan,and Isuzu--all underwent strikes. Toyota faced bankruptcydue to inventory mismanagement, until it was bailed out by Bank of Japan. The Korean War broke out in June 1950, endingthe overallrecession and leadingto ordersfor trucksand contracts for vehiclerepair,paid for in U.S. dollars. But while outputincreased rapidly,the autofirmswere reluctant to add to their work force. It was unclear how long the boom would last and memories of the confrontation with unions overlayoffs were still fresh. So firmssubcontracted production that haduntil thenbeencarriedout in-house to other,generally smaller firms. Thus,whileToyota's output rosefive-fold during1952-1957, employment roseonly 12%. Two factorsenabledthis shift. Because outputwaslow, production dependedupon general purposetools and skilled workers rather than production linesor otherdedicated assets.In fact,individual manufacturing steps, suchas the drillingof holesand the hand deburring of castings, had longbeen "putout"to smallworkshops. Second, until the late 1950sthere wassignificant excess capacity in manufacturing (particularly in machining andstamping), andthe autoindustry wassmall relative to manufacturing as a whole. It was thus relatively easyto find suppliers for simpleparts at competitiverates. This made subcontracting doubly advantageous.By turning to outsidesuppliers, the auto firms were able to increasetheir output without new investment. Instead,they could devotetheir limited fmandalresources to final assembly, newmodeldesign, andotheractivities
which remained in-house.

166

The management of the early subcontracting system was straightforward. Pricingwaseasy: the purchaser simply paid the goingrate for that typeof product.But logistics andscheduling were a nightmare and qualityremained a problem. Suppliers had little incentive to investon the auto ftrm'sbehalfbecause of their explicitstatusas a safetyvalve. In a recession, theyfully expected the automakers to pull work backin-house. But because prices werereadilyobserved, existing suppliers couldbe offered the right of f'trstrefusalon new business, in contrast to Detroit, where orders were shifted frequently among f'trms. In most respects, subcontracting initially resembledthe "low-tech" strategywhich Helper describes for the post-WorldWar II U.S. auto industry. But work and workerswere not periodically shiftedin and out of the ftrm. Instead, the subcontracting paradigm shifted. Contrary to expectations, the Japanese passenger car industry sawtwenty-five yearsof continual,indeed phenomenal,growth. Rather than facing periodic recessions, output doubled every two to threeyears, slowing onlyafter 1970. The truck and three wheel vehiclemarketsboth grew more slowlyand remainedmore cyclical. In addition, new entrykept frmsin a constant race to improvequalityand lower costs--in contrast to the stableoligopoly that formed in the post-war U.S. industry. By the late 1950sthe volume made it difficultto purchase partsfrom localjob shops. Nor were such shops able to provideadequate quality. But rather than vertically integrate--and increase directemployment and bank borrowings--the auto plantschanged the contentand methodof subcontracting.Assemblers thus increasingly relied upon their suppliers for not only simplemachining but alsofor the manufacture of parts and graduallyof entire subassemblies. On the one hand, in the face of the rapid increase in output,suppliers shiftedfrom general-purpose machine tools to proper productionlines. In some casessubassembly lines were physically shifted to the factories of the more reliablesuppliers.In others, the autoftrmsencouraged their suppliers to growin scaleandsophistication by providing assurances of continued orders. On the otherhand,assurances of future ordershad to be givenmerelyto obtainthe interestof suppliers. The overall Japanese economy wasgrowing, and the entrepreneurial owners of smallsubcontractors neededa positive incentive not to moveelsewhere. By 1960 both the content of subcontracting and the nature of the relationship had shifted. Toyota,the mostsuccessful of the early entrants, set the pacein the mid-1950s.Later entrants, suchas Mitsubishi, changed only in the early 1960s. With the shiftto the subcontracting of more complex operations, the natureof the contracting relationship alsoevolved. Contractcontinuity, as noted above,was usedto encourage investment and to lock in supplier capacity. But as dedicatedtooling and even productionlines became widespread, it wasno longerpossible to switch ordersat low cost. Indeed, in the shortrun the purchaser becamedependent on the existing supplier, and suppliers,generally small firms, were likewise dependenton their automotive customer for an ever-increasing share of their sales. Furthermore, as production volumewas still low compared to that of the

167

U.S., economiesof scale worked against multiple sourcing,even for relatively simpleparts.
As the transaction cost literature stresses,the auto firm and the

supplierof a part thus found themselves in a bilateral monopoly[13]. Contracting in this one-on-one environment was fraughtwith potential disputes overpricing andotherdetails; a supplier couldholdout at the last moment for a priceincrease, or the purchaser couldthreaten to take the business elsewhere or (in Detroit) pull it in-house. Furthermore, it was ideally an ongoing relationship,requiring investmentsand manifold adaptations overtime. An inability to manage such contracting relationships is widelyheld to lie behindthe bias in the U.S. towardverticalintegration [8]. (But seeHelper'schapter and [4, pp. 44-46].) In anyevent, written contracts were inadequate; transactions weresimply too complex to specify in muchdetail,andthe courtsystem too unwieldy to resolve disputes. A newframework wasrequired if the strategy of subcontracting complex parts production was to be maintained over the longhaul.
Innovation: Managing Interdependence

In their attemptto copewith interdependency--and to develop more capable suppliers--managers in the Japanese autoindustry were forcedto innovate. In theU.S., asHelperdiscusses, FordandGM combined vertical integration and the purchase of simpleparts,relyingupon hierarchy and contracts respectively. In Japan, however, the autoindustry avoided vertical integration while relyingupon suppliers for complex parts,and thushad recourse to neithermarketnor hierarchy.Insteadtheydeveloped a hybrid mechanism for governing transactions that relied cruciallyupon trust. Personal trustalonewasinadequate for governing the complex interactions of two firms, but it provided an important starting point. For interdependency to develop, both parties had to placethemselves at risk. Personal relationshipsprovided the assurance needed to initiate subcontracting. But the purchasing relationship alsoevolved gradually as volume,variety,and complexity of subcontracted work all slowly increased. Second, overtimenorms for pricing, delivery, quality, andotherdetails were developedwhich lessenedroom for dispute. Third, well-specified mechanisms for inter firm communication ("voice", as Helper wouldcall it) helped prevent incipient disputes. Finally, conscious investments weremade to build and maintainreputation. In a multi-firm,repeatedcontracting environment, thishelped provide sanctions for bothsides to remainfaithful. Thusa verycomplicated relationship evolved, in whichtrust,reputation, and interdependency made commitmentscredible, while the norms and expectations that arose helpeddelineate the implicit termsfor transactions. Trustandnorms wereexactly the sortof problems, for instance, that Mitsubishi Motors(thenCentral Japan Heavy Industries) faced in 1958-1959 at its Mizushima Plant [10, Chapter2]. The finn, whichat the time made three-wheel trucks,pulledwork in-house duringthe 1958 recession.A boomfollowedandthe firm haddifficulties finding suppliers. Other rums already had faced and overcomesimilar problemswith subcontractors. Mitsubishi wastherefore able to call in outside consultants in 1961to help

168

turn aroundits subcontracting system.Theseconsultants from the Nagoya areawerefamiliarwith Toyota's supplier management practices, whichwere systematized between 1952-1954. In line with their recommendations, Mitsubishi committed itselfto suppliers, developing a long-term purchasing plan to help convince them that it wouldnot pull work from them in the future. It also clarifiedthe responsibility for interactions with suppliers internallyand revised pricingpolicies. Thesehelpedthe firm make a start at settingnormsand rebuilding trust. An importanttool of the new policywas the suppliercooperation association or kyoryoku kai whichMitsubishi set up, following the example of Toyota. One function was to maintain reputation. The association broughtthe management of all local suppliers togetheron a regularbasis, and so all suppliers wouldsoonknow of any deviation by Mitsubishi from its statedpolicies.But the cooperation associations alsofacilitated efforts by each auto firm to bring a measureof organi7ation to its supplier network. The associations thus helped the auto firms achieve the advantages of coordination whichare typically seenas the peculiar province of internal organi7ation[3]. For example, through the cooperation associations the auto firms discussed future product strategyand sales forecasts with suppliers, coordinating investment plansacross firms. They also discussed changes in productdesignand automotive technology which might make currentsuppliers' capacity redundantor require new entry. Finally,as detailedbelbw,they were the institutional locusthroughwhich the autofirmstaughtsuppliers bettermanagement andproduction methods. By working closelywith suppliers,an auto assembler could focus its attentionon new model development and other strategic decisions, while leavingadaptation and implementation to others. The poor technicaland managerial capabilities of suppliers led the autofirmsto develop their suppliers. Throughthe cooperation associations the autocompanies ran seminars andworkshops on a widevarietyof topics for supplierengineers and managers.Early effortsfocused on the industrial engineering techniques neededto set up and run production lines and on the costaccounting to monitorthem. Along with bringingin consultants, the auto firms involved their own engineering staff, sometimes seconding themfor a half-year or more. Laborrelations techniques (e.g.,QC circles),

statistical quality control, andjust-in-time (JIT) process control were all

taughtin thisway. (Note that the development of JIT wasinduced by the logistical problems mentioned earlier.) Finally,in the 1970s, the autofirms stressed the use of value analysis (VA) and value engineering (VE), industrial engineering methodologiesfor coordinating the systematic examination of designparameters by cross-functional teams in order to improvevalueand/or decrease cost. VA/VE programs enabledsuppliers to undertake more of the initial design process and retain greatercontrol overcosts andquality.In fact,Japanese autofirmscanneithermanufacture nor design a new car nowwithoutthe inputof their currentsuppliers.The cooperationassociations were thus central in introducingimproved management methods andbetterproduction technology, andin coordinating designacross firms.

169

Finally, norms for pricing were crucial in making subcontracting workable. While purchasing took placeon an on-going basis,all the auto firms produced multiplemodels,and hencecouldpurchase similar parts from two to three differentsuppliers. But the autofirms did not rely upon directrivalryamong these firms. Instead, suppliers wererequired to submit bidswith cost breakdowns by process, andnotjust a quoteon the final unit price. Since at anygiven time manysuppliers used the same manufacturing process--stamping, boring, plating--detailed comparisons at the process level werepossible across firmsand(the current method) across time [1, 2]. The bid systemthus provideda relativelyobjectivestartingpoint for price negotiations. But the bid alsoprovided detailedinformation whichhelped pinpointthe source of highcosts and henceto directengineering effortsto problem areas. In turn, low bids for a given process could signalnew techniques whichthe auto firms wouldthen try to ferret out and teachto others. The pricing mechanism thusnot onlyhelpedfirmsto avoiddisputes but also encouraged technical and management innovation. While there was rivalryamongsuppliers, it was restrained and often indirect; the currentsupplier for a part typically wouldhavethe rightof first refusalfor the equivalent part when a new model was introduced. This providedthe assurance firms neededto investin plant and equipment withouthaving to front-load capitalcosts onto the work at hand. (U.S. automotive suppliers often requirea two year payback on any significant capitalinvestment, due to pastbitter experiences with work being pulled in-house by the Big Three [7, andmy owninterviews].) But in addition, the pricingformula meant that a firm which implemented innovations faster thatits rivalsearned highprofits, andfirmswereexplicitly compensated for designinnovations stemmingfrom their VA/VE programs. Failure to remain competitive resultedin lost orders. While innovative firms were quite profitable,many parts firms exited in the 1960sand early 1970s. There were thusmanyincentives, positive and negative, for the partsfirms to improvetheir operations.
Implications and Extensions

Supplieroperations improvedrapidly. This was critical for the success of the industry, since purchases by the automakerscomprised 70% of manufacturing costs. From 1958-1965, when costsfell by half, lower parts pricesaccounted for 54% of the reduction(and lower steel and materials prices for 16%),whileonly32% accrued frominternal costsavings of the auto makersthemselves [10, Table 3.7]. In 1955 the Japanese passenger car industrywas competitive in neither cost nor quality with imported vehicles despite significant tradebarriers. It is now competitive throughout the world,in largemeasure through the mobilization of outside suppliers.The industry achieved this not throughhierarchyor market,but by developing a sophisticated and effident mechanism for coordinating activities with independent suppliers. The systematic reliance on strategic alliances wasnot restricted to the auto industry. The postwar business environment encouraged Japanese fu-ms in manysectors to avoidvertical integration. In the assembled goods

170

industries, thiswasachieved by developing closetieswith suppliers, drawing in part uponthe automotive example.As Helper argues [7], a partnership with suppliers is not unknown in the U.S., and eventhe auto industry is making halting efforts to changeits strategy[11]. Coordination is not important in all industries and in such casespure market purchasing is adequate. Similarly, in some instances it is not possibleto maintain a semblance of rivalry,whichunderlies the viabilityof the Japanese pricing mechanism.Thus Japanese auto firms make their own engineblocksand largebodystarepings. But wherecoordination is neededthe Japanese auto examplesuggests that partnerships with suppliers can be managedso as to maintain order, without resort to bureaucraticcoordinationunder vertical integration. Why is the strategy not more widespread?The primary barrier, particularly when a strategy is adoptedwidely,is that it cannotalways be reversed.In the Japanese case, the widespread utilization of subcontracting encouraged the growthof small firms. Machiningdistricts, reminiscent of the garment district in Manhattan, are stillwidespread. One careerpattern of an ambitious youngster, in fact,is to apprentice in a series of smallshops and then set up on his own. Historically, manyforemenleft largefirms to become suppliers to theirformeremployers. Because of thisthereremains a large reservoir of entrepreneurswho can be turned to as subcontractors--though the auto industryis no longer prized as an end customer.In the U.S., vertical integration limitedthe growthpotential of smallfirms,and the low tech partspurchasing strategy relegated them to stagnant andunstable linesof business. In manyAmerican industries there are thus few capable suppliers,rendering it difficult to adopt a subcontracting strategy.On the otherhand,Japanese firms cannotreadily beginpullingwork in-house, as it wouldthreaten the cooperation on which their entiresystem depends.Several Japanese consumer electronics firms rely upon morn-and-pop storesfor the majorityof their salesand are findingit hard to shift to discount distributors. Another reasonwhy the strategyis not more widespread is that business cultureis moldedby strategy. In Japan trust and reputationare wellunderstood. The Japanese equivalents of Dun & Bradstreet are careful

to listkeysuppliers andcustomers, whilethe owner/operators of small firms (and their counterparts in largeones)are schooled in running cross-firm partnerships. In the U.S., hardball contracting practices haveleft a legacy
of distrust,and American business culturelacks imagesof and presumed familiaritywith partnerships. Once adopted, therefore, verticalintegration andnon-integration maybothbe equilibrium strategies--even if theybecome dysfunctional. Continuedresearchon Japanese business historypromises many interesting extensions. In the Japanese legalenvironment the courtsystem
is not a viable means of conflict resolution. Part of the reliance on trust

in subcontracting may thus representan adaptationto the inability to enforce contracts. Accounting alsoappears to havedeveloped in a divergent direction: formal capitalmarketshave been unimportant, and so reporting requirements havenot beenrelevant for mostfirms. Instead, management accounting has been more central[12]. Little has been written on this

171

topic, evenin Japan. The area of corporate services differs;small firms utilizeneitheraccountants nor lawyers.Some services are provided by their customers and suppliers, but there is an array of business consultants in Japanwhichhasyet to be described in English. Finally,how precisely do firmscommunicate with eachother? I havesketched two key areasabove, thebid system (for pridng)andvalue analysis (for technical change). There are clearlyothertechniques that helpthe invisible handshake of smallfirm managersand their customers to substitute for the visiblehand of middle
management.
References

1.

Banri Asanuma, "The Organization of Parts Purchases in the Japanese Automotive

Industry," JapaneseEconomicStudies, 13 (Summer 1985), 32-53.

2. ,
3.
4.

"TheContractual Framework for PartsSupplyin the Japanese Automotive

Industry," JapaneseEconomicStudies,13 (Summer 1985), 54-78. Alfred D. Chandler,Jr. Strategyand Structure: Chaptersin the Histon/of the American Industrial Enterprise(Cambridge,MA, 1962).
Ronald Coase, "The Nature of the Firm: Influence," Journal of Law, Economics, &

Organization, 4 (Spring1988),33-47.
5. 6. MichaelCusumano,TheJapaneseAutomobile Industn/: Technology and Management

at Nissanand Toyota(Cambridge, MA, 1985).

Susan Helper, "Comparative Supplier Relations in the U.S. and Japanese Auto

7. 8.

Industries:An Exit/Voice Approach,"Business and Economic Histon/,2nd series, 19 (1990). , "Strategy and Irreversibility in SupplierRelations The Case of the U.S. Auto Industry," Business Histon/Review(forthcoming). BenjaminKlein,R.A.Crawford,and ArmenAlchian, 'ertical Integration, Appropriable Rentsand the Competitive Contracting Process," Journalof Law and Economics,21
(October 1978), 297-326.

9.

MorrisSilver,Enterprise and the Scopeof the Firm (Cambridge, MA, 1984).

10. Michael Smitka, Competitive Ties:Subcontracting in theJapaneseAutomotive Industn/ (New York, forthcoming, 1991). 11. , "AmericanManagement: Reformationor Revolution? The Transfer of JapaneseManagementTechnolo9yto the U.S.,"Working Paper No. 37, Center on Japanese Economyand Business,Columbia University,November 1989. 12. H. Thomas Johnson and Robert S. Kaplan, Relevance Lost The Rise and Fall of

Management Accounting (Cambridge, MA, 1987). 13. OliverWilliamson, The Economic Institutions of Capitalism (NewYork, 1985).