Hevkn 1994, Vbt 9. Na 4. 63&4SS2. Formulating Vertical Integration Strategies^ KATHRYN RUDIE HARRIGAN Columbia University
A framework is proposed that develops the dimensions of vertical integra- tion strated and proposes key factors that might augment their uses within various scenarios. These represent new hypotheses and conjectures about make-or-buy decisions that require empirical testing. If the framework is valid, strategists could formulate better hybrid vertical integration strategies by recogniang the hypothesized effects of these forces on the industries that might be linked.
Managers need a ready supply of raw matoials and tbey have not recognized how to make this a more services, as wdl as a ready market for their firms' durable and keen competitive weapon. Because suc- outputs. The business arratigements tbat are used to cessful vertical integration strategies require the control tbese risks are forms of vertical integration. cooperation of several strategic business units Tbey tnay include vertical acquisitions, or intemal (SBUs), the formulation of such strategies is in the devdo{«ient of supplying or distributing units, or province of the chief executive officer (CEO). In otber means of extending firms' control over out- some cases, effective vertical integration may even siders. require temporary subsidization of one business unit at the expense of another. Decisions regarding sucb The Phenomenon of Vertical Integration SBU coordination (and resource allocations among Vertical integration is a corporate strategy tbat has them) must be made by the chief strategists. Tbus been misunderstood. It has lotig been a key force in effective vertical integration strategies need to refiect tbe development of bigb productivity and managerial botb business unit and corporate level strategy scqibistication in U.S. business (Cbandler, 1977). Ver- requirements. tically integrated corporations bave been key engines Tbis paper proposes a framework for developing of cbange in tbe past and bave enbanced sbarebolder effective vertical integration strategies. It was wealtb (Lubatkin, 1982). Yet earUer findings tbat developed by syntbesizing tbe tbeoretical foundations "dominant verticals" (Rumelt, 1974) and vertical establisbed by tbe industrial economics and strategic mergers were least successful as diversifications management literatures witb firms' observed behav- (Baker, Miller, & Ramsperger, 1981) may bave iors. Thus it suggests a new way to look at vertical soured mana^rs and academic researcbers on tbe integration and the forces that affect firms' cboices usefulness of tbis strategy unnecessarily. Oftentimes concerning vertical linkages. It develops normative researcbas did not recognize tbat votical integration propositions conceming wbicb generic vertical could be an effective strategy, provided it was used strategies migbt be more appropriate under different pnutently, because tbey often took an overly aggre- competitive circumstances, and it uses examples of gated view of it. Because critics bave not discemed firms' successes or failures in using vertical integra- bow tbe impOTtant dimensions of vertical integraticm tion to suggest bow traditional concepts of this be adapted over time (as industries cbange). strategy mi^t be amended to refiect effective in- dustry practices. These suggestions are new hypo- theses, which will require empirical testing. Vertical 'The research was st^iported by tbe Strategy Researcb Center, Cdurabia University. Comments and oitidsm from William H. integration is one of tbe first diversification strategies Newman, Ian MacMillan, and Donald C. Hambrick are also tbat firms embrace. Unfortunatdy, some firms seem graulaHy acknowledged. The author is indebted to Donald C. to use it in a manner that seems inappropriate for I&iabrick for suggesting the graphic configuration in Fipire 1. tbdr circumstances. Tbe issue of vertical inte^ation 638 deserves additional analysis because it bas been sified firms he bad classified in otber ways. If misunderstood in tbe past, and because tbe develop- previous studies of votical inte^ation used dassifica- ment of a more rigorous means of analyzitig this tion rules simiiar to Rumelt's, tbey also may bave strategy (and the performance it promises) cotild understated its tme activity level. For example. result in the formulation of more effective industry Federal Trade Commission data for 1948 to 1972 linkages, more rapid technological improvements, conceming large acquisitions (assets larger tban $10 stronger global strategies, and better use of vertical million) indicate tbat less than 15 percent of these integration. Table 1 summarizes many ofthe advan- were vertical mergers; and Salter and Wdnhold's tages and risks associated with vertical integration. (1979) list of $100 million or more acquisitions made The narrow case in which a firm's major diversi- during 1975-1978 indicated that only 4 percent were fication strategy has been only vertical is not the focus vertical. Thus an illusion was perpetuated that ver- of this inquiry. Instead, it considers the larger tical integrations were rare, except in the ral, mbber, universe of firms that have linked two or more SBUs basic metals, and forest products industries. But, in through vertical relationsbips and bave diversified in fact, acquisitions that are classified as bdng other- otber ways as well. Rumeh (1974) would bave found wise "related" to tbe firm's core businesses may also tbat many of tbe more diversified firms in his For- bave provided new distribution cbannds or other tune 500 sample also bad vertical transfers of goods assets that are vertically related to them. or services in-bouse, bad tbat been tbe focus of bis Acquired firms are bundles of assets rather than inquiry. Tbus Rumelt reported tbat 22 percent of tbe single business units, as Ocddental Petroleum firms in bis sample embraced a dominant vertical discovered when courting Cities Service (and as Du strategy in 1969 (up from 20 percent in 1949), but Pont learned while absorbing Conoco). Acquisitions vertical linkages also existed witbin tbe bighly diver- can offer vertical linkages as wdl as nonvertical diver- sifications. Corporate strategists must dedde whether to retain the business units acqtiired inddentally in Table 1 this fashion; and if tbey are vertically rdated, Some Advantages and Disadvantages strategists must dedde wbetber to encourage intra- of Vertical Integration firm commerce (subsidy) or demand arms-length transactions between vertical sister units. Thus, more seemingly unrelated mergers bave vertical elements Advantages Disadvantages to tbem tban is generally recognized, but oftentimes Internal benefits Internat costs strategists see no advantage to encouraging vertical Integration econoniies reduce Need for overbead to coordinate relationsbips between in-bouse units. Tbis may oc- costs by eliminating steps, re- vertical integration increased ducing duplicate overbead, costs cur because tbe opportunities for vertical competitive and cutting costs (tecbnology Burden of excess capacity from advantage bave passed in some industries. In otber dependent) unevenly balanced tninimum Iniproved coorditiation of activ- efficient scale plants (technol- cases, bowever, tbis occurs because managers did not ities reduces inventorying and ogy dependent) recognize bow to exploit tbe advantages of vertical other costs Poorly organized vertically in- Avoid time-consuming tasks, tegrated firms do not enjoy integration effectively. such as price shopping, com- sipergies tbat compensate for municating design details, or bigber costs negotiating contracts The Legacy of Vertical Integration Cortipetitive benefits Competitive dangers Avoid forecbsure to inputs, ser- Obsolete processes may be per- Vertical integration bas been an important mana- vices, or markets petuated gerial innovation and a necessary technological step Improved marketitig or techno- Creates mobility (or exit) bar- logical intelligence riers in developing certain industries, but it may not be Opportunity to create product links firm to sick adjacent busi- appropriate in tbe same form under all circum- differentiation (increased value added) Lose access to itiformation from stances. For example, ownership of ore mines, ships, Superior control of firm's eco- suppliers or distributors foundries, rolling mills, and fabricating plants was nomic environment (market Synergies created through ver- power) ' tical im^ration may be over- necessary for steel companies to lower costs and im- Create credibility for new prod- rated prove productivity in the 1890s. In its early years. ucts Managers integrated before Ford Motor Company owned and operated every Synergies could be created by tbinking tbrough the most ap- coordinatbig vertical activities propriate way to do so stage of processing from iron ore to Hnisb-and-trim skillfully operations (except tires and glass). In these early 639 years, suppliers may not have berai as willing to share on tbeories of market power and tbe ideal of perfectly Ford's gamble in persua^ng consumers to purcbase competitive industries (Addman, 1979; Blair & "horsdess carriages," so Ford bad to devdop com- Kaserman, 1978; Comanor, 1967; Dennison, 1939; ponraits to its spedfications for itsdf. Tliere were Frank, 1925; Jewkes, 1930; and Lavington, 1925). substantial economies assodated with vertical in- Tbese scholars largely have not recognized that dif- tegration once FOTd bad overcome the public's ferent motives for vertical integration—such as resistance to ptu'ehasing this tiovel product, and these technological leadersbip, to secure access to raw cost advantage rewarded Ford's gamble. Ford's in- materials, or competitive preemption—migbt exist tegrated strate^ and logi^ical system enabled large witbin tbe same industry; nor bave tbey considered numbers of consumers to afford low priced and tbe diversity of ways in wbicb vertical integration reliable automobiles in 1910 (CbamSa, 1977) because strategies migbt be formed (Adams & Dirlam, 1964; it lowered FcH-d's costs of (^-ocurement, standardized Qevenger & Campbell, 1977; Greenbut & Obta, its components, and fadlitated an end-to-end pro- 1979; Larson, 1978; Mancke, 1972; Perry, 1980). For duction process. Tbis is tbe type of vertical integra- example, firms vary in bow many tasks tbey perform tion bebavior one tnigbt expect witbin emerging in- in-house, in the number of buyer-seller linkages dustries wbere firms must provide tbdr own infra- downward in a vertical chain they forge, and in the stmctures and otber supplies. form of control employed. Few economic scholars, except perhaps Bork (1954), McCSee and Bassett By 1983, bowever, tbe automobile industry bad (1976), anti Porter (1980), have recognized the ways matured sucb tbat uncertainties regarding generic in which vertical integration could make industries product demand were reduced. Ford's outside sup- more competitive (rather tban less so). Most econo- l^iers were willing to invest in tooling and otber assets mic scbolars bave beld one view of vertical integra- to supply tbe autotnakers, and bigb degrees of inter- tion, a view based beavily on tbe convenient assump- nal transfers were no longer necessary if unecono- tion of a monopolist, instead of considering bow mic. (And tbe throughput of U.S. automakers was firms migbt use tbis strategy differently. not large enough for vertical integration to remain Because industry structures differ, it is not surpris- as economic as it once was.) The challenge from ing tbat many approacbes to vertical control could Japanese automakers was difficult to meet when satisfy managers' needs for a ready supply of raw firms sucb as Ford were so strategically infiexible. materials or a ready market for tbeir factories' out- Moreover, vertical integration bad lost some of its put. Tbe successes some firms bad witb strategies of attraction because managers (wbo often resented bav- full integration, long vertical cbains, and otber varia- ing to buy from sister units) did not understand tbe tions is surprising, bowever. Some firms, such as role tbat vertical integration played in tbdr firm's Robert Hall or Botany Industries, have suffered corporate scbeme. Often finns did not have the sup- notable failures from vertical integration of the porting mechatiisms needed to reap the maximum wrong type and/or its use at the wrong time (Harri- synergies that migbt be available from vertically in- gan, 1983a). But if managers better understood the tegrated linkages, or tbey tnisapidied tbem in otber many dimensions of vertical integration and the key ways (Williamson, 1975). In brief, in tbeory, many forces that affect their abilities to execute vertical firms favor making ratber tban buying key resources strategies well, they could better avoid fundamental and se^ces, but their inabilities to matiage integra- errors assodated with vertical integration and maxi- tion taint thdr appredation of this strategy. More- mize the benefits available in joining dissimilar but over, the use of vertical int^ratim must change with rdated businesses. Briefiy, managers would not at- time. The competitive damage created by excessive tempt to create synergies in cases where external integration can be substantial, as in the examples of forces made integration too risky or thdr intemal the U.S. automobile and sted industries in 1983. systems made communicatiotis inadequate. Antitrust dedsions also have created a tarnished image of vertictd int^'ation. Economists, wbo did A New Look at Vertical Integration not consider tbe particular requisities of diverse The dd concept of vertical integration as being 100 firm's corporate strategies, bave rdnforced a percent owned operations that are physically inter- unidimensicmal view of vertical integraticHi based connected to supply 1(X) percent of a firm's needs is 640 outinoded. Under !^q>r(9riate drcumstances, quality Breadth of Integrated Activittes control and access to stable supplies can be obtain- Tbe breadtb of integrated activities is tbe numbs ed through quasi-int^ration arrangements. Firms of tasks tbat firms perform in-bouse. Firms perform- could contract for R&D services, for example, to ing many upstream or downstream tasks in-bouse are utilize the technology of genetic engineering in prod- broadly integrated; firms performing few vertically uct devdopment, or they could form joint ventures related tasks are tiarrowly integrated. to obtain this capability. Finns could bave com- ponents engineered to tbdr tigbt and bigbly spedfic Traditional concepts of vertical int^ration did not instmctions by outsiders, as do Japanese automobile address tbe number of integrated activities tbat firms manufacturers, for example. And if tbeir bargain- migbt undertake. Figure 1 contrasts tbe old view, ing power is sufficient, firms can use a kanban or represented by firm A, witb examples of tbe ex- "just in time" system of inventory control tbat sbifts panded concept of vertical integration proposed bere. tbe burden of bolding costs to tbeir suppliers In Figure 1, firm A is not as broadly ititegrated as (Obmae, 1982). are firms B and C. Circumstances in wbicb firms migbt cboose tbe broadest integrations successfully If firms prefer not to use outsiders as extensions are suggested elsewbere. of tbdr corporate entity, a variety of otber vertical arrangements are possible. Some firms may conclude Stages of Integrated Aetivities that they need not undertake certain activities at all. Eli Lilly, for example, uses outsiders exclusively with Tbe number of stages undertaken in tbe dimension success to merchandise its ethical Pharmaceuticals. of vertical integration tbat many traditional views Tandy/Radio Shack, by contract, uses primarily its have embraced. Figure 1 shows that firm A has more own retail outiets to distribute personal computers, stages than firms B or C, because its activities extend but it has been increasing its use of outsiders. In other from ultra-raw materials to retiiil ouUds, but firm situations, firms may find that tbey can enjoy tbe in- C is engaged in a greater number of steps in the va- tegration economies, uncertainty reduction, compe- tical chain. Although Rgure 1 depicts the transfor- titive intdligence, and otber benefits tbat intemal ver- mation process as an extension of adjacent stages ac- tical linkages may provide tbrotigb outsiders. The key tivities, it is possible for firms to skip a stage in the in using vertical integration is recognizing which ac- chain (by using outsiders for an intermediate process- tivities to perform in-house, how to relate these ac- ing step) in order to monitor costs better, to save on tivities to each other, how much of its needs the firm asset investments for fadlities that would be under- sbould satisfy in-bouse, bow mucb ownersbip equity utilized if brought in-house, or for otber reasons. needs to be risked in doing so, and wben tbese dimen- Degree of Internal Transfers sions sbould be adjusted to accommodate new com- petitive conditions. Briefiy, tbe concept of vertical Tbe degree of integration is tbe proportion of a integration sbould be expanded to encompass a vari- resource transferred intemally, and fully integrated ety of arrangements by wbicb tbe firm cim use out- firms transfer almost 100 percent of a particular ser- siders (as well as its own business units) to forge an vice or material in-bouse. In Figure 1 only firm C optimal vertical system for supplying goods, services, is "taper integrated" with respect to services and and capabilities. materials upstream and downstream. Firms A and B are "fully integrated." The Dimensions of Vertical Strategy Form of Ownersbip Arraagement In any vertical integration strategy, consdous (or The form of integrated ownersbip indicates the unconsdous) dedsions are made regarding: (1) tbe proportion of a firm's equity invested in a vertically breadtb of integrated activities undertaken; (2) tbe linked venture, and in some environments carefully number of stages of integrated activities; (3) tbe specified contracts, franchises, jinnt vmtures, or degree of intemal transfers for eacb vertical linkage; other forms of quasi-integration can be good alter- and (4) tbe form of ownersbip used to control tbe natives to wholly-owned ventures. Figiu-e 1 does mx vertical rdationsbip. illustrate different f Components The Warehousing Ultra- Product Processed Process and Strategic Assembly in and Pbysical Raw Develoianent Raw Innovation Finisbed Business Marketing Downstream Wbolesating Distribution Retail Materials Services Materials Services Materials Unit Services Product Services Services Outlets ^ Key:- - Solid lines indicate intrafirm transfers. — Dotted lines indicate external purcbases or sales. A In A, tbe strategic business unit is flanked by integrated sister tmits upstreani and downstream. Tbe firm is engaged in tnany stages of integrated activity, and all transfers of products and services witbin the vertical chain are made in-house (high degree of integration). The relationship between any two business units in A is "fully integrated." B in B, the business unit's upstream atid downstream sisters engage in many activities. Tbus the firm is broadty integrated. Tbe firm is engaged in many stages of integration, but tbe tengtb of tbe vertical chain differs for various inputs. As drawn, tbe firm is "fully-integrated** for the inputs it does supply in B because it transfers those inputs in-bouse. C In C, the business unit purcbases more inputs^ or services from (or sells more to) outsiders than in B. Tbe firm is less broadly integrated tban in B. It is engaged in many stages of activity downstream for one output. It is taper integrated for many inputs, induditig product development services, because the business unit purchases some inputs and services from outsiders and sells some outputs to outsiders. alternatives could be identified by the percentage of to tbinking about this problem. total eqtiity firms risked in a particular vertical rda- Nonintegration. Strategies for attaining materials tionship. There are situations in which partial or no and markets with no internal transfers and no owner- ownership may be preferred to wholly-owned vertical ship are like contracts. They are espedally attractive linkages. when firms are reluctant to buy spedalized assets, need to lower breakeven points because of Vetical Integration Strategy Alternatives underdeveloped demand, or can arratige delivery schedules with suppliers (or distributors) as though How can firms best manage thdr needs for scarce they were extensions of the firm's assets. Koppers and supplies or access to distribution channels? Several Motisanto both used this approach to vertical int^ra- altematives are suggested that encompass the dimen- tion successfully in genetic engineering, in which de- sicms of vertical integraticm strategy. The mix or com- mand was highly uncertain and technological cbange binations of tbese approacbes are bypotbesized to occurred rapidly, synergies were low witb ongoing cbange over time, as industry conditions cbange or businesses in 1981, and tbe figure bstd bigb bargain- as firms' needs to control adjacent industries tightiy ing power with respect to upstream and downstream change (Sichel, 1973). Tbese altematives are: nonin- markets. Firms risk the lowest proportion of thdr tegration, quasi-integration, taper integration, and assets in vertical arrangements involving noninte- fuU integratim. Previous tbeories of vertical int^ra- grated controls. don did not recognize tbe different dimensions com- Quasi-integration. Quasi-integrated firms need not prising it; tbtis combinitig tbdr use to create generic own 100 percent of the adjacent business units in tbe strategy dtematives represents a new vertical cbain to enjoy tbe benefits of bonding tbdr 642 interests to other firms' interests. The bond between provided them with access to enough crude oil to rirms could take the form of cooperative ventures, keep thdr plants running economically, and that sdl- minority equity agreements, loans or loan guarantees, ing a portion of thdr primary petrocbonicals to othra prepurchase credits, spedalized logistical facilities or firms allowed them to gain scale economies through "understandings" concerning customary arrange- spedalization in processing downstream. Taper inte- ments (Blms, 1972; Porter, 1980). Downstream quasi- gration represents a useful compromise between de- integration arrangements enable firms to retain a net- sires to control adjacent businesses and needs to re- work of qualified distributors to maintain quality im- tain strategic fiexibility. ages. Upstream "take-or-pay contratts" andkanban Full Integration. Physically interconnected arrangements enable firms to enjoy the advantages technologies usually involve high degrees of internal of vertical integration without assuming the risks of transfers, but full int^ration also can be used effec- it. Whiskey distillers used quasi-integration suc- tively if price competition is not fierce, disecononmies cessfully to penetrate diverse geographical markets, from temporary imbalances are not significant, and and microcomputer producers used it to obtain soft- little hardship occurs from bdng cut off from out- ware and distribution of their products. The com- petitive scanning advantages of quasi-integration can side market or tecbnological intelligence. Transfer- be especially effective if firms using it devise in- ring all of tbe firm's needs for a particular good or telligence gathering mechanisms to use the informa- service in-bouse exposes it to increased risks of ex- tion that adjacent firms and competitors might cess capadty, competitive infiexibility, and loss of provide. infomiation concerning customer or competitive cbanges. Firms also face bigber capital costs and Quasi-integrated arrangements place greater pro- bigber exit barriers with this strategy (Harrigan, 1981, portions of ownership equity at risk, but they also 1983b). Nevertheless, Brooks Brothers sells its own provide greater fiexibility in responding to changing tailored suits with success, Courtaulds used its own conditions than a contract may provide. Tbe tbird rayon fiber in textiles, and PPG Industries used its ownersbip alternative, full ownersbip, is most fre- own synthetic soda ash to make glass. These firms qently observed. It assumes tbat tbe firm exerts com- were fully integrated with respect to the materials plete control over tbe activities of tbe vertically linked businesses. Full ownersbip risks tbe greatest propor- named above without encountering the problems tion of equity, but many firms believe it is easier to otber firms bave faced witb tbis strategy altemative. manage tban contractual or quasi-integrated rdation- In general, it would seem tbat full integration works sbips and prefer it over tbem (Harrigan, fortb- best witbin stable environments, but for corporate coming). strategy reasons it may be necessary if outside sup- pliers or distributors are inadequate. Altbougb it Taper Integration. Wben firms are backward or seems generally wise to bave competitive antennae forward integrated but rely on outsiders for a por- tion of tbdr supplies or distribution, tbey are' 'taper collecting intelligence upstream and downstream by integrated." Sucb firms can monitor tbe R&D de- enga^ng in some commerce witb outsiders, taper in- velopments of outsiders, reduce vulnerability to tegration may not be necessary in some settings in strikes and shortages witbin their systems, and ex- wbicb full integration is tbe more profitable strategy. amine the products of competitors while enjoying tbe Breadth and Stages of Integration. Firms tbat per- lower costs and greater advantages (and profit mar- form many activities involved in making a particular gins) or vertical integration. Under certain drcum- product (sucb as Pfizer in pbarmaceuticals and Ten- stances taper integration is not necessary and firms neco in coal gasification) may enjoy synergies witb can add substantial value tbrougb upstream or down- tbdr other businesses. Being broadly integrated also stream activities, taper integration can be used effec- offers tbem opportunities to capture large profit tively, as American Cyanamid did in etbical pbar- margins by adding more vaiue themsdves. Firms that maceuticals by supplying basic and finisbedcbemical s engage in several vertical stages for each integrated to its Lederle Laboratories subsidiary wben it was activity tbey undertake (such as Texas Instmments convenient to do so but relying on outsiders to sut^ly in miaocomputers and MOIHI in p^rcrieum refining) chemicals in otber cases. Similarly Amoco (Standard can enjoy these synergies at sevoal diverse levds Oil of Indiana) and many otber petroleum refiners within thdr organizations. They also can control found tbat upstream taper int^ration arrangemoits cmdid aspects of tbdr produds' quality by par- 643 tidpatii% at several stages in tbe ebain of process- investments in tbem and ovQ-come customers' reluc- ing. Ahbough it serans evident tbat integrating broad tance to adopt new products (or new generations of and long rangii% operaticms could be complex and products). Tecbnologic^ iimovation is a major cause costly, it is bypotbesized tbat in some cases it could of accderated industry evolution and of increased de- be bigbly rewarding because sucb strategies lever:^ tnand uncertainty. Different vertical integration firms' abilities to enter new markets, exploit new strategies will be more appropriate if tecbnology tecbtiolo^es, and evaluate tbe impact of evolutionary cbanges rapidly (or slowly), depending on wbetber changes faster. firms would be tecbnological leaders or followers. Particular cinnbitiations of the breadth of activities Pioneering firms would be more likely to integrate integrated, number of stages undertaken, degree of tban would tecbnological followers. Witb tbis excep- intemal transfers, and form of ownership control are tion, bowever, less vertical integration is expected more likdy to be successful for certain firms than early (and late) in an industry's evolution in contrast others, depending on: (1) the uncertainty surround- witb tbe scenario Stigler (1951) envisioned, because ing sales growth, industry infrastmcture, and other of (1) tbe risks of demand uncertainty and (2) dif- market traits; (2) tbe likelibood tbat tbe industry in fering needs to prove a new product's wortb. Tbere- question will undergo radical tecbnological cbange, fore, tbe most likely pattern of integration bebavior severe price warfare, or otber stmctural cbanges caus- one migbt expect to see overtime (bolding otber fac- ing competition to be volatile; (3) tbe power of firms tors constant) is an inverted U-sbape. to bargain, cajole, or pressure suppliers (or distribu- tors) into performing value adding tasks for tbem; Demand Uncertainty. Wben demand conditions and (4) firms' strategy needs. (For example, more in- become stable, bigber degrees of intemal integration tegration migbt be used if tbe firm's CEO concluded migbt be undertaken witb ease because one firm's that it needed greater control over adjacent parties sales volumes can become large (and regular) enough to attain tedmological leadership or other objectives.) to absorb the output of adjacent plants without in- It is important to nctfe tbat tbe most appropriate ver- curring costly excess capadty penalties. Because it tical integration strate^es will cbange over time as would take time for the experience spurring sales industry conditions cbange, as corporate strategy growth to occur, one would not expect firms within needs cbange, and as firms' capabUities evolve. Tbe many embryonic industries (as well as declining in- CEO must assess tbe rdative wortb of tbe strategy dustries) to be broadly integrated or engaged in many altematives sketcbed above in ligbt of tbe forces tbat stages of integrated opo^ons. Demand for products key factors exert on tbe dimensions of vertical in embryonic and declining settings may be highly integration. uncertain. The chief strategist may elect for tbe firm to undertake more activities or more integrated stages Factors Affecting Vertical Strategies in settings in wbicb pioneering investments are nec- essary to acbieve otber corporate objectives, sucb as Four key factors are bypotbesized to affect tbe ver- tbe creation of infrastmdures, particularly if existing tical integration strategies tbat firms embrace: distribution cbannds are blocked or inappropriate for (1) forces propelling industry evolution and exacer- tbe firm's needs. bating demand uncertainty; (2) tbe nature of com- Creating Credibility for New Industries. Vertical petition in the linked industries; (3.) the bargaining integration bebaviors in new industries would be ex- power of suppliers or distributors (and customers); pected to differ from those in established industries and (4) corporate strategy requirements. Table 2 and to differ from behaviors within raibryonic indus- details the effects of these factors on the dimensions tries that began in the previous century. There were comprising vertical strategy altematives. From this significant differences in the need for infrastmc- [ffesentation and the discussion bdow, certain com- tures—cbamiels of distribution, standard means of binations of these dimensions are shown to be more assessing quality, and so on—supporting tbe devdop- apptopnate tban otbers wben tbe key factors occur ment of tbe embryonic steel, automobile, and tobac- togetber in certain ways. co industries of tbe last century compared witb tbose surrounding tbe embryotiic industries of tbe 1980s. Fwces Propeili^ bdnsbr Evolutioii In newly devdoping countries and earlier in tbe Induaries evdve in Aructure as firms make diverse development of U.S. business, it frequently was 644 Table 2 An niustration of the Strategic Framework Strategy Dining Characteristics When Appropriate Advantages Risks Degree of integration Full All of a particular in- In mature aiut stable envirtm- Superior ctmtrol of ecwiomic Decrease market power integration put (or output) is trans- ments protect proprietary pro- environment Technolc^ does not provitfe ferred in-bouse to a sister cesses from es[Honage Avoid foreclosure to inputs many integration eccmcHnies business unit. Often ad- Maintain tight quality control or markets Asset inflexibility jacent, integrated stages at all stages Guarantee quality is consis- Price renegotiati(His difficult are in balance in their If Hrm seeks technological or tent with corporate image Minimum efficient scale throughput volumes. quality leader^ship position Capture integration econo- plants are rarely the same up- Diseconomies from imbal- mies (espedally advantageous if stream and downsueam. tluts ance are not significant market share leader) scnne portion of linkage is like- Technological changes occur ly to be out of balance slowly Tapered Some portion (but not Physical interconnection un- Enables firm to moiutor out- Subcontractors will not be integration all) of Hrm's require- necessary side R&D and mark^ing prac- available to abso-b fluctuatirais ments fcM- an input is sup- Diseconomies in minimum tices. Could incorporate out- in production and demand plied in-house or some efficient scale [dant that are un- siders' innovations while gain- Access to best suppliers (or portion of outputs is sdd derutilized are substantial ing capability in-house, as well distributors) will be cut off (consumed) in-house. If firm seeks technological, Enables firrn to understand (competitive ftHcdosure) quality, or market share leader- suppliers' (or distributors') cost Pay prnniums ftxc access to ship in volatile competitive set- structures and profit margins outsiders* goods and services tings Increases bargaining power Lower pri(»ity as a cust(nn« New products needing expla- (implied threat of full integra- (or vendor), because outsido-s nations or infrastructure no out- tion) are overflow outlets primarily siders provide Same as full integration with Same as full inte^iuion but If firm desires to prod in- less risk more advantagMus house units Nonintegra- No internal transfers Better quaiity/prices avaii- No penalties from underuti- Quality contrtri may ncH be as tion of inputs (or outputs) to able from outsiders lized plants high and market power of firm in-house sister units Costs of investitig to make Avoids purchasing highly may be unable to exert control components is too formidable specific assets (av(nds inflexibil- over adjacent Hrms throu^ or volume consumed is too iow ity) contract (and selling to outsiders would Avoids purchasing large ca- Subcontractor will not be be difficult or a serious diversi- pacity when firm's needs are avEulable to pn-form needed fication from core business of small tasks firm) Lowers overhead and break- Firm loses cost advantages of If technology is changing even points integration economies (where rapidly Allows preplanned delivery these exist) If price competition (and schedules to reduce inventory- other competitive tactics) causes ing costs (kanban) market shares to fluctuate wild- ly I f demand is highly uncertain Breaith of integration Broadly Many activities (in- Product differentiability is Maintains intelligence con- No scale Mon(»nies available integrated puts, services, or chiui- high cerning component costs and at the small volumes needed for nels of distributiOT or No outsiders yet produce ways to streamline product de- in-house (and market) COT- consumption) related to goods or services needed sign (experience curve) sumption the needs of a particular Industry structures beccnning Maintains product quality Cwtly setup costs associated business tmit are en^ig^ established and economies be- and design secrecy with short production runs for in. (Broadly integrated c 645 Table 2 (continued) Strategy Defining Characteristics When Appropriate Advantages Risks Stages of integrated activity Many sU^es Firm is eng^ed in When product (or compo- Captures more of value add- Synergies foregone if com- many vertically related nents) are state of the art and ed in veitical chain of process- munications systems do not ex- activities—from ultra- firm seeks technological leader- ing ploit these linkages welt raw materials to distribu- shihip Firm can create substantial Creates exit barriers due to titMi to final ccHisumers— When life of product is ex- improvements in supplying obsolescence in which the buyer-seller pected to be eight years or technology or distribution prac- Risks of throughput imbal- relationships could longer tices ance magnified for each st^e occur. V/hen firm is foreclosed by Could create highly differen- added to vertical chain competitors tiated and high quality products Subcontractors needed to al- When firm seeks quality lead- Preanpt nonintegrated com- leviate imbalances in through- ership petitors by forcing industry put (due to technology or structures to evolve to firm's ad- changing demand) will not be vantage available Control proprietary advan- Costly and inefficient if firms tages do not manage complexity well Reach ultimate consumers Involves firms in very diverse more effectively activities that may be far from its core strengths Few stages Firm is engaged in few When technological or cus- Outsiders' itinovations can be Firm's product perceived as vertKaliy relatKl activi- tomer scanning is less important harnessed to improve product a "me-too" entry ties in the chain from When product lives are ex- quality or design Integrated competitors will ultra-raw materials to pected to be short or the new Firm can piggyback on the enjoy superior cost structures distribution obsolescing technology will be marketing or brand image ex~ and intelligence very unlike past processes penditures of outsiders When product is very young or industry structure is embry- onic When demand is declining When physical interconnec- tions are few or not necessary Form of ownership Wholly- Businesses are wholly- When contracts or other Proceeds from irivestment Same risks as being too inte- owned owned by firm. quasi-integrated forms of con- need not be shared with others grated on other strategy dimen- trol are inadequate Greater strategic flexibility sions above; risk exposure is When firm desires to protect because business units are fully maximum technology or trade secrets from owned and controUed outsiders When demand is declining (except as a "phase-out" tactic) iu^i- Less than full owner- If risk of failure and invest- Reduces asset investments Could create mobility or exit integration ship and control. Couid ment costs are high Allows firm to create barriers if partnws are impor- include joint ventures, If economies of integration "spider's web" of quasi-inte- tant to other businesses of the franchises, minority are insignificant but need for grations to sam^de several firms' firm equity investments, loan control is strong approaches to a technological or Costs of managing quasi-in- guarantees, or an "un- Technology changes rapidly nuu'keting problem tegrated rela,tionship exceed derstanding" regarding Industry structures are still Creates bargaining power benefits of this control system customer relationships embryonic over adjacent business units Contractual problems could To transfer ownership of an ln^)roves access to new mate- stymie strategic flexibility and undesirable business unit to out- rials or processes run up administration costs siders Scanning advantages of taper If corporate mission does not integration with less asset risk require tight control over qual- Integration economies of ity lapw integratiori (provided firm If competition is vt^atile manages quasi-integrated rela- tionship effectively) Lowers fixed costs while pro- viding access to adjacent firms' intelligence and skills net^^sary for firms to undertake many stages of in- once integrated vertically to build them. The major tcgx&ted activities (and to provide the necessary in- reason for pioneering firms within embryonic indus- frastructures) in order to help an industry to develop. tries to undertake many stages of integrated actiwtiy But ROW nmny new industries can use the same in- now would be to create credibility for a radical new frastructures devdoped in an earlier era by firms that industry. This was once the case in persuading textile 646 firms to use rayon as well as cotton and wool on their tures have been developed in detail elsewhere (Har- looms; and Celanese forward integrated from rayon rigan, 1980; Porter, 1980). The elements of industry to yarn, textiles, and garment manufacture to prove structure that affect vertical int^ration include: pro- to consumers, as well as textile firms, that its new duct traits, supplier traits, consumer traits, manufac- fiber (rayon acetate) was viable. Similarly ALCOA turing technology traits, and competitor traits. once forward integrated beyond its current number Product Traits. Because differentiated products of integrated stages to fabricate aluminum products can justify higher prices (Bain, 1968), higher degrees to sell to consumers when other metals fabricators of intemal transfer can be undertaken for such prod- would not use its new metal. In the 1980s, however, ucts with greater success even when integration there seem to be few new industries (except, perhaps, economies are not substantial. In choosing which genetic engineering) for which this need to create new components or services to produce in-house, how- market conduits is as high as it was in earlier eras. ever, it is important to understand which attributes When the risks of launching an embryonic industry of a product create those qualities for which con- are quite high, firms can form joint ventures to link sumers are willing to pay a premium. Noncritical supplies and distribution charmels; and when demand components and services (and those offering poor conditions become stable and certain patterns of economics) could be purchased from outsiders and competition are recognized as being more successful sensitive components and services (and those offer- than others, more internal integration can be safely ing the best economics) produced in-house. By free- undertaken. When demand is declining, however, the ing plant space and resources that formerly were first linkage that firms in declining industries might devoted to noncritical and uneconomic components be expected to sever are those with owned distribu- and services, firms can undertake a more profitable tion channels, because making demand dependent on mix of activities with their resources while tying up an independent market enables firms to assess more the assets of outsiders for low profit activities. clearly whether pockets of enduring demand exist for If trade secrets protect some aspect of a firm's the product in question (Harrigan, 1980). This is what products, higher degrees of integration are necessary, Celanese did in acetate, IHamond Shamrock did in as in the case of Polaroid, which stopped purchas- acetylene, and Brown Shoe did in leather tanning. ing its negative materials from Kodak when its ins- In summary, except for the unique situations men- tant photography patent expired. (Too much pro- tioned above, specialized suppliers or distributors are prietary information was contained there to let com- better suited to provide goods and services to firms petitors produce it.) Similarly, Schlumberger ac- in embryonic industries on a contractual basis until quired its own custom logic semiconductor house uncertainties concerning demand and competitive (Fairchild Camera & Instrument) to protect its pro- viability are resolved. prietary knowledge conceming well-logging services, and Dow Chemical often is fully integrated to pre- vent other firms from learning too much about its Volatility of Competition processes and designs. Supplier Traits. The principal motives for firms to Individual business units would not be expected to integrate backward often include capturing high pro- favor vertically integrated strategies in settings in portions of value added, controlling product quality which industry structures exacerbate the likelihood (and proprietary knowledge), or overcoming compe- of price warfare and depressed profit margins. Be- titors' advantages if the best suppliers are already cause volatile industry structures increase the under contract to others. If competition is escalating likelihood that competition will degenerate into the on the basis of innovations, however, firms should use of tactics that devastate long term profitability be wary about embracging high degrees of intemal and sap the irmovative resourcefulness of firms, ver- transfers because they cut off their access to the tical integration generally should be avoided in such benefits of outsidtfs' innovations in an enviroiunoit settings. Other things held constant, the greatest in which flexibility is crucial to competitive abiUty. number of successful linked stages and the greatest In such settings, it may be desirable to hdp create successful breadths of integrated activities would be another new supplier (through quasi-integration ar- expected within settings in which competition is rangements, which allow the new entity to serve stable. The characteristics of hostile industry struc- others as well as sponsoring taper integratioa firm) 647 rather that fall into the trap of technological inflexi- perfonn key intermediate processing steps reduces the bility by fuUy owning such suppliers and buying in- likelihood that firms will be stuck with vertical puts only from than. resources that become high exit barriers. As long as Consumer Traits. The principal reasons to in- demand is increasing and firms can avoid price wars tegrate forward often indude capturing high propor- in selling thdr output, vertical linkages will not ex- tions of value added, controlling the quality and im- acerbate the tendency for competition to become vo- 3^ of one's products, and raising customers' switch- latile; but unless spedal efforts are made to overcome ing cost barriers (Porter, 1980). Compiex products these forces, vertical linkages can become high exit that require substantial demonstration or explana- barriers as industries mature (Harrigan, 1983b). tions and servicing (as microcomputers did in 1978) Competitor Traits. Efforts made to diminish the are strong candidates for downstream linkages. In pressures of other structural traits toward price war- doing so, firms must be cautious to ensure: (1) that fare can be done by competitors who (1) compete on they are not overly dependent on in-house merchan- the basis of price or (2) use vertical integration as a dise for resale and (2) that they do not stay forward means of foreclosure. "Dominant verticals," the integrated after the advantages of being so integrated group of narrowly diversified, integrated firms that have expired. When products become successful Rumelt (1974) identified, are most likely to possess enough to create high customer switching costs, the the types of strong commitments to vertical integra- need for forward integration is lowered, and firms tion strategies that function like exit barriers, caus- should move away from battlefronts unless they are ing them to act irrationally, by cutting prices to niain- well positioned to win price wars. tain high throughputs in thdr integrated fadlities to Manufacturing Technology Traits. The technology the detriment of other competitors. used in manufacturing must offer substantial integra- Firms that use their vertical linkages as a means tion economies in order for vertical integration to be of foreclosing nonintegrated firms from access to advantageous (Khandwalla, 1974). Because the mini- materials, markets, innovations, and competitive in- mum efficient scale of some technologies is so much telligence also are damaging competitors because they larger than firms' needs for that component, firms can escalate the evolution of the industry towards that integrate such activities may be forced to enter defensive vertical integrations. When many firms merchant sales to dispose of their unused outputs have integrated, all face similar pressures to keep from the oversized plant (or run the plant at uneco- thdr vertical chains efficiently utilized, and price nomic volumes). Integration should be avoided if the competition becomes more likely than if noninte- costs of excess capacity caimot be offset by charg- grated firms were allowed to supply or purchase ex- ing premium prices. Instead, firms should use sub- cess volumes of tnaterials and services to alleviate im- contractors to perform those tasks that require assets balances in vertically related technologies. Thus it that most firms wouid use infrequently at present may be preferable for an industry to have some non- operating scales. Thus, producers of solar collector integrated firms to absorb other firms' excess capac- panels send out for chrome plating, and ethical ity, lest industry bloodshed result instead. pharmaceutical firms send out for the bromine In summary, high degrees of internal transfer, long chemistry step in production. vertical chains, and many integrated activities are ex- Firms must keep aware of their distinctive compe- pected in settings in which industry structures do not tences in production to ensure that: (1) their critically exacerbate price warfare or rapid rates of change in skilled laborers—sdentific researchers and en- products or processes. Because competitors must be gineers—are employed Qest they be hired away by able to change tactics rapidly in turbulent industry ctmpetitors), and (2) they avoid being stuck with ob- settings, a highly integrated posture in such settings solete assets. If firms are constrained by plant space, could reduce a firm's maneuverability and damage it may be wise to purchase simple or low volume com- its profitability. Even the partial reprieve of substan- ponoits from outsiders. Critically skilled onployees ti^ integration economies or the ability to charge thereby can be kept busy on difficult (but challeng- premium prices to pay for costly idle capadty may ing} tasks that lev^age firms' future capabilities to not offset the long tom corporate damage that being cmnpMe, ud the burden of thdr salaries am be too highly integrated in such settings could create. qxead ova high vduiae supidying activities. Finally, The tradeoffs that firms will make depend on their if technology is changing raindly, using outsiders to overall strategies and market power. 648 could be used in several dissimilar i Firms that possess the bargaining power needed to voticai iitt^ration strategies that increase or enhaace obtain secure access to suppliers and distribution innovations by sharing technological infonnation channels without damaging thdr strategic flexibility common to separate stages of int^raticm may requre could reduce thdr asset exposure and inflexibility by more int^ration than the framework suggests. redudng ownership stakes in supplying or distribut- The number of stages undertaken, for example, ing business units. Another way is to use the firm's would be expected to be highest if significant bargaining power to persuade sequent businesses to synergies are gained or other corporate needs are assume the duties that a firm wishes to avoid (Mac- served. The key determinants of whether (or not) a Millan, Hambrick, & Pennings, 1982). firm should skip a particular stage in its int^rated The most important determinants of bargaining chain of activities are the task's importance to its cor- power are: (1) product specifidty to the industry in porate mission and the quality of goods or services question; (2) existence of alternative outlets or provided by outsiders. A firm's position within its sources of suppliers; (3) ability to self-manufacture industry also suggests how many integrated stages it the good/service in question; and (4) dependence of would perform, and firms on the fringes of an in- the supplier (or distributor) on the business unit dustry would be more likdy to purchase (rather than (Porter, 1976a). If firms possess the power to leverage produce) materials or services from leader firms thdr market positions, they could use this power to whose upstream plants produce in excess of their control adjacent firms' assets without owning them. downstream plants' capadties. Firms that control brand names or patents, for ex- Although firms will vary considerably in which ample, could hire outsiders to market thdr products tasks they choose to do in-house, thdr needs to cap- if communications with downstream parties are less ture more value added would mean that they per- important than other advantages that such arrange- formed in-house those tasks for which their expen- ments might provide. But if firms' suppliers or dis- sive and criticai resources were best suited. Firms also tributors possess bargaining power (and if they can- will integrate those tasks that woidd enable them to not be persuaded to perfonn useful services for firms) enjoy synergies with other business units, those that then those activities may have to be undertaken in- are important to thdr business missions, or that of- house in order to attain the control that some firms fer high profit margins for them. They most likely desire. If this situation occurs in settings in which de- would own outright those aspects of thdr businesses mand is unstable and competition is volatile, the out- that were most important to them, and they would come of doing so could be disastrous. pool the risks (or extend thdr control over adjacent firms) for less critical activities supplied by outsiders Corporate Strategy Needs through quasi-integration arrangements. The most scarce resource that firms possess is their The foregoing arguments that less integration is entrepreneurial ability. Rather than sedng thdr mix preferable to more must be moderated by considera- of businesses as streams of cash fiows, diief execu- tions of corporate strategy needs. Because vertical in- tives should consider them as reservoirs of capabili- tegration can be costly if used imprudently, corporate ties. Thus vertical integration strategies would en- strategists must scrutinize the advantages they hope courage activities and rdationships for which person- to capture by condoning (or denying) the creation of nel with crudal skills (or other scarce capabilities) certain vertical rdationships. Votical integration may might otherwise not be retained. be part of a larger strategy involving shared resources In previous sections, less vertical integration has and experience curve economies for some businesses, been antidpated within turbulent settings, particidar- for example, requiring the firm to sustain relation- ly those in which techncdc^cal duutge occurs rapi(By. ships that the strategy framework otherwise would Firms pursuing techndogical leadersUp strate^es of- not recommend. fer an important exception to this hypoth^s, how- Some vertical integrations promise to improve long ever, because they often m:e wiBing to endue the ton- term synergies for the entire firm, although they ap- porary imbalances of full integration whm {«oduc- pear to penalize a particular business unit. Supply ing sensitive components, and thef are wining to M>- side economies, for example, could be gained by sorb the risks of many int^-ated stages (as detailed sharing manufacturing fadlities for components that in Table 2) in order to be prased References Abell, D. F. Strategic windows. Journat of Marketing, 1978,42(3), Chandler, A. D. The visibte hand: The manageriat rvsotution in 21-26. American business. Cambridge, Mass.: Harvard Univenity Press. 1977. Adams, W.. & Dirlam, J. Steel imports and the vertical oligopoly power. American Economic Review, 1964, 54, 640-655. Qevenger, T. C, Campbell, G. R. Vertical organization: A neglected element in market structure-performance models. In- Adelman, M. A. Integration and antitrust policy. Harvard Law dustriat Organizationat Review, 1977, 5(1), 259-262. Review, 1979, 63(1), 27-77. Bain, J. S. Industriat organization. 2nd ed. New York: Wiley, 1968. Comanor. W. Vertical mergers, market power, and antitrust laws. American Economic Review, 1967, 57, 259-262. Baker, H. K., MiUer, T. O., & Ramsperger, B. J. An inside look at corporate mergers and acquisition. MSU Business Topics, Dennison, S. R. Vertical integration and the iron and steel industry. 1981, 29(1), 49-57. Economic Joumat, 1939, 49, 244-258. Frank, L. K. The significance of industrial integration. Joumat Blair, R., & Kaserman, O. Vertical integration, trying and antitrust of Political Economy, 1925, 33, 179-195. policy. American Economic Review, 1978, 68, 397-402. Greenhut, M. L., & Ohta, H. Vertical integration of sticcessive Blois, K. J. Vertical quasi-integration. Journat of Industriat oligopolist. American Economic Review, 1979, 69,137-141. Economics, 1972, 20, 253-272. Harrigan, K. R. Strategy formulation in decKmng industries. Bork, R. Vertical integration and the Sherman act: The 1^ history Academy of Mangement Review, 1980, 5, 599-604. of an economic misconception. University of Chicago Law Harrigan, K. R. Deterrents to divestiture. Acaderrty of Manage- Review, 1954, 22(1), 158-201. ment Joumat, 1981, 24, 306-323. 651 Hltrigaa, K. R. Strategies for verticat integration. Lexington, McGee, 3.S.,& Bassett, L. R. Vertical imegration revisited. Joa^ Man.: Heath, l»83a. nat of Law and Economics, 1976, 19, 17-38. Hkrrigaii, K. R. Exit barriers and vertical integration. Academy Ohmae, K. The mind of the strategist. New York: McGraw-Hill, <^ Management Proceedings, 1983b, 32-36. 1982. Hannan, K. R. Strategies for Joint ventures, forthcoming. Perry, M. K. Forward integration by ALCOA: 1888-1930. Jour- Jewkes, J. Factors in industrial integration. Quarterty Joumat of nal of Industriat Economics, 1980, 29, 37-53. Economics, 1930, 44, 621-638. Porter, M. E. Intertmmd choice, strategy, and bilateral market KhandwaUa, P. N. Mass output orientation of operations power. Cambridge, Mass.: Harvard University Press, 1976a. tectmotogy and organizational structure. Administrative Science Porter, M. E. Please note location of nearest exit: Exit barrien Qtiarterly, 1974, 19, 74-79. and strategic and organizational planning. Catifornia Manage- Larson, D. A. An empirical test of the relationship between market ment Review, 1976b, 19(2), 21-33. concentration and vertical integration. Industriat Organization Porter, M. E. Competitive strategy: Technotogies for analyzing Review, 1978, 6(1), 71-74. industries and competitors. New York: Free Press, 1980. Lavington, F. Technical influences on vertical integration. Economica, 1925, 7(19), 27-36. Rumelt, R. P. Strategy, structure and economic performance. Lubatkin, M. H. >4 market modet anatysis of diversification Boston: Division of Research, Graduate School of Business Ad- strategies and administrative experience on the performance of ministration, Harvard University, 1974. merging firms. Unpublished doctoral dissertation. University of Tennessee, Knoxville, 1982. Salter, M. S., & Weinhold, W. A. Diversification through acquisi- tion: Strategies for creating economic value. New York: Free MacMiUan, I. C. Preemptive strategies. Journat of Btjstness Press, 1979. Strategy, 1983, 4(2), 16-26. MaeMiUan, I. C, Hambrick, D. C,&Pennings, J. H. Backward Sichel, W. Vertical integration as a dynamic industry coticept. Anti- vertical integration and interorganizational dependence—A trust Buttetin, 1973, 18, 463-482. FIMS-based analysis of strategic business units. Working paper, Stigler, G. J. The division of labor is litnited by the extent of the Columbia University Business School, 1982. market. Joumat of Potiticat Economy, 1951, 59, 185-193. Maticke, R. B. Iron ore and steel: A case study of the economic Williamson, O. F. Markets and hierarchies. New York: Free Press, causes of vertical integration. Journat of Industriat Economics, 1975. 1972, M, 220-229. Kathryn Rudie Harrigan is Associate Professor of Manage- ment ofOrganizations in the Graduate Schoot of Business, Cotumbia University. 652