The Wheel of Retailing• STANLEY C. HOLLANDER

New types of retailing frequently start off with crude facilities, littleprestige, and a reputation for cutting prices and margins. As theymature, they often acquire more expensive buildings, provide moreelaborate services, impose higher margins, and become vulnerable tonew competition.

The author examines the history of numerous retail institutions todetermine if this process really constitutes a "natural law of retailing."

E WHEEL of retailing" is thename Professor Malcolm P. McNair

has suggested for a major hypothesis con-cerning patterns of retail development.This hypothesis holds that new types ofretailers usually enter the market as low-status, low-margin, low-price operators.Gradually they acquire more elaborateestablishments and facilities, with bothincreased investments and higher operat-ing costs. Finally they mature as high-cost, high-price merchants, vulnerable tonewer types who, in turn, go throughthe same pattern. Department-store mer-chants, who originally appeared as vigor-ous competitors to the smaller retailersand who have now become vulnerable todiscount house and supermarket compe-tition, are often cited as prime examples ofthe wheel pattern.^

Many examples of conformity to thispattern can be found. Nevertheless, we

• ABOUT THE AUTHOR. Stanley O. Hollander(Ph.D.. University of Pennsylvania) is Professor ofBusiness Administration at Michigan State Univer-sity. He is editor of a recent book of readings inretailing theory, "Explorations in Retailing" (Bureauof Business Research, Michigan State University,1959], and was a panelist on retail pricing beforethe Joint Economic Commit+ee, U. S. Congress, in1958.

His previous wort in marketing history has in-cluded several studies of d'iscount retailing, oneof which M.S.U. recently published as "The Riseand Fall of a Buying Club." He compiled theAmerican Marketing Association's "Bibliography onDiscount Selling" (1956] and has published articlesand monographs on other aspects of retailing andmarketing.

The author is indebted to the participants in the1959 Marketing Theory Seminar at Boulder, Colorado,for many penetrating comments on an earlier draftof this paper.

may ask: (1) Is this hypothesis validfor all retailing under all conditions?(2) How accurately does it describe totalAmerican retail development? (3J Whatfactors cause wheel-pattern changes inretail institutions?

The following discussion assembles someof the slender empirical evidence availablethat might shed some light on these threequestions. In attempting to answer thethird question, a number of hypothesesshould be considered that marketing stu-dents have advanced concerning the forcesthat have shaped retail development.

TENTATIVE EXPLANATIONS OF THE WHEEL

(A) Retail Personalities. New types ofretail institutions are often established byhighly aggressive, cost-conscious entrepre-neurs who make every penny count andwho have no interest in unprofitable frills.But, as P. D. Converse has suggested,these men may relax their vigilance andcontrol over costs as they acquire ageand wealth. Their successors may be lesscompetent. Either the innovators or theirsuccessors may be unwilling, or unable,to adjust to changing conditions. Conse-quently, according to this view, deteriora-tion in management causes movementalong the wheel.̂

^ M. P. McNair, "Significant Trends andDevelopments in the Postwar Period," inA. B. Smith (editor). Competitive Distribu-tion in a Free, High-Level Economy and ItsImplications for the University (Pittsburgh:University of Pittsburgh Press, 1958), pp.1-25 at pp. 17-18.

' P. D. Converse, "Mediocrity in Retailing,"JOURNAL OF MARKETING, Vol. 23 (April,1959), pp. 419-420.

37

38 JOURNAL OF MARKETING July 1960

(B) Misguidance. Hermann Levy hasadvanced the ingenious, if implausible,explanation that retail trade journals,seduced by profitable advertising from thestore equipment and supply industry, coaxmerchants into superfluous "moderniza-tion" and into the installation of overlyelaborate facilities.^

(C) Imperfect Competition. Althoughretail trade is often cited as the one typeof business that approaches the AdamSmith concept of perfect competition, someeconomists have argued that retailing ac-tually is a good example of imperfect com-petition. These economists believe thatmost retailers avoid direct price competi-tion because of several forces, including re-sale price maintenance, trade associationrules in some countries, and, most impor-tant, the fear of immediate retaliation.Contrariwise, the same retailers feel thatservice improvements, including improve-ments in location, are not susceptible todirect retaliation by competitors. Hence,through a ratchet process, merchants inany established branch of trade tend toprovide increasingly elaborate services atincreasingly higher margins.*

(D) Excess Capacity. McNair attrib-utes much of the wheel effect to the devel-opment of excess capacity, as more andmore dealers enter any branch of retailtrade." This hypothesis rests upon an im-perfect competition assumption, since, un-der perfect competition excess capacitywould simply reduce margins until the ex-cess vendors were eliminated.

(E) Secular Trend. J. B. Jefferys haspointed out that a general, but uneven,long-run increase in the British standardof living provided established merchantswith profitable opportunities for tradingup. Jefferys thus credits adjustments tochanging and wealthier market segmentsas causing some movement along thewheel. At the same time, pockets of oppor-tunity have remained for new, low-margin

' Hermann Levy, The Shops of Britain(London: Kegan Paul, Trench, Trubner &Co., 1947), pp. 210-211.

* D. L. Shawver, The Development of The-ories of Retail Price Determination, (Ur-bana; University of Illinois Press, 1966),p. 92.

•̂ Same reference as footnote 1.

operations because of the uneven distribu-tion of living-standard increases.^

(F) Illusion. Professor B. Holdren hassuggested in a recent letter that presenttendencies toward scrambled merchandis-ing may create totally illusory impressionsof the wheel phenomenon. Store-wideaverage margins may increase as new,high-markup lines are added to the prod-uct mix, even though the margins chargedon the original components of that mixremain unchanged.

DIFFICULTIES OF ANALYSIS

An examination of the actual develop-ment of retail institutions here and abroaddoes shed some light on both the wheelhypothesis and its various explanations.However, a number of significant diffi-culties hinder the process.

(1) Statements concerning changes inretail margins and expenses are the cen-tral core of the wheel hypothesis. Yetvalid information on historical retail ex-pense rates is very scarce. Long-runchanges in percentage margins probablydo furnish fairly reliable clues to expensechanges, but this is not true over shortor intermediate periods. For example,1957 furniture-store expense rates wereabout 5 percentage points higher thantheir 1949-1951 average, yet gross mar-gins actually declined slightly over thesame period.'

(2) Historical margin data are some-what more plentiful, but these also haveto be dredged up from fragmentarysources.^

(3) Available series on both expensesand margins merely note changes in re-tailers' outlays and receipts. Th«y do notindicate what caused those changes andthey do not report changes in the costs

•J. B. Jefferys, Retail Trading in GreatBritain, 1850-1950 (Cambridge: CambridgeUniversity Press, 1954), various pages, es-pecially p. 96.

' Cited in Fabian Linden, "DepartmentStore Operations," Conference Board Bust-ness Record, Vol. 14 (October, 1958), pp.410-414, at p. 411.

* See Harold Barger, Distribution's Placein the American Economy Since 1869(Princeton: Princeton University Press,1955).

THE WHEEL OF RETAILING 39

borne by suppliers, consumers, or the com-munity at large,

(4) Margin data are usually publishedas averages that may, and frequently do,mask highly divergent tendencies.

(5) A conceptual difficulty presents aneven more serious problem than the pau-city of statistics. When we talk about"types" of retailers, we think of classi-fications based upon ways of doing busi-ness and upon differences in price policy.Yet census categories and other systemsfor reporting retail statistics are usuallybased upon major differences in commod-ity lines. For example, the "pineboard"druggists who appeared in the 1930s area "type" of retailing for our purposes.Those dealers had cruder fixtures, chargedlower prices, carried smaller assortments,gave more attention to turnover, and hadless interest in prescriptions than did con-ventional druggists. Yet census reportsfor drugstores necessarily included all ofthe pineboards that maintained any sortof prescription department.

Discount houses provide another ex-ample of an important, but amorphous,category not reflected in census classifi-cations. The label "discount house" coversa variety of retailers. Some carry stocks,others do not. Some have conventionalstore facilities, whereas others operate inoffice buildings, lofts, and warehouses.Some feature electrical appliances andhard goods, while others emphasize softgoods. Some pose as wholesalers, andothers are practically indistinguishablefrom all other popular priced retailers intheir fields. Consequently discount dealers'operating figures are likely to be mergedinto the statistics reported for other ap-pliance, hardware, or apparel merchants.

EXAMPLES OF CONFORMITYBritish

British retailing provides several ex-amples of conformity to the wheel pattern.The grocery trade has gone through sev-eral wheel-like evolutions, according to adetailed analysis made by F. G. Pennanceand B. S. Yamey." Established firms did

"F. G. Pennance and B. S. Yamey, "Com-petition in the Retail Grocery Trade, 1850-1939," Economica, Vol. 22 (March, 1955),pp. 303-317.

initiate some changes and some marginreductions, so that the pattern is ob-scured by many cross currents. But themajor changes seem to have been dueto the appearance and then the matura-tion, first, of department-store food coun-ters; then, of chain stores; and finally,of cut-price cash-and-carry stores. Nowsupermarkets seem to be carrying thepattern through another evolution.̂ o

Jefferys also has noted a general long-run upgrading in both British depart-ment stores and chains.̂ ^ Vague com-plaints in the co-operative press and adecline in consumer dividend rates sug-gest that wheel-like changes may haveoccurred in the British co-operative move-ment. ̂ ^

American

Very little is known about retail mar-gins in this country before the Civil War.Our early retail history seems to haveinvolved the appearance, first, of hawkers,walkers, and peddlers; then, of generalstores; next, of specialty stores; and fi-nally, of department stores. Each of thesetypes apparently came in as a lower-margin, lower-price competitor to the es-tablished outlets, and thus was consistentwith the wheel pattern. We do not know,however, whether there was simply a long-run decline in retail margins throughsuccessive improvements in retail effici-ency from one type to another (contraryto the wheel pattern), or whether eachof the early types was started on a low-margin basis, gradually "up-graded," andso provided room for the next entrant(in accordance with the pattern).

The trends toward increasing marginscan be more easily discerned in manybranches of retailing after the Civil War.Barger has described increases over theyears 1869-1947 among important retail

"• "La Methode Americaine," Time, VoL 74(November 16, 1959), pp. 105-106.

" Same reference as footnote 6.""Battle of the Dividend," Co-operative

Review, Vol. 86 (August, 1956), p. 183;"Independent Commission's Eeport," Co-operative Review, Vol. 38 (April, 1958), pp.84-89; "£52 Million Dividend in 1957," Co-operative Review (August, 1958), pp. 171-172.

40 JOURNAL OF MARKETING July 1960

segments, including department stores,mail-order firms, variety stores, andjewelry dealers. He attributes much ofthe pre-World War I rise in department-store margins to the ahsorption of whole-saling functions. Changes in merchandisemix, such as the addition of soda foun-tains and cafeterias to variety stores andthe upgrading of mail-order merchandise,seem to have caused some of the otherincreases. Finally, he believes changes incustomer services have been a major forcein raising margins.'^ Fabian Linden hasextended Barger's observations to notesimilar 1949-1957 margin increases fordepartment stores, variety chains, andappliance dealers.^^

Some other examples of at least partialconformity to the wheel pattern may becited. Many observers feel that bothdiscount-house services and margins haveincreased substantially in recent years.̂ -̂ ^One major discount-house operator hasstated that he has been able to keep hisaverage markup below 12%, in spite ofconsiderable expansion in his facilitiesand commodity mix.'*! However, the con-census seems to be that this probably isan exception to the general rule.

A study of gasoline pricing has pointedout how many of the so-called "off-brand"outlets have changed from the "traekside"stations of pre-war days. The traeksidedealers typically maintained unattractiveand poorly equipped installations, at out-of-the-way locations where unbrandedgasoline was sold on a price basis. Todaymany of them sell well-promoted regionaland local brands, maintain attractive, effi-cient stations, and provide prompt andcourteous service. Some still offer cutprices, but may have raised their pricesand margins up to or above national brandlevels.1^ Over time, many of the pineboard

" Same reference as footnote 8, p. 82." See footnote 7."D. A. Loehwing, "Resourceful Mer-

chants," Barron's, Vol. 38 (November 17,1958), p. 3.

" S . Masters, quoted in "Three Conceptsof Retail Service," Stores, Vol. 41 (July-August, 1959), pp. 18-21.

" S. M. Livingston and T. Levitt, "Com-petition and Retail Gasoline Piices," TheReview of Econom.ics and Statistics, Vol. 41(May, 1959), pp. 119-132 at p. 132.

druggists also seem to have become con-verted to fairly conventional operations.^^

NON-CONFORMING EXAMPLESforeign

In underdeveloped countries, the rela-tively small middle- and upper-incomegroups have formed the major marketsfor "modern" types of retailing. Super-markets and other modern stores havebeen introduced in those countries largelyat the top of the social and price scales,contrary to the wheel pattern.^^ Some non-conforming examples may also be foundin somewhat more industrialized environ-ments. The vigorous price competitionthat developed among Japanese depart-ment stores during the first three decadesof this century seems directly contraryto the wheel hypothesis.^o B. S. Yamey'shistory of resale price maintenance alsoreports some price-cutting by traditional,well-established British merchants whodeparted from the wheel pattern in the1880s and 1890s.2i Unfortunately, ourignorance of foreign retail history hindersany judgment of the representativenessof these examples.

American

Automatic merchandising, perhaps themost "modern" of all American retail in-stitutions, departed from the wheel pat-tern by starting as a high-cost, high-margin, high-convenience type of retail-

Paul C. Olsen, Tke Marketing of DrugProducts (New Brunswick: Rutgers Univer-sity Press, 1948, pp. 130-132.

" H. S. Hettinger, "Marketing in Persia,"JOURNAL OF MARKETING, Vol. 15 (January,1951), pp. 289-297; H. W. Boyd, Jr., R. M.Clewett, & R. L. Westfall, "The MarketingStructure of Venezuela," JOURNAL OF MAR-KETING, Vol. 22 (April, 1958), pp. 891-397;D. A. Taylor, "Retailing in Brazil," JOURNALOF MARKETING, Vol. 24 (July, 1959), pp. 54-58; J. K. Galbraith and R. Holton, Market-ing Efficiency in Puerto Rico (Cambridge:Harvard University Press, 1955), p. 35.

™ G. Fukami, "Japanese DepartmentStores," JOURNAL OF MARKETING, Vol. 18(July, 1953), pp. 41-49 at p. 42.

^ "The Origins of Resale Price Mainte-nance," The Economic Journal, Vol. 62(September, 1952), pp. 522-545.

THE WHEEL OF RETAILING 41

ing.-- The department-store branch move-ment and the concomitant rise of plannedshopping centers also has progressed di-rectly contrary to the wheel pattern. Theearly department-store branches consistedof a few stores in exclusive suburbs andsome equally high-fashion college andresort shops.

Only in relatively recent years havethe branches been adjusted to the chang-ing and more democratic characteristicsof the contemporary dormitory suburbs.Suburban shopping centers, too, seem tohave appeared first as "Manhasset MiracleMiles" and "Ardmores" before reachingout to the popular price customers. Infact, complaints are still heard that theregional shopping centers have displayedexcessive resistance to the entry of reallyaggressive, low-margin outlets.^^ E. R. A.Seligman and R. A. Love's study of retailpricing in the 1930s suggests that pres-sures on prices and margins were gener-ated by all types of retailers. The massretailing institutions, such aa the depart-ment and chain stores, that had existedas types for many decades were responsi-ble for a goodly portion of the price cut-ting.-^ As McNair has pointed out, thewheel operated very slowly in the caseof department stores.

Finally, Harold Barger has describedthe remarkable stability of over-all dis-tributive margins during the years 1919-1947.2̂ Some shifting of distributive workfrom wholesalers to retailers apparentlyaffected their relative shares of the totalmargins during this period, but this isnot the type of change contemplated by

==W. S. Fishman, "Sense Makes Dollars,"1959 Directory of Automatic Merchandising(Chicago: National Automatic Merchandis-ing Association, 1959), p. 52; M. V. Mar-shall, Automatic Merchandising (Boston:Graduate School of Business Administration,Harvard University, 1954), pp. 108-109, 122.

^ P. E. Smith, Shopping Centers (New-York: National Retail Merchants' Associa-tion, 1956), pp. 11-12; M. L. Sweet, "Tenant-Selection Policies of Regional Shopping Cen-ters," JOURNAL OP MARKETING, Vol. 23 (April,1959), pp. 399-404.

=" E. R. A. Seligroan and R. A. Love, PriceCutting a-)id Price Maintenance (New York:Harper & Brothers, 1932).

*̂ Same reference as footnote 8, pp. ix, x.

the wheel pattern. Of course, the stabilityBarger notes conceivably could have beenthe result of a perfectly smooth function-ing of the pattern, with the entrance oflow-margin innovators providing exactlythe right balance for the upcreep of mar-gins in the longer established types. Buteconomic changes do not come in smoothand synchronized fashion, and Barger'sdata probably should indicate consider-ably wider oscillations if the wheel reallyset the mold for all retailing in the post-war period.

CONCLUSIONS

The number of non-conforming exam-ples suggests that the wheel hypothesisis not valid for all retailing. The hypoth-esis, however, does seem to describe afairly common pattern in industrialized,expanding economies. Moreover, the wheelis not simply an illusion created by scram-bled merchandising, as Holdren suggests.Undoubtedly some of the recent "upcreep"in supermarket average margins is dueto the addition of nonfood and other highmargin lines. But in recent years thewheel pattern has also been characteristicof department-store retailing, a field thathas been relatively unreceptive to newcommodity groups.̂ f

In some ways, Jefferys' secular trendexplanation appears most reasonable. Thetendency of many established retailers toreduce prices and margins during depres-sions suggests also that increases may bea result of generally prospering environ-ments. This explanation helps to resolvean apparent paradox inherent in the wheelconcept. Why should reasonably skilledbusinessmen make decisions that consist-ently lead their firms along seeminglyprofitable routes to positions of vulnera-bility? Jefferys sees movement along thewheel as the result of sensible, business-like decisions to change with prosperingmarket segments and to leave the poorercustomers to low-margin innovators. Hisexplanation is supported by the fact that

^ R. D. Entenberg, The Changing Competi-tive Position of Department Stores in theUnited States by Merchandise Lines (Pitts-burgh : University of Pittsburgh Press,1957), p. 52.

42 JOURNAL OF MARKETING July 1960

the vulnerability contemplated by thewheel hypothesis usually means only aloss of market share, not a loss of ab-solute volume. At least in the UnitedStates, though, this explanation is par-tially contradicted by studies showingthat prosperous consumers are especiallyprone to patronize discount houses. Alsothey are equally as likely to shop insupermarkets as are poorer consumers.-^

The imperfect competition and excesscapacity hypotheses also appear highlyplausible. Considerably more investigationis needed before their validity can beappraised properly. The wheel patterndeveloped very slowly, and very recentlyin the department-store field. Yet marketimperfections in that field probably weregreater before the automobile gave theconsumer shopping mobility. Major por-tions of the supermarket growth in foodretailing and discount-house growth inappliance distribution occurred during

"R. Holton, The Supply and DemandStructure of Food Retailing Services, A CaseStudy (Cambridge: Harvard UniversityPress, 1954).

periods of vastly expanding consumption,when excess capacity probably was atrelatively low levels. At the moment thereis little evidence to suggest any clear-cutcorrelation between the degree of marketimperfection and the appearance of thewheel pattern. However, this lack maywell be the result of the scarcity of em-pirical studies of retail competition.

Managerial deterioration certainly mustexplain some manifestations of the wheel,but not all. Empires rise and fall withchanges in the quality of their leadership,and the same thing seems true in busi-ness. But the wheel hypothesis is a hy-pothesis concerning types of retailing andnot merely individual firms. Consequently,the managerial-deterioration explanationholds true only if it is assumed that newpeople entering any established type ofretailing as the heads of both old andnew companies are consistently less com-petent than the first generation. Again,the fact that the wheel has operated veryslowly in some fields suggests that sev-eral successive managerial generations canavoid wheel-like maturation and decay.

Reprints of every article in this issue are available (as long as thesupply lasts) at the following prices:

Single reprint $1.00 Four to 99, each $ 0.50Two reprints 1.50 First 100 40.00Three reprints 1.80 Additional lOO's, each.. 20.00

Quantity DiscountSpecial prices for large quantities.

Send your order to:AMERICAN MARKETING ASSOCIATION27 East Monroe Street, Chicago 3, Illinois