Effects of the Private Label Invasion in Food Industries

8/3/2019 Effects of the Private Label Invasion in Food Industries

1/14

Agricultural & Applied Economics Association

Effects of the Private-Label Invasion in Food IndustriesAuthor(s): Michael B. Ward, Jay P. Shimshack, Jeffrey M. Perloff, J. Michael HarrisSource: American Journal of Agricultural Economics, Vol. 84, No. 4 (Nov., 2002), pp. 961-973Published by: Oxford University Press on behalf of the Agricultural & Applied Economics Association

Stable URL: http://www.jstor.org/stable/1244738 .Accessed: 12/05/2011 05:17

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .http://www.jstor.org/action/showPublisher?publisherCode=aaea . .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

Agricultural & Applied Economics Association and Oxford University Press are collaborating with JSTOR todigitize, preserve and extend access to American Journal of Agricultural Economics.

http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=ouphttp://www.jstor.org/action/showPublisher?publisherCode=aaeahttp://www.jstor.org/stable/1244738?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=aaeahttp://www.jstor.org/action/showPublisher?publisherCode=aaeahttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/1244738?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=aaeahttp://www.jstor.org/action/showPublisher?publisherCode=oup


2/14

EFFECTS OF THE PRIVATE-LABEL INVASIONIN FOOD INDUSTRIES

MICHAEL B. WARD, JAY P. SHIMSHACK, JEFFREY M. PERLOFF,AND J. MICHAEL HARRIS

Using supermarket scanner data, we test a variety of hypotheses from trade journals about the in-vasion of private-label food products. According to conventional industry wisdom, name-brand firmsdefended their brands against new private-label products by lowering their prices, engaging in addi-tional promotional activities, and increasingly differentiating their products. Our empirical evidenceis inconsistent with these beliefs.

Key words: private label, entry, price, promotional activity, differentiation, supermarket.

When the share of private-label processedfoods and beverages increases, brand-namefirms' prices tend to rise and their promo-tional activities all contrary o widely held be-liefs about his industry. We use recent groceryscanner data to examine he effects of private-label entry on prices, promotional activities,and product diversification n thirty-two oodand beverage ndustries.

Although the food industry rade journalsand newspapers aveextensively discussed he"invasion" of private-label products over thelast decade, academics have largely ignoredthe effects of this entry on prices, advertising,and product diversification. A few researchershave discussed why firms produce privatelabels (Bontems, Dilhan, and Requillart;Galizzi, Venturini, nd Boccaletti), why retail-ers sell them (Mills, 1995; Dhar and Hoch;Narasimhan and Wilcox), and why pricesdiffer between private-label and brandedgoods (Conner and Peterson, Hinloopenand Martin).

The only academic papers that have explic-itly examined the magnitude and speed ofprivate-label penetration or the response ofbrand-name irms to such entry tend to sup-port the conventional business wisdom. Raoand Mills (1999) derive optimal heoretical e-sponses to the entry of private-label goods instylized models, and conclude hat name-brandfirms should increase promotion and prod-uct diversification. The only previous empiri-cal studies, Putsis and Cotterill and Putsis, hatfocus on brand proliferation nd demand andmarket structure responses respectively, on-clude that private-label penetration loweredthe 1991-92 average price of national brands.Although this finding may have been accuratein the early 1990swhen the "invasion" tarted,we find the opposite using more recent data.

In the first section we summarize he foodindustry's onventional wisdom about private-label products in a series of stylized facts.Although these stylized facts about the ef-fects of entry are consistent with standardeconomic intuition, we describe in the sec-tion Economic Theory about Entry and Prices,four economic theories explaining why pricesin noncompetitive markets may rise despiteentry of new firms and products. In the sec-tion Data, we describe our grocery scannerdata set covering thirty-two ood and bever-age industries. We discussour empirical esultsand econometric methodology n the sectionsEvidence on the Private-Label nvasion andEvidence on Name-brand Responses respec-tively. In Summary and Conclusions, we sum-marize our findings and draw conclusions.

Michael B. Ward is an assistant professor in the Depart-ment of Agricultural and Resource Economics, University ofCalifornia, Berkeley and member of the Giannini Foundation;Jay P. Shimshack is an assistant professor in the Department ofEconomics, Thffs University and member of the Giannini Foun-dation, University of California, Berkeley; Jeffrey M. Perloff is aprofessor in the Department of Agricultural and Resource Eco-nomics, University of California, Berkeley and member of theGiannini Foundation; J. Michael Harris is an economist in theEconomic Research Service, U.S. Department of Agriculture.

The authors are very grateful to Mark Denbaly for suggestingthe project and supporting it. The authors thank Veronica Jonesand John Hession for helping to produce the data in a useable form.The authors' names are listed in reverse alphabetical order. Twoanonymous reviewers made very helpful suggestions.

Amer. J. Agr. Econ. 84(4) (November 2002): 961-973Copyright 2002 American Agricultural Economics Association


3/14

962 November 2002 Amer. J. Agr. Econ.

Stylized Facts

By searching Lexis-Nexis files of newspaperarticles and trade journals, we identified thereceived wisdom from the food and beverageindustries. We summarize these common be-liefs as stylized "facts". Although we reportthese claims here as though they were facts,we will show that many of them are not truetoday (even if they were true at the time thestatements were made).

Private-Label Invasion

Discount brands were introduced intoAmerican supermarkets in the late 1970s(Janofsky). As one article stated, consumers

long regarded such a product as "a cheapand nasty generic substitute for the realthing, rolled out by retailers during recessionsand discarded once the economy picked upagain" (The Economist). However, as thatarticle notes, consumers changed their viewwhen high-quality private-label products wereintroduced in the late 1980s and early 1990s.

The substantial quality improvement of pri-vate labels resulted from technological ad-vances and production by name-brand firms.Technological advances allowed competitorsto "come close" to replicating successful na-tional brands (Kennedy). Additionally, somenational brands started producing private-label versions of their name-brand prod-ucts, often to employ the excess capacityin their plants. Examples include CampbellSoup (Vlasic pickles), Union Carbide (garbagebags), Hershey Foods (Ronzoni pasta), andDel Monte (canned fruits and vegetables)(Janofsky, Beckett). Other name-brand firmsproduced private-label brands in different cat-egories from their own. For example, H. J.Heinz Company, the leading producer ofketchup, makes about 80% of the private-labelsoups sold in supermarkets (Beckett).

Consumers were becoming increasinglyaware of private-label improvements as well.A Gallup Poll commissioned by the PrivateLabel Manufacturers Association (PLMA) re-ported that consumers' store brand awarenessrose from 86% in i991 to 91% in 1995, andthe percentage of consumers who regularlybought private-label brands increased from77% to 83% (Responsive Database ServicesInc.). Of consumers surveyed, 76% agreed thatstore brands were "brands just like nationalbrands." Of consumers who had bought pre-

mium private label items, 90% rated the prod-ucts equal to or better than national brandsand planned to buy private-label products inthe future (Doug, Palmer).

Supermarkets started prominently display-ing house brands while dropping second-tiernational brands in many categories (Janofsky).According to some reports, brands rankedthird, fourth, or fifth were particularly hard hit.

Supermarket chains learned that privatelabels provided higher profits than nationalbrands. According to Jonathan Ziegler, aretail-industry analyst at Sutro & Co., the grossmargins on store brands may be 35 % or more,versus an average of 25% on other products(Beckett). Supermarket profits also rise fromcarrying private-label brands because they cre-

ate loyalty to a particular supermarket chainrather than to a national brand: Customers re-turn to Safeway if they prefer the chain's Selectbrands.

Response of Name-Brand Firmsto Private-Label Entry

Trade journals and newspaper articles re-ported many round-table discussions by indus-try executives in which they bemoaned the in-vasion of

private-label productsand discussed

how they responded. In addition to lower-ing prices and engaging in promotional activ-ities, these executives said that they engagedin brand building (Nijssen and Van Trijp) byincreasingly differentiating their products. Forexample, when consumers started switchingfrom Kellogg's to private-label cereals simi-lar to popular Kellogg's brands (that sell atroughly half the price), Kellogg announcedthat it was issuing more coupons to make itsprices more competitive with generic brands,

improvingits

advertising,and further diversi-

fying its products (MacDonald, 1998).

Price responses. Many national brand exec-utives reported that the private-label invasionwas killing brand loyalty, so that they had tocut prices to compete. This reasoning was givenby Philip Morris when it cut its price for Marl-boro cigarettes, Procter & Gamble when it re-duced the price of Pampers diapers by a quar-ter, and Kraft General Foods when it loweredits cheese prices by 8% (Beckett, de Jonquieresand Tait). Many other firms reported that theylowered their prices indirectly by means ofsales and discount coupons.


4/14

Ward t al. Private-Label nvasion n Food Industries 963

Promotional activities. Name-brand man-ufacturers reported increasing point-of-purchase promotional activities in responseto the new competition.1 The share of promo-tional budgets allocated to point-of-purchaseexpenditures and advertising were 73% and27% in 1992, compared to 62% and 38%respectively in 1960 (Lenius). From 1980to 1992, U.S. food manufacturers' pendingon promotional schemes, such as money-offoffers and coupons, rose from half to three-quarters of total marketing budgets, whileadvertising's hare fell from 44% to 25% (deJonquieres and Tait).

Differentiation. The executives contend hattheir brands are maintained through differ-

entiation. According to Quaker Oats' CEORobert S. Morrison, "Leading brands pos-sess great long-term value only if they canevolve over time to respond to the tastesand needs of new generations" Martin andKubomura). Dale F Morrison, president andCEO of Campbell Soup, said "We are mak-ing our brands more contemporary, more rele-vant and more convenient o consumers of allages" (Martin and Kubomura). For instance,Campbell produced new soup varieties andpackaging ormats such as its ready-to-serve

tomato soup in resealable plastic containersand Campbell's Soup To Go in microwaveablesingle-serve bowls.

One might hink of constantly roviding ewproducts as a flagpole strategy: "Let's run t upthe flagpole and see who salutes it." Productsthat are not accepted by consumers re quicklydropped.

In recent years, an average of 27% of Gen-eral Mills' volume has come from productsfive years old or less (Martin and Kubomura)."Our iscal 2000 plans call for higher evels of

new product nnovation across our U.S. busi-nesses," eports Stephen Sanger, hairman ndCEO of General Mills. To accelerate he flowof innovative and creative products, GeneralMills' operating divisions now focus at least25% of total resources on new products andnew business deas. The company has recentlyintroduced Sunrise ertified organic ereal andYoplait Go-Gurt, a portable yogurt snack n aflexible squeeze tube that s ready o eat refrig-erated or frozen.

Many managers reported that they in-creased the rate at which they innovate n re-sponse to the challenge of private abels.2Forexample, firms ntroduced 16,143products n1991, ncluding 12,398 ood products and 3,745nonfood products such as diapers and sham-poo, which was 22% more than they intro-duced the previous year (Kennedy).

In summary, ccording o conventional wis-dom, private-label products have been persis-tently and rapidly penetrating rocery marketssince the late 1980s.This invasion has been atthe expense of national brands, and second-tier national brands were particularly hardhit. Company executives and industry expertscontend that name-brand irms responded toprivate-label entry in three ways: Firms low-ered prices, engaged in more promotionalactivities, and further differentiated theirproducts.

Economic Theories about Entry and Prices

All of these stylized acts appear o be reason-able. Indeed, economists usually assume thatentry leads to more competition with lowerprices and possibly more promotional activi-ties and greater differentiation.3 However, wewillpresent evidence that many of the stylizedfacts are not currently rue in processed foodand beverage ndustries. n particular, rivate-label entry s correlated with increased name-brand prices and reduced promotional activi-ties. We start our analysis by briefly noting thatat least four economic theories allow for priceincreases n response to entry.

The simplest of these theoretical explana-tions nvolves quality. Manufacturers f name-brand products may raise the quality of their

goodswhen faced with

private-label ntry.This

response eads to higher name-brand roductprices especially f it is more costly to producehigher quality goods.4

I For example, Nelson and Hilke found that featuring by retail-ers can provide strategic advantages to a dominant coffee manu-facturer in fending off a new entrant or smaller competitor thattries to expand.

2 The number of new supermarket and drugstore items wentfrom a mere 1,281 in 1964 to 20,076 in 1994, though many of theseinnovations presumably would have occurred in the absence ofprivate-label competition. "Quantum Leap." Investor's BusinessDaily, Executive Update CEO Briefing, January 20 1995, p. A3.

3 The standard analyses can be found in any good introductoryor intermediate undergraduate microeconomics textbook (e.g.,Lipsey, Courant, and Ragan, pp. 219, 254-260). In a common text-book example, entry into a competitive market shifts the market

supplycurve to the

right, loweringthe

equilibrium price. Similarly,when entry occurs, noncompetitive firms may increase their as ad-vertising and product differentiation to appeal to rivals' consumers.

4Several industry executives report that they fight back by rais-ing quality. General Mills Inc. Vice President John Hallberg, saidthat, because his company feared technological advances that


5/14


6/14

Ward et al. Private-Label Invasion in Food Industries 965

of 166product categories rom 5 major super-market departments edible groceries, rozenfood, bakery, dairy, and deli).

We were provided with data rom hirty-fourrandomly selected categories-a little overone-fifth of the available categories. We in-clude only thirty-two f these categories n ouranalyses because one category, baby food, hadno private-label or generic purchases duringthe relevant period, and another, wine, hadonly a trivial amount (0.3% of sales).

Although the data set contains nformationon each ndividual randed tem sold, t reportsonly aggregated data for generic and private-label items per time period. Thus, we do notknow how many private-label tems or firmsthere were, but we do know the average priceand the total quantity old.Local promotion nformation s collected byIRI field auditors on a weekly basis and s usedto develop physical and sales volume measuresof food products sold under promotion andmerchandising. he auditors rack and classifythe use of displays, etail ads, and any other re-tailer merchandising fforts. Promotion nfor-mation is assembled each week and mergedwith weekly scanner data. The informationallows IRI to differentiate regular everydaysales from sales made under promotion or spe-cial merchandising. The Economic ResearchService data base provides en promotion mea-sures (five for dollar sales and five for physicalvolume), which reflect the share of sales andphysical volume sold under price reduction,display, eature, eature and display, nd any n-dividual or combined use of these promotions.Price reduction refers o items with temporarysale prices; display, o aisle or end displays; ea-ture, to items that are primarily advertised nlocal papers or paper inserts; eature and dis-play, to items that are both advertised and ondisplay.

We measure price for a brand as the totalrevenue divided by the total quantity (mea-sured n units of weight or volume).6We deriveour nominal prices rom sales data that are netof discounts such as from coupons. The nomi-nal prices are converted nto real prices usingthe Consumer Price Index (where the indexequals one in the first period).

The data set covers a little over two years.The sample is divided into twenty-nine timeperiods of four weeks each, which we call

months. Thus, there are thirteen IRI monthsper year. The first month n the sample endedon 8 December 1996, and the last one on 31January 1999.

Evidence on the Private-Label nvasion

Have private-label roducts ubstantially en-etrated ood and beverage sectors and do theycontinue to enter rapidly? We find that thequantity and revenue shares of private-labelproducts and the rates of increase of theseshares are high in many (but not all) food andbeverage ndustries.

Averaging ver all thirty-two f our remain-ing categories (and weighting each categoryequally), the revenue share of private-labeland generic items is 14.3% and the quantityshare s19.0%.Private-label nd generic goodsare nearly two-thirds of the quantity share offrozen poultry, but only 1% of pickles and rel-ish. Generics' share was 0.6% of hot cerealsales, 0.5% of shortening and oil, and 0.3%or less for all other categories. Moreover, nogenerics were sold in roughly half of the cat-egories. Henceforth, we treat both private a-bel and generics as one group, which we callprivate-label oods.

To determine whether he rate at which thequantity hare of private-label oods increasedover our time period, we regressed the log-arithm of their share on a time trend andmonthly seasonal dummies. In about half ofour categories, here wasno substantial hangein the revenue share of private-label goods.In only two, English muffins and ice cream,did we find a statistically significant (at the0.05level) decreasing ate.Private-label oodspenetration ncreased at a statistically ignif-icant rate in the remaining 40% of the cat-egories. Double-digit growth rates occurredin slightly more than one-quarter of the cat-egories, which are categories with typicallysmall private-label hares.

Evidence on Name-Brand Responses

Are stylized "facts" about the reaction ofname-brand irms to increased private-labelcompetition correct? In particular, id name-brand firms defend their brands against theprivate-label nvasion by lowering heir prices,conducting sales, engaging in other promo-tional activities, and increasingly differentiat-ing their products?

6 We experimented with a variety of other measures of price (e.g.,average price per item, rather than by weight), but found that ourqualitative results are unaffected by the weighting used.


7/14


Prices

How do prices respond to private-label en-try? Each of the four theories that we dis-cussed above explains why brand-name pricesmay rise or fall after entry of private-labelproducts. We lack the data to distinguish be-tween these theories, so we take a reduced-form approach that is consistent with all thetheories. We start by examining the correlationbetween name-brand and private-label sharedirectly. We then consider various alternativereduced-form tests of this relationship. Finally,we discuss the correlation between private-label share and prices for all goods and private-label products.

Brand-name prices. We inspected the his-

togram for each brand price in each category.Generally, these distributions look lognormal,with a single peak and a long right tail. Thus,we summarize these distributions by the firsttwo moments of the log price.7

We start by examining the effect of private-label penetration on the pricing behavior of in-dividual name-brand firms. Table 1 shows theresults of the regressions of the log (real) priceof each of the eight largest firms and the log av-erage price of the other branded firms in eachcategory on the log of the share of private-label

goods and three seasonal (quarterly) dummies(adjusted for first-order autocorrelation). Onereason that we control for seasonality in theseregressions is MacDonald's finding that foodprices tend to fall in periods in which demandpeaks.

We have a remarkable result: When theshare of private-label goods rises, the pricesof name-brand goods tend to rise (even aftercontrolling for the possible endogeneity of theprivate-label share). Despite the large number(288) of estimated coefficients, every statisti-

cally significant coefficient (0.05 criterion) ispositive.

A similar phenomenon has been observed inthe pharmaceutical market (Caves, Whinston,and Hurwitz; Grabowski and Vernon; Frankand Salkever, 1992, 1997). When generic phar-maceutical manufacturers are allowed to sellan exact clone of a previously proprietary drug,they sell at a price far below the original name-

brand product price. Although price-sensitiveconsumers switch to the generics, the brand-conscious consumers who continue to buy thename-brand drug are frequently charged ahigher price than they paid originally. For ex-

ample, Frank and Salkever (1997) reportedthat the share of prescriptions sold by retailpharmacies that was accounted for by genericproducts roughly doubled during the 1980s.Using a sample of thirty-two drugs that lostpatent protection during the early to mid 1980s,they found that many name-brand prices in-creased after generic entry and were accompa-nied by large decreases in the price of genericdrugs.

Sensitivity experiments. Are the regressionsin table 1 reasonable? Three obvious concernsare that our equations omit other relevant fac-tors, that the rise in the average brand-nameprice variable reflects a change in composition,and that the share of private-label goods maybe an endogenous variable.

We experimented with augmenting ourbasic regression in table 1 (the log of brand-name price regressed on the log of private-label share and seasonal dummies) with var-ious combinations of the following groups of(log) measures of diversity and market power:

(A) Gini indexesfor items and for

firms,8(B) number of name-brand firms, (C) num-ber of brands, (D) number of items,9(E) fraction of births and deaths, (F) shares ofthe two largest, four largest, and eight largestname-brand firms, and the share of the name-brand firms other than the eight largest ones.We added each of these groups of "structure-conduct-performance" (SCP) variables sepa-rately to the equation used in table 1. We alsoexperimented with adding several groups ofthese variables at the same time. None of these

SCP variables substantially changed our quali-tative results for the coefficient on the private-label share. These extra variables were rarelystatistically significant and their coefficientsfollowed no obvious pattern.10

7Because we determined no clear patternsto the second mo-

ment in our regression analyses, we only briefly mention the resultshere. As the share of private-label goods rises, the variance of theprice of branded goods does not change statistically significantlyin fourteen categories, falls in six categories, and rises in twelvecategories.

8 The Gini coefficient was calculated as

G= a [(n - rj))- (rj- 1)]

i=1where sj is the share of the jth item, rJ is its rank (the largest sharehas rank 1), and there are n items.

9One might be concerned that an increase in private-label shareinduces brand-name firms to reduce the number of items

theysell

(thereby resulting in higher prices). However, the evidence gener-ally rejects this hypothesis, as we show in the following section onproduct diversity.

')We also tried regressing brand-name prices on theSCP variables and seasonal dummies (that is, dropping the


8/14

Ward t al. Private-Label nvasion n Food Industries 967

Table 1. Elasticity of Name-Brand Firm's Price with Respect to Private-Label Share

Eight Largest Brand Firms and Other Brand Firms

1 2 3 4 5 6 7 8 9+

Baked beans0.22b 0.42b 0.54b 0.28a 0.25b 0.19b 0.09 0.10b 0.31bButter 0.29 0.53 -0.82 0.38 -0.13 -0.84 1.15 -0.41 -0.28

Canned ham -0.03 -0.07 -0.04 0.02 -0.01 0.02b -0.02 0.00 0.04Canned uices 0.09b 0.17b 0.37b 0.37b 0.02 -0.57 0.07 -.18 -0.05Cottage cheese -0.14 -0.28 -0.53 -0.09 0.47a -0.11 -0.13 0.03 0.45bCrackers 0.19b 0.26b -0.34 -0.17 0.04 -0.03 0.14a 0.03 -0.08Desserts 0.07b 0.09b 0.03 -0.03 0.05 0.09a 0.04 0.01 0.03English muffins -0.59 -0.10 0.14 -0.03 -0.24 -0.55 -0.01 -0.55 0.55bFrosting 0.19b 0.10b 0.04 0.05 -0.13 0.01 0.47a -0.01 -0.08Frozen baked goods 0.17b 0.00 0.49b 0.17b -0.05 0.32b 0.23b 0.28 -0.11Frozen breakfast ood 0.05b 0.06b 0.08b 0.26b 0.00 0.07 0.12b 0.08b -0.03Frozen ruit 0.23 0.07 0.10 0.83b 0.66b -0.12 0.18 0.57b 0.83bFrozen poultry -0.12 0.03 -0.07 0.14 0.10 0.18 0.12 0.05 0.25aGelatin mixes -0.16 -0.22 0.13b 0.11 0.20 -0.01 0.17a -0.10 0.63bHot cereal 0.19b -0.13 0.06b -0.14 0.05a -0.01 -0.04 0.17a -0.06Ice cream 0.09 0.46b 0.57a 0.10 -0.13 -0.43 0.25 0.52b 0.23bInstant potatoes 0.13 0.01 0.02 -0.15 -0.02 0.02 0.11 0.21a -0.02Mustard & ketchup 0.31b 0.22b 0.17b 0.20b 0.06b 0.28b -0.02 0.06 0.29bPeanut butter -0.01 0.29a 0.21a 0.01 0.24a 0.17 -0.10 -0.02 0.64bPickles and relish 0.09b 0.08 0.09b 0.05 0.06b 0.02 0.07b 0.12b 0.09bPizza products 0.05b 0.07 0.02 0.01 0.01 0.15b 0.06b 0.09b 0.05Pizza, refrigerated 0.02 -0.02 -0.01 -0.16 -0.19 0.03 -0.03 -0.03 -0.03Popcorn/popcorn il 0.25b 0.01 0.07a 0.02 0.13 -0.07 -0.09 0.00 0.14Rice/popcorn akes 0.20b -0.02 0.07 0.01 -0.95 -0.02 -0.01 0.04a -0.18Shortening nd oil 0.51b 0.02 -0.08 0.17 -0.54 0.08 -0.17 0.03 0.19

Snack/granolaars 0.02 0.16b -0.01 0.01 0.12b 0.09b 0.13b 0.15 0.06

Spaghetti/Italian auce 0.11a 0.02 0.06 0.06 0.02 0.01 0.01 0.30 0.08Sugar substitutes 0.04b -0.02 0.07a 0.33a 0.12b 0.38b 0.05a 0.03a 0.12aTea, ready to drink 0.21b 0.20b 0.45b 0.02 0.27b 0.19 0.15 0.00 0.18Tomato products 0.18 0.55 0.08 0.21 0.10b 0.23 0.50b 0.09 0.06Vinegar -0.24 -0.06 0.17 0.06 -0.83 0.08 0.42b -0.04 -0.06Yogurt 0.37b 0.15b -0.01 0.25a 0.11 -0.07 0.38 0.04 0.00

aReject he hypothesis hat he elasticity coefficient) s zero at 0.10confidence evel.bReject he hypothesis hat he elasticity coefficient) szero at 0.05confidence evel.Notes:Not reported re hree quarterly ummies nd a constant. irst-order utocorrelation orrection mposed.

Similarly, one might be concerned thatmergers by grocery store chains played an im-portant role. Because our time period is rela-tively brief, there was relatively ittle changein concentration at the national level (unlikeat the regional level). We proxied for thesechanges by adding a time trend, which wasnot statistically significant and had negligi-ble effects on the private-label share coeffi-cient. Moreover, hanges n grocery tore mar-ket power are unlikely to differentially affectprices within categories of food and beverages.

We were concerned that we lacked goodproxies for costs. By excluding uch variables,

we implicitly assumed that there were no dra-matic shifts in costs during our two and one-third year sample period. It is also possible thatthere were shifts in demand for reasons havingnothing to do with private labels. As a partialadjustment for unobserved shifts in costs anddemand, we included a time trend. Again thistrend was not statistically significant and hadnegligible qualitative effects.

We examined whether our main price resultis due to a change in the composition of itemssold by brand as private-label share increases.Suppose that prices did not change but thatprivate-label products are disproportionatelylarge and that larger sizes have lower prices perounce. Then, the introduction of private labelswill result in higher average brand prices.

private-label share). Again, the SCP variables were rarely statisti-cally significant and had no obvious pattern.


9/14


We examined he quantity weighted averagesize of branded tems to determine whetherthey rose or fell over our sample period. Wefound that weighted average item sizes in-creased for sixteen categories and fell for six-teen. For those sixteen categories with sig-nificantly positive branded price responses(table 2), the weighted average size rose forten and fell for six categories. For those sevencategories with extremely trong positive priceresponses (a significant branded price elastic-ity of >0.2), the weighted average tem sizesincreases for six and fell for one. Moreover,the sizes are comparable etween branded and

Table 2. Elasticities of Price with Respectto the Share of Private-Label Goods

PrivateAll Branded Label

Baked beans -0.03 0.04 -0.36bButter -0.25 0.16 -0.47Canned ham -0.10b -0.08a -0.39bCanned juices 0.02 0.13b -0.14bCottage cheese -0.46b -0.02 -0.84bCrackers -0.27b -0.12b -0.27bDesserts 0.06a 0.06a -0.04English muffins -0.54b -0.10 -0.67b

Frosting 0.09b 0.11b -0.02Frozen baked goods 0.11 0.22b -0.10Frozen breakfast ood -0.02a -0.01 0.11bFrozen fruit 0.48b 0.24 0.21Frozen poultry 0.34b 0.35b -0.02Gelatin mixes 0.26 0.30a -0.06Hot cereal -0.08b 0.07a 0.01Ice cream -0.02 0.18b 0.06Instant potatoes -0.09b -0.05 0.00Mustard and ketchup 0.15b 0.30b -0.01Peanut butter -0.08 0.01 -0.15aPickles and relish 0.05b 0.06b -0.11aPizza

products-0.06b -0.04b

-0.19bPizza, refrigerated 0.00 -0.02 0.21Popcorn and popcorn oil -0.43b -0.13b -0.09Rice and popcorn cakes 0.02 0.09a -0.15bShortening and oil 0.12a 0.31b -0.07aSnack bars/granola bars 0.00 0.01 0.04bSpaghetti/Italian sauce 0.11b 0.14b -0.15bSugar substitutes -0.35b -0.06b -0.74bTea, ready to drink 0.30b 0.32b 0.07Tomato products 0.10 0.43b -0.20bVinegar -0.63b 0.31 -0.30aYogurt 0.07 0.19b -0.05

aReject the hypothesis that the elasticity (coefficient) is zero at 0.10 confi-dence level.bReject the hypothesis that the elasticity (coefficient) is zero at 0.05 confi-dence level.Note: Not reported are three quarterly dummies. First-order autocorrelationcorrection imposed.

private-label irms; private-label irms do notdisproportionately roduce only large sizes.Thus, we conclude hat the hypothesis hat theentry of private abels eads to greater sales ofbranded tems that are smaller s false, partic-ularly or those categories of most interest.

Finally, we considered the possibility thatthe share of private-label oods sendogenous.That s,we need to consider alternative xpla-nations or why private-label hare and brand-name prices are positively correlated. We havepresented economic theories that are consis-tent with the possible explanation hat largerprivate-label hare eads to higher brand-nameprices. Alternatively, igher prices of brandedgoods could increase the demand for privatelabels. To disentangle hese two stories we useinstruments o control the possible simultane-ity of the private-label hare variable n ourprice equation.11

Various actors may affect the private-labelshare. First, over time, consumers, aving riedprivate-label roducts, become convinced hattheir quality rivals that of name-brand prod-ucts, so the consumers ncreasingly buy theseproducts cf. the "try t you'll ike it" model ofKohn and Shavell). Second, both the price ofname-brand oods and the likelihood of entryof private-label oods depend on the existenceof unfilled niches n product pace.Third, irmsmay introduce new private-label tems whenthe price of name-brand goods is unusuallyhigh. Fourth, onsumers may increasingly uyprivate-label products if branded prices rise(substitution ffect). If the increase n the shareof private-label goods is due to only factorsthat are exogenous or at least predetermined,then we are justified n treating the share ofprivate-label oods as exogenous in our priceregressions. However, to be sure, we conductHausman ests of endogeneity of private-labelshare using nstruments.

To derive nstruments, we hypothesized hatthe same factors causing the private-labelshare o grow n one food category affect otherfood categories. Thus, we used the growth ofthe share of private-label oods in other mar-kets as instruments n each given market. Wedivided the thirty-two industries into seven

11As an alternative to treating private-label share as endogenous,we tried regressing price on a one period or two-period lag ofthe private-label share-so that share was "predetermined." Notsurprisingly, the results were virtually the same as when we usedcurrent share. We also experimented with a Granger-type modelwith short lags of both prices and shares, which was not informative(we have too short a period for such a temporal ordering test tomake sense).


10/14


groups: canned goods, frozen goods, sweets,condiments, dairy, Italian, and other.12 Wetook the average of the share of private-labelgoods in each of these groups and used theseas instruments dropping he one to which the

given industry belongs). Thus, each equationhad six instruments.We then regressed he log price of branded

goods in each category on the share of private-label goods and three seasonal dummies withand without using nstruments. he nstrumen-tal variables estimates were very close to theoriginal estimates. For example at the cate-gory level, we rejected exogeneity at the 0.05level on the basis of a Hausman test in onlytwo of the thirty-two ndustries frozen poul-try and rice and popcorn cakes), and even in

those industries the efficient and consistentpoint estimates were very close. Consequently,we report only our generalized east squaresestimates.

Relativeprices. We have demonstrated hatname-brand rices rise in response o private-label entry. We now ask, do private-label ricesand overall market prices rise or fall with in-creased private-label penetration?

If private-label goods usually sell for lessthan do branded goods, the overall aver-age price in the market place should lie be-tween that of private-label and name-brandgoods and should fall with entry of private-label goods. The private-label price is lessthan the price for branded goods in allcategories except frozen poultry, where theprivate-label prices are substantially higherthan hose of name-brand oods. (Frozen poul-try private-label oods have the largest hare-nearly wo-thirds f industry quantity-of anycategory.)

In table 2, we investigate the effects ofprivate-label entry on overall, branded, andprivate-label prices. Table 2 reports the en-try elasticities rom the regressions of the logsof average overall price, the average brandedprices, and the average private-label priceon the same right-hand side variables as intable 1:private-label hare and three quarterly

dummies (where we control for first-orderautocorrelation).

We considered hree possible heories aboutthe effects of increased penetration n private-label prices:

1. If the private-label goods are essentiallyidentical generic goods that are priced com-petitively at constant marginal ost, entry ofadditional private-label oods should haveno effect on their price.

2. If the private-label tems are monopolisti-cally competitive as described by a Cham-berlinian model, where all goods competewith each other symmetrically that is, notasymmetrically s they would in a spatialmodel), entry s likely to lower the averageprice.3. If the private-label goods are effectivelybranded (e.g., Safeway's well-regarded e-lect brand) and spatially differentiated romother branded goods, entry could raise orlower price.

We expected that either the first or secondstories apply n these food industries, o we ex-pect private-label rices o stay constant or fallwith entry. The empirical vidence sconsistentwith these predictions. n virtually very cate-

gory, private-label penetration has no statisti-cally significant ffect or a negative effect onthe price of private-label goods. The only ex-ceptions are frozen breakfast oods and snackand granola bars.

From table 1, we already know that pene-tration has either no statistically ignificant f-fect or raises the price of individual brandedgoods.13Although conventional ndustry wis-dom predicts hat the overall average price willfall with increased private-label penetration,the question remains whether t does so giventhe increase in name-brand prices. The over-all price has a statistically ignificant negativeelasticity n ten categories 31%), a positive ef-fect in six categories 19%), and is unchangedin the remaining ones. In all but one indus-try where the overall elasticity s positive, thename-brand lasticity is an even larger posi-tive number. We conclude that an increase nprivate-label share is correlated with either noeffect or a decrease in the price of private-labelgoods, no effect or an increase in the price of

12The categories in our groups are canned: canned ham, cannedjuices; dairy: butter, cottage cheese, yogurt; frozen: frozen break-fast food, frozen fruit, frozen poultry, and frozen baked goods;Italian: pizza products, refrigerated pizza, spaghetti/Italian sauces;sweets: desserts, frosting, gelatin mixes, ice cream, snack/granolabars, sugar substitutes; condiments: mustard and ketchup, picklesand relish, vinegar; other: baked beans, crackers, English muffins,hot cereal, instant potatoes, peanut butter, shortening and oil, pop-corn and popcorn oil, ready-to-drink tea, rice and popcorn cakes,tomato products.

13However in table 2, the elasticity for several of the aggregatedname-branded categories is statistically significantly negative. Pre-sumably this inconsistent result reflects the aggregation of severalnegative (but statistically insignificant) elasticities for the individ-ual firms in table 1.


11/14

970 November 002 Amer. .Agr. Econ.

branded oods,and usually but not always)noeffect or a negative effect on the overall price.

Promotional Activities

According o our stylized "facts", ame-brandfirms engage in sales-price reductions--or nonprice promotions in response to theprivate-label nvasion. Unfortunately, we donot have information about these firms' na-tional advertising, ut we do know about ocaland in-store promotions-which are increas-ingly the dominant orms of promotion.

Name-brand irms requently ngage n pro-motional activities. In a typical month, theshare of name-brand tems with temporaryprice reductions anges rom 5% in hot cerealto 18% n butter.

The nonprice promotional activities ncludelocal feature ads and n-store displays.The ow-est share of items with nonprice promotionsof merchandising s 4% for sugar substitutesand vinegar, but the share exceeds a third forcanned ham, crackers, ce cream, and ready-to-drink ea. The share of all merchandising--price and nonprice promotions combined--reaches nearly a half for canned ham and icecream.

There is a pronounced downward rend inpromotional ctivities.Except for refrigeratedpizza and rice and popcorn cakes, he quantityshare of items with price promotions s stableor declining. There s little change n the shareof nonprice promotions, with statistically ig-nificant decreases n only four categories andincreases n only four categories.

We regressed he share of items on sale onthe share of private-label goods and the sea-sonal dummies. Table 3 shows the elasticityof the share of price reductions and nonpricepromotions with respect to the private-labelshare. Every statistically ignificant elasticityof price reductions with respect to private-label share is negative: Name-brand irms ei-ther make no change or have fewer sales. Asprivate labels expand, name-brand irms aresubstantially ess likely to engage in nonpricepromotions using eature ads and displays.Notonly are virtually all the statistically signifi-cant elasticities negative (the pizza productcategory is the only exception), but most ofthese negative elasticities are larger than onein absolute value-in many cases much larger.For example, the elasticities for rice and pop-corn cakes, yogurt, mustard and ketchup, andhot cereals are between -4 and -5, while theelasticity for butter is nearly -11. The statisti-

Table 3. Elasticity of the Share of Promo-tions with Respect to the Private-Label hare

Price NonpriceReductions Promotions

Baked beans -0.89b -3.33bButter 4.17a -10.94bCanned ham 0.14 0.19Canned uices -0.58 -0.84bCottage cheese -0.45 1.73Crackers -0.17 -0.48bDesserts -0.49a -0.32English muffins 0.78 -1.5lbFrosting -2.17b -2.93bFrozen baked goods -0.82b -1.70bFrozen breakfast food -0.25b -0.02Frozen ruit -1.13 -0.97Frozen

poultry-0.11 -0.21a

Gelatin mixes -0.29 -1.70bHot cereal -1.87b -4.90bIce cream -1.09b -1.17bInstant potatoes -0.99a -3.89bMustard nd ketchup 0.22 -4.48bPeanut butter -0.14 -3.10bPickles and relish -0.78b -0.53bPizza products -0.08 0.76bPizza, refrigerated -1.02b -0.03Popcorn and popcorn oil -1.66b -3.92bRice and popcorn cakes -1.32a -4.08bShortening nd oil -0.27 -2.09bSnack bars/granola ars -0.17 -3.02Spaghetti/Italian sauce -0.38a -0.96bSugar ubstitutes -0.40 -2.29bTea, ready to drink -0.37b -0.45bTomato products -0.86a -1.15Vinegar -0.12 0.00Yogurt -2.90b -4.05b

a Reject the hypothesis that the elasticity (coefficient) is zero at 0.10 confidencelevel.bReject the hypothesis that the elasticity (coefficient) is zero at 0.05 confidencelevel.Note: Not reported are three quarterly dummies. Autocorrelation correction

imposed.

cally significant sales elasticities are also neg-ative, though usually smaller in absolute valuethan the nonprice promotion elasticities. Thus,our findings once again repudiate the conven-tional wisdom: Name-brand firms react to in-creased private-label competition by holdingfewer sales and reducing the share of itemswith nonprice promotions.

Differentiation

As we lack information about the detailedcharacteristics and qualities of the giganticnumber of items produced, we cannot directlylook at the degree of differentiation. Instead,


12/14


we use the number of items, the number ofbirths and deaths of items, and the ratio ofitems to firms as proxies for the attemptsby name-brand firms to differentiate theirproducts.

What has happened o the number of name-brand items and firms? Have the private-label goods driven some name-brand firmsand items out of business? Or, have name-brand firms-attempting to further differen-tiate their products-increased the number ofitems they sell?

The number of firms, brands, and itemsvaries substantially cross ndustries. n a typ-ical month, the canned ham industry had anaverage of thirty irms, ourty-one brands, andeighty-six tems, whereas the ice cream ndus-try had more than eleven times as many firms(342), more than fifteen times as many brands(626), and nearly eighty-five times as manyitems (7,294).

Unlike at the beginning of the 1990s, thenumber of branded tems and firms s fallingin the majority of categories. Although thenumber of items and firms is more likely tobe increasing n categories where total quan-tity is growing, he relationship etween quan-tity growth and growth n the number of itemsor firms s not a strong one. Counterexamplesinclude ready-to-drink ea products, wherequantity grew at 13.2% per year, yet the num-ber of branded irms elling these products ellby 8.3% per year and the number of brandeditems dropped at a 7.9% annual rate.

Does the number of items per firm fallas private-label penetration ncreases? To an-swer this question, we regressed the log ofthe ratio of items to firms on private-labelshare and three seasonal dummies, orrectingfor first-order autocorrelation. Table 4 showsthat differentiation-as measured by itemsper firm-does not increase with private-labelcompetition, contrary to our stylized "fact".There is a statistically ignificant lasticity ofthe items per firm with respect o private-labelshare in only seven categories. Of these, onlyone elasticity s positive. Of course, it is stillpossible that name-brand irms increasinglymodify their products' characteristics or raisethe quality of existing products.

Another indicator of differentiation by firmsis the rate at which they create new products(and kill off old ones). The average number ofbirths and deaths per firm are roughly equal inmost categories, which is consistent with rel-atively small rates of increase or decrease inthe number of items per firm over time. Addi-

Table 4. Elasticities of Number of Name-Brand Items, Number of Firms, and Items perFirm with Respect to the Private-Label Share

Number Number Itemsof of

perItems Firms Firm

Baked beans -0.07a -0.02 -0.03Butter -0.19b -0.12 -0.06Canned ham 0.00 0.02 -0.01Canned uices -0.09 0.02 -0.23bCottage cheese -0.22b -0.28b 0.08Crackers -0.03 0.02 -0.05desserts 0.02 -0.01 0.00English muffins 0.14 0.19 -0.08Frosting -0.06 0.00 -0.07Frozen baked goods -0.04 -0.01 -0.06Frozen breakfast food 0.12b 0.12b 0.02Frozen ruit -0.27 -0.41 0.25Frozen poultry 0.03a 0.14b -0.13bGelatin mixes 0.02 -0.01 -0.03Hot cereal -0.06a 0.04 -0.11bIce cream -0.01 0.07 -0.16bInstant potatoes -0.01 -0.08 0.08Mustard and ketchup 0.03 0.05a -0.02Peanut butter 0.09 0.14 -0.05Pickles and relish -0.01 0.00 -0.03aPizza products 0.03 0.04 -0.01Pizza, refrigerated 0.10 0.15b 0.03Popcorn and popcorn oil 0.10 0.27a -0.10Rice and popcorn cakes -0.03 0.15b -0.19bShortening nd oil -0.02 0.07 -0.10Snack bars/granola bars 0.07 0.11b 0.12bSpaghetti/Italian auce 0.02 0.03 -0.03Sugar substitutes -0.27b -0.18b -0.10bTea, ready o drink -0.13a -0.06 0.03Tomato products 0.00 -0.02 0.06Vinegar -0.17 -0.32b -0.02Yogurt -0.02 -0.08 0.02

aReject the hypothesis that the elasticity (coefficient) is zero at 0.10 confi-dence level.bReject the hypothesis that the elasticity (coefficient) is zero at 0.05 confi-dence level.Note: Not reported are three quarterly dummies. Autocorrelation correctionimposed.

tionally, this observation is consistent with theflagpole theory in which firms are constantlycreating new items and removing old items totry to stay abreast of changing consumer tastes.

The quantity share of new and deceasedproducts out of all branded items is relativelysmall. (The share of births is from the cur-rent period and the share of deceased goods isfrom the previous period.) One might expectthat firms would choose to eliminate unsuc-cessful products with small shares. Given thatthere are roughly an equal number of birthsand deaths, one might infer that the collectiveshare of births would be larger than that of


13/14

972 November 2002 Amer J. Agr. Econ.

deceased goods. In only a few sectors is theshare of births s substantially arger han thatof cancelled products. For example, he shareof new canned ham tems and new frozen poul-try items are roughly twice that of deceased

items. In most sectors, he shares of the birthsand the deaths are roughly equal.We might also expect that the prices of both

new and eliminated tems would be lower thanthe prices of other items. New products mighthave low introductory rices.Before eliminat-ing an unsuccessful product, a firm might trylowering ts price. However, he prices of bothnew and eliminated tems are roughly equaland are usually fairly close to the prices ofother goods (the price ratios in the table areroughly equal to one). One notable exceptionis popcorn and popcorn oil where the pricesof new and eliminated tems are only abouttwo-thirds hat of continuing tems. Anotherinteresting exception is mustard and ketchupwhere new (more specialized?) tems typicallycost twice as much as do existing products, ndeven eliminated tems are 50% more expen-sive than continuing tems.

Summary nd Conclusions

Using statements by industry xperts, we col-lected a set of stylized facts about the inva-sion of private-label products nto supermar-kets. We then used supermarket canner datato examine the validity of these beliefs.

The first belief is that private-label roductscontinue to expand their share of food indus-tries. This view is correct in only some foodcategories. We find that the quantity share ofprivate-label and generic) goods is increasingin less than half of food and beverages cate-gories, but this share s growing at double-digitannual rates for one in four categories.The second set of beliefs concerns the re-sponse of name-brand irms to the increasedcompetition from the private labels. Theconventional ndustry wisdom is that name-brand irms defended their brands against heprivate-label nvasion by lowering heir prices,engaging in additional promotional activities,and increasingly differentiating their products.Our empirical evidence is inconsistent withthese beliefs.

Increases in the share of private-label goodsare correlated with a rise in the price of name-brand goods. Because an increase in private-label share has no effect on or decreases theprice of private-label goods, the overall price

level usually (but not always) remains un-changed or falls. As the share of sales toprivate-label products rise, name-brand firmshold fewer sales and decrease the share ofitems with nonprice promotions. When the

private-label share increases, the number ofname-brand items per firm-a measure ofdifferentiation--is unchanged or falls.

[Received October 2000; final revisionreceived September 2001.]

References

Anderson, S.P.,A. de Palma, and J.-E Thisse. Dis-crete Choice Theory of Product Differentiation,

Cambridge, MA: MIT Press, 1992.Beckett, J. "Grocers Revamp Brands: Food

Chains Improve Look of Private Labels." SanFrancisco Chronicle, 2 March 1992, p. B1.

Bontems, P, M. Dilhan, and V. Requillart. "Strate-gic Effects of Private Labels." Eur. Rev. Agr.Econ. 26(June 1999):147-65.

Caves, R.E., M.D. Whinston, and M.A. Hurwitz."Patent Expiration, Entry, and Competition nthe U.S. Pharmaceutical ndustry." rookingsPapers on Econ. Activity (Microeconomics),1991,pp. 1-48.

Conner, J.M., and E.B. Peterson. "Market Struc-ture Determinants f National Brand-PrivateLabel Price Differences of Manufactured oodProducts." . Indust. Econ. 40(June 1992):157-71.

Cotterill, R.W., and W.P Putsis, Jr. "Market Shareand Price Setting Behavior or Private Labelsand National Brands." Rev. Indust. Org.17(August 000):17-39.

de Jonquieres, G., and N. Tait. "A Trolley Full ofTroubles." inancial Times, May, 1993,p. 17.

Dhar, S.K., and S.J. Hoch. "Why Store Brand

Penetration Varies by Retailer." Mktg. Sci.16(3)(1997):208-27.

Dierker, E., and H. Dierker. "Product Differenti-ation and Market Power." Adv. Math. Econ.1(1999):39-67.

Doug, K. "Brand Ho! Bakery Industry Trends or1996." Bakery Production and Marketing, 15May 1996.

Economist, The. "Make It Your Own." 4 March1995,p. 8.

Frank, R.G., and D.S. Salkever. "Pricing, PatentLoss and the Market for Pharmaceuticals."

S. Econ. J. 59(October 1992):165-79.-

. "Generic Entry and the Pricing of Phar-maceuticals." J. Econ. and Manage. Strategy.6(Spring 1997):75-90.


14/14

Effects of the Private Label Invasion in Food Industries

Documents

Transcript of Effects of the Private Label Invasion in Food Industries