Jmkg.73.6.154

Alexander Krasnikov, Saurabh Mishra, & David Orozco Evaluating the Financial Impact of
Branding Using Trademarks: A
Framework and Empirical Evidence
Firms spend considerable efforts to build brand awareness and associations among consumers. Yet there is alimited understanding of the financial returns of such investments. In this article, the authors present a frameworkthat uses trademarks as measures of firms’ branding efforts. They classify trademarks into two categories—brand-identification trademarks and brand-association trademarks—and propose that they are indicators of firm efforts tobuild brand awareness and associations among consumers, respectively. The authors then evaluate the chain ofeffects linking such assets with metrics of firms’ financial value. A longitudinal analysis of data collected fromsecondary sources reveals that the stock (i.e., total number) of brand-association trademarks available to firms intime period t increases their cash flow, Tobin’s q, return on assets, and stock returns and reduces their cash-flowvariability in period t + 1. Furthermore, the authors observe that the stock of brand-identification trademarks ownedby firms in period t – 1 influences the effects of brand-association trademarks on cash flow, Tobin’s q, and stockreturns. Together, these findings provide useful insights into the financial value of branding.
Keywords: branding, trademarks, brand awareness, brand associations, financial performance, random-coefficients Marketingmanagers spend a considerable portion of focuses on the financial outcomes of brand equity, such
their budgets to build and manage brand equity as shareholder value (e.g., Kerin and Sethuraman 1998).
(Madden, Fehle, and Fournier 2006). However, However, because measures that capture consumer brand they face increasing pressures to justify the financial returns awareness and associations are often difficult to link on such expenses (Srivastava, Shervani, and Fahey 1999).
with financial-market outcomes (Ailawadi, Lehmann, and Consequently, brand valuation has taken a central role in Neslin 2003), there have been limited efforts to link the two both academic research and practitioner research (Salinas and Ambler 2008), with researchers and managers alike Some researchers have attempted to bridge the two stressing continuing investigations in this area (Marketing views (e.g., Mizik and Jacobson 2008; Shankar, Azar, and Science Institute–Emory Marketing Institute 2007).
Fuller 2008). The focus of such studies, however, has pri- Among the many issues studied, researchers have marily been on finding the financial value of consumer emphasized the importance of linking the consumer-based brand associations, with limited attention devoted to under- and financial-market-focused perspectives on brand equity standing the role of brand awareness. Consumers’ brand (Keller and Lehmann 2006). The consumer-based view awareness is an important dimension of brand equity emphasizes efforts that build brand awareness and associa- (Keller 1993) and is a prerequisite for building brand asso- tions among consumers to enhance brand equity (e.g., ciations (Krishnan 1996). Consequently, mere valuation of Keller 1993). In contrast, the financial-market perspective brand associations may be insufficient. Rather, an investiga-tion into the following question becomes important: What Alexander Krasnikov is Assistant Professor of Marketing, School of Busi- are the financial returns to firms from building brand aware- ness, George Washington University (e-mail: [email protected]).
Saurabh Mishra is Assistant Professor of Marketing, Desautels Faculty of We attend to this question in this research. Specifically, Management, McGill University (e-mail: [email protected]).
we present a framework that builds on the close relationship David Orozco is Assistant Professor of Business Law, School of Business between brands and trademarks (Aaker 1991; Cohen 1986) and Economics, Michigan Technological University (e-mail: dorozco@mtu.
edu). The authors acknowledge the financial assistance provided by the to argue that firms’ trademark activities capture a significant Marketing Science Institute and the Emory Marketing Institute to conduct portion of their branding efforts. We classify trademarks this research. The study was motivated by discussions with faculty mem- into two broad categories—brand-identification trademarks bers at the Kellogg School of Management’s Center for Research in Tech- and brand-association trademarks—and propose that they nology and Innovation. The authors also thank the three anonymous JM are indicators of firms’ efforts to build brand awareness and reviewers and the special issue editor Rajendra K. Srivastava for their associations, respectively, among consumers. We then eval- valuable comments. The authors contributed equally to the article and arelisted alphabetically.
uate the links of these two types of trademarks with multi-ple metrics of firms’ financial performance.
2009, American Marketing Association
Journal of Marketing
ISSN: 0022-2429 (print), 1547-7185 (electronic)
Vol. 73 (November 2009), 154–166
Our analysis confirms that both brand-identification and Trademarks: Definition, Types, and Link with
brand-association trademarks affect firms’ financial value.
Brand Equity
In particular, the stock (i.e., total number) of brand- The U.S. federal law (Lanham Act, 15 U.S.C. § 1127 association trademarks that firms own in time period t are [1982]) defines trademark as “any word, name, symbol or observed to increase their cash flows, Tobin’s q, return on device, or combination thereof, adopted and used by a assets (ROA), and stock returns and to reduce their cash manufacturer or merchant to identify his goods and distin- flow variability in period t + 1. Furthermore, we observe guish them from those manufactured or sold by others.” This that the stock of brand-identification trademarks owned by definition parallels that of brand, which has been defined as firms in period t – 1 influences the effects of brand- “[a] name, term, design, symbol, or any other feature that association trademarks on cash flow, Tobin’s q, and stock identifies one seller’s good or service as distinct from those of other sellers” (Bennett 1995, p. 27), thus implying the Together, our findings provide useful insights into the close relationship between trademarks and brands.
financial value of branding and result in multiple theoretical Although trademarks have been recognized under U.S.
and managerial contributions. First, we provide a compre- law for a long time, the federal Trademark Act of 1946, hensive understanding of the chain of effects that links commonly known as the Lanham Act, provides important firms’ efforts to build consumer brand awareness and asso- guidelines for their precise definition and eligibility require- ciations with their financial performance. Second, by using ments (Oathout 1981). Under the Lanham Act, to register trademarks in brand-valuation research, we encourage mar- trademarks with the U.S. Patent and Trademark Office keting scholars to focus more on these objective and easily (PTO), the onus lies on the registrant firm to prove (1) that accessible assets in their research. Last, many firms report its trademarks are unique and/or (2) that consumers readily that managers pay inadequate attention to trademarks identify them with the firm (Melton 1979). Firms often (Bosworth 2003). Our findings should encourage managers establish these conditions in courts by documenting adver- to focus more on their firms’ trademark activities. Next, we tising and promotional spending that links their brands with present our conceptual framework, discuss our methodol- trademarks, by providing survey reports of consumers’ ogy and results, and close with implications and limitations mental representations of their trademarks, or by presenting expert testimony (Jacoby 2001). These requirements ensurethat a firm’s trademarks are uniquely and strongly linked Conceptual Framework
with its brands in the minds of consumers, thus lending sup-port to our thesis that trademarks are a measure of a firm’s Brand Equity and Its Dimensions
Brand equity has been defined from multiple perspectives in Furthermore, the legal criteria set by the Lanham Act the extant literature (Keller and Lehmann 2006). In this and numerous subsequent court decisions have led to the research, we follow the consumer-based perspective, which registrations of various trademarks, which can be classified defines brand equity as “the differential effect that brand into two broad categories that parallel the brand-awarenessand brand-association dimensions of brand equity discussed knowledge has on consumer response to the marketing of previously (see also the Appendix). The first category com- that brand” (Keller 1993, p. 2). To understand how con- prises trademarks that are related to brand names (e.g., sumers store and access brand knowledge, researchers have Nike), brand logos or brand symbols (e.g., Nike’s swoosh), built on extant memory frameworks laid out in the psychol- or a combination of the two. Because such trademarks rep- ogy literature (e.g., Wyer and Srull 1989).
resent brand identifiers, they capture a significant portion of With respect to brand knowledge, these frameworks firms’ efforts to build brand awareness among consumers.
imply that the brand name or identifier is a central concept We refer to names, logos, and symbols as brand- (Nedungadi 1990), which, when developed in memory, identification trademarks in this research.
helps consumers recall important attribute and nonattribute The second category of trademarks encompasses the associations they attach to the brand (Krishnan 1996).
different attribute and nonattribute associations that con- Attribute-based associations are the links that consumers sumers attach to brands. Trademarks that capture attribute hold between brands and product characteristics (or attrib- and nonattribute associations can be decorative or informa- utes), such as unique product color, scent, package, shape, tional (U.S. Trademark Manual of Examination Procedures and sound (e.g., Srinivasan, Park, and Chang 2005). Non- §§ 1202.03–1202.04). Decorative trademarks extend to attribute-based associations include consumer perceptions design elements (e.g., color, motion, packaging, scent, of brand-use imagery or brand personality (Mizik and shape), whereas informational trademarks include slogans (e.g., Nike’s “Just Do It”), which may communicate image- The previously mentioned frameworks suggest that related messages that complement the brand (Jacoby 2001).
brand equity, as represented in consumers’ memory, has two Together, this category of brand-association trademarks dimensions—namely, brand awareness and brand associa- reflects a significant portion of a firm’s efforts to develop tions, with brand awareness being a necessary precondition consumer brand associations.1 We subsequently describe for the creation of strong brand associations (Keller 1993).
Trademarks capture a significant portion of firms’ efforts tobuild brand awareness and associations among consumers.
1Certain aspects of a firm’s branding strategy affect consumer We explain this argument in more detail next.
brand associations but cannot be protected through trademarks.
Evaluating the Financial Impact of Branding / 155
how the two types of trademarks, given their close relation- brand equity from counterfeiting or trademark infringement ship to consumer brand awareness and associations, may (e.g., Morrin, Lee, and Allenby 2006).
Finally, trademark value can also be inferred from the manner in which firms often leverage trademarks through Brand Equity, Trademarks, and Firm Value
licensing (Jacoby 2001). Licensing involves a firm granting Industry experts often suggest that trademarks are important permission to a third party (i.e., a licensee) to use the firm’s intellectual property assets that enhance shareholder value trademarks in association with the licensee’s products and (e.g., Davidson 2004). Several observations of industry services in exchange for royalties. For example, the leading practice also support this view. For example, in 2007, the pharmaceutical firm AstraZeneca licensed the trademark for retail company Sears Holding Corp. transferred some of the its Prilosec heartburn medication and the color purple asso- brand equity residing in its trademarks to a wholly owned ciated with the product to Procter & Gamble (P&G), which legal vehicle to issue $1.8 billion in securities (Business- launched it as an over-the-counter medicine (U.S. Trade- In examining the source of this value, we direct our In summary, we expect that both brand-identification attention to the two types of trademarks discussed previ- and brand-association trademarks are linked with firm ously. With respect to brand-identification trademarks, such financial value. Furthermore, with respect to formulating as brand names, logos, and symbols, extant literature points the chain of effects that links these two types of trademarks to at least two characteristics that attest to their value poten- with financial performance, extant consumer-based brand tial. First, such trademarks enable consumer recognition of equity frameworks (e.g., Keller 1993; Krishnan 1996) indi- brands in the crowded marketplace (Henderson and Cote cate that the stock of brand-identification trademarks a firm 1998). Second, brand identifiers also often serve as impor- owns provides a base (or a precondition) on which the firm tant predictive cues of product performance to consumers can build a strong regime of brand-association trademarks (Erdem and Swait 1998). Therefore, we can expect that a strong regime of brand-identification trademarks positivelyaffects consumer preference for the brand.
Methodology
Beyond brand-identification trademarks, brand- association trademarks may provide significant value to Database Overview
firms. Brand associations, as reflected in such trademarks, To evaluate our conceptual model, we compiled a data set have been argued to positively affect consumers’ brand- by integrating information from several secondary sources, related attitudes (Keller 1993). Furthermore, a more com- including Standard & Poor’s COMPUSTAT database, the plex network of associations has been shown to give con- University of Chicago’s Center for Research in Security sumers greater confidence in their attitudes (Pullig, Prices (CRSP), the PTO’s Trademark Electronic Search Netemeyer, and Biswas 2006), which makes them less System (TESS), and firms’ annual reports.
prone to attitude change (Pham and Muthukrishnan 2002) The firms in the final sample were from the following and helps attenuate the effect of competitors’ persuasion industries: beverages, apparel, computer and communica- attempts (Pechmann and Ratneshwar 1991). In addition, tion equipment, confectionary, department stores, eating firms may leverage consumer brand associations to intro- places, grocery stores, home appliances, jewelry, motor duce brand extensions, which can help firms enter new, vehicles, packaged food, perfumes and cosmetics, software, often more profitable product markets (e.g., Dacin and and wine and malt beverages. We focused on multiple industries to enhance the generalizability of our findings.
Furthermore, there is evidence that a firm’s efforts to From these industries, we selected firms for which informa- build strong brand equity that is secured against dilution tion required for the analysis was available in the COMPU- through trademarks have a positive impact on its financial STAT and CRSP databases for the period 1995–2005.
value. Prior research has suggested that strong brand equity We focused on the 1995–2005 period for two reasons.
helps firms generate higher revenue premiums, reduce vul- First, a longitudinal assessment helps increase the robust- nerability to competitive actions, increase long-term effec- ness of our findings and is in line with previous marketing tiveness of promotions, and lower promotional expenditures research on brand valuation (e.g., Bahadir, Bharadwaj, and for subsequent brand extensions (e.g., Ailawadi, Lehmann, Srivastava 2008; Rao, Agarwal, and Dahlhoff 2004). Sec- and Neslin 2003; Slotegraaf and Pauwels 2008; Srinivasan, ond, and more important, several important legal develop- Park, and Chang 2005), all of which positively influence ments in the early 1990s (e.g., the 1989 Trademark Revi- firms’ financial value (Mizik and Jacobson 2008). Indeed, sion Act, the 1992 U.S. Supreme Court ruling on researchers have emphasized the importance of trademarks nontraditional trademarks) increased the legal certainty of in protecting this value by drawing attention to the threats to certain trademarks such as sound, scent, shape, and motionmarks. Such trademarks enable firms to establish and pro-tect important aspects of attribute-based consumer brandassociations. Consequently, an accurate measurement of Nevertheless, we believe that trademark registration activities brand-association trademarks, one of our key independent reflect a significant portion of firms’ efforts to create brand asso- variables, necessitated the inclusion of these trademarks in ciations among consumers. We thank an anonymous reviewer for our analysis. Our assessment of trademark registrations by firms in our sample confirmed that the period following 156 / Journal of Marketing, November 2009
1995 most accurately captured sound, scent, shape, and Overall, the foregoing criteria resulted in a sample of ijk = the difference between stock return for firm i in industry j and month k and the return on 108 firms, with an approximately 50–50 split of manufac- turing and service firms. In the following sections, we k = the value-weighted return on all stocks listed in the New York Stock Exchange,American Stock Exchange, and NASDAQ Measures
less the one-month T-bill rate in month k, Financial performance. To evaluate the financial value HMLk = the book-to-market factor adjusted for of branding efforts, we focused on multiple measures of firms’ financial performance. First, we included cash flow SMBk = the size-based risk factor in month k, and cash flow variability as the main dependent variables of UMDk = the monthly momentum factor (up minus interest in this research. Economic theory posits that a firm’s ability to generate future cash flows determines the ωijk = residuals for firm i in industry j and month firm’s value (Rappaport 1986). Indeed, given the impor- tance of cash flow to firms, a significant volume of research Moreover, as Luo (2009) suggests, we corrected for serial has emerged on this topic in the fields of finance and correlations by regressing ωijk on its lagged value and by accounting (e.g., Dechow, Kothari, and Watts 1998; Ismail and Choi 1996). In the marketing literature, Gruca and returns. Then, following Mizik and Jacobson (2008), we Rego (2005) document the importance of cash flows as aggregated the monthly stock return data to annual for firm measures of financial performance and provide guidelines i in industry j and year t (RSTijt) using Equation 2: to capture these metrics. We followed their approach in cal-culating cash flow levels and variability in this research, using information collected from COMPUSTAT.
Second, although we focused on cash flow and cash flow variability, we recognize that there are alternative As a result, we obtained 924 annual stock return values for measures of shareholder value. Therefore, to enhance the validity of our findings, we also included Tobin’s q, ROA,and stock returns as additional measures of financial value.
Trademark-based measures of brand awareness and Tobin’s q, the ratio of a firm’s market value to the brand associations. We extracted detailed information onall trademarks registered by the 108 firms in our sample replacement cost of its assets, is a forward-looking measure from the PTO TESS database up to the year 2005. During that summarizes investors’ expectations regarding a firm’s this process, we retrieved the following information for potential to generate future revenues (Lindenberg and Ross each trademark: serial number, registration number, year of 1981). Previous research in brand valuation has also used registration, owner name, trademark description, trademark Tobin’s q to evaluate the financial value of different brand- drawing codes, and status (live or canceled). While retriev- ing strategies (e.g., Rao, Agarwal, and Dahlhoff 2004). We ing the trademark data, two authors with the help of two calculated Tobin’s q using the methodology that Chung and research assistants manually verified that the trademarks Pruitt (1994) outline, with information collected from corresponded to the firm in question. Overall, we identified 22,060 live and registered trademarks. Two independent In addition to Tobin’s q, extant research on financial coders then coded the trademarks as either brand- statement analysis (e.g., Fairfield, Sweeney, and Yohn 1996) identification or brand-association trademarks. A detailed implies that financial ratios, such as ROA, provide informa- coding plan was developed to assist the coders. The plan tion about future profitability to investors. Therefore, we included the definition and examples of the two categories also included ROA as a measure of financial performance, of trademarks (examples were the same as those given in using information obtained from COMPUSTAT.
the Appendix) and a coding nomenclature.
Finally, we evaluated stock returns as an additional In developing our coding plan, we followed the classifi- measure of financial value. Researchers in marketing have cations discussed previously. Trademarks that included increasingly adopted this forward-looking measure to eval- brand names and/or symbols in the description field were uate the value relevance of marketing activities and con- coded as brand-identification trademarks. Furthermore, the sumer brand perceptions (Luo 2009; Mizik and Jacobson coders identified trademarks specifying brand attribute or 2008). For our analysis, we used the Fama and French image as brand-association trademarks. For example, they (1993) momentum multirisk market model to measure stock coded Target Corp.’s trademark Archer Farms as a brand returns. Specifically, we calculated the monthly stock identifier but coded its slogan “Expect more. Pay less.” as a returns for firm i in industry j and month k (STijk) usingEquation 1: Fama–French factors were downloaded from http://mba.tuck.
dartmouth.edu/pages/faculty/ken.french/index.html.
Evaluating the Financial Impact of Branding / 157
nonattribute trademark within the category of brand- industry, and period). Therefore, we used a growth model, a association trademarks. Similarly, they coded Apple’s iPod special case of random coefficients models, to estimate the shape trademark as an attribute-based trademark, also in the impact of the two types of trademarks on firms’ value. Such category of brand-association trademarks.
an approach is relevant for multilevel data and helps both Through this process, we identified 4146 brand- model different sources of observed heterogeneity, which is association trademarks and 17,914 brand-identification common in nested designs, and account for any unobserved trademarks. Initial coder agreement was high (89%), and firm- and industry-specific effects (Bryk and Raudenbush the coders resolved their differences to produce a final set of codes. Furthermore, to verify consistency of our trade- We modeled variation in the financial outcome variable mark classification plan with experts in trademark law, we Indijt (i.e., cash flow, cash flow variability, Tobin’s q, ROA, contacted a large midwestern law firm through a survey.
and stock returns) for firm i in industry j in period t as a Thirty lawyers completed and returned the survey.3 Their function of lagged earnings (Earnijt – 1) and the stock of all overall agreement with our coding scheme was high (88%).
live brand-association trademarks (BrAijt – 1) in period t – 1: Finally, we verified the coding of a subset of trademarks with the classification plan used by the PTO. For motion, ijt = β0ij + β1ij × BrAijt – 1 + β2ij × EARNijt – 1 + εijt.
sound, scent, or shape trademark registrations, the PTO Furthermore, it is possible that there is heterogeneity assigns a specific drawing code. We extracted all trade- related to firm-level characteristics. Consequently, in Equa- marks with this drawing code and evaluated their corre- tion 4, we modeled the variation in intercept and slopes as spondence with our coding. We identified no discrepancies with our classification. Together, these steps gave us confi- dence that our measures adequately captured the branding 0ij = γ00j + γ01j × SIZEij + γ02j × RDij + γ03j × ADij efforts of firms in our sample. In the final analysis, we used the stock of all live brand-association and brand- identification trademarks available to a firm in a given year 1ij = γ10j + γ11j × BrAwijt – 2 + η1ij, and Controls. In addition, following the work of Gruca and Rego (2005), we included several firm- and industry- specific controls in our analysis. We outline these in detail in the next section. Tables 1 and 2 present details on the ij = research and development intensity of firm i in ADij = advertising intensity of firm i in industry j, and Model Formulation
ij = trademark intensity of firm i in industry j.
Both conceptual and empirical considerations shaped our The normally distributed error terms η0ij, η1ij, and η2ij model formulation. First, our conceptual framework sug- captured unobserved effects specific to firm i in industry j.
gested that investments in building a strong network of Because we observed that advertising, research and devel- brand associations increase a firm’s financial performance.
opment, and trademark intensities did not vary much during This is in line with previous research in marketing that has the 11-year time frame covered in our research, we aver- established that certain types of brand associations enhance aged the values for each firm. We then specified the estima- shareholder value (e.g., Mizik and Jacobson 2008). In our tion for capturing the variation in slope for BrAijt – 1 (β1ij).
model, we accounted for this relationship by introducing a Previously, we posited that brand associations are built on a main effect of brand associations on the outcome variables.
firm’s prior efforts to build brand awareness among con- However, we also posited that research in consumer behav- sumers. As such, we expected that the variation in the ior suggests that the effectiveness of brand associations impact of brand-association trademarks in period t – 1 depends on consumers’ brand awareness (Krishnan 1996).
(BrAijt – 1) on financial performance and shareholder value We accounted for this relationship by incorporating an in period t would be contingent on the stock of brand- interaction between brand-identification trademarks and identification trademarks available in period t – 2 (BrAwijt – 2).
brand-association trademarks on firm performance in our To incorporate this interaction, we adopted recommenda- tions for conducting moderation analysis in random coeffi- Second, our sample design also determined our model- cients models (Hofmann and Gavin 1998) and performed ing approach. We used a cross-sectional and longitudinal within-firm centering of our Level 1 variables in Equation 3 panel sample with multiple levels of observations (by firm, Furthermore, because our sample draws from multiple industries, there may be heterogeneity in parameter esti-mates in Equation 4 as a result of industry effects. There-fore, in Level 3 (Equation 5), we modeled the variation infirm-level effects (γ 3On average, respondents were familiar with trademark laws 00j, γ10j, and γ20j) using industry-level (reporting an average of 5.43 on a seven-point scale measuring variables. The terms ξ00j, ξ10j, and ξ20j are normally distrib- “How familiar are you with trademark laws?”) and had an average uted and account for unobserved effects that are specific to 158 / Journal of Marketing, November 2009
Description of Variables
Variable
Description
Cash flows from operations (in millions of dollars) Ratio of firm’s quarterly cash flow coefficient of variation (standard deviation divided by mean for a given year) to market’s quarterly cash flow The following ratio: [(number of common shares outstanding × share price + liquidating value of preferred stock + book value of long-term debt + short-term liabilities – short-term assets)/(book value of total assets)] Ratio of net income before extraordinary items to total assets Abnormal stock price fluctuations after controlling for the average market portfolio returns in three stock exchanges Net income before extraordinary items (in millions of dollars) Ratio of firm’s annual advertising expenditures to total assets divided by the industry’s average ratio of advertising expenditures to total assets Ratio of firm’s R&D expenditures to total assets divided by the industry’s average ratio of R&D expenditures to total assets Logarithm of total number of firm employees Average five-year sales growth for industry Standard deviation of five-year sales growth for industry Herfindahl concentration index (HHI) (i.e., sum of squared shares of firms Dummy for service (1) versus manufacturing (0) firm Stock of all live brand-identification trademarks a firm owned in a given Stock of all live brand-association trademarks a firm owned in a given year Ratio of firm trademark registrations to total number of trademark registrations in the industry in a given period Notes: R&D = research and development.
γ00j = α001 × DMNDjt + α002 × INSTjt + α003 × HHIjt Finally, we applied an autoregressive structure to the residuals in the growth model to account for autocorrelation in our data.4 We also conducted White’s test to ensure that residuals are homoskedastic for all outcome variables.
To estimate the growth model (Equations 3–5), we used a stepwise approach and modeled different sources of varia- jt = overall demand in industry j in period t, tion (i.e., within firm, industries, and over time) (Bryk and jt = demand instability in industry j in period t, HHIjt = market concentration ratio in industry j in SERVj = indicator for service (1) versus manufactur- 4We conducted a pooled Durbin–Wood test and found evidence Evaluating the Financial Impact of Branding / 159
/ Journal of Marketing,
No
vember 2009
Descriptive Statistics
BrAw EARN
Notes: All correlations greater than .05 and less than –.05 are significant at p < .05.
Raudenbush 1992). Table 3 provides detailed findings from inputs in Equations 3–5, we evaluated a model including only newly registered live association and awareness trade- Consistent with our conceptual framework, we observed marks in a given year. The significance patterns were robust that brand-association trademarks increase cash flows (β = to this specification as well. However, the use of stock of all 7.820, t-value = 3.58) and decrease cash flow variability trademarks rather than only stock of newly registered live (β = –.044, t-value = –1.97). Moreover, we observed that trademarks as inputs provided better fit with the data. Third, brand-association trademark activity is positively associated we explored whether incorporating more lags in brand- with Tobin’s q (β = .007, t-value = 2.41), ROA (β = .051, identification and brand-association trademarks provided a t-value = 3.16), and stock returns (β = .003, t-value = 5.01).
better fit. By applying Davidson and MacKinnon’s (1981) Together, the findings imply that a firm’s efforts aimed to J-test, we confirmed that including the alternative number establish consumer brand associations enhance its financial of lags does not invalidate our model. Finally, we reesti- value. Next, we found that the interaction between brand- mated our model, including the square of lagged advertising identification and brand-association trademarks was posi- intensity and square of lagged branding efforts, to investi- tively associated with cash flows (γ = .019, t-value = 1.98).
gate nonlinear effects (not reported in Table 3).7 The However, its impact on cash flow variability (γ = .001, parameter estimates for the square of brand-association t-value = .45) and ROA (γ = –.0004, t-value = –1.05) was trademarks in Equation 3 were not significant with respect not significant. Furthermore, we observed that the inter- to all outcome measures except for stock returns, for which action term was negatively associated with Tobin’s q (γ = the estimate was negative and marginally significant (p < –.0001, t-value = –1.96) and stock returns (γ = –.0001, .08). The parameter estimates for the square of brand- t-value = –2.02). As such, our findings provide mixed identification trademarks and the square of advertising effects of brand awareness on firm value.5 intensity (ADij) in Equation 4 were also nonsignificant.
In addition to the effects of primary interest, we Moreover, inclusion of these terms did not improve the observed that prior earnings are positively associated with overall model fit or change the significance of other esti- cash flows (β = .248, t-value = 16.10) and ROA (β = .001, mates reported in Table 3. We discuss these findings in t-value = 3.27); however, prior earnings had a marginally detail in the “Implications” section.
negative effect on stock returns (β = –.001, t-value = –1.80).
Endogeneity. Shugan (2004) suggests that a firm’s past Next, with respect to the role of firm characteristics, our performance determines its marketing investments; there- analysis revealed that larger firms generated greater cash fore, successful firms might engage in more trademark reg- flows (γ = 937.23, t-value = 4.42) and greater ROA (γ = istrations.8 Consequently, it was critical to test for potential 2.249, t-value = 2.86). We also found evidence that relative endogeneity in our model specification. Unfortunately, branding efforts, as captured by a firm’s trademark inten- there are no established instruments that adequately capture sity, are associated with higher cash flows (γ = 978.62, firms’ branding investments. However, on the basis of t-value = 3.85) and Tobin’s q (γ = .478, t-value = 2.97). We extant research in this area, we employed two sets of instru- observed that advertising intensity was marginally nega- ments to assess endogeneity in our analysis. First, empirical tively associated with ROA (γ = –3.360, t-value = –1.83).
findings indicate that a firm’s number of brands and its Among industry characteristics, market demand was posi- sales, general, and administrative expenditures are related to tively associated with cash flows (α = 1576.49, t-value = its branding efforts (e.g., Bahadir, Bharadwaj, and Srivas- 4.29) and stock returns (α = .231, t-value = 3.35), but its tava 2008; Rao, Agarwal, and Dahlhoff 2004). Therefore, impact was not significant for the other outcome variables.
we used these as instruments. We also employed the ratio of The results also suggest that stock returns were greater in firm intangible assets to total assets as an additional instru- less stable industries (α = .181, t-value = 2.52). Finally, we ment in this step. The Hausman test failed to reject the null found that Tobin’s q was marginally greater in service- hypothesis of exogeneity for all five outcome variables.
based industries than in manufacturing industries (α = .641, Second, following McAlister, Srinivasan, and Kim’s (2007) approach, we checked for endogeneity using predicted val- Robustness Check
ues of brand-association trademarks, earnings, and brand-identification trademarks, which we obtained from regres- Model specification. We also estimated alternative mod- sion of those variables on their respective one-year lagged els to test the robustness of our model specification. First, values as instruments. The results of this variable estimation instead of distinguishing between brand-identification and procedure were consistent with our findings reported in brand-association trademarks, we used a composite mea- Table 3. The Hausman tests also failed to reject the exo- sure capturing the stock of all live trademarks available to a geneity hypothesis for all dependent variables except for firm in a given year as our key predictor variable. Compari- cash flows. In addition to instrument variable analysis, we son of fit indexes of the two approaches supported our conducted a bivariate Granger causality test to evaluate model (Equations 3–5) for all five outcome variables.6 Sec- noncausality among the dependent variables and the mea- ond, instead of using the stock of all live trademarks as sures capturing brand identification and association trade- 5Following our conceptual framework, we did not evaluate the main effects of brand-identification trademarks in our estimations.
6Model comparison details are available from the first author on 7We thank an anonymous reviewer for this suggestion.
8We thank an anonymous reviewer for this comment.
Evaluating the Financial Impact of Branding / 161
/ Journal of Marketing,
No
vember 2009
Impact of Brand-Identification and Brand-Association Trademarks on Firm Performance and Shareholder Value
Cash Flows
Cash Flow Variability
Tobin’s q
Stock Returns
Predictors/
Financial
Parameter Estimate
Parameter Estimate
Parameter Estimate
Parameter Estimate
Parameter Estimate
Indicators
(t-Value)
(t-Value)
(t-Value)
(t-Value)
(t-Value)
Firm Characteristics
Industry Characteristics
marks. We used one- and two-year lags of the outcome accounting information for small or newly established firms variables for the test (Greene 2003). The test did not sup- from secondary databases, we were restricted to relatively port the hypothesis that our outcome measures cause large firms with established brands in our sample. Conse- changes in the two types of trademark registrations. Finally, quently, investors might discount the brand-awareness we asked lawyers in our survey to indicate their agreement with the following statement: “Only successful firms Second, the investor recognition hypothesis offered in engage in trademark registrations.” More than 89% of the finance literature may also explain our results (Merton respondents disagreed with this statement. On the basis of 1987). According to that hypothesis, advertising and other these observations, we believe that endogeneity does not marketing efforts that increase a firm’s visibility among present a serious problem in our analysis. However, it consumers also attract individual investors to the firm’s should be noted that because there are no established instru- stocks (Grullon, Kanatas, and Weston 2004; McAlister, ments that adequately capture firms’ branding investments, Srinivasan, and Kim 2007). It has been documented that and given the nature of our data, we cannot completely rule individual investors, as opposed to institutional investors, often face cognitive constraints and information collectioncosts, which restrict their ability to incorporate all relevant Discussion and Implications
information (e.g., changes in cash flows) accurately andinstantaneously in their stock valuation (Merton 1987).
Extensive research in marketing has encouraged marketing Consequently, financial models predict that increased indi- managers to focus on building brand equity by enhancing vidual investor ownership of a firm’s stock decreases its consumers’ awareness of and associations with brands short-term stock returns (Merton 1987). Therefore, it is pos- (Keller 1993). However, the extant literature offers limited sible that the brand-awareness efforts of firms attract more insights into the financial returns of such efforts (Keller and individual investors to their stocks, thereby attenuating the Lehmann 2006). The primary focus of our research was to stock returns and Tobin’s q value of such firms. To evaluate offer a better understanding of the benefits of such a focus.
this thesis, we obtained stock-ownership information of the Specifically, we evaluated the financial value of brand- firms in our sample from the Thomson Reuters database.
ing by linking trademark registrations of firms with their Using median split, we classified our sample into two financial performance. We broadly classified trademarks groups (high and low) according to the number of brand- into two categories—brand-identification and brand- identification trademarks weighted by the number of brands association trademarks—and proposed that they are indica- the firms owned. A comparison of the proportion of stocks tors of firms’ efforts to build consumer brand awareness and owned by individual (as opposed to institutional) investors associations, respectively. Our examination of 22,060 trade- in the two groups demonstrated that the proportion of indi- marks registrations of 108 firms, across multiple industries, vidual investors owning the stocks was greater for the group and in the period 1995–2005, confirms that efforts aimed to of firms with more brand-identification trademarks (z = build brand awareness and associations among consumers 3.01, p < .05). Though not conclusive, this observation have significant financial implications for firms. Overall, provides preliminary support for the investor recognition we observed that brand-association trademarks positively affect firm cash flows, Tobin’s q, ROA, and stock returns.
Furthermore, brand-association trademarks help reduce the Research Implications
variability of future cash flows. Of particular significance is By using trademark registration information to capture the the observation that, on average, each additional brand- financial value of branding, our study makes several contri- association trademark is associated with $7.8 million of butions to marketing theory. First, it addresses researchers’ future cash flows, a .05% increase in future ROA, and a .3% recent calls to formulate new methodological approaches increase in the future stock returns of a firm. In addition, the that bridge the gap between consumer-based and financial- findings confirm that by improving consumers’ awareness market-focused perspectives on brand equity (Keller and of brands, firms enhance the future cash flows generated by Lehmann 2006). Extant research focusing on linking consumer-based measures of brand equity with shareholder Although we expected a similar positive effect of value has primarily relied on proprietary data (e.g., Mizik consumer-brand awareness on stock market measures, our and Jacobson 2008; Shankar, Azar, and Fuller 2008), which analysis revealed otherwise. We observed that increasing may not be readily available to many researchers. In con- consumer brand awareness diminishes the positive effects trast, to capture brand equity, we provide a framework that of brand-association trademarks on stock returns and uses trademark registration information, which is objective Tobin’s q. There are two potential explanations for our find- and easily available to all firms operating in the United ings.9 First, because it is difficult to find financial and States. To the best of our knowledge, our framework is thefirst to discuss a categorization of trademarks within abranding framework. Future researchers may find our clas- 9Following the suggestion of an anonymous reviewer, we also evaluated whether the presence of outliers in our data may explain Second, extant research on brand valuation has provided this result. We evaluated the presence of outliers using Cook’s dis-tance and restricted likelihood distance parameters. Reestimation rich insights into the financial value of different types of of the model without potentially influential observations did not brand associations held by consumers (e.g., Mizik and Jacobson 2008) but has paid limited attention to the role of Evaluating the Financial Impact of Branding / 163
consumer brand awareness in influencing firm financial additional iPod product shape trademarks. The shape trade- performance. By confirming that consumer brand aware- marks enabled Apple to secure the iPod against competitive ness influences the financial value of brand associations, threats on its innovative design. On the flip side, under cur- our findings may encourage researchers to include brand rent law, a firm may lose its trademark rights if it no longer awareness as an important dimension in models that capture uses the trademark in commerce. Some firms have devel- oped businesses that resurrect unused trademarks, which Finally, as mentioned previously, we did not find evi- negatively affects the original trademark owner and appro- dence for quadratic effects of branding efforts and advertis- priates any brand equity remaining in the unused trademark.
ing intensity on firms’ financial performance. It may be that For example, when P&G discontinued its White Cloud there is an optimal level of branding and advertising efforts, brand of toilet tissue, Wal-Mart, in a move unknown to beyond which such efforts lead to diminishing financial P&G, overtook the trademark and adopted it to market its returns. However, McAlister, Srinivasan, and Kim (2007) private label of paper goods. Marketers can avoid such suggest that firms often lack the appropriate tools to deter- potentially damaging trademark strategies of competitors mine the optimal level of marketing spending and thus by keeping a close eye on their own activities and on their rarely reach the maximum level of financial performance.
competitors’ trademark activities.
This reasoning may potentially explain our findings as well.
However, although our analysis gives us confidence in our Limitations and Future Research
implications, our findings reflect the marginal effects ofbranding at the levels represented in our data. Further Directions
research might explore other firm and industry contexts for Although we offer several important implications, our which nonlinear effects can be observed.
research suffers from a few limitations that bring to lightavenues for future research. First, our sample was restricted Managerial Implications
to larger firms. This was primarily a result of our focus on This research also offers several important managerial the metrics of financial value as dependent variables, for insights. First, our findings should assist marketing man- which longitudinal measures are available only for rela- agers in more cogently communicating the financial value tively large firms. Further research might include smaller of branding to management. This becomes especially firms with newly established brands to investigate whether important during lean economic conditions, when firms the value associated with branding efforts of such firms dif- may be inclined to make cuts in their brand-related invest- fers from what we observed in our sample.
ments. The results imply that such moves may lead to a Second, we propose that trademarks are measures of potential loss of future financial value for firms. Indeed, firm branding efforts. However, we realize that brand- several instances from business practice also reveal that association trademarks do not capture all the dimensions of firms with strong brands, such as McDonald’s, are able to consumer brand associations. For example, as we noted pre- raise prices despite a weakening economic environment viously, firms are often limited in what they can protect through trademarks. Many aspects of marketing strategy Second, scholars have noted that managers rarely work affect a brand’s image but cannot be protected through closely with the legal function of their firms (Bosworth trademarks (e.g., celebrity endorsements, sponsorship 2003). Therefore, it is likely that marketers may not be events, taste tests). Furthermore, it is not possible to capture aware of the many categories of trademarks, such as color, the favorability or uniqueness of brand associations through scent, sound, shape, and motion, that have gained legal trademarks. Capturing such dimensions likely requires a precedence only recently. We hope that, given the value multimethod approach. In the future, researchers should such trademarks generate, marketers will be motivated to focus on complementing trademark information with con- work more closely with their firms’ legal counsel.
sumer attitudinal information to better capture the financial Finally, this research may motivate firms to review their trademark portfolios more closely to uncover unappreciated Finally, analysis that centers on financial performance trademark opportunities and to benchmark against the best as the dependent variable and uses secondary measures to performers. Several examples from industry attest that a capture branding efforts of firms is prone to endogeneity forward-looking and well-executed trademark strategy can concerns. Although we conducted several tests to ascertain help build and protect a firm’s brand equity. For example, whether endogeneity was a serious concern in our analysis, part of Apple’s success in building the strong iPod brand the lack of established instruments to capture firms’ brand- has been attributed to its trademark strategy (Orozco and ing investments did not enable us to completely rule out the Conley 2008). Apple was proactive in registering the iPod endogeneity problem. Further research might aim to find product name, and then it built on that name by registering better instruments for capturing branding efforts of firms.
164 / Journal of Marketing, November 2009
APPENDIX
Trademarks Classification and Examples
Brand-Identification Trademarks
Category
Brand-Association Trademarks
Colors that are part of the physical good progression of the notes D flat, D flat, G REFERENCES
Aaker, David A. (1991), Managing Brand Equity: Capitalizing on
University of Manchester Institute of Science and the Value of a Brand Name. New York: The Free Press.
Ailawadi, Kusum L., Donald R. Lehmann, and Scott A. Neslin Bryk, Anthony S. and Stephen W. Raudenbush (1992), Hierarchi- (2003), “Revenue Premium as an Outcome Measure of Brand cal Linear Models: Application and Data Analysis Methods.
Equity,” Journal of Marketing, 67 (October), 1–17.
Newbury Park, CA: Sage Publications.
Bahadir, S. Cem, Sundar G. Bharadwaj, and Rajendra K. Srivas- BusinessWeek (2007), “The New Alchemy at Sears,” (April 16), tava (2008), “Financial Value of Brands in Mergers and Acqui- (accessed July 8, 2009), [available at http://www.business sitions: Is Value in the Eye of the Beholder?” Journal of Mar- week.com/magazine/content/07_16/b4030071.htm].
keting, 72 (November), 49–64.
Bennett, Peter D. (1995), Dictionary of Marketing Terms, 2d ed.
Chung, Kee H. and Stephen W. Pruitt (1994), “A Simple Approxi- Chicago: American Marketing Association.
mation of Tobin’s q,” Financial Management, 23 (Autumn), Bosworth, Derek L. (2003), “The Importance of Trademarks to Capital Raising and Financial Performance: Lessons from Cohen, Dorothy (1986), “Trademark Strategy,” Journal of Market- SMEs,” working paper, Manchester School of Management, Evaluating the Financial Impact of Branding / 165
Colvin, Geoff (2009), “Yes, You Can Raise Prices,” Fortune, 159 2009), [available at http://www.msi.org/research/index.cfm? Dacin, Peter A. and Daniel C. Smith (1994), “The Effect of Brand McAlister, Leigh, Raji Srinivasan, and Min Chung Kim (2007), Portfolio Characteristics on Consumer Evaluations of Brand “Advertising, Research and Development, and Systematic Risk Extensions,” Journal of Marketing Research, 31 (May), of the Firm,” Journal of Marketing, 71 (January), 35–48.
Melton, Carol A. (1979), “Generic Term or Trademark? Confusing Davidson, Russell and James G. MacKinnon (1981), “Several Legal Standards and Inadequate Protection,” American Univer- Tests for Model Specification in the Presence of Alternative sity Law Review, 29 (1), 109–133.
Hypotheses,” Econometrica, 49 (May), 781–93.
Merton, Robert C. (1987), “A Simple Model of Capital Market Davidson, Scott (2004), “When Trademarks, Trade Names, and Equilibrium with Incomplete Information,” Journal of Finance, Brands Get Valued,” CA Magazine, 137 (7), 36–38.
Dechow, Patricia M., S.P. Kothari, and Ross L. Watts (1998), “The Mizik, Natalie and Robert Jacobson (2008), “The Financial Value Relation Between Earnings and Cash Flows,” Journal of Impact of Perceptual Brand Attributes,” Journal of Marketing Accounting and Economics, 25 (2), 133–68.
Research, 15 (February), 15–32.
Erdem, Tulin and Joffre Swait (1998), “Brand Equity as a Signal- Morrin, Maureen, Jonathan Lee, and Greg M. Allenby (2006), ing Phenomenon,” Journal of Consumer Psychology, 7 (2), “Determinants of Trademark Dilution,” Journal of Consumer Research, 33 (September), 248–57.
Fairfield, Patricia M., Richard J. Sweeney, and Teri Lombardi Nedungadi, Prakash (1990), “Recall and Consumer Consideration Yohn (1996), “Accounting Classification and the Predictive Sets: Influencing Choice Without Altering Brand Evaluations,” Content of Earnings,” The Accounting Review, 71 (July), Journal of Consumer Research, 17 (December), 263–76.
Oathout, John D. (1981), Trademarks: A Guide to Selection, Fama, Eugene F. and Kenneth R. French (1993), “Common Risk Administration, and Protection of Trademarks in Modern Busi- Factors in the Returns on Stocks and Bonds,” Journal of Finan- ness Practice. New York: Charles Scribner’s Sons.
cial Economics, 33 (February), 3–56.
Orozco, David and James G. Conley (2008), “Shape of Things to Greene, William H. (2003), Econometric Analysis, 5th ed. Upper Come,” The Wall Street Journal, (May 12), R6.
Saddle River, NJ: Pearson Education.
Pechmann, Cornelia and S. Ratneshwar (1991), “The Use of Com- Gruca, Thomas S. and Lopo L. Rego (2005), “Customer Satisfac- parative Advertising for Brand Positioning: Association Versus tion, Cash Flow, and Shareholder Value,” Journal of Marketing, Differentiation,” Journal of Consumer Research, 18 (2), Grullon, Gustavo, George Kanatas, and James P. Weston (2004), Pham, Michel Tuan and A.V. Muthukrishnan (2002), “Search and “Advertising, Breadth of Ownership, and Liquidity,” Review of Alignment in Judgment Revision: Implications for Brand Posi- Financial Studies, 17 (2), 439–61.
tioning,” Journal of Marketing Research, 39 (February), 18–30.
Henderson, Pamela W. and Joseph A. Cote (1998), “Guidelines for Pullig, Chris, Richard G. Netemeyer, and Abhijit Biswas (2006), Selecting and Modifying Logos,” Journal of Marketing, 62 “Attitude Bias, Confidence, and Challenge Alignment: A Case of Negative Brand Publicity,” Journal of the Academy of Mar- Hofmann, David A. and Mark B. Gavin (1998), “Centering Deci- keting Science, 34 (4), 528–42.
sions in Hierarchical Linear Models: Implications for Research Rao, Vithala R., Manoj K. Agarwal, and Denise Dahlhoff (2004), in Organizations,” Journal of Management, 24 (5), 623–41.
“How Is Manifest Branding Strategy Related to the Intangible Ismail, Badr and Kwan Choi (1996), “Determinants of Time Value of a Corporation?” Journal of Marketing, 68 (October), Series Properties of Earnings and Cash Flows,” Review of Financial Economics, 5 (2), 131–45.
Rappaport, Alfred (1986), Creating Shareholder Value. New York: Jacoby, Jacob (2001), “The Psychological Foundations of Trade- mark Law: Secondary Meaning, Genericism, Fame, Confusion Salinas, Gabriela and Tim Ambler (2008), “A Taxonomy of Brand and Dilution,” Trademark Reporter, 91 (4), 1013–1071.
Valuation Methodologies: How Different Types of Methodolo- Keller, Kevin L. (1993), “Conceptualizing, Measuring, and Man- gies Can Help to Answer Different Types of Questions,” Mar- aging Customer-Based Brand Equity,” Journal of Marketing, keting Science Institute Report No. 08-204.
Shankar, Venkatesh, Pablo Azar, and Matthew Fuller (2008), ——— and Donald R. Lehmann (2006), “Brands and Branding: “BRAN*EQT: A Multicategory Brand Equity Model and Its Research Findings and Future Priorities,” Marketing Science, Application at Allstate,” Marketing Science, 27 (July–August), Kerin, Roger A. and Raj Sethuraman (1998), “Exploring the Brand Shugan, Steven M. (2004), “Endogeneity in Marketing Decision Value–Shareholder Value Nexus for Consumer Goods Compa- Models,” Marketing Science, 23 (Winter), 1–3.
nies,” Journal of the Academy of Marketing Science, 26 (4), Slotegraaf, Rebecca J. and Koen Pauwels (2008), “The Impact of Brand Equity and Innovation on the Long-Term Effectiveness Krishnan, H.S. (1996), “Characteristics of Memory Associations: of Promotions,” Journal of Marketing Research, 45 (June), A Consumer-Based Brand Equity Perspective,” International Journal of Research in Marketing, 13 (4), 389–405.
Srinivasan, V., Chan S. Park, and Dae R. Chang (2005), “An Lindenberg, Eric B. and Stephen A. Ross (1981), “Tobin’s q Ratio Approach to the Measurement, Analysis, and Prediction of and Industrial Organization,” Journal of Business, 54 (Janu- Brand Equity and Its Sources,” Management Science, 51 (9), Luo, Xueming (2009), “Quantifying the Long-Term Impact of Srivastava, Rajendra K., Tasadduq A. Shervani, and Liam Fahey Negative Word of Mouth on Cash Flows and Stock Prices,” (1999), “Marketing, Business Processes, and Shareholder Marketing Science, 28 (1), 148–65.
Value: An Organizationally Embedded View of Marketing Madden, Thomas J., Frank Fehle, and Susan Fournier (2006), Activities and the Discipline of Marketing,” Journal of Market- “Brands Matter: An Empirical Demonstration of the Creation ing, 63 (Special Issue), 168–79.
of Shareholder Value Through Branding,” Journal of the Acad- Trademark Act of 1946 (1982), 15 U.S.C. §§ 1051–1127.
emy of Marketing Science, 34 (2), 224–35.
Wyer, Robert S., Jr., and Thomas K. Srull (1989), “Person Mem- Marketing Science Institute–Emory Marketing Institute (2007), ory and Judgment,” Psychological Review, 96 (1), 58–83.
“Marketing Strategy Meets Wall Street,” (accessed July 8, 166 / Journal of Marketing, November 2009

Source: http://ws36.manag-volhovsky.pu.ru/files/upload/niim/seminar/kmo_2009_financial_impact_of_trademarks.pdf