Principles of Marketing Research 3

Article

Brand Name Types and Consumer Demand: Evidence from China’s Automobile Market

Fang Wu, Qi Sun, Rajdeep Grewal, and Shanjun Li

Abstract Brand naming challenges are more complex in logographic languages (e.g., Chinese), compared with phonographic languages (e.g., English) because the former languages feature looser correspondence between sound and meaning. With these two dimensions of sound and meaning, the authors propose a four-way categorization of brand name types for logographic languages: alphanumeric, phonetic, phonosemantic, or semantic. Using automobile sales data from China and a discrete choice model for differentiated products, the authors relate brand name types to demand, with evidence showing that Chinese consumers preferred vehicle models with semantic brand names (7.64% more sales than alphanumeric) but exhibited the least preference for phonosemantic names (4.92% lower sales than alphanumeric). Domestic Chinese firms benefited from semantic brand names, whereas foreign firms gained from using foreign-sounding brand names. Entry-level products performed better with semantic brand names, and high-end products excelled when they had foreign-sounding brand names. Thus, the four-way categorization of brand name types should help multinational firms and domestic Chinese firms understand and leverage the association between brand name types and consumer demand.

Keywords automobile market, brand name types, Chinese consumers, consumer demand, linguistics

Online supplement: https://doi.org/10.1177/0022243718820571

Developing a brand name that resonates with customers is a

major marketing challenge (Keller and Lehmann 2006),

especially in an era when brands compete across different

countries, cultures, and value systems (Mittal and Tsiros

1995). Linguistic systems that rely on different phonetics,

semiotics, and semantics—defined as logographic lan-

guages—exacerbate the brand naming challenges. For

example, the complexity of Chinese as a logographic lan-

guage makes it difficult to choose a brand name (Fetscherin

et al. 2012) such that:

Localizing a brand name for China is a mind-boggling challenge.

Ideally, the name should convey the brand’s story, set out its local

positioning and be memorable. It should sound similar to the orig-

inal, and have a good ring to it. It shouldn’t evoke unintended

meanings in Mandarin or major dialects. (Doland 2017)

Some brands got it right; “Sai Bai Wei” (which means

“more tasty than hundreds of flavors”) signals Subway with

just three characters, and “Ke Kou Ke Le” (“tastes good and

makes you happy”) refers to Coca-Cola with four charac-

ters, both of which can be spoken smoothly in Chinese and

resonate with the meaning of the characters (Zhang and

Schmitt 2001, p. 313). The Chinese name for Airbnb, “Ai

Bi Ying,” contains three characters that translate to “love,”

“each other,” and “welcome,” yet it prompted considerable

backlash because the phrase is hard to pronounce, did not

make sense to consumers, and evoked pornographic associa-

tions (Doland 2017).

Theory anticipates such challenges for logographic lan-

guages such as Chinese, because the correspondence between

speech and writing is less isomorphic in these cases (Zhang and

Schmitt 2001), unlike phonographic languages such as English

that feature strong correspondence between speech and writing.

In addition, logographic languages contain many homonyms;

as Schmitt, Pan, and Tavassoli (1994) note with regard to the

Fang Wu is Associate Professor, College of Business, Shanghai University of

Finance and Economics (email: [email protected]). Qi Sun

(corresponding author) is Associate Professor, College of Business, Shanghai

University of Finance and Economics (email: [email protected]).

Rajdeep Grewal is the Townsend Family Distinguished Professor of

Marketing, Kenan-Flagler Business School, University of North Carolina at

Chapel Hill (email: [email protected]). Shanjun Li is Associate

Professor and Kenneth L. Robinson Chair of Agricultural Economics and

Public Policy, Charles H. Dyson School of Applied Economics and

Management, Cornell University, and a faculty research fellow, National

Bureau of Economic Research (email: [email protected]).

Journal of Marketing Research 2019, Vol. 56(1) 158-175

ª American Marketing Association 2018 Article reuse guidelines:

sagepub.com/journals-permissions DOI: 10.1177/0022243718820571

journals.sagepub.com/home/mrj

pronunciation of “gong” in Chinese, it corresponds to at least

ten distinct characters with widely divergent meanings, includ-

ing “work,” “bow,” “public,” “meritorious service,” “attack,”

“supply,” “palace,” “respectful,” and a personal surname. The

separation of meaning and pronunciation creates more nuance

and makes it difficult to choose brand names, especially for

multinational corporations that aim to standardize their brand

names across countries.

Following extant branding research that focuses on logo-

graphic languages (e.g., Schmitt, Pan, and Tavassoli 1994;

Zhang and Schmitt 2004), we consider the two key dimensions

of sound and meaning (Table 1) and thus specify four types of

brand names. First, alphanumeric names consist of simple

numbers or logographical alphabets, with neither semantic

meaning nor any particular sound in Chinese or foreign lan-

guages (e.g., CX-7 from Mazda, which is the same as its West-

ern name and contains letters and numbers). Second, phonetic

names combine Chinese characters to mimic a foreign-

sounding word but have no meaning in Chinese (e.g., You-

Lee-O by the Chinese automaker Chery, which “sounds”

English but has no meaning in Chinese). Third, semantic names

convey a meaning in Chinese but do not have any foreign sound

(e.g., Nissan adopts Chinese brand name “Lou Lan,” the name

of an ancient city in western China, for its Murano sport utility

vehicle [SUV]). Fourth, phonosemantic names retain both a

foreign sound and a meaning in Chinese (e.g., Volkswagen uses

the brand name “Tu-Rui” in Chinese for its Touareg SUV,

which sounds like its English name Touareg and means “path

with spirit” in Chinese). Multinational corporations must

understand these four brand name types before they can stan-

dardize or adapt their brand names appropriately across

countries.

We use this four-way classification of logographic lan-

guages to examine how the sales impact of each type of brand

name varies with the product’s country of origin (COO; foreign

vs. domestic) and market segment (high-end vs. entry-level),

using monthly new vehicle sales data during 2008–2012 in the

Chinese automobile market. With a discrete-choice model

based on market-level data (e.g., Berry 1994), we quantify the

influence of brand name types on consumer demand and find

that Chinese consumers, on average, preferred vehicle models

with semantic brand names (7.64% more sales than alphanu- meric brand names) but exhibited the lowest preferences for

phonosemantic ones (4.92% lower sales than alphanumeric brand names). The association of brand name type and prefer-

ence also varied with the COO of a brand and across market

segments (entry-level or high-end). For example, among

domestic brands, Chinese consumers preferred semantic brand

names, but for foreign brands, they preferred alphanumeric

brand names and disliked semantic brand names. Sematic

brand names were preferred for entry-level models, but alpha-

numeric and phonetic brands names benefited high-end car

models.

These findings extend extant laboratory research that has

provided useful insights but relatively low generalizability.

Typical lab studies examine a single dimension (e.g., mean-

ingfulness, foreign vs. nonforeign sounding; Keller, Heckler,

and Houston 1998; Leclerc, Schmitt, and Dubé 1994) and

derive results from survey-based intention measures. However,

intentions do not always translate to behaviors (Chandon, Mor-

witz, and Reinartz 2005), so we supplement prior work by

showing how brand names relate to actual sales. Furthermore,

we provide insights into the mechanisms that underlie this

effect by investigating the moderating roles of both COO and

Table 1. Framework of Brand Name Types for Logographic Languages.

Sound

No Yes

Meaning

No Brand Name Type: Alphanumeric Brand Name Type: Phonetic Definition: Brand names that contain referential or nonsense

mixtures of (phonographic) letters and/or digits. Definition: Brand names that correspond as much as possible to

the phonetic structure of foreign language but contain no specific meaning in Chinese.

Examples: 3M, 360 (a popular internet security software brand in China).

Examples: Motorola’s Chinese name is “Mou-Tuo-Luo-La”; Galanz (the largest microwave oven brand in China) uses “Ge-Lan-Shi,” which sounds like “Galanz” in English.

Yes Brand Name Type: Semantic Brand Name Type: Phonosemantic Definition: Brand names that have actual meanings in Chinese

but do not represent foreign sounds. Definition: Brand names with Chinese characters that have

both the phonetic structure of the foreign language and a concrete meaning in Chinese.

Examples: Microsoft’s Chinese name is “Wei-Ruan,” which means “tiny and soft” in Chinese; Lenovo’s Chinese name is “Lian-Xiang,” which means “association” in Chinese.

Examples: Coca-Cola’s Chinese name is “Ke-Kou-Ke-Le,” which sounds like “Coca-Cola” and means “tastes good and makes you happy”; E-Land (a popular apparel brand in China) uses “Yi-Lian,” which sounds like “E-Land” and means “loving clothes.”

Notes: We use the Roman transliteration known as the Pin-Yin system to present the Chinese brand names. We provide two examples of each brand name type, one for a multinational firm and the other for a domestic Chinese firm.

Wu et al. 159

market segments. Finally, we augment literature on the stan-

dardization versus adaptation of brand names in international

markets, which thus far has largely emphasized elements of the

marketing mix other than branding (Sandler and Shani 1992).

We conceptualize brand name types along a standardization–

adaptation continuum for logographic languages.

The findings in turn provide actionable insights for manag-

ers in China. Domestic Chinese firms should have stressed

semantic brand names to enjoy 10.28% more sales than they would have earned with the alphanumeric brand names, but

foreign firms should have adopted brand names that feature

foreign language elements (e.g., semantic brand names led to

5.37% lower sales than alphanumeric brand names). Firms should have adapted their brand naming strategies across mar-

ket segments, because consumers preferred semantic brand

names for entry-level products (3.85% more sales than alpha- numeric brand names) but disliked semantic brand names for

higher-end products (semantic brand names led to 8.90% lower sales than alphanumeric brand names).

In the next section, we present relevant literature on brand

name types. We then discuss the research context in more

detail, to provide a foundation for our empirical analysis. After

we illustrate our data and lay out the empirical model, together

with the identification and estimation strategies, we present our

estimation results. Finally, we conclude by elaborating on the

implications of our research.

Background and Literature Review

As we mentioned previously, brand name types vary along two

dimensions, sound and meaning, which creates the four-way

categorization of brand name types that we depict in Table 1.

With this classification, we aim to address the standardization–

adaptation challenge for multinational corporations that oper-

ate in countries with phonographic and logographic language

systems (Alashban et al. 2002). In our context, alphanumeric

names are the most standardized; they remain unchanged

across markets. Semantic names are the most adaptable and

reflect each local language. Phonetic and phonosemantic brand

names lie in the middle, though phonosemantic names are more

adaptable than phonetic names. On the one hand, prior research

has suggested that multinational corporations tend to adapt

their brand names in logographic languages (e.g., Francis, Lam,

and Walls 2002), especially when competition is intense

(Krishna and Ahluwalia 2008). On the other hand, the foreign

language elements embedded in alphanumeric, phonetic, and

phonosemantic brand names might trigger stronger foreignness

effects and lead to favorable consumer attitudes in China, given

that foreignness effects tend to be stronger in Asian countries

(Maheswaran, Chen, and He 2013). Our results show that both

competition and foreignness effects are at play, whereby Chi-

nese consumers prefer more standardized brand names (i.e.,

alphanumeric, phonetic, or phonosemantic) for vehicles from

foreign manufacturers and more adapted names (i.e., semantic)

for entry-level brands for which competition tends to be more

intense.

In Table 2, we summarize the literature on these four brand

name types. 1

Most of the studies are set in the U.S. context and

use lab experiments, with consumer as the unit of analysis and

survey-based dependent measures as opposed to actual con-

sumer choice. In contrast, our study is set in the Chinese con-

text, where we use observational field data with brand (e.g.,

Honda CR-V) as a unit of analysis and estimate a demand

model based on unit sales.

Early research focused on semantic brand names, before

moving on to phonetic and alphanumeric brand names. These

studies generally investigate one brand name type (semantic

[Peterson and Ross 1972]; phonetic [Leclerc, Schmitt, and

Dubé 1994]; alphanumeric [Gunasti and Ross 2010]) and refer

solely to a phonographic language. Subsequent extensions fea-

tured logographic languages (semantic [Lee and Kim Soon

Ang 2003]; phonetic [Melnyk, Klein, and Volckner 2012];

alphanumeric [Yan and Duclos 2013]). Studies on phonose-

mantic brand names only emerged with Zhang and Schmitt’s

(2001, 2004) studies in China, in which these authors used the

microfoundations of logographic languages, including individ-

ual characters and separate sound and meaning (e.g., Schmitt,

Pan, and Tavassoli 1994). The microfoundation approach also

supported the simultaneous consideration of semantic, pho-

netic, and phonosemantic brand names (Zhang and Schmitt

2001). We build on this laboratory-based research to examine

the four brand names type with observational field data.

Alphanumeric brand names contain referential or random

mixtures of (phonographic) letters and/or digits, such as 3M

and 7 UP (e.g., Boyd 1985). Prior research has shown that

consumers generally associate alphanumeric brand names with

technical, formulated, chemical, or powerful products (e.g.,

automobiles, bicycles, appliances; Pavia and Costa 1993). In

China, the inclusion of lucky numbers (e.g., 6, 8, 9) and aus-

picious letters (e.g., A, S) also enhance consumers’ evaluations

of brand quality (Ang 1997). Moreover, alphanumeric names

are the most standardized, as they remain unchanged across

countries, making them an important branding strategy for

multinational firms (Alashban et al. 2002); thus, we assess the

relative efficacy of alphanumeric brand names in a logographic

language context.

Research on semantic brand names generally examines the

relationship between the meaning of a brand name and consu-

mers’ brand perceptions and attitudes, mainly in phonographic

language contexts. Research has shown that brand names that

use familiar words or convey a product benefit (e.g., “Lifelong”

brand luggage) enhance brand outcomes such as attitudes, per-

ceptions, and recall (Keller, Heckler, and Houston 1998; Lee

and Swee Hoon Ang 2003; Peterson and Ross 1972).

1 To gain a historical perspective, we organize Table 1 in chronological order.

For tractability, we gathered studies from primary marketing journals: Journal

of Marketing Research, Journal of Marketing, Journal of Consumer Research,

Journal of Consumer Psychology, Journal of the Academy of Marketing

Science, International Journal for Research in Marketing, Marketing Letters,

and Psychology & Marketing (Marketing Science did not publish any relevant

article).

160 Journal of Marketing Research 56(1)

Table 2. Representative Research Examining Brand Name Types.

Article

Theoretical Emphasis Study Setting Sample Size(s)

Unit of Analysis Methodology

Dependent VariableAlphanumeric Phonetic Semantic Phonosemantic

Peterson and Ross (1972)

X X P X United States

Consumer Experiment Perceived measure

Robertson (1987)

X X P X United States

152*216 Consumer Experiment Perceived measure

Meyers-Levy (1989)

X X P X United States

79, 100 Consumer Experiments Perceived measure

Pavia and Costa (1993)

P X X X United States

48–64 (focus) 300 (survey)

Consumer Focus groups, survey

Perceived measure

Leclerc, Schmitt, and Dubé (1994)

X P X X United States

40, 184, 42 Consumer Experiments Perceived measure

Keller, Heckler, and Houston (1998)

X X P X United States

160 Consumer Experiments Perceived measure

McCracken and Macklin (1998)

X X P X United States

45, 143 Consumer Experiments Perceived measure

Sen (1999) X X P X United States

127, 125 Consumer Experiments Perceived measure

Zhang and Schmitt (2001)

X P P P China 183, 120, 240 Consumer Experiments Perceived measure

Lee and Swee Hoon Ang (2003)

X X P X Singapore 176 Consumer Experiments Perceived measure

Lee and Kim Soon Ang (2003)

X X P X Singapore 44, 88 Consumer Experiments Perceived measure

Zhang and Schmitt (2004)

X P P P China 368 Consumer Experiments Perceived measure

Miller and Kahn (2005)

X X P X United States

143 Consumer Experiments Perceived measure

Gunasti and Ross (2010)

P X X X United States

51, 60, 74 Consumer Experiments Perceived measure

Samu and Krishnan (2010)

X X P X United States

154, 245 Consumer Experiments Perceived measure

Melnyk, Klein, and Volckner (2012)

X P X X France, United States, Taiwan

105, 88, 181 Consumer Experiments Perceived measure

Yan and Duclos (2013)

P X X X Hong Kong 145, 96, 142, 145 Consumer Experiments Perceived measure

Gunasti and Devezer (2016)

P X X X United States

189, 176, 206 Consumer Experiments Perceived measure

Gunasti and Ozcan (2016)

P X X X United States

107*183 Consumer Experiments Perceived measure

Current study P P P P China 270 brands over five years

Brand Field data Sales

Notes: The phonetic brand name type indicates foreign-sounding brand names. We report the sample size for each study; multiple numbers signify multiple studies. All experimental studies are lab experiments.

Wu et al. 161

Furthermore, such effects are stronger for logographic lan-

guages (e.g., Chinese) than phonographic languages (e.g., Eng-

lish; Lee and Swee Hoon Ang 2003), because logographic

language characters are mapped more directly onto meanings

(Biederman and Tsao 1979). Thus, we aim to assess the

strength of semantic brand names on demand in a logographic

language context.

Research on phonetic brand names focuses on the effects of

phonetic symbolism, which exists when the mere sound of a

word, beyond its actual definition, conveys meaning (French

1977) and thus influences brand preferences (e.g., Coulter and

Coulter 2010; Lowrey and Shrum 2007; Yorkston and Menon

2004). From an international marketing perspective, foreign-

sounding brand names induce COO effects (Leclerc, Schmitt,

and Dubé 1994). Melnyk, Klein, and Volckner (2012) find that

incongruence between the actual COO of a brand and the COO

implied by its foreign-sounding name decreases consumers’

purchase likelihood, and the effect is stronger (weaker) for

emerging (developed) markets. Therefore, foreign-sounding

phonetic brand names may lead to higher sales for brands from

foreign firms, but not for domestic firms.

Phonosemantic brand names are specific to logographic lan-

guages, which create a possibility of combining semantic and

phonetic characteristics. This brand naming strategy is popular

with multinational corporations that seek a phonographic-to-

logographic brand name translation in Chinese markets (Zhang

and Schmitt 2001). Many managers accordingly believe that

phonosemantic brand names combine the advantages of pho-

netic and semantic names, because they feature both foreign

sounds and actual meanings in local languages. However, this

effect may depend on the context. As Zhang and Schmitt

(2001) show, consumers’ evaluations of phonosemantic brand

names (vs. phonetic or semantic names) depend on whether

they mentally code meaning or sound. The foreign language

proficiency of these consumers also moderates this effect;

populations with less proficiency express greater preferences

for semantic over phonosemantic brand names (Zhang and

Schmitt 2004). Therefore, we expect phonosemantic brand

names to lead to higher demand for brands targeted at consu-

mers proficient in foreign languages.

Research Context

China and Chinese Language

China offers a unique and important context for studying the

impact of brand name types on consumer demand. It is the

second-largest economy in the world and is still growing;

moreover, the Chinese language is a logographic system.

Unlike phonographic Western languages (e.g., English), its rich

word representations can result from sound (phonetic represen-

tation), meaning (semantic representation), or their combina-

tion (phonosemantic representation). Thus, the Chinese

language allows for varied brand naming options for both mul-

tinational and domestic firms. One-fifth of the world’s popu-

lation speaks some form of Chinese as their first language, and

Chinese characteristics and pronunciations are common in

other languages, such as Japanese, Korean, and Vietnamese

(Ramsey 1987). Marketing literature accordingly has devoted

increasing attention to the Chinese market and various strategic

issues, including consumers’ information processing (Pan and

Schmitt 1996), brand name translation and creation (Zhang

and Schmitt 2001, 2004), and market entry challenges (Johnson

and Tellis 2008). Our analysis contributes to this research by

providing managerial insights for marketers regarding brand

naming strategies in China, which should apply to the Asia-

Pacific region.

Brand Name Categorizations and the Chinese Automobile Market

China’s trademark law requires that all firms use brand names

in Chinese for all business activities, with the exception of

certain alphanumeric names. Therefore, brand names in China

always fall into either of two categories: those consisting of

(logographic) Chinese characters and those consisting of a mix

of (phonographic) letters and/or numbers, which are alphanu-

meric (Pavia and Costa 1993). The four-way classification in

Table 1 is consistent with Chinese trademark laws.

For this research, China’s automobile market in particu-

lar offers an appealing context. It has transformed from

being virtually nonexistent to the largest automobile market

in the world, in just two decades: It produced and sold more

than 23 million vehicles in 2014, compared with 16.5 mil-

lion in the United States (Li, Xiao, and Liu 2015). As is

common in emerging markets, approximately 60%–70% of vehicle purchasers are first-time buyers, without experience

to draw on, unlike buyers in mature markets. Therefore,

brand names have more room and opportunity to convey

information and influence consumer choices. Because of

China’s mandate that all firms use brand names in Chinese

for all business activities (cf. certain alphanumeric brand

names), choosing Chinese brand names for vehicle models

is important for both multinational and domestic firms. This

market also offers detailed, longitudinal data on new vehicle

sales across different market segments (e.g., foreign and

domestic brands, high-end, entry-level), so we can examine

the heterogeneous effects of brand names on consumer

demand along these dimensions.

Brand naming decisions in the automobile industry take

place at the vehicle model level (e.g., Chevrolet Cruze, Chev-

rolet Malibu), so by examining the impact of brand names on

consumer demand at this level, we control for confounding

effects from unobserved heterogeneity at the firm (e.g.,

General Motors) and parent-brand (e.g., Chevrolet) levels

(e.g., brand image embedded in parent brand names).

Finally, foreign and domestic automobile manufacturers in

China use all four types of brand names, such rich varia-

tions are crucial to our identification of the effects of brand

name types (see Table 3).

162 Journal of Marketing Research 56(1)

Data

We rely on market-level sales data obtained from R.L. Polk &

Company, a leading provider of information in automobile

markets. The data contain monthly unit sales and characteris-

tics of all new vehicle models sold in China at the vintage and

model level (e.g., 2005 vintage Toyota Camry) from January

2008 until December 2012.

We gathered the English brand names for each vehicle

model directly from the data set, then obtained the correspond-

ing Chinese brand names from the database of the Ministry of

Industry and Information Technology of China (MIITC). To

identify vehicle models with alphanumeric names, we verified

whether their Chinese names consist of only foreign alphabets

and/or numbers. For the other three brand name types, we

provided a list of all English and Chinese brand names for each

vehicle model to two professors from the English department in

a Chinese university (not associated with this research). They

independently identified the type of brand name (semantic,

phonetic, or phonosemantic) for each model, using definitions

provided by Zhang and Schmitt (2001). If they disagreed about

any name types (which occurred for 9 of the 209 nonalphanu-

meric names), a third professor from the same department pro-

vided an opinion, and we based the final decision on the simple

majority. 2

For several reasons, we code phonosemantic and

phonetic brand names according to the pronunciation of Chi-

nese brand names in Mandarin Chinese. First, our data concern

automobile sales in mainland China. The Law of the People’s

Republic of China on the Standard Spoken and Written Chinese

Language (Article 12) requires the use of Mandarin Chinese

(Putonghua) in mainland China in all broadcast programs and

advertising. Therefore, brand names identified on television or

radio are all in Mandarin Chinese, even in areas where con-

sumers speak other languages (e.g., Tibet, Inner Mongolia).

Second, all schools in China teach Mandarin Chinese, so nearly

all Chinese consumers can speak and understand it, even

though some also speak other languages.

We excluded premium brands (e.g., Audi, BMW), which

predominantly use alphanumeric names (cf. Land Rover). We

also excluded two nonpremium brands that use only one type

of brand name for all models, Skoda and Peugeot, which

account for less than 4% of the Chinese market during this sample period. After the exclusions, the data set contains 270

vehicle models and 10,607 observations of monthly sales at

the vintage–model level. In a Web Appendix, we summarize

the distribution of name types across brands, vehicle classes,

vehicle types (i.e., sedan, SUV, and multipurpose vehicle

[MPV]), and COO.

To illustrate the pattern of vehicle sales by brand name

types, we plot vehicle sales over several categories. In Figure

1, we plot average monthly vehicle sales by four name types;

sales of vehicle models with phonosemantic names slightly

trail those with the other three name types, especially during

the latter part of our sample period. In Figure 2, we plot the

average monthly vehicle sales by four name types for Chinese

and foreign vehicle models. Chinese vehicle models with

semantic names appear to have higher sales (Panel A), but

foreign models with semantic names achieve lower sales (Panel

B). In Figure 3, we plot average monthly vehicle sales by name

types for entry-level and high-end vehicle models (classified by

the number of engine cylinders, as we discuss subsequently).

Entry-level models with alphanumeric names tend to have

lower sales (Panel A), whereas high-end models with phonetic

names tend to have higher sales (Panel B). In Figure 4, we plot

the number of vehicle model introductions (128 of 270 models)

for the four brand name types in our sample period (2008–

2012) for Chinese and foreign brands. Chinese manufacturers

introduced more vehicle models with semantic brand names,

but foreign brands introduced more models with phonetic

brand names.

Our data contain rich descriptions of vehicle characteristics

for each vintage and model, including vehicle price, fuel con-

sumption, vehicle size, weight, power, vehicle type, and an

indicator for domestic brands. In Tables 4 and 5, we present

Table 3. Example Brand Name Types for Vehicle Models.

Vehicle Brand

Vehicle Model Name in English Vehicle Model Name in Chinese (Pinyin)

Brand Name Types

Foreign Brands

Honda CR-V CR-V Alphanumeric Toyota Camry Kai Mei Rui (sounds like “Camry” in English and has no specific meaning in

Chinese) Phonetic

Volkswagen Touran Tu An (sounds like “Touran” in English and means “safe trip” in Chinese) Phonosemantic Ford Explorer Tan Lu Zhe (Chinese translation for explorer) Semantic

Domestic Brands

BYD F6 F6 Alphanumeric Geely Uliou You Li Ou (sounds like “Uliou” in English and has no specific meaning

in Chinese) Phonetic

Chery Riich Rui Qi (sounds like “rich” in English and means “prosperous kirin,” a mythical chimerical creature, in Chinese).

Phonosemantic

Foton View Feng Jing (meaning “View” in Chinese) Semantic

Notes: Phonetic, phonosemantic, and semantic brand names are presented in Chinese Pinyin, as in Zhang and Schmitt (2001, 2004).

2 In an empirical analysis in which we excluded these nine models, the

substantive results did not change. In Web Appendix Table A1, we list the

brand name type for each of the 270 vehicle models in our data set.

Wu et al. 163

the descriptive statistics and correlations, respectively. To

compute vehicle prices, we use manufacturer suggested retail

prices (MSRP) and sales tax. Set by manufacturers, the MSRPs

generally are constant across locations and for each model year.

The sales tax is normally 10%, but temporary cuts reduced it to 5% and 7.5% for vehicles with displacements of less than 1.6 liters in 2009 and 2010, respectively (the end of this tax cut in

December 2010 leads to a sharp increase in vehicle sales

towards the end of 2010, Figure 1). 3

We base prices on MSRP

rather than transaction prices, so a potential bias could arise if

we were to ignore time-varying discounts and promotions.

However, in contrast with the promotion-heavy U.S. market,

the Chinese automobile market features infrequent manufac-

turer or dealer promotions, and retail prices generally are close

to MSRP (Li, Xiao, and Liu 2015).

Empirical Model

Brand Name Types and Consumer Demand for New Vehicles

We use a discrete choice model for differentiated products to

study the impacts of brand name types on consumer demand

(Berry 1994; Berry, Levinsohn, and Pakes 1995). These discrete

choice models are parsimonious, yet they account for many

competing products and can address endogeneity from unob-

served product characteristics (Dubé et al. 2002). As a result

of this parsimony and flexibility, marketing researchers have

applied these discrete choice models for differentiated products

to model demand for several product categories, including

automobiles (Sudhir 2001), ready-to-eat cereals (Goldfarb, Lu,

and Moorthy 2009), PC servers (Chu and Chintagunta 2009),

and sports drinks (Chen, John, and Narasimhan 2008).

In our empirical model, we assume that the indirect utility

function of consumer i, gained from purchasing vehicle model j

(i.e., vintage–model combination) at time t, is as follows:

uijt ¼ X4

k¼2 gk Namejk þ a lnðpjtÞþ xjtbþ parentb þ typeg

þ importd þ timet þ xjt þ eijt; ð1Þ

where Namejk indicates the brand name type of model j, such

that k ¼ 1 for alphanumeric names (the reference group), k ¼ 2 for phonetic names, k ¼ 3 for phonosemantic names, and k ¼ 4 for semantic names. In our utility specifications, the vehicle

model’s brand name is a vehicle characteristic, because consu-

mers derive utility directly from brand names (Keller 1993). In

turn, gk is the taste parameter for a brand name of type k, pjt is the price of vehicle j at time t, and xjt is other observed charac-

teristics of vehicle j at time t (e.g., engine power, fuel efficiency).

Brand equity drives consumer demand and may confound

brand name type effects (Chu and Chintagunta 2009; Goldfarb,

Lu, and Moorthy 2009; Keller 1993). In Equation 1, we include

parent-brand fixed effects ( parentb) to control for any time-

invariant factor that may affect demand; that is, parentb absorbs parent-level brand equity (e.g., Toyota). Then, we can

use the variations in brand name types at the model level,

within each parent brand, to identify brand name type effects.

To control for unobserved heterogeneity among different

vehicle models, we include other fixed effects in consumers’

utility functions. The vehicle type (i.e., sedan, SUV, or MPV)

fixed effect typeg captures time-invariant heterogeneity at the

vehicle type level, such as consumers’ taste preferences over

vehicle functionality. Another indicator importd distinguishes

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

A ve

ra ge

M on

th ly

S al

es

Time (Months)

Alphanumeric Phonetic Phonosemantic Semantic

Figure 1. Average monthly vehicle sales by brand name types. Notes: This figure depicts the average monthly vehicle sales at the vintage–model level by four brand name types (i.e., alphanumeric, phonetic, phonosemantic, and semantic brand names).

3 As we explain subsequently, our vehicle price variable includes sales taxes

and our model specification includes time fixed effects. Therefore, we account

for variation in consumer demand resulting from tax cuts.

164 Journal of Marketing Research 56(1)

imported from domestically built vehicle models (most foreign

vehicle brands in China sell both domestically built and

imported models; because there are many joint ventures [e.g.,

Beijing Hyundai, Shanghai Volkswagen], if a vehicle model is

from a joint venture, we classify it as a foreign brand). Further-

more, we use time (year–month) fixed effects timet to control

for time-variant factors that affect all vehicle models.

Some other vehicle characteristics (e.g., vehicle–model

level brand equity, interior or exterior styling, promotions,

advertising) are unobserved by us but may be observed by

consumers and firms. To address this concern, we include the

disturbance term xjt and acknowledge that such unobservable characteristics may correlate with the price variable and brand

name type indicators, which could lead to endogeneity. We

adopt an instrumental variable approach to control for such

endogeneity. Finally, the Eijt error terms are independently and identically distributed across consumer i, vehicle j, and time t.

The demand function in Equation 1 may follow a nesting

structure, with the outside good defined as not buying any

vehicle (our empirical setup is robust to alternative definitions

of the outside good, as long as it has no cross-sectional varia-

tion; Berry 1994). Each of the six vehicle classes (mini, small,

compact, medium, large, and premium) is a nest (Deng and Ma

2010; Klier and Linn 2012). With this nest structure, the error

term Eijt takes the functional form of Eijt ¼ oict þð1 � sÞZijt, where oict is the shock to all the vehicles in class c; s is the similarity coefficient, which represents the extent to which

consumers receive similar shocks to individual vehicles in a

class c; and Zijt is the idiosyncratic shock to consumer i, vehicle j, and time t. We provide the detailed derivation of the

A

B

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

A ve

ra ge

M on

th ly

S al

es

Time (Months)

Alphanumeric Phonetic Phonosemantic Semantic

0

2,000

4,000

6,000

8,000

10,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

A ve

ra ge

M on

th ly

S al

es

Time

Alphanumeric Phonetic Phonosemantic Semantic

Figure 2. Average monthly vehicle sales by brand name types: Chinese vs. foreign models. Notes: Panel A depicts the average monthly vehicle sales of Chinese vehicle models, and Panel B depicts the average monthly vehicle sales of foreign vehicle models, both at the vintage–model level by four brand name types (i.e., alphanumeric, phonetic, phonosemantic, and semantic brand names).

Wu et al. 165

empirical model in Web Appendix B. Assuming that Eijt fol- lows a Type I extreme value distribution, we can express the

market share of vehicle j at time t as (e.g., Berry 1994)

lnðsjtÞ� lnðs0tÞ¼ X4

k¼2 gk Namejk þ a lnðpjtÞþ xjtb

þs lnðsjtjctÞþ paren tb þ typeg þ importd þ timet þxjt:

ð2Þ

The left-hand side of Equation 2 is the difference between the

log market share of vehicle j and the log market share of the

outside good (which we define as not buying a vehicle). Here,

lnðsjtjctÞ is the log market share of vehicle j in class c at time t. We estimate Equation 2 in a linear regression framework, using

instrumental variables to address endogeneity (e.g., Berry 1994). 4

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

A ve

ra ge

M on

th ly

S al

es

Time (Months)

Alphanumeric Phonetic Phonosemantic Semantic

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59

A ve

ra ge

M on

th ly

S al

es

Time (Months)

Alphanumeric Phonetic Phonosemantic Semantic

A

B

Figure 3. Average monthly vehicle sales by brand name types: Entry-level vs. high-end models. Notes: Panel A depicts the average monthly vehicle sales of entry-level models, and Panel B depicts the average monthly vehicle sales of high-end vehicle models, both at the vintage–model level by four brand name types (i.e., alphanumeric, phonetic, phonosemantic, and semantic brand names), where an entry-level model is one with four or fewer engine cylinders, and a high-end model is one with six or more engine cylinders.

4 Our model features two compromises. First, it has the advantage of

relaxing the undesired property of the independence of the irrelevant

166 Journal of Marketing Research 56(1)

Heterogeneous Impacts of Brand Name Types on Vehicle Demand

To obtain deeper managerial insights, we investigate heteroge-

neity in the impacts of brand name types on vehicle demand

based on COO and targeted market segment. We introduce

these moderators in our model in a linear manner:

gk Namejk ¼ dkðNamejk � DOMjÞþ wkðNamejk � FORjÞ þykðNamejk � EntryjÞ þfkðNamejk � HighjÞ; 8 k ¼ 1; 2; 3; 4;

ð3Þ

where DOMj signals domestic models, FORj is an indicator of

foreign models, Entryj indicates entry-level models (four or

fewer cylinders), 5

and Highj signals high-end models (six or

more cylinders). Furthermore, as taste parameters, we include

A

B

0

1

2

3

4

5

6

7

8

9

2008 2009 2010 2011 2012

N um

be r

of V

eh ic

le M

od el

s

Year

Alphanumeric Phonetic Phonosemantic Semantic

0

1

2

3

4

5

6

7

8

9

2008 2009 2010 2011 2012

N um

be r

of V

eh ic

le M

od el

s

Year

Alphanumeric Phonetic Phonosemantic Semantic

Figure 4. Number of vehicle models with each brand name type introduced by year (2008 to 2012): Chinese vs. foreign models.

Table 4. Descriptive Statistics.

Variables Mean SD Min Max

Alphanumeric names .205 .403 0 1 Phonetic names .269 .443 0 1 Phonosemantic names .214 .410 0 1 Semantic names .312 .464 0 1 Monthly sales 3,506.195 4,883.020 1 75,307 Vehicle price (in 10,000 RMB,

2008 level) 14.246 13.357 2.080 173.350

Fuel consumption (Liter/100 km)

8.437 1.836 4.0 15.6

Vehicle size (m 2 ) 12.183 2.118 7.779 19.730

Vehicle power (kW) 97.344 35.483 30 246.359 Vehicle weight (ton) 1.815 .384 1.095 3.291 If sedan .670 .470 0 1 If SUV .169 .375 0 1 If MPV .161 .367 0 1 If imported .125 .331 0 1 Number of observations per

year 2008 1,301 2009 1,822 2010 2,307 2011 2,476 2012 2,701

Notes: The total number of observations is 10,607. Each observation is a vin- tage–model–year–month combination. Vehicle size is the product of vehicle length, width, and height.

alternatives for vehicle models across nests, but it maintains the

independence of the irrelevant alternatives assumption within a nest (i.e.,

vehicle class). Second, it allows for consumer heterogeneity at the nest

level (i.e., vehicle class) but not at the product level (i.e., vehicle

model). These compromises are practical: in a model with multiple

endogenous explanatory variables (e.g., vehicle price, log within-class

market share, brand name indicators; we elaborate on these

subsequently), it often is not possible to estimate the many coefficients

with consumer heterogeneity, because of the limited number of available

instruments (Klier and Linn 2012).

5 We use the number of engine cylinders to categorize vehicle models in our

data set into two segments: entry-level models, which have four or fewer

engine cylinders, and high-end models, which include six or more. Most

vehicle models in our data set have four or six cylinders. Of the 270 models,

six have three cylinders, 198 have four cylinders, 64 models have six cylinders,

and 2 models have eight or more cylinders. Most entry- and mid-level vehicle

models in China rely on four-cylinder engines. In our sample period, SUVs

(usually equipped with larger engines) were not popular, accounting for less

than 8% of sales. Most sales of nonpremium brands involve sedans, for which

six-cylinder models tend to be high-end models.

Wu et al. 167

dk for domestic models, wk for foreign models, yk for entry- level models, and fk for high-end models. Domestic models with alphanumeric brand names are the reference group when

we model COO, and entry-level models with alphanumeric

names are the base group for the targeted market segments.

Identification

Two notable identification issues arise with Equations 2 and 3.

First, the vehicle price variable lnðpjtÞ might correlate with the unobserved characteristics xjt, in which case the log within- class share lnðsjtjctÞ would correlate with xjt, because it is a function of pjt. For example, vehicles with better quality (unob-

served characteristic in our data and embedded in xjt) can induce higher willingness to pay, and manufacturers may

charge higher prices because of higher marginal costs or oli-

gopolistic market power. We follow Berry (1994) and Berry,

Levinsohn, and Pakes (1995) and use two sets of instruments

for both lnðpjtÞ and lnðsjtjctÞ. The first set includes sums of observed nonprice characteristics (cf. name types) of the other

vehicle models of the same type (e.g., sedan, SUV) sold by the

same firm. The second set includes nonprice characteristics of

other vehicle models of competing firms that belong to the

same vehicle type. The identification assumption is that these

nonprice characteristics are exogenous. If we assume that the

Chinese automobile market is an oligopoly with differentiated

products, optimal price-setting decisions by automakers imply

that the price of one product correlates with the attributes of

other products in the market, thus satisfying the instrument

relevance assumption. That is, these instruments would corre-

late with vehicle prices and within-class share but not with the

error term that comprises unobserved characteristics.

Second, the indicators for the brand name types Namejk and

interaction terms could correlate with the unobserved charac-

teristic xjt, because firms might use name types to convey

certain characteristics that the researchers cannot observe (Kel-

ler, Heckler, and Houston 1998). For example, Labbrand, a

leading brand naming consultancy in China, created the Chi-

nese brand name for the Spanish automobile brand SEAT to

emphasize its European tradition (see http://www.labbrand.

com.cn/work/seat, in Chinese [retrieved October 2018]) and

derived Chinese and English brand names for the Volkswagen

Phideon sedan to emphasize its luxury aspects (see http://www.

labbrand.com.cn/work/phideon, in Chinese [retrieved October

2018]). These variables (e.g., European tradition of SEAT

brands, luxury nature of Volkswagen Phideon sedan) are

omitted from our estimation equation (but summarized in the

unobserved characteristics xjt), though they correlate with the brand name type indicators Namejk. Because vehicle manufac-

turers usually make brand naming decisions for each vehicle

model only once when they introduce it to the market, these

omitted variables also should be time invariant.

The instruments for brand name type indicators for a given

vehicle model are based on the brand naming options of other

vehicle models. For vehicle model j, we record the year of its

introduction to the Chinese market. 6

The first set of instruments

for j’s brand name indicators includes the number of vehicle

models from the same parent brand that were introduced before

j, calculated separately as those with alphanumeric, phonetic,

phonosemantic, or semantic brand names. For example, the

Honda Accord was introduced in China in 1999, so the first

set of instruments for its brand name type are the number of

Honda vehicle models introduced before 1999 in the Chinese

Table 5. Correlation Matrix.

ABS PHO PHN SEM SAL PR FU SZ PW WE SE SU MP IM

Alphanumeric names (ABS)

1.00

Phonetic names (PHO) �.31** 1.00 Phonosemantic names

(PHN) �.26** �.32** 1.00

Semantic names (SEM) �.34** �.41** �.35** 1.00 Monthly sales (SAL) �.04** .03** .01 �.00 1.00 Vehicle price (PR) �.04** .13** .06** �.15** �.16** 1.00 Fuel consumption (FU) �.04** .09** �.00 �.05** �.12** .64** 1.00 Vehicle size (SZ) �.10** .14** .05** �.09** �.13** .62** .65** 1.00 Vehicle power (PW) �.02 .12** .06** �.15** �.15** .84** .71** .70** 1.00 Vehicle weight (WE) �.09** .17** .06** �.14** �.16** .75** .70** .92** .79** 1.00 If sedan (SE) .13** �.08** �.05** .01 .17** �.23** �.23** �.61** �.19** �.52** 1.00 If SUV (SU) �.08** .19** �.04** �.08** �.11** .35** .37** .57** .36** .52** �.64** 1.00 If MPV (MP) �.08** �.09** .10** .07** �.10** �.07** �.09** .20** �.13** .14** �.62** �.20** 1.00 If imported (IM) .08** .26** .11** �.28** .15** .49** .36** .33** .52** .44** �.45** .13** �.13** 1.00

*p < .10. **p < .05.

6 We record the introduction year at the vehicle model level (instead of

vintage–model level). By law, all manufacturers must file with MIITC when

they introduce new vehicle models in China. Therefore, the introduction year is

the earliest year each vehicle model’s information appears in the MIITC

database.

168 Journal of Marketing Research 56(1)

market with each of the four brand name types. The second set

of instruments pertains to the number of vehicle models from

other parent brands introduced before j, again according to the

four brand name types.

The rationale for leveraging these instruments is that firms

may have incentives to follow or deviate from current trends in

brand naming decisions when introducing new vehicle models.

For example, they might adopt their competitors’ practices if

certain brand name types are already popular in the market, to

reduce the risks associated with brand naming decisions. In

contrast, firms could have incentives to deviate and adopt less

popular brand names to stimulate consumer awareness. There-

fore, for each vehicle model j, the brand name types of existing

vehicle models may correlate with its own brand name type. In

addition, firms choose brand names according to their compet-

itive strategy, so the brand name types of vehicles from the

same and other parent brands may have different impacts on

each individual model’s choice. Therefore, we use two sets of

instruments for brand name types (i.e., brand naming decision

from own and other parent brands) to allow for these poten-

tially different impacts. After controlling for vehicle-invariant

time effects, the unobserved product attributes of a particular

vehicle model j at year t should not correlate with the brand

name types of other vehicles at the time of the introduction

model j, which is often many years before t.

Finally, as in Tables A2 and A3 in the Web Appendix,

vehicle models from different brands, in different vehicle

classes (e.g., mini, compact), of different vehicle types (i.e.,

sedan, SUV, and MPV), or from different COO (i.e., domestic

vs. foreign) tend to use certain brand name types. The set of

fixed effects in our specifications addresses the potential com-

mon confounding factors across these dimensions. For exam-

ple, the parent brand fixed effects (parentb) control for

unobserved heterogeneity at both the brand and COO levels;

the vehicle-type fixed effects ( typeg) control for unobserved

heterogeneity at the vehicle type level. Our vehicle classifica-

tion is size-based, so including size of the vehicle model as an

explanatory variable can capture confounding factors specific

to its class (e.g., mini, compact).

Empirical Results

Impacts of Brand Names on Consumer Demand for Vehicles

We present the estimation results for our baseline Equation 2 in

Table 6. Parameter estimates of the logit model are in Column

1, and those for the nested logit specification are in Column 2.

The estimation results for the two specifications are similar; on

average, Chinese consumers preferred vehicle models with

semantic names, but they exhibited the least preference for

models with phonosemantic names (p < .05). However, con- sumers were largely indifferent between models with alphanu-

meric and phonetic brand names. Meanwhile, the own-price

elasticity of vehicle demand implied by the results in Column

2 is �6.89, in line with the estimates in prior literature (e.g.,

Deng and Ma 2010; Xiao and Ju 2014). The coefficients for the

other vehicle characteristics also exhibit the expected signs:

consumers disliked fuel-inefficient vehicles and preferred

larger, more powerful models. Quantitatively, our estimation

results suggest that a vehicle model with a semantic brand

name would have sold approximately 7.64% more than a model with an alphanumeric brand name, whereas one with a phono-

semantic brand name would have sold approximately 4.92% less, ceteris paribus.

7 Our estimated demand effects of brand

name types are within a reasonable range, according to related

studies. For example, with the estimated own-price elasticity of

�6.89, the compensating demand effects of brand name types would be equivalent to approximately 1.11% (semantic vs. alphanumeric) and .71% (phonosemantic vs. alphanumeric) differences in vehicle prices. Sullivan (1998) finds that differ-

ent brand names result in an approximately 3.63% difference in vehicle prices, using twin automobiles in the U.S. used vehicle

market. It makes sense that our estimated price effects would

be smaller than Sullivan’s, because we focus on the brand name

element of individual models (i.e., brand name types at the

vehicle model level), whereas Sullivan considers the brand

effects of both parent brand names (e.g., Toyota) and individual

models (e.g., Corolla).

In Columns 3 and 4 of Table 6, we split our sample into two

groups with observations before and after 2010, respectively,

and estimate the model with these two samples separately. By

doing so, we can account for the possible shift in consumer

preferences for vehicles from different COO that might have

stemmed from the 2010 Diaoyu/Senkaku Islands dispute 8

between Japan and China and possible subsequent anti-

Japanese sentiment among Chinese consumers. We observe

that the estimated effect of phonosemantic brand names is less

negative than that in Column 3. This less negative effect might

be due to the increased consumer ethnocentrism after the 2010

Diaoyu/Senkaku Islands dispute, because the increased ethno-

centrism may drive consumer preference toward products with

more adapted (less foreign) brand elements, such as those with

phonosematic brand names (compared with alphanumeric or

phonetic brand names). Moreover, the implied price elasticity

in Column 4 is smaller (in magnitude) than that in Column 3.

As we have discussed, Chinese consumers rushed to purchase

vehicles in 2010 before the tax cuts ended, and such purchase

behaviors may lead to lower estimated price elasticity in our

2010 to 2012 sample.

In Column 5 of Table 6, we present results from another test

for preference invariance by adding the COO-specific time

7 To calculate the effects of brand name types on vehicle demand, we calculate,

for each vehicle model with an alphanumeric brand name, the changes in its

market share if it were to take a semantic brand name, according to our demand

estimates. Then, we determine the average of all changes for all vehicles with

alphanumeric brand names. 8

This territorial dispute involves a group of uninhabited islands known as the

Senkaku Islands in Japan and as the Diaoyu Islands in China. The Japanese

government’s purchase of Diaoyu Islands from a private owner led to

widespread anti-Japanese demonstrations in China beginning in 2010.

Wu et al. 169

T a b

le 6 .

E ff e ct

s o f B ra

n d

N am

e T

yp e s

o n

V e h ic

le D

e m

an d .

(1 )

L o gi

t (2

) N

e st

e d

L o gi

t (3

) N

e st

e d

L o gi

t 2 0 0 8 – 2 0 0 9

S am

p le

(4 )

N e st

e d

L o gi

t 2 0 1 0 – 2 0 1 2

S am

p le

(5 )

N e st

e d

L o gi

t w

it h

C O

O -S

p e ci

fi c

T im

e T

re n d

(6 )

N e st

e d

L o gi

t w

it h

V e h ic

le T

yp e

as A

lt e rn

at iv

e N

e st

s

(7 )

N e st

e d

L o gi

t w

it h

V e h ic

le P ri

ce R

an ge

as A

lt e rn

at iv

e N

e st

s

L o g

ve h ic

le p ri

ce �

1 .9

9 1 2 **

* (. 1 7 4 4 ) �

2 .9

5 9 8 **

* (. 8 8 6 9 ) �

4 .3

3 7 3 **

* (. 8 8 5 2 ) �

2 .3

9 3 1 **

* (. 7 2 3 6 ) �

2 .5

4 2 1 **

* (. 8 2 7 2 ) �

1 .4

8 3 0 **

* (. 2 6 0 4 ) �

1 .5

2 1 7 **

* (. 1 4 9 6 )

L o g

w it h in

-c la

ss sh

ar e

.7 4 6 7 **

* (. 0 2 5 0 )

.6 2 4 0 **

* (. 0 3 9 9 )

.7 9 5 1 **

* (. 0 2 1 1 )

.7 4 6 9 **

* (. 0 2 2 2 )

.8 8 6 2 **

* (. 0 0 9 8 )

.9 2 3 0 **

* (. 0 0 5 9 )

P h o n e ti c

.0 6 3 5

(. 0 6 7 0 )

� .0

0 4 9

(. 0 6 3 4 )

� .1

0 5 0

(. 1 3 2 8 )

.2 0 7 0

(. 4 9 9 6 )

� .0

8 0 2

(. 0 6 0 2 )

.0 1 3 1

(. 0 1 5 2 )

.0 0 6 1

(. 0 1 0 0 )

P h o n o se

m an

ti c

� .1

5 3 3 **

(. 0 6 7 3 ) �

.1 8 7 6 **

* (. 0 6 6 3 ) �

.3 0 4 5 **

* (. 0 9 4 4 )

� .0

9 5 8 *

(. 0 5 1 3 ) �

.1 6 8 0 **

* (. 0 6 4 0 )

� .0

3 1 5 *

(. 0 1 8 8 )

� .0

3 2 4 **

(. 0 1 2 9 )

S e m

an ti c

1 .0

5 6 6 **

* (. 0 7 7 0 )

.2 4 7 2 **

* (. 0 4 9 8 )

.2 3 7 6 **

* (. 0 8 4 1 )

.3 2 4 9 **

* (. 0 3 5 9 )

.2 1 2 6 **

* (. 0 5 1 3 )

.0 4 8 6 **

* (. 0 1 4 4 )

.0 5 9 5 **

* (. 0 0 9 8 )

F u e l co

n su

m p ti o n

(l it e r/

1 0 0

k m

) �

.0 4 7 7 **

(. 0 2 1 8 )

� .0

4 0 1 *

(. 0 2 1 8 )

� .0

5 9 2 **

(. 0 2 4 3 ) �

.0 4 4 3 **

* (. 0 1 9 0 )

� .0

4 4 2 **

(. 0 1 9 8 ) �

.0 3 7 4 **

* (. 0 0 5 7 )

� .0

2 3 2 **

* (. 0 0 3 7 )

V e h ic

le si

ze (m

2 )

.0 9 3 5 **

* (. 0 3 1 2 )

.1 7 2 1 **

* (. 0 3 7 7 )

.0 4 0 1

(. 0 5 1 3 )

.1 4 7 3 **

* (. 0 3 5 0 )

.1 4 2 8 **

* (. 0 3 4 5 )

.0 4 1 3 **

* (. 0 0 9 6 )

.0 1 6 9 **

* (. 0 0 5 5 )

P o w

e r

(k W

) .0

1 4 2 **

* (. 0 2 2 6 )

.0 1 7 9 **

(. 0 0 8 4 )

.0 2 6 7 **

* (. 0 0 8 1 )

.0 1 2 5 *

(. 0 0 7 1 )

.0 1 3 4 *

(. 0 0 7 9 )

.0 1 4 1 **

* (. 0 0 2 5 )

.0 1 2 2 **

* (. 0 0 1 4 )

W e ig

h t

(t o n )

.3 0 1 8

(. 2 0 9 3 )

.0 8 9 1

(. 3 8 9 7 )

.4 9 5 0

(. 4 5 8 9 )

.5 7 5 2

(. 4 1 8 1 )

.2 0 2 3

(. 3 7 7 8 )

.6 8 9 2

(4 0 5 9 )

.2 4 6 8 **

* (. 0 6 6 0 )

If im

p o rt

e d

3 .7

6 4 8 **

* (. 0 8 9 1 )

2 .6

7 8 7 **

* (. 6 5 5 8 )

3 .7

3 2 8 **

* (. 8 3 2 8 )

3 .2

2 2 2 **

* (. 4 8 2 1 )

2 .9

5 7 3 **

* (. 7 5 9 0 )

1 .0

8 6 9 **

* (. 1 7 5 5 )

1 .5

6 9 5 **

* (. 1 0 2 3 )

C o n st

an t

� 8 .5

2 0 4 **

* (. 2 6 7 8 ) �

1 2 .8

5 7 6 **

* (. 3 2 1 1 ) �

1 2 .6

4 7 0 **

* (. 5 1 0 9 ) �

1 2 .2

7 4 5 **

* (. 2 9 5 8 ) �

1 1 .5

3 3 0 **

* (. 5 4 8 9 ) �

1 1 .8

3 2 5 **

* (. 1 0 6 5 ) �

1 1 .7

7 1 7 **

* (. 0 6 7 1 )

P ar

e n t

b ra

n d

fi x e d

e ff e ct

s Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

V e h ic

le ty

p e

fi x e d

e ff e ct

s Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

T im

e (y

e ar

-m o n th

) fi x e d

e ff e ct

s Y

e s

Y e s

Y e s

Y e s

Y e s

Y e s

Y e s

F ir

st -s

ta ge

F -s

ta ti st

ic , p h o n e ti c

6 4 9 .5

3 4 8 5 .5

6 1 2 1 1 .9

5 6 5 2 .5

1 5 3 1 .2

2 5 3 7 .9

4 6 2 9 .4

6 F ir

st -s

ta ge

F -s

ta ti st

ic ,

p h o n o se

m an

ti c

2 6 1 .5

4 1 9 2 .2

2 5 0 6 .6

0 1 2 6 .7

4 2 6 6 .6

5 2 3 7 .6

1 3 0 1 .5

8

F ir

st -s

ta ge

F -s

ta ti st

ic , se

m an

ti c

2 ,1

7 2 .7

1 1 ,6

2 0 .1

4 2 ,1

3 6 .9

3 3 ,5

4 0 .9

1 1 ,3

9 5 .2

1 1 ,9

4 7 .5

1 2 ,4

6 1 .9

0 O

ve ri

d e n ti fi ca

ti o n

te st

1 6 .7

4 (p ¼

.1 6 )

1 4 .6

2 (p ¼

.2 1 )

1 3 .5

1 (p ¼

.2 6 )

1 5 .4

2 (p ¼

.1 6 )

8 .0

6 (p ¼

.2 4 )

1 3 .5

4 (p ¼

.2 6 )

1 5 .3

6 (p ¼

.1 7 )

R 2

.6 9 3 9

.8 4 3 8

.8 0 6 6

.8 8 9 2

.8 3 2 5

.8 6 9 4

.9 0 3 5

N u m

b e r

o f o b se

rv at

io n s

1 0 ,6

0 7

1 0 ,6

0 7

3 ,1

2 3

7 ,4

8 4

1 0 ,6

0 7

1 0 ,6

0 7

1 0 ,6

0 7

*p <

.1 0 .

** p <

.0 5 .

** *p <

.0 1 .

N o te

s: R

o b u st

st an

d ar

d e rr

o rs

ar e

in p ar

e n th

e se

s. D

e p e n d e n t

va ri

ab le

is ln ðs

jt Þ�

ln ðs

0 tÞ

.

170

trend into our model to allow consumers’ preferences for vehi-

cles from distinct COO to vary differently over time. We thus

make our model more flexible, such that it can account for

possible shocks in consumers’ preferences for vehicles from

not only Japan but also other countries. In Columns 6 and 7 of

Table 6, we test the robustness of our empirical results against

alternative definitions of product nests. In column 6, we define

a nest as a vehicle type (sedan, SUV, and MPV). In Column 7,

we categorize vehicle models in our data set into four quartiles

according to their prices, and we define each quartile as a nest.

The results (Columns 3–7, Table 6) are qualitatively the same

as our baseline results (Column 2, Table 6), in that Chinese

consumers most (least) preferred vehicles with semantic (pho-

nosemantic) brand names. Finally, in all the specifications with

instrumental variables, the first-stage F-statistics suggest that

our instruments for brand name types are not weak (p < .01), and we fail to reject overidentification. Thus, from an empirical

perspective, our instruments seem to be valid.

As we have noted, semantic names are the most meaningful

and easy for consumers to recognize and remember (e.g., Kel-

ler, Heckler, and Houston 1998), whereas phonetic names carry

foreign language elements and benefit from the foreignness

effect, such that they provide a sense of modernity or social

status (e.g., Leclerc, Schmitt, and Dubé 1994). Alphanumeric

names tend to be simple and easy to recognize, and premium

vehicles predominantly use them; consumers might prefer their

prestigious image. However, phonetic and alphanumeric names

require greater foreign language proficiencies to process than

do semantic brand names. Therefore, the effects of each type of

brand name on consumer demand may depend on consumers’

foreign language (English in our context) proficiency. Consu-

mers with weaker proficiency may prefer semantic brand

names, but those with greater proficiency could prefer nonse-

mantic ones (Kum, Lee, and Qiu 2011). On average, Chinese

consumers have low English proficiency, so intuitively, they

might exhibit more preference for semantic than other name

types, 9

which could translate into more sales.

Finally, Chinese consumers might exhibit weaker prefer-

ence for vehicle models with phonosemantic names compared

with other brand name types. Although Chinese bilinguals with

high proficiency in both English and Chinese may prefer pho-

nosemantic names to other name types (because they can pro-

cess the words both phonologically and semantically; Kum,

Lee, and Qiu 2011; Zhang and Schmitt 2004), the foreignness

effect in phonetic and alphanumeric names might dominate any

gains from bilingual processing. This reasoning implies that the

demand effect could vary from consumers’ brand name type

preferences.

Heterogeneity in the Impact of Brand Name Types

Foreign versus domestic brands. In Table 7, we present the esti- mation results when we allow the impacts of brand name types

to vary between domestic and foreign brands. The results

demonstrate that when brands’ COOs come into play, the

impact of brand name type on sales changes. For domestic

brands, consumers most preferred vehicle models with seman-

tic brand names, but they were indifferent to the other three

brand name types. For foreign brands, Chinese consumers pre-

ferred alphanumeric names (p < .05), followed by phonetic and phonosemantic names, and they least preferred models with

semantic names (p < .05). A potential explanation for these findings (Zhang and

Schmitt 2004) is that domestic vehicles in China tend to be

Table 7. Heterogeneous Effects of Brand Name Types on Vehicle Demand.

Domestic Versus Foreign Models Entry-Level Versus High-End Models

Phonetic � Domestic .5968 (.6124) Phonetic � Entry-level �.3904*** (.1433) Phonosemantic � Domestic .1936 (.2855) Phonosemantic � Entry-level �.2887* (.1476) Semantic � Domestic 1.4527*** (.3563) Semantic � Entry-level .2401** (.1175) Alphanumeric � Foreign 2.7206*** (.5614) Alphanumeric � High-end �.6136 (.5883) Phonetic � Foreign 2.4275*** (.4926) Phonetic � High-end �.3488 (.5240) Phonosemantic � Foreign 2.4367*** (.5671) Phonosemantic � High-end �1.7910** (.6096) Semantic � Foreign 1.9143*** (.5564) Semantic � High-end �1.4972*** (.4700) Parent brand fixed effects Yes Yes Vehicle type fixed effects Yes Yes Indicator for imported vehicles Yes Yes Time fixed effects Yes Yes R

2 .8650 .8408

*p < .10. **p < .05. ***p < .01. Notes: The total number of observations is 10,607. Entry-level models are those with four or fewer engine cylinders, and high-end models are those with six or more engine cylinders. All specifications include the log of vehicle price, log of within-segment market share, fuel consumption, vehicle size, power, and weight, and a constant as explanatory variables. The results are from the two-stage least squares estimation in which we instrument for vehicle prices, within-class market share, and brand-name types. Robust standard errors are in parentheses.

9 Education First’s English Proficiency Index, a ranking of English-language

abilities worldwide, ranked China 36th among 44 countries in which English is

not the native language. It ranks second-lowest in Asia (Yue 2012).

Wu et al. 171

of lower quality and less expensive than foreign brands (Deng

and Ma 2010), so consumers purchasing domestic brands also

may have lower levels of income, education, and English pro-

ficiency and thus rely more on meaning to process brand

names. In turn, these consumers preferred semantic brand

names. Consumers who purchased foreign brand vehicles

instead may have higher levels of English proficiency and rely

more on sound to process brand names. Therefore, they gener-

ally preferred nonsemantic brand names and showed the least

preference for semantic brand names.

Our results highlight the importance of the foreignness

effect (Maheswaran, Chen, and He 2013) of brand names. A

preference for foreign-made products may stem from consu-

mers’ ability to achieve higher levels of self-enhancement by

purchasing foreign brands (Chen, Brockner, and Katz 1998).

Therefore, foreignness may have a stronger effect on consu-

mers purchasing foreign brands, and these consumers prefer

nonsemantic brand names.

Our findings also are consistent with Krishna and Ahluwa-

lia’s (2008) assertion: domestic consumers may be surprised

when multinationals use localized (e.g., semantic) brand names

or domestic firms use foreign-sounding (e.g., alphanumeric,

phonetic, or phonosemantic) brand names, because they

develop expectations about brand name choices on the basis

of the firms’ COO. Unexpected brand naming strategies may

spark consumers’ skepticism and negatively affect their brand

perceptions. As Melnyk, Klein, and Volckner (2012) show,

incongruence between the actual COO of a product and the

implied COO of a foreign-sounding name decreases consu-

mers’ purchase likelihood. Our results support this argument

by demonstrating that such negative effects on consumers’

brand perceptions also translate into brands’ sales performance.

Entry-level versus high-end models. For entry-level vehicle mod- els, consumers preferred models with semantic brand names (p

< .05) over those with phonetic or phonosemantic brand names (p < .05). For high-end models, they least preferred models with semantic or phonosemantic (p < .05) and most preferred models with alphanumeric or phonetic (p < .05) brand names (see Table 7).

These results again are consistent with our preceding dis-

cussions. Consumers who purchased entry-level vehicles may

be less educated and proficient in English; consumers who

purchased high-end models may be more educated and profi-

cient. Therefore, they preferred semantic brand names for

entry-level models but alphanumeric or phonetic brand names

for high-end models.

Krishna and Ahluwalia (2008) indicate that both multina-

tionals and domestic firms adapt their brand names more

toward local languages as competition increases. Our results

support this point: Chinese consumers preferred semantic

brand names for entry-level vehicles, for which competition

is more intense than for high-end vehicles. Such preferences

for semantic brand names should encourage foreign and

domestic manufacturers to adopt semantic brand names in

China for their entry-level vehicles.

Discussion

Brand naming challenges with logographic languages arise

because of a weak correspondence between sound and mean-

ing. To address this correspondence, we propose a four-way

classification for brand name types in logographic languages

(Table 1). The classification is beneficial for multinational cor-

porations to decide the degree of standardization/adaptation in

brand naming decisions, whereby alphanumeric names are

most standardized and semantic names are most adapted.

We rely on market-level sales data from China, a country

with a logographic language, to quantify the effect of brand

name types on new vehicle demand. We examine heterogeneity

in the effect of brand name types across domestic and foreign

brands and across market segments (entry-level vs. high-end).

Consumers in China preferred vehicle models with semantic

brand names; they least preferred those with phonosemantic

names. However, the impact of brand names on vehicle sales

was heterogeneous: Consumers preferred semantic brand

names for domestic or entry-level vehicles and alphanumeric

or phonetic (foreign-sounding) names for foreign or high-end

vehicles.

Theoretical Implications

In quantifying the effect of brand name type on consumer

demand using field sales data, we build on existing research

that relies on lab experiments and considers brand name effects

on consumer perceptions or attitudes. We thus provide more

generalizable effect-size estimates for brand name type. That

is, lab studies likely produce greater effect sizes because they

use attitudes and intentions instead of actual behavior (Chan-

don, Morwitz, and Reinartz 2005). For example, Gunasti and

Ross (2010) suggest that as the number in an alphanumeric

brand name increases (e.g., Canon DC-800MX from Canon-

DC700MX), choice share could increase by 30%. For phonetic brands, Leclerc, Schmitt, and Dubé (1994) find effect sizes of

28.6%–52.9% for consumer attitudes. In China, Zhang and Schmitt (2001, 2004) document 20%–40% changes in con- sumer attitudes across semantic, phonetic, and phonosemantic

brand name types. 10

All these effect sizes are substantially

greater than the 4%–7% effect we find. Despite their high internal validity, it is important to acknowledge that laboratory

studies overestimate effect sizes.

Our four-way categorization of brand names fits well with

the general theoretical framework of brand naming research,

such that it supports a comparison of brand name effects across

different market segments in distinct contexts (e.g., high-end

vs. low-end products, foreign vs. domestic manufacturers).

10 Standardized effect size measures (e.g., r, Cohen’s d, odds ratio) may be

appropriate for consumer attitudes and intentions, but lab studies rarely report

these standardized measures or provide sufficient information to calculate

them. Therefore, we use the difference in the group means as a rough

measure of effect size. Gunasti and Ross (2010) use consumer choice of

hypothetical products as a dependent measure, so their effect size

(approximately 30%) is more directly comparable to our finding.

172 Journal of Marketing Research 56(1)

Semantic brand names seem to work for entry-level segments;

alphanumeric and phonetic brands seem preferable for the

high-end market. Semantic names are also preferable for

domestic brands, but alphanumeric ones are better for foreign

brands. These findings suggest that consumer heterogeneity

(e.g., education level, such that more education likely enables

consumers to earn more and seek high-end foreign brands)

drives brand name preferences.

Finally, our research provides several generalizable find-

ings. For example, meaningful and suggestive brand names,

on average, have favorable attitudinal effects, in China as in

Western language settings (Lee and Ang 2003). Alphanumeric

brand names are also fitting for high-end products (Ang 1997).

Language choices for brand names differ across foreign and

domestic brands, as noted in other Asian countries as well

(Krishna and Ahluwalia 2008).

Managerial Implications

Our empirical results provide managerial insights to inform

brand naming decisions in China. In addition to shaping con-

sumers’ perceptions and attitudes toward a product, brand

names significantly affect product sales. This finding implies

that manufacturers, both domestic and multinational, need to

exercise caution in choosing appropriate name types. In partic-

ular, when targeting average consumers in China, semantic

brand names can increase product sales, but one should use

phonosemantic brand names with caution in China.

Although our data are from 2008 to 2012 and do not catch

the latest trends in the Chinese automobile market, we believe

that our results are still relevant in guiding the brand naming

strategies and in spurring new research. For example, after

years of double-digit growth, China’s automobile market is

witnessing two significant trends: (1) changing competitive

dynamics and (2) changing consumer behaviors and attitudes

toward cars. Although domestic automakers dominate the low-

end segment, growth in this segment is increasingly attracting

foreign brands (Jullens 2014). A 2015 survey shows that Chi-

nese car buyers are becoming more practical and less status-

conscious (Gao et al. 2016). As our results suggest, consumers

at the lower end of the Chinese automobile market preferred

semantic brand names, and the foreignness effect was weaker

among these functionally oriented consumers. Thus, we expect

wider adoption of semantic names by foreign automobile

brands in China that target low-end segments.

High-end consumers preferred foreign-sounding names, so

alphanumeric and phonetic names may be more suitable for

ultraluxury automobile segments (e.g., Ferrari, Porsche, Bent-

ley). This result is consistent with the observation that many

classic brands in the ultraluxury segment use alphanumeric or

phonetic brand names in China (e.g., Ferrari 488; Porsche Cay-

enne, which uses a phonetic translation in Chinese). In contrast,

Bentley is thriving on its semantic brand names. For example,

Bentley uses semantic Chinese name “Tian Yue,” which means

“beyond excellence,” for its Bentayga SUV. These divergent

effects for ultraluxury automobiles might stem from the rich

traditions and history of the parent brands, which could trans-

cend their COO and make the brand name types (488 for Ferrari

vs. Tian Yue for Bentley) less relevant.

Consumers clearly identify with COO when choosing auto-

motive brands, so the choice of brand name type depends on the

brand’s COO (e.g., foreign firms with foreign-sounding name

type, domestic firms with semantic name type). However, in

markets for consumer electronics, apparel, or furniture, COO

effect may be less apparent, so firms might use brand names

strategically to disguise their true COO and manipulate con-

sumer perceptions (Zhang 2015). In markets in which the

manipulation of COO is feasible, a tendency to adopt

foreign-sounding brand names might increase, especially in

developing countries where consumers are more susceptible

to influences of foreignness effects embedded in foreign-

sounding brand names.

Further Research

Our study represents an initial step toward a clearer understand-

ing of the effects of brand naming in the marketplace. There are

several directions of further research. First, it would be worth-

while to examine other markets such as high-tech (e.g., elec-

tronics) and fashion (e.g., cosmetics, luxuries) industries, in

which foreign brands offer advantages for signaling quality

or social status, relative to industries that produce necessities.

Culture is an important moderating factor that may lead to

varying brand name effects across countries, so future research

could explore the demand effects of brand names in other

countries as well. Second, future research could pay more

attention to the mechanisms underlying the impacts that we

identify in this analysis. Linguists suggest other rules for devel-

oping brand names (e.g., phonetic devices such as alliteration

and assonance, orthographic devices such as abbreviations or

acronyms, morphological devices such as affixation and com-

pounding, semantic devices such as metaphor; Nilsen 1979;

Van den Bergh, Adler, and Oliver 1987), which provide alter-

native ways to categorize brand names (Chan and Huang

1997). Finally, we study brand name types at the product level

(e.g., Camry), after controlling for brand equity and brand

name effects embedded in the parent brand names (e.g.,

Toyota). Future research could develop appropriate identifica-

tion strategies to study brand name effects at the parent brand

level to gain additional insights.

Acknowledgments

The authors thank the JMR review team for their constructive

feedback. The authors also thank Junhong Chu, Yan (Lucy) Liu,

Shrihari (Hari) Sridhar, Jan-Benedict Steenkamp, Juanjuan Zhang,

and conference participants at the 2015 Marketing Science Confer-

ence for their helpful comments and suggestions.

Author Contributions

All authors contributed equally.

Wu et al. 173

Editorial Team

Joel Huber served as guest editor, and Gerard Tellis served as associ-

ate editor.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to

the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for

the research, authorship, and/or publication of this article: Fang Wu,

Qi Sun, and Shanjun Li acknowledge the financial support from the

National Natural Science Foundation of China (Grants 71402088,

71303148, and 71628303).

References

Alashban, Aref A., Linda A. Hayes, George M. Zinkhan, and Anne L.

Balazs (2002), “International Brand-Name Standardization/Adap-

tation: Antecedents and Consequences,” Journal of International

Marketing, 10 (3), 22–48.

Ang, Swee Hoon (1997), “Chinese Consumers’ Perception of Alpha-

numeric Brand Names,” Journal of Consumer Marketing, 14 (3),

220–33.

Berry, Steven (1994), “Estimating Discrete-Choice Models of Product

Differentiation,” RAND Journal of Economics, 25 (2), 242–62.

Berry, Steven, James Levinsohn, and Ariel Pakes (1995),

“Automobile Prices in Market Equilibrium,” Econometrica, 63

(4), 841–90.

Biederman, Irving, and Yao-Chung Tsao. (1979), “On Processing

Chinese Ideographs and English Words: Some Implications from

Stroop-Test Results,” Cognitive Psychology, 11 (2), 125–32.

Boyd, Colin W. (1985), “Point of View: Alphanumeric Brand

Names,” Journal of Advertising Research, 25 (5), 48–52.

Chan, Allan K.K., and Yue Yuan Huang (1997), “Brand Naming in

China: A Linguistic Approach,” Marketing Intelligence & Plan-

ning, 15 (5), 227–34.

Chandon, Pierre, Vicki G. Morwitz, and Werner J. Reinartz (2005),

“Do Intentions Really Predict Behavior? Self-Generated Validity

Effects in Survey Research,” Journal of Marketing, 69 (2), 1–14.

Chen, Xinlei (Jack), George John, and Narasimhan Om (2008),

“Assessing the Consequences of a Channel Switch,” Marketing

Science, 27 (3), 398–416.

Chen, Ya-Ru, Joel Brockner, and Tal Katz (1998), “Toward an Expla-

nation of Cultural Differences in In-Group Favoritism: The Role of

Individual Versus Collective Primacy,” Journal of Personality and

Social Psychology, 75 (6), 1490–1502.

Chu, Junhong, and Pradeep K. Chintagunta (2009), “Quantifying the

Economic Value of Warranties in the U.S. Server Market,” Mar-

keting Science, 28 (1), 99–121.

Coulter, Keith S., and Robin A. Coulter (2010), “Small Sounds, Big

Deals: Phonetic Symbolism Effects in Pricing,” Journal of Con-

sumer Research, 37 (2), 224–37.

Deng, Haiyan, and Alyson C. Ma (2010), “Market Structure and Pric-

ing Strategy of China’s Automobile Industry,” Journal of Indus-

trial Economics, 58 (4), 818–45.

Doland, Angela (2017), “As Airbnb Just Learned, It’s Really Hard to

Localize Your Brand Name for China,” AdAge (March 23), http://

adage.com/article/cmo-strategy/hard-localize-brand-china-airbnb-

learned/308405/.

Dubé, Jean-Pierre, Pradeep Chintagunta, Bart Bronnenberg, Ron

Goettler, Amil Petrin, and P.B. Seetheraman, et al. (2002),

“Structural Applications of the Discrete Choice Model,” Market-

ing Letters, 13 (3), 207–20.

Fetscherin, Marc, Illan Alon, Romie Littrell, and Allan Chan (2012),

“In China? Pick Your Brand Name Carefully,” Harvard Business

Review (September), https://hbr.org/2012/09/in–china-pick–your-

brand–name-carefully.

Francis, June N.P., Janet P.Y. Lam, and Jan Walls (2002), “The

Impact of Linguistic Differences on International Brand Name

Standardization: A Comparison of English and Chinese Brand

Names of Fortune-500 Companies,” Journal of International Mar-

keting, 10 (1), 98–116.

French, Patrice L. (1977), “Toward an Explanation of Phonetic

Symbolism,” Word, 28 (3), 305–22.

Gao, Paul, Sha Sha, Daniel Zipser, and Wouter Baan (2016), “Finding

the Fast Lane: Emerging Trends in China’s Auto Market,” survey,

McKinsey & Company, https://www.mckinsey.com/industries/

automotive-and-assembly/our-insights/finding-the-fast-lane-emer

ging-trends-in-chinas-auto-market

Goldfarb, Avi, Qiang Lu, and Sridhar Moorthy (2009), “Measuring

Brand Value in an Equilibrium Framework,” Marketing Science,

28 (1), 69–96.

Gunasti, Kunter, and Berna Devezer (2016), “How Competitor Brand

Names Affect Within-Brand Choices,” Marketing Letters, 27 (4),

715–27.

Gunasti, Kunter, and Timucin Ozcan (2016), “Consumer Reactions to

Round Numbers in Brand Names,” Marketing Letters, 27 (2),

309–22.

Gunasti, Kunter, and W. Ross (2010), “How and When Alphanumeric

Brand Names Affect Consumer Preferences,” Journal of Market-

ing Research, 47 (6), 1177–92.

Johnson, Joseph, and Gerard J. Tellis (2008), “Drivers of Success for

Market Entry into China and India,” Journal of Marketing, 72 (3),

1–13.

Jullins, John (2014), “It’s a Race to the Bottom in China’s Auto

Market,” Strategy þ Business (September 11), https://www.strat egy-business.com/blog/Its-a-Race-to-the-Bottom-in-Chinas-Auto-

Market?gko¼2b27c. Keller, Kevin Lane (1993), “Conceptualizing, Measuring, and Man-

aging Customer-Based Brand Equity,” Journal of Marketing, 57

(1), 1–22.

Keller, Kevin Lane, Susan E. Heckler, and Michael J. Houston (1998),

“The Effects of Brand Name Suggestiveness on Advertising

Recall,” Journal of Marketing, 62 (1), 48–57.

Keller, Kevin Lane, and Donald R. Lehmann (2006), “Brands and

Branding: Research Findings and Future Priorities,” Marketing

Science, 25 (6), 740–59.

Klier, Thomas, and Joshua Linn (2012), “New-Vehicle Characteristics

and the Cost of the Corporate Average Fuel Economy Standard,”

RAND Journal of Economics, 43 (1), 186–213.

174 Journal of Marketing Research 56(1)

Krishna, Aradhna, and Rohini Ahluwalia (2008), “Language Choice

in Advertising to Bilinguals: Asymmetric Effects for Multina-

tionals Versus Local Firms,” Journal of Consumer Research, 35

(4), 692–705.

Kum, Doreen, Yih Hwai Lee, and Cheng Qiu (2011), “Testing to

Prevent Bad Translation: Brand Name Conversions in Chinese-

English Context,” Journal of Business Research, 64 (6),

594–600.

Leclerc, France, Bernd H. Schmitt, and Laurette Dubé (1994),

“Foreign Branding and Its Effects on Product Perceptions and

Attitudes,” Journal of Marketing Research, 31 (2), 263–70.

Lee, Yih Hwai, and Kim Soon Ang (2003), “Brand Name Suggestive-

ness: A Chinese Language Perspective,” International Journal of

Research in Marketing, 20 (4), 323–35.

Lee, Yih Hwai, and Swee Hoon Ang (2003), “Interference of Picture

and Brand Name in a Multiple Linkage Ad Context,” Marketing

Letters, 14 (4), 273–88.

Li, Shanjun, Junji Xiao, and Yimin Liu (2015), “The Price Evolution

in China’s Automobile Market,” Journal of Economics and Man-

agement Strategy, 24 (4), 786–810.

Lowrey, Tina M., and L.J. Shrum (2007), “Phonetic Symbolism and

Brand Name Preference,” Journal of Consumer Research, 34 (3),

406–14.

Maheswaran, Durairaj, Cathy Yi Chen, and Junhong He (2013),

“Nation Equity: Integrating the Multiple Dimensions of Country

of Origin Effects,” in Review of Marketing Research, Vol. 10,

Naresh K. Malhotra, ed. Bingley, UK: Emerald Group Publishing,

153–89.

McCracken, J. Colleen, and M. Carole Mcklin (1998), “The Role of

Brand Names and Visual Cues in Enhancing Memory for Con-

sumer Packaged Goods,” Marketing Letters, 9 (2), 209–26.

Melnyk, Valentyna, Kristina Klein, and Franziska Volckner (2012),

“The Double-Edged Sword of Foreign Brand Names for Compa-

nies from Emerging Countries,” Journal of Marketing, 76 (6),

21–37.

Meyers-Levy, Joan (1989), “The Influence of a Brand Name’s Asso-

ciation Set Size and Word Frequency on Brand Memory,” Journal

of Consumer Researh, 16 (2), 197–207.

Miller, Elizabeth G., and Barbara E. Kahn (2005), “Shades of Mean-

ing: The Effect of Color and Flavor Names on Consumer Choice,”

Journal of Consumer Research, 32 (1), 86–92.

Mittal, Vikas, and Michael Tsiros (1995), “Does Country of Origin

Transfer Between Brands?” in Advances in Consumer Research,

Vol. 22, Frank R. Kardes, and Mita Sujan, eds. Provo, UT: Asso-

ciation for Consumer Research, 292–96.

Nilsen, Don L.F. (1979), “Language Play in Advertising: Linguisitic

Invention in Product Naming,” in Language in Public Life, James

Alatis and Richard Tucker, eds. Washington, DC: Georgetown

University Press.

Pan, Yigang, and Bernd Schmitt (1996), “Language and Brand Atti-

tudes: Impact of Script and Sound Matching in Chinese and Eng-

lish,” Journal of Consumer Psychology, 5 (3), 263–77.

Pavia, Teresa M., and Janeen Arnold Costa (1993), “The Winning

Number: Consumer Perceptions of Alphanumeric Brand Names,”

Journal of Marketing, 57 (3), 85–98.

Peterson, Robert A., and Ivan Ross (1972), “How to Name New

Brands,” Journal of Advertising Research, 12 (6), 29–34.

Ramsey, S. Robert (1987), The Language of China. Princeton, NJ:

Princeton University Press.

Robertson, Kim (1987), “Recall and Recognition Effects of Brand

Name Imagery,” Psychology and Marketing, 4 (1), 3–15.

Samu, Sridhar, and H. Shanker Krishnan (2010), “Brand Related

Information as Context: the Impact of Brand Name Characteristics

on Memory and Choice,” Journal of Academy of Marketing Sci-

ence, 38 (4), 456–57.

Sandler, Dennis M., and David Shani (1992), “Brand Globally but

Advertise Locally? An Empirical Investigation,” International

Marketing Review, 9 (4), 18–31.

Schmitt, Bernd H., Yigang Pan, and Nader T. Tavassoli (1994),

“Language and Consumer Memory: The Impact of Linguistic Dif-

ferences Between Chinese and English,” Journal of Consumer

Research, 21 (3), 419–31.

Sen, Sankar (1999), “The Effects of Brand Name Suggestiveness and

Decision Goal on the Development of Brand Knowledge,” Journal

of Consumer Psychology, 8 (4), 431–55.

Sullivan, Mary W. (1998), “How Brand Names Affect the Demand for

Twin Automobiles,” Journal of Marketing Research, 35 (2).

154–65.

Sudhir, K. (2001), “Competitive Pricing Behavior in the Auto Market:

A Structural Analysis,” Marketing Science, 20 (1), 42–60.

Van den Bergh, Bruce, Keith Adler, and Lauren Oliver (1987),

“Linguistic Distinction Among Top Brand Names,” Journal of

Advertising Research, 27 (4), 39–44.

Xiao, Junji, and Heng Ju (2014), “Market Equilibrium and the Envi-

ronmental Effects of Tax Adjustments in China’s Automobile

Industry,” Review of Economics and Statistics, 96 (2), 306–17.

Yan, Dengfeng, and Rod Duclos (2013), “Making Sense of Numbers:

Effects of Alphanumeric Brands on Consumer Inference,” Inter-

national Journal of Research in Marketing, 30 (2), 179–84.

Yorkston, Eric, and Geeta Menon (2004), “A Sound Idea: Phonetic

Effects of Brand Names on Consumer Judgments,” Journal of

Consumer Research, 31 (1), 43–51.

Yue, Zhang (2012), “China’s English Ability Lagging Behind,” China

Daily (November 3), http://english.peopledaily.com.cn/90882/

8003546.html.

Zhang, Kaifu (2015), “Breaking Free of a Stereotype: Should a

Domestic Brand Pretend to Be a Foreign One?” Marketing Sci-

ence, 34 (4), 539–54.

Zhang, Shi, and Bernd H. Schmitt (2001), “Creating Local Brands in

Multilingual International Markets,” Journal of Marketing

Research, 38 (3), 313–25.

Zhang, Shi, and Bernd H. Schmitt (2004), “Activating Sound and

Meaning: The Role of Language Proficiency in Bilingual Con-

sumer Environments,” Journal of Consumer Research, 31 (1),

220–28.

Wu et al. 175

Copyright of Journal of Marketing Research (JMR) is the property of American Marketing Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.