Journal of Retailing and Consumer Services 23 (2015) 7076

Contents lists available at ScienceDirect

Journal of Retailing and Consumer Services

journal homepage: www.elsevier.com/locate/jretconser

Consumer choice of store brands across store formats: A panel data

analysis under crisis periods
Joseph Kaswengi a,n, Mbaye Fall Diallo b,1
a
b

University of Orlans, VALLOREM Lab, Orlans, France

University of Lille 2 (Institute of Retail Marketing and ManagementIMMD), The LSMRC Lab (Universit de Lille-Skema Business School), Roubaix, France

art ic l e i nf o

a b s t r a c t

Available online 13 January 2015

Purpose: This paper investigates the effect of marketing variables and consumer personal characteristics
on store brand choice over national brands in times of crisis in France. It also seeks to clarify how store
formats affect consumption strategies towards brands in turbulent times. Based on a large sample (panel
of 4500 households, N 79,789), we used a binary logit model to assess consumer choice of store brands
over national brands across two different store formats (hypermarket and supermarket). Results show
that, overall, marketing variables and consumer characteristics affect signicantly store brand choice
over national brands. However, while crisis intensity clearly moderates the relationships between
marketing policy variables and store brand choice, it does not affect so extensively the relationships
between consumer characteristics and store brand choice over national brands. Furthermore, hypermarket and supermarket formats are not affected similarly by crisis. This research highlights the diversity of
consumer strategies developed to cope with economic crises. Theoretical and managerial implications of
these ndings are discussed.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
Store brand
National brand
Marketing mix
Socio-demographics
Crisis
Store format

1. Introduction
The strategic role of store brands (SBs) has been emphasised in
prior European studies (Lamey et al., 2007; Martos-Partal and
Gonzlez-Benito, 2011) and especially in France (Ataman et al.,
2007; Diallo et al., 2013). According to Nielsen Scan Track, store
brand market shares increased in value from 22% in 2002 to 30% in
2012 in France. Some researchers attribute this growth to the
economic crisis in the European context (Lamey et al., 2007). Other
researchers showed that during this economic shock, consumers
switched to store brands and decreased their consumption expenditures (Kaytaz and Gul, 2014). However, although prior studies
emphasized the importance of store formats for consumers in
different contexts (Martnez-Ruiz et al., 2010; Seock, 2009; Tripathi
and Dave, 2013; Zielke, 2010), it is not clear how consumers make
their choice of brands in different store formats during crisis periods.
We use the term format to distinguish between supermarket and
hypermarket retailers. Nevertheless, tough previous studies gave no
indication about brand choice in different store formats in crisis, we
can assume a migration phenomenon between formats to cope with
the economic situation. Over the past years, smaller formats such as

Corresponding author.
E-mail addresses: joseph.kaswengi@univ-orleans.fr (J. Kaswengi),
mbayefall.diallo@univ-lille2.fr (M.F. Diallo).
1
Address: IMMD, 6 rue de lhtel de ville, 59050 Roubaix, France.
http://dx.doi.org/10.1016/j.jretconser.2014.12.003
0969-6989/& 2014 Elsevier Ltd. All rights reserved.

supermarkets have gained greater popularity in Europe as they

allow time saving compared to bigger formats (i.e., hypermarket).
However, loss in purchasing power is likely to bring back some
consumers to hypermarkets where there are more deal opportunities and wider store brand ranges. In such a changing environment, the following questions appear to play a relevant role from a
retailer and a manufacturer perspective:
1. What key marketing and consumer factors inuence store
brand choice over national brands during economic slowdown
and economic recession?
2. Do these effects differ across store formats?

A better understanding of these issues can help to identify the

appropriate marketing strategies for retail chains, thereby optimising
retailer resources in turbulent times. It may also help national brand
managers who face increased store brand competition when crisis
hits. Previous research on store brand drivers focused on marketing
mix variables (Ngobo, 2011a), store image (Diallo et al., 2013), retailer
factors (Dhar and Hoch, 1997), consumer characteristics (Richardson
et al., 1996), perceived risk (Liljander et al., 2009). These studies do
not point out the moderating role of economic crisis, nor consumer
store format adjustments to cope with difcult economic situation.
Therefore, the aim of this research is to investigate the impact of
marketing variables (e.g., price, quantity, and non-price promotion)
and consumer characteristics (e.g., gender, age, and income) on store

J. Kaswengi, M.F. Diallo / Journal of Retailing and Consumer Services 23 (2015) 7076

brand choice over national brands in times of crisis while taking into
account different store formats. Thus, our research intends to understand consumer and retailer adjustments in times of crisis. By doing
so, we complement previous studies on the relationship between
business cycle and store brand purchase (Hoch and Banerji, 1993;
Lamey et al., 2012).
This paper adds to the existing literature in the following ways.
First, whereas prior research yielded important insights about store
performance, to the best to our knowledge, the effects of the
economic crisis on the relationships between marketing variables,
consumer personal factors and store brand choice relative to
national brands have not yet been examined. Our research contributes to understand how retailers should adapt specic marketing
mix variables and how they should manage changes in sociodemographic variables when crisis hits. In this respect, it intends
to complete more general recent studies analysing how marketing
variables affect retail strategies in times of crisis (Hampson and
McGoldrick, 2013; Kaytaz and Gul, 2014; Ferguson, 2014). Second,
although prior research addressed the issue of store format in
relation to consumer behaviour (e.g., Kumar and Karande, 2000;
Zielke, 2010), it did not investigate store format adjustments in
times of crisis. However, retailers need to understand quickly how to
adapt their store format management strategies when economic
crisis leads consumers to make changes in store frequentation.
Therefore, we intend to clarify the role of store format in times of
crisis by analysing two retail formats, in an attempt to extend and
bolster recent previous studies on the subject (e.g., Kaswengi, 2013;
Kumar and Roy, 2013).
The rest of the paper is structured as follows: rst, we examine
previous studies on store brands and present our research hypotheses. Second, we present the research methodology. Third, the
results are analysed. Finally, the paper concludes with a discussion
of the implications of the ndings and suggests avenues for further
research.

2. Theoretical framework
2.1. Research background
The central tenet of this research is that marketing elements,
socio-demographics and times of crisis play a role in store brand
choice over national brands. There are numerous studies that model
the utility function of the consumer as a function of marketing
variables (e.g. feature, display, ) and nd that these actions affect
utility and thus, brand choice (Balachander and Ghose, 2003). An
example is Erdem et al. (2004) who examined consumer choice
behaviour with regard to store brands in the United States, the United
Kingdom, and Spain. They found that consumer preferences for
quality and price explain consumers store brand choices. Previous
research also suggested that store brand choice can be linked to
demographic proles (Dhar and Hoch, 1997). For example, Ngobo
(2011b) showed that the rate of organic store brands increases
according to the presence of a working female. However, these
studies do not consider some key aspects of store brand consumption
such as the business cycle, which affects consumer purchasing power.
Therefore, by considering the variable crisis times, it is possible to
further understand relevant marketing actions that inuence customer behaviour. Hampson and McGoldrick (2013) review many
studies showing that consumers shopping attitudes and behaviours
are sensitive to recession. The behaviours may reect consumer
consciousness, especially in terms of price (Lichtenstein et al., 1993).
Sinha and Batra (1999) showed that store brands are an excellent
alternative for price conscious consumers. In line with Kaytaz and
Gul (2014), the economic crisis is summed up to play a role in the

71

relationship between marketing and socio-demographics elements

and store brand choice.
2.2. Hypotheses development
We developed research hypotheses based on a comprehensive
review of previous studies on store brands and national brands
(e.g., Sinha and Batra, 1999; Erdem et al., 2004; Ngobo, 2011a), on
the business cycle in relation to marketing variables (e.g., Lamey
et al., 2007; Hampson and McGoldrick, 2013) and on the relationships between store format and brand purchase behaviour (e.g.,
Kaswengi, 2013; Kumar and Roy, 2013).
2.2.1. Effect of product price and quality
Price and quantity constitute important factors that consumers
take into account in the shopping context. Price is a major positioning tool to differentiate a product (Yoo et al., 2000). Heavy store
brands users have lower incomes and are in general more price
sensitive (Dhar and Hoch, 1997). Krishnamurthi and Raj (1988)
suggest that a consumer with certain personal characteristics faced
with a set of brands with known attributes chooses not only a brand,
but also the quantity to purchase. Hampson and McGoldrick (2013)
review several studies showing that consumers shopping attitudes
and behaviours are sensitive to recession. Specically, they found
that during the shock, consumers decreased consumption expenditures and switched to cheaper goods. These behaviours may reect
consumer price consciousness (Lichtenstein et al., 1993). In a
recessionary period, store brands are an excellent alternative for
price conscious consumers (Sinha and Batra, 1999).
Prior research also showed that shares of store brands increase
more during recessions (Lamey et al., 2007; Tellis and Tellis, 2009).
The periods of economic contractions give incentives for consumers to experiment with lower-priced brands (e.g., store brands).
The higher-prices of national brands give the opportunity for the
price-sensitive consumers to prefer store brands. This experience
can lead consumers to close the perceived quality gap between the
brands. For example, Jones (2014) found that many consumers
who traded down to store brands during the 20072009 recessions found them to be quality products and maximised their
utility by purchasing larger quantities of store brands. Therefore,
we expect that cost savings will motivate consumers to purchase
larger quantities of store brands compared to national brands in
times of crisis. We can formulate the following hypotheses:
Hypothesis 1. Product price has a positive effect on store brand
choice over national brands in times of crisis.
Hypothesis 2. The number of products purchased has a positive
effect on store brand choice over national brands in times of crisis.
2.2.2. Effect of national brand feature and display
Academic research showed that as consumers are exposed to
promotion (e.g. feature advertising) more frequently, they develop
brand awareness, associations and a more positive perception of
brands (Balachander and Ghose, 2003; Zhang, 2006). By increasing
promotions intensity, manufacturers reinforce consumers brandrelated beliefs and attitudes, which lead to brand choice (Yoo et al.,
2000). Mela et al. (1998) found that national brands can use feature
and display advertising to isolate themselves from the competition. A
similar research showed that feature and display separate evaluations
of high-tier brands from low-tier brands, and therefore, increase the
choice of a national brand more than that of a store brand (Lemon
and Nowlis, 2002). Moreover, feature and display advertising play a
greater role in periods of contraction than in expansions. In fact, most
rms adjust their behaviour in response to recessions by increasing
features and displays (Tellis and Tellis, 2009). Differentiating national
brands from store brands should reduce price sensitivity and curtail

72

J. Kaswengi, M.F. Diallo / Journal of Retailing and Consumer Services 23 (2015) 7076

switching from national brands to lower-priced private labels (Lamey

et al., 2012). It is therefore expected that national brand display and
feature advertising will negatively affect store brand choice over
national when crisis hits. We hypothesise:
Hypothesis 3. National brand feature advertising has a negative
effect on store brand choice over national brands in times of crisis.
Hypothesis 4. National brand display has a negative effect on
store brand choice over national brands in times of crisis.
2.2.3. Effect of socio-demographic variables
Previous research suggested that store brand choice can be linked
to demographic proles (Dhar and Hoch, 1997; Burton et al., 1998).
For example, Ngobo and Jean (2012) showed that the rate of organic
store brands increases according to the presence of a working female.
However, they found a negative effect of income and occupation on
store brand product choice. Basing on previous studies, we posit that
socio-demographic factors will differ in their inuence on the store
brand choice in times of crisis. Engemann and Wall (2009) showed
that the last recession (20072009) had different impacts across
demographic groups in United States. For example, teen employment
fell by 23.8% during the recession, whereas employment of those
aged 55 and older rose by 7.4%. Hoynes (1999) examined the effect of
business cycles on the employment, earnings, and income of persons
in different demographic groups. The results of the 1982 recession
show that there are sizeable and statistically signicant differences in
the responses to cycles across skill groups. Jones (2014) found market
shares of store brand coffee to be much larger for lower-income
shoppers than for higher-income ones. Following this line of thought,
we expect to nd mixed effects of demographics on store brand
choice during the crisis. Hence, we hypothesise:
Hypothesis 5. Socio-demographic variables (age, profession,
income and family size) are related to consumer choice of store
brands over national brands in times of crisis.
2.2.4. Moderation of crisis intensity
According to Van Heerde et al. (2013), marketing researchers have
paid increasing attention in recent years to the link between marketing phenomena and macroeconomic uctuations. For instance, Lamey
et al. (2007) documented the effect of the business cycle on the
evolution of private-label share in several Western countries whereas
Millet et al. (2012) showed that consumers motivational orientations
differ across economic expansion and contraction periods. In line with
Kaytaz and Gul (2014), the economic crisis is hypothesized to play an
important role in the relationship between marketing elements and
store brand choice. Lamey et al. (2012) found that store brand share is
function of marketing policy variables (e.g. advertising and promotional pressure) and business cycle uctuations. Therefore, we expect
that the interaction between crisis intensity and marketing variables
leads to signicant effects on store brand choice over national brands.
Previous literature also showed that times of crisis led to consumption changes based on consumer demographics (Engemann and
Wall, 2009). As an example, income is likely a factor that will affect
more strongly store brand choice during recessions. Particularly, as
household income decreases, consumers switch from national brands
to store brands. Hampson and McGoldrick (2013) found signicant
differences across age, gender, employment and income in consumer
shopping adaptation in a recession period. Along these lines, crisis
intensity and socio-demographic characteristics can interact and lead
to signicant effects on consumer choice of store brands over national
brands. Specially, we expect to nd that a combination of a sociodemographic characteristics and crisis intensity will affect consumer
choice of store brands over national brands. Based on the discussion
above, the following hypotheses are proposed:

Hypothesis 6. Crisis intensity moderates the relationships

between marketing variables (price, quantity, brand display and
brand feature) and consumer choice of store brands over national
brands.
Hypothesis 7. Crisis intensity moderates the relationships
between socio-demographic factors (age, profession, income and
family size) and consumer choice of store brands over national
brands.
2.2.5. Brand choice in different store formats in crisis periods
Empirical studies have demonstrated signicant relations
between demographics, marketing variables and store format
(Zielke, 2010; Hsieh and Stiegert, 2012; Kumar and Roy, 2013). For
example, Kaswengi (2013) found that the impact of the stores
marketing elements (e.g. access, price) on brand equity is different
depending on store formats. Specically, the inuence of store access
on brand equity is statistically more signicant in supermarkets than
in hypermarkets. Moreover, Seock (2009) examined the inuence of
Hispanic consumers perceived importance of apparel retail store
environmental cues and demographic characteristics. The results
showed that customer service is a signicant determinant in Hispanic consumers decision to shop at department stores, specialty
stores, and mass merchant stores. This study led to another interesting result, namely that the respondents shopping frequency at
department stores, Internet websites, and catalogues was signicantly different based on the respondents age and number of years
lived in the US. In periods of crisis, consumers are likely to shop more
often in hypermarkets than in supermarkets. In fact, hypermarket
chains offer more product variety and lower prices in general than
supermarkets (Martnez-Ruiz et al., 2010). Recently, Tripathi and
Dave (2013) have indicated differences between store formats in
apparel retail. Based on the preceding discussion, we expect that the
impact of marketing variables and demographics on consumers
store brands choice in periods of crisis differ between store formats.
Hypothesis 8. The relationships between marketing variables
(price, quantity, national brand display and feature) and consumer
choice of store brands over national brands differ depending on
store format in times of crisis.
Hypothesis 9. The relationships between socio-demographic factors (age, profession, income and family size) and consumer choice
of store brands over national brands differ depending on store
format in times of crisis.

3. Research methodology
3.1. Model specication
The deterministic component of latent utility of household h
for format f on trip t is given by:
U t h f h; f X t h; f h t h; f ;

where h, f is the xed component of utility, X h, f h represents

the variable component of utility which is dependent on the
composition of shopping basket for the specic trip and th,f is
the error term.
We use the binary logit model to study the impact of marketing
variables and consumer characteristics on consumer choice of
store brands. The binary logit model has been extensively applied
in marketing literature (e.g. Ngobo, 2011b). It has this general
form:



T
T
Pr yi 1 exp x= 1 exp x
2
t

We consider that household h (h1,, H) buying in format

f (f 1,, F), in store s (s1,, S), in category g and in week

J. Kaswengi, M.F. Diallo / Journal of Retailing and Consumer Services 23 (2015) 7076

t (t 1,, T) chooses between two types of brand products denoted k,

where k1 for store brand products and 0 for national brand
products. A product choice by household h for format c, in store s
and in week t (hkcst) is described by a binary choice model:
8
>
< 1 if 4 0
;
hbcst
>
:
0 Otherwise

3.2. Data description and sample

We use purchase records from MarketingScans Behavior Scan
panels (a GFK & Mediametrie company) in Angers (France). Angers
(152,337 inhabitants) is the 17th largest French city. Its panels
represent stratied samples of about 4500 households in France. The
housing bubble in the United States (US) burst in August 2007; in the
United Kingdom (UK), Northern Rock failed in September 2007.
However, the European economies and the France in particular were
not affected right away by these developments. Ofcial gures from
INSEE (French national statistical department) showed that France was
slightly hit by crisis in 2008 (0.2% of annual growth in volume) and
deeply touched by it in 2009 ( 2.9% of annual growth in volume).
Ination decreases from 2.8% in 2008 to 0.1% in 2009. Consistent with
the research purposes, we selected two time periods: a slowdown
period, but no crisis ofcially (January 1June 30, 2008) and an ofcial
recession period (January 1June 30, 2009) in France based on ofcial
gures from INSEE. Each period of time covers 26 weeks. We therefore
compare a period of low crisis intensity to a period of high crisis
intensity in France. Overall, we analysed N79,789 observations
across two retail formats (hypermarket and supermarket) and across
four product categories (toothpaste, shower gel, dry pasta and
shampoo). These consumer packaged goods were chosen because of
their more necessary nature, even in times of crisis. Table 1 provides
detailed description of the panel in terms of household purchases in
each period investigated.

Table 1
Socio-demographic description of the panel.
Freq. (%)

Gender
Men active
Women active
Women inactive
Men inactive
Total
Age of the family head
1829
3039
4049
5059
More than 59 years
Total
Profession
Categories (  )
Categories ( )
Total
Family size
1 Person
2 Persons
3 Persons
4 Persons
5 Persons
67 Persons
Total

Low crisis

High crisis

3,772
28,669
9,368
1,094
42,903

(8.8)
(66.8)
(21.8)
(2.5)
(100.0)

3,322
24,221
8,384
959
36,886

(9.0)
(65.7)
(22.7)
(2.6)
(100.0)

1,536
7,456
14,273
11,002
8,636
42,903

(3.6)
(17.4)
(33.3)
(25.6)
(20.1)
(100.0)

1,227
6,411
12,401
9,179
7,668
36,886

(3.3)
(17.4)
(33.6)
(24.9)
(20.8)
(100.0)

24,044 (56.0)
18,859 (44.0)
42,903 (100.0)

20,680 (56.1)
16,206 (43.9)
36,886 (100.0)

5,464
12,048
8,083
10,107
6,045
1,158
42,903

4,869
10,192
6,806
8,922
4,991
1,106
36,886

(12.8)
(28.1)
(18.8)
(23.6)
(14.1)
(2.7)
(100.0)

(13.2)
(27.6)
(18.5)
(24.2)
(13.5)
(3.0)
(100.0)

Crisis intensity: we used a dummy variable with 0 Low and

1 High.
Store format: we used a dummy variable with 0 Hypermarket
and 1 Supermarket.
Brand choice (dependent): a binary variable was used with
0 National brand and 1 Store brand.

3.3. Variables included in the model

4. Analysis and results

We included marketing policy variables (product price, number

of products purchased, brand display and brand feature) and
consumer personal characteristics (age, profession, income and
family size) in the model. The variables are operationalised based
on previous research (e.g., Dhar and Hoch, 1997; Balachander and
Ghose, 2003; Bonfrer and Chintagunta, 2004; Ngobo, 2011b). The
operationalisation of the variables used in the model is as follows:

4.1. Main effects and interaction

Product price: the price of a given product in store s that is

patronised by household h in week t, in category g, in store s
and format f.
Quantity of products purchased: the number of products purchased by household h in week t, in category g, store s and
format f.
National brand display: we employed a variable coded 0
(absence) or 1 (presence) in week t, in category g, in store s
and format f.
National brand feature: we employed a variable coded 0
(absence) or 1 (presence) in week t, in category g, store s and
format f.
Age: the age of the family head for household h.
Profession: households were divided into two categories coded
1 Lower profession categories (CSP-) and 2 Higher profession categories (CSP).
Income: the total income of household h per month (categorisation with eight classes).
Family size: the number of persons in household h.

73

In this section, we analyse descriptive statistics on consumer

behavior and promotion variables in the panel, before presenting the
results of the logistic regression model for brand choice in times of
crisis. Table 2 highlights three main trends during crisis periods: rst
the number products purchased decreases for both national and
store brands. However, national brand products are overall more
affected by crisis times than store brands which market shares
remain stable for some product categories (e.g. pasta). Second,
during the crisis periods investigated the average prices remain
almost the same. Only average shampoo price (national brands)
decreased signicantly when crisis became harder (more intense).
This trend seems surprising as one would expect more price cuts
when crisis hits further. Finally, we notice signicant differences in
promotion usage. While the number of products purchased under
brand display decreased over turbulent times, products purchased
under brand feature increased harshly. These trends illustrate
different consumer reactions to manufacturers and retailers actions
when crisis intensies.
The results of the logistic regression model are presented in
Table 3 (NNational Brand 54,431 and NStore Brand 25,358). In terms of
goodness of t, the model was able to classify correctly 75.7% of the
observations with respect to their chosen brand. Besides, the
Nagelkerke R2 (0.43) value indicated that the independent variables
explain a substantial amount of variance in the dependent variable.
Finally, the log-likelihood test,  2log 70,417.84 is signicant at

74

J. Kaswengi, M.F. Diallo / Journal of Retailing and Consumer Services 23 (2015) 7076

Table 2
Descriptive statistics across product categories investigated.
Low crisis (2008) N 42,903

High crisis (2009) N 36,886

P1

P2

P3

P4

P1

P2

P3

P4

Number of brand
purchases

NB: N 7170
SB: N 1399

NB: N 5991
SB: N 2443

NB: N 10,427
SB: N 8,148

NB: N 6162
SB: N 1163

NB: N 6154
SB: N 1077

NB: N 4691
SB: N 2025

NB: N 9,089
SB: N 8,143

NB: N 4747
SB: N 960

Price product

NB: M(sd) 2.35

(1.16)
SB: M(sd) 1.23
(0.49)

NB: M(sd) 2.97

(1.50)
SB: M(sd) 1.46
(0.64)

NB: M(sd) 1.43

(0.79)
SB: M(sd) 1.00
(0.45)

NB: M(sd) 3.49

(1.51)
SB: M(sd) 1.55
(0.77)

NB: M(sd) 2.35

(1.15)
SB: M(sd) 1.27
(0.59)

NB: M(sd) 2.99

(1.28)
SB: M(SD) 1.48
(0.71)

NB: M(sd) 1.44

(0.79)
SB: M(SD) 1.07
(0.45)

NB: M(sd) 3.23

(1.48)
SB: M(SD) 1.53
(0.61)

Brand display

No: N 8219
Yes: N 350

No: N 7287
Yes: N 1147

No: N 17,628
Yes: N 947

No: N 6570
Yes: N 755

No: N 6920
Yes: N 311

No: N 5756
Yes: N 960

No: N 15,968
Yes: N 1,264

No: N 5193
Yes: N 514

Brand feature

No: N 8534
Yes: N 53

No: N 8386
Yes: N 48

No: N 17,673
Yes: N 48

No: N 6605
Yes: N 720

No: N 7092
Yes: N 139

No: N 6295
Yes: N 421

No: N 16,123
Yes: N 1,109

No: N 5404
Yes: N 303

Note: P1 toothpaste; P2 shower gel; P3 pasta; P4 shampoo; NB national brand; SB store brand: M(sd) means (average price).

Table 3
Effects of marketing variables and socio-demographic factors, (N 79,789 with NCrisis

low 42,903

and NCrisis

high 36,886).

Independent variables

Main effects model

Model with interaction effects

Marketing variables

SE

Wald

Sig.

SE

Wald

Sig.

Intercept
Product price
Quantity of products purchased
Brand display
Brand feature
Crisis  Product price
Crisis  Quantity of products
Crisis  Brand display
Crisis  Brand feature

3.50
 2.124
1.662
 1.322
 1.122

.089
.019
.030
.041
.051

1,547.453
13,145.885
2,991.636
1,053.372
484.672

.000
.000
.000
.000
.000

3.451
 2.058
1.447
 1.230
 .988
 .150
.468
 .138
 .217

.090
.024
.040
.055
.068
.035
.060
.070
.080

1480.460
7323.010
1300.346
501.824
213.516
18.074
61.691
3.941
7.432

.000
.000
.000
.000
.000
.000
.000
.047
.006

Socio-demographics variables
Age
Profession
Income
Family size
Crisis  Age
Crisis  Profession
Crisis  Income
Crisis  Family size

 .063
 .096
 .024
.024

SE
.004
.022
.005
.005

Wald
283.610
18.617
19.523
19.511

Sig.
.000
.000
.000
.000

 .065
 .127
 .025
.018
.003
.067
.001
.014

SE
.005
.030
.007
.007
.007
.044
.010
.011

Wald
167.071
17.707
11.471
5.592
.179
2.360
.019
1.805

Sig.
.000
.000
.001
.018
.673
.124
.891
.179

Table 4
Similarities and differences across store formats.
Independent variables

Intercept
Product price
Quantity of products purchased
Brand display
Brand feature
Age
Profession
Income
Family size

Crisis low

Crisis high

Hyper
(N 25,876)

Sig.

Super
(N 17,027)

Sig.

Hyper
(N 22,035)

Sig.

Super
(N 14,851)

Sig.

3.084
 2.155
1.716
 1.059
 .810
 .066
 .122
 .035
.034

.000
.000
.000
.000
.000
.000
.002
.000
.000

4.033
 1.943
.962
 1.454
 1.152
 .065
 .119
 .011
 .008

.000
.000
.000
.000
.000
.000
.015
.371
.515

3.404
 2.239
2.054
 1.422
 1.131
 .053
 .097
 .028
.041

.000
.000
.000
.000
.000
.000
.024
.006
.000

3.391
 2.186
1.522
 -.938
 1.257
 .077
.025
 .021
.015

.000
.000
.000
.000
.000
.000
.628
.096
.245

Note: Hyper hypermarket; Super supermarket.

the 0.001 level. Consequently, we can consider that the binary

regression model has a good t to the data.
The results show that product price affects negatively store brand
choice while quantity of products purchased has a positive effect on it
(H1 and H2 supported). National brand display and feature have also

negative effects on store brand choice, meaning that manufacturers

should focus on promotion activities during economic crisis (H3 and H4
supported). The effects of consumer personal characteristics are also
signicant. While age, profession and income affect negatively store
brand choice, family size is found to inuence positively store brand

J. Kaswengi, M.F. Diallo / Journal of Retailing and Consumer Services 23 (2015) 7076

choice over national brands (H5 supported). We also assessed the

moderating role of crisis intensity on the investigated relationships. Our
results indicate two main trends. On the one hand, crisis intensity
affects the relationships between marketing policy variables and store
brand choice (H6 supported). On the other hand, crisis intensity affects
less extensively the relationships between socio-demographic variables
and store brands choice over national brand (H7 partially supported).
These results are discussed in Section 5.
4.2. Effects across store formats
As several prior studies emphasised the role of store format on
brand choice (e.g., Hsieh and Stiegert, 2012; Zielke, 2010), we
assessed the investigated relationships in two different store formats: hypermarket (N 47,911) and supermarket (N31,878). These
retail formats were chosen because they constitute the main formats
in the French market. We obtained a good model t for each retail
format. For instance, the Nagelkerke R2 value is 0.47 in hypermarket
and 0.39 in supermarket,  2log values are respectively 19,171.32
and 13,279.75. These values are both signicant at the 0.001 level.
Results (see Table 4) indicate signicant differences in income and
family size when crisis intensity is low, while the other variables are
only affected by store format in terms of strength. When crisis
intensity is high, we notice similar patterns in consumer behaviour
across store formats. However, three variables (instead of two) then
relate to store format (profession, income and family size). Marketing variables (price, quantity, display and feature) variables are not
signicantly affected by store format in either crisis period considered. These results show that consumers adjust their format choice
based mainly on personal characteristics to cope with crisis situations. They give support to H9, but reject H8. We discuss these
results in the next section.

5. Discussion and implications

5.1. Theoretical and managerial implications
The aim of this study was to understand how marketing variables
and consumer characteristics affect consumer choice of store brands
over national brands in two different retail formats in crisis periods.
This research highlights three main theoretical contributions. First, it
underlines the role of marketing policy variables and sociodemographic factors on store brand purchase over national brand in
times of crisis. These results are in line with previous studies
indicating the greater importance of store brands when crisis strikes
consumers (Hoch and Banerji, 1993). Consequently, consumers and
retailers make adjustments in times of crisis to cope with limited
resources. Our study shows that point of sale promotion of national
brands (brand display and brand feature) is a key variable that affects
negatively store brand choice in times of crisis. This nding conrms
prior studies on store brand and the business cycle adopting a longterm orientation (Lamey et al., 2007, 2012). The effect of sociodemographic variables on store brand choice was also analysed by
Ngobo (2011a) who reported similar effects. However, our research
adds further insights by focusing on the effect of crisis period. Second,
as far as we know, this paper is the rst to analyse the moderating
role of crisis intensity on the relationships between marketing
variables and consumer characteristics and store brand choice over
national brands in different store formats. By emphasising the
differences on the moderating effect of crisis intensity depending on
the nature of the variables, this research extends previous studies on
the business cycle (Claeys and Cauberghe, 2014; Hoch and Banerji,
1993; Lamey et al., 2007), and adds a new understanding of consumer
behaviour towards brands in times of crisis. Third, this paper also
underlined the role of store format in times of crisis. We show that

75

both store formats are affected by crisis. We point out different

consumer adjustments depending on store format to face nancial
constraints in hard economic situations. In fact, our results indicate
signicant differences between store format for income and family
size when crisis intensity is low, while the other variables are only
affected by store format in terms of strength. When crisis intensity is
high, differences in store format arise not only for income and family
size, but also for profession. Overall, consumers tend to base their
format adjustment strategies on personal characteristics instead of
marketing policy variables when crisis intensity is high. Therefore,
this research completes previous studies on the relationship between
store format and brand purchase behaviour (Zielke, 2010; Kaswengi,
2013; Kumar and Roy, 2013) and helps to better understand how
specic retail formats are affected when crisis hits.
Several managerial implications can also be derived from our
ndings. First, marketers should be aware that when crisis strikes
consumers, they are likely to make adjustments based on both
marketing variables (e.g., product price and quantity purchased)
and their personal characteristics (e.g., age, profession, income).
Consequently, retailer strategies during times of crisis should be
diversied in order to meet the need of diverse consumer segments.
For manufacturers, it is clear from this research that they should
strengthen their promotions/advertising strategies (brand display,
brand feature) to counter store brand market share rise in times of
crisis. Consequently, in line with Van Heerde et al. (2008), we caution
high-end market players (i.e., national brands) about the risk of using
price as a competitive weapon in dire economic situations because
this may increase price sensitivity. The overall image of national
brands would also be affected in a price war with store brands.
Rather, they should innovate further to strengthen their market
position. Second, as crisis intensity has different moderation effects
depending on the nature of the variables, retailers should adjust
consistently their management during times of crisis. For instance,
the focus should be laid on marketing variables as they are found to
affect more extensively consumer brand choice when crisis intensity
is high. Third, as store formats are differently affected in terms of
store brand choice over national brands, they should be managed
more carefully in a period of crisis. For instance, our research shows
that consumers with lower income and higher family size tend to
shop in hypermarkets instead of supermarkets when crisis intensity
is low. However, when crisis intensity is high, not only adjustments
are made based on income and family size, but profession is also
affected. In fact, lower profession categories tend to switch store
frequentation in favor of hypermarkets where they can nd promotion and cheaper products. Consequently, supermarket stores should
especially pay attention to lower income consumers by focusing on
marketing variables (mainly price, brand display and brand feature)
to avoid market shares losses in a challenging period.
5.2. Limitations and future research studies
Our research has some limitations and these provide suggestions
for future research. First, our analysis has concerned one test market in
France. Many countries of the world, especially developed countries
also faced slowdown and economic depression. Therefore, generalisability can be enhanced, replicating this study with more countries,
including different types of consumers, and other cultures. Culture
differences may moderate the effect of marketing variables on store
brand choice. Recently, Steenkamp and Geyskens (2014) have identied cross-national regularities as to the role of nine manufacturer and
retailer factors in explaining SB market share. It would be interesting
to understand how global economic crisis affects the three strategies
underlined by these authors (global integration, local adaptation, and
worldwide learning). Second, our model tested only a few marketing
factors. Additional research should incorporate more marketing
actions (such as TV advertising) and perceptual variables such as

76

J. Kaswengi, M.F. Diallo / Journal of Retailing and Consumer Services 23 (2015) 7076

satisfaction (Martnez-Ruiz et al., 2010). In fact, advertising researchers

found that repetitive advertising schedules increase the probability for
a brand to be included in the consideration set and then the
probability of its being chosen. We also excluded promotion with
private labels because their marketing support is typically governed by
very different decision processes than national brands (Van Heerde
et al., 2013). Third, because marketing and brand policies differ across
product categories (Wedel and Zhang, 2004), future research needs to
have deeper understanding of brand choice with wider product
categories in retail chains. Despite these limitations, our research
contributes to a better understanding of the drivers of store brand
choice in times of crisis.
References
Ataman, B., Mela, C.F., van Heerde, H.J., 2007. Consumer packaged goods in France:
national brands, regional chains, and local branding. J. Marketing Res. 44 (1),
1420.
Balachander, S., Ghose, S., 2003. Reciprocal spillover effects: a strategic benet of
brand extensions. J. Marketing 67 (1), 423.
Bonfrer, A., Chintagunta, P., 2004. Store brands: who buys them and what happens
to retail prices when they are introduced? Rev. Ind. Organ. 24, 195218.
Burton, S., Lichtenstein, D., Netemeyer, R., Garretson, J., 1998. A scale for measuring
attitude toward private label products and an examination of its psychological
and behavioral correlates. J. Acad. Marketing Sci. 26 (4), 293306.
Claeys, A.S., Cauberghe, V., 2014. What make crisis response strategies work? The
impact of crisis involvement and message framing. J. Bus. Res. 67 (2), 182189.
Dhar, S.K., Hoch, S.J., 1997. Why store brand penetration varies by retailer?
Marketing Sci. 16 (3), 208227.
Diallo, M.F., Chandon, J.L., Cliquet, G., Philippe, J., 2013. Factors inuencing
consumer behaviour towards store brands: evidence from the French market.
Int. J. Retail Distrib. Manage. 41 (6), 422441.
Engemann, K., Wall, H.J., 2009. The Effects of Recessions Across Demographic
Groups. Working Paper 2009-052 A, Federal Reserve Bank of St Louis, available
at SSRN: http://research.stlouisfed.org/wp/2009/2009-052.pdf.
Erdem, T., Zhao, Y., Valenzuela, A., 2004. Performance of store brands: a crosscountry analysis of consumer store-brand preferences, perceptions, and risk.
J. Marketing Res. 41 (1), 86100.
Ferguson, J.L., 2014. Implementing price increases in turbulent economies: pricing
approaches for reducing perceptions of price unfairness. J. Bus. Res. 67 (1),
27322737.
Hampson, D.P., McGoldrick, P.J., 2013. A typology of adaptive shopping patterns in
recession. J. Bus. Res. 66 (7), 831838.
Hsieh, M.F., Stiegert, K.W., 2012. Store format choice in organic food consumption.
Am. J. Agric. Econ. 94 (2), 307313.
Hoch, S.J., Banerji, S., 1993. When do private labels succeed? Sloan Manage. Rev. 34,
5767.
Hoynes, H.W., 1999. The Employment, Earnings, and Income of Less Skilled Workers
Over the Business Cycle. NBER Working Paper No. w7188, available at SSRN:
http://ssrn.com/abstract=171542.
Jones, E., 2014. Consumer preferences for national brands and private labels: do
business cycles matter? National Brands and Private Labels in Retailing. In: J.C.
Gazquez-Abad et al. (Eds.), Springer Proceedings in Business and Economics.
DOI 10.1007/978-3-319-07194-7_10.
Kaswengi, J., 2013. Les effets de limage du magasin sur le capital-marque: une
tude travers les circuits de magasin. Management et Avenir 60 (2), 102119.
Kaytaz, M., Gul, M.C., 2014. Consumer response to economic crisis and lessons for
marketers: the Turkish experience. J. Bus. Res. 67 (1), 27012706.
Krishnamurthi, L., Raj, S.P., 1988. A model of brand choice and purchase quantity
price sensitivities. Marketing Sci. 7 (1), 120.
Kumar, V., Karande, K.W., 2000. The effect of retail store promotions on brand and
store substitution. J. Bus. Res. 49 (2), 167181.
Kumar, A., Roy, S., 2013. Store attribute and retail format choice. Adv. Manage. 6
(11), 2733.
Lamey, L., Deleersnyder, B., Dekimpe, M.G., Steenkamp, J.B.-E.M., 2007. How
business cycles contribute to private-label success: evidence from the United
States and Europe. J. Marketing 71 (1), 115.
Lamey, L., Deleersnyder, B., Dekimpe, M.G., Steenkamp, J.B.-E.M., 2012. The effect of
business-cycle uctuations on private-label share: what has marketing conduct
got to do with it? J. Marketing 76 (1), 119.
Lemon, K.N., Nowlis, S.M., 2002. Developing synergies between promotions and
brands in different price-quality tiers. J. Marketing Res. 39 (2), 171185.
Lichtenstein, D.R., Ridgway, N.M., Netemeyer, R.G., 1993. Price perceptions and
consumer shopping behavior: a eld study. J. Marketing Res. 30 (2), 234245.
Liljander, V., Polsa, P., van Riel, A., 2009. Modelling consumer responses to an
apparel store brand: store image as a risk reducer. J. Retailing Consum. Serv.
16 (4), 281290.
Martnez-Ruiz, M.P., Jimnez-Zarco, A.I., Izquierdo-Yusta, A., 2010. Customer satisfactions key factors in Spanish grocery stores: evidence from hypermarkets
and supermarkets. J. Retailing Consum. Serv. 17 (4), 278285.

Martos-Partal, M., Gonzlez-Benito, ., 2011. Store brand and store loyalty: the
moderating role of store brand positioning. Marketing Lett. 22 (3), 297313.
Mela, C.F., Gupta, S., Jedidi, K., 1998. Assessing long-term promotional inuences on
market structure. Int. J. Res. Marketing 15 (2), 89107.
Millet, K., Lamey, L., Van den Bergh, B., 2012. Avoiding negative vs. achieving
positive outcomes in hard and prosperous economic times. Organ. Behav. Hum.
Decis. Processes 117 (2), 275284.
Ngobo, P.-V., 2011a. Private label share, branding strategy and store loyalty.
J. Retailing Consum. Serv. 18 (4), 259270.
Ngobo, P.-V., 2011b. What drives household choice of organic products in grocery
stores? J. Retailing 87 (1), 90100.
Ngobo, P.-V., Jean, S., 2012. Does store image inuence demand for organic store
brands? J. Retailing Consum. Serv. 19 (6), 621628.
Richardson, Jain, A., P., Dick, A., 1996. Household store brand proneness: a framework. J. Retailing 72 (2), 159186.
Seock, Y.K., 2009. Inuence of retail store environmental cues on consumer
patronage behavior across different retail store formats: An empirical analysis
of US Hispanic consumers. J. of Retailing and Consum. Serv. 16 (5), 329339.
Sinha, I., Batra, R., 1999. The effect of consumer price consciousness on private label
purchase. Int. J. Res. Marketing 16 (3), 237251.
Steenkamp, J.-B.E.M., Geyskens, I., 2014. Manufacturer and retailer strategies to
impact store brand share: global integration, local adaptation, and worldwide
learning. Marketing Sci. 33 (1), 626.
Tellis, G.J., Tellis, K., 2009. A critical review and synthesis of research on advertising
in a recession. J. Advertising Res. 49 (3), 304327.
Tripathi, G., Dave, K., 2013. Store format choice and relationship quality in apparel
retail: a study of young and early-middle aged shoppers in New Delhi region.
J. Retailing Consum. Serv. 20 (5), 479487.
Van Heerde, H.J., Gijsbrechts, E., Pauwels, K., 2008. Winners and losers in a major
price war. J. Marketing Res. 45 (5), 499518.
Van Heerde, H.J., Gisjenberg, M.J, Dekimpe, M., Steenkamp, J.B.E.M., 2013. Price and
advertising effectiveness over the business cycle. J. Marketing Res. 50 (2),
177193.
Wedel, M., Zhang, J., 2004. Analyzing brand competition across subcategories.
J. Marketing Res. 41, 448456.
Yoo, B., Donthu, N., Lee, S., 2000. An examination of selected marketing mix
elements and brand equity. J. Acad. Marketing Sci. 28 (2), 195211.
Zhang, J., 2006. An integrated choice model incorporating alternative mechanisms
for consumers reactions to in-store display and feature advertising. Marketing
Sci. 25 (3), 278290.
Zielke, S., 2010. How price-image dimensions inuence shopping intentions for
different store formats? Eur. J. Marketing 44 (6), 748770.

Dr Mbaye Fall Diallo is an Assistant professor at University of Lille 2 (Institute of

Retail Marketing and ManagementIMMD) and a researcher at the LSMRC Lab
(Univ Lille Nord de France-Skema Business School). He is member of scientic
associations including Association for Consumer Research, Association for Education and Research in Commercial Distribution, European Marketing Academy and
the French Association of Marketing (AFM). His research interests lie in the eld of
retail management, store brands and quantitative data analysis (structural equation
models and econometrics of panel data). His work has been published in journals
such as European Business Review, Journal of Retailing and Consumer Services,
International Business Research, International Journal of Retail and Distribution
Management, Service Industries Journal, The Business Review, Revue Management &
Avenir, Revue Franaise de Gestion, Recherche et Applications en Marketing, and in
conferences proceedings such as INFORMS Marketing Science, European Association
for Education and Research in Commercial Distribution, European Marketing Academy,
International Conference of Marketing Trends, Academy of International Business,
Association Franaise de Marketing Congress, etc. He is the author/co-author of two
books and ve book chapters on retail branding and management. Additionally, he
has been a visiting scholar at University of Brasilia (Brazil) two times. He serves as a
reviewer for several academic journals (International Journal of Retail & Distribution
Management, Service Industries Journal, Journal of Consumer Behaviour, International
Review of Retail, Distribution and Consumer Research, Economies et Socits, ) and
scientic conferences (AMA, EMAC, EAERCD, AFM, Etienne THIL, ).

Dr Joseph Kaswengi is an Assistant professor at University of Orlans and member

of VALLOREM Lab (University of Orlans and University of Tours). Email: joseph.
kaswengi@univ-orleans.fr; Address: Vallorem, Rue de Blois BP 6739 45067 Orlans
Cedex 2, France. He obtained his Ph.D. at University of Orlans (Vallorem). He is
member of scientic associations such as the Gaston Berger Federation and the
French Association of Marketing. His research lies in the areas of retail marketing,
brand management and quantitative techniques applied to marketing, as well as
revenue management. He has published in journals such as Revue Management et
Avenir and presented papers in international and national conferences such as
French Association of Marketing (AFM), International Symposium on National Brands
and Private Labels, Mthodes Marketing, International Conference of Marketing Trends,
etc. He is the author of a book on brand equity-point of sale relationship and a
chapter on national-store brands. He is also involved in consulting.