(electronic) How Business Cycles Contribute to Private-Label Success: Evidence from the United States and Europe

Abstract

The growth of private labels over the past decades has been attributed to various factors. This article formally addresses the link between private-label success and economic expansions and contractions using recently developed time-series/econometric techniques. The findings confirm conventional wisdom that a country's private-label share increases when the economy is suffering and shrinks when the economy is flourishing. However, asymmetries are found in the extent to and speed with which private-label share changes in cyclical up-versus downturns. Consumers switch more extensively to store brands during bad economic times than they switch back to national brands in a subsequent recovery. In addition, the switch to private-label brands is faster than the opposite movement to national brands after the recession ends. Finally, not only are consumers more prone to buy private labels during economic downturns, but some keep buying them when bad economic times are long over as well, leaving permanent " scars " on national brands' performance level. The authors argue that national-brand manufacturers can mitigate the effect of an economic downturn on their shares by intensifying their marketing-support activities in recessions. Such a proactive strategy is not often observed. On the contrary, available evidence suggests that many manufacturers exacerbate their predicament by cutting back on their marketing expenses when the economy turns sour. Most retailers invest more strongly in their private-label program when the economy deteriorates, making it even more difficult for national brands to catch up with the share lost during contractions.

Full text

1
Journal of Marketing
Vol. 71 (January 2007), 1–15
© 2007, American Marketing Association
ISSN: 0022-2429 (print), 1547-7185 (electronic)
Lien Lamey, Barbara Deleersnyder, Marnik G. Dekimpe, &
Jan-Benedict E.M. Steenkamp
How Business Cycles Contribute to
Private-Label Success: Evidence
from the United States and Europe
The growth of private labels over the past decades has been attributed to various factors. This article formally
addresses the link between private-label success and economic expansions and contractions using recently
developed time-series/econometric techniques. The findings confirm conventional wisdom that a country’s private-
label share increases when the economy is suffering and shrinks when the economy is flourishing. However,
asymmetries are found in the extent to and speed with which private-label share changes in cyclical up- versus
downturns. Consumers switch more extensively to store brands during bad economic times than they switch back
to national brands in a subsequent recovery. In addition, the switch to private-label brands is faster than the
opposite movement to national brands after the recession ends. Finally, not only are consumers more prone to buy
private labels during economic downturns, but some keep buying them when bad economic times are long over as
well, leaving permanent “scars” on national brands’ performance level. The authors argue that national-brand
manufacturers can mitigate the effect of an economic downturn on their shares by intensifying their marketing-
support activities in recessions. Such a proactive strategy is not often observed. On the contrary, available evidence
suggests that many manufacturers exacerbate their predicament by cutting back on their marketing expenses when
the economy turns sour. Most retailers invest more strongly in their private-label program when the economy
deteriorates, making it even more difficult for national brands to catch up with the share lost during contractions.
Lien Lamey is a doctoral candidate in Marketing, Catholic University Leu-
ven, Belgium (e-mail: lien.lamey@econ.kuleuven.be). Barbara Deleersny-
der is Assistant Professor in Marketing, Rotterdam School of Manage-
ment, Erasmus University, the Netherlands (e-mail: bdeleersnyder@rsm.
nl). Marnik G. Dekimpe is Research Professor of Marketing, Tilburg Uni-
versity, the Netherlands, and Professor of Marketing, Catholic University
Leuven (e-mail: m.g.dekimpe@uvt.nl). Jan-Benedict E.M. Steenkamp is
C. Knox Massey Distinguished Professor of Marketing and Market-
ing Area Chair, University of North Carolina, Chapel Hill (e-mail:
JBS@unc.edu). The authors thank the three anonymous
JM
reviewers,
Christophe Croux, Philip Hans Franses, and participants of the 2005 Mar-
keting Science Conference for their valuable comments. They also thank
AiMark, Stephen Hoch, and Hilde Vanderheyden (ACNielsen) for provid-
ing the data. Finally, the authors gratefully acknowledge financial support
from the Flemish Science Foundation FWO (Grant No. G.0116.04), from
the Marketing Science Institute through the 2005 Alden G. Clayton Doc-
toral Dissertation Competition, and from the Zyman Institute of Brand Sci-
ence through the 2005 ZIBS Doctoral Dissertation Competition.
To read or contribute to reader and author dialogue on
JM
, visit
http://www.marketingpower.com/jmblog.
Private-label products now account for more than 20%
of global grocery sales and are expected to grow to
30% by 2020 (M+M Planet Retail 2004). Although
considerable cross-country differences still exist—a 2003
ACNielsen survey reports an aggregate private-label share
of approximately 15% in the United States, which is below
the level in many Western European countries, such as
Switzerland (38%), Spain (23%), and France (21%)—all
developed countries have witnessed a steady increase in the
share commanded by store brands over the past decades. In
the United Kingdom, for example, private-label share rose
from 21.5% in 1980 to 39.3% in 2003, and Belgium wit-
nessed a growth from 11.4% in 1983 to 30.1% in 2003. This
growing success poses a serious challenge to manufacturers
of consumer goods and has been attributed to factors such
as a gradual shift in the communication budget from adver-
tising to sales promotions (Hoch, Montgomery, and Park
2002), the growing concentration in the retail sector (Hoch
and Banerji 1993), an improvement in the quality of private
labels over time (Steenkamp and Dekimpe 1997), and the
increasing efforts that retailers put into their private-label
programs (Hoch 1996).
Various authors have also linked private-label perfor-
mance to economic conditions. For example, Quelch and
Harding (1996, p. 99) observe that “private-label market
share generally goes up when the economy is suffering and
down in stronger economic periods.” Likewise, Nandan and
Dickinson (1994) state that during difficult economic times,
the popularity of private labels tends to increase, whereas in
periods of relative economic prosperity, the share of
national brands increases. A similar feeling is echoed in
many business reports. A Deloitte & Touche (2003, p. 2)
report argues that “private labels have typically experienced
significant growth in times of recession, due to their low
prices, and the reduced disposable income of households.”
This does not bode well for national-brand manufacturers.
Unlike other drivers of private-label success, the general
economic conditions are largely beyond their control. It is
possible, however, that the opposing effects during contrac-
tions and expansions will cancel out each other. However,
some sources have speculated that despite economic recov-

2 / Journal of Marketing, January 2007
ery, legions of consumers are not willing to switch back to
manufacturer brands, even when they can again afford them
(The Wall Street Journal 1993). If this pattern is systematic,
the sequence of recessions and expansions would contribute
to the prolonged upward evolution in private-label share
over repeated business cycles.
Furthermore, is the rate of change in private-label share
symmetric across expansion versus contraction periods? For
example, do grocery shoppers quickly switch to less expen-
sive private-label brands as the economy deteriorates, and
are they more hesitant to switch back to national brands
when economic prospects begin to improve? Such an asym-
metric pattern would be the net result of actions by individ-
ual consumers, retailers, and/or manufacturers in response
to aggregate economic fluctuations.
The purpose of this article is to shed light on these
issues by formally investigating the relationship between
private-label success and the aggregate business cycle.
Specifically, we address the following three key research
questions:
1. Does private-label success behave countercyclically?
2. Does private-label performance behave differently (asym-
metrically) over expansion and contraction periods?
3. Does the aggregate business cycle contribute to long-term
private-label success?
Our empirical analysis encompasses four countries: Bel-
gium, the United Kingdom, the United States, and West
Germany. These countries tend to follow clearly different
private-label strategies, ranging from a focus on discount
private labels in Germany to a highly sophisticated, three-
tier private-label strategy in the United Kingdom, with Bel-
gium and the United Sates in between (Kumar and
Steenkamp 2006). The international setting contributes to
the empirical generalizability of the results (Steenkamp
2005).
We organize the remainder of this article as follows:
First, we review the literature on the relationship between
private-label success and economic conditions. Second, we
develop a framework for understanding the impact of busi-
ness cycles on private-label share. Third, using this frame-
work, we develop predictions about the effect of business-
cycle fluctuations on private-label performance. Fourth, we
describe the methodology and the data, followed by the
empirical results. Finally, we present managerial implica-
tions and suggestions for further research.
Previous Research
Some prior literature has already suggested that there is a
negative link between private-label success and the business
cycle. Quelch and Harding (1996) note that U.S. private-
label success increased remarkably during the 1981–1982
economic recession. During that period, U.S. private-label
share peaked at 17%, compared with 14% during the previ-
ous years. In a study of the impact of the recent Asian eco-
nomic crisis, Ang, Leong, and Kotler (2000) observe an
increasing trend toward buying store brands during a reces-
sion. In the only quantitative study to date, Hoch and
Banerji (1993) conclude that in the United States, variation
in private-label share is closely linked to the business cycle.
Using annual data from 1971 to 1993, they find that
changes in aggregate private-label share are negatively
related to changes in disposable income.
Our study extends previous work in several important
ways. First, to the best of our knowledge, no study has sys-
tematically analyzed the relationship between private-label
share and the economy from a business-cycle perspective.
Existing knowledge of this relationship is often limited to
the analysis of a single recession period (e.g., Ang, Leong,
and Kotler 2000), which does not allow for rigorous conclu-
sions. Only Hoch and Banerji (1993) examine a longer time
span, covering multiple contraction and expansion phases.
However, they emphasize the short run rather than the
business-cycle relationship. They difference their data
before regressing private-label share on disposable income.
The first-difference filter reweights strongly toward the
higher frequencies (associated with short-term fluctuations)
and downweights lower frequencies that are typically asso-
ciated with business-cycle fluctuations (Baxter and King
1999).
Second, unlike prior work, we do not presume that
private-label performance reacts to the same extent and with
the same speed to increases and decreases in economic
activity. We explicitly allow for asymmetric performance in
economic up- and downward periods.
Third, although several academic and business writers
have theorized that private labels grow in economic down-
turns, no study has formally investigated the extent to which
this increase is cancelled out in the subsequent recovery
period or whether part of it is permanent. This is another
key question of interest to both national brand manufactur-
ers and retailers.
Thus, to study the link between private-label success
and the general state of the economy, we (1) apply a
business-cycle filter to the data to analyze the extent of
comovement between private-label share and the aggregate
business cycle over repeated recession and expansion peri-
ods; (2) consider the potentially asymmetric nature of this
relationship; (3) explore the presence of permanent upward
shifts in private-label performance due to business cycles;
and (4) do this for several key private-label countries,
broadly covering the range of private-label strategies fol-
lowed in the marketplace. These issues have not been com-
prehensively analyzed in previous research, but they are sig-
nificant to both manufacturers and retailers.
Cyclical Sensitivity of Private-Label
Share
We propose a framework to understand the cyclical sensi-
tivity of private-label share to business-cycle fluctuations.
Our framework distinguishes between four major compo-
nents of cyclical fluctuations—cyclical comovement, deep-
ness asymmetry, steepness asymmetry, and asymmetric
growth—and three marketplace parties—consumers,
national-brand manufacturers, and retailers.
“Cyclical comovement” refers to the extent to which
fluctuations within private-label-share series behave in step,
or comove, with the country’s aggregate business cycle.
Three types of cyclical comovement may exist (Stock and

How Business Cycles Contribute to Private-Label Success / 3
Watson 1999): procyclical (movement in the same direction
as the aggregate economy), countercyclical (movement in
the opposite direction), and acyclical (no comovement).
Moreover, economic studies have found cyclical fluctua-
tions in macroeconomic aggregates, such as gross domestic
product (GDP) (Razzak 2001), unemployment (Bodman
2001), and consumption (Holly and Stannett 1995), to
behave asymmetrically across expansion versus contraction
periods. Two kinds of cyclical asymmetries are distin-
guished: deepness asymmetry and steepness asymmetry
(Sichel 1993). “Deepness asymmetry” pertains to the phe-
nomenon that the deepness of series’ troughs may differ
from the height of their peaks over business-cycle periodic-
ities. “Steepness asymmetry” refers to asymmetry in the
speed of adjustment, in that series drop faster (or slower)
than they increase over business-cycle periodicities. We
explore whether analogous asymmetries are present in
private-label share.
Even if deepness and/or steepness asymmetries are pre-
sent across expansion and contraction periods, the question
remains regarding the extent to which private-label-share
gains during a contraction are permanent, leaving a perma-
nent scar on national brands’ performance, or whether the
economic recovery is strong/long enough for the latter to
recoup fully. We estimate an asymmetric growth model
(Beaudry and Koop 1993; Cover 1992) to investigate
whether cyclical shocks cause only temporary deviations
around an underlying trend driven by other factors (e.g., a
gradual increase in the private labels’ quality, the growing
concentration in the retail industry, an expansion in the
number of chains and product categories carrying private
labels; see, e.g., Kumar and Steenkamp 2006) or whether
these cyclical disturbances alter the private labels’ long-
term growth path, with a differential persistent impact for
expansion and contraction periods.
We use these four components to understand how con-
sumers’ (demand side) and manufacturers’ and retailers’
(supply side) responses to aggregate business-cycle fluctua-
tions may result in predictable effects on private-label sales.
We believe that this framework provides a useful tool for
developing predictions about the impact of business-cycle
fluctuations on private-label share. Note that we test the
expected pattern of variation in private-label share rather
than the psychological and managerial processes that give
rise to these predictions.
Demand Side: Consumer Purchase Adjustment
over the Business Cycle
Cyclical comovement. As consumers’ abilities and will-
ingness to buy goods decrease during economic contrac-
tions, they may want to economize on their expenditures by
reducing the quantity bought or by postponing their pur-
chases until prospects become better (Katona 1975). This
strategy is less of an option for consumer packaged goods
(CPGs) because of their more necessary nature. To save on
CPG purchases, consumers are more likely to economize on
price (Shama 1981). Given that, on average, store brands
are priced 25%–30% below national brands (e.g., Kumar
and Steenkamp 2006), they become an obvious candidate to
switch to when the economy turns sour. This behavior is 1We thank an anonymous reviewer for this suggestion.
expected to lead to countercyclical movements in private-
label share, which is consistent with the finding that the
propensity to buy private labels is inversely related to
income (Ailawadi, Neslin, and Gedenk 2001). Moreover,
during economic slowdowns, consumers become more
inclined to acquire price information (Wakefield and Inman
1993), making them more price conscious (Estelami,
Lehmann, and Holden 2001). Because price consciousness
has been found to be a good predictor of private-label suc-
cess (Ailawadi, Neslin, and Gedenk 2001), this further
increases the countercyclical behavior of private-label
share.
Cyclical asymmetries. At the beginning of a recession,
consumers have a strong incentive to limit their spending
and wait for better times (Gale 1996). This is enhanced by
the notion that people lose their trust and confidence in the
economy relatively quickly, which has been found to be a
strong predictor of purchase-adjustment decisions (Allenby,
Jen, and Leone 1996). Accordingly, consumers’ willingness
to buy premium-priced national brands is expected to
decrease sharply at the onset of a recession.
Consumers may continue to economize on their expen-
ditures even if the economic climate begins to improve
again, resulting in a slower return to national brands in the
subsequent expansion. This delayed reverse movement may
emerge because (1) consumers typically find themselves at
the lowest level of their actual income or ability to buy right
after the downturn (Gale 1996); (2) any initial rise in their
income may be used to pay off debts and rebuild a precau-
tionary stock of assets or capital (Carroll 1992); and (3) the
general pessimism regarding the recent recession and their
breach of trust has made consumers more prudent in updat-
ing their income expectations, so that the anticipated
increase in their future income and wealth remains con-
servative (Katona 1975). Indeed, it is a well-known princi-
ple that has been observed in areas ranging from economics
to interorganizational relationships that it takes much less
time to breach trust than to restore it (Katona 1975; Noote-
boom, Berger, and Noorderhaven 1997). The old adage
“once bitten, twice shy” applies here too.1This gradual
return to national brands during expansions, together with
the sharp decrease in the willingness to buy national brands
during contractions, results in asymmetric steepness.
Furthermore, people change their purchase behavior
more drastically in response to a contraction than to an
expansion (Bowman, Minehart, and Rabin 1999). In this
vein, we expect that the switch to private labels during the
recession will be stronger or more extensive than the subse-
quent return to national brands in an expansion period,
resulting in higher peaks than troughs in the cyclical pattern
of private-label share (i.e., deepness asymmetry).
Asymmetric long-term growth. Overall, the aforemen-
tioned factors contribute to a pattern in which consumers
switch quickly and more strongly to private labels during
contractions but return to national brands more gradually in
subsequent expansions. However, will all consumers even-
tually switch back to national brands, resulting in the
absence of a long-lasting impact of an economic downturn?

4 / Journal of Marketing, January 2007
2As we discuss in more detail in the “Methodology” section,
deepness and steepness asymmetry refer to the nature of the cycli-
cal fluctuations around private labels’ overall trend, whereas asym-
metric growth models test whether economic contractions and
expansions alter the growth rate of that trend.
After all, the presence of steepness and deepness asymme-
try is not a sufficient condition for such a permanent effect
to occur (Beaudry and Koop 1993).2If expansions last
longer than contractions, national brands might still recover
fully and regain their precontraction market share.
However, we expect that a full recovery is not likely,
because the perceived quality of private labels is typically
lower than their actual quality (Kumar and Steenkamp
2006; Richardson, Dick, and Jain 1994). As we argued pre-
viously, an economic recession induces several national-
brand buyers to switch to a private label. From actual prod-
uct experience, which is the most important source of
information for consumers to form their quality perceptions
(Steenkamp 1989), they may learn that actual private-label
quality exceeds their prior perceptions. On the basis of con-
sumers’ positive experiences with private labels, we expect
that some will not switch back to national brands after the
recession, giving rise to a permanent, positive effect of an
economic contraction on private-label performance. In a
similar vein, Ailawadi and Keller (2004) argue that if store
brands can change consumers’ quality perceptions (e.g.,
through direct consumption experience), they will gain
more customers, and it will be difficult for manufacturers to
win them back.
Supply Side: Manufacturer and Retailer Behavior
over the Business Cycle
Cyclical comovement. Manufacturer and retailer actions
may reinforce consumers’ tendencies to switch to private
labels during contractions. Most manufacturers decrease
their brand support during bad economic times (e.g., Srini-
vasan, Rangaswamy, and Lilien 2005). Moreover, manufac-
turers are loath to reduce prices in a recession (Backus and
Kehoe 1992; Deleersnyder et al. 2004), even though price
sensitivity is higher at such a time (Estelami, Lehmann, and
Holden 2001). This behavior may contribute to the counter-
cyclical behavior in private-label share.
Conversely, retailers tend to rejuvenate their own label
when the economy sours (Hoch 1996). When the economy
softened in the early 1980s and again at the beginning of the
1990s, numerous U.S. supermarkets revamped their private-
label programs with new logos, new items, and more shelf
space (Hoch 1996). Similar patterns have been observed in
Europe (Walters 1994) and Asia (Davies 2000). Through
such countercyclical private-label-support activities, retail-
ers contribute to the countercyclical performance of store
brands.
Cyclical asymmetries. Prior research has found that
firms adjust their business strategies asymmetrically over
different business-cycle phases (Mascarenhas and Aaker
1989). Brand support tends to be cut more quickly during
recessions than it is restored when good economic times
arrive again (Axarloglou 2003). Prices have also been found
to be more flexible going up than going down (Ball and
Mankiw 1994) and to be higher during contractions than
during expansions (Backus and Kehoe 1992; Deleersnyder
et al. 2004). As a consequence of such asymmetric manu-
facturer and retailer behavior, consumers’ switch to private
labels during the recession will be faster and more extensive
than their subsequent return to national brands in an expan-
sion period. This results in steepness and deepness asym-
metry patterns in private-label share.
Asymmetric long-term growth. It is less clear whether
long-term growth is also related to manufacturer and
retailer actions. Even when manufacturers and retailers
adjust their business strategies significantly and asymmetri-
cally over different business-cycle phases, it is unclear
whether brand manufacturers can regain lost ground by
ramping up marketing investments in economic expansions
or whether their retrenchment during recessions generates a
decline in expertise (e.g., in research and development) that
they cannot fully recoup during the subsequent expansion
(Fatás 2002; Hillier and Baxter 2001).
Table 1 summarizes the expected effect of demand- and
supply-side behavior on private-label performance along
the four cyclical characteristics: nature of the comovement,
deepness and steepness asymmetry, and asymmetric
growth. Although it would be of interest to quantify the
impact from each of the preceding members separately, it is
Net Effect on
Private-Label
Business Cycle Consumer Manufacturer Retailer Share
Cyclical Comovementa––––
Cyclical Asymmetries
Deepness asymmetry + + +
Steepness asymmetry + + +
Asymmetric Growth
Contraction ++ +
Expansion –
aA negative sign means that private-label share is expected to behave countercyclically.
TABLE 1
Framework to Relate the Business Cycle to Private-Label Share
Determinants of Private-Label Share
Demand Side Supply Side

How Business Cycles Contribute to Private-Label Success / 5
important to understand first the overall net impact of the
aggregate business cycle on private-label share.
Methodology
Our research methodology consists of four steps. First, we
apply the well-known Hodrick and Prescott (1997) filter
(hereinafter, the HP filter) to isolate the cyclical component
in the various time series of interest. Second, we quantify
the extent of cyclical sensitivity in private-label success
through the cyclical comovement elasticity. Third, we
derive statistics to test for deepness and steepness asymme-
try. Fourth, we assess whether expansion and contraction
periods affect the long-term growth rate in private-label
share differently.
Extracting the Cyclical Component
Not all over-time variation present in a series can be attrib-
uted to business-cycle movements. Therefore, in line with
economic studies (e.g., Cook 1999; Holly and Stannett
1995), we adopt the HP filter to extract from each individ-
ual series the fluctuations that occur at business-cycle peri-
odicities. The HP filter decomposes a time series, yt, into a
trend component, which varies smoothly over time, and
a cyclical component, by fitting a smooth curve through
a set of data points. The variance of the cyclical component
is minimized subject to a penalty for variation in the second
difference of the trend component. The cyclical component,
which fluctuates around that trend, is then obtained by sub-
tracting the long-term trend from yt—that is,
More formally, the HP filter obtains by minimizing
where λis a penalty parameter that determines the degree
of smoothing; the larger its value, the smoother is the result-
ing growth component. Because business cycles are of vary-
ing length that tend to last no longer than eight years (Chris-
tiano and Fitzgerald 1998), we choose our smoothing
constant to generate a trend that accounts for all fluctua-
tions longer than eight years. We follow the work of Baxter
and King (1999), who recommend a value of λequal to ten
for annual series. This value produces a good correspon-
dence between the HP filter and an ideal band-bass filter
that passes through cycles between two and eight years.
Furthermore, to enhance the comparability across series, we
analyze ytin logarithms, so that the units of when mul-
tiplied by 100, represent percentage deviations from the
series’ growth path (Stock and Watson 1999).
The proposed HP filter has several attractive features.
First, the HP filter does not induce any asymmetries in the
cyclical component when they were absent in the original
series (Sichel 1993). This is a necessary property when
exploring the presence of cyclical asymmetries (see the
“Quantifying the Extent of Cyclical Sensitivity” subsec-
tion). Second, because it removes a stochastic trend from a
series, the HP filter induces stationarity in a series that is
originally difference stationary (Baxter and King 1999;
Holly and Stannett 1995). This avoids the well-known
yt
c,
() ( ) [(1
1
21
2
1
yy y y
t
t
T
tt
t
T
=
+
=
−
∑∑
−+ −
lλltttt
yy
lll
)( )],−−
−12
yt
lyyy
t
ctt
=−
l.
yt
c,
yt
l,
3It is not necessary to include an intercept, because both the
private-label-share and the GDP-per-capita series are zero-
reverting after filtering. We thank an anonymous reviewer for
bringing this to our attention.
spurious-correlation problem when correlating integrated
variables. Finally, the HP filter can be regarded as a special
case of a structural time-series model that consists of a
trend and a cyclical component. This property will prove to
be useful in our validation exercises when we want to con-
trol for a potential break caused by the German reunifica-
tion in 1990 (for details, see Appendix A).
Nevertheless, as with any business-cycle filter, the HP
filter suffers from some drawbacks. First, it is unable to dis-
tinguish short-term from business-cycle fluctuations
(Canova 1998; Reeves et al. 2000). Indeed, the HP filter is a
high-pass filter, in that it filters out fluctuations with small
frequencies. Because business cycles tend to have a period-
icity between two and eight years (Christiano and Fitzger-
ald 1998), band-pass filters, which can filter out both high
and low frequencies, may seem better suited. However, with
annual data (as we use), band-pass and high-pass filters
become equivalent (Baxter and King 1999) because the
Nyquist frequency (i.e., the highest frequency about which
there is direct information) then corresponds to a compo-
nent of two years (Granger and Hatanaka 1964). Second, it
has been argued that the HP filter may (as do many other
filters) induce spurious cycles within a series, thus causing
spurious cross-correlations among filtered series (Harvey
and Jaeger 1993). However, in an extensive simulation
study, Kaiser and Maravall (1999) find that this problem is
very small in HP-filtered series. Third, it has been argued
that the results may be sensitive to the choice of smoothing
parameter λ(Kaiser and Maravall 2001, p. 116). Thus, we
conduct an extensive robustness analysis with other values
that have been suggested in the literature.
Quantifying the Extent of Cyclical Sensitivity
To quantify the extent of cyclical sensitivity in private-label
success, we derive the cyclical comovement elasticity. We
regress the cyclical component extracted from a country’s
private-label share, on the corresponding cyclical
component filtered from that country’s GDP per capita,
Although, strictly speaking, the business cycle is
defined in terms of the joint movement of several economic
series across multiple sectors (see www.nber.org/cycles.
html), fluctuations in aggregate output have been found to
be at the core of the business cycle, making it a good proxy
for the country’s economic activity as a whole (Stock and
Watson 1999, p. 15). This results in the following test
equation:
Because both cyclical components are expressed in per-
centage deviations, the resulting parameter βis an elasticity
estimate.3The sign and significance of βindicate whether
private-label share evolves pro-, counter-, or acyclical. Con-
versely, its magnitude reflects the extent to which fluctua-
tions in the general economy are attenuated or amplified in
() ,2 pls gdpc
t
ct
ct
=+βε
gdpct
c.
plst
c,

6 / Journal of Marketing, January 2007
4In line with prior studies in the economics literature (Bodman
2001; Cook 1999; Razzak 2001), we use a univariate detrending
procedure (i.e., the HP filter) to derive the cyclical component,
from which we subsequently assess the skewness. Although more
complicated multivariate procedures have also been used in prior
studies, Canova (1998) finds that the third moment (reflecting the
skewness) is not sensitive to the choice of detrending method, and
he cautions against the use of multivariate detrending procedures
that impose additional restrictions (e.g., common stochastic or
deterministic trends) that are not satisfied in the data. In unre-
ported analyses, we found that the various private-label series have
a unit root but are not cointegrated with the GDP-per-capita series.
private-label share. The HP filter may induce serial correla-
tion in the data (Engle 1974). To account for this, we can
add autoregressive error terms in estimating Equation 2, the
number of which we determine on the basis of information
criteria (Judge et al. 1988).
Examining the Presence of Deepness and
Steepness Asymmetry
To test whether private-label share displays cyclical (a)sym-
metries, we explore whether the filtered series is posi-
tively skewed or, in other words, skewed to the right.4In the
case of deepness asymmetry, we expect that the positive
deviations from the mean or trend during contraction peri-
ods are larger in absolute value than the negative deviations
during expansion periods, whereas the number of observa-
tions above the mean or trend is smaller than the number
below. Similarly, if a time series exhibits steepness asym-
metry, its first difference, representing the slope or rate of
change, will exhibit positive skewness. As such, increases in
the series’ growth corresponding to contractions should be
larger, but less frequent, than the more moderate decreases
during expansions.
To identify potential asymmetries in the evolution of
private-label success over contraction and expansion peri-
ods, we adopt the nonparametric triples test that Randles
and colleagues (1980) propose. In this test, all possible
triples (yi, yj, yk) of observations of a series ytare consid-
ered. For a series with T observations, such triples can
be identified. A triple is a right (left) triple when the middle
observation is closer to the smaller (larger) observation than
to the larger (smaller) observation. In a symmetric distribu-
tion, there are as many right as there are left triples. If there
are relatively more right triples, the underlying distribution
is skewed to the right. Formally, the triples test statistic is
given by
which can be shown to equal
() ˆ
ˆ,
ˆ[
32
η
σ
η
η
T
where = (number of right tripples) (number of left triples)]−
⎛
⎝
⎜⎞
⎠
⎟
⎡
⎣33
T
⎢⎢ ⎤
⎦
⎥
,
T
3
⎛
⎝
⎜⎞
⎠
⎟,
plst
c
and
We refer to Appendix B for the derivation of The
asymptotic distribution of the test statistic is standard nor-
mal, so we can use conventional critical values. In case of
deepness asymmetry, we expect a right-skewed distribution,
reflected in a positive value of the triples test statistic, for
As for steepness asymmetry, we derive a comparable
asymmetry statistic on the first difference of
Again, a right-skewed distribution will be reflected in a
positive test statistic.
Following previous research (see, e.g., Bodman 2001;
Cook 1999; Deleersnyder et al. 2004; Sichel 1993), we
apply the univariate asymmetry tests to (or for
steepness asymmetry). As an alternative approach, we could
consider using the forecasted values of Equation 2 (i.e.,
as input to the triples test. However, this procedure
suffers from two deficiencies. First, it ignores that is a
valid but imperfect indicator of the business cycle. As such,
may not capture all fluctuations in private-label
share that can be attributed to the business cycle. Second,
this procedure assumes that the third-moment (skewness)
properties of automatically pertain to the private-
label series as well because a linear transformation (rescal-
ing) of cannot alter the presence/absence of asymme-
tries (Sichel 1993). This assumption appears unduly
restrictive. Indeed, prior research has established that sev-
eral economic variables (e.g., unemployment, industrial
production, consumer price level) that are conceptually
linked to the business cycle have third-moment properties
that differ from those in GDP (Bodman 2001; Sichel 1993;
Verbrugge 1998).
Assessing the Presence of Asymmetric Growth
Univariate deepness and steepness asymmetries describe
the behavior of the cyclical fluctuations around the series’
overall trend or mean level and are studied on the filtered
series, that is, where the underlying trend (mean) has been
removed. As such, they cannot address whether this differ-
ential behavior also contributes to long-term private-label
success (e.g., by altering the series’ growth rate) or whether
they represent only temporary deviations from the under-
lying trend or mean that eventually cancel out one another
(Beaudry and Koop 1993). For example, if all customers
who switched to private labels during the contraction would
gradually return to buying national brands after the econ-
omy recovers, no long-term growth would be observed that
could be attributed to the contraction. As such, deepness
and steepness asymmetries can exist “with or without there
being asymmetry in persistence” (Beaudry and Koop 1993,
p. 151).
gdpct
c
gdpct
c
ˆ
βgdpct
c
gdpct
c
ˆ)βgdpct
c
Δplst
c
plst
c
Δplst
c.
plst
c,
plst
c.
ˆ.ςc
()
ˆ
51
3
33
3
2
1
3
ση
TTc
T
c
c
=⎛
⎝
⎜⎞
⎠
⎟
⎛
⎝
⎜⎞
⎠
⎟−
−
⎛
⎝
⎜⎞
⎠
⎟
=
∑ˆˆ .ςc
() ˆ{[ ( )432
1
η=⎛
⎝
⎜⎞
⎠
⎟+−
−
<<
∑
Tsign y y y
ij k
ijk
+++−++−sign y y y sign y y y
ik j jk i
()()]},223/

How Business Cycles Contribute to Private-Label Success / 7
5Through this operationalization, all values for exptand contrt
will be nonnegative, making the interpretation of their correspond-
ing coefficients in Equation 7 more straightforward.
6For the sake of simplicity, we do not explicitly model that
private-label share is bounded between zero and one, because it
could be argued that private-label share operates well away from
these boundaries. Nonetheless, we obtained the same substantive
findings when we replaced Δplstby Δ[(plst)/(1 – plst)] in Equation 7.
To assess formally whether cyclical shocks affect the
private labels’ long-term growth and to check whether this
effect differs for contraction and expansion periods, respec-
tively, we specify an asymmetric growth model. In line with
the work of Beaudry and Koop (1993), Cover (1992), and
Thoma (1994), we define two new variables that reflect the
general state of the economy at a certain point in time t:
Decreases (increases) in the cyclical component of GDP
per capita correspond to contractions (expansions). The
variable exptmeasures the magnitude of the expansion by
calculating how much the business cycle, reflected in the
filtered GDP-per-capita series, has increased relative to its
previous trough. Similarly, the variable contrtmeasures the
magnitude of a contraction by calculating how much the
business cycle has dropped relative to its previous peak
when the economy is on a downturn.5Private-label growth,
Δplst, is subsequently linked to current and lagged values of
exptand contrt. By assessing whether they have additional
explanatory power over lagged growth terms Δplst– j (j = 1,
…, J), we test whether the business cycle Granger (1969)
causes private-label growth:
with lag lengths J, K, and L determined on the basis of
information criteria (Judge et al. 1988).6Note that our
asymmetric growth model is specified in differences. Pre-
liminary unit-root tests can be used to determine whether
the series is indeed nonstationary (as implicitly assumed in
Equation 7). If not, no long-term effects can be detected.
However, because the presence of a unit root is a necessary
but not sufficient condition for (cyclical) shocks to have a
long-term impact (Dekimpe and Hanssens 1995), we still
need to test for the significance of in
Equation 7. Through the intercept α, we control for unob-
()Σll=+
0
Lϕ
Σk
Kk
=−
0ϕ
()7
10
ΔΔpls pls
tj
j
J
tj k
k
K
=+ +
+
=
−−
=
−
∑∑
αβ ϕcontrtk
ϕϕε
ll
l
+−
=
∑+expt
L
t
0
,
(6) expt
t
cc
gdpc prior trough in gdpc if gdpc=−(),Δtt
c
t
c
t
if gdpc
contr if gdp
>
=≤
⎧
⎨
⎪
⎩
⎪
=
0
00
0
,,
,
Δ
Δ
and
cc
prior peak in gdpc gdpc if gdpc
t
c
ct
ct
c
>
=−≤
0
(),Δ00
⎧
⎨
⎪
⎩
⎪.
7For an in-depth discussion on the role of the drift term in unit-
root models, see Juselius and Hargreaves (1992).
8We are indebted to Stephen Hoch for making the U.S. data
available to us.
served factors, such as the gradual increase in private-label
quality over time.7
By splitting the business cycle into two phases, we test
for asymmetries in the long-term response of private-label
growth to expansions and contractions. Thus, the effect of a
recession is not necessarily cancelled out in a subsequent
expansion. Unlike Equation 2, we no longer assess the link-
age between the cyclical movements in private-label share,
or the cyclical movements in the economy as a whole.
Instead, we now test whether changes in the latter con-
tribute to the private labels’ market-share growth (and, thus,
their long-term level). The sum of the parameters
associated with recessionary (expansionary) periods—that
is, = = —gives the combined
short-term impact on private-label growth. Because Equa-
tion 7 contains lags of the dependent variable, we can show
that the impact on the series’ long-term or steady-state
growth rate becomes, respectively (Franses 2005),
We can derive standard errors of these ratios using the well-
known delta method. Under the assumption that a contrac-
tion period stimulates private-label growth, we expect that
its impact is positive; thus, > 0. When the impact
of an expansion on private-label growth is negative—thus,
< 0)—the size of the expansions will determine the
extent to which the growth-stimulating effect of previous
contractions will be offset.
Data
Annual data on the aggregate value share of private labels in
CPGs are provided by ACNielsen for Belgium (1983–
2004), by Taylor Nelson Sofres for the United Kingdom
(1980–2003), by Selling Area Markets Inc and ACNielsen
for the United States (1971–2003),8and by GfK for West
Germany (1975–2002). The data span 20–30 years, which
is comparable in length to other studies on business-cycle
sensitivity (e.g., Cook 1999; Mills 2001).
Although private labels have grown their share in total
grocery sales in all four countries, there are some notable
differences as well, especially in terms of how this growth
was obtained (Kumar and Steenkamp 2006). In Germany, to
a large extent, private-label growth is driven by the success
of the hard-discount format, as implemented by chains such
as Aldi, Lidl, and Penny. For example, Aldi sells store
brands almost exclusively (>95%) and is one of the fastest-
growing retailers in the German grocery market. By 2002, it
Σll=+
0
Lϕ
Σk
Kk
=−
0ϕ
ϕ
β
ϕ
ϕ
β
k
k
K
j
j
JLT
L
j
j
−
=
=
−
+
=
=
∑
∑
∑
−
=
−
0
1
0
11
()and
l
l
11
JLT
∑
=+
().ϕ
ϕST
+)
(Σll=+
0
Lϕ
ϕST
−
Σk
Kk
=−
0ϕ
()ϕl
+
ϕk
−
plst
c,

8 / Journal of Marketing, January 2007
9Note that the pooling test is conducted on the filtered, rather
than the original, series. As such, we account for differences in the
long-term growth rates, which may distort the cyclical inferences.
had already captured close to 18% of total grocery retailing
sales in Germany. The absence of manufacturer brands from
most hard discounters’ assortments leads to an almost unob-
structed growth of private labels. In contrast, in the United
Kingdom, private-label growth is mainly fueled by retailers’
ability to offer an elaborate three-tier private-label range,
including value, standard, and premium store brands.
Finally, the United States and Belgium (the latter country
being similar in private-label landscape to other European
countries, such as France, Spain, and the Netherlands) are
located in between Germany and the United Kingdom.
Their growth in private-label sales can be attributed both to
the increasing success of discount operations and to the
growing importance of elaborate, quality-oriented, private-
label programs by mainstream retailers.
Data on real GDP per capita are used as a proxy for the
general economic activity in a particular country. Business-
cycle fluctuations across many sectors are reflected in
aggregate output, making the cyclical component of GDP
(per capita) a good indicator for the overall economic cycle
(Stock and Watson 1999). We obtained GDP per capita,
expressed in constant prices, from the United Nations Sta-
tistic Divisions for Belgium, the United Kingdom, and the
United States. For West Germany, we obtained these data
from the VGRDL (Volkswirtschaftliche Gesamtrechnungen
der Länder).
Empirical Results
The Extent of Cyclical Sensitivity
To control for serial correlation in the data, we added two
autoregressive terms to Equation 2. The Chow test on com-
mon slopes and autoregressive error terms revealed that we
could pool the four countries in our sample (F(9, 86) = 1.24,
p> .10).9We obtained a significant and negative comove-
ment elasticity (β= –.96, p< .01; R2= .42), indicating that
private-label share behaves countercyclically. It increases
during contractions and decreases during expansions. This
10We obtained the same substantive findings when working with
log-transformed private-label-share series instead of the original
series.
confirms conventional wisdom from a business-cycle per-
spective and is in line with the negative short-term relation-
ship that Hoch and Banerji (1993) find.
Cyclical Asymmetries
In addition to the extent of cyclical sensitivity in private-
label share, we consider the nature of cyclical ups and
downs, as reflected in the respective cyclical-asymmetry
statistics. The test results for the nonparametric triples test
appear in Table 2. To avoid potential distortion that may
arise from deriving asymmetry statistics after log-
transforming the series (DeFusco, Karels, and Muralidhar
1996), we based all inferences about these univariate asym-
metries on the cyclical component extracted from the origi-
nal (nontransformed) series.10
In line with our expectations, cyclical fluctuations in
private-label share evolve asymmetrically across expansion
and contraction phases. For all countries, the triples test
showed the expected positive sign for both deepness and
steepness asymmetry, even though the test was significant
in few individual cases. However, note that the power of
asymmetry tests in business-cycle research has been found
to be low, especially when working with annual data (Mills
2001; Razzak 2001). Psaradakis and Sola (2003) show that
whenever evidence of asymmetry in the cycle is established
using HP-filtered series, this can be taken as strong evi-
dence in favor of asymmetry.
Meta-analysis is a recommended way to address the
challenge posed by the low power of individual tests. It
offers a more powerful test for the presence of business-
cycle asymmetries than do the individual test statistics
because it combines evidence from all countries, allowing
for an overall significance test. Using the meta-analytic
method of adding weighted Zs (Rosenthal 1991), we found
support for asymmetric behavior in both the extent (deep-
ness) to which and the speed (steepness) with which
private-label share fluctuates over the business cycle (both
ps < .05). Thus, these results indicate that increases in
TABLE 2
Results on Deepness and Steepness Asymmetry
Belgium United Kingdom United States West Germany Meta-
(n = 22) (n = 24) (n = 33) (n = 28) Analysisa
Deepness 0.050 0.043*0.039 0.033*1.92**
(1.13) (1.16) (.93) (.76)
Steepness .042 .061* .034 .067* 2.13**
(.83) (1.31) (.74) (1.50)
*
p
< .10 (one-tailed test).
**
p
< .05 (one-tailed test).
aThe meta-analysis reports z-values that we obtained by the method of adding weighted Zs (Rosenthal 1991).
Notes: Z-statistics are in parentheses.

How Business Cycles Contribute to Private-Label Success / 9
11As a further test for our assertion that the observed asymme-
tries in plsc
tare driven by the business cycle, we applied the triples
test to the residuals of Equation 2. We found no evidence of any
remaining asymmetry in these residuals. Detailed test results are
available on request.
private-label share during economic contractions occur
more extensively and faster than declines in private-label
share in subsequent expansion phases.11
Do Economic Contractions Lead to Permanent
Private-Label Growth?
The question still remains whether the net result of a series
of cyclical ups and downs evokes long-term private-label
growth. To draw this conclusion, (1) the private-label-share
series should be evolving, and (2) long-term changes in
private-label share should be related to business-cycle
swings. Unit-root tests (Enders 1995) indicated that the
private-label series are indeed evolving. We could not reject
the null hypothesis of the presence of a unit root for any of
the four countries (all ps > .10).
To determine whether long-term changes in private-
label share can be attributed to business-cycle swings, we
estimated the asymmetric growth model of Equation 7.
Before estimation, a pooling test confirmed that common
slopes were allowed (F(9, 82) = 1.38, p> .10) but that
country-specific intercepts were needed (F(3, 91) = 2.40,
p< .10). We subsequently pooled the data across the four
countries, including country-specific intercepts. We
included one lag for private-label-share growth, but we did
not include any lags for real-GDP-per-capita expansion and
contraction (thus, in Equation 7, J = 1, and K, L = 0). We
derived standard errors using White’s robust covariance
estimation method (Greene 2003). The results appear in
Table 3.
Three of the country-specific intercepts were signifi-
cant—αBelgium = .022 (p< .10), αU.K. = .012 (p< .10), and
αWest Germany = .018 (p< .05)—indicating the presence of a
positive drift in private-label share. The country-specific
intercept for the United States was not significant (αU.S. =
–.003, p> .10). Furthermore, we found that previous-year
private-label evolution had a significant, positive impact on
private-label evolution this year (β1= .34, p< .05).
Most important, we found that economic expansions
and contractions affect private-label-share evolution to a
different degree. Contractions cause a substantial positive
impact on private-label growth that is not offset in subse-
quent expansions. This important finding applies to changes
in both short-term and long-term private-label growth.
Impact on short-term private-label growth. On the basis
of Equation 7, we found a negative, albeit insignificant,
impact on the trend or growth rate in private-label share
during economic expansions (ϕ+= –.18, p> .10), whereas a
contraction induced a significant, positive effect on short-
term private-label growth (ϕ–= .80, p< .05). Because we
work in log-log space, these parameters are elasticities.
Thus, when GDP per capita decreases 1% compared with
the peak just before the recession, private-label growth
increases .80% in the short run. Conversely, when GDP per
capita increases 1% compared with the trough just before
the expansion, private-label growth declines .18% in the
short run. In summary, a contraction causes a short-term
upward lift in the trend of private-label share, which is
(partly) maintained afterward, because this shift is not fully
recouped in the subsequent expansion.
Impact on long-term private-label growth. The impact
of a recession on the long-term or steady-state growth in
private-label share was again positive and significant =
1.22, p< .05), whereas during the expansion, the long-term
growth rate was not significantly affected = –.28, p>
(ϕLT
+
(ϕLT
−
TABLE 3
Results of the Asymmetric Growth Model
Estimate
p
Value
Country-Specific Drift αBelgium .022 .055a
αU.K. .012 .071a
αU.S. –.003 .666a
αWest Germany .018 .029a
Lagged Private-Label Growth β1.34 .023a
Business-Cycle Effects
Contraction
.80 .006b
1.22 .009b
Expansion
–.18 .223b
–.28 .216b
R2= .36, N = 98
aTwo-sided
p
value.
bOne-sided
p
value.
ϕLT
+
ϕST
+
ϕLT
−
ϕST
−

10 / Journal of Marketing, January 2007
.10). Thus, when, in recessionary times, GDP per capita
decreases 1% compared with the peak just before the reces-
sion, this causes a long-term upward lift in private-label
growth of 1.22%. Conversely, when GDP per capita
increases 1% compared with the trough just before that
expansion, this generates a long-term growth reduction of
only an insignificant .28%. Thus, a fraction of the con-
sumers who switched to store brands in a contraction stay
with their choice even after the economy improves.
Robustness Checks
Alternative smoothing parameters. We followed the rec-
ommendation of Baxter and King (1999) to set λequal to
10. However, other researchers have recommended smaller
values. Canova (1994) considers a value of λequal to 4,
Ravn and Uhlig (2001) suggest a value of 6.25, and Harvey
and Jaeger (1993) work with a value of 7. Different values
of λlead to the derivation of slightly different cyclical ele-
ments. We conducted extensive robustness analyses to
examine the sensitivity of our results to a wide range of λs,
namely, the three aforementioned smaller values (4, 6.25,
and 7) and the three larger ones (15, 20, and 30). The results
(see Table 4) show that our main findings are robust to the
choice of the smoothing value.
Alternative filtering approach. Different filtering proce-
dures extract slightly different types of business-cycle infor-
mation from the original series (Canova 1998). We reana-
lyzed our data, using another widely used detrending
procedure, namely, the band-pass filter of Baxter and King
(1999). The results based on brand-pass-filtered data con-
firmed the findings we obtained from HP-filtered data. The
comovement-elasticity results revealed the same general
pattern as before; there is combined evidence for counter-
cyclical movement in private-label share (β= –.70, p< .01),
and its value closely resembled the value of –.96 we found
previously. Moreover, the analyses confirmed the presence
of cyclical asymmetries, both in terms of deepness (p< .05)
and in terms of steepness (p< .10). Finally, for the asym-
metric growth model, we again obtained in both the short
and the long run a significant, positive effect on private-
label growth of a contraction = 1.05, p< .01; =
1.45, p< .05), which was not offset in the subsequent
expansion period = –.11, p> .10; = –.16, p> .10).
ϕLT
+
(ϕST
+
ϕLT
−
(ϕST
−
Error in the cyclical component. Any filtered series may
contain some noise. Different smoothing parameters in the
HP filter and alternative filters extract somewhat different
business-cycle information from the same series, with dif-
ferent amounts of noise (Kaiser and Maravall 2001). As we
reported previously, we replicated our findings with (1) six
alternative smoothing values and (2) the band-pass detrend-
ing procedure. This gives more confidence in our findings
and indicates that error in the extracted cyclical component
does not seriously affect our key insights.
Alternative specification of the asymmetric growth
model. In estimating the asymmetric growth model (Equa-
tion 7), we accounted for the magnitude of expansions and
contractions (see Equation 6). We assessed the robustness
of our findings to an alternative specification of the expan-
sion and contraction variables, focusing on the occurrence
of an expansion or contraction per se rather than on their
magnitude. We defined a dummy variable that equals one
when the economy is downturning and zero when the econ-
omy is expanding.
In the corresponding pooled asymmetric growth model
with country-specific fixed intercepts, we again found a
positive impact of a recession on private-label-share growth
= .020, p< .05). Because the country intercepts were
all positive and significant (except for the United States, in
which the intercept was insignificant at the 10% level), we
conclude that the additional positive effect found during the
contraction is not cancelled out by the subsequent expan-
sion. We also estimated a model without intercepts but with
an extra dummy variable to identify expansions. The
dummy for recessions was positive and significant, whereas
the dummy for expansions was positive but insignificant,
leading again to the conclusion that increases in private-
label share during recessions are not compensated for in the
subsequent expansion. In summary, these alternative model
specifications support our finding that the sequence of
aggregate business cycles contributes to long-term private-
label success.
German reunification in 1990. Germany experienced a
major change during the time span under investigation,
namely, its reunification in 1990. After the unification, West
German investments in production and equipment fell
because of major capital transfers to former East German
(ϕST
−
TABLE 4
Robustness Checks: Different Smoothing Parameters for the HP Filter
λ= 4 λ= 6.25 λ= 7 λ= 10 λ= 15 λ= 20 λ= 30
Comovement Elasticity –1.00** –.98** –.98** –.96** –.94** –.93** –.90**
Cyclical Asymmetriesa
Deepness 1.76* 1.88* 1.87* 1.92* 1.96* 1.91* 1.84*
Steepness 1.96* 2.07* 1.64* 2.13* 1.95* 1.91* 2.48**
Asymmetric Growth
β1.36* .46** .36* .34* .44* .43** .42**
.98** .79* .89** .80* .65* .56* .53*
–.15 –.17 –.11 –.18 –.23 –.25 –.25
*
p
< .05.
**
p
< .01.
aWe obtained the reported z-values of the meta-analysis by the method of adding weighted Zs (Rosenthal 1991).
ϕST
+
ϕST
−

How Business Cycles Contribute to Private-Label Success / 11
regions (Lütkepohl, Teräsvirta, and Wolters 1999). In line
with prior studies on the Germany economy (e.g., Hubrich
and Vlaar 2004), we control for a potential break in the
GDP-per-capita and the private-label-share series in 1990
when we derive fluctuations that can be attributed to the
business cycle (for more details, see Appendix A). The
parameters associated with the pulse dummy in Equation
A3 were insignificant and do not affect the cyclical compo-
nents derived from the West German private-label-share and
GDP-per-capita series, respectively. Thus, the German
reunification does not alter any of our prior findings.
Discussion
In this study, we examined the sensitivity of private-label
share to the aggregate business cycle for several key
private-label countries. In the beginning of the article, we
identified three research questions that this study attempted
to address. We summarize our findings in relation to these
questions. First, we find that private-label success indeed
behaves countercyclically. Several academics and practi-
tioners have theorized that there is an inverse relationship
between the state of the economy and private-label perfor-
mance, but with the exception of Hoch and Banerji (1993),
our study is the first to test this hypothesis rigorously. Our
analysis confirms this conventional wisdom from a
business-cycle perspective; that is, a country’s private-label
share increases when the economy is suffering and shrinks
when the economy is flourishing.
Second, private-label performance behaves differently
over expansion and contraction periods. This is a new find-
ing that has not been examined in previous research. We
find that business-cycle fluctuations induce asymmetries in
both the extent and the speed of up- versus downward
movements in private-label share.
Third, the aggregate business cycle contributes to long-
term private-label success. Not only does the sequence of
contractions and expansions induce temporary upward and
downward swings in private-label share, but we also find
that part of the share gain during contractions is permanent.
This is also an important new finding, which contradicts a
long-held opinion that private labels “are discarded once the
economy picks up again” (Ward et al. 2002, p. 962). Thus,
not only are consumers more prone to buy private labels
during economic downturns, but some of them also keep
buying these store-brand alternatives when bad economic
times are long over.
Managerial Implications
Economic recessions contribute to the prolonged upward
evolution in private-label share, leaving permanent scars on
national brands’ performance levels. Business-cycle fluctu-
ations are beyond the control of individual brand managers.
However, managers might mitigate the effect of an eco-
nomic downturn on their brands’ positions by engaging in
proactive marketing, a strategy in which the firm views the
recession as an opportunity and invests aggressively in mar-
keting activities during downturns. This strategy has been
found to result in improved performance not only during
contractions but also in subsequent expansion periods
(Hillier and Baxter 2001; Srinivasan, Rangaswamy, and
Lilien 2005).
Unfortunately, however, the behavior of most brand
managers appears to be anything but proactive. Indeed, their
behavior might actually exacerbate the negative effects of
an economic downturn on their brands’ positions. The key
driver behind this is managers’ short-term focus. After all,
most managers spend no longer than three years in the same
job (Kerin 2005) and thus may be prone to underweigh
long-term implications of their short-term decisions. In this
context, Milgrom and Roberts (1992, p. 471) observe that a
manager “will be tempted to put too much emphasis on
activities that boost short-term performance compared to
those whose benefits will be hidden … for a long period of
time.”
To protect short-term profits, managers want fast ways
to cut costs in times of slow demand. Two common ways to
achieve this are scaling back of innovation activity (Axar-
loglou 2003) and cutting advertising budgets (Picard 2001).
Table 5 offers additional evidence on advertising retrench-
ment during a recession for ten major national CPG brands
in the United Kingdom. Expenditure on these activities is
often discretionary, and cutting them directly affects the
bottom line, though their adverse effects may be revealed
only in the longer run.
The apparently widespread practice of reducing brand
support during bad economic times could reinforce the
impact of the business cycle in favor of store-brand alterna-
tives. However, is it true that innovation and advertising can
reduce the negative effect of an economic downturn for
brand manufacturers? If these instruments are not effective,
managers’ behaviors might be perfectly rational. Although
there is little hard evidence, there is indirect support that
innovation and advertising can indeed serve this role for
manufacturer brands. Several authors have speculated that
private labels are less successful in categories characterized
by intense innovation activity (Quelch and Harding 1996;
TABLE 5
Average Yearly Growth in Advertising
Expenditures for Several Top Brands in the United
Kingdom
Expansion Period Contraction Period
(1997–1998)a, b (2001–2002)c
Aquafresh +.03 –.25
Colgate +.11 –.18
Danone +2.80 –.01
Guinness +.04 –.15
Heinz +.68 –.25
Kellogg’s +.03 –.34
Kraft +.24 –.30
Nescafé +.21 –.51
Nestlé +.87 –.16
Persil +.11 –.11
aWe determined contraction and expansion years on the basis of
the HP-filtered GDP-per-capita data of the United Kingdom.
bAverage yearly growth between 1997 and 1998.
cAverage yearly growth between 2001 and 2002.
Notes: We obtained advertising-expenditure data from ACNielsen
Media Research U.K.

12 / Journal of Marketing, January 2007
12Examples of Reckitt Benckiser brands in the United States are
Lysol, Woolite, Calgon, Veet, Air Wick, Electrosol, and Spray ’n
Wash.
Steenkamp and Dekimpe 1997). Moreover, Hillier and Bax-
ter (2001) show that firms that increase their product devel-
opment spending during a recession perform better in terms
of profitability and market share during the subsequent
recovery. Although this does not prove that innovation is
effective as a barrier against private-label success in reces-
sions, at the very least, it suggests that reducing innovation
activity in economic downturns is not conducive to fight
private labels.
Because most CPG categories are mature, the positive
role of advertising may appear less obvious. However,
advertising is crucial to maintain product differentiation
with private labels, which is necessary to justify the price
difference (Hoch and Banerji 1993). Frankenberger and
Graham (2003) report that firms that increased their adver-
tising expenditure in a recessionary period created added
value that extended through the year following a recession.
However, the company is advised not to rely on advertising
in isolation but rather to use it in combination with the
introduction of new products. Lodish and colleagues (1995)
report that the advertising elasticity for new products is five
times higher than it is for established products. Similarly,
Steenkamp and Gielens (2003) find that new products that
were heavily advertised achieved a trial rate that was 72%
higher than the trial rate of new products that did not
receive much advertising support.
Taken together, these findings imply that proactive mar-
keting involving new-product introductions supported by
heavy advertising may be a powerful strategy to mitigate
the effect of a recession on the position of manufacturer
brands. Proactive marketing in times of recession may offer
the key for national-brand manufacturers to prevent con-
sumers from (permanently) switching to private-label offer-
ings. One company that has successfully followed this route
is U.K.-based Reckitt Benckiser.12 Despite difficult eco-
nomic times in many of its markets, it has managed real
top-line growth (and doubled profitability) through hefty
product innovation—40% of sales are from products
launched in the last three years—and by investing much
more heavily in advertising than most of its competitors.
Notably, another “discretionary” marketing activity, mar-
keting research, plays a key role in this process. Product
development is based on ideas generated by marketing
research and are tested with consumers before launch
(Financial Times 2005).
Contrary to most manufacturers, anecdotal evidence
suggests that retailers actually exploit the benefits of proac-
tive marketing because they tend to increase their private-
label support in recessions. For example, the British super-
market chain J. Sainsbury responded to difficult trading
times in the beginning of the 1990s by increasing its pro-
motional emphasis on own-label products with price reduc-
tions on some 300 of these items (Walters 1994). As another
illustration, more than one-third of all new-product
launches in Germany during the economically depressed
period of 2001–2003 were private-label brands (Kumar and
Steenkamp 2006). Through such countercyclical private-
label support activities, retailers may significantly con-
tribute to long-term private-label performance. Thus,
although the impact of the aggregate business cycle may
seem uncontrollable, a proactive marketing strategy by
manufacturers and retailers may either mitigate or accentu-
ate the observed dependency of private-label success on
general economic conditions.
Limitations and Directions for Further Research
Our study has several limitations that offer interesting
avenues for further research. Future studies should examine
the interplay between consumer actions and manufacturer
or retailer behavior to assess the relative contribution to the
increased popularity of private labels during contractions. If
this phenomenon is driven mainly by consumer behavior in
response to changes in the economic climate, this does not
bode well for national-brand manufacturers, because part of
private-label success is beyond their control.
Conversely, if manufacturer behavior is an important
factor in explaining private-label success over the business
cycle, what can managers do to reverse this? Can increased
new-product development mitigate the effect of a recession
for manufacturers? Can advertising staunch the loss of mar-
ket share in an economic downturn, and if so, what adver-
tising medium and content are most effective in recession-
ary times? Is it better to spend the money on price
promotions, despite their often modest pass-throughs?
Which marketing instrument is most effective and cost effi-
cient? Can synergies be obtained by using several instru-
ments simultaneously? Finally, why does available evi-
dence, admittedly anecdotal, indicate that managers
apparently behave in ways that hurt the long-term prospects
of the company? Do they not know this, or do they unrea-
sonably discount future profits? How can performance
evaluation criteria and career paths be adapted to prevent
unreasonable discounting?
On the part of the retailer, it is fruitful to consider
whether actions during recessionary periods really reinforce
the cyclical sensitivity of private labels and what kind of
actions are most successful to strengthen the boost in
private-label sales. Retailers should also give closer consid-
eration to the profitability of private labels. Even if their
behaviors increase private-label success in recessions, are
they better off financially after the relevant criterion—
annual dollar contribution per square foot of shelf space—is
considered? From a more technical point of view, further
research could develop an integrated (potentially nonlinear)
model that enables the investigation of the four components
of the business cycle in a single step rather than our current
multistep procedure.
In summary, our study formally examines how the
popularity of private-label alternatives is related to aggre-
gate business-cycle fluctuations in a CPG setting and offers
evidence on the much-debated and speculative idea that pri-
vate labels are only for recessionary times. We propose a
framework that helps understand and predict the observed
net effect on private-label performance. However, more
research is called for to identify which part of the observed
negative effect of a recession on national-brand sales can be
offset by manufacturers and which factors underlie the

How Business Cycles Contribute to Private-Label Success / 13
observed patterns. We hope that our article provides an
impetus to this much-needed research effort.
Appendix A
The HP filter can be rationalized as the optimal estimator of
the trend in the following structural model (Harvey and
Jaeger 1993):
where the trend component is modeled in general state-
space form as
where = where λis the smoothing parameter or
penalty that determines the degree of smoothing. In line
with the work of Boone and Hall (1999), we normalize
equal to 1. By imposing a value for λequal to the smooth-
ing parameter in Equation 1 and setting equal to 0 in
Equation A2, this general structural time-series model pro-
duces a trend and a cyclical component that are identical to
those obtained using the well-known HP filter. The conver-
sion of the HP filter into Equation A2 and A3 enables us to
adopt the Kalman filter to calculate maximum likelihood
estimates of the trend component.
To control for a potential break in the series after a cer-
tain point in time (τ), we can extend the HP filter formu-
lated in Equations A1 and A2 with two pulse dummies that
can account for a potential change in the level and the trend
of the series. The resulting state-space form of the trend,
in Equation A2 then becomes
() ,
,
Ayy D
D
tt t
tt tt
3111
12
ll
=++
=+ +
−−
−
βγ
ββ γ εε
τ
τ~~,,N0 1
λ
⎛
⎝
⎜⎞
⎠
⎟
yt
l,
ση
2
σyc
2
1/ ,λ
σσ
ε
22
/yc
() , ~(, )Ayy N
tt t t t
tt
20
11 2
1
ll
=++
=+
−−
−
βη η σ
ββ ε
η
ttt
N, εσ
ε
~(, ),02
yt
l
() , ~(, ),AyyyyN
ttt
ct
cyc
10
2
=+
lσ
For a similar extension of the business-cycle filter with
break dummies, see Boone and Hall (1999).
Thus, when we filter out the fluctuations due to the
business cycle, the state-space formulation (Equation A3)
enables us to control for a potential break in a series (e.g.,
the German reunification in 1990).
Appendix B
We express the quantities and in terms of proba-
bilities and computed them as
fXXX
sign X
ijk
i
(,, )
(
*=
+
1
3
XXX
sign X X X
sign X X X
jk
ik j
jk i
−
++−
++−
⎡
⎣
⎢2
2
2
)
()
()
⎢⎢
⎢
⎢
⎢
⎤
⎦
⎥
⎥
⎥
⎥
⎥
=
;
() ˆ
B2 2
ζ11
2
22
2
TfXX
fX
jk
jk
j
(
)
−
∑∑
<
(ˆ(, )ˆ),
ˆ(
*
*
where
η
,, ) ( , , )
*;and
(
XTfX X X
kijk
i
ijk
ik
=−=
≠≠
≠
∑
1
21
BB3) ˆˆ.ζη
32
1
9
=−
() ˆ[ˆ*
BTf
i
T
11
11
1
ζ=
=
∑(()ˆ],
ˆ() (, ,
**
X
fX TfX X
i
iij
−
=−
(
)
η2
1
1
1
2
where andXk
jk
ijk
),
∑∑
<
≠≠
ˆ
ζ3
ˆ,ζ2
ˆ,ζ1
.D if t
if t
0
1
where τ
τ
τ
=≠
==
⎧
⎨
⎩
ACNielsen (2003), “The Power of Private Label: A Review of
Growth Trends Around the World,” (accessed August 28,
2006), [available at http://www2.acnielsen.com/reports/
documents/2003_privatelabel.pdf].
Ailawadi, Kusum L. and Kevin Lane Keller (2004), “Understand
Retail Branding: Conceptual Insights and Research Priorities,”
Journal of Retailing, 80 (Winter), 331–42.
———, Scott Neslin, and Karen Gedenk (2001), “Pursuing the
Value-Conscious Consumer: Store Brands Versus National
Brand Promotions,” Journal of Marketing, 65 (January),
71–89.
Allenby, Greg M., Lichung Jen, and Robert P. Leone (1996),
“Economic Trends and Being Trendy: The Influence of Con-
sumer Confidence on Retail Fashion Sales,” Journal of Busi-
ness and Economic Statistics, 14 (January), 103–111.
Ang, Swee Hoon, Siew Meng Leong, and Philip Kotler (2000),
“The Asian Apocalypse: Crisis Marketing for Consumers and
Business,” Long Range Planning, 33 (February), 97–119.
Axarloglou, Kostas (2003), “The Cyclicality of New Product
Introductions,” Journal of Business, 76 (January), 29–48.
Backus, David K. and Patrick K. Kehoe (1992), “International
Evidence on the Historical Properties of Business Cycles,”
American Economic Review, 82 (September), 864–88.
Ball, Laurence and Gregory N. Mankiw (1994), “Asymmetric
Price Adjustment and Economic Fluctuations,” Economic Jour-
nal, 104 (March), 247–61.
Baxter, Marianne and Robert G. King (1999), “Measuring Busi-
ness Cycles: Approximate Band-Pass Filters for Economic
Time Series,” Review of Economics and Statistics, 81 (Novem-
ber), 575–93.
Beaudry, Paul and Gary Koop (1993), “Do Recessions Perma-
nently Change Output?” Journal of Monetary Economics, 31
(April), 149–63.
Bodman, Philip M. (2001), “Steepness and Deepness in the Aus-
tralian Macroeconomy,” Applied Economics, 33 (February),
375–82.
Boone, Laurence and Stephen G. Hall (1999), “Stylized Facts of
the Business Cycle Revisited: A Structural Modeling
Approach,” International Journal of Financial Economics, 4
(July), 253–68.
REFERENCES

14 / Journal of Marketing, January 2007
Bowman, David, Deborah Minehart, and Matthew Rabin (1999),
“Loss Aversion in Consumption-Savings Model,” Journal of
Economic Behavior & Organization, 38 (February), 155–78.
Canova, Fabio (1994), “Detrending and Turning-Points,” Euro-
pean Economic Review, 38 (April), 614–23.
——— (1998), “Detrending and Business Cycle Facts,” Journal
of Monetary Economics, 41 (May), 475–512.
Carroll, Christopher D. (1992), “The Buffer Stock Theory of Sav-
ing: Some Macroeconomic Evidence,” Brookings Papers on
Economic Activity, 2, 61–155.
Christiano, Lawrence J. and Terry J. Fitzgerald (1998), “The Busi-
ness Cycle: It’s Still a Puzzle,” Economic Perspectives, 22 (4),
56–83.
Cook, Steven (1999), “Cyclicality and Durability: Evidence from
U.S. Consumers’ Expenditure,” Journal of Applied Economics,
11 (November), 299–310.
Cover, James P. (1992), “Asymmetric Effects of Positive and
Negative Money-Supply Shocks,” Quarterly Journal of Eco-
nomics, 4 (November), 1261–82.
Davies, Keri (2000), “The Asian Economic Recession and Retail
Change: The Implication for Retailer Strategies in Asia,” Inter-
national Review of Retail, Distribution and Consumer
Research, 10 (October), 335–53.
DeFusco, Richard A., Gordon V. Karels, and Krishnamurty
Muralidhar (1996), “Skewness Persistence in U.S. Common
Stock Returns: Results from Bootstrapping Tests,” Journal of
Business Finance & Accounting, 23 (October), 1183–95.
Dekimpe, Marnik G. and Dominique M. Hanssens (1995), “The
Persistence of Marketing Effects on Sales,” Marketing Science,
14 (Winter), 1–21.
Deleersnyder, Barbara, Marnik G. Dekimpe, Miklos Sarvary, and
Philip M. Parker (2004), “Weathering Tight Economic Times:
The Sales Evolution of Consumer Durables over the Business-
Cycle,” Quantitative Marketing and Economics, 2 (December),
347–83.
Deloitte & Touche (2003), “The Growing Importance of Private
Label Products,” (Winter), (accessed August 28, 2006), [avail-
able at http://www.deloitte.com/dtt/cda/doc/content/Food_For_
Thought_Winter_2003_04.pdf].
Enders, Walter (1995), Applied Econometric Time Series. New
York: John Wiley & Sons.
Engle, Robert F. (1974), “Band Spectrum Regression,” Interna-
tional Economic Review, 15 (February), 1–11.
Estelami, Hooman, Donald R. Lehmann, and Alfred C. Holden
(2001), “Macro-Economic Determinants of Consumer Price
Knowledge: A Meta-Analysis of Four Decades of Research,”
International Journal of Research in Marketing, 18 (Decem-
ber), 341–55.
Fatás, Antonio (2002), “The Effects of Business Cycles on
Growth,” in Economic Growth: Sources, Trends and Cycles,
Norman Loayza and Raimundo Soto, eds. Santiago, Chile:
Central Bank of Chile, 191–219.
Financial Times (2005), “‘Craziness’ Makes Merger Work for
Reckitt Benckiser,” (February 9), 21.
Frankenberger, Krisina D. and Roger C. Graham (2003), “Should
Firms Increase Advertising Expenditures During Recessions?”
Report No. 03-115, Marketing Science Institute, 65–85.
Franses, Philip Hans (2005), “On the Use of Econometric Models
for Policy Simulation in Marketing,” Journal of Marketing
Research, 42 (February), 4–14.
Gale, Douglas (1996), “Delay and Cycles,” Review of Economic
Studies, 63 (April), 169–98.
Granger, Clive W.J. (1969), “Investigating Causal Relations by
Econometric Models and Cross-Spectral Methods,” Economet-
rica, 37 (3), 424–38.
——— and Michio Hatanaka (1964), Spectral Analysis of Eco-
nomic Time Series. Princeton, NJ: Princeton Press.
Greene, William H. (2003), Econometric Analysis, 5th ed. Upper
Saddle River, NJ: Prentice Hall.
Harvey, Andrew C. and A. Jaeger (1993), “Detrending, Stylized
Facts and the Business Cycle,” Journal of Applied Economet-
rics, 8 (July–September), 231–47.
Hillier, Tony and Marilyn Baxter (2001), “How to Prevent a Hang-
over: Brave Strategies for Difficult Times,” Market Leader,
(Autumn), 1–5.
Hoch, Stephen J. (1996), “How Should National Brands Think
About Private Labels?” Sloan Management Review, 37 (Win-
ter), 89–102.
——— and Shumeet Banerji (1993), “When Do Private Labels
Succeed?” Sloan Management Review, 34 (Summer), 57–67.
———, Alan L. Montgomery, and Young-Hoon Park (2002),
“Why Private Labels Show Long-Term Market Share Evolu-
tion,” Working Paper No. 2000-E18, Graduate School of Indus-
trial Administration, Carnegie Mellon University.
Hodrick, Robert J. and Edward C. Prescott (1997), “Postwar U.S.
Business Cycles: An Empirical Investigation,” Journal of
Money, Credit and Banking, 29 (February), 1–16.
Holly, Sean and Mike Stannett (1995), “Are There Asymmetries in
UK Consumption? A Time Series Analysis,” Applied Econom-
ics, 27 (August), 767–72.
Hubrich, Kristin and Peter Vlaar (2004), “Monetary Transmission
in Germany: Lessons for the Euro Area,” Empirical Econom-
ics, 29 (May), 383–414.
Judge, George G., R. Carter Hill, William E. Griffiths, Helmut
Lütkepohl, and Tsoung-Chao Lee (1988), Introduction to the
Theory and Practice of Econometrics. New York: John Wiley
& Sons.
Juselius, Katerina and Colin P. Hargreaves (1992), “Long-Run
Relations in Australian Monetary Data,” in Macroeconomic
Modeling of the Long Run, Colin P. Hargreaves, ed. Brookfield,
VT: Edward Elgar Publishing, 249–85.
Kaiser, Regina and Agustin Maravall (1999), “Estimation of the
Business Cycle: A Modified Hodrick-Prescott Filter,” Spanish
Economic Review, 1 (2), 175–206.
——— and ——— (2001), Measuring Business Cycles in Eco-
nomic Time Series. New York: Springer-Verlag.
Katona, George (1975), Psychological Economics. New York:
Elsevier Scientific Publishing.
Kerin, Roger A. (2005), “Strategic Marketing and the CMO,”
invited essay in “Marketing Renaissance: Opportunities and
Imperatives for Improving Marketing Thought, Practice, and
Infrastructure,” Journal of Marketing, 69 (October), 1–25.
Kumar, Nirmalya and Jan-Benedict E.M. Steenkamp (2006), Pri-
vate Label Revolution. Boston: Harvard Business School Press.
Lodish, Leonard M., Magid Abraham, Stuart Kalmenson, Jeanne
Livelsberger, Beth Lubetkin, Bruce Richardson, and Mary
Ellen Stevens (1995), “How T.V. Advertising Works: A Meta-
Analysis of 389 Real-World Split Cable T.V. Advertising
Experiments,” Journal of Marketing Research, 32 (May),
125–40.
Lütkepohl, Helmut, Timo Teräsvirta, and Jürgen Wolters (1999),
“Investigating Stability and Linearity of a German M1 Money
Demand Function,” Journal of Applied Econometrics,14
(September–October), 511–25.
M+M Planet Retail (2004), “Private Label Trends Worldwide,
2004,” (accessed November 11, 2004), [available at http://
www.planetretail.net].
Mascarenhas, Briance and David A. Aaker (1989), “Strategy over
the Business Cycle,” Strategic Management Journal,10
(May–June), 199–210.
Milgrom, Paul and John Roberts (1992), Economics, Organiza-
tion, and Management. Englewood Cliffs, NJ: Prentice Hall.
Mills, Terence C. (2001), “Business Cycle Asymmetry and Dura-
tion Dependence: An International Perspective,” Journal of
Applied Statistics, 28 (August), 713–24.

How Business Cycles Contribute to Private-Label Success / 15
Nandan, Shiva and Roger Dickinson (1994), “Private Brands:
Major Brand Perspective,” Journal of Consumer Marketing,11
(December), 18–28.
Nooteboom, Bart, Hans Berger, and Niels G. Noorderhaven
(1997), “Effect of Trust and Governance on Relational Risk,”
Academy of Management Journal, 40 (April), 308–338.
Picard, Robert G. (2001), “Effects of Recessions on Advertising
Expenditures: An Exploratory Study of Economic Downturns
in Nine Developed Nations,” Journal of Media Economics, 14
(1), 1–14.
Psaradakis, Zacharias and Martin Sola (2003), “On Detrending
and Cyclical Asymmetry,” Journal of Applied Econometrics,
18 (May–June), 271–89.
Quelch, John A. and David Harding (1996), “Brands Versus Pri-
vate Labels: Fighting to Win,” Harvard Business Review,37
(Winter), 99–109.
Randles, Ronald H., Michael A. Fligner, George E. Policello, and
Douglas A. Wolfe (1980), “An Asymptotically Distribution-
Free Test for Symmetry Versus Asymmetry,” Journal of the
American Statistical Association, 75 (March), 168–72.
Ravn, Morten and Harald Uhlig (2001), “On Adjusting the HP-
Filter for the Frequency of Observations,” Discussion Paper
Series No. 2858, 1–19, Centre for Economic Policy Research.
Razzak, Weshah A. (2001), “Business Cycle Asymmetries: Inter-
national Evidence,” Review of Economic Dynamics, 4 (Janu-
ary), 230–43.
Reeves, Jonathan J., Conrad A. Blyth, Christopher M. Triggs, and
John P. Small (2000), “The Hodrick-Prescott Filter, a General-
ization, and a New Procedure for Extracting an Empirical
Cycle from a Series,” Studies in Nonlinear Dynamics and
Econometrics, 4 (April), 1–16.
Richardson, Paul S., Alan S. Dick, and Arun K. Jain (1994),
“Extrinsic and Intrinsic Cue Effects on Perceptions of Store
Brand Quality,” Journal of Marketing, 58 (October), 28–36.
Rosenthal, Robert (1991), Meta-Analytic Procedures for Social
Research. Newbury Park, CA: Sage Publications.
Shama, Avraham (1981), “Coping with Stagflation: Voluntary
Simplicity,” Journal of Marketing, 45 (Summer), 120–34.
Sichel, Daniel E. (1993), “Business Cycle Asymmetry: A Deeper
Look,” Economic Inquiry, 31 (April), 224–36.
Srinivasan, Raji, Arvind Rangaswamy, and Gary Lilien (2005),
“Turning Adversity into Advantage: Does Proactive Marketing
During a Recession Pay Off?” International Journal of
Research in Marketing, 22 (June), 109–125.
Steenkamp, Jan-Benedict E.M. (1989), Product Quality: An Inves-
tigation into the Concept and How It Is Perceived by Con-
sumers. Assen, the Netherlands: Van Gorcum/Hendron.
——— (2005), “Moving Out of the U.S. Silo: A Call to Arms for
Conducting International Marketing Research,” invited essay in
“Marketing Renaissance: Opportunities and Imperatives for
Improving Marketing Thought, Practice, and Infrastructure,”
Journal of Marketing, 69 (October), 1–25.
——— and Marnik G. Dekimpe (1997), “The Increasing Power of
Store Brands: Building Loyalty and Market Share,” Long
Range Planning, 30 (December), 917–30.
——— and Katrijn Gielens (2003), “Consumer and Market Dri-
vers of the Trial Rate of New Consumer Products,” Journal of
Consumer Research, 30 (December), 368–84.
Stock, James H. and Mark W. Watson (1999), “Business Cycle
Fluctuations in U.S. Macroeconomic Time Series,” in Hand-
book of Macroeconomics, John Taylor and Michael Woodford,
eds. Amsterdam: Elsevier Science, 3–64.
Thoma, Mark A. (1994), “Subsample Instability and Asymmetries
in Money-Income Causality,” Journal of Econometrics,64
(September–October), 279–306.
Verbrugge, Randal J. (1998), “A Cross-Country Investigation of
Macroeconomic Asymmetries,” Economic Working Paper
Archive, Macroeconomics, No. 9809017.
Wakefield, Kirk. L. and Jeffrey J. Inman (1993), “Who Are the
Price Vigilantes? An Investigation of Differentiating Character-
istics Influencing Price Information Processing,” Journal of
Retailing, 69 (Autumn), 216–33.
The Wall Street Journal (1993), “Bargain Hunters Bag More Store
Brands,” (April 15), B1–B10.
Walters, David (1994), “The Impact of the Recession on Retailing
Management Decisions and Performance,” International Jour-
nal of Retail & Distribution Management, 22 (4), 20–31.
Ward, Michael B., Jay P. Shimshack, Jeffrey M. Perloff, and J.
Michael Harris (2002), “Effect of the Private-Label Invasion in
Food Industries,” American Journal of Agricultural Economics,
84 (November), 961–73.