Content uploaded by Giuseppe Di Vita
Author content
All content in this area was uploaded by Giuseppe Di Vita on Jun 20, 2016
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=wifa20
Download by: [Newcastle University] Date: 11 May 2016, At: 12:25
Journal of International Food & Agribusiness Marketing
ISSN: 0897-4438 (Print) 1528-6983 (Online) Journal homepage: http://www.tandfonline.com/loi/wifa20
When Consumers and Products Come From the
Same Place: Preferences and WTP for Geographical
Indication Differ Across Regional Identity Groups
Luca Panzone, Giuseppe Di Vita, Stefania Borla & Mario D’Amico
To cite this article: Luca Panzone, Giuseppe Di Vita, Stefania Borla & Mario D’Amico (2016):
When Consumers and Products Come From the Same Place: Preferences and WTP for
Geographical Indication Differ Across Regional Identity Groups, Journal of International Food &
Agribusiness Marketing
To link to this article: http://dx.doi.org/10.1080/08974438.2016.1145611
Published online: 11 May 2016.
Submit your article to this journal
View related articles
View Crossmark data
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING
http://dx.doi.org/10.1080/08974438.2016.1145611
When Consumers and Products Come From the Same
Place: Preferences and WTP for Geographical Indication
Differ Across Regional Identity Groups
Luca Panzonea , Giuseppe Di Vitab, Stefania Borlac, and Mario D’Amicob
aNewcastle University, Newcastle upon Tyne, UK; bUniversity of Catania, Catania, Italy; cWarwick
University, Warwick, UK
ABSTRACT
This article contributes to the existing literature on geographical
indications by observing consumers’ stated preference for
extra-virgin olive oil in two groups differing in their regional
identity. In particular, consumers from two groups were asked
to rank products in a contingent ranking survey. One group
(“insiders,” Sicilian consumers) shared origin with a good
product (Sicilian oil); the other group (“outsiders,” Rome and
Milan) presented “no association” consumer-product. Results
indicate that insiders are willing to pay more for goods
originating from the region they identify with compared with
a region associated with outsiders. Identity seems to give a bias
by which a local product is not necessarily perceived as superior
in absolute terms, but in relative terms: outside products are
never considered better than inside options but are either
inferior or equal in perceived value.
KEYWORDS
Contingent ranking;
extra-virgin olive oil;
geographical indication;
regional identity
Geographical indications (GIs) are an important component of the agricultural
and food economy in European Union (EU) countries. The unique
combination of human, biological, and historical resources that are embedded
in traditional food products from specific locations makes these products
unique and highly valuable to consumers (Rangnekar, 2004). To clearly ident-
ify the link with their place of origin, these products generally bear the name of
the location (country, region, or even locality) where the good is produced
(e.g., Bordeaux wines), and use regulated GI labels.
1
Earlier research has com-
prehensively explored the importance of GI labels (Josling, 2006; Moschini,
Menapace, & Pick, 2008) with a primary focus on adverse selection and the
welfare consequences of the imposition of quality standards (e.g., Marette,
Crespi, & Schiavina, 1999; Zago & Pick, 2002). From a policy perspective,
the importance of GIs is reflected in the incentive they provide in the develop-
ment of individual (Kreps & Wilson, 1982; Shapiro, 1983) as well as collective
reputation systems (Tirole, 1996; Winfree & McCluskey, 2005). GIs are parti-
cularly important for food products, which require specific local knowledge of
CONTACT Luca Panzone luca.panzone@newcastle.ac.uk School of Agriculture, Food and Rural
Development, Agriculture Building, Newcastle University, Newcastle upon Tyne NE1 7RU, UK.
© 2016 Taylor & Francis Group, LLC
Downloaded by [Newcastle University] at 12:25 11 May 2016
applied food technology, such as wine and fresh products (e.g., Scarpa,
Philippidis, & Spalatro, 2005; Scarpa, Thiene, & Marangon, 2007; Stanziani,
2004).
Research has comprehensively explored the interest of consumers for
regional foods and their attitudes and perception associated to the place of
origin of food products. The main focus of this research has been the implica-
tions of product differentiation, the demand for information on origin and its
impact on habits, and the anthropology of food consumption (e.g., Brunori,
2007; Guerrero et al., 2009; Iaccarino, Di Monaco, Mincione, Cavella, & Masi,
2006; Kuznesof, Tregear, & Moxey, 1997). Yet, an unexplored aspect of
consumer behavior in the current literature on GI is the relation between con-
sumers and location of origin. In fact, consumers use the products they pur-
chase to define and communicate their personal and social identity (Hogg &
Williams, 2000; Tajfel, 1979), and being part of a defined social group plays
an important role in the well-being of consumers. The products that consu-
mers choose then help them signal their group membership. Social identity
can then conceivably be important in the consumption of GI-labeled products
because these goods are sold with a geographical signal that can be linked to
group membership. Specifically, a GI on the label of a product allows consu-
mers to identify themselves as insiders (i.e., sharing origin with the goods) or
as outsiders (i.e., not sharing the origin with the goods) (see Akerlof & Kranton,
2000, for more general definitions of insiders and outsiders). In this article,
social identity is defined in its regional aspects, and it reflects a broad corre-
spondence between the origin of consumers and a related agro-food product.
Because GIs clearly and specifically inform on the geographical origin of
goods, they can activate feelings of self-identity in those consumers who share
the same origin of the goods.
2
The choice of a product originating from the
same locality of the consumer can be seen as a social standard of choice: by
giving priority to the local good, consumers not only purchase something they
surely like (Loureiro & Umberger, 2007; van der Lans, Van Ittersum, De Cicco,
& Loseby, 2001) but also sustain the socioeconomic system in which they live
(Tregear, Arfini, Belletti, & Marescotti, 2007). At the same time, insiders are
likely to prefer “inside” options because of exposure to local food from early
age (Birch & Marlin, 1982). This association can lead to an ingroup bias
(Ahmed, 2007; Giannakakis & Fritsche, 2011), particularly a home-country-
of-origin (H-COO) bias (Schooler, 1965). The positive utility from both taste
preference for an “inside” food and membership in a social group (Chen & Li,
2009; Klor & Shayo, 2010; Leonardelli, Pickett, & Brewer, 2010; Tajfel, 1974)
expectedly results in a significant willingness to pay (WTP) for one’s own
GIs (Caswell & Mojduszka, 1996; Loureiro & Umberger, 2007; van der Lans
et al., 2001).
Knowledge of the origin can also lead to cognitive processes that negatively
affect the choice of a local food. The result of this mental process would be an
2 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
out-of-home country-of-origin bias (OOH-COO). For instance, consumers
might associate better taste and/or reputation with a foreign good, or they
might want to show a positive predisposition to origin different from their
own (particularly if preferences are stated) to signal xenophilia (Perlmutter,
1954). More generally, consumers could perceive a higher level of affinity with
the good “from outside” on grounds that differ from preference for the local
food (Oberecker, Riefler, & Diamantopoulos, 2008). On the other hand, out-
siders would be expected to be indifferent to the geographical origin of goods,
which would be purely valued for its ability to satisfy taste preferences.
This article is a first attempt to explore explicitly differences in consump-
tion of GI between outside and inside groups. In fact, while differences in
behavior in different broad geographical groups have been examined (Scarpa
& Del Giudice, 2004), there is no clear intuition on how consumers decide
when they can directly associate their origin to the origin of product. For
instance, it is unclear whether H-COO bias dominates, is dominated, or coex-
ists with OOH-COO bias. Hence, the main objective of this article is to extend
the current understanding of consumer behavior in the choice of food by
assessing preferences for origin in food for insiders and outsiders separately.
The empirical analysis consists of four groups of consumers, two of which
consist of insiders, ranking a set of nine olive oil products. Three products
come from their same region of insiders (Sicily). The other two groups consist
of outsiders and have no direct association with the geographical origin of
any products. The GI signal is expected to activate a sense of belonging to
the regional group, causing differences in rankings to the advantage of
own-regional products.
Earlier research supports the intuition that proximity to the origin of food
can increase WTP for food (e.g., Hu, Batte, Woods, & Ernst, 2012), as well as
for olive oil (Scarpa & Del Giudice, 2004). However, this literature did not
allow consumers to directly identify with a specific GI. Previous research also
highlights that consumers have positive WTP for GIs (see, e.g., Rangnekar,
2004; Van der Lans et al., 2001), without, however, considering whether
and how WTP differs between inside and outside groups. The present article
represents an attempt to close these gaps: in the empirical exercise, prefer-
ences for region of origin in the choice of olive oil are collected separately
for two groups of insiders and two groups of outsiders. The olive oil market
presents a lower level of differentiation compared with other markets with GIs
(e.g., cheese, wine), making the experimental fieldwork simpler to implement.
Respondents were not aware of the rationale of the data collection process,
and insiders could identify products originating from their same region.
Results support the intuition to the extent that own-regional products are
those valued the most in both groups of insiders, and the same region is con-
sistently the lowest in outsiders. Preferences for other attributes (protected
designation of origin [PDO] labels and organic) are of comparable magnitude.
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 3
Downloaded by [Newcastle University] at 12:25 11 May 2016
The remainder of the report is as follows. The next section presents the
econometric model used in the article. Then the data collection process is
described, followed by the results and a discussion of the findings, and then
a conclusion is given.
Econometric model
Contextualization of the economic decision-making process
The study starts by defining a simple model of consumer behavior. Imagine
two markets g ¼I (for insiders) and O (for outsiders) differing in their
geographical location. Each market is composed by N goods i that differ in
their unobservable (to the econometrician) tastes n
i
, other observable
attributes X
i
and by their GI GI
i
. Goods originate from only two locations
k ¼s, −s, both recognizable by consumers. Among the N goods, some are
produced in the same location of one of the two markets (i.e., g ¼s), sharing
origin with insiders I, and information is communicated on the label. The
market for outsiders O instead satisfies the condition g ¼−s, implying the
absence of identification with any good in the market.
Utility is defined as a probabilistic utility model (see, e.g., McFadden 1974),
consisting of a determinist component V(·) and a random component ε:
Ui¼Vðni;Xi;GIiÞ þ eið1Þ
Consumers maximizing this utility function manifest a WTP for
GI depending on whether g ¼s. In particular, GI reflects preferences
associated to a specific location (Tirole, 1996; Winfree & McCluskey, 2005),
and WTP is
WTPOðGI ¼sÞ>
<WTPOðGI ¼ sÞ ð2aÞ
This is the condition for outsiders: with no regional identity, the difference
in WTP between inside and outside products reflects pure preferences for GI
and is not predictable a priori. For insiders, GI activates feeling of shared
regional identity, so that WTP equals
WTPIðGI ¼ sÞ ¼ WTPOðGI ¼ sÞ ð2bÞ
WTPIðGI ¼sÞ>WTPOðGI ¼sÞ ð2cÞ
In other words, if insiders and outsiders hold identical preferences, both
groups should have the same WTP for outside products, while insiders would
be prepared to pay more for inside goods, ceteris paribus. The next section
outlines the model used to estimate WTP.
4 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
Econometric analysis of rankings
Imagine a market where consumers j evaluate N options i differing in their
price P, regional origin of the good GI, and other attributes X (PDO and
organic labels). For ease of reporting, attributes are grouped in a vector Z
i
¼
[P
i
, GI
i
, X
i
]. Utility may vary across individual following respondent-specific
variables D
j
. Preferences are estimated defining a utility function in the form
Uij ¼ZibþDjciþeij ð3Þ
where residuals e
ij
¼dn
i
þu
ij
contain unobservable tastes n
i
and a purely ran-
dom component u. As usual, true utility U* is treated as a latent variable. This
specification assumes consumers hold fixed preferences over attributes Z
i
and
n
i
(the coefficients b and d), while the impact of personal characteristics differs
across options (the coefficient c
i
). Residuals e0
j¼ ðe0
j;1;. . . ;e0
j;NÞ MVNð0;RÞ
are assumed multivariate normal with mean zero and variance-covariance
matrix Σ specific to the ranking of each individual j. This matrix relaxes
the independence of irrelevant alternatives (IIA) assumption (e.g., Dow &
Endersby, 2004; Hausman & Ruud, 1987; Schechter, 2010): the probability
of a rank depends on common shocks n
i
in the residual of all ranks, which
correspond to subjective expected product quality at respondent level.
Because the market offers N products, consumers can rank them from the
lowest to the highest utility expected on consumption going from 1 (the least
preferred option) to N (the most preferred option). The full ranking of
products provides additional information on preferences compared with a sin-
gle choice: a stated choice provides information on the item giving the highest
utility, treating all remaining products as equal; a ranking instead allows con-
sumers to state different levels of expected utility for all options, including
those that are not chosen. As a result, a rank-ordered probit model uses an
ordinal dependent variable, contrary to the binary nature of dependent
variables in choice models. The probability of observing a specific ranking
corresponds to the product of the probability of ranking each option first in
a progressively shrinking choice set: the consumer sequentially allocates prefer-
ences by determining the best option in the full set of N options, then the best
of the remaining N − 1 options, and so on (e.g., Fok, Paap, & Van Dijk, 2012).
In detail, the probability of the ranking provided by consumer j is the
probability that U
i, k þ1
− U
i, k
>0, for k ¼1,…, N − 1 (given b and c
i
).
3
This
inequality leads to a differenced utility
Djk ¼Uj;kþ1Uj;k¼ ðZkþ1ZkÞbþDjðckþ1ckÞ
þej;kþ1ej;k¼WkbþDjpkþnjk ð4Þ
where njk MVNð0;RjÞand k ¼1,…, N − 1. If k
ik
¼W
k
b þD
j
π
k
is
the deterministic part of Equation 4, the probability of the rank
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 5
Downloaded by [Newcastle University] at 12:25 11 May 2016
U
jN
>… >U
j2
>U
j1
equals
PrðN;N1;...;1Þ ¼ PrðDj;10;. . . ;Dj;N10Þ
¼Pðnj1 kj1;. . . ;nj;N1 kj;N1Þ
¼ ð2pÞ ðN1Þ=2Rj
1=2Zkj1
1
. . . Z
kj;N1
1
exp 1
2W0R1
jW
@Wð5Þ
which is the function to integrate numerically (e.g., Dow & Endersby, 2004).
The same approach can use a logistic link function (Lareau & Rae, 1989),
assuming independent and identically distributed extreme value distributed
residuals ε. In this case, the probability of a rank follows the rank-ordered logit
likelihood function (Beggs, Cardell, & Hausman, 1981; Foster & Mourato,
2002; Hausman & Ruud, 1987)
PrðUjN . . . Uj2Uj1Þ ¼ Y
N1
i¼1
expðZibÞ
PiexpðZibÞ
ð6Þ
Importantly, the rank-ordered logit assumes the validity of the IIA assump-
tion. Results are presented also for this option to allow interested readers to
compare estimates.
Extending Equation 3, the estimated utility function in each group
corresponds to
Uij ¼b1Piþb2GIiþb3XiþDjciþeij ð3′Þ
For both rank-ordered probit and logit, the WTP for the region of origin GI
is derived from the parameters of Equation 3′ as the marginal rate of substi-
tution between price and the characteristic (see Foster & Mourato, 2000;
Lareau & Rae, 1989) as
WTP ¼@Pj
@GIj¼ @Uij=@GIi
@Uij=@Pi¼ b2
b1ð7Þ
Noticeably, WTP for Sicily is expected to vary according to group member-
ship. While the questionnaire does not measure the perceived identity of the
consumer, identity is captured by the design of the survey: it equals 1 for the
two samples of insiders and 0 for the set of outsiders.
Data
Data to test the empirical propositions of the previous section have been
collected through a survey to 1,000 Italian consumers on extra-virgin olive
oil consumption. Data were collected through face-to-face interviews on four
subsamples: two groups of insiders (250 Sicilian consumers each in Palermo
and Catania) and two groups of outsiders from different regions (250
6 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
respondents each in Rome and Milan). This approach was chosen because of
its simplicity in collecting data and the high completion rate (Hox & De
Leeuw, 1994). The sample included only consumers residing permanently
in the corresponding city
4
(self-reported), and approximately 93% of indivi-
duals who started the survey completed it. The choice of urban locations
aimed at excluding consumers with direct connection with the production
of extra-virgin olive oil. Questionnaires were administered to food shoppers
in mall intercepts at a large retail store in each of the four study areas. The
structure of the final questionnaire was developed using results and infor-
mation derived from previous focus groups.
5
The survey collected infor-
mation on motivations and attitudes for the purchase of olive oil in general
and extra-virgin olive oil in particular. The questionnaire also inquired about
economic barriers and drivers to olive oil consumption, as well as the socio-
economic characteristics of the respondent. The choice of a mall intercept was
guided by the need to capture a random population of consumers (i.e., indi-
viduals responsible for household provisions) in a real shopping environment,
obtaining a varied sample of individuals.
6
Table 1 presents summary statistics
for the respondents included in the final analysis.
The survey included a contingent ranking experiment (see, e.g., Bateman,
Cole, Georgiou, & Hadley, 2006; Foster & Mourato, 2000; Lareau & Rae,
1989). In this study, consumers were presented with nine different olive oil
products, differing in terms of price, origin, organic label, and PDO. The
choice set was obtained orthogonalizing all attributes (including price levels)
to remove collinearity. To limit the occurrence of an investigator bias, consu-
mers performed the conjoint analysis alone, allowing respondents to ponder
and rank options with the questionnaire in front of them.
7
Each of the three
regions considered has different PDO labels, but the experimental choice card
(in Table 2) refers to a generic PDO to ensure consumers could understand the
Table 1. Demographic characteristics of the sample.
Category Variable
Catania Palermo Milan Rome
% % % %
Gender Female 52.1 52.5 69.1 57.4
Male 47.9 47.5 30.9 42.6
Age (yr) 18–30 14.4 7.2 12.4 10.0
31–45 40.7 46.2 41.6 50.6
46–60 34.7 36.7 34.3 22.1
>60 10.2 10.0 11.8 17.3
Education Primary 22.4 14.1 24.2 31.3
Secondary 42.4 48.9 51.7 41.8
Graduate/postgraduate 35.2 37.1 24.2 26.9
Income (euros) 10,000 8.0 5.0 2.2 4.0
−10,000–20,000 44.1 36.2 27.0 36.5
−20,000–40,000 35.2 44.3 48.3 48.2
40,000 12.7 14.5 22.5 11.2
Respondents 234 221 178 249
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 7
Downloaded by [Newcastle University] at 12:25 11 May 2016
choice task; estimated coefficients then capture the average value consumers
assign to the label, while their standard errors account for the heterogeneity
in consumer perception. Consumers were then asked to rank products accord-
ing to their tastes, from 1 (“least preferred”) to 9 (“most preferred”). Tied ranks
were not allowed. The final choice set is presented in Table 2.
Results
This section presents the results of the contingent ranking experiment fitted on
the four samples of consumers presented in the previous section. Specifically,
Table 3 reports estimates from a rank-ordered logit, while Tables 4 and 5 are
estimates from a rank-ordered probit that relaxes the IIA assumption. Results
in Table 5 differ from those in Table 4 by including stated personal preferences
for PDO, organic, and own-regional products (these are explained next). The
rank-ordered probit was estimated using an unstructured variance–covariance
matrix with J × (J − 3)/2 þ1 correlation parameters. Estimation uses the
GHK algorithm to approximate the multivariate distribution function, using
option 1 as the utility-normalizing option (setting its standard deviation to
1 and its correlations with other errors to 0), and option 2 as the utility
scale–normalizing option. WTP values have been estimated according to
Equation 7, using Tuscany as the baseline regional dummy. Because results
are fairly consistent across model specification, the analysis follows primarily
the rank-ordered probit.
Before proceeding, some insight could be gathered from observing the aver-
age rankings per sample. The ranking of each of the nine options in the four
samples (Figure 1) indicates that Sicilian oils (3, 7, and 8) occupy relatively
high ranks in Sicilian samples, where option 7 (the least expensive) is always
ranked at the top. The sample of outsiders instead preferred Tuscan options,
leaving Sicilian products in the third or fourth position at the most. Both Rome
and Milan groups also present a Sicilian option (number 8, the most expensive)
as the least preferred option in the list. Figure A1 in Appendix A shows that
Sicilian consumers tend to be more likely to rank top a Sicilian option, while
the link region-top rank is less clear in the groups of outsiders. These figures
Table 2. Description of the choice set.
Option Price (€) Origin Organic PDO
1 10.5 Tuscany Yes Yes
2 10.5 Apulia Yes No
3 8.5 Sicily Yes No
4 8.5 Apulia No Yes
5 8.5 Tuscany Yes Yes
6 6.5 Apulia Yes Yes
7 6.5 Sicily Yes Yes
8 10.5 Sicily No Yes
9 6.5 Tuscany No No
8 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
Table 3. Estimated parameters of rank-ordered logit.
Catania Palermo Milan Rome
Coefficient S.E. Coefficient S.E. Coefficient S.E. Coefficient S.E.
Price −0.3052*** 0.0178 −0.3083*** 0.0190 −0.0996*** 0.0188 −0.1294*** 0.0161
Organic 0.7048*** 0.0584 0.8356*** 0.0601 0.8205*** 0.0675 0.8053*** 0.0565
PDO 0.9818*** 0.0598 1.4581*** 0.0688 0.7298*** 0.0657 0.8793*** 0.0569
Apulia 0.1227* 0.0628 −0.1253* 0.0659 0.3010*** 0.0730 0.2864*** 0.0612
Sicily 0.8462*** 0.0642 0.2754*** 0.0658 −0.0154 0.0740 −0.1060** 0.0622
WTP Organic 2.3096 2.7101 8.2372 6.2248
WTP PDO 3.2173 4.7289 7.3268 6.7972
WTP Apulia 0.4019 −0.4065 3.0216 2.2136
WTP Sicily 2.7728 0.8932 −0.1542 −0.8192
Respondents 236 221 176 247
Options 9 9 9 9
LR v
2
(5) 747.45*** 878.15*** 340.34*** 553.23***
Log likelihood −2647.51 −2390.13 −2082.95 −2885.43
Significance is as follows: *P ¼.10, **P ¼.05, ***P ¼.01.
9
Downloaded by [Newcastle University] at 12:25 11 May 2016
Table 4. Estimated parameters of rank-ordered probit.
Catania Palermo Milan Rome
Coefficient S.E. Coefficient S.E. Coefficient S.E. Coefficient S.E.
Price −0.2881*** 0.0343 −0.2223*** 0.0332 −0.9988*** 0.1273 −0.7922*** 0.0831
Organic 0.5709*** 0.0885 0.7399*** 0.0706 1.7336*** 0.2296 1.5156*** 0.1580
PDO 0.9108*** 0.0875 1.1991*** 0.0927 1.5675*** 0.2239 1.6289*** 0.1545
Apulia −0.0409 0.0582 −0.2877*** 0.0592 0.9222*** 0.2176 0.6468*** 0.1411
Sicily 0.8057*** 0.0912 0.0964 0.0602 0.7779*** 0.2483 0.1063 0.1542
WTP organic 1.9814 3.3275 1.7356 1.9130
WTP PDO 3.1613 5.3929 1.5693 2.0561
WTP Apulia −0.1420 −1.2940 0.9233 0.8164
WTP Sicily 2.7964 0.4334 0.7788 0.1342
Respondents 236 221 176 247
Options 9 9 9 9
Wald v
2
(5) 193.83*** 202.84*** 85.49*** 166.19***
Log likelihood −2388.5438 −2036.43 −1528.64 −2228.82
Significance is as follows: *P ¼0.10; **P ¼0.05; ***P ¼0.01. Note: option 1 is the alternative normalizing location; option 2 is the alternative normalizing scale.
10
Downloaded by [Newcastle University] at 12:25 11 May 2016
Table 5. Estimated parameters of rank-ordered probit with interest for region, PDO, and organic.
Catania Palermo Milan Rome
Coefficient S.E. Coefficient S.E. Coefficient S.E. Coefficient S.E.
Price −0.2473*** 0.0447 −0.3575*** 0.0556 −1.3155*** 0.1676 −0.6336*** 0.0998
Organic 0.6782*** 0.0843 0.4133*** 0.1349 2.0585*** 0.2737 1.4280*** 0.1864
PDO 1.1888*** 0.1094 0.8973*** 0.1351 1.7924*** 0.2652 1.4520*** 0.1804
Apulia −0.3764*** 0.0824 −0.1294 0.0986 1.0363*** 0.2546 0.5503*** 0.1838
Sicily −0.0073 0.0825 0.7812*** 0.1356 0.8191*** 0.2776 0.0619 0.2015
WTP organic 2.7426 1.1561 1.5648 2.2538
WTP PDO 4.8071 2.5097 1.3625 2.2918
WTP Apulia −1.5220 −0.3618 0.7878 0.8686
WTP Sicily −0.0294 2.1850 0.6226 0.0977
Option 1 Baseline Baseline Baseline Baseline
Option 2 Region 0.2431 0.2595 0.4146** 0.2080 −0.2332 0.5150 −0.1824 0.2151
PDO −0.5405** 0.2366 −0.1958 0.2077 −0.6384** 0.2884 0.1056 0.2435
Organic 0.5763* 0.3068 −0.2761 0.2311 0.7730** 0.3465 −0.4686* 0.2532
Option 3 Region 0.2284 0.3274 0.4573 0.2797 −0.3099 0.8846 0.3722 0.3509
PDO −0.5791** 0.2950 −0.6574** 0.2730 −2.7217*** 0.5507 −0.1244 0.3950
Organic 0.5927 0.3891 −0.6400** 0.3006 1.0478 0.6535 0.1353 0.3913
Option 4 Region 0.0584 0.2777 0.0923 0.2356 −0.4138 0.8071 0.5346* 0.3008
PDO −0.2158 0.2471 −0.0771 0.2312 −0.8950* 0.4920 −0.3433 0.3364
Organic −0.0139 0.3292 −0.6860*** 0.2606 −0.5815 0.5906 0.7681** 0.3390
Option 5 Region 0.0896 0.2164 0.0005 0.1790 0.7155 0.7757 0.0970 0.2456
PDO −0.2293 0.1876 −0.3326* 0.1837 −1.3710*** 0.4066 −0.2471 0.2710
Organic −0.0686 0.2627 −0.0105 0.2113 0.1071 0.4910 0.9941*** 0.2850
Option 6 Region 0.4350 0.3790 −0.1780 0.2763 0.1973 1.8568 −0.8495 0.5981
PDO −0.4365 0.3319 −0.3918 0.2691 −6.7825*** 1.1544 −1.7092** 0.7049
Organic 0.0500 0.4456 0.3147 0.2942 −0.5340 1.3318 1.1353 0.7163
Option 7 Region 0.6695 0.4376 0.2747 0.3649 0.1794 1.8928 −0.5873 0.5942
PDO −0.3822 0.3768 −0.8132** 0.3530 −6.6403*** 1.1518 −1.6022** 0.6997
Organic 0.2447 0.5174 0.3905 0.3951 −0.7229 1.3552 0.8801 0.7107
Option 8 Region −0.0348 0.2273 0.2789 0.2480 −0.2844 0.7507 0.1319 0.2636
PDO 0.0081 0.2030 −0.2444 0.2484 0.4374 0.3884 −0.1402 0.3046
(Continued)
11
Downloaded by [Newcastle University] at 12:25 11 May 2016
Table 5. Continued.
Catania Palermo Milan Rome
Coefficient S.E. Coefficient S.E. Coefficient S.E. Coefficient S.E.
Organic −0.0629 0.2754 −0.7969*** 0.2834 −1.8005*** 0.5458 −1.0169*** 0.3238
Option 9 Region 0.1613 0.4577 0.2348 0.4004 6.7859*** 2.4191 −1.8549** 0.7550
PDO −0.7031* 0.4067 −0.5231 0.3881 −7.0286*** 1.5072 −1.5766* 0.8723
Organic −0.0994 0.5467 −1.1374*** 0.4237 −2.6403 1.7809 0.9946 0.8712
Respondents 221 236 176 247
Options 9 9 9 9
LR chi2(29) 206.32*** 195.71*** 124.50*** 205.41***
Log likelihood −2009.10 −2353.99 −1436.16 −2181.3308
Significance is as follows: *P ¼.10, **P ¼.05, ***P ¼.01.
Note: option 1 is the alternative normalizing location; option 2 is the alternative normalizing scale.
12
Downloaded by [Newcastle University] at 12:25 11 May 2016
seem to provide preliminary support to the intuition that the “Sicily” brand is a
more important quality signal to insiders than to outsiders. Further, it suggests
that insiders do not necessarily rank a Sicilian option as first, while ranking
down options from other regions (i.e., a negative OOH-COO bias). However,
these initial considerations are only speculative, and only the ceteris paribus
analysis that follows can lead to more accurate considerations on consumer
behavior.
WTP for regional oils in inside and outside groups
Results from the estimated rank-ordered logit and probit are presented in
Table 3 and 4 respectively, while Table 5 reports estimated parameters for
the rank-ordered probit with demographics included. Results indicate that
preferences for regions vary across regional identity groups, albeit presenting
some common features: both Sicilian samples rank highest Sicilian oils (in
Palermo jointly with Tuscan oils), and Apulian oils always feature last (in
Catania jointly with Tuscan oils). The specific rank of WTP by region can
be found in Appendix B. The same pattern, expectedly, characterize the
WTP for different regions (Table 6 and Figure 2). The sample for Catania
appears to be the one with the highest interest in Sicilian oils, also registering
the highest WTP for the region (relative to Tuscany). The different pattern of
WTP in insiders could accounts for two items: first, the Sicilian capital
Palermo could have higher exposure to continental products compared with
Catania, which is a smaller city and has a smaller port; second, western Sicily,
where Palermo is located, can count on a larger olive-growing area and has
easier access to olive oil in local markets or in farms, providing a stronger
provincial rather than regional identity.
Figure 1. Average rank of options, by sample. Square ¼Sicilian option; triangle ¼Tuscan option;
rhombus ¼Apulian option. Bars represent bootstrapped standard errors (1,000 replications).
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 13
Downloaded by [Newcastle University] at 12:25 11 May 2016
In terms of the remaining covariates, price is negatively related to utility, with
outsiders presenting higher price sensitivity for oil. This higher sensitivity to
changes in price reflects the fact that, despite being embedded in the national
Figure 2. Average WTP by sample and model used. Bars represent bootstrapped standard errors
(100 replications).
Table 6. Estimated WTP for Sicily and Apulia.
WTP Logit Probit 1 Probit 2
Insiders
Catania Sicily € 2.77*** € 2.80*** € 2.19***
Apulia € 0.40 −€ 0.14 −€ 0.36
Palermo Sicily € 0.89 *** € 0.43 −€ 0.03
Apulia −€ 0.41* −€ 1.29*** −€ 1.52***
Outsiders
Milan Sicily −€ 0.15 € 0.78*** € 0.62**
Apulia € 3.02*** € 0.92*** € 0.79**
Rome Sicily −€ 0.82 € 0.13 € 0.10
Apulia € 2.21*** € 0.82*** € 0.87**
Significance is as follows: *P ¼.10, **P ¼.05, ***P ¼.01.
Logit refers to rank-ordered logit; probit 1 refers to rank-ordered probit without demographics; and probit 2
refers to rank-ordered probit with demographics. Confidence intervals have been estimated using 100
bootstrap replications.
14 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
food culture, olive oil is less than a commodity in urban areas, especially in
regions with limited olive production. Expectedly, organic and PDO certifica-
tions represent important quality signals and strongly influence the ranking
given to products. WTP for a PDO label is slightly higher for insiders compared
with outsiders: the limited interface insiders have with producers in large retail
stores compared with other sources (e.g., at farm) could lead to problems of
imperfect information, and a WTP for PDO labels could help guarantee the
truthfulness of the unobservable origin of the product ex ante, thus preventing
ex post dissatisfaction. Conversely, WTP for organic labeling appears close
across the four samples. The difference in estimated coefficients between probit
and logit (i.e., after relaxing the IIA assumption) supports the notion that unob-
servable preferences for taste matter in the determination of consumer prefer-
ences: the upward bias in the coefficient of price in outsiders before adjusting
for n suggests unobservable characteristics are valued positively, and consumers
use price to infer unobservable quality (e.g., Panzone, 2012; Wolinsky, 1983).
Inclusion of consumer preferences for own-region, organic, and PDO
products
The results of the rank-ordered probit in Table 4 have been extended to incor-
porate personal stated preferences for own-region olive oil, PDO, and organic
products (Table 5), which were recorded in the questionnaire. In particular,
consumers were asked the following questions:
“Are you interested to quality certifications in the olive oil you purchase? [yes/no
question] If yes, which ones?”
Consumers could choose one or more of PDO label, PGI label, and organic
label. Preferences for PDO and organic products were then coded as binary
variables for those consumers indicating the interest in these two labels.
Consumers were also asked the following question:
“Where does the olive oil you habitually buy come from?”
They could choose only one option from “local,” “regional,” “national,” or
“outside the EU.” Preferences for own-regional olive oil were coded as a binary
variable equal to 1 if consumers answered “regional” or “local.” Noticeably,
“local” differs from “regional” in spatial terms, as it refers to a stronger link with
land and its rural economy (e.g., Hinrichs, 2000); however, “local” is a “subset”
of the region where the individual resides, and they are considered jointly.
Appendix C shows that consumers have different interests in own-region,
PDO, and organic products, supporting the need for this further analysis.
Results indicate that stated interests for region, PDO, and organic influences
the rank assigned to each product (Table 5). Specifically, Sicilian consumers do
not manifest their stated interest for own-regional products: coefficients can be
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 15
Downloaded by [Newcastle University] at 12:25 11 May 2016
positive and large for Sicilian products in both Catania and Palermo, but they
are not significant. Only in Catania is this variable significant, but it favors a
Tuscan option (option 2). In Milan, an interest in regional oils favors Tuscan
option 9, while in Rome they give a significant advantage to Apulian option 4
and a disadvantage to Tuscan option 9. These results suggest that consumers
who reported to habitually look for regional products use region as a key driver
of rank (and choice), independent of the specific region of origin. Consumers
instead use their stated interest for PDO against both PDO and non-PDO
options: consumers expect lower quality from PDOs compared with the base-
line (a high-priced Tuscan PDO). Conversely, preferences for organic labels
tend to favor organic products in Rome and discourage nonorganic options
in other samples. There also seems to be a synergy between organic and
PDO labels: interest in organic oils leads to low rankings for organic labels
whenever it has no PDO label as well.
Discussion: I like it more if it comes from the same place as me
GIs are an important tool to provide information to consumers on the origin of
the food they purchase. The general literature on GIs considers consumers as
interested in the label purely on the basis of the information it provides. This
article extends this knowledge by exploring preferences for origin by consu-
mers with different regional identities. With the objective of measuring the
WTP for a Sicilian origin of olive oils, the empirical analysis assessed consumer
preferences using a contingent ranking exercise on two samples of Sicilian
(insiders) and two samples of non-Sicilian consumers (outsiders from Rome
and Milan). WTP was estimated using rank-ordered probit and logit models.
The use of rankings provides more information than a choice, because it allows
respondents to report different levels of utility for each product in the choice
set. Methodologically, this exercise is also one of the first use of a rank-ordered
probit in empirical consumer behavior research, adjusting for the presence of
unobservable tastes in the residuals of each consumer–product combination
(see also Schechter, 2010). The analysis focuses on the olive oil market due
to the relevance of GIs to consumers in this market (e.g., Espejel et al., 2008;
Menapace, Colson, Grebitus, & Facendola, 2011). Results indicate that prefer-
ences are fairly similar within identity groups, while differing across identity
groups. Preferences for other characteristics are fairly stable across markets.
Identity theory predicts that a link between origin of the consumer and
origin of the good leads to an additional positive contribution to the utility
insiders’ estimate for a good. This component adds to pure preference for a
region, favoring the evaluation of products that originate in the same location
of the consumer. Results support this intuition: regional preferences for Sicily
are higher in insiders than in outsiders. Part of this difference can be ascribed
to the perception the consumer has of the region (e.g., Van der Lans et al.,
16 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
2001; Winfree & McCluskey, 2005), which for insiders also accounts for the
value of the social group (Ahmed, 2007; Chen & Li, 2009) and the economic
well-being of local communities (Tregear et al., 2007). This preference is acti-
vated from feelings of affinity between product and consumer (Oberecker
et al., 2008). Taste preferences can undoubtedly influence the ranking
decisions: exposure is known to increase liking (Birch & Marlin, 1982) through
product familiarity (Wansink, 2002) and affect (Van der Lans et al., 2001),
and the priority given to an “inside” product is going to increase taste prefer-
ences in the long run. The modeling via rank-ordered probit accounts for the
presence of expected unobservable tastes in the variable n, presenting an
estimate that only accounts for preferences for region of origin.
Results highlight that regional identity does not always lead to a dominant
role of the own-regional product: regional identity also allows for some level
of xenophilic preferences (Perlmutter, 1954), with some insiders preferring pro-
ducts originating from outside (i.e., “Tuscany” generates the same utility as
“Sicily” in Palermo). Conversely, regional identity appears to lead to a dislike
of outside products: products originating outside the region are always less valu-
able, or equal to at best, than own-regional products. As a result, an H-COO bias
might work asymmetrically: it does not increase the value of inside products but
reduces the value of outside products. This asymmetry is consistent with
research on the critical judgment over controversial inside matters: insiders
value an opinion asymmetrically, whereby the same statement is considered
neutral if coming from insiders and negative if coming from outsiders (Hornsey,
Oppes, & Svensson, 2002). From a marketing perspective, results indicate that
the nature of GI labeling differs across identify group, and consumers are more
interested in own-regional products that outsiders. Different markets may differ
in their reaction toward outside options, and retailers should consider more-
decentralized strategies: for instance, they could supply mostly inside goods in
markets with strong social identity, complementing the market with only a small
number of outside options, while in markets with weaker social identity, choice
sets could be able to accommodate larger regional variety. Conversely, markets
with no regional production of olive oil should accept more freely any outside
options. Alternatively, supplier could try to explore mechanisms suitable to
overcome barriers linked to social identity, for instance introducing impartial
measures of quality (e.g., ratings from specialized journals) or using promotional
strategies such as tasting sessions or targeted discounts.
Results might not fully hold in different distribution systems. In fact, con-
sumers shopping in large retail stores would be expected to rely on heuristics
that allow them to detect the unobservable quality of a product. In this context,
the information on the label plays an important role in the definition of quality
(Broniarczyk & Alba, 1994), while different cues might apply if the product was
sourced elsewhere (e.g., directly from a supplier). In support of this point,
research observes that different retail channels influence consumer choices
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 17
Downloaded by [Newcastle University] at 12:25 11 May 2016
and price acceptability (Baker, Grewal, & Parasuraman, 1994; D’Amico, Di
Vita, Chinnici, Pappalardo, & Pecorino, 2014; Degeratu, Rangaswamy, &
Wu, 2000; Grewal & Baker, 1994). While the identity of the consumer would
be expected to be relevant in all segments of the market, it would be taken
as given in certain channels (local market) and less so in more impersonal
marketplaces, particularly as consumers may choose large retail stores for
other identity-unrelated objectives, such as saving money (see, e.g., Di Vita,
D’Amico, La Via, & Caniglia, 2013). Finally, while the results presented in this
work can be considered expected given the widespread consumption of olive
oil in Italy, the same results might not apply to GIs in other more differentiated
products (e.g., cheese or wine).
8
Because the current data set does not compare
behavior in different retail channels or different products, this is a testable
implication left for future research.
In terms of more general research, the findings of this article indicate that
choices are not a dry representation of consumer preferences but are inter-
twined with the personality of respondents (see Akerlof & Kranton, 2000).
Specifically, the presence of a GI label can activate emotions and feelings that
influence choices beyond the pure information they convey. Moreover,
choices maximize not only personal utility but also social utility, to the extent
that consumers choose to satisfy socially agreed standards of behavior (real or
perceived). These choices inevitably contribute to the personal development
of the consumer, reinforcing preferences for the label over time. As a result
of this social utility, GIs influence choices beyond pure preferences: some
insiders identify with a product with a GI label, which very likely represent
the socially responsible choice (i.e., the choice that can maintain standards
of living and welfare in the area). While sharing a regional identity with a
product is not sufficient to increase the value assigned to a GI, it can devalue
products from outside the area.
Finally, the results indicate that GIs are valued more by insiders than by out-
siders. While information is likely to be relevant for both segments, consumers
are likely to view the label as a more important feature when they perceive extra
benefits from their choices to fall within the remit of their own locality. This
point is supported by observing the coefficients of Tables 4 and 5: for insiders,
the utility derived from own-region labels can be larger than the disutility of
price, an effect that is not present in outsiders. To this extent, regions play a
prominent role in the Italian socioeconomic context, and consumers expect
the benefits from purchasing an inside products to stay within the economy.
The PDO is instead valued similarly across samples, implying limited differ-
ences in terms of the value associated with a guaranteed origin. The direct
implication is that current labels are not neutral to the eye of a consumer but
are valued differently across identity groups. As a result, the current regulation
of GIs should be improved by accounting for the role of social identity in
choices and should ensure that earnings from GIs are used to the direct benefit
18 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
to local economies in terms of employment and innovation. Importantly, there
is a general lack in research evaluating current GI policies in terms of their long-
term impact on local governance, consumer welfare, and behavioral change.
Conclusions
This article highlights the importance of regional identity in the behavior of
consumers, particularly with respect to their WTP for GIs. GIs are an impor-
tant element where regional identity can be observed because consumers from
a production area can identify with goods that bear the same name as the
location of origin. The utility they derive then stems not only from knowledge
of and familiarity with the taste of the final goods but also from a broader set of
preferences that includes local identity. Results indicate that preferences for a
specific GI depend on the ability of consumers to associate with that same
location. As a result, the relation between choices and the perceived standard
of the social group where the consumer belongs should be explored further in
the future. Social identity is rarely considered in an applied model of consumer
behavior, and previous research focused primarily on the implications on labor
markets (see Akerlof & Kranton, 2000, 2005). Nevertheless, social utility
appears to have an influence on different areas of personal choice and
consumer behavior and the potential for research in this area is vast. Further
research should develop a more accurate model of economic consumer
behavior consistent with that incorporate social identity, in order to improve
the predictive power of existing models and to provide more powerful insights
for policymaking and research.
Notes
1. Currently, the EU identifies 1,321 food products that are awarded a regulated GI, ranking
from more general Traditional Speciality Guaranteed (TSG) to Protected Geographical
Indication (PGI) and to the highest level of Protected Designation of Origin (PDO).
2. Part of this process is likely to be automatic and driven by the presence of an identifiable
geographical name (i.e., a priming process). In experimental exercises, individuals are
primed with sentences containing selected keywords that relate to the targeted emotion
(e.g., Epley & Gilovich, 1999). In the case of social identity, pronouns such as “we” or
“they” can be sufficient to prime feelings of social identity (e.g., Brewer & Gardner,
1996).
3. While the rank-ordered logit allows for the presence of tied ranks (i.e., the utility of two
ranks can be the same), the data set used in the analysis contains no ties.
4. Rome and Milan are cities with large communities from southern Italian regions. The
survey did not collect this information but screened respondents who were not residents.
These groups of migrant are still minorities in the cities surveyed, and some people with a
regional link are very likely present in the data set (and we cannot discriminate them
from the data). Nonetheless, the literature shows the general tendency of immigrants to
absorb the norms of the host country very quickly to signal integration (Dustmann,
1996), limiting the occurrence of this problem.
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 19
Downloaded by [Newcastle University] at 12:25 11 May 2016
5. Two preliminary focus groups aimed at selecting the broad items to include in the final
questionnaire. The first focus group interviewed producers, technical consultants (agrono-
mists and agricultural economists), public officers of the agricultural regional department,
and a producers’ association (PDO committee of different geographic areas). In a second
focus group, a group of consumers were invited to express their opinion with respect to
their attitudes toward olive oil (its use, shopping places, and so on) and the most important
attributes they consider when shopping (color, transparency, price, method of production,
and so on). Focus groups only discussed “Sicilian olive oil,” in order to detect and identify
main technical and economic attributes of Sicilian olive oil productions.
6. www.stata.com/manuals13/rasroprobit.pdf.
7. We thank an anonymous referee for suggesting the inclusion of this point.
8. We thank an anonymous referee for suggesting this point.
Acknowledgments
We are particularly grateful to Elena Caniglia for excellent and dedicated support in question-
naire design and data collection. We would also like to thank Luisa Monaco, Carlo Reggiani,
and Gianni Cicia for useful comments.
Funding
We are indebted to the Agriculture Department of the Sicilian Regional Government for
financial support through the project “Technical, Economic, and Trade Analysis of Olive
Oil Supply Chain in Sicily” (director: Prof. Mario D’Amico).
Contributors
Luca Panzone is a Lecturer in Consumer Behaviour at Newcastle University. His research interests
focus on the analysis of social problems related to agriculture, food and the environment. In
particular, his interest is mainly focused on the quantitative analysis of food consumption in
households and consumers. His research focuses on the drivers of consumer choices; the role of
consumer motivation and perceptions in making healthy and sustainable choices; and the design
of sustainable food policy and retail environments.
Giuseppe Di Vita is Research Fellow in Agro-food economics at the Department of Agri-
culture, Food and Environment (Di3A), University of Catania and at the Mediterranean
University of Reggio Calabria; and visiting researcher at the University of Turin (Italy). His
current research interests focus on agricultural and food economics, farm business, agro-food
marketing, and rural development.
Stefania Borla is an independent researcher in microeconomics. At the time of this research she
was Teaching Fellow at the University of Warwick. Her research interests focus on mathematical
modelling in Industrial Organisation, particularly in the area of retailing and labour economics.
Mario D’Amico is Associate Professor of Agricultural Economics and Policy at the Depart-
ment of Agriculture, Food and Environment (Di3A), University of Catania (Italy). Its current
research interests include: micro economics, Agro-food management, Consumer preferences,
Agro-food marketing and Fisheries economics.
ORCID
Luca Panzone http://orcid.org/0000-0003-2382-3635
20 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
References
Ahmed, A. M. (2007). Group identity, social distance and intergroup bias. Journal of Economic
Psychology, 28(3), 324–337. doi:10.1016/j.joep.2007.01.007
Akerlof, G. A., & Kranton, R. E. (2000). Economics and identity. The Quarterly Journal of
Economics, 115(3), 715–753.
Akerlof, G. A., & Kranton, R. E. (2005). Identity and the economics of organizations. The
Journal of Economic Perspectives, 19(1), 9–32. doi:10.1257/0895330053147930
Baker, J., Grewal, D., & Parasuraman, A. (1994). The influence of store environment on quality
inferences and store image. Journal of the Academy of Marketing Science, 22(4), 328–339.
doi:10.1177/0092070394224002
Bateman, I. J., Cole, M. A., Georgiou, S., & Hadley, D. J. (2006). Comparing contingent
valuation and contingent ranking: A case study considering the benefits of urban river water
quality improvements. Journal of Environmental Management, 79(3), 221–231. doi:10.1016/
j.jenvman.2005.06.010
Beggs, S., Cardell, S., & Hausman, J. (1981). Assessing the potential demand for electric cars.
Journal of Econometrics, 16, 1–19. doi:10.1016/0304-4076(81)90056-7
Birch, L. L., & Marlin, D. W. (1982). I don’t like it; I never tried it: Effects of exposure on two-
year-old children’s food preferences. Appetite, 3(4), 353–360.
Brewer, M. B., & Gardner, W. (1996). Who is this “We”? Levels of collective identity and self
representations. Journal of Personality and Social Psychology, 71(1), 83–93. doi:10.1037/
0022-3514.71.1.83
Broniarczyk, S. M., & Alba, J. W. (1994). The role of consumers’ intuitions in inference
making. Journal of Consumer Research, 21(3), 393–407. doi:10.1086/209406
Brunori, G. (2007). Local food and alternative food networks: A communication perspective.
Anthropology of Food, (S2). Retrieved from http://aof.revues.org/430
Caswell, J. A., & Mojduszka, E. M. (1996). Using informational labeling to influence the
market for quality in food products. American Journal of Agricultural Economics, 78,
1248–1253. doi:10.2307/1243501
Chen, Y., & Li, S. X. (2009). Group identity and social preferences. American Economic Review,
99(1), 431–457.
D’Amico, M., Di Vita, G., Chinnici, G., Pappalardo, G., & Pecorino, B. (2014). Short food
supply chain and locally produced wines: factors affecting consumer behavior. Italian
Journal of Food Science, 26(3), 329–334.
Degeratu, A. M., Rangaswamy, A., & Wu, J. (2000). Consumer choice behavior in online and
traditional supermarkets: The effects of brand name, price, and other search attributes. Inter-
national Journal of Research in Marketing, 17(1), 55–78. doi:10.1016/s0167-8116(00)00005-7
Di Vita, G., D’Amico, M., La Via, G., & Caniglia, E. (2013). Quality perception of PDO extra-
virgin olive oil: which attributes most influence Italian consumer? Agricultural Economics
Review, 2, 46–58.
Dow, J. K., & Endersby, J. W. (2004). Multinomial probit and multinomial logit: A comparison
of choice models for voting research. Electoral Studies, 23, 107–122. doi:10.1016/s0261-3794
(03)00040-4
Dustmann, C. (1996). The social assimilation of immigrants. Journal of Population Economics,
9(1), 37–54.
Epley, N., & Gilovich, T. (1999). Just going along: Nonconscious priming and conformity to
social pressure. Journal of Experimental Social Psychology, 35(6), 578–589.
Espejel, J., Fandos, C., & Flavian, C. (2008). Consumer satisfaction: A key factor of consumer
loyalty and buying intention of a PDO food product. British Food Journal, 110(9), 865–881.
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 21
Downloaded by [Newcastle University] at 12:25 11 May 2016
Fok, D., Paap, R., & Van Dijk, B. (2012). A rank-ordered logit model with unobserved
heterogeneity in ranking capabilities. Journal of Applied Econometrics, 27(5), 831–846.
doi:10.1002/jae.1223
Foster, V., & Mourato, S. (2000). Valuing the multiple impacts of pesticide use in the UK: A
contingent ranking approach. Journal of Agricultural Economics, 51(1), 1–21. doi:10.1111/
j.1477-9552.2000.tb01206.x
Foster, V., & Mourato, S. (2002). Testing for consistency in contingent ranking experiments. Jour-
nal of Environmental Economics and Management, 44(2), 309–328. doi:10.1006/jeem.2001.1203
Giannakakis, A. E., & Fritsche, I. (2011). Social identities, group norms, and threat: On the
malleability of ingroup bias. Personality and Social Psychology Bulletin, 37(1), 82–93.
doi:10.1177/0146167210386120
Grewal, D., & Baker, J. (1994). Do retail store environmental factors affect consumers’ price
acceptability? An empirical examination. International Journal of Research in Marketing,
11(2), 107–115. doi:10.1016/0167-8116(94)90022-1
Guerrero, L., Guàrdia, M. D., Xicola, J., Verbeke, W., Vanhonacker, F., Zakowska-Biemans, S.,
… Hersleth, M. (2009). Consumer-driven definition of traditional food products and inno-
vation in traditional foods. A qualitative cross-cultural study. Appetite, 52(2), 345–354.
doi:10.1016/j.appet.2008.11.008
Hausman, J., & Ruud, P. (1987). Specifying and testing econometric models for rank-ordered
data. Journal of Econometrics, 34, 83–104. doi:10.1016/0304-4076(87)90068-6
Hinrichs, C. C. (2000). Embeddedness and local food systems: Notes on two types of direct agri-
cultural market. Journal of Rural Studies, 16(3), 295–303. doi:10.1016/s0743-0167(99)00063-7
Hogg, M. A., & Williams, K. D. (2000). From I to we: Social identity and the collective self.
Group Dynamics: Theory, Research, and Practice, 4(1), 81–97. doi:10.1037/1089-2699.4.1.81
Hornsey, M. J., Oppes, T., & Svensson, A. (2002). “It’s OK if we say it, but you can’t”:
Responses to intergroup and intragroup criticism. European Journal of Social Psychology,
32(3), 293–307. doi:10.1002/ejsp.90
Hox, J. J., & De Leeuw, E. D. (1994). A comparison of nonresponse in mail, telephone, and
face-to-face surveys. Quality and Quantity, 28(4), 329–344. doi:10.1007/bf01097014
Hu, W., Batte, M. T., Woods, T., & Ernst, S. (2012). Consumer preferences for local
production and other value-added label claims for a processed food product. European
Review of Agricultural Economics, 39(3), 489–510. doi:10.1093/erae/jbr039
Iaccarino, T., Di Monaco, R., Mincione, A., Cavella, S., & Masi, P. (2006). Influence of
information on origin and technology on the consumer response: The case of soppressata
salami. Food Quality and Preference, 17(1), 76–84. doi:10.1016/j.foodqual.2005.08.005
Josling, T. E. (2006). What’s in a name? The economics, law and politics of geographical indica-
tions for foods and beverages. IIIS Discussion Paper, No. 109. Retrieved from http://ssrn.
com/abstract=922267
Klor, E. F., & Shayo, M. (2010). Social identity and preferences over redistribution. Journal of
Public Economics, 94, 269–278. doi:10.1016/j.jpubeco.2009.12.003
Kreps, D., & Wilson, R. (1982). Reputation and imperfect information. Journal of Economic
Theory, 27, 253–279. doi:10.1016/0022-0531(82)90030-8
Kuznesof, S., Tregear, A., & Moxey, A. (1997). Regional foods: A consumer perspective. British
Food Journal, 99(6), 199–206. doi:10.1108/00070709710181531
Lareau, T. J., & Rae, D. A. (1989). Valuing WTP for diesel odor reductions: An application of con-
tingent ranking technique. Southern Economic Journal, 55(3), 728–742. doi:10.2307/1059585
Leonardelli, G. J., Pickett, C. L., & Brewer, M. B. (2010). Optimal distinctiveness theory: A
framework for social identity, social cognition, and intergroup relations. In M. P. Zanna
& J. M. Olson (Eds.), Advances in experimental social psychology (Vol. 43, pp. 63–113).
New York, NY: Elsevier.
22 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
Loureiro, M. L., & Umberger, W. J. A. (2007). A choice experiment model for beef: What US
consumer responses tell us about relative preferences for food safety, country-of-origin
labeling and traceability. Food Policy, 32(4), 496–514. doi:10.1016/j.foodpol.2006.11.006
Marette, S., Crespi, J. M., & Schiavina, A. (1999). The role of common labeling in a context of
asymmetric information. European Review of Agricultural Economics, 26(2), 167–178.
doi:10.1093/erae/26.2.167
McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior. In P. Zarembka
(Ed.), Frontiers in econometrics (pp. 105–142). New York, NY: Academic Press.
Menapace, L., Colson, G., Grebitus, C., & Facendola, M. (2011). Consumers’ preferences for
geographical origin labels: Evidence from the Canadian olive oil market. European Review
of Agricultural Economics, 38(2), 193–212. doi:10.1093/erae/jbq051
Moschini, G., Menapace, L., & Pick, D. (2008). Geographical indications and the competitive
provision of quality in agricultural markets. American Journal of Agricultural Economics, 90
(3), 794–812. doi:10.1111/j.1467-8276.2008.01142.x
Oberecker, E., Riefler, P., & Diamantopoulos, A. (2008). The consumer affinity construct:
Conceptualization, qualitative investigation, and research agenda. Journal of International
Marketing, 16(3), 23–56. doi:10.1509/jimk.16.3.23
Panzone, L. (2012). Alcohol tax, price quality proxy and discounts: A reason why alcohol taxes
may rebound. Journal of Agricultural Economics, 63(3), 715–736. doi:10.1111/j.1477-9552.
2012.00351.x
Perlmutter, H. (1954). Some characteristics of the xenophilic personality. The Journal of
Psychology, 38, 291–300. doi:10.1080/00223980.1954.9712938
Rangnekar, D. (2004). The socio-economics of Geographical Indications. A review of empirical
evidence from Europe. UNCTAD-ICTSD Project on IPRs and Sustainable Development,
Issue Paper No. 8. Retrieved from http://ictsd.org/downloads/2008/07/a.pdf
Scarpa, R., & Del Giudice, T. (2004). Market segmentation via mixed logit: extra-virgin olive
oil in urban Italy. Journal of Agricultural and Food Industrial Organization, 2(7), 1–20.
doi:10.2202/1542-0485.1080
Scarpa, R., Philippidis, G., & Spalatro, F. (2005). Estimating patriotic preferences for Mediter-
ranean products under taste heterogeneity: a nation-wide survey of Italian households.
Agribusiness, 21, 329–349.
Scarpa, R., Thiene, M., & Marangon, F. (2007). The value of collective reputation for environ-
mentally-friendly production methods: The case of Val di Gresta carrots. Journal of Agricul-
tural and Food Industrial Organization, 5(1), Article 7. doi:10.2202/1542-0485.1149
Schechter, L. (2010). The apple and your eye: Visual and taste rank-ordered probit analysis
with correlated errors. Food Quality and Preference, 21(1), 112–120. doi:10.1016/j.foodqual.
2009.08.009
Schooler, R. D. (1965). Product bias in Central American common market. Journal of
Marketing Research, 2(4), 394–397.
Shapiro, C. (1983). Premiums for high quality products as returns to reputations. The
Quarterly Journal of Economics, 4, 659–679.
Stanziani, A. (2004). Wine reputation and quality controls: The origin of the AOCs in 19th
century France. European Journal of Law and Economics, 18(2), 149–167. doi:10.1023/b:
ejle.0000045079.03968.07
Tajfel, H. (1974). Social identity and intergroup behaviour. Social Science Information, 13,
65–93. doi:10.1177/053901847401300204
Tajfel, H. (1979). Individuals and groups in social psychology. British Journal of Social and
Clinical Psychology, 18, 183–190. doi:10.1111/j.2044-8260.1979.tb00324.x
Tirole, J. (1996). A theory of collective reputations (with applications to the persistence of cor-
ruption and to firm quality). Review of Economic Studies, 63(1), 1–22. doi:10.2307/2298112
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 23
Downloaded by [Newcastle University] at 12:25 11 May 2016
Tregear, A., Arfini, F., Belletti, G., & Marescotti, A. (2007). Regional food and rural develop-
ment: the role of product qualification. Journal of Rural Studies, 23, 12–22. doi:10.1016/j.
jrurstud.2006.09.010
Van der Lans, I. A., Van Ittersum, K., De Cicco, A., & Loseby, M. (2001). The role of the
region of origin and EU certificates of origin in consumer evaluation of food products.
European Review of Agricultural Economics, 28(4), 451–477. doi:10.1093/erae/28.4.451
Wansink, B. (2002). Changing eating habits on the home front: Lost lessons from World War II
research. Journal of Public Policy & Marketing, 21(1), 90–99. doi:10.1509/jppm.21.1.90.17614
Winfree, J. A., & McCluskey, J. J. (2005). Collective reputation and quality. American Journal
of Agricultural Economics, 87(1), 206–213. doi:10.1111/j.0002-9092.2005.00712.x
Wolinsky, A. (1983). Prices as signals of product quality. Review of Economic Studies, 50(4),
647–658. doi:10.2307/2297767
Zago, A. M., & Pick, D. (2002). The public provision of information. In B. Krissoff, M. Bohman, &
J. A. Caswell (Eds.), Global food trade and consumer demand for quality (pp. 229–243).
New York, NY: Springer.
Appendix A: Probability of ranking each option first by sample
This first appendix shows the estimated probability of ranking highest a
specific option. Overall, Sicilian samples seem to show a strong preference
for one Sicilian option, while outsiders manifest less clear preferences for
Figure A1. Fitted probability that each option is preferred, by sample.
24 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
origin. Specifically, the fitted probability that each option is ranked first
(Figure A1) indicates that Sicilian samples have a single Sicilian favorite
(option 7), while for outsiders one option in each region have similar
probability of being ranked first (options 5, 6, and 7).
Appendix B: Rank of WTP for different regions in the sample
Table B1 ranks different regions within each identity group through a series of
Wald tests. The table reports the probability to reject the null hypothesis of
equality of the marginal utility of two regions, with pairings reported in
column 1. Specifically, outsiders like Sicilian products equally or more
than the baseline Tuscany (except the rank-ordered logit in Milan),
with @UO
@GI ðGI ¼sÞ @UO
@GI ðGI ¼ sÞand less than Apulian ones, with
@UO
@GI ðGI ¼sÞ<@UO
@GI ðGI ¼ sÞ. These inequalities reflect preferences specific
to this exercise and cannot be fully generalized. On the other hand, insiders
value the own GI no less than any other option on display: Apulian options
are always preferred less than Sicilian ones, while Tuscan options differ sig-
nificantly only in Catania. As a result, Sicilian options are always first, either
alone or jointly with Tuscan options, and @UI
@GI ðGI ¼sÞ @UI
@GI ðGI ¼ sÞ.
Because rank-ordered probit estimates accounts for unobservable (expected)
product characteristics, these equalities are corrected for pure taste expecta-
tions. In terms of the relations in Equations 2a–2c, Figure A1 indicates that
Catania has a significantly higher WTP for Sicily than do outsiders, and
Table B1. Preferences for region of origin within each identity group.
Catania Palermo Milan Rome
Rank-ordered logit
Prob. region
1 ¼Region 2
Apulia >Tuscany Apulia <Tuscany Apulia >Tuscany Apulia >Tuscany
0.0508 0.0570 0.0000 0.0000
Prob. region
1 ¼Region 2
Sicily >Tuscany Sicily >Tuscany Sicily ¼Tuscany Sicily <Tuscany
0.0000 0.0000 0.8356 0.0882
Prob. region
1 ¼Region 2
Sicily >Apulia Sicily >Apulia Sicily <Apulia Sicily <Apulia
0.0000 0.0000 0.0000 0.0000
Rank-ordered probit (demographics excluded)
Prob. region
1 ¼Region 2
Apulia ¼Tuscany Apulia <Tuscany Apulia >Tuscany Apulia >Tuscany
0.4818 0.0000 0.0000 0.0000
Prob. region
1 ¼Region 2
Sicily >Tuscany Sicily ¼Tuscany Sicily >Tuscany Sicily ¼Tuscany
0.0000 0.1095 0.0017 0.4905
Prob. region
1 ¼Region 2
Sicily >Apulia Sicily >Apulia Sicily <Apulia Sicily <Apulia
0.0000 0.0000 0.0811 0.0000
Rank-ordered probit (demographics included)
Prob. region
1 ¼Region 2
Apulia ¼Tuscany Apulia <Tuscany Apulia >Tuscany Apulia >Tuscany
0.1895 0.0000 0.0000 0.0028
Prob. region
1 ¼Region 2
Sicily >Tuscany Sicily ¼Tuscany Sicily >Tuscany Sicily ¼Tuscany
0.0000 0.9298 0.0032 0.7588
Prob. region
1 ¼Region 2
Sicily >Apulia Sicily >Apulia Sicily <Apulia Sicily <Apulia
0.0000 0.0000 0.0176 0.0000
Note: Probabilities refer to the probability to reject the null hypothesis of equality of the estimated coefficient
of two regions through a Wald test.
JOURNAL OF INTERNATIONAL FOOD & AGRIBUSINESS MARKETING 25
Downloaded by [Newcastle University] at 12:25 11 May 2016
Palermo has significantly lower WTP for Apulia than do outsiders:
WTPIðGI ¼sÞ WTPOðGI ¼sÞand WTPIðGI ¼ sÞ WTPOðGI ¼ sÞ.
Appendix C: Determinants of the interest in own-region, PDO, and
organic olive oil
Probit regressions on the determinant of the interest in own-region, PDO,
and organic products are presented in Table C1. Covariates include: the
logarithm of income and age; gender; education (equal to 1 if the individual
holds a high school diploma or university degree); household size; and
geography: for PDO and region, a dummy equal to 1 if respondents indicate
origin as one of the two most important criteria (out of six) of choice; and a
dummy equal to 1 if the person stated a previous purchase of organic
products. Table C1 indicates that an interest in the origin of oils is a key
determinant for preference for own-regional products and PDO in both
Sicilian samples, for PDO in the Milan sample and for region in the Rome
sample. Similarly, a previous organic purchase is an important predictor of
stated interests for organic labels in all samples. Consumers appear to perceive
PDO and organic as non-necessities: a reported interest for organic labels
increases in income in Catania, and decreases with household size in Catania
and Rome, while income matters for PDO among outsiders. Household size
is also negatively associated to region in outsiders. In terms of age, younger
consumers in Rome state higher preferences for organic labels and PDO,
while older respondents in Milan pay more attention to the PDO label.
Finally, education favors preferences for PDO oils in Catania, Milan and
Rome; and gender favors preferences for region in Palermo and for PDO
labels in Rome.
26 L. PANZONE ET AL.
Downloaded by [Newcastle University] at 12:25 11 May 2016
Table C1. Determinants of stated preferences for regional, DOP, and organic products.
Catania Palermo Milan Rome
Coefficient S. E. Coefficient S. E. Coefficient S. E. Coefficient S. E.
Regional
Intercept 0.2518*** 1.1342 −0.7756 1.5134 −0.8901 2.1709 −1.4492 1.3752
ln(income) 0.1796 0.1650 0.1722 0.2150 −0.1962 0.2821 −0.0753 0.2070
Male 0.2120 0.1753 0.5003*** 0.1935 0.0418 0.3463 0.0861 0.1782
ln(age) −0.2390 0.2677 −0.4092 0.3655 0.4036 0.5553 0.5164 0.3309
Education −0.3205 0.2355 0.4291 0.3702 −0.6515 0.4124 −0.4075* 0.2236
Household size −0.0284 0.0784 0.0137 0.0838 −0.3277* 0.1742 −0.2030** 0.0919
Geography 0.8047*** 0.1989 1.0002*** 0.2177 −0.4448 0.3423 0.5307*** 0.1720
Observations 234 221 176 247
LR χ
2
(6) 22.62*** 44.00*** 9.47 37.09***
Log likelihood −146.73 −118.90 −36.41 −144.54
Pseudo R
2
0.0716 0.1561 0.1150 0.1137
DOP
Intercept −1.4906 1.1433 −1.4501 1.4817 −8.2787*** 1.7033 −1.6398 1.4396
ln(income) 0.1146 0.1724 0.1555 0.2082 0.4704* 0.2428 0.8889*** 0.2675
Male 0.0705 0.1795 −0.0895 0.1876 0.2075 0.2353 0.7824*** 0.1998
ln(age) −0.0960 0.2671 0.0271 0.3578 1.4899*** 0.4316 −0.7541** 0.3851
Education 1.1402*** 0.2934 0.3037 0.3325 0.4973* 0.2991 0.7351*** 0.2808
Household size −0.0421 0.0797 −0.1076 0.0810 −0.0354 0.1137 0.0640 0.1029
Geography 0.3807*** 0.2172 0.7370*** 0.2042 0.6760*** 0.2162 −0.1032 0.1898
Observations 234 221 176 247
LR χ
2
(6) 30.92*** 20.80 *** 39.41*** 58.19***
Log likelihood −137.3009 −126.8102 −95.6587 −121.7062
Pseudo R
2
0.1012 0.0758 0.1708 0.1929
Organic
Intercept −0.9509 1.2544 −5.0534* 2.7800 −1.4503 1.8057 8.5238*** 2.1277
ln(income) 0.6027*** 0.2020 0.2958 0.3517 0.0801 0.2601 0.2250 0.2693
Male −0.0528 0.1974 −0.1122 0.3072 −0.3794 0.2809 −0.1655 0.2360
ln(age) −0.3274 0.3042 0.3843 0.6437 −0.1202 0.4754 −2.3355*** 0.5297
Education −0.1037 0.2667 0.6284 0.6101 0.4464 0.3677 0.0451 0.3102
(Continued)
27
Downloaded by [Newcastle University] at 12:25 11 May 2016
Table C1. Continued.
Catania Palermo Milan Rome
Coefficient S. E. Coefficient S. E. Coefficient S. E. Coefficient S. E.
Household size −0.1645* 0.0889 0.0943 0.1425 0.0220 0.1215 −0.5197*** 0.1333
Purchaser 1.4601*** 0.2034 3.0377*** 0.3798 1.3418*** 0.2526 2.5747*** 0.3482
Observations 234 221 176 247
LR χ
2
(6) 78.58*** 127.35*** 38.62*** 114.97***
Log likelihood −108.9537 −42.3301 −69.8588 −88.8229
Pseudo R
2
0.2650 0.6007 0.2166 0.3929
Significance is as follows: *P ¼.10, **P ¼.05, ***P ¼.01.
28
Downloaded by [Newcastle University] at 12:25 11 May 2016
