ASSOCIATION FOR CONSUMER RESEARCH
Labovitz School of Business & Economics, University of Minnesota Duluth, 11 E. Superior Street, Suite 210, Duluth, MN 55802
Economic Characterization of Low Income Families in the City of Sao Paulo Using Electricity Consumption As a Predictive
Variable
Eduardo Francisco, FGV, Brazil
Eric Cohen, Ibmec, Brazil
Felipe Zambaldi, UMESP, Brazil
Francisco Aranha, FGV, Brazil
This study examines the relationship between energy consumption and family income in the city of Sao Paulo, with 384 low income
households sampled in a field survey in 2005. Using decision trees and multiple regression, the energy consumption was tested as a
variable to complement Brazil’s Criterion for Economic Classification to predict family purchasing power. The combined used of the
Criterion, the electric bill amount, and the number of bedrooms shows an improvement from 0.222 to 0.300 in the explanatory power
of the model, which may be useful to retailers, government planners and financial agents in understanding socio-economic
characteristics.
[to cite]:
Eduardo Francisco, Eric Cohen, Felipe Zambaldi, and Francisco Aranha (2008) ,"Economic Characterization of Low Income
Families in the City of Sao Paulo Using Electricity Consumption As a Predictive Variable", in LA - Latin American Advances in
Consumer Research Volume 2, eds. Claudia R. Acevedo, Jose Mauro C. Hernandez, and Tina M. Lowrey, Duluth, MN :
Association for Consumer Research, Pages: 219-221.
[url]:
http://www.acrwebsite.org/volumes/14087/la/v2_pdf/LA-02
[copyright notice]:
This work is copyrighted by The Association for Consumer Research. For permission to copy or use this work in whole or in
part, please contact the Copyright Clearance Center at http://www.copyright.com/.
Latin American Advances in Consumer Research (Volume 2) / 219
Johansson, Johny K., Douglas, Susan P., and Ikujiro Nonaka (1985), “Assessing the Impact of Country of Origin on Product
Evaluations: a New Methodological Perspective,” Journal of Marketing Research, 22 (4), 388-96.
Maheswaran, Durairaj and Cathy Yi Chen (2006), “Nation Equity: Incidental Emotions in Country-of-Origin Effects,” Journal of
Consumer Research, 33 (4), 370-76.
Maheswaran, Durairaj (1994), “Country of Origin as a Stereotype: Effects of Consumer Expertise and Attribute Strength on Product
Evaluations,” Journal of Consumer Research, 21 (2), 354-65.
Samiee, Saeed, Shimp, Terence A., and Subhash Sharma (2005), “Brand Origin Recognition Accuracy: Its Antecedents and Consumers’ Cognitive Limitations,” Journal of International Business Studies, 36, 379-97.
Economic Characterization of Low Income Families in the City of Sao Paulo Using Electricity
Consumption as a Predictive Variable
Eduardo Francisco, FGV, Brazil
Eric Cohen, Ibmec, Brazil
Felipe Zambaldi, UMESP, Brazil
Francisco Aranha, FGV, Brazil
Summary
In 2000, São Paulo’s population numbered approximately 10.5 million inhabitants, or roughly 3 million families (IBGE, 2002), out
of which 30% earned below four minimum wages (R$604.00 at the time).
Income has traditionally been adopted as an indicator of life condition and poverty, and it has been used extensively in marketing
research. Nonetheless, one problem is that often individuals are reluctant or unwilling to declare income (Bussab; Ferreira, 1999).
The most commonly used indicator used in the country is the Brazil Criterion, proposed in 1996 by the National Association of Market
Research Firms–ANEP. This criterion is criticized as not characterizing accurately households in the extremities of the income distribution
curve (Mattar, 1996; Silva, 2004).
In this study we used an additional variable, energy consumption, with the objective of improving the family income model. The
choice of this variable is justified first by the fact that virtually all families, regardless of the remoteness of the household, are served by
utility companies. Secondly, collecting electricity consumption is easily done. Third, electricity consumption may reflect the possession
of durable goods (Guerreiro et al., 1996) and is indirectly related to the number of bedrooms and people living in the household
(Pompermayer; Charnet, 1996).
Conceptualization
Despite the multiple and sometimes conflicting definitions of income, class and social class in the literature, Mattar (1995) points
out there are some universal characteristics.
The concept of family and household itself is not free of different perspectives; to be sure, IBGE (2004) and Fundação SEADE (2005)
have slightly different definitions for family and household: SEADE does not consider domestic employees as members of the household,
whereas IBGE does consider them.
The widespread use of income as an indicator for purchasing power in market research is criticized by Mattar (1994, 1996), who notes
that grouping together people with the same income but diverse cultural, educational and aspiration characteristics will result in “averaging
out” purchasing behavior, likes and dislikes of radically distinct individuals.
In Brazil, there was no single and objective socioeconomic classification criterion until 1970 (Mattar, 1996). The usual practice then
by firms in the attempt to use segmentation as a basis for marketing strategy was to develop and use their own customer classification.
But this meant that firms and advertising agencies could not transport concepts, make market comparisons, or use media to target customer
segments across. Finally, the Brazilian Association of Advertisers–ABA introduced in 1970 the first socio-economic criterion in Brazil,
which was subject to criticism. A new criterion emerged with the ABA-ABIPEME criterion (Mattar, 1995), which had easier data
collection and coding procedures, and better representation of the population in the extremes of the distribution (Pereira, 2004; Mattar,
1995).
Later, ABEP–Brazilian Association of Market Research Firms proposed in 2004 the Brazil Criterion. One of the criticisms of this
criterion is its inability to capture life style and social conditions. This criterion uses possession of goods and instruction of the family head,
using a 34-point scale that is finally converted into seven economic classes (A1, A2, B1, B2, C, D and E).
Method
The object of this study is the low income population of the city of São Paulo. Low income families are defined as those with income
up to four minimum wages, including income from all family members. The choice of four wages is justified by thresholds used in the
Brazilian Institute of Geography and Statistics (IBGE, 2004) to characterize low income families. Therefore, the frame of reference is
defined as the set of single family households earning up to four minimum wages, located in nine neighborhoods, living in independent
households, and served by electric utility companies.
443 household interviews were made in August 2005, and information regarding electric consumption, age, education, income,
household type, employment, and possession of durable goods. Respondents were asked to provide the electric bill, and that information
was used to retrieve billing information, resulting in a final sample of 384 households that had usable and accurate information for this
study.
220 / WORKING PAPERS
The study used linear regression, with household income as the dependent variable, and economic classification as the independent
variable together with the electricity consumption variable. The latter was supplied by Eletropaulo, the local utility company. The number
of bedrooms and people living in the household were introduced in the regression as control variables.
The study also used decision trees to complement the analysis and help determine interactions among variables (Berry and Linoff,
1997).
Major Findings
Of the 384 respondents in the sampled in this study, only 253 produced electric bills during the field survey. The bills enabled the
collection of electricity consumption from the utility company. The variables collected were consumption of energy, 12-month
consumption, electric bill amount in local currency, and 12-month amount. Of the 253 households that produced the bill, the average
amount was R$ 59 and the standard deviation was R$ 33.
Income data was collected, with household head income ranging from R$ 458 to R$ 663, and total income averaging R$ 655.
Distribution in terms of the Brazil Criterion was 5% class B, 45% class C, 49% class D and 1% class E.
Initially, a regression used Brazil’s Criterion to predict Total Family Income. A polynomial regression showed higher predictive
power over exponential, logarithmic and logistic regression models (R2=0.222 ANOVA: F=108.89, P-value=0,000).
The next regression incorporated the electric bill consumption as well as the number of inhabitants and the number of bedrooms in
the household, with a resulting R2= 0.300, ANOVA with F=54.34 and P-value=0.000. Stepwise regression indicated that the number of
inhabitants did not significantly contribute to the explanatory power of income variability.
The final regression is given by the following results, with p-values lower than 0.4% and ANOVA p-value 0%. The Variance Inflation
Factor for all variables is under 5, which indicates absence of multicollinearity:
a=
b1 =
b2 =
b3 =
Intercept =
Brazil’s Criterion =
Electric bill amount =
Number of inhabitants =
409.27
-7.752
0.87527
83.59
These results show that the incorporation of the electric consumption and number of bedrooms improved the R2 of the model from
0.222 to 0.300.
In conclusion, results indicate that the relationship between income and predictor variables, with analysis of residuals, contribute in
the understanding of the phenomenon under investigation.
The collection of the electric consumption is done on a regular basis, without data inaccuracies that result from the unwillingness
to disclose personal information (Francisco, 2002).
In addition, information collected from utility companies can also be used to calibrate or validate government demographic data. It
can also be used in planning distribution, a practice carried out by U.S. utility firms (Noonan; Johnson, 2005).
Lastly, another potential benefit is the availability of reliable data for credit decisions, which in emerging economies can minimize
the risk of income informality, and lack of accurate credit behavior information. This can be used by financial institutions and consumer
businesses to achieve better segmentation, and increase company profitability.
References:
ABEP–Associação Brasileira de Empresas de Pesquisa. (2004). Critério de Classificação Econômica Brasil. São Paulo, Brasil:
Access Dec 22, 2004, http://www.anep.org.br.
Almeida, Pergentino M. and Hilda Wickerhauser (1991). O Critério ABA/ABIPEME-em Busca de Uma Atualização. Um estudo e
uma proposta submetidos à ABA e à ABIPEME. São Paulo, Brasil: ABA; ABIPEME.
Berry, Michael J. A. and Gordon S. Linoff (1997). Data Mining techniques: for marketing, sales and customer support. New York:
Wiley Computer Publishing.
Bussab, W. O., Ferreira, M. (1999). Critério Brasil de Estratificação Socioeconômica: Aspectos Demográficos. Anais do CLADEA,
San Juan, Porto Rico, 34.
Chapin, Stuart (1933). The Measurement of Social Status. Minneapolis, MN: University of Minnesota Press.
Coleman, Richard P. and Lee Rainwater (1978). Social Standing in America: New Dimensions of Class. New York: Basic Books.
Francisco, Eduardo (2002). Customer Franchise: A Mina de Ouro do Geomarketing. InfoGEO, 25, ago., 56-57.
Guerreiro, A. G., Serra, S. T., Carvalho, M. L. R. de, Silva Filho, M. C. de. (1996). A influência da Venda de Eletrodomésticos na
Expansão do Consumo Residencial de Energia Elétrica. Anais do Congresso Brasileiro de Energia–CBE, Rio de Janeiro, Brasil,
7, 22-33.
IBGE–Instituto Brasileiro de Geografia e Estatística. (2002). Censo Demográfico 2000. Documentação dos Microdados da Amostra.
Rio de Janeiro, Brasil:
IBGE–Instituto Brasileiro de Geografia e Estatística. (2004). Pesquisa de Orçamentos Familiares, 2002-2003. Primeiros Resultados.
Rio de Janeiro, Brasil:
Krauss, I. (1976). Stratification, Class and Conflict. New York: The Free Press.
Mattar, Fauze N. (1994). Estratificação Socioeconômica e Pesquisas de Marketing. Anais do ENANPAD, São Paulo, Brasil, 18, v. 7,
307-325. Access 13 June 2005, http://fauze.com.br.
Mattar, Fauze N. (1995). Estudo sobre Estratificação Social para Utilização em Marketing e Pesquisas de Marketing. Proposta de
Novo Modelo para Estratificação Sócio-Econômica. Estudo para Título de Livre-Docência, Faculdade de Economia,
Administração e Contabilidade da Universidade de São Paulo, São Paulo, Brasil.
Latin American Advances in Consumer Research (Volume 2) / 221
Mattar, Fauze N. (1996). Porque os Métodos de Classificação Socioeconômicos utilizados no Brasil não funcionam. Anais
eletrônicos do ENANPAD, São Paulo, Brasil, 20. Access 13 jun. 2005, http://fauze.com.br.
Noonan, Jessica C. and Amy L. Johnson (2005). GIS boosts T&D planning for Asset Management. Transmission & Distribution
World, Oct., 28-32.
Pereira, Vinicius R. (2004). Métodos Alternativos no Critério Brasil para Construção de Indicadores Sócio-Econômicos: Teoria de
Resposta ao Item. Dissertação de Mestrado, Pontifícia Universidade Católica do Rio de Janeiro, Rio de Janeiro, Brasil.
Pompermayer, Máximo L. and R. Charnet (1996). Determinantes da Demanda Residencial de Energia Elétrica. Anais do Congresso
Brasileiro de Energia–CBE, Rio de Janeiro, Brasil, 7, 102-115.
Schaninger, Charles M. (1981). Social Class Versus Income Revisited: An Empirical Investigation. Journal Of Marketing Research,
XVIII, May, 192-208.
SEADE–Fundação Sistema Estadual de Análise de Dados. (2005). Pesquisa de Condições de Vida: Apresentação da Base das
Variáveis do Questionário e dos Indicadores Gerados. São Paulo, Brasil:
Silva, N. L. (2004). Critério Brasil: o mercado falando a mesma língua. Access 22 Dec. 2004, http://www.anep.org.br.
Automatic Lover2: Linking Consumer Practice to Cultural Texts about the Vibrator
Luciana Walther, Coppead Graduate School of Business, Brazil
This working paper presents the early stages of an exploratory interpretive study on an emerging type of Brazilian woman as a more
overt consumer of sexual products. By using Hirschman, Scott and Wells’ Model of Product Discourse (1998), we analyze consumption
practices in comparison to cultural texts about the vibrator. We intend to further understand, with a sociological approach, how women
in Brazil create and negotiate meaning through sexually-related consumption; and to examine, with a marketing approach, the
effectiveness of product placement and of stealth branding as strategies to market sex to women. Managerial recommendations are also
provided.
Conceptualization
People are having sex. At this very moment, millions of people are either engaging in sexual intercourse or planning to (Mackay 2000).
And while they are at it, it is very likely that some sort of consumption might be involved. Be it of preservatives, lubricants, magazines,
toys, erectile dysfunction drugs or the Internet, in present day western societies, one can hardly have sex and not be a consumer at the same
time. So why does literature on consumption rarely address sexuality?
Although sexuality has been an important topic in the sociology literature for decades, it has to date received relatively little attention
in the marketing field. The existing marketing research that, to some extent, relates to sexuality has more frequently focused on topics like
homosexuality, gender issues, hedonism, desire, and sex in advertising (Hirschman and Holbrook 1982; Stern 1991; Elliot and Ritson
1995; Kates 1998, 2002; Reichert and Lambiase 2002, Belk, Ger and Askegaard 2003), while the actual consumption of sex products and
services has been addressed by very few (Gould 1991; Kent 2005; Langer 2007).
The idea for this exploratory study arose during a previous research on the emergence of a more overt female consumer of sexual
products (Walther 2007). The ethnographic method was used then, comprising in-depth interviews and participant observation at womenonly sex shops in Rio de Janeiro, Brazil, which is popularly and scholarly known as a highly sexualized culture (Parker 1991). During the
interviews with shop attendants, an episode of North-American TV series “Sex and the City” was repeatedly mentioned as the reason for
which the so-called Rabbit vibrator became the best sold item. Therefore, studying the vibrator as a commoditized product seemed like
a promising path of investigation.
Thus, the aim of this study is twofold: (1) to further understand, with a sociological approach, how women in Brazil create and
negotiate meaning through sexually-related consumption (Gould 1991); and (2) to examine, with a marketing approach, the effectiveness
of product placement (Balasubramanian 1994) and of stealth branding (Holt 2002) as strategies to market sex to women.
Method
In this study, we applied Hirschman, Scott and Wells´ Model of Product Discourse (1998) to consumption practices and cultural texts
about the vibrator. This model assumes that, in order to acquire meaning, any good must go through a process that will culturally construct
it as a symbol and hierarchically place it amidst a product constellation (McCracken 1988) or object system (Baudrillard 1968). Therefore,
understanding the process, that is, how goods acquire meaning, is as important as understanding what they mean. This is why data analysis
through the Model of Product Discourse may help reaching the aims of this study.
Hirschman, Scott and Wells used data from primetime commercials and programs. They also identified the influence of history on
contemporary product meaning. Here we analyzed three Sex and the City episodes–the ones in which a vibrator played a relevant part–
comparing them to practices, as described by respondents during our previous study (Walther 2007). Later, we intend to use, as non-fiction
texts, articles from a Brazilian women’s magazine that focuses on sexuality.
After comparing text to practice, we arrived at several interpretive questions that equated scenes from Sex and the City with issues
regarding Brazilian sexual identity. Then it was time to resort to history, referring to works like Laqueur’s “Making Sex” (1990) and
Brazilian social scientist Gilberto Freyre’s seminal book “Casa-Grande e Senzala” (1933). This step provides a conceptual map for
consumption scenes and helps answering questions about how things came to be the way they are.
During all steps of this research, it was important to identify cultural and legal differences between the USA and Brazil, and how they
reflect on practices and representations of sexually-related consumption.
2
“Automatic Lover” is the title of a 1978 disco song by Dee D. Jackson, in which a woman falls in love with a robot.