(This is a sample cover image for this issue. The actual cover is not yet available at this time.)
This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party websites are prohibited.
In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information regarding Elsevier's archiving and manuscript policies are encouraged to visit:
http://www.elsevier.com/copyright

Lifecycle effects on consumer ﬁnancial product portfolios in South
Africa: An exploratory analysis of four ethnic groups
Mthunzi A. Ngwenyaa,1, Leonard J. Paasb,⇑
aAfrican Infrastructure Investment Managers (Pty.) Ltd., Colinton House, The Oval, 1 Oakdale Road, Newlands, 7700 Cape Town, South AfricabDepartment of Marketing, Faculty of Economics and Business Administration, VU University Amsterdam, De Boelelaan 1105, 1081 HV Amsterdam, The Netherlands
article info
Article history:
Received 11 February 2011
Received in revised form 12 September
2011
Accepted 15 September 2011
Available online 19 September 2011
JEL classiﬁcation:
C49
F39
PsycINFO classiﬁcation:
2240
3900
Keywords:
Consumer psychology
Saving behavior
International ﬁnance
abstract
This paper assesses ownership of 16 ﬁnancial products by households in different lifecycle
stages amongst four ethnic groups (Africans, Coloureds, Asians, and Whites) in South
Africa. The lifecycle hypothesis indicates younger households should own more debt-
related ﬁnancial products, whereas households in intermediate lifecycle stages should
own more ﬁnancial products to accumulate assets; both these claims are disconﬁrmed
for all groups. However, White households in intermediate household stages own more
ﬁnancial products than younger and older households, consistent with previously reported
lifecycle ﬁndings in Western countries. Consistent with the literature on innovation adop-
tion we ﬁnd that younger, afﬂuent and highly educated households amongst the other
three ethnic groups tend to own more ﬁnancial products than older Africans, Coloureds
and Asians. These results indicate that innovation adoption literature may better describe
ﬁnancial product ownership in developing countries than the lifecycle hypothesis.
2011 Published by Elsevier B.V.
1. Introduction
Consumer ﬁnancial product portfolios are combinations in which individuals own products such as checking accounts,
saving accounts, loans and investment products. Such product portfolios have received considerable attention in the extant
literature (e.g.,Gunnarson & Wahlund, 1997; Kamakura, Ramaswami, & Srivastava, 1991; Stafford, Kasulis, & Lusch, 1982).
This is because combinations in which products are owned can be useful for gaining insight into household’s ﬁnancial strat-
egies (Gunnarson &Wahlund, 1997), for detecting consumers that are interested in speciﬁc ﬁnancial products (Kamakura
et al., 1991; Paas & Molenaar, 2005) and for assessing the priorities that consumers have for owning various ﬁnancial prod-
ucts (Paas, 1998; Stafford et al., 1982).
In this paper we assess applicability of the lifecycle hypothesis and ﬁndings from the literature on the adoption of
innovative products and services for explaining ﬁnancial product ownership in a country with a developing economy, South
Africa. The lifecycle hypothesis has been the prominent theory in explaining borrowing and saving behavior of consumers.
However, the lifecycle hypothesis has traditionally been applied to explain ﬁnancial behavior in Western countries
0167-4870/$ - see front matter2011 Published by Elsevier B.V.
doi:10.1016/j.joep.2011.09.008
⇑Corresponding author. Tel.: +31 20 5986003; fax: +31 20 5989870.
E-mail addresses:mthunzi.ngwenya@macquarie.com,mthunzi.ngwenya@gmail.com(M.A. Ngwenya),l.j.paas@vu.nl(L.J. Paas).
1Tel.: +27 82 924 0543.
Journal of Economic Psychology 33 (2012) 8–18
Contents lists available atSciVerse ScienceDirect
Journal of Economic Psychology
journal homepage: www.elsevier.com/locate/joep

(Browning & Lusardi, 1996; Wärneryd, 1999), which we deﬁne as Northern America, Western Europe, Australia and New
Zealand. It remains unclear whether the lifecycle hypothesis applies outside Western countries. In the theoretical section
of this paper, Section2, we propose that literature on innovation adoption may also explain ﬁnancial product ownership out-
side Western countries, as ﬁnancial products in newly emerging economies can be considered as innovations for most con-
sumers (Roos et al., 2005; Steenkamp & Burgess, 2002). Note that we provide one of the few studies of consumer ﬁnancial
behavior outside Western countries (cf.Davies, Easaw, & Ghoshray, 2009; Roos et al., 2005). Researching the relevance of
theories, which were developed on data from Western countries only, in newly emerging economies not only provides in-
sights into the generalizability of such theories but also knowledge about the conditions in which such theories apply or not
(Burgess & Steenkamp, 2006).
In an empirical study, we analyze a data set of 18,965 respondents from the four main ethnic groups in South Africa: Afri-
cans, Coloureds, Asians, and Whites - we use the generally accepted terms for these ethnic groups in South Africa, according
to the South African Bureau of Statistics;www.statssa.gov.za, see alsoSteenkamp and Burgess (2002). The data indicate, for
each respondent, ownership of seven products used for borrowing purposes, such as loans and mortgages, and nine products
for saving or investing, such as savings accounts and stocks. Using data on age, marital status, and parental status, we deﬁne
the lifecycle stages in line withMurphy and Staples’s (1979)lifecycle operationalization, with necessary adjustments to de-
scribe the South African context. We then apply the lifecycle phase of each respondent to predict household ﬁnancial product
portfolios among Africans, Coloureds, Asians, and Whites according to a segmentation technique, latent class analysis (LCA;
Fernandez-Blanco, Orea, & Prieto-Rodriguez, 2009; Oppewal, Paas, Crouch, & Huybers, 2010; Wedel & Kamakura, 2000). In
our application, the LCA technique simultaneously segments households on the basis of their ﬁnancial product portfolios and
estimates the effects of household lifecycle phase and ethnic group membership on the likelihood for ﬁnancial product port-
folios to occur. On the basis of our ﬁndings, we consider the applicability of the lifecycle hypothesis in developing economies
or, alternatively, the ﬁndings from the innovation adoption literature that are discussed below.
Accordingly, in Section2, we discuss the lifecycle hypothesis and its applicability in South Africa. In Section 2 we also
discuss relevance of ﬁndings from the innovation adoption literature for ﬁnancial product ownership in South Africa. In Sec-
tion 3 we introduce our data set, describe the revised lifecycle coding that we apply, and discuss the conducted LCA analysis.
Section4contains the results of this analysis, including the household ﬁnancial product portfolios we found and the effects
of lifecycle stage, ethnic group, educational level, and income on the occurrence of these product portfolios. We conclude
with a discussion and implications of our ﬁndings, in Section5.
2. Household lifecycle phase and ﬁnancial product ownership in South Africa
2.1. The lifecycle hypothesis in South Africa
The lifecycle hypothesis (Modigliani & Brumberg, 1954) and the related permanent income hypothesis (Friedman, 1957)
have been prominent theories for describing, explaining, and predicting ﬁnancial behavior (Browning & Lusardi, 1996;
Wärneryd, 1999). Their basic assumption indicates that households avoid income ﬂuctuations across their lifetimes to sus-
tain a constant consumption pattern. As a result, young households should have debts, households in intermediate lifecycle
phases save and invest to pay off debts and build a reserve for their old age, and older households tend to spend assets that
they accumulated earlier in life (Browning & Lusardi, 1996; Wärneryd, 1999).
Various studies in Western countries show that the variables that constitute the lifecycle hypothesis, such as age, marital
status, and having children, all inﬂuence the household’s ﬁnancial product portfolio (e.g.,Gunnarson & Wahlund, 1997; Paas,
Bijmolt, & Vermunt, 2007; Soutar & Cornish-Ward, 1997; Stanley, Ford, & Richards, 1985; Tin, 2000). Households in the inter-
mediate lifecycle stages own ﬁnancial products for saving or investing purposes, which is in line with the lifecycle hypothesis
(Browning & Lusardi, 1996; Wärneryd, 1999). Findings with regard to debts indicate that products such as loans and mort-
gages are more prominent among younger households (Gunnarson & Wahlund, 1997; Stanley et al., 1985). Furthermore,
households in intermediate lifecycle phases own more ﬁnancial products than others, i.e., they are ﬁnancially the most active
groups (Gunnarson & Wahlund, 1997; Paas et al., 2007; Soutar & Cornish-Ward, 1997).
It remains unclear whether the ﬁndings regarding ﬁnancial product ownership across different lifecycle phases also ap-
ply to non-Western countries. Theories developed using only data from Western countries do not necessarily apply in
newly emerging markets (Burgess & Steenkamp, 2006; Steenkamp, 2005.Deaton (1989)andGersovitz (1988)suggest sev-
eral reasons the lifecycle affects may differ. First, in developing countries, households usually are indecomposable units
that survive beyond the individual members, implying that ﬁnancial decisions occur at the household level rather than
the individual level. Saving for retirement becomes less relevant under such circumstances, because children will take care
of the elderly members of a family. Second, households in developing countries have lower and less certain incomes, so it
may be very difﬁcult, if not impossible, for them to save. Third, borrowing constraints may be stronger for low-income
households in developing nations, and thus, borrowing also becomes more difﬁcult if not completely impossible. Fourth,
savings may serve as a buffer against uncertain and unpredictable events, instead of for inter-temporal consumption
smoothing.
The conditions in developing countries that lead to a questionable applicability of the lifecycle hypothesis apply most
strongly to Africans and least to Whites in South Africa (Nga, 2007). The ethnic groups in South Africa have different incomes,
M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–189

are culturally diverse, and have highly divergent historical backgrounds (Burgess & Steenkamp, 2006; Nga, 2007; Roos et al.,
2005; Steenkamp & Burgess, 2002). In terms of income, Whites have the highest incomes, then Asians and Coloureds; Afri-
cans have the lowest incomes (Casale & Desmond, 2007). Income levels affect household ﬁnancial product portfolios of
households in Western countries (Paas et al., 2007; Soutar & Cornish-Ward, 1997), and we predict they may have similar
inﬂuences in South Africa. With regard to cultural background, Africans in particular often invest some of their assets in
informal institutions, such as Stokvels (Finscope, 2009) based on the African principle ofubuntu, which encompasses reci-
procity and socially shared responsibility for the ﬁnancial distress of any member, such as when there is death in a family
(Nga, 2007). Moreover, members from historically disadvantaged ethnic groups lack an opportunity to extract knowledge
from direct observation of their parents, the mass media, or conversations with other adults. They thus may ﬁnd themselves
in situations in which they have the ﬁnancial means to acquire ﬁnancial products and services but lack an understanding of
the implications of these products (Roos et al., 2005).
2.2. Consumer innovativeness and ﬁnancial product ownership
In Section 2.1, we pointed out that lifecycle based predictions may not hold for the previously disadvantaged groups in
South Africa: Africans, Coloureds and Asians. Most relevant herein is that ﬁnancial products and services are relatively new
for many Africans, Coloureds and Asians in South Africa, as a result of its history of apartheid (Roos et al., 2005). In fact,
Steenkamp and Burgess (2002)used ﬁnancial product ownership to assess consumer innovativeness in South Africa. There-
fore, we conjecture that from the perspective of Africans, Coloureds and Asians ﬁnancial products can be considered as inno-
vations, i.e., ideas, practices or objects perceived to be new by a person or another adopting entity (Chakravarty & Dubinsky,
2005; Rogers, 1995).
When an innovation is introduced to a market there are innovators, who adopt the product very early, and others who
adopt the product later. Innovators generate the initial levels of penetration of the products and also inﬂuence others by
using the product, possibly displaying it and encouraging others to try the product. Researchers have investigated many
antecedents of innovativeness (Chakravarty & Dubinsky, 2005). Socio-demographics play a role in the adoption of innova-
tions. It has been found that young, relatively wealthy, and relatively highly educated consumers often adopt innovations
before others (Arts, Frambach, & Bijmolt, 2011; Gatignon & Robertson, 1985; Im, Bayus, & Mason, 2003; Steenkamp &
Gielens, 2003). Innovators are relatively wealthy, because the wealthy can better absorb the loss resulting from the acqui-
sition of a failing innovation (Rogers, 1995). That is, not all innovations are successful and wealthy consumers can take the
risk that they allocate resources to an unsuccessful innovation. Furthermore, innovators are less dogmatic, more rational and
more intelligent (Rogers, 1995). These are characteristics of the higher educated. Next to this innovators have more favorable
attitudes towards change (1995), which is an attribute distinguishing younger individuals from their older counterparts
(Musteen, Berker, & Baeten, 2006).
Assuming that young, relatively wealthy and highly educated consumers are more likely to adopt ﬁnancial products
that are perceived to be relatively new is not only consistent with the general literature on innovation adoption, but also
with the literature on the adoption of novel ﬁnancial products and services. In a series of previous studies the acquisition
of relatively new ﬁnancial products and services was studied from the perspective of innovation adoption (e.g.,Black,
Lockett, Winklhofer, & Ennew, 2001; Hoffmann & Broekhuizen, 2010; Lockett & Littler, 1997). The most recent of these
studies shows that higher levels of afﬂuence and higher educational levels increase the probability of owning new invest-
ment products (Hoffmann & Broekhuizen, 2010). Furthermore, in an empirical investigation of the Turkish consumer’s
acceptance of Internet banking services (Polatoglu & Ekin, 2001, p. 161), which is an innovation in the ﬁnancial domain,
it was found that young, afﬂuent and highly educated groups accept this new service more readily. Another type of inno-
vation, direct banking, was also found to be adopted more readily by young and afﬂuent consumers in the UK (Lockett &
Littler, 1997).
Based on the discussion above, we propose that young, relatively wealthy, relatively highly educated Africans, Colour-
eds and Asians may own more ﬁnancial products than other members of these three ethnic groups. That is, due to afﬁr-
mative action policies, which require companies to meet minimum employment percentages of historically disadvantaged
groups in their employment proﬁles (South Africa, 1998), the emerging situation allows younger Africans, Coloureds, and
Asians to earn incomes that their older counterparts did not earn at their ages, even after taking inﬂation into account
(Maisonnave, Decaluwé, & Chitiga, 2009). Unlike Whites, who always had access to these employment and income oppor-
tunities, such that cohort behavior across different generations likely has reached a stationery state, the non-White
groups, especially Africans, indicate signiﬁcant changes across generations. Besides, this younger Africans, Coloureds
and Asians have had more access to educational institutions than the older members of their ethnic groups (Maisonnave
et al., 2009).
In sum, the lifecycle based prediction that young households own more credit related ﬁnancial products and interme-
diate aged households own more ﬁnancial products, particularly products for saving and investing, may hold for white
South African consumers. For the three other ethnic groups, Africans, Coloureds, and Asians, it is could also be the case
that younger households own more ﬁnancial products than the older households of the same ethnic group. This latter ten-
dency would be consistent with the literature on innovation adoption and inconsistent with literature on the lifecycle
hypothesis.
10M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–18

3. Methodology
3.1. Data
The analyzed database comes from the All Media and Products Survey and was collected between July 2007 and June
2008 (SAARF, 2008). The survey covered both the purchase of media products and a wide range of non-media products,
including ownership of ﬁnancial products. The survey population in South Africa was anyone aged 16 years or older, in both
urban and rural areas. All questions dealing with demographic factors and ownership of ﬁnancial services were close-ended.
The total sample size is 19,755 respondents. We only analyzed people aged 18 years or older, resulting in a sample of 18,965
respondents.
The sample is not representative for the South African population. Most notably, Africans are underrepresented, as are
individuals with lower income and educational levels. The sample is representative with regard to gender, about 50% of
the sample is female, which reﬂects the South African population. Weighting has been applied to correct for discrepancies
between the sample and the South African population. Moreover, because the sample is large and heterogeneous across the
target population, we conclude that after the application of weighting it represents various segments across all four ethnic
groups.
We list the product ownership percentages for each of the ﬁnancial products by the four ethnic groups inTable 1.
2As
Table 1shows, Africans tend to have lower propensities to own each of the ﬁnancial products, Coloureds and Asians take inter-
mediate positions, and Whites have the highest propensities for owning all products. Members of some ethnic groups, Africans
in particular, may own products in an informal banking sector (Cross, 1987).Finscope (2009)estimates that about 10% of South
African adults rely only on informal products, such as Stokvels and Mashonisa (an informal, usually unregistered money lender
who lends small amounts). Thus, the product penetrations reported inTable 1may underestimate ﬁnancial product ownership
to some degree.
3.2. Coding the household lifecycle phase
Data on age, marital status, and dependent and independent children are available in the analyzed database. We used
these data to allocate the interviewed households into 13 categories, according toMurphy and Staples’s (1979)household
lifecycle operationalization, as we show inTable 2. Note that theMurphy and Staples (1979)approach has not been without
criticism (e.g.,Derrick & Lehfeld, 1980; Wagner & Hanna, 1983). Some have noted that it does not add much to a simple
categorization of households according to age (e.g.,Wagner & Hanna, 1983). However, in research concerning the ownership
of ﬁnancial assets it was found that the other lifecycle variables, next to age (e.g., presence of children in the household,
marital status), signiﬁcantly affect ﬁnancial product ownership (Baek & Hong, 2004; Tin, 2000; Xiao, 1996). Nevertheless,
in our application we found that theMurphy and Staples (1979)lifecycle model needed adjustments. Using their precise
operationalization led to missing values regarding the household lifecycle for 22% of the respondents in our data set—that
is, these households could not be placed in any of the 13 categories inTable 2.
We ﬁnd that most of the missing values consist of (1) young (<35 years) singles with children, (2) middle-aged
(35–64 years) singles with children, and (3) middle-aged singles without children. Hence we add these three categories
Table 1
Product ownership across ethnic groups.
Product African (%) Coloured (%) Asian (%) White (%) Overall (%)
Micro loan 0.3 0.4 0.2 0.2 0.3
Vehicle ﬁnance 1.5 2.9 2.7 6.9 3.4
Mortgage 1.5 4.2 2.9 7.3 3.7
Student loan 0.3 0.4 0.4 0.8 0.5
Credit card 3.7 7.5 10.3 24.9 11.1
Overdraft facility 1.1 2.0 3.0 7.6 3.3
Other loan 0.6 2.2 0.9 1.3 1.1
Mzansi 4.7 1.3 0.2 0.5 2.6
Savings account 54.6 59.6 62.3 70.1 60.5
Investment trust or mutual fund 2.8 4.2 4.9 11.3 5.7
Stocks 0.8 1.2 2.4 4.9 2.2
Retirement annuity 4.8 6.9 6.3 17.4 9.0
Endowment investment saving – no life insurance 1.4 2.4 2.2 4.4 2.5
Endowment investment saving – with life insurance 3.4 5.9 7.1 12.7 6.8
Life coverage policy 10.6 16.7 16.9 32.5 18.5
Shares 1.1 1.6 1.9 8.2 3.4
2The Mzansi Account is a low income, transactional banking account, developed to fulﬁll the commitments of South Africa’s Financial Sector Charter.
Financial ﬁrms are required to enhance people’s access to ﬁnancial products and services and speciﬁcally to increase the reach of banking services toall
communities, including those that historically have been underserved.
M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–1811

toMurphy and Staples’s (1979)classiﬁcation. Furthermore, we deleted four categories inTable 2, referring to divorce status:
young divorced with children (3c inTable 2), middle-aged divorced without children (4b), middle-aged divorced with young
children (4d), and middle-aged divorced without dependent children (4f). These categories are highly uncommon in our
database. Reported numbers of divorced women in South Africa have been historically low and between 2.2% and 3.2% since
1996; the reported ﬁgure for men has been even lower and in the range of 1.2% and 1.9% (StatisticsSouth Africa, 2003,
2007).
3
We present the revisedMurphy and Staples (1979)lifecycle operationalization inTable 3, in which 97.8% of the respon-
dents can be categorized. The incorporation of single parents and middle-aged singles may be an adjustment necessary in
Western countries as well, which also have experienced increases in single person/parent households (Michael, Fuchs, &
Scott, 1980; Quintano & D’Agostino, 2006).
3.3. Latent class analysis
We apply the LCA model for dichotomous response variables (Vermunt & Magidson, 2007; Wedel & Kamakura, 2000). To
deﬁne this model, consider a sample of respondents, denotedi=1,...,I. For each respondenti, we record whether they own a
speciﬁc ﬁnancial productk,k=1,...,K. The complete product portfolio of respondentiacross the 16 ﬁnancial products dis-
cussed is represented by the vectorY
i. In this vector, elementyiktakes a value of 1 when individualiowns productk, and 0
otherwise. We assume a limited number of segments, denoteds=1,...,S, also referred to as latent classes. The vectorX
irep-
resents segment membership probabilities of respondenti, andP(X
i=s) is the probability that respondentibelongs to a spe-
ciﬁc segments. These probabilities are determined by the observed product ownership patterns inY
i. Respondents with
highly similar product portfolios are more likely to belong to the same segment, compared with those with more dissimilar
product portfolios. Covariates also can be incorporated in the model. The vector of values respondentihas for these covar-
iates is deﬁned asZ
i. The model is formally deﬁned as
Table 3
South African household lifecycle phases.
StageAge Marital status Children
1a. Young single without children <35 Single No
1b. Young single with children <35 Single Yes
2. Young married without children <35 Married No
3. Young married with children <35 Married Yes
4a. Middle-aged single without children 35–64 Single No
4b. Middle-aged single with children 35–64 Single Yes
4c. Middle-aged married without children 35–64 Married No
4d. Middle-aged married with young and adolescent children 35–64 Married Young and adolescent
4e. Middle-aged married without dependent children 35–64 Married Independent
5a. Older married >64 Married Not speciﬁed
5b. Older unmarried >64 Divorced Widowed Not speciﬁed
Table 2
Murphy and Staples’s (1979)household lifecycle phases.
StageAge Marital status Children
1. Young single <35 Single No
2. Young married without children <35 Married No
3a. Young married without children <35 Married No
3b. Young married with children <35 Married Yes
3c. Young divorced without children <35 Divorced No
4a. Middle-aged married without children 35–64 Married No
4b. Middle-aged divorced without children 35–64 Divorced No
4c. Middle-aged married with young and adolescent children 35–64 Married Young and adolescent
4d. Middle-aged divorced with young and adolescent children 35–64 Divorced Young and adolescent
4e. Middle-aged married without dependent children 35–64 Married Independent
4f. Middle-aged divorced without dependent children 35–64 Divorced Independent
5a. Older married >64 Married Not speciﬁed
5b. Older unmarried >64 Divorced widowed Not speciﬁed
3Ziehl (2002)traced the divorce rate per 1000 South African Whites from 1920 to 1996 and showed that it had been steadily increasing, from 0.47% in 1920,
to a peak of 3.97% in 1990, then down to 3.57% in 1996. The reported percentage of divorced women may be an underestimate, because women from families
whose tradition looks down on divorce may not report themselves as divorced (Hosegood, McGrath, & Moultrie, 2009), and those with failed marriages may not
divulge this information, in the belief that never being married would increase their chances of remarriage. Nevertheless, the divorced group remains a small
minority in South Africa, which supports our decision to replace the divorced categories within an ‘‘other’’ category.
12M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–18

PðYijZiÞ¼XS
s¼1PðXi¼sjZiÞYK
k¼1PðYikjXi¼sÞ:ð1Þ
Eq. (1) is a model forP(Yi|Zi). It refers to the probability density relevant for the occurrence of a particular ﬁnancial prod-
uct portfolio for individuali,Y
i, assuming that this person has valuesZion the covariates. In turn,P(Xi=s|Zi) refers to the
probability of belonging to segments, given the values for personion the vector of covariatesZ
i. Moreover,P(Yik|Xi=s)is
the probability that personiowns productk, given this person’s probability to be in segments, X
i=s.
Allocation to segments is probabilistic, implying a respondentimay have nonzero segment membership probabilities for
multiple segments. The probability that respondentiis a member of segmentsis calculated as
PisðdsÞ¼expðdsYiÞ
PS
s0¼1expðds0YiÞ;ð2Þ
wheredsis the vector of parameters to be estimated. The conventional modal assignment rule allocates respondents to a
single segment, such that respondentienters the segment for which s/he has the highest membership probability (Wedel
& Kamakura, 2000).
The model incorporates the effects of the covariates on the probability of belonging to a particular segments, using a mul-
tinomial logit equation (Vermunt & Magidson, 2007). For the covariate effects, the linear model for the log of the ratio of the
probability of being in latent classs, relative to being in the reference classS, takes the following form:
logPðXi¼sjZi
PðXi¼SjZi

¼
csþXC
c¼1csczic;for 16s6S1;ð3Þ
wherecsdenotes an intercept;csc,16c6C, is the slope for thecth covariate; andzicis the value for respondention thecth
covariate. Because the distribution of the log of the ratio of the probabilities in Eq. (3) is known, we can test the signiﬁcance
of covariate effects (Vermunt & Magidson, 2007).
We estimated the LCA model parameters using maximum likelihood estimation. Maximization of the likelihood function
was realized by the EM algorithm (Dempster, Laird, & Rubin, 1977), implemented in Latent Gold 4.5 (Vermunt & Magidson,
2007). The relative ﬁt of alternative model speciﬁcations is evaluated for different numbers of segments using the minimum
Bayesian Information Criterion (BIC) and Constant Akaike Information Criterion (CAIC) rules (Wedel & Kamakura, 2000). An-
other important criterion is whether the solution can be interpreted and facilitates insight.
We ran three models that included various sets of covariates. In the ﬁrst model, we included one variable to represent
both the ethnic group of respondentiand his or her household lifecycle phase, such as ‘‘African single without children’’.
In the second model, we incorporated ethnic group and household lifecycle phases as separate variables to assess their sig-
niﬁcance independently. In the third model, we included household income and educational level as control variables for
ethnic group and household lifecycle phase. These control variables are highly relevant for ﬁnancial behavior (Browning
& Lusardi, 1996; Wärneryd, 1989, 1999), and people across South Africa’s ethnic groups and lifecycle stages may differ on
these important variables (Casale & Desmond, 2007; Statistics South Africa, 2007). The aim of the third analysis thus is to
assess whether the effects of ethnic group and household lifecycle remain signiﬁcant even when we include these other rel-
evant variables as controls.
4. Results
4.1. South African ﬁnancial product portfolios
We ran models with 1–10 latent classes. To avoid local minimums, we ran each of the models 10 times, using 25 different
starting sets each time. The minimum values for both the BIC and CAIC statistics emerged from a four-class model,
BIC = 110,097 and CAIC = 110,305. The four-class model can be interpreted clearly, and the covariate effects, as we discuss
in Section4.3, are also interpretable. Therefore, the four-class solution has high face validity.
Table 4deﬁnes the four prototypical household ﬁnancial product portfolios across the four analyzed ethnic groups. To
illustrate our interpretation of the model, consider the ﬁrst cell in the segment 1 column, according to which 0% of the house-
holds in segment 1 own micro loans.
Using their segment-speciﬁc product penetrations, we label the segments inTable 4as follows:
1. Inactives: Generally do not own ﬁnancial products. They are 39% of the sample.
2. Savers: Generally own savings accounts only. They are 32% of the sample.
3. Savings + 1: Often own a savings account and a single additional product, generally life insurance, a credit card, or a
retirement annuity. They are 23% of the sample.
4. Actives: Relatively high propensities to own all ﬁnancial products, except Mzansi. They are 5% of the sample.
The average South African is not highly active in the ﬁnancial market. Note that this supports the notion ofRoos et al.
(2005)andSteenkamp and Burgess (2002)that ﬁnancial products are innovations to many South Africans, as previously
M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–1813

discussed in the current paper. In particular, 39% appear in a segment characterized by a general tendency to own none of the
products under research. Furthermore, a savings account is the only product commonly owned in single-product portfolios
of savers inTable 4, 32% of the sample. According to acquisition pattern analysis theory, this latter ﬁnding in an LCA model
implies that savings accounts generally are acquired before the other ﬁnancial products that we research (Bijmolt, Paas, &
Vermunt, 2004; Paas et al., 2007), which is similar to the results in studies conducted in Western countries, where house-
holds also acquire savings accounts before other ﬁnancial products (Paas, 1998; Paas et al., 2007; Soutar & Cornish-Ward,
1997). Thus, the South African context does not seem to inﬂuence this tendency. Furthermore, the segmentation solution
mostly reﬂects different levels of ﬁnancial activity in terms of product ownership; that is, the number of ﬁnancial products
owned is the distinguishing factor across different South African households, not the tendency to borrow or save. This differs
strongly from a previous study on household ﬁnancial product portfolios conducted in Sweden (Gunnarson & Wahlund,
1997). The Swedish study found ﬁnancial product portfolios reﬂecting different ﬁnancial strategies.
4.2. A comparison of alternative model speciﬁcations
As we noted in Section3.3, we ran three alternative models, including different covariates. In each model the segmenta-
tion structure reported inTable 4was included, with variation only in the incorporated covariates. The covariate effects in
LCA models assess the inﬂuence of exogenous variables on the probabilities of respondents being allocated to various
segments.
In the ﬁrst model we included one variable representing both the ethnic group and household lifecycle. The effect of this
single covariate is signiﬁcant (d.f. = 30, Wald = 2203.465,p<0.001). We ran a second model to assess whether the signiﬁ-
cance of the variable representing the interaction between ethnic group and lifecycle stage results from the signiﬁcance
of both factors or just one of the two variables. Therefore, in the second model, ethnic group and household lifecycle are
incorporated as separate variables, both of which have signiﬁcant effects on the segment membership probabilities of indi-
vidual respondents (ethnic group: d.f. = 3, Wald = 1972.160,p<0.001; lifecycle stage: d.f. = 10, Wald = 886.392,p<0.001). In
the third model we included household income and educational level as control variables, next to ethnic group and house-
hold lifecycle. Both control variables signiﬁcantly affect segment membership probabilities (household income: d.f. = 7,
Wald = 903.020,p<0.001; educational level: d.f. = 1, Wald = 1282.796,p<0.001). Higher incomes and higher educational
levels imply a greater probability of appearing in the savings + 1 or active segments. The effects of ethnic group and house-
hold lifecycle remain signiﬁcant after we include these control variables (ethnic group: d.f. = 3, Wald = 396.552,p<0.001;
household lifecycle stage: d.f. = 11, Wald = 929.360,p<0.001). This is an important ﬁnding, as it implies that the effects
of lifecycle and ethnic group membership on allocation to segments are not explained by multicollinearity with income
and/or educational level.
4.3. Ethnic differences in household lifecycle effect on ﬁnancial product portfolios
We report on the allocation of members from the different ethnic groups to the segments inTable 4, according to the
covariate effect in the ﬁrst model, which we discussed in Section4.2.Table 5Areports lifecycle stage-speciﬁc segment mem-
bership probabilities for Africans. InTable 5A, we also proﬁle each lifecycle group according to average income in thousands
of Rands and the percentage with at least a high school education. The most active groups among Africans are (1) young mar-
Table 4
Segment-speciﬁc product penetrations.
Product Cluster 1 Cluster 2 Cluster 3 Cluster 4
Inactives Savers Savings + 1 Actives
(39%) (32%) (23%) (5%)
Micro loan 0% 0% 1% 2%
Vehicle ﬁnance 0% 0% 4% 47%
Mortgage 0% 0% 5% 46%
Student loan 0% 0% 1% 1%
Credit card 0% 0% 28% 78%
Overdraft facility 0% 0% 5% 39%
Other loan 0% 0% 2% 9%
Mzansi 6% 0% 1% 0%
Savings account 24% 98% 73% 55%
Investment trust or mutual fund 0% 1% 14% 40%
Stocks 0% 0% 5% 19%
Retirement annuity 0% 1% 23% 63%
Endowment investment – without life ins. 0% 1% 6% 15%
Endowment investment – with life insurance 0% 2% 17% 45%
Life cover policy 1% 10% 50% 63%
Shares 0% 1% 6% 31%
Average # of products per segment member 0.31 1.14 2.41 5.53 14M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–18

ried without children, (2) young married with children, and (3) middle-aged married with young and adolescent children.
Households in these lifecycle phases are relatively less often in the inactive segment. We shade the relatively active groups
to enhance interpretations ofTable 5A. The three shaded lifecycle phases all have average incomes of at least 7000 Rand. The
only other lifecycle phase with this average income level or higher is the middle-aged, married, independent children only
group. The average educational level of this group, however, is lower than that of the three relatively active lifecycle groups.
We posit that age, income level and educational level are all relevant for ﬁnancial product ownership at different lifecycle
stages among Africans.
The lifecycle hypothesis prediction thus does not hold for Africans in South Africa, because it would imply that young
households own more debt-related products, such as loans and mortgages, whereas more mature households would have
more products for saving and investing. The hypothesis also would indicate that middle-aged households should be rela-
tively ﬁnancially active in terms of product ownership. Contrarily,Table 5Ashows that the young married without children,
young married with children, and middle-aged married with young and adolescent children groups own more ﬁnancial
products than others. This result is consistent with ﬁndings in the innovation literature, i.e., young, wealthy, highly educated
households are more likely to acquire innovative products, as discussed in Section2.2.
InTable 5B, we report the results for the Coloured group. In general, Coloureds in all lifecycle stages appear less often in
the inactive cluster than Africans. Their position was comparatively better under apartheid than the position of Africans, be-
cause ‘‘the creation of the ‘colour bar’ rigorously excluded all Black South Africans (i.e. Africans) from any skilled or semi-
skilled work. Similar but not quite so stringent restrictions were applied to Coloured and Asian workers’’ (Feinstein, 2005,
p. 74). Yet despite the historical and cultural differences between Africans and Coloureds, we ﬁnd that the most active life-
cycle stages for Coloureds are the same as those for Africans, namely, (1) young married without children, (2) young married
with children, and (3) middle-aged married with young and adolescent children, as we show in the shaded areas inTable 5B.
Households in these lifecycle phases less often appear in the inactive segment and are relatively more likely to constitute the
savings + 1 and active segments. We ﬁnd relatively high incomes but low educational levels for the group of middle-aged,
married, independent children only respondents. Despite their high income level, this group does not belong to the most
active lifecycle groups, as was also the case for Africans. This ﬁnding indicates that educational level is critical to the own-
ership of ﬁnancial products and services, again consistent with innovation adoption literature ﬁndings addressed in Section
2.2.
Table 5A
Cluster membership across lifecycle phases for Africans.
Notes: Shaded groups <50% in segment 1 (inactives).
Table 5B
Cluster membership across lifecycle phases for Coloureds.
Notes: Lightly shaded groups <50% in segment 1 (inactives); darker shaded groups > 30% in segments 3 and 4 (savings + 1 and actives).M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–1815

We also perceive a lifecycle effect inTable 5Bthough. Coloured young singles have relatively high incomes and high edu-
cational levels, but they are not among the most ﬁnancially active groups. Perhaps they will develop their ﬁnancial product
portfolios later in their household lifecycles, as expected by the lifecycle hypothesis. This topic also is worthy of further
investigation, though the lifecycle hypothesis currently cannot explain the ﬁnancial product portfolios of Coloureds across
different cohorts.
For the Asian group, inTable 5C, we ﬁnd that the following lifestyle groups are relatively less likely to appear in the inac-
tive segment but more likely to show up in the savings + 1 or active segments: (1) young married without children, (2) young
married with children, (3) middle-aged married with young and adolescent children, and (4) middle-aged married without
dependent children. Thus, this group is highly similar to the Coloured group, if somewhat more active in terms of product
ownership, particularly by middle-aged marrieds without dependent children. The previously mentioned lifecycle effect for
Coloureds also holds for Asians, such that Asian young singles have relatively high incomes and high educational levels but
are not among the most ﬁnancially active groups.
InTable 5D, we report the ﬁndings for Whites, which reveal stark differences with the three other ethnic groups. Whites
own more ﬁnancial products, than members of other ethnic groups, across all lifecycle stages. Moreover, the ﬁndings for the
White group are somewhat consistent with the lifecycle hypothesis and ﬁndings in Western countries (Gunnarson & Wahl-
und, 1997; Paas et al., 2007; Soutar & Cornish-Ward, 1997). Younger groups own relatively few ﬁnancial products, middle-
aged groups own the most (most often in the relatively active segments), and older households tend to own more ﬁnancial
products than younger households but less than the middle-aged households. Beyond these trends, income levels are con-
sistent with those found in Western countries across household lifecycle stages: The intermediate stages have the higher
incomes, but the older and younger groups tend to have lower incomes. Also, educational levels are more similar across dif-
ferent household stages, with only the older groups exhibiting relatively lower levels of education. However, one ﬁnding is
similar to those for the three other ethnic groups: Whereas the lifecycle hypothesis predicts younger households should own
more debt-related products, this expectation was disconﬁrmed for Whites, as for the other ethnic groups.
5. Discussion
In this paper, we analyzed an extensive data set pertaining to the ownership of 16 ﬁnancial products by members of four
ethnic groups in South Africa: Africans, Coloureds, Asians, and Whites. We ﬁnd that the level of ﬁnancial activity is the main
Table 5C
Cluster membership across lifecycle phases for Asians.
Notes: Lightly shaded groups <50% in segment 1 (inactives); darker shaded groups >30% in segments 3 and 4 (savings + 1 and actives).
Table 5D
Cluster membership across lifecycle phases for Whites.
Notes: Lightly shaded groups <50% in segment 1 (inactives); darker shaded group >30% in segments 3 and 4 (savings + 1 and actives); darkest shaded group
>60% in segments 3 and 4. 16M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–18

differentiating factor between household ﬁnancial product portfolios in South Africa. Different ﬁnancial strategies do not
seem to play a role as was found to be the case in a previously researched Western country, Sweden (Gunnarson & Wahlund,
1997). We also ﬁnd that the lifecycle hypothesis-based expectation (Browning & Lusardi, 1996; Wärneryd, 1999) that youn-
ger households will have more debts and households in intermediate lifecycle stages own more assets does not hold for any
of the four researched ethnic groups. However, White households in intermediate household stages own more ﬁnancial
products, compared with younger and older households, which is consistent with ﬁndings from studies in Western countries
(Gunnarson & Wahlund, 1997; Paas et al., 2007; Soutar & Cornish-Ward, 1997). This tendency is not found amongst Africans,
Coloureds and Asians. Amongst these ethnics groups younger, relatively wealthy and educated, households tend to own
more ﬁnancial products. In a previous study,Ando, Guiso, Terlizzese, and Dorsainville (1992)reported that young house-
holds in Japan and Italy saved more than expected and that consumption smoothing was not as strong as expected. However,
in their sample, young households still saved less than households in intermediate lifecycle stages. Our ﬁnding is more pro-
nounced, in the sense that young households own more ﬁnancial products than households in intermediate stages among
Africans, Coloureds, and Asians.
The ﬁnding that young African, Coloured, and Asian households with higher incomes and higher educational levels tend
to own relatively many ﬁnancial products is consistent with literature on innovation adoption (Arts et al., 2011; Gatignon &
Robertson, 1985; Im et al., 2003; Steenkamp & Gielens, 2003), which expects such households to be most inclined to adopt
new innovations. Given the economic and historical backgrounds of these three ethnic groups, ﬁnancial products are inno-
vations for them (Roos et al., 2005; Steenkamp & Burgess, 2002). The effects of recent government policies, such as afﬁrma-
tive action, to address inequalities of the past might also favorably discriminate younger non-Whites from their
predecessors, in terms of their ability to own ﬁnancial services products (Maisonnaiveet al., 2009). However, young, single
Coloureds and Asians, even with their relatively high incomes and educational levels, tend to own fewer products than inter-
mediate lifecycle stage households from the same ethnic groups, which is consistent with the Western ﬁndings (Gunnarson
& Wahlund, 1997; Paas et al., 2007; Soutar & Cornish-Ward, 1997). Thus, for young Coloureds and Asians, lifecycle effects
may play a role. This relevance of the lifecycle hypothesis may imply that it will apply more aptly to these groups in the
future. Young Coloured and Asian households of today will be the intermediate households in one or two decades. They
may continue following lifecycle hypothesis predictions; research could test this prediction by following different South Afri-
can cohorts over time. However, for the largest ethnic group in South Africa, Africans, we did not ﬁnd any such lifecycle effect
in our database. Furthermore, at the time our data were collected, between July 2007 and June 2008, the innovation adoption
literature better explains ownership of ﬁnancial products amongst Africans, Coloureds and Asian than the lifecycle
hypothesis.
Our research thus has implications for studies of household ﬁnancial product portfolios in non-Western countries, as well
as for the lifecycle hypothesis. For non-Western countries, longitudinal research should assess the applicability of ﬁndings
from the innovation adoption literature, together with the lifecycle hypothesis, for households’ product portfolios. Overall,
we caution that lifecycle hypothesis should not be taken for granted in non-Western countries. Further research is required
to assess the extent to which local conditions affect its applicability.
Another implication is that we adjusted the Western lifecycle operationalization provided byMurphy and Staples (1979)
to the South African context, by including categories of singles without children and single parents, as well as excluding the
divorced categories. Singles and single parents are growing groups in Western countries (Michael, Fuchs, & Scott,1980;
Quintano & D’Agostino, 2006), so this adjustment may be relevant for both Western and non-Western countries. More gen-
erally, our study points out thatMurphy and Staples’s (1979)lifecycle operationalization and other lifecycle categorizations,
as developed in Western countries, require adjustments in non-Western countries.
Acknowledgements
We would like to express our gratitude to Ruud Frambach, Hester van Herk, two anonymous reviewers and the associate
editor of this paper for the useful suggestions on a previous draft of this paper. We are also grateful for participants of the
ICABEEP/IAREP/SABE 2011 session in which this paper was presented for their useful comments.
References
Ando, A., Guiso, L., Terlizzese, D., & Dorsainville, D. (1992). Saving amongst young households: Evidence from Japan and Italy.Scandinavian Journal of
Economics, 94(2), 233–250.
Arts, J. W. C., Frambach, R. T., & Bijmolt, T. H. A. (2011). Generalizations on consumer innovation adoption: A meta-analysis on drivers of intention and
behavior.International Journal of Research in Marketing, 28(2), 134–144.
Baek, E., & Hong, G.-S. (2004). Effects of life-cycle stages on consumer debts.Journal of Family and Economic Issues, 25(3), 359–385.
Bijmolt, T. H. A., Paas, L. J., & Vermunt, J. K. (2004). Country and consumer segmentation: Multi-level latent class analysis of ﬁnancial product ownership.
International Journal of Research in Marketing, 21(4), 323–340.
Black, N. J., Lockett, A., Winklhofer, H., & Ennew, C. (2001). The adoption of Internet ﬁnancial services: A qualitative study.International Journal of Retail &
Distribution Management, 29(8), 390–398.
Browning, M., & Lusardi, A. (1996). Household savings: Micro theories and micro facts.Journal of Economic Literature, 34(4), 1797–1855.
Burgess, S. M., & Steenkamp, J.-B. E. M. (2006). Marketing renaissance. Research in emerging markets advances marketing science and practice.International
Journal of Research in Marketing, 23(4), 1–25.
Casale, D., & Desmond, C. (2007). The economic well-being of the family: Household’s access to resources in South Africa, 1995–2003. In A. Y. Amoateng&T.
B. Heaton (Eds.),Families and households in post-apartheid South Africa: Socio-demographic perspectives. Cape Town: HSRC Press. M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–1817

Chakravarty, S., & Dubinsky, A. (2005). Individual investors’ reactions to decimalization: Innovation diffusion in ﬁnancial markets.Journal of Economic
Psychology, 26(1), 89–103.
Cross, C. (1987). Informal lending: Do-it-yourself credit for black rural areas.Indicator South Africa, 4(3), 711–725.
Davies, S., Easaw, J., & Ghoshray, A. (2009). Mental accounting and remittances: A study of rural Malawian households.Journal of Economic Psychology, 30(3),
321–334.
Deaton, A. (1989). Saving in developing countries: Theory and review.Proceedings of the World Bank Annual Conference on Developmental Economics, 61, 96.
Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood estimation from incomplete data via the EM algorithm (with discussion).Journal of
the Royal Statistical Society, Series B, 39(1), 1–38.
Derrick, F. W., & Lehfeld, A. K. (1980). The family life cycle: An alternative approach.Journal of Consumer Research, 7(2), 214–217.
Feinstein, C. H. (2005).An economic history of South Africa: Conquest, discrimination and development. New York: Cambridge University Press.
Fernandez-Blanco, V., Orea, L., & Prieto-Rodriguez, J. (2009). Analyzing consumer heterogeneity and self reported tastes: An approach consistentwith the
consumer’s decision making process.Journal of Economic Psychology, 30(4), 622–633.
Finscope (2009).Fiscope SA 2009. Johannesburg: FinMark Trust.
Friedman, M. (1957).A theory of the consumption function. Princeton: Princeton University Press.
Gatignon, H., & Robertson, T. S. (1985). A propositional inventory for new diffusion research.Journal of Consumer Research, 11(March), 849–867.
Gersovitz, M. (1988). Saving and development. In T. N. Srinivasan (Ed.).Handbook of development and economics(Vol. 1). Amsterdam: Elsevier.
Gunnarson, J., & Wahlund, R. (1997). Household ﬁnancial strategies in Sweden: An exploratory study.Journal of Economic Psychology, 18(2–3), 211–233.
Hoffmann, A. O. I., & Broekhuizen, T. H. L. (2010). Understanding investors’ decisions to purchase innovative products: Drivers of adoption timing and range.
International Journal of Research in Marketing, 27, 342–355.
Hosegood, V., McGrath, N., & Moultrie, T. (2009). Dispensing with marriage: Marital and partnership trends in rural KwaZulu-Natal, South Africa, 2000–
2006.Demographic Research, 20(13), 312–320.
Im, S., Bayus, B. L., & Mason, C. H. (2003). An empirical study of innate consumer innovativeness, personal characteristics, and new-product adoption
behavior.Journal of the Academy of Marketing Science, 31(1), 61–73.
Kamakura, W. A., Ramaswami, S. N., & Srivastava, R. K. (1991). Applying latent trait analysis in the evaluation of prospects for cross-selling of ﬁnancial
services.International Journal of Research in Marketing, 8(4), 329–349.
Lockett, A., & Littler, D. (1997). The adoption of direct banking services.Journal of Marketing Management, 13(8), 791–811.
Maisonnave, H., Decaluwé, B., & Chitiga, M. (2009).Does South African afﬁrmative action policy reduce poverty? A CGE analysis. Centre Interuniversitaire sur le
Risque, les Politiques Economiques et l’Emploi, Working paper 09-36.
Michael, R. T., Fuchs, V. R., & Scott, S. R. (1980). Changes in the propensity to live alone: 1950–1976.Demography, 17(1), 39–56.
Modigliani, F., & Brumberg, R. (1954). Utility analysis and the consumption function: An interpretation of the cross section data. In K. Kurihari (Ed.),Post-
Keynesian economics(pp. 388–436). New Brunswick, NJ: Rutgers University Press.
Murphy, P. E., & Staples, W. A. (1979). A modernized family lifecycle.Journal of Consumer Research, 6(2), 12–22.
Musteen, M., Berker, V. C., & Baeten, V. L. (2006). CEO attributes associated with attitude toward change: The direct and moderating effects of CEO tenure.
Journal of Business Research, 59, 604–612.
Nga, M. -T. (2007).An investigative analysis into the saving behavior of poor households in developing countries, with speciﬁc reference to South Africa.
Unpublished MSc thesis, Department of Economics, The University of the Western Cape.
Oppewal, H., Paas, L. J., Crouch, G. I., & Huybers, T. (2010). Segmenting consumers on how they spend a tax rebate: An analysis of the Australian stimulus
payment.Journal of Economic Psychology, 31(4), 510–519.
Paas, L. J. (1998). Mokken scaling characteristic sets and acquisition patterns of durable- and ﬁnancial-products.Journal of Economic Psychology, 19(3),
353–376.
Paas, L. J., Bijmolt, T. H. A., & Vermunt, J. K. (2007). Acquisition patterns of ﬁnancial products: A longitudinal investigation.Journal of Economic Psychology,
28
(2), 229–241.
Paas, L. J., & Molenaar, I. W. (2005). Analysis of acquisition patterns: A theoretical and empirical evaluation of alternative methods.International Journal of
Research in Marketing, 22(1), 87–100.
Polatoglu, V. N., & Ekin, E. (2001). An empirical investigation of the Turkish consumers’ acceptance of Internet banking services.International Journal of Bank
Marketing, 19(4), 156–165.
Quintano, C., & D’Agostino, A. (2006). Studying inequality in income distribution of single-person households in four developed countries.Review of Income
and Wealth, 52(4), 525–546.
Rogers, E. M. (1995).Diffusion of innovations. New York: The Free Press.
Roos, V., Chiroro, P., Van Coppenhagen, C., Smith, I., Van Heerden, E., Abdoola, R. E., et al (2005). Money adventures: Introducing economic conceptsto
preschool children in the South African context.Journal of Economic Psychology, 26(3), 243–254.
SAARF (2008).All media and products survey 2008B. Johannesburg: South African Advertising Research Foundation.
Soutar, G. N., & Cornish-Ward, S. (1997). Ownership patterns for durable goods and ﬁnancial assets: A Rasch analysis.Applied Economics, 29(11), 903–911.
South Africa (1998).Employment equity act no. 55 of 1998. Pretoria: Government of South Africa.
Stafford, E. F., Kasulis, J. J., & Lusch, R. L. (1982). Consumer behaviour in accumulating household ﬁnancial assets.Journal of Business Research, 10(4), 397–417.
Stanley, T. O., Ford, J. K., & Richards, S. K. (1985). Segmentation of bank customers by age.International Journal of Bank Marketing, 3(3), 56–63.
Statistics South Africa (2003).Population census 1996: Community proﬁle databases. Pretoria: Statistics South Africa.
Statistics South Africa (2007).Community survey 2007: Methodology, processes, and highlights of key results. Pretoria: Statistics South Africa.
Steenkamp, J.-B. E. M. (2005). Moving out of the US silo: A call to arms for conducting cross-country marketing research.Journal of Marketing, 69(October),
6–8.
Steenkamp, J.-B. E. M., & Burgess, S. M. (2002). Optimum stimulation level and exploratory consumer behavior in an emerging market.International Journal
of Research in Marketing, 19(2), 131–150.
Steenkamp, J.-B. E. M., & Gielens, K. (2003). Consumer and market drivers of the trail probability of new consumer packaged goods.Journal of Consumer
Research, 30(December), 368–384.
Tin, J. (2000). Life-cycle hypothesis, propensities to save, and demand for ﬁnancial assets.Journal of Economics and Finance, 24(2), 110–121.
Vermunt, J. K., & Magidson, J. (2007).Technical guide for latent GOLD 4.5: Basic and advanced. Belmont, MA: Statistical Innovations Inc..
Wagner, J., & Hanna, S. (1983). The effectiveness of family life cycle variables in consumer expenditure research.Journal of Consumer Research, 10(3),
281–291.
Wärneryd, K.-E. (1989). On the psychology of saving: An essay on economic behaviour.Journal of Economic Psychology, 10(4), 515–541.
Wärneryd, K.-E. (1999).The psychology of saving: A study of economic psychology. Northampton, UK: Edward Elgar Publishing.
Wedel, M., & Kamakura, W. A. (2000).Market segmentation: Conceptual and methodological foundations(2nd ed.). Amsterdam, The Netherlands: Kluwer.
Xiao, J. J. (1996). Effects of family income and life-cycle stages on ﬁnancial asset ownership.Financial Counseling and Planning, 7, 21–30.
Ziehl, S. (2002). Divorce statistics – A case of wool being pulled over our eyes?SA Journal of Demography, 8(1), 201–202. 18M.A. Ngwenya, L.J. Paas / Journal of Economic Psychology 33 (2012) 8–18