Advances in Applied Sociology

2012. Vol.2, No.4, 280-291

Published Online December 2012 in SciRes ([Link] [Link]

Determinants of Saving among Low-Income Individuals in Rural

Uganda: Evidence from Assets Africa
Gina A. N. Chowa1*, Rainier D. Masa1, David Ansong2
1
School of Social Work, University of North Carolina, Chapel Hill, USA
2
Brown School of Social Work, Washington University, St. Louis, USA
Email: *chowa@[Link], rmasa@[Link]

Received September 8th, 2012; revised October 10th, 2012; accepted October 22nd, 2012

Although research has shown that poor people in sub-Saharan Africa (SSA), including those living in ru-
ral areas save, little is known about the factors that influence saving and asset accumulation among this
population. Using three theoretical perspectives on saving and asset accumulation, this study examines the
broader determinants of saving and asset accumulation among low-income individuals in rural Uganda.
Compared with the individual-oriented and sociological perspectives, institutional theory explains a large
part of the variance in saving outcome among rural, low-income households. Wealth, proximity to finan-
cial institutions, financial education, and financial incentives are positively associated with higher saving
performance. Findings suggest that poor people can and do save, particularly when institutional barriers to
saving are removed. Institutional structures, which encourage low-income individuals to save, may con-
tribute to a poverty reduction policy that shifts from just income supplementation to a more inclusive
wealth promotion policy that assists people in creating their own pathways out of poverty.

Keywords: Saving; Sub-Saharan Africa; Rural Uganda; Theory; Asset Building; Institutional Theory

Introduction lifetime consumption and improve a family’s well-being over

several generations. Savings and assets provide a cushion to fall
Numerous reasons, including low and irregular income and
back on during hard times and emergencies. Asset-poor fami-
lack of access to financial services, have been posited to con-
lies in SSA become more vulnerable to unexpected economic
tribute to sub-Saharan Africa’s (SSA) low formal savings rate1.
events or natural disasters and as a result, to their long-term
Access to financial services, including deposit or savings ac-
adverse consequences (Hoddinott, 2006). For the economy at
counts, remains a privilege for most of the population (Consul-
large, savings represents an important source for the financing
tative Group to Assist the Poor, 2010). In rural Uganda, where
of investment in developing countries. The savings rate of a
86% of the country’s 31 million people live, only 10% of the
population has access to basic financial services (Chemonics country has been found to be strongly correlated with invest-
International, 2007). Inadequate physical infrastructure, oner- ment and growth rates (Attanasio & Banks, 2001).
ous documentation requirements and high account fees have Although two decades of experimentation and research on
been found to be associated with lower levels of formal finan- saving and asset building suggest that linking low-income
cial service use in developing regions (Beck, Demirguc-Kunt, families with saving and asset-building strategies has the poten-
& Peria, 2008). In spite of the barriers, empirical research has tial to positively influence family well-being outcomes, includ-
shown that poor people in SSA, including those living in rural ing economic, educational, and health-related results, (Chowa,
areas, save (Chowa, Masa, & Sherraden, 2012; Collins, Mur- Ansong, & Masa, 2010; Chowa, Masa, & Sherraden, 2012;
doch, Rutherford, & Ruthven, 2009; Dupas & Robinson, 2009). Erulkar & Chong, 2005; Lerman & McKernan, 2010; Schreiner
However, limited empirical research has been conducted to & Sherraden, 2007; Ssewamala & Ismayilova, 2009; Williams
understand the factors that influence saving and asset accumu- Shanks, Kim, Loke, & Destin, 2010), most of the studies on
lation among rural individuals and households in SSA. determinants of household saving have focused on middle- to
Savings and assets are of interest to policymakers and schol- upper-income families. Few empirical studies have investigated
ars across disciplines because of their importance to individuals, savings’ determinants among low-income individuals and fami-
households, and the economy. For individuals and households, lies, and those that have have primarily been completed within
economic security throughout the life course is inherently the United States (Curley, Ssewamala, & Sherraden, 2009; Han
linked not only to income but also to asset ownership. Savings & Sherraden, 2009). Building on existing theories and empiri-
and assets are important because, unlike income, they are what cal evidence, this study contributes to knowledge in two ways.
individuals and families accumulate and hold over time. Assets, First, the study explores factors that affect saving among par-
such as savings, also generate returns that generally increase ticipants of an asset-building program in rural areas of Uganda.
As mentioned earlier, little is known about the factors affecting
1
In SSA, there are 163 bank accounts (compared with 635 bank accounts in saving and asset accumulation in rural, low-income populations
other developing regions) and 28 bank loans (compared with 245 bank
loans in other developing regions) per 1000 adults (Consultative Group to
in developing economies, particularly in SSA. Data used in this
Assist the Poor, 2010). study came from an asset-building intervention targeted to
*
corresponding author low-income individuals in rural communities in a developing

280 Copyright © 2012 SciRes.

 G. A. N. CHOWA ET AL.

country. Observed savings determinants from a quasi-experi- temporary or transitory income is channeled into savings; and
mental setting that intentionally provides an opportunity to save households are freely able to save and borrow to smooth their
or build assets may be quite different than those observed from consumption (Friedman, 1957). Both economic theories view
national data sets. Second, the study examines whether diverse savings primarily as a function of income.
theories of saving, which have been primarily developed in Due to the limited availability of relevant data on low-in-
more industrialized societies, can explain savings behavior come households in developing economies, literature cited in
among low-income households in developing countries. This this study also includes data on middle- to upper-income
study adds to a growing body of research (Curley, Ssewamala, households. Evidence from SSA and other developing countries,
& Sherraden, 2009; Han & Sherraden, 2009) that examines albeit mostly from middle- to upper-income households, sug-
competing theories of saving in a single model with a compara- gests that income positively influences saving and in ways con-
tive approach. To the authors’ knowledge, this is the first study sistent with standard economic theories (Schmidt-Hebbel,
to concurrently examine diverse theories of saving using quasi- Webb, & Corsetti, 2002). In Kenya, household income was
experimental household-level data from rural, low-income found to be a statistically significant predictor of savings
populations in SSA. among rural farmers, entrepreneurs, and teachers (Kibet, Mutai,
Ouma, Ouma, & Owuor, 2009). In Uganda, higher permanent
Theories of Savings and Evidence from SSA and and transitory incomes significantly increased the level of net
Other Developing Economies deposits among households that reported owning bank deposit
accounts (Kiiza & Pederson, 2001). Income was also a signifi-
Examining and explaining determinants of saving and asset cant predictor of improved savings in India (Agrawal, Sahoo, &
building have garnered attention from scholars across numerous Dash, 2007; Athukorala & Sen, 2004), Morocco (Abdelkhalek,
disciplines. Economic theories put primary emphasis on income Arestoff, de Freitas, & Mage, 2009), Pakistan (ur Rehman,
and age as predictors of saving and asset accumulation (Modi- Bashir, & Faridi, 2011), and the Philippines (Bersales & Mapa,
gliani & Ando, 1957). Behavioral economists and economic 2006). These findings suggest that households save a larger
psychologists have recognized the role of self-control, motives, share of their income when that income is higher.
and other personality characteristics on saving (Katona, 1975; Aside from income, age, including the age of the head of
Thaler & Shefrin, 1981; Wӓrneryd, 1999). Sociologists have household and other household members, is an important pre-
been interested in how class and social stratification influence dictor of saving, according to neoclassical economic theories.
saving and asset accumulation (D’Souza, 1981; Sorensen, Economic researchers commonly use dependency ratio, or
2000). Social workers have examined the effects of institutional those under age 15 and over 65 as a share of the total household
factors such as access, incentives, expectation, and facilitation composition, as an explanatory demographic variable. In the
in promoting saving (Beverly & Sherraden, 1999; Sherraden, LCH, households with more children at home may save less
1991; Sherraden, Schreiner, & Beverly, 2003). Similar to Han until the children leave home, which would raise the per capita
and Sherraden (2009), this section classifies existing theories income of the household. Thus, higher dependency ratio nega-
into three perspectives: 1) an individual-oriented perspective; 2) tively affects savings rates. Evidence from SSA and other de-
a social stratification perspective; and 3) an institutional per- veloping countries shows contradictory results. An increase in
spective. dependency ratio decreased saving, while a decrease in de-
pendency ratio increased saving among households in Kenya
Individual-Oriented Perspectives (Kibet et al., 2009), Indonesia (Johansson, 1998), India (Ang,
2009), China (Ang, 2009), Morocco (Abdelkhalek et al., 2009),
The individual perspective in this study includes neoclassical and Pakistan (ur Rehman et al., 2011). This result suggests that
economics, economic psychology, and behavioral economics. a reduction in the number of dependents relative to a working
Neoclassical economic theory assumes that individuals are age population has eased household budget constraints, thus
rational beings who respond in predictable ways to changes in increasing savings, which is consistent with the LCH. However,
incentives; and assumes that individuals have perfect knowl- other researchers have found no significant relationship be-
edge and access to perfect markets. Two prominent neoclassical tween savings rates and dependency ratio in developing coun-
economic theories include: 1) the life cycle hypothesis (LCH; tries (Cornia & Jerger, 1982; Deaton, 1992; Schmidt-Hebbel et
Modigliani & Ando, 1957); and 2) the permanent income hy- al., 2002). These nonsignificant findings are similar to Mason’s
pothesis (PIH; Friedman, 1957). Both theories assume that (1987, 1988) studies, which challenged findings that depend-
individuals and households are concerned about long-term ency ratio has a strong negative effect on saving. Mason quali-
consumption opportunities and therefore, explain saving and fies the negative effect of dependency ratio on savings by in-
consumption in terms of expected future income. The LCH troducing the age factor. He demonstrates that, in some cases,
posits that savings will be used to smooth consumption when the effects of the dependency ratio depend on the age of the
income varies by age. A main idea of the LCH is that working dependents. In the Philippines, for instance, researchers found
people are savers, whereas children and retired people are not. that the percentage of young dependents had a negative effect
Thus, differences in consumption and saving among house- on savings, whereas, the percentage of the elderly had a sig-
holds are attributed to age differences (Modigliani & Ando, nificant positive effect on savings (Bersales & Mapa, 2006).
1957). While people are working, they use their income to pro- Although evidence suggests that neoclassical economic theo-
vide for the household consumption, while at the same time ries can predict savings behavior of households in developing
they are saving to provide for their retirement. On the other countries, some researchers have argued that the application of
hand, the PIH suggests that savings decisions are based on in- LCH and PIH to explain savings behaviors of low-income
come being perceived as either permanent or temporary. households in developing countries can be problematic
Households mainly spend the permanent income, while the (Rosenzweig, 2001). For instance, PIH’s assumption that

Copyright © 2012 SciRes. 281

 G. A. N. CHOWA ET AL.

households are freely able to save and borrow to smooth their unrealistic assumptions of standard economic models of human
consumptions may not be true in developing countries where behavior, such as unbounded rationality, unbounded willpower,
low-income households have very limited access to a well- and unbounded “selfishness” (Shefrin & Thaler, 1988; Thaler,
developed insurance and credit market (Rosenzweig, 2001). In 1994). According to this perspective, common human charac-
addition, distinction between permanent and temporary income teristics such as self-control and ability to delay gratification,
may not be evident in many parts of the developing countries mental accounting, use of rules-of-thumb, default options, and
where household income is minimal and irregular. Savings in hyperbolic discounting shape financial behaviors and economic
low-income settings for long-term purposes, such as retirement, decisions (Ainsle, 1975; Angeletos, Laibson, Repetto, To-
may not be substantial given that many households struggle to bacman, & Weinberg, 2001; Laibson, 1997; Mullainathan &
meet subsistence consumption level, especially in times of Thaler, 2000; Shefrin & Thaler, 1988; Thaler, 1981). These
emergencies and other income shocks. Furthermore, the LCH characteristics can lead households to behave in ways that are
may not effectively predict long-term savings in low-income inconsistent with their own priorities. However, little is known
settings in developing countries because many households, about the explanatory powers of these factors on savings be-
particularly rural households, include more than two adults, and havior of low-income households in developing countries. On
adults of different generations. Since the life-cycle of the the other hand, studies from industrialized countries have
household is not the same as the life-cycle of the individual, it shown that self-control (Ameriks, Caplin, Leahy, & Tyler, 2004;
is not clear whose age matters for savings decisions (Deaton & Moffitt et al., 2010; Romal & Kaplan, 1995), and use of default
Paxson, 2000; Rosenzweig, 2001). In SSA, for instance, evi- options (Madrian & Shea, 2001; Thaler & Benartzi, 2004)
dence suggests that households change size and composition in shape saving behaviors of middle- and upper-income individu-
response to income fluctuations (Fafchamps, Udry, & Czukas, als and households.
1998).
Unlike neoclassical economic theories, the other two indi- Sociological Perspective
vidual perspectives on savings in this study—economic psy-
chology and behavioral economics—do not assume that people Social stratification theory refers essentially to a distribution
are rational and all-knowing. These two perspectives assume of power in society. The divisions in society, based on eco-
that personality characteristics and attitudinal variables affect nomically conditioned power, are called classes, which refer to
saving and asset accumulation. The inclusion of psychological any group of people that is found in the same economic situa-
factors on savings research has been the subject of investiga- tion (D’Souza, 1981; Weber, 1967). Class and social stratifica-
tions by early economic thinkers such as Jevons (1965), Mar- tion have strengths in explaining the factors affecting savings
shall (1961), and Fisher (1977). Although they recognized that behavior among low-income households because class relates
savings depend on economic factors, particularly income and to the possession (or lack) of resources (economic or otherwise)
its size and frequency, they also believe that there are various necessary for individuals and households to save and build up
psychological characteristics that influence the temptation to their assets. Individuals and families in lower economic classes
spend and forego saving. Although fewer psychologists have have limited access to information, resources, and services that
investigated psychological determinants of saving behavior can help them save and accumulate assets over time. When
than economists, there are some established psychological low-income families have assets, they are less likely to have
models on savings behaviors, including those by Katona (1975), access to additional resources that they can use to generate
Ölander and Seipel (1970), and Lindqvist (1981). For instance, positive returns on the assets they already own. The economis-
Katona’s theory of saving (1975) is partly determined by in- tic approach to class and social stratification suggests that a
come and partly by some independent intervening factors. Two major explanation of class inequalities rests in the nature of
important factors are the ability to save (mostly objective data) access to, and take up of, material resources, as well as the
and willingness to save (a variety of psychological variables). institutions that govern such access (Crompton, 2008). In addi-
Ability to save refers to those who can save, whereas willing- tion to control and possession of economic resources, class and
ness to save is related to the degree of optimism or pessimism social stratification are powerful determinants of outcomes that
of economic conditions (Katona, 1975). Thus, ability to save can further shape saving and asset accumulation patterns. De-
does not guarantee savings because savings also depend on an mand from social network members, particularly family mem-
individual’s willingness to save. Due to constraints on the bers, can make it difficult to save and accumulate assets (Stack,
availability of relevant data, there is no evidence thus far on the 1974). Evidence suggests that poverty in extended families can
effects of psychological factors on the saving behavior of impede saving and asset accumulation (Caskey, 1997; Chiteji &
low-income households in developing countries. Although Hamilton, 2005; Heflin & Patillo, 2002). Some scholars have
some evidence from industrialized countries shows that psy- attributed asset poverty to other factors beyond the control of an
chological predictors tend to have low explanatory power individual, such as cultural origins (Al-Awad & Elhiraika,
(Furnham, 1985; Lindqvist, 1981; Lunt & Livingstone, 1991), 2003), gender biased-cultural norms (Chowa, 2008), and finan-
there is also some evidence suggesting that personality charac- cial socialization in families, schools, or communities (Chiteji
teristics, including optimism/pessimism about economic condi- & Hamilton, 2005; Chiteji & Stafford, 1999; Cohen, 1994), and
tion ( Lunt & Livingstone, 1991), perceived locus of control race (Oliver & Shapiro, 1995; Shapiro, 2004).
(Lunt & Livingstone, 1991; Perry & Morris, 2005), perceived Evidence from SSA and other developing countries suggests
ability to save (Sherraden & McBride, 2010), and future orien- that class-related factors can explain savings behaviors of
tation (Webley & Nyhus, 2006) are associated with saving be- low-income households. Education has been found to be a sig-
haviors. nificant predictor of savings in Kenya (Kibet et al., 2009), and
Behavioral economics integrates insights from psychology the Philippines (Bersales & Mapa, 2006) but not in India
and economics. Behavioral economics qualifies some of the (Agrawal et al., 2007). Higher education level translates to

282 Copyright © 2012 SciRes.

 G. A. N. CHOWA ET AL.

higher savings level. Occupation, which can be predicted by a that institutional factors are associated with saving and asset
person’s level of education, was also found to be a significant building. In Uganda, proximity of the financial institution to the
predictor of savings rates in rural Kenya (Kibet et al., 2009). household is associated with the probability of whether or not a
With regard to access to other resources that can increase in- household will open a formal saving account, as well as the
come and generate positive returns on existing assets, evidence level of net deposits among households owning a bank account
from SSA is mixed. In Uganda, increases in the availability of (Kiiza & Pederson, 2001). In the same study, urban households
credit resulted in higher savings levels (Kiiza & Pederson, were more likely to open a deposit account than their rural
2001). Households with access to credit consistently hold counterparts. Higher transaction costs (due to reduced accessi-
higher net balances of savings than household without access to bility) were also found to have significant negative effects on
credit. However, in rural Kenya, improved access to credit re- the level of savings deposits among Ugandan (Kiiza & Peder-
sulted in savings reduction (Kibet et al., 2009). The authors son, 2001) and rural Kenyan households (Dupas & Robinson,
cited that this contradictory finding could be due to observed 2009). Dupas and Robinson (2009) found that subsidizing the
savings being used for consumption purposes and rarely con- opening fees for a savings account on behalf of a random sam-
verted to income-generating assets. ple of small business owners in rural Kenya increased the sav-
Furthermore, evidence also suggests that class-related factors, ings of women, many of them market vendors who opened the
such as education, not only affect savings rate but also owner- account compared to women who were not offered the account.
ship of a formal savings account. In Uganda, the education Also in Kenya, Kibet and colleagues (2009) found that higher
level of the head of household was found to be a statistically transport costs to saving institutions had a negative impact on
significant predictor on whether a household will acquire a the saving habits of teachers in rural areas. These results sug-
formal savings account (Kiiza & Pederson, 2001). Owing to the gest that poor people can be better off if it is much cheaper to
sufficiently low-income of many poor households in develop- start a bank account. In these circumstances, accumulation of
ing countries, they tend to use informal savings mechanisms, savings and other assets is not solely based on individual char-
which are less secure and safe than formal savings accounts acteristics and choices; some people have greater access than
(Collins et al., 2009). For example, in Pakistan, increases in others, and this disparity in access is evident in many parts of
income led to a higher rate of participation in both formal and the developing world.
informal savings sectors. However, at higher levels of income, Aside from access, information, particularly general infor-
formal institutions (such as banks) become more widely used mation about financial institutions and their products and ser-
than informal institutions (e.g. bisi, an informal savings com- vices, was found to be associated with owning a bank account
mittee similar to a rotating-savings-and-credit-association) among households in SSA. In Uganda, researchers found that
(Carpenter & Jensen, 2002). Further, the same study found that the likelihood of owning a savings account increases by
bank use was strongly influenced by literacy and numeracy, roughly 33 times when a household becomes well informed
suggesting that access to and use of formal savings institutions about a particular bank and its services (Kiiza & Pederson,
could face severe constraints in countries with low educational 2001). In the Philippines, restrictions (i.e. prohibitions or rules
attainment and literacy rates (Carpenter & Jensen, 2002). that restrict access to or use of savings or assets) attached to a
commitment savings product helped low-income women save
Institutional Perspective (Ashraf, Karlan, & Yin, 2006). However, the impact of other
institutional-level factors on saving behaviors among low-in-
Institutional theory posits that individuals and households are come households in developing countries is not yet known due
faced with institutional-level factors that make it impossible or to lack of relevant data. This study aims to provide evidence on
difficult to save. The main hypothesis of institutional theory the impact of two other institutional-level factors: incentives
assumes that low-income individuals and families are unable to and expectations on savings2.
save and accumulate assets primarily because they do not have
the same institutional opportunities that higher-income indi-
viduals and households receive (Beverly & Sherraden, 1999; Methodology
Sherraden, 1991). Otherwise, given access to the same institu- Data and Sample
tional support for saving and asset building that their wealthier
counterparts use, low-income families can be in a position to This paper studies participants of AssetsAfrica program in
save and accumulate assets. Institutions in the institutional the- rural Uganda. AssetsAfrica is a demonstration and research
ory refer to “purposefully-created policies, programs, products, 2
and services that shape opportunities, constraints, and conse- Evidence from the United States and other industrialized countries sug-
gests that other institutional constructs are associated with financial behav-
quences” (Beverly et al., 2008: p. 10). Institutional theory hy- iors and decisions. Information (or financial education), for instance, has
pothesizes that institutions affect worldviews, which in turn, been found to increase participation and contribution in savings plans
affect financial behaviors and decisions (Beverly & Sherraden, (Bernheim & Garrett, 2003; Clancy, Grinstein-Weiss, & Schreiner, 2001;
1999). Seven institutional-level dimensions have been hypothe- Nyce, 2005). Incentives, particularly in the form of matching contributions,
sized to influence saving and asset accumulation. These dimen- have been found to increase savings (Duflo, Gale, Liebman, Orszag, & Saez
2006; Han & Sherraden, 2009) and participation in savings plan (Duflo, et
sions are access, information, incentives, facilitation, expecta- al., 2006; Nyce, 2005). Research has also found that facilitation (i.e. assis-
tions, restrictions, and security (Beverly & Sherraden, 1999; tance with participation and savings), whether in the form of precommit-
Beverly et al., 2008; Sherraden & Barr, 2005; Sherraden et al., ment constraints (Thaler & Benartzi, 2004), automatic enrollment (Madrian
2003; Sherraden, Williams Shanks, McBride, & Ssewamala, & Shea, 2001), or direct deposit (Han & Sherraden, 2009; Sherraden &
2004). This study focuses on the institutional factors (access, McBride, 2010) shapes saving and asset accumulation. Evidence also sug-
gests that expectations (i.e. implicit or explicit suggestions about desired
information, incentives and expectations) that have variations in saving; Han & Sherraden, 2009; Schreiner & Sherraden, 2007) and restric-
the asset-building program used in the study. tions (Sherraden & McBride, 2010) help individual and families increase
Evidence from SSA and other developing countries suggests their savings.

Copyright © 2012 SciRes. 283

 G. A. N. CHOWA ET AL.

initiative designed to test asset-building innovations in Africa. predictors, and then the institutional-level predictors.
Implemented in Masindi, Uganda between 2004 and 2008, the
Ugandan pilot project used a quasi-experimental design, com- Outcome Variable
paring across treatment and comparison villages. This design
was chosen in part because of high risk of contamination across The outcome variable in this study was similar to previous
treatment and control subjects if randomly assigned in the same evaluations of asset-building programs: the average quarterly
village. The AssetsAfrica study sample consisted of 400 indi- net saving (AQNS). We used AQNS to measure saving per-
viduals assigned to the intervention group (n = 200) or the formance at the end of the program. AQNS is defined as net
comparison group (n =200). Individuals in the intervention savings per quarter and is calculated by dividing net savings by
group were selected by village committees, (i.e., economically the number of participation quarters (Schreiner et al., 2001). In
needy individuals in the community were chosen) and offered this study, the number of participation quarters was 12.
the opportunity to participate in the project. The intervention
implemented in this project was a structured asset-building Explanatory Variables
program offered to half of the study sample for a 3-year period.
The intervention comprised a comprehensive program that In this study, we used key constructs of the saving theories
included matched funds for participants’ savings, financial discussed in the earlier sections. Aside from the explanatory
education, and training on how to manage the asset they variables, we also included three demographic variables: gender,
planned to acquire with their savings. Treatment group partici- marital status, and gender of youngest child. Gender is a di-
pants opened savings accounts in a commercial bank. Due to chotomous variable, having 1 for female and 0 for male. Mari-
the absence of banks in the six intervention-site villages and the tal status is also a dichotomous variable, having 1 for married
distance to the financial institution in the business district of and 0 for not married. Gender of youngest child is a dichoto-
Masindi, Stanbic bank (an international company that operates mous variable, having 1 for male and 0 for female.
in 16 African countries) established a mobile bank that visited Using HMR, four models were estimated in the study. The
the villages every week to collect savings. Participants who first model we ran was based on the individual-oriented per-
wanted to complete their own banking transactions could either spectives. Controlling for demographic variables, we estimated
wait for the weekly mobile bank visit or travel to the bank in the impact of explanatory variables based on economic and
the Masindi business district. Deposits had to remain in the psychological theories of saving. Age and wealth were used as
accounts for a minimum of 6 months before participants were measures based on neoclassical economic theories. Wealth was
eligible to receive matched funds. Restrictions were made for used as a substitute for income. Measuring income was a chal-
lump sum deposits to encourage more regular savings over the lenge in this study as most respondents were seasonal income
participation period. Participants who successfully reached their earners who only earned income part of the year. As a result,
goals had their savings matched at a ratio of 1:1 after which respondents could not accurately recall how much income they
each participant purchased their desired asset. To encourage earned in a year. Therefore, wealth was used as a substitute as it
sustainability and viability of the assets, only purchases of pro- was easy to measure and could be verified on the spot. Wealth
ductive assets (i.e., those that would generate income) were was measured using an index, which included the total value in
eligible for matching funds. USD of financial and productive assets reported by participants
Data were collected by 12 locally trained interviewers who at baseline survey. Optimism/pessimism about future economic
conducted face-to-face surveys. Questionnaires were adminis- expectation, perceived locus of control, attitude toward saving,
tered twice (baseline and follow-up) over a 13-month interval. and self-control were used to examine how psychological fac-
Because this study focuses on examining the extent to which tors and common human characteristics were related to saving.
different theories of saving explain saving in an asset-building Future economic expectation was measured using a 3-item,
program for low-income households, only participants in the 7-point Likert scale ranging from 1 (extremely worse) to 7 (ex-
treatment group were used for analysis (n = 200). Missing data tremely better). Respondents were asked how well they ex-
on one or more explanatory variables reduced the sample to 127. pected their life and food supply to be next year, in addition to
Given the reduction of sample size, we analyzed the extent to how well they expected to be living next year. Perceived locus
which respondents (n = 127) and nonrespondents (n = 73) dif- of control was also measured using a 3-item, 7-point Likert
fered. Bivariate analyses showed no significant differences (p scale ranging from 1 (strongly disagree) to 7 (strongly agree).
> .05) between the two groups in all of the demographic and Respondents were asked if they felt pretty sure that their lives
explanatory variables.
would work out the way they wanted them to; if they usually
performed tasks the way they expected to; and if they always
Data Analysis finished doing something once they started it. Attitude toward
Through the use of hierarchical multiple regressions (HMR), saving was a dummy variable, with respondents choosing from
we assessed the degree to which savings among AssetsAfrica 1 (agree) that saving takes too long or 0 (do not agree).
participants could be explained by demographic variables and Self-control was measured using a 1-item, 7-point likert scale
three theoretical perspectives. HMR allows researchers to de- ranging from 1 (never) to 7 (always). Respondents were asked
cide which order to use for a list of explanatory variables by how often they were hesitant to spend money that they had
putting the predictors or groups of predictors into blocks of saved.
variables. Unlike stepwise regression, the groups of predictors The second model based on the sociological perspective
are based on theoretical grounds. Consistent with prior studies takes into account class and social stratification variables in
(Curley et al., 2009; Han & Sherraden, 2009), we entered the addition to the control variables we added in the first model.
individual-oriented predictors first, followed by sociological We included four different predictors to measure the relation-

284 Copyright © 2012 SciRes.

 G. A. N. CHOWA ET AL.

ships between savings and class and social stratification. First, Results
social support was measured using a 3-item, 7-point Likert
scale ranging from 0 (no help at all) to 7 (a lot of help). Re- Descriptive Results
spondents were asked how much help they get to meet the A total of 127 individuals were included in the sample of this
needs of the household from family and friends, from organiza- study. Table 1 presents descriptive statistics for the variables in
tions, and from the community. Second, economic strain was the analysis. The median amount of savings per quarter among
measured using a 1-item, 7-point Likert scale ranging from 1 AssetsAfrica participants was US $60. The median for savings
(extremely hard) to 7 (extremely easy). Respondents were asked per quarter are reported because the variables for savings ex-
how hard or easy it was to meet the needs of their household. pectations, wealth, number of financial education hours at-
Third, education was a dummy variable, with respondents tended, and amount of matching funds received are skewed.
choosing from 1 (secondary education or higher) or 0 (primary More females (59.84%) than males (40.16%) are represented in
education or lower). The dummy education variable was cre- the sample. The sample had more married individuals (79.53%)
ated based on the frequency distribution of the original variable. than unmarried individuals (20.47%). Participants’ youngest
And finally, type of employment was a dummy variable, with children included mostly girls (55.91%) compared to boys
respondents choosing from 1 (formal employment) and 0 (oth- (44.09%). Study participants were, on average, was about 35
ers). years old, while their median wealth was valued at US $372.
The third model adds the three institutional theory constructs Regarding self-control, 18.90% of participants reported that
(access, information, and expectation) to assess the relationship they were rarely to never hesitant to spend money that they had
between saving outcomes and institutional factors, after all the saved, whereas 30.71% and 50.4% of participants reported that
predictors in the previous models are taken into account. First, they were sometimes and more than often hesitant to spend
expectation was measured by the total amount of money re- their savings. Sixty-four percent of participants agreed that
spondents said they would like to save. Second, incentive was saving takes too long. On average, participants expected that
measured by the total amount of matching funds that each par- their economic conditions in the following year would be better
ticipant received at the end of the program. Third, information than the current year. Further, participants, on average, tended
was measured by the total number of hours that participants to agree that they had control over their lives.
spent in financial education. And fourth, access was divided On average, AssetsAfrica participants received some help
into two sub-domains: ease of visiting financial institutions and from family, friends, organizations, and the community. Re-
ease of depositing money. Both explanatory variables were garding the ability to meet the needs of their household, 45.7%
dummy variables. We created two groups for each access indi- of the participants reported having some degree of difficulty
cator: 1 for (easy ratings) and 0 for (hard ratings). meeting the needs of the household; 28.3% reported being in
In the fourth model, we included incentive as an institutional the middle, and 26% reported having an easier time meeting
theory construct to assess the relationship between saving out- their households’ needs (25.98%). Fifty-four percent of the
comes and financial incentives after all the predictors in the participants had at least a secondary education or higher,
first three models were considered. Incentives were separated whereas 45.67% had a primary education or lower. A lower
from the rest of institutional theory constructs to assess how percentage (20.47%) of participants was formally employed.
much of the outcomes are explained by financial incentives Regarding institutional features, the median amount of
when incentives are a key feature of the intervention. The deci- money participants said they wanted to save was US $300. The
sion to separate incentives was also made because incentives median number of financial education hours that AssetsAfrica
might inflate the AQNS when included and might therefore participants attended was 5. A higher percentage (63.78%) of
provide a different picture of the effects on savings. Separating participants reported having a hard time visiting a financial
them provided a clearer picture for participants about the actual institution. On the other hand, a lower percentage (33.86%) of
money being saved without regard to incentives. participants reported having a hard time making deposits. The
In HMR, a full model with all theoretical perspectives was median matching funds that participants received was US $290.
compared to models without each block of perspective. Partial Bivariate tests were conducted to examine the association
F-tests were conducted to determine to what extent each theo- between AQNS and each explanatory variable. Table 1 shows
retical perspective explained saving outcomes in an asset- results of bivariate analysis. Four of five institutional-level
building program. Results of the diagnostic tests and residual predictors were statistically significant (p < .10). However, this
analyses showed that the outcome variable and three predictors analysis only serves an exploratory purpose because it did not
were highly skewed and not normally distributed. To reduce the control for other explanatory variables.
influence of extreme observations on regression coefficient
estimates, we used two types of data transformation. We trans- Hierarchical Multiple Regression Analysis
formed AQNS and wealth using inverse hyperbolic sine trans-
formation to handle observed zero values (Burbidge, Magee, & Table 2 presents results of the HMR analysis. Results indi-
Robb, 1988). We transformed information (hours of financial cate that a significant amount of variance in savings perform-
education attended) and expectation (desired total amount of ance in AssetsAfrica is explained by model 4 or the final model
money in savings) variables using logarithmic transformation. in which all explanatory variables are considered. All R2 values
Further, because the financial incentive variable violated the were statistically significant (p < .05). Two R2 increments were
assumption of constant variance, we used weighted least also statistically significant (p < .05), (i.e. when institutional
squares regression in the final model to correct for the bias theory variables are added in the models). The R2 in the final
caused by heteroscedastic data (Kutner, Nachtsheim, & Neter, model indicates that 61% of the variance in an individual’s
2004). For each multi-item predictor, a composite score was saving performance can be accounted for by the three theoreti-
calculated by taking the average score of all items. cal perspectives on saving and demographic characteristics. The

Copyright © 2012 SciRes. 285

 G. A. N. CHOWA ET AL.

Table 1. Table 2.
Descriptive statistics and bivariate analysis results. Hierarchical multiple regression analysis on average quarterly net sav-
ings.
Variables % or Ma (SD) AQNS
Explanatory
Outcome variable Model 1 Model 2 Model 3 Model 4
variables
Average quarterly net saving (AQNS) in 2008 US$ 60.00 - Individual
Individual-oriented perspective Age .022 .021 .010 .006
Age 34.91 (10.40) .007 Future economic
.364 .476^ .325 .287
Wealth 372.00 .448*** Expectation

Expectation of future economic condition 5.30 (0.75) .254 Perceived locus

−.124 −.075 −.095 −.063
of control
Perceived locus of control 5.35 (1.08) .041
Wealth .443*** .470*** .483*** .284***
Self-control 4.61 (1.61) .216
Self-control .042 .048 −.032 −.011
Rarely to never hesitant to spend savings (%) 18.90
Sometimes hesitant to spend savings (%) 30.70 Attitude toward
saving (saving takes .028 −.048 −.324 −.151
Often to always hesitant to spend savings (%) 50.40 too long = 1)

Attitude toward saving (saving takes too long) (%) 63.78 Sociological
Social support −.095 .004 −.042
Sociological perspective
Social support 3.79 (1.51) .073 Economic strain −.049 −.005 −.059

Economic strain 3.62 (1.28) .009 Education (secondary

education or higher = −.358 −.553 −.217
Education .084 1)
Primary or lower (%) 45.67 Employment status
.272 .095 .070
Secondary or higher (%) 54.33 (formal = 1)
Type of employment .121 Institutional
Formal employment (%) 20.47 Expectation .144 .013
Others (%) 79.53
Ease of depositing
−.360 .179
Institutional perspective (easy =1)
Savings expectation (in USD) 300.00 .158^ Ease of visiting
.598^ .467^
Ease of visiting bank .594^ bank (easy = 1)

Easy to visit bank (%) 36.22 Number of hours of

financial education .506* .381*
Hard to visit bank (%) 63.78
attended
Ease of making deposits −.088 Financial incentives
.003***
Easy to make deposits (%) 66.14 (match funds)
Hard to make deposits (%) 33.86 Demographics
*
Hours of financial education attended 5.00 .508 Gender (female = 1) −.205 −.268 −.096 −.228
Financial incentive (match funds received, in Marital status
290.00 .004*** −.142 −.189 −.213 −.113
USD) (married = 1)
Demographics Gender of youngest
.218 .225 .323 .339
Gender −.270 child (male = 1)
Female 59.84 F value 3.10** 2.22* 2.40** 9.34***
Male 40.16 R2 .1926 .2031 .2722 .6089
Marital status −.074 2 *
Incremental R .0105 .0691 .3367***
Married 79.53
^
p < .10, *p < .05, **p < .01, ***p < .001, two-tailed test.
Not married 20.47
Gender of youngest child .344 greatest increase in variance occurred in the final model, which
Male 44.09 include a financial incentive variable. R2 increased from .2722
Female 55.91 in Model 3 to .6089 in Model 4, or a statistically significant
increase of .34.
Number of subjects 127
Results of each regression model are presented next. When
a
Each entry is percentage of AssetsAfrica participants in the categorical predictor individual-oriented perspective and demographic characteristics
variable or the mean of the continuous variable. Standard deviations are presented are held constant, a test of model fit suggests that there is a
in parentheses. Median is presented for AQNS, wealth, expectations, financial
education hours, and financial incentives. ^p < .10, *p < .05, **p < .01, ***p
regression relation (p < .01) and at least one β is not equal to
< .001. zero. Among the individual-level predictors, wealth is the only

286 Copyright © 2012 SciRes.

 G. A. N. CHOWA ET AL.

explanatory variable that is statistically significant (p < .001). centive, which in the current study is provided through match-
For instance, when other variables are held constant, and when ing funds in the saving performance of low-income individuals
wealth increases by 10%, AQNS is expected to increase by in rural Uganda. In the final model, four explanatory variables
about 4%. Consistent with previous research, other individ- were statistically significant. When other variables are held
ual-oriented variables in the study are not statistically signifi- constant, and when wealth increases by 10%, AQNS is ex-
cant. Model 1 explains 19% of the variance in AQNS. Results pected to increase by about 3% (p < .001). This result is 1% to
of model 2 also showed that the model fits with the data and 2% lower than the estimated values from the three earlier mod-
that there is a regression relation in the population (p < .05). els. Similarly, when the number of financial education hours
Wealth remains statistically significant (p < .001). When other attended increases by 10%, AQNS is expected to increase by
variables are held constant, and when wealth increases by 10%, about 4%, holding other variables constant (p < .05). Further,
AQNS is expected to increase by about 5%. In addition, future the expected increase in AQNS from those individuals who
economic expectation showed a statistical trend (p < .10). In the found it hard to visit a bank to those who found it easy to visit a
study sample, when other variables are held constant, a one-unit bank is about 60%, holding other variables constant (p < .10).
increase in future economic expectation, (i.e., when individuals Finally, for a one dollar increase in matching funds received,
expect next year to be better than the current year), AQNS is AQNS is expected to increase by about .3% (p < .001).
expected to increase by 61%. This finding contradicts what
economic psychologists have posited, which is that individuals Discussion
are likely to save more when they expect their economic future
to be worse. However, this finding is consistent with neoclassi- Conventional wisdom dictates that saving is more difficult
cal theory, which posits that when people have more disposable for low-income individuals and households than their wealthier
income, they tend to save more. How this translates to expecta- counterparts. However, this study adds evidence that rural,
tions of more disposable income rather than actual disposable low-income individuals in SSA can and do save, especially
income requires more research. Further, results of model 2 when given the opportunities. Further, results of this study pro-
showed no statistically significant sociological perspective vide initial evidence about the factors affecting saving among
predictors. This finding does not support results of previous low-income participants of an asset-building program in rural
studies in developing countries that have shown some socio- areas of Uganda. All three theoretical perspectives, albeit in
logical-oriented variables such as education and employment to varying degrees, explain variation in saving performance
be statistically significant predictors of savings. Model 2 ex- among rural, low-income households. The final model, which
plains 20% of the variance in AQNS. However, the incremental includes all three theoretical frameworks and demographics,
R2 between model 2 and model 1 is not statistically significant, accounted for the highest variance in AQNS. The statistically
suggesting that explanatory variables related to class and social significant incremental R2 in models 3 and 4 suggests that in-
stratification may not be important to saving performance stitutional theory of saving is important in predicting savings
among low-income rural Ugandans who participated in Assets- performance in an asset-building intervention for low-income,
Africa. rural households. Further, the substantial increase in variance
Consistent with the first two models, an F test model fit that occurred in the final model, which included amount of
showed that model 3 fit the data, and that there is also a regres- matching funds, suggests that financial incentives are important
sion relation in the population (p < .01). Model 3 explains 27% in encouraging low-income individuals to save.
of the variance in the outcome variable. The incremental R2 The strong explanatory power of institutional theory in pre-
between models 3 and 2 is statistically significant (p < .05); dicting saving outcomes among participants of AssetsAfrica is
suggesting the importance of information, access, and expecta- not surprising. Institutional theory of saving has been devel-
tion to the saving performance of low-income rural individuals oped with the objective of increasing our knowledge of how
who participated in AssetsAfrica. Wealth remains statistically individuals, especially the poor, can save (Sherraden, 1991).
significant (p < .001). When other variables are held constant, Although institutional theory emerged primarily from asset-
and when wealth increases by 10%, AQNS is expected to in- building research in the United States, results of the current
crease by 5%. Financial education and ease of visiting a bank study suggest that institutional theory can also explain substan-
(proximity) are also statistically significant. Holding other tially the factors affecting saving performance among rural
variables constant, for any 10% increase in the number of fi- low-income individuals in SSA. In other words, consistent with
nancial education hours attended, AQNS is expected to increase results from studies conducted in the United States, institutional
by about 5%. Further, holding other variables constant, the factors, such as information, access, and incentives, influence
expected increase in AQNS from those individuals who found saving behaviors of rural, low-income individuals in Uganda.
it hard to visit a bank to those who found it easy to visit a bank Thus, saving outcomes among the poor are not solely attributed
is about 82%. This finding showed a statistical trend (p < .10), to individual characteristics and choices. In AssetsAfrica, all
the two other measures of institutional theory—expectation and participants were required to take financial education, primarily
ease of depositing—were not statistically significant. because the intervention was a package that included access,
Results of the final model showed that the model fit with the incentives, and information through financial education. A
data. Overall, all three theoretical perspectives and demo- Chemonics International report (2007) cited limited financial
graphic variables explain 61% of the variance in AQNS. The information and lack of financial education as the two most
models also showed that there is a significant regression rela- common barriers to increasing savings rates in rural Uganda. In
tion in the population (p < .001). As stated earlier, the greatest this study, results suggest that financial education (one method
variance occurred in the final model, resulting in a statistically of providing financial information) is positively associated with
significant incremental R2 of .34 between model 3 and 4 (p higher savings. In other words, offering financial education to
< .001). This finding suggests the importance of financial in- low-income individuals in developing countries can lead to

Copyright © 2012 SciRes. 287

 G. A. N. CHOWA ET AL.

positive outcomes, as research on financial behaviors among does not necessarily provide a reliable measure of well-being
low-income individuals in the United States have shown (Hirad because many people in developing countries, especially in
& Zorn, 2002; Lusardi, 2002). Further, the current study sug- rural areas, engage in informal labor markets where incomes
gests that access (or proximity to financial institutions) is also are often highly variable and can be seasonal. Wealth provides
influential in encouraging low-income individuals to save. Dis- a better and more stable picture of long-term living standards
tance remains a major barrier to formal financial services and than an income snapshot because wealth or assets are what
other markets in rural areas in SSA. In rural Uganda, only 10% individuals and households accumulate over time and what last
of the population has access to basic financial services longer. In the final model, other than wealth, no other individ-
(Chemonics International, 2007). Although 80% of rural ual-oriented variables were found to be a statistically significant
Ugandans were active savers, only 13% of those active savers predictor of saving. These findings are consistent with results of
had a savings account in a formal financial institution (Che- prior studies that found no or weak statistical relationship be-
monics International, 2007). Even though financial institutions tween individual-oriented (economic and psychological) factors
may exist, transaction costs in reaching them can make them and savings outcome in both developed and developing econo-
unavailable. These transaction costs typically refer to time, mies. Further, when sociological variables were added, results
effort, and money spent to reach a bank. This finding supports showed a statistically nonsignificant incremental R2. This find-
previous findings about the positive association between prox- ing indicates that the four sociological constructs (education,
imity to financial institutions and higher savings in SSA (Kibet employment, level of social support, and degree of economic
et al., 2009; Kiiza & Pederson, 2001). However, unlike prior strain) in the study have a weak association in explaining sav-
studies, the current study provides initial evidence that prox- ing among rural, low-income individuals in AssetsAfrica. In
imity to financial institutions also influences positive saving other words, results suggest that class and social stratification
behaviors of low-income individuals in rural SSA. Further, the are less important than wealth and institutional factors to an
positive association between proximity and saving suggests that individual’s saving performance. These findings contradict
there can be benefits to both financial providers and users when what other researchers have found. These divergent findings
financial services are not only expanded but are easy to access, may be attributed to the following: 1) observed savings deter-
particularly in rural areas. Finally, financial incentives in the minants from a quasi-experimental setting that intentionally
form of matching funds increased AQNS, albeit not as much as provides an opportunity to save might be quite different than
expected, among AssetAfrica participants. Schreiner and Sher- those observed from national, aggregated data sets; 2) the sam-
raden (2007) argued that matches in asset-building programs ple in the current study was primarily rural, low-income indi-
may encourage low-income people to save for at least three viduals, whereas the samples in prior studies included a more
reasons: 1) matches increase the reward to saving and may help representative sample of the general population; and 3) meas-
compensate for the sacrifice required to defer consumption; 2) urement-related limitations might have confounded the find-
matches may motivate people by translating a given level of ings. For instance, economic strain is measured by only one
saving into a stock of wealth that is large enough to use for a question.
major asset; and 3) matches may be the program feature that
catches the participant’s attention and motivates him or her to Limitations and Conclusion
participate in the first place. The use of matching funds in As-
setsAfrica is perhaps the first of its kind in rural SSA. The posi- There are several limitations to this study. First, participants
tive relationship between the matching funds and saving out- in AssetsAfrica are not a representative sample of the overall
come indicates that low-income individuals will respond posi- low-income population in Uganda and SSA. Selection bias,
tively when given the opportunity to save and save more. This particularly self-selection, limits the generalizability of the
finding is identical to other studies in the United States that results. Second, this study examines saving performance only
investigated the role of matching funds on the savings per- in the account provided by the program. Thus, we cannot say
formance of low-income individuals (Duflo et al., 2006)The whether this is new savings or whether they have saved in other
benefits of matching funds to AssetsAfrica participants go be- ways that were not reported in the study. Third, although we
yond encouraging them to save; individuals are rewarded for used relevant measures of each theory, this study cannot con-
saving more. Higher savings mean that low-income individuals firm the validity of each theory especially when tested in a
and households are better prepared to weather income shocks different geographic and socioeconomic population. Results
and other emergencies. Savings are particularly important to may be an artifact of weak measures of key constructs of the
families living in developing economies, like Uganda, because saving theories. Future studies should employ, among other
formal social safety nets that families in more developed things, a wider range of key measures and more reliable scales
economies can rely on to buffer against emergencies are not to measure important constructs. Fourth, the study’s model
widely available. specifications may have omitted important predictors. Omission
Although saving performance in AssetsAfrica is influenced of important predictors may lead to different results. Similarly,
primarily by institutional factors, wealth is positively associated the functional form between the outcome and predictors might
with savings performance. In all four models, wealth is a statis- not have been specified correctly. For most of the variables, a
tically significant predictor, which suggests that wealth has a linear relationship with the saving outcome was assumed. Fu-
strong power in explaining saving among low-income indi- ture studies should also test for interaction effects among con-
viduals in the AssetsAfrica program. The current study supports structs of interest. For instance, it could be that expectations of
the findings of previous studies, albeit mostly on higher income total savings amount are not predictive in themselves, but the
individuals, in developing economies. A majority of studies on interaction with financial incentives or access might have an
saving behaviors in developing economies have included only effect. Fifth, the study has a small sample size relative to the
income in their analyses for various reasons. However, income number of explanatory variables included in the analyses. Thus,

288 Copyright © 2012 SciRes.

 G. A. N. CHOWA ET AL.

the study findings may have been biased because of the rela- for everyone? Barriers to bank access and use around the world.
tively small sample size. World Bank Economic Review, 22, 397-430.
Despite the limitations, some of the findings are statistically doi:10.1093/wber/lhn020
Bernheim, B. D., & Garrett, D. M. (2003). The effects of financial
and substantively significant. Results provide initial evidence education in the workplace: Evidence from a survey of households.
on the factors that affect saving by rural, low-income individu- Journal of Public Economics, 87, 1487-1519.
als. Controlling for other theoretical perspectives, the institu- doi:10.1016/S0047-2727(01)00184-0
tional theory of saving has independent and significant ex- Bersales, L. G. S., & Mapa, D. S. (2006). Patterns and determinants of
planatory power in saving outcomes. Unlike the institutional household saving in the Philippines (EMERGE Technical Report).
theory of saving, which emerged from studies of saving behav- Makati City: EMERGE.
Beverly, S., & Sherraden, M. (1999). Institutional determinants of
iors of low-income individuals and households, individual- saving: Implications for low-income households and public policy.
oriented and sociological perspectives were not developed to Journal of Socio-Economics, 28, 457-473.
explain saving by low-income populations. Findings also sug- doi:10.1016/S1053-5357(99)00046-3
gest that institutional structures of saving matter to low-income Beverly, S., Sherraden, M., Cramer, R., Williams Shanks, T., Nam, Y.,
individuals in rural Uganda. Given access to the same institu- & Zhan, M. (2008). Determinants of asset holdings. In S. M. McKer-
tional support for saving and asset building that their wealthier nan, & M. Sherraden (Eds.), Asset building and low-income families
(pp. 89-152).Washington DC: Urban Institute Press.
counterparts use, low-income individuals can be in a position to
Burbidge, J. B., Magee, L., & Robb, A. L. (1988). Alternative trans-
save and accumulate assets. Unlike the other theoretical per- formations to handle extreme values of the dependent variable. Jour-
spectives that may have limited practical implications, institu- nal of the American Statistical Association, 83, 123-127.
tional theory of saving can directly inform future programs and doi:10.1080/01621459.1988.10478575
policies designed to promote financial inclusion and help Carpenter, S. B., & Jensen, R. T. (2002). Household participation in
low-income individuals and families save and build assets. formal and informal savings mechanisms: Evidence from Pakistan.
Review of Development Economics, 6, 314-328.
doi:10.1111/1467-9361.00157
Acknowledgements Caskey, J. P. (1997). Beyond cash-and-carry: Financial savings, finan-
cial services, and low-income households in two communities.
Support for AssetsAfrica comes from the Ford Foundation. Swarthmore, PA: Swarthmore College.
The authors thank Grace Kemirembe for her valuable assistance Consultative Group to Assist the Poor (2010). Financial access 2010
in AssetsAfrica, and Susan White for her careful review and report. Washington DC: CGAP.
insightful comments on the manuscript. Chemonics International (2007). Improving access to financial services
in rural Uganda: Rural SPEED final report. Washington DC:
Chemonics.
REFERENCES Chiteji, N. S., & Hamilton, D. (2005). Family matters: Kin networks
and asset accumulation. In M. Sherraden (Ed.), Inclusion in the
Abdelkhalek, T., Arestoff, F., de Freitas, N. E. M., & Mage, S. (2009). American dream: Assets, poverty, and public policy (pp. 87-111).
A microeconometric analysis of households saving determinants in New York: Oxford University Press.
Morocco. The 1st GDRI DREEM Conference, Istanbul, 21-23 May Chiteji, N. S., & Stafford, F. P. (1999). Portfolio choices of parents and
2009. their children as young adults: Asset accumulation by African-
Agrawal, P., Sahoo, P., & Dash, R. K. (2007). Savings behavior in American families. American Economic Review, 98, 377-380.
India: Co-integration and causality evidence. The Singapore Eco- doi:10.1257/aer.89.2.377
nomic Review, 55, 273-295. doi:10.1142/S0217590810003717 Chowa, G. A. N. (2008). The impacts of an asset-building intervention
Ainsle, G. (1975). Specious reward: A behavioral theory of impulsive- on rural households in Sub-Saharan Africa (Doctoral Dissertation).
ness and impulse control. Psychological Bulletin, 82, 463-496. ProQuest Dissertations and Theses Database (UMI No. 3332075).
doi:10.1037/h0076860 Chowa, G. A. N., Ansong, D., & Masa, R. (2010). Assets and child
Al-Awad, M., & Elhiraika, A. (2003). Cultural effects and savings: well-being in developing countries: A research review. Children &
Evidence from immigrants to the United Arab Emirates. Journal of Youth Services Review, 32, 1508-1519.
Development Studies, 39, 139-151. doi:10.1016/[Link].2010.03.015
doi:10.1080/00220380412331333179 Chowa, G. A., Masa, R. D., & Sherraden, M. (2012). Wealth effects of
Ameriks, J., Caplin, A., Leahy, J., & Tyler, T. (2004). Measuring self- an asset-building intervention among rural households in Sub-Saha-
control. Cambridge, MA: National Bureau of Economic Research. ran Africa. Journal of the Society for Social Work and Research, 3,
Ang, J. (2009). Household saving behavior in an extended life cycle 329-345. doi:10.5243/jsswr.2012.20
model: A comparative study of China and India. Journal of Devel- Clancy, M., Grinstein-Weiss, M., & Schreiner, M. (2001). Financial
opment Studies, 45, 1344-1359. education and savings outcomes in Individual Development Accounts.
doi:10.1080/00220380902935840 St. Louis, MO: Washington University, Center for Social Develop-
Angeletos, G. M., Laibson, D., Repetto, A., Tobacman, J., & Weinberg, ment.
S. (2001). The hyperbolic buffer stock model: Calibration, simulation, Cohen, S. (1994). Consumer socialization: Children’s saving and
and empirical evaluation. Journal of Economic Perspectives, 15, 47- spending. Childhood Education, 70, 244-246.
68. doi:10.1257/jep.15.3.47 Collins, D., Morduch, J., Rutherford, S., & Ruthven, O. (2009). Portfo-
Ashraf, N., Karlan, D., & Yin, W. (2006). Tying odysseus to the mast: lios of the poor: How the world’s poor live on $2 a day. Princeton,
Evidence from a commitment savings product in the Philippines. The NJ: Princeton University Press.
Quarterly Journal of Economics, 121, 635-672. Cornia, G., & Jerger, G. (1982). Rural vs. urban saving behavior: Evi-
doi:10.1162/qjec.2006.121.2.635 dence from an ILO collection of household surveys. Development
Athukorala, P., & Sen, K. (2004). The determinants of private saving in and Change, 13, 123-157. doi:10.1111/j.1467-7660.1982.tb00115.x
India. World Development, 32, 491-503. Crompton, R. (2008). Class and stratification. Malden, MA: Polity
doi:10.1016/[Link].2003.07.008 Press.
Attanasio, O. P., & Banks, K. (2001). The assessment: Household sav- Curley, J., Ssewamala, F., & Sherraden, M. (2009). Institutions and
ing—Issues in theory and policy. Oxford Review of Economic Policy, savings in low-income households. Journal of Sociology and Social
17, 1-19. doi:10.1093/oxrep/17.1.1 Welfare, 36, 9-32.
Beck, T., Demirguc-Kunt, A., & Peria, M. S. (2008). Banking services D’Souza, V. S. (1981). Inequality and its perpetuation: A theory of

Copyright © 2012 SciRes. 289

 G. A. N. CHOWA ET AL.

social stratification. New Delhi: Manohar. Lusardi, A. (2002). Preparing for retirement: The importance of plan-
Deaton, A. (1992). Household saving in LDCs: Credit markets, insur- ning costs. National Tax Association Proceedings (pp: 148-154).
ance and welfare. Scandinavian Journal of Economics, 94, 253-273. Washington DC: National Tax Association.
doi:10.2307/3440451 Madrian, B., & Shea, D. (2000). The power of suggestion: Inertia in
Deaton, A., & Paxson, C. (2000). Growth and saving among individu- 401(k) participation and savings behavior. Quarterly Journal of
als and households. The Review of Economics and Statistics, 82, Economics, 116, 1149-1187. doi:10.1162/003355301753265543
212-225. doi:10.1162/003465300558740 Marshall, A. (1961). Principles of economics: An introductory volume
Duflo, E., Gale, W., Liebman, J., Orszag, P., & Saez, E. (2006). Saving (9th ed.). London: Macmillan.
incentives for low- and middle-income families: Evidence from a Mason, A. (1987). National saving rates and population growth: A new
field experiment with H&R Block. Quarterly Journal of Economics, model and new evidence, In D. G. Johnson, & R. D. Lee (Eds.),
121, 1311-1346. doi:10.1162/qjec.121.4.1311 Population growth and economic development: Issues and evidence
Dupas, P., & Robinson, J. (2009). Savings constraints and microenter- (pp. 523-560). Madison, WI: University of Wisconsin Press.
prise development: Evidence from a field experiment in Kenya Mason, A. (1988). Saving, economic growth, and demographic change.
(NBER Working Paper No. 14693). Cambridge, MA: National Bu- Population and Development Review, 14, 113-144.
reau of Economic Research. doi:10.2307/1972502
Erulkar, A., & Chong, E. (2005). Evaluation of a savings and micro- Modigliani, F., & Ando, A. K. (1957). Tests of the life cycle hypothesis
credit program for vulnerable young women in Nairobi. Nairobi: of savings. Bulletin of the Oxford Institute of Statistics, 19, 99-124.
Population Council. Moffitt, T. E., Arseneault, L., Belsky, D., Dickson, N., Hancox, R. J.,
Fafchamps, M., Udry, C., & Czukas, K. (1998). Drought and saving in Harrington, H. et al. (2010). A gradient of childhood self-control
West Africa: Are livestock a buffer stock? Journal of Development predicts health, wealth, and public safety. PNAS Proceedings of the
Economics, 55, 273-305. doi:10.1016/S0304-3878(98)00037-6 National Academy of Sciences of the United States of America, 108,
Fisher, I. (1930). The theory of interest. London: Macmillan. 2693-2698.
Friedman, M. (1957). A theory of the consumption function. Princeton, Mullainathan, S., & Thaler, R. (2000). Behavioral economics. Cam-
NJ: Princeton University Press. bridge, MA: National Bureau of Economic Research.
Furnham, A. (1985). Why do people save? Attitudes to, and habits of Nyce, S. A. (2005). The importance of financial communication for
saving money in Britain. Journal of Applied Social Psychology, 15, participation rates and contribution levels in 401(k) plans. Philadel-
354-373. doi:10.1111/j.1559-1816.1985.tb00912.x phia, PA: University of Pennsylvania, Wharton School, Pension Re-
Kibet, L. K., Mutai, B. K., Ouma, D. E., Ouma, S. A., & Owuor, G. search Council.
(2009). Determinants of household saving: Case study of smallholder Ölander, F., & Seipel, C. M. (1970). Psychological approaches to the
farmers, entrepreneurs and teachers in rural areas of Kenya. Journal study of saving (Studies in Consumer Savings No. 7). Urbana-
of Development and Agricultural Economics, 1, 137-143. Champaign, IL: Bureau of Economic and Business Research, Uni-
Kiiza, B., & Pederson, G. (2002). Household financial savings mobili- versity of Illinois.
zation: Empirical evidence from Uganda. Journal of African Econo- Oliver, M., & Shapiro, T. (1995). Black wealth/white wealth: A new
mies, 10, 390-409. doi:10.1093/jae/10.4.390 perspective on racial inequality. New York: Routledge.
Kutner, M. H., Nachtsheim, C. J., & Neter, J. (2004). Applied linear Perry, V. G., & Morris, M. D. (2005). Who is in control? The role of
regression models (4th ed.). New York: McGraw-Hill Irwin. self-perception, knowledge, and income in explaining consumer fi-
Han, C. K., & Sherraden, M. (2009). Do institutions really matter for nancial behavior. The Journal of Consumer Affairs, 39, 299-313.
saving among low-income households? A comparative approach. doi:10.1111/j.1745-6606.2005.00016.x
Journal of Socio-Economics, 38, 475-483. Romal, J. B., & Kaplan, B. J. (1995). Differences in self-control among
doi:10.1016/[Link].2008.07.002 spenders and savers. Psychology: A Journal of Human Behavior, 37,
Heflin, C., & Patillo, M. (2002). Kin effects on Black-White account 8-17.
and home ownership. Sociological Inquiry, 72, 269-292. Rosenzweig, M. R. (2001). Savings behavior in low-income countries.
doi:10.1111/1475-682X.00014 Oxford Review of Economic Policy, 17, 40-54.
Hirad, A., & Zorn, P. M. (2002). Prepurchase homeownership counsel- doi:10.1093/oxrep/17.1.40
ing: A little knowledge is a good thing. In N. P. Retsinas, & E. S. Schmidt-Hebbel, K., Webb, S. B., & Corsetti, G. (1992). Household
Belsky (Eds.), Low-income homeownership: Examining the unexam- saving in developing countries: First cross-country evidence. The
ined goal (pp. 146-174). Washington DC: Brookings Institution World Bank Economic Review, 6, 529-547.
Press. doi:10.1093/wber/6.3.529
Hoddinott, J. (2006). Shocks and their consequences across and within Schreiner, M., & Sherraden, M. (2007). Can the poor save? Saving and
households in rural Zimbabwe. Journal of Development Studies, 42, asset building in Individual Development Accounts. New Brunswick,
301-321. doi:10.1080/00220380500405501 NJ: Transaction Publishers.
Jevons, W. S. (1965). The theory of political economy (5th ed.). London: Schreiner, M., Sherraden, M., Clancy, M., Johnson, L., Curley, J.,
Macmillan. Grinstein-Weiss, M. et al. (2001). Savings and asset accumulation in
Johansson, S. (1998). Life cycles, oil cycles, or financial reforms? The individual development accounts (CSD Report 01-23). St. Louis, MO:
growth in private savings rates in Indonesia. World Development, 26, Washington University, Center for Social Development.
111-124. doi:10.1016/S0305-750X(97)10011-0 Shapiro, T. (2004). The hidden cost of being African American: How
Katona, G. (1975). Psychological economics. New York: Elsevier. wealth perpetuates inequality. New York, NY: Oxford University
Laibson, D. (1997). Golden eggs and hyperbolic discounting. The Press.
Quarterly Journal of Economics, 112, 443-477. Shefrin, H., & Thaler, R. (1988). The behavioral life-cycle hypothesis.
doi:10.1162/003355397555253 Economic Inquiry, 26, 609-643.
Lerman, R. I., & McKernan, S. M. (2009). Benefits and consequences doi:10.1111/j.1465-7295.1988.tb01520.x
of holding assets. In S. M. McKernan, & M. Sherraden (Eds.), Asset Sherraden, M. (1991). Assets and the poor: A new American welfare
building and low-income families (pp. 175-206). Washington DC: policy. Armonk, NY: M.E. Sharpe, Inc.
Urban Institute Press. Sherraden, M., & Barr, M. (2005). Institutions and inclusion in saving
Lindqvist, A. (1981). A note on determinants of household saving be- policy. In N. Retsinas & E. Belsky (Eds.), Building assets, building
havior. Journal of Economic Psychology, 1, 37-59. credit: Creating wealth in low-income communities (pp. 286-315).
doi:10.1016/0167-4870(81)90004-0 Washington DC: Brookings Institution Press.
Lunt, P., & Livingstone, S. (1991). Psychological, social and economic Sherraden, M., Schreiner, M., & Beverly, S. (2003). Income, institu-
determinants of saving: Comparing recurrent and total savings. tions, and saving performance in individual development accounts.
Journal of Economic Psychology, 12, 621-641. Economic Development Quarterly, 17, 95-112.
doi:10.1016/0167-4870(91)90003-C doi:10.1177/0891242402239200

290 Copyright © 2012 SciRes.

 G. A. N. CHOWA ET AL.

Sherraden, M. S., & McBride, A. M. (2010). Striving to save: Creating havioral economics to increase employee saving. Journal of Political
policies for financial security of low-income families. Ann Arbor, MI: Economy, 112, 164-187. doi:10.1086/380085
University of Michigan Press. Thaler, R. H., & Shefrin, H. M. (1981). An economic theory of self-
Sherraden, M. S., Williams Shanks, T., McBride, A. M., & Ssewamala, control. Journal of Political Economy, 89, 392-401.
F. (2004). Overcoming poverty: Supported saving as a household doi:10.1086/260971
development strategy (CSD Working Paper 04-13). St. Louis, MO: ur Rehman, H., Bashir, F., & Faridi, M. Z. (2011). Saving behavior
Washington University, Center for Social Development. among different income groups in Pakistan: A micro study. Interna-
Sørensen, A. (2000). Toward a sounder basis for class analysis. Ameri- tional Journal of Humanities and Social Science, 1, 268-277.
can Journal of Sociology, 105, 1523-1558. doi:10.1086/210463 Wärneryd, K. E. (1999). The psychology of saving: A study on eco-
Ssewamala, F. M., & Ismayilova, L. (2009). Integrating children sav- nomic psychology. Northampton, MA: Edward Elgar.
ings accounts in the care and support of orphaned adolescents in rural Weber, M. (1967). Class, status and party. In R. Bendix, & S. M. Lipset
Uganda. Social Service Review, 83, 453-472. doi:10.1086/605941 (Eds.), Class, status, and power: Social stratification in comparative
Stack, C. (1974). All our kin: Strategies for survival in a black commu- perspective (pp. 21-28). London: Routledge & Kegan Paul.
nity. New York: Harper and Row. Webley, P., & Nyhus, E. (2006). Parents’ influence on children’s future
Thaler, R. H. (1981). Some empirical evidence on dynamic inconsis- orientation and saving. Journal of Economic Psychology, 27, 140-
tency. Economic Letters, 8, 201-207. 164. doi:10.1016/[Link].2005.06.016
doi:10.1016/0165-1765(81)90067-7 Williams Shanks, T. R., Kim, Y., Loke, V., & Destin M. (2010). Assets
Thaler, R. H. (1994). Psychology and savings policies. American Eco- and child well-being in developed countries. Children & Youth Ser-
nomic Review, 84, 186-192. vices Review, 32, 1488-1496.
Thaler, R. H, & Benartzi, S. (2004). Save more tomorrow: Using be- doi:10.1016/[Link].2010.03.011

Copyright © 2012 SciRes. 291