Microeconomics Term Paper
“Early influences on saving
behavior: Analysis of British panel data” research paper is the work of Sarah
Brown and Karl Taylor and it was published in Journal of Banking and Finance 62 (2016) 1-14 in 2015. The purpose
of their study is to find out the determinants of saving behavior of children,
the relationship between childhood saving behavior and adulthood saving
behavior and the influence of parent’s saving behavior on children saving
behavior. This study contributes to the existing empirical works on saving by
investigating the saving behavior of individuals overtime from childhood (aged
11-15) to early adulthood (aged 17-30).
The authors begin their study
by analyzing the determinants of saving behavior during childhood. They use the
British Household Panel Survey (BHPS) data from 1997-2008 to conduct their
analysis. The data is taken from the Institute for Social and Economic
Research. According to the data, 43% of
their British kids save money to buy things; 35% save and not spend; and 22% spend
immediately. The random effects binary probit framework has been used to
examine the determinants on the children saving. The dependent variable in the
saving behavior during childhood analysis is whether the child saves and the
independent variables comprise of the monthly amount saved by the child’s
parent and the parental financial expectations. The monthly amount saved by the
child’s parents variable is included in the set of explanatory variables to see
whether an intergenerational saving behavior relationship exists between
parents and children. For parental financial expectation variable, it is used
to explore whether expected future finances of parents have any influence on
the saving behavior of the offspring. Parents
can have three prospects regarding their future financial performance: whether
decreasing (pessimistic) or improving (optimistic) or no change (neutral).
This study finds out that
child allowance is inversely associated with the probability of saving. A one
percent increases in the child’s pocket money is associated with a decrease in
the probability of the child saves by 2.2 percentage points. On the other hand,
the weekly pays that the child receives form part-time work are positively associated
with the probability of the child saves. In this sense, a one percent increases
in weekly pay increases the probability of the child saves by 1.37 percentage
point. Hence, different sources of income received by children appear to affect
the child’s saving behavior in two contrast ways. According to this study,
there seems to be no intergeneration effect of saving behavior between parents
and children since the intergeneration coefficient on the amount of monthly
savings of the parents is statistically insignificant. Nevertheless, the positive
and/or neutral financial expectation of parents proves to have a negative
impact on the possibility that the child saves. If parents are neutral or
optimistic about their future financial performance, this will reduce the
probability that their kid saves by 2-3 percentage points respectively. The
peripheral findings in the first part of this research include future
orientation kids i.e. having an intention to go to college, whether the child
is the natural offspring of his/her parents, the age of the kids, the talk
between parents and children about issues that matter on a daily basis, the
education level of parents, the household income, and having parents possess a
home outright are all positively related to the possibility of child saves. On
the contrary, not having a computer in the household and being in a single
parent household are negatively associated with the probability of the child
saving.
After having studied about
the determinants of saving behavior during childhood, the authors continue to
study about the influence of saving behavior as a child on the probability that
the individuals saves on a monthly basis during early adulthood. They continue
to use the BHPS Youth Survey data from 1997-2008 in conjunction with the UK
Household Longitudinal Study (UKHLS) from Understanding society in order to
extend the span of their study period to 2013/14. 31% of the sample saved both
as a child and in early adulthood, whereas 15% did not save as a child and in
early adulthood. In addition, 37% of the participants in the survey save on a monthly
basis. Four different models have been used to test the saving behavior during
early adulthood, namely (i) a static random effects probit model, (ii) a
dynamic random effects probit model, (iii) a random effects tobit model, and
(iv) instrumental variable models. The dependent
variable in these four models is whether the adult saves. In the static random
effects probit model, the independent variable is having saved as a child. State dependence, time invariant
characteristics, and group means of the time varying covariates are the
independent variable in the dynamic random effects probit model, while whether
having saved as a child, the number of times the individual saved during
childhood, and whether the individual saved during childhood can be decomposed
into whether the child saved to buy things or whether the child saved not to
spend are the explanatory variables for the random effects tobit model. An
instrumental variable analysis is employed to avoid the potential for
endogeneity of saving as a child for later-in-life saving. In the instrumental
variable analysis, the independent variables are the parental financial
expectations, whether the adult saved in the final period, and whether the
individual ever saved during their childhood.
From the static random
effects probit analysis, the study finds out that permanent income is
positively associated with the probability of saving on a monthly basis, while
income volatility is vice versa. Furthermore, the more education the individual
receives, the more likely that he/she will save during his/her adulthood. Those
who are in the labor force tend to save more than those who are not. The
outcomes of this analysis also disclose that individuals who own a home have a
higher probability of saving a monthly basis by 13 percentage points, while
those who have saved during their childhood also tend to save more in their
adulthood by 14.5 percentage points. By decomposing the binary indicator of
whether the individual ever saved as a child according to saving motive, the
authors interestingly finds out that those who saved to buy things during their
childhood have a higher probability of saving in their adulthood than those who
saved not to spend by 1.72 percentage point, i.e. 15.3 percentage points
compared to 13.58 percentage points. The authors deem this may be because
individuals saved as a child specifically to buy things have acquired important
skills in budgeting and setting goals at an early age. Finally, the authors discover
that the higher the number of time the individual saved as a child, the more
likely that s/he will save during his/her adulthood, ranging from 4.5
percentage points (saved once as a child) to 14.6 percentage points (saved four
or more times as a child).
With the inclusion of state
dependence in the dynamic random effects probit analysis, this has changed some
covariates that is statistically significant in the static random effects
probit analysis to be statistically insignificant. Nonetheless, the influence
of whether the individual saved as a child, the number of times the individual
saved as a child, and the effect of whether individual saved to buy things or
not spend remain statically significant at almost identical levels. Given that
the dependent variable is logged and whether the individual saved as a child is
a binary variable in the random effects tobit analysis, the results still
support the previous notable discoveries that whether the individual ever saved
as a child is associated with higher level of monthly savings during adulthood,
the level of savings is increasing monotonically in the number of times saved
as a child, and saving to buy things has a dominant effect on saving and not
spend. In addition to these remarkable findings, the instrumental variable
analysis conceals that individual who had a parent that was financially pessimistic
was less likely to save as a child. To be precise, a young adult whose parent
was financially optimistic was around 4.6 percentage points less likely to save
as a child.
After a thorough
reading and analyzing, I could find several shortcomings of this research. First
of all, the authors do not explain why the negative relationship between
whether a child saved and the head of household expects no change in finances
is higher than the negative relationship between whether a child saved and the
head of household expects finances to improve in the analysis of the childhood
saving behavior. Intuitively, the negative relationship between the former
variables should be less than the negative relationship between the variables.
Secondly, there are several other control variables I could think of that
should be included into this study such as the level of risk tolerance and the
quality of education that the child received. The level of risk tolerance and
the quality of schooling are likely to have explanatory power on the saving
behavior both during childhood and adulthood.
This research fails to control for the communications between parents and
children about financing and savings due to the limitation of the data. Even
though the authors find out there is a positive relationship between children’s
saving behaving and talks to child about issues that matter on a daily basis, the
use of “the effect of regular communication between the parent and the child on
important matters” in lieu of the effect the communications between “parents
and children about financing and savings” casts a doubt on the reliability of
their suggestion to parents to communicate with the offspring about financial
and saving matters in order develop the children financial literacy skills. They
should expand their data collection to control for the communications between
parents and children about financing and savings in order to make the
aforementioned suggestion trustworthy. In addition, the authors are unable to
give suggestions on the exact form(s) of allowance the parent should give to
their children in order to improve their saving behavior albeit they recommend
parents to consider the form in which they give allowances to children if
parents wish to encourage their children to save. Finally, this research
studies only on British households and hence, the results of this study may be
different for households in different countries or regions. For instance,
factors that influence saving behavior during childhood for British kids and
Asian kids may be different from each other. It would be better if the authors
could expand the scope of their study with the inclusion of data from other
countries in different continents.
To sum
up, this study wants to find out the factors that encourage children to save,
the effect of childhood saving behavior on adulthood saving behavior, and the
influence of parent’s saving behavior on children’s saving behavior. This
research contributes to the existing empirical literature on saving by
exploring the implications of saving behavior at the early stages of the life
cycle, from childhood through to early adulthood. The results of the saving
behavior during childhood analysis illustrate that income associated with work
is positively associated with the likelihood that the child saves, whilst
parental allowances/pocket money has the opposite effect. Furthermore, the positive
or neutral financial expectations of the parent can influence the saving
behavior of the offspring. There seems to be no correlation between the saving
behavior of parents and the saving decision of children. Form the saving
behavior during adulthood analysis, the authors find out that having saved as a
child increase the chance of saving during adulthood, the level of savings is increasing
monotonically in the number of times saved as a child, and saving to buy things
has a dominant effect on saving and not spend. There are some rooms for improvement in this study that I
can identify. This study can be improved by incorporating more control variables,
adding more data to control for the communications between parents and children
about financing and savings, specifying the exact type(s) of allowance that can
develop the saving behavior of children, and enlarging the scope of the study with
the inclusion of household data from various countries/regions. Plus, the
authors should explain why the negative relationship between whether a child
saved and the head of household expects no change in finances is higher than
the negative relationship between whether a child saved and the head of
household expects finances to improve in the saving behavior during their
saving behavior during childhood analysis.
Reference
Sarah B. and Karl T., 2015. “Early influences on saving behavior: Analysis
of British panel data”.
Journal of Banking and Finance 62 (2016), pp. 1-14.
No comments:
Post a Comment