2017

An Analysis of the Factors Affecting the Spending and Saving Habits of College Students
Samantha Villanueva Skidmore College
An Analysis of the Factors Affecting the Spending and Saving Habits of College Students
By Samantha Villanueva
A Thesis Submitted to Department of Economics Skidmore College
In Partial Fulfillment of the Requirement for the B.A Degree Thesis Advisor: Qi Ge
May 2, 2017
I. INTRODUCTION
Managing personal finances has shown to be a growing issue, particularly
within American culture. Within the first three months of 1999, consumer
spending increased by 6.7% while savings reached an all-time low of -0.5%
(Roberts & Jones, 2001). At an early age, individuals are exposed to various
methods of handling personal finances, which can often lead to the
development of poor habits (Gutter, Garrison & Copur, 2010). It is not until the
university level where most consumers begin to experience a large degree of
financial independence. Attaining money has become especially important to
college students, a generation of individuals raised in a credit card society
(Roberts & Jones, 2001). As they grow accustomed to this, students develop
their own beliefs regarding spending and saving habits, many of whom have
little regard for incurring debt (Pritchard, Myers & Cassidy, 1989). Having easy
access to credit allows students to delay paying off purchases, resulting in large
debt balances. While the majority of studies have been conducted in the United
States, few researchers have expanded the scope of this topic across various
cultures including Malaysia (Sabri & McDonald, 2010), Australia (Phau & Woo,
2008), and the United Kingdom (Furnham, 1999). As young adults become
enculturated in a credit-heavy culture, findings have shown dependence on this
method of payment, resulting in the development of poor habits and personal
financial issues (Hayhoe et al., 2000). The further analysis of personal finance
among young adults can help identify methods to resolve common issues that
may arise as well as develop strategies that promote better financial practices.
This study utilizes an original survey distributed to students at Skidmore
College, a private Liberal Arts college located in upstate New York. The survey
captures demographic 5 characteristics and personal attitudes toward
spending and saving. Particular attention is drawn towards connecting existing
economic theories of consumption to the behaviors of college students, as
limited research has been conducted in this realm of personal finance. The data
collected includes numerical as well as descriptive statistics that provide
insights into certain class years, ethnicities and gender that tend to spend more
over others. It was found that overall spending has fluctuated more than
savings over the course of one’s college career, so the focus of this study is on
average monthly spending. Findings from the study provide strong evidence
for ethnic differences in spending habits. There is support for the idea that
Whites and Asians spend significantly more than other ethnicities, which is in
line with my predictions of this sample population. Furthermore, the results of
the analysis on the Permanent Income Hypothesis also provide evidence that
this hypothesis can be applied to the sample population of Skidmore College
students. About 54.5% of respondents lie on the threshold of support for the
 Permanent Income Hypothesis, which is the idea that these students are willing
to spend more now in anticipation of earning a greater future income. I find that
there is very limited evidence of Hyperbolic Discounting, as 74.5% of students
reported a preference for receiving a delayed and higher reward rather than
one that was more immediate but less in value. There is reason to believe that
females employ more savings mechanisms, which supports the findings of
Furnham (1999). The utilization of money management techniques is obvious,
as women are found to spend less on average by 4.1%. By carrying out this
study within the context of a small, private liberal arts college, this paper
contributes to the growing body of literature on the role of personal finance in
the context of college students. To my knowledge, there has been no prior
research of personal 6 finance habits conducted on a liberal arts population.
Previous studies such as that of Cummins et al. (2009), indicates a geographic
factor that may play into the development of financial habits. Certain regions
may encourage particular behaviors, resulting in location-specific effects,
especially if students selected to participate are mostly from that homogenous
population. The implementation of this study specifically using individuals from
various ethnicities allows for greater applicability that is representative of
colleges with a diversified student body representation. This paper highlights
the impact of cultural norms that inevitably impact the establishment of
spending behaviors, many of which are all appropriated based on differing
social norms across the globe. As this topic is relevant across various
continents, there is increasing desire to further the examination of financial
habits among college-aged individuals. A majority of the literature reviewed
focuses on demographic factors or socially constructed models of consumption
to explain spending behavior. However, there is very limited knowledge on the
applicability of existing economic theories that explain consumers’ spending
habits, particularly within the setting of college students. As such, contributions
can be made from this study’s attempts to draw connections between college
level spending and theories of Hyperbolic Discounting and the Permanent
Income Hypothesis. Subsequent sections of this paper are organized in the
following manner: Section II provides a framework of pertinent literature to
contextualize the existing findings on college students’ spending and saving
habits; Section III includes methodology and data, where a description of the
survey questionnaire is provide along with data collection methods; Section IV
provides results from the regression analysis; Section V is a discussion, which
draws 7 conclusions and relates findings from the study back to existing
literature; and Section VI concludes with final remarks and policy implications.

II. LITERATURE REVIEW Current literature explores the impacts of demographic

factors that influence the financial habits of college students. Some researchers
have also extended the literature outside the U.S., finding cultural aspects that
may influence participants’ responses. However, there has also been very
limited research on the relationship between financial behaviors of college
students while also account for existing economic theories of spending, which
this study aims to address. The literature reviewed provides theoretical
framework on the permanent income hypothesis and hyperbolic discounting as
well as sufficient background on the financial habits of college students.
Theoretical Framework I: Hyperbolic Discounting The model of hyperbolic
discounting accounts for a time-inconsistent mechanism of choices. Given two
options, humans tend to show a preference for a more immediate, smaller
value reward rather than waiting for a later, higher value reward. These
individuals are often described as being present-oriented, with more regard for
current or immediate satisfaction rather than delayed satisfaction. As a result,
people end up delaying certain decisions, such as saving mechanisms. This
conflict of time preferences results in hyperbolic discounting, or intertemporal
preferences (Angeletos et al., 2001). A study by Angeletos, Laibson, Repetto,
Tobacman, and Weinberg (2001) analyzes the theory of hyperbolic discounting
in the context of 8 consumer behavior. They apply this model towards different
households and assume that economic life begins at age 20 and retirement
starts at age 63 (Angeletos et al., 2001). The findings of their simulation study
conclude that at an earlier stage of life, consumers tend to exhibit hyperbolic
discounting mannerisms (Angeletos et al., 2001). There is a preference for
instant gratification and participants were found to put off saving, assuming they
would eventually be able to start budgeting more efficiently at a later date
(Angeletos et al., 2001). Younger consumers in particular are linked to high
volumes of consumption financed by credit cards (Angeletos et al., 2001). The
study can be critiqued in terms of its approach. Angeletos et al. (2001) employ
a simulation mechanism which lays out a generic profile of a consumer and a
typical spending pattern. However, this may be too generalized and too
simplistic of a model. Consumer behaviors change and adapt over time, and
this model does not reflect for any sorts of income shocks or savings
mechanisms. Similarly to Angeleto’s (2001) findings, Laibson (1998)
establishes a model of an individual consumer who is an autonomous, temporal
being. He uses his model to study how one would act through various periods
of control over a consumption decision (Laibson, 1998). Age, income and
wealth are all found to be highly correlated with measures of patience (Laibson,
1998). More specifically, the expectation that one will be earning more in the
near future drives more willingness to spend. Young consumers who expect
rising income paths and consumers with low levels of cash on hand are more
likely to have a higher Marginal Propensity to Consume (Laibson, 1998).
Though able to account for certain behavioral mannerisms in consumers, the
model of hyperbolic discounting fails to consider some important factors. For
instance, bounded rationality and impulsive buying are two concepts that may
influence 9 consumer spending, but are not accounted for in the idea behind
hyperbolic discounting. Bounded rationality is the idea that when faced with a
decision, consumers are weighing their options under certain constraints
including limited information about alternative options and certain
consequences that may come with their chosen decision. On the other hand,
impulsive buying explains the instinctual behavior of a consumer, which may
probe him or her to make a spur of the moment decision. The model of
hyperbolic discounting may fall short in that it does not encompass other factors
that lie behind consumption behavior. Theoretical Framework II: Permanent
Income Hypothesis Another theory that also explains consumer behavior is the
Permanent Income Hypothesis. The idea behind this hypothesis is that people
spend their money in a way that is consistent with their expected long-term
income. Over time, consumption is said to be smoothed by changes in income.
The Permanent Income Hypothesis emphasizes that consumers will spend
money based on lifetime income, not just current income. Hayashi (1985)
conducts a study where Japanese families are questioned about their spending
habits, revisiting the participants every three months for an entire year. Findings
from the study conclude that the Permanent Income Hypothesis applied to
about 85% of the sample population (Hayashi, 1985). Wage earners are found
to exhibit consumption smoothing mechanisms (Hayashi, 1985). Further review
shows that changes in income explained only a small portion of expenditure
differences, warranting support for the Permanent Income Hypothesis
(Hayashi, 1985). However, the theory has been criticized for being limited in
the type of goods consumers purchased. Much of the previous literature
focused on changes in consumption of perishable 10 goods. This gives rise to
the speculation that the Permanent Income Hypothesis may not be applicable
towards non-perishable items. Perhaps consumption habits change when it
comes to more durable goods, thus the need for further research. Within the
context of my study, the Permanent Income Hypothesis is tested in a manner
that investigates the willingness of individuals to spend more given current
income, in anticipation of earning a higher, more stable income in the future.
There is regard for both durable and non-durable goods to account for different
types of goods that are purchased over a consumption period. Factors
Influencing Financial Behaviors I: The Role of Ethnicity Much of the existing
body of literature on the financial habits and attitudes of college students
focuses on the increasing importance of this topic and its implications on the
consumer culture. A study by Chen and Volpe (1998) is conducted with three
main purposes in mind: to provide evidence of personal financial literacy among
college students, examine why some students may be more knowledgeable
than others, and explain how a student’s knowledge influences his or her
opinions and decisions on financial matters. The researchers utilize a survey
questionnaire that asks participants about their knowledge of personal finance,
opinions and decisions about financial matters and demographic data (Chen &
Volpe, 1998). The survey was sent to students from 14 different colleges, both
public and private schools across the United States. Chen and Volpe (1998)
incorporate a multivariate model with independent variables being academic
discipline, class rank, gender, race, nationality, work experience, age, income
and type of major. Results from the survey reveal much about the population
of college students. At a young age, most of the money source available to
college 11 students is used for consumption rather than savings (Chen & Volpe,
1998). This may include personal allowances to cover the costs of food, school
supplies, and other miscellaneous items. Ethnicity also plays a significant role,
as African-Americans were less knowledgeable on financial matters across the
board, and foreign students in general performed worse than American
students (Chen & Volpe, 1998). Researchers conclude that students with less
knowledge on finance tend to have wrong opinions and make incorrect
decisions (Chen & Volpe, 1998). The survey questionnaire was designed to
test financial literacy as well as opinions towards financial decisions. However,
this may pose as a cultural bias, as the survey was only conducted across the
United States, drawing conclusions from an American cultural perspective. It is
highly possible that international students, being enculturated under different
standards, hold very different opinions and make opposing decisions when
dealing with finances compared to American students. The survey design may
have been biased toward a typical American’s mindset on spending and saving.
To better understand factors that influence basic adolescent spending and
saving, Pritchard, Myers and Cassidy (1989) gather data from high school
students from private and public schools across the nation. An administered
questionnaire determines whether students are savers, necessity spenders
and discretionary spenders (Pritchard et al., 1989). Findings from the study
report specifics on ethnic and gender differences. Students who tend to
predominantly save are female and white (Pritchard et al., 1989). This category
of participants performs better on standardized tests, receives better grades,
are rated as hard workers and plan to enroll in further education (Pritchard et
al., 1989). These kinds of individuals are said to be more future-oriented, with
a regard for higher savings and delayed gratification (Pritchard et 12 al., 1989).
For students who are primarily necessity spenders, most of the population are
females and black (Pritchard et al., 1989). Overall, they have poorer
performance on tests, receive lower grades and do not feel that saving is as
important given their lower socioeconomic status (Pritchard et al., 1989). There
is a lower drive for future goals in the workplace, as most of the students seem
to be present-oriented individuals. This preference for immediate gratification
is associated with the theory of hyperbolic discounting, where necessity
spenders prefer to receive a sooner, instant reward. With discretionary
spenders, the majority of this population are males and whites, which
researchers attributed to car ownership (Pritchard et al., 1989). Discretionary
spenders feel that having a lot of money is important (Pritchard et al., 1989).
Generally, there is more interest in attaining success in the workplace,
 accounting for poor educational performance (Pritchard et al., 1989). However,
the sample population taken for this study only utilized employed high school
seniors. While this sheds light on the habits of students entering college, it is
not very representative of high school students overall. Seniors may employ
different financial mechanisms, thus the results would not be very generalizable
to a larger population.
III. Factors Influencing Financial Behaviors II: A Growing Credit-Dependent
Culture Much of the literature that covers credit card usage draws similar
conclusions in that the presence of credit makes students highly dependent on
this payment method. In a study by Roberts and Jones (2001), the primary
focus revolves around the role of money attitudes and credit card use on
compulsive buying among U.S. college students. They emphasize that the
desire to be part of the consumer culture is constantly increasing, especially as
American 13 students are raised in a society where credit card usage is at an
all-time high (Roberts & Jones, 2001). They utilize a money attitudes dimension
scale to further understand the factors which most commonly lead to
compulsive buying habits (Roberts & Jones, 2001). The researchers employ a
model developed by Yamauchi and Templer to demonstrate the most powerful
factors behind money attitudes (Roberts & Jones, 2001). The Money Attitude
scale (MAS) consists of three dimensions: power, distrust, and anxiety (Roberts
& Jones, 2001). The dimension, power, was defined as individuals who use
money as a tool of influence and as a means of impressing others to exhibit
success (Roberts & Jones, 2001). Distrust is linked to price sensitivity, with
individuals who are hesitant, suspicious and doubtful in situations regarding
money (Roberts & Jones, 2001). The last factor, anxiety, is used to identify
people who view money as a source of stress or use it to cope with anxiety
(Roberts & Jones, 2001). With the use of a survey and the Money Attitudes
Scale, Roberts and Jones (2001) find that many college students use credit
cards irresponsibly and in the future, tend to suffer both financially and even
psychologically. These types of attitudes seem to carry on after college and
can often worsen. The model that the researchers employ in this study seems
limited in its scope of attitudes towards money. It only considers three
dimensions that were tested in the survey, leaving out many other factors that
could very well influence a participant’s financial attitudes, such as early
exposure to financial management and formal financial education.
Furthermore, the present study’s sample population appears homogenous,
utilizing students all from one college at a given point in time. To better assess
the relationship between the proposed attitude dimensions and spending
habits, Roberts and Jones (2001) address the need for further longitudinal
research.
A key component of the consumer culture is the materialistic attitudes held by
many college students. This segment of shoppers is a particularly attractive
market for credit card companies, made evident by the annual increase in on-
campus solicitations. Palan, Morrow, Trapp and Blackburn (2011) discuss the
issue of credit card misuse among college students, stating that students
associate credit cards with spending. Their findings support the idea that credit
cards promote compulsive buying behaviors and incentivize purchases that
may not otherwise be bought when using cash (Palan et al., 2011). Palan et
al.’s (2011) study utilized an online survey method that only tested senior
business majors enrolled in a capstone course. The fact that participants were
all business majors completing their culminating class gives a very biased view
of results. These individuals have been educated with a business degree,
which can provide conclusive results for business-minded students but not
those who do not fit into this category. Another study in support of these
arguments is one by Norvilitis, Merwin, Osberg, Roehling, Young and Kamas
(2006). Researchers claim that students often do not understand the financial
implications of their behaviors (Norvilitis et al., 2006). Those with credit cards
tend to spend less time and more money when making purchases (Norvilitis et
al., 2006). Presently, students are captivated by the consumer culture, and
having the means to delay paying off purchases is enticing. In the long run this
can result in higher debt and money management issues. Norvilitis et al. (2006)
use a comprehensive study mechanism, which includes a 173-item omnibus
questionnaire. While the study design was intentionally extremely detail-
oriented, capturing everything from financial well-being, attitudes toward debt,
psychological measures and materialism, it was extremely taxing to complete,
which may have deterred some students 15 from being fully engaged
throughout the entire survey. Furthermore, participants were instructed to
complete the survey outside of class and return with the completed
questionnaire at a subsequent meeting. There is a high possibility of
extraneous variables that could have affected the survey results. Students
inevitably completed the survey at different times, locations and could have
done so at multiple intervals rather than in one sitting. These variances may
cause results to vary in ways that were unintended by the researchers. Factors
Influencing Financial Behaviors III: Gender Differences In a separate study of
differences in spending habits and credit use, Hayhoe, Leach, Turner, Bruin
and Lawrence (2000) study the relationship between affective credit attitudes
and gender on purchasing habits. The researchers introduce a multivariate
model that includes variety of purchases, financial management practices,
financial stressors, affective credit attitude, and number of credit cards with a
balance (Hayhoe et al., 2000). They use this model to analyze behavioral
differences, particularly with regards to gender. Hayhoe et al. (2000) survey
college students over the age of eighteen from six different public universities.
The sample consists of an even split between males and females, the majority
of which are full-time students (Hayhoe et al., 2000). To analyze the results,
the researchers use a logistic regression analysis when studying the effect of
credit purchases and apply an OLS regression model when studying financial
practices (Hayhoe et al., 2000). Afterwards, Hayhoe et al. (2000) use a path
analysis model to show the relationship between credit attitudes, variety of
purchases, number of financial stressors, number of financial management
practices, and number of credit cards on which the student carried a balance.
As the researchers use an exploratory analysis, they 16 first run a saturated
model where all paths are initially specified, followed by a restricted model,
where they only include significant paths (Hayhoe et al., 2000). Results show
that there is a strong influence of gender and affective credit attitudes. In
particular, financial management practices, financial stressors, affective credit
attitudes and number of credit cards differed by gender (Hayhoe et al., 2000).
The responses received allow researchers to draw conclusions about the ways
men and women vary in their financial habits. Females tend to use credit cards
on appearance goods, like clothing, while males use credit cards for
electronics, entertainment and food (Hayhoe et al., 2000). Women are also
found to exhibit more financial practices such as keeping a written budget,
planning spending and saving regularly (Hayhoe et al., 2000). However, both
genders feel that overall, they do a good job managing their finances (Hayhoe
et al., 2000). The shortcoming of this study is the lack of testing for differences
in ethnicity. The variable has proven to have significant effects among other
studies in the field and is also a variable included in the model used for this
study. While there was a large sample size in Hayhoe et al.’s (2000) analysis,
it is unclear the various backgrounds of individuals that were captured by this
study, which may provide further understanding and examination of the results.
Among various pieces of literature, gender often acts as a strong determining
factor. In a study of money attitudes and credit card debt, women report having
a self-imposed budget more frequently than men (Norvilitis et al., 2006).
Contrary to Norvilitis et al. (2006) and Hayhoe et al. (2000), Roberts (2000)
finds that women have been raised and enculturated to find satisfaction from
shopping. Thus, they are more likely to exhibit spending behaviors, particularly
compulsive buying, as compared to men. This finding suggests that spending
for 17 females may only be greater than that of males in the context of shopping
for appearance goods, such as clothing, shoes, accessories and cosmetics.
The conclusions from different researchers focusing on gender appear to agree
on these statements across the board. Roberts (2000) acquires data strictly
from students at Baylor University, a Baptist university in Texas. The student
body of Baylor University may attract many students of this religious
demographic, which can hold certain beliefs towards spending and saving.
Researchers could have extended their research to a non-denominational
university where this religious aspect would not have been an issue. Literature
outside the U.S. While most studies focus on American college students, some
researchers have extended the literature outside the U.S., highlighting the fact
that the financial attitudes and behaviors of college students are also an
international focus. Sabri and MacDonald (2010) analyze the relationship of
savings behavior and financial issues among college students in Malaysia.
They find that financial experience prior to college often fosters poor habits
(Sabri & MacDonald, 2010). As the majority of students first experience
financial independence at the university level, there is overall low financial
literacy among the participants. The sample consists of both private school and
public school students, which later proves to be a significant factor in the study
(Sabri & MacDonald, 2010). Participants that come from private schools are
more likely to come from wealthier backgrounds, which can account for the high
volume of spending among these students (Sabri & MacDonald, 2010).
Moreover, Sabri and MacDonald (2010) are also able to identify that those of
Chinese descent are a specifically wealthy 18 population in Malaysia, and
much of the spending is linked to this group of students. Overall, respondents
in this sample are more prone to spending than saving; more than half of the
respondents choose to spend money that is received for scholarships or
education loans (Sabri & MacDonald, 2010). Often, this money is spent on
personal shopping, most of which is consumed before the end of one semester
(Sabri & MacDonald, 2010). The consumer culture is not only growing rapidly
in the U.S. but also seems to have taken shape in both developed and
developing economies around the globe. Phau and Woo (2008) investigate
money attitudes and credit card usage among Young Australians using a mall
intercept method in a popular shopping complex. Participants of the study are
administered an eight-question survey which asks about demographics, money
attitudes, compulsive buying habits, credit card usage and shopping patterns
(Phau & Woo, 2008). It is found that young adults tend to associate money with
a high-status image (Phau & Woo, 2008). Frequent spending habits are
associated with an individual’s desire to achieve a certain social status.
Moreover, Phau and Woo (2008) identify cultural and social norms that may
have varying effects when the study is conducted in different countries. There
exists both present oriented and future oriented societies, which can be a
strong determinant in whether individuals are more likely to spend or save
(Phau & Woo, 2008). As Australia is a melting pot of cultures, the observed
attitudes towards money matters are varied (Phau & Woo, 2008). Researchers
conclude that attitudes and behaviors toward spending and saving are a
function of both age and cognitive maturation (Phau & Woo, 2008). The
younger a student, the less they are inclined to save, as there is no immediate
worry of covering financial costs (Phau & Woo, 2008). There is regard for
attaining a constant stream of income that will account for accumulated debt.
This 19 finding by Phau and Woo (2008) provides support for the Permanent
Income Hypothesis. Younger consumers are less mindful of covering costs
now, as they anticipate earning money from a future employer that will allow
them to smooth out consumption habits over time. However, the sample was
taken from a single, homogenous Australian population in a popular shopping
complex. It is unknown whether there is an environmental effect factoring into
the results of this study. Conducting a study on financial attitudes and behaviors
in a shopping mall may have adverse effects on consumer responses.
Considering that many of the participants had made or were planning to make
a purchase can influence their views on their personal habits, skewing the
results of this study.

In a study conducted in London, Furnham (1999) observes the spending and

saving habits of British adolescents. It is interesting to note the findings of such
a study, as not many researchers have attempted to investigate the financial
habits of children. Focusing on a younger age bracket may provide insights into
reasons as to why college students spend or save the way they do. Furnham
(1999) is able to suggest why an individual may be more susceptible to
spending, as early exposure to certain attitudes and parental treatment can
largely factor into the development of spending habits. The study on British
children asks participants to complete a questionnaire which asks about
sources of income, how much money is generally put into savings, where it is
stored and the purpose it is intended for (Furnham, 1999). The main
demographics Furnham (1999) focuses on are gender, age and class, with the
first two proving to be highly significant. Researchers conclude that age is the
most powerful predictor of saving (Furnham, 1999). The older a child is, the
more money he or she will receive and save. In terms of gender, females are
better at money management, as they are less comfortable with 20 handling
debt (Furnham, 1999). However, this could be due to differences in
socialization, as it is found that at a younger age, boys are receiving more
pocket money and are allowed to take on part-time jobs before girls (Furnham,
1999). This finding by Furnham (1999) may explain the gender differences that
appear within multiple studies on financial attitudes. The socialization and
upbringing of boys in comparison to girls builds a separate framework for
handling money issues. Finally, social class differences appear to be a difficult
demographic to measure. It is predicted that higher socioeconomic status
implies more savings, but the sample turned out to be a homogenous
population of children from middle class backgrounds (Furnham, 1999). This
limitation to the study did not allow for full investigation of the range of
demographics that were initially intended for study. After examining a broad
range of literature, it can be concluded that the main contributions of my study
stem from a connection of economic theories to the spending and saving habits
of college students. Many researchers have focused on examining different
variables that may have different effects on a college student’s financial habits,
but few have analyzed whether the results have shown support for existing
theories that account for consumer behavior. More specifically, theories of
Hyperbolic Discounting and the Permanent Income Hypothesis serve to
understand financial attitudes among consumers, and minimal research has
been conducted to further investigate this relationship within the context of
 personal finance. My study seeks to provide a bridge between the spending
and saving habits of college students and the theoretical framework behind
consumers’ financial habits.

IV. METHODOLOGY & DATA Data Collection For the purposes of this study, I
designed my own survey using Qualtrics and distributed the questionnaire to
students via convenience sampling at Skidmore College across all class years.
Convenience sampling is a method which recruits volunteers to participate in a
study, selected due to their availability and easy access. Students were
approached in common spaces around campus including the library and Case
Student Activity Center, and those who agreed to participate were emailed the
link to the online survey. A conscious effort was made to capture the
demographics of the population. To account for this, students from various
ethnicities and genders were asked to participate, such that it is a reflective
sample of the Skidmore College student body. For instance, approximately
70.6% of students at Skidmore identify as White, followed by Asian at 6.2%,
then Black at 4.4%, all of which were factored into the recruitment of
participants (Skidmore College - CollegeData College Profile). The data
collection process was conducted over a two-week period, where a total of 66
surveys were sent out and 55 complete responses were recorded – a response
rate of 83.3%. Survey measurement instrument The seven-minute survey was
divided into three separate sections. The first asked about demographic factors
such as class year, gender and ethnic background. Students selected the
choices that best reflect how they self-identify under these categories. The
second section consisted of a set of four randomized questions that asked
about financial sources, satisfaction with current spending, plans for future
spending and Hyperbolic Discounting. The question 22 pertaining to financial
sources involved a multiple-choice selection with options being a) I receive
most of my money from an allowance (from parents, guardians, etc.) b) I earn
money from an employer c) I have a debit card and d) I have a credit card.
Participants were then asked subsequent questions on whether they planned
for future spending as well as satisfaction with current spending. Both were
measured on a 5-point Likert scale with options ranging from Always, Often,
Sometimes, Seldom and Never. Hyperbolic Discounting was measured by
asking if a student would prefer to receive $15 today or $20 tomorrow. The final
section was a set of five randomized questions regarding most frequently
purchased items, financing of leisure expenditures, a numerical average of
monthly spending, changes of spending and saving that may have occurred
over college career, and the Permanent Income Hypothesis. To understand
what students spent most of their budget on, participants were asked to rank
among five categories in the order of 1, being most frequently purchased, to 6,
being least frequently purchased. Options provided were clothing, electronics,
cosmetics, entertainment, and food and beverage. Individuals were also asked
what source they used to make personal or leisure purchases, and the survey
provided options including money from a parent or guardian, money earned on
their own, credit card and other. Students were also asked to estimate average
monthly spending and were given options such as a) Less than $100 b) $100-
199 c) $200-399 d) $400-599 or e) $600+. The final two questions were posed
in a graphic slide manner, which asked participants to select an option on a
given numerical scale that ranged from 0 to 100. The first question asked, “On
a scale of 0 (no change at all) to 100 (lots of change), have you experienced
changes in your spending or saving habits over your time at Skidmore?” The
second question asked how much participants agreed with 23 the statement, “I
am willing to spend more now because I anticipate getting a decently paid job
after graduation”. 0 indicated strongly agree while 100 denoted strongly
disagree. Model After collecting results from the survey questionnaire, two
baseline models were developed. The models were built upon specific factors
being tested within the survey instrument. Upon initial analysis, it appeared that
spending had a much more significant effect than saving on a college student’s
financial habits. When asked to consider changes in both saving and spending
that have occurred over an individual’s time in college, there were more
reported changes in spending while few offered the same degree of changes
in saving. Specifically, 25.5% of participants recorded responses in the range
of 50-100, indicating significant changes in spending. Conversely, when asked
about changes in savings, only 16.4% of respondents reported much change
in savings. As a result, I have chosen to focus the analysis on spending habits.
I utilized two probit regressions as my baseline models. Model 1 tests the
effects of class year, gender and race on average spending while Model 2
analyzes two economic theories of consumption, namely Hyperbolic
Discounting and the Permanent Income Hypothesis, in relation to average
monthly spending. The separation of variables allows for different types of
analyses that focus either solely on demographics or theoretical framework.
Pr (���_�����.) = �(�3�������. + �9������. +
�:����. + e.) (1) Pr (���_�����.) = �(�3��. + �9���. + e.)
(2) where HD denotes Hyperbolic Discounting and PIH indicates Permanent
Income Hypothesis. 24 Measurement of Variables The dependent variable,
which analyzes average spending, was measured through a multiple-choice
question that asked students to estimate their average monthly spending.
Options given to students included a) less than $100 b) $100-199 c) $200-399
d) $400-599 or e) $600+. The results were aggregated and categorized to run
the probit regression model. It was found that most of the respondents
answered either d) $400-599 or e) $600+. Thus, those results were taken to
create the dummy variable ���_�����, where responses that were
either d) $400-599 or e) $600+ were assigned a value of 1, and 0 otherwise.
The independent variables each correspond to a specific factor being tested in
relation to an individual’s spending mechanisms. The variable �������
accounts for whether students in different class years currently attending
Skidmore College have different effects on spending. This provides a
comparison among age groups to test spending mechanisms. The variable
������� included the Classes of 2020, 2019, 2018 and 2017. I predict
that younger students, or freshmen, would have a positive relationship with
average spending, while older class years would have a negative relationship
with average spending. At the onset, I predict that freshmen are only starting
to get acclimated with financial independence and may not be able to manage
spending as well as students in other class years. ������ tests for
differences in spending that may arise depending on whether the participant
was male or female. My predictions for ������ are that females are more
discretionary with spending as previous literature has shown them to employ
more savings mechanisms. This hypothesis is supported by the results of
Norvilitis et al. (2006). The third variable ���� is a composite of the various
ethnicities included within the testing parameters. The survey included options
of White, Black or African American, 25 Asian, Hispanic, American India or
Alaska Native, Native Hawaiian or Pacific Islander, and Other. Similar to the
findings of Chen and Volpe (1998), my predictions for ethnicity are that Whites
will spend more than other demographics while Blacks will spend less. Finally,
�� and ��� correspond with Hyperbolic Discounting and the Permanent
Income Hypothesis, respectively. The last two variables in the equation test for
the applicability of those economic theories in the context of personal finance.
Hyperbolic Discounting was tested by asking survey participants a question
with regards to their preferences for receiving a reward. In particular, the
question asked if the individual would prefer to receive $15 today or $20
tomorrow. To test the application of the Permanent Income Hypothesis,
students were asked to rank their sentiments toward a statement. It was framed
in such a manner that asked participants how much they agreed with the
statement, “I am willing to spend more now because I anticipate getting a
decently paid job after graduation”. The question was presented in a graphic
slide manner on a scale ranging from 0 to 100. Along the scale, there were
various markers that indicated seven options from Strongly Agree to Strongly
Disagree. I predict that there will be some evidence of both the Permanent
Income Hypothesis as well as Hyperbolic Discounting among the sample
population of college students. Researchers such as Angeletos et al. (2001)
and Laibson (1998) both find that at a younger age, consumers tend to be
present oriented and exhibit mannerisms that are in line with both theoretical
models. Tables 1, 2, and 3 provide summary statistics of the data collected
from the survey.

Based on the summary statistics, it is evident that the variable for ���� has
the highest standard deviation, indicating the largest variance of values. As
there are four different subsections to the variable race and a purposefully
disproportionate amount between each, it would make sense that ����
holds the highest standard deviation value. On the other hand,
���_����� and ������ both have fairly low standard deviations,
as these variables were encoded to be dummy variables. The summary
statistics also show that the breakdown of ��ℎ������ is such that it is
proportionate to the actual demographic representation of the Skidmore
College student body. Whites are the most prominent ethnicity, with a
representation of 51.8% among survey participants, followed by Blacks and
Asians, each composing 12.5% of the participants in the study. The ��ℎ��
category of ethnic background represents 23.2% of the participants, as there
were many respondents who reported a mixed ethnic background. In terms of
gender, there is a larger population of female students, as was also reflected
during the recruitment of participants. To draw some conclusions based on
class year, it was ensured that a fairly even number of participants were
recruited from each class. 27.3% of participants were from the Class of 2017,
25.5% were from the Class of 2018, and the Class of 2019 and 2020 comprised
23.6% each of the total responses. Data Cleaning Prior to conducting
regression analyses, the raw data collected from the survey were cleaned and
encoded. First, an initial sweep through the responses indicated that one
participant did not fully complete the survey. This single incomplete response
was deleted, so 28 as not to skew the results and only include a portion of this
individual’s answers. There also existed a column of text response for the
���� variable, where participants had the option to specify a more particular
ethnic background if they chose to do so. Since there were no recorded
responses in this column, it was also deleted. The raw data output also included
columns for amount of time taken to finish the survey, confirmation of consent,
and confirmation of submission, all of which were unnecessary and omitted in
preparation for the data analysis process. Since most of the recorded
responses were in non-numerical form, most of the data had to be encoded
into numerical values that denoted specific characteristics. The variable
��ℎ������ was initially encoded into ten separate categories, which
resulted in too many subsets to analyze. Therefore, the variable ���� was
created which assigned numbers 1-4 to correspond with different ethnic
categories; 1 = White, 2 = Black or African American, 3 = Asian, and 4 = Other.
The category Other is comprised of individuals who recorded responses of
Hispanic descent or mixed ethnicities. For the variable �������, the
same process was applied where 1 = Class of 2017, 2 = Class of 2018, 3 =
Class of 2019, and 4 = Class of 2020. With regards to gender, 1 corresponds
to female participants and 0 otherwise, resulting in the creation of a dummy
variable. The Permanent Income Hypothesis relates to the idea of individuals
exhibiting consumption smoothing mechanisms over their lifetime. To test for
this theory, participants were asked the degree to which they agreed with the
statement, “I am willing to spend more now because I anticipate getting a
decently paid job after graduation”. On a scale of 0 being Strongly Agree to 100
meaning Strong Disagree, survey participants were given a graphic slider 29
to indicate their attitudes toward the statement. Therefore, four separate value
ranges were established to aggregate the data for the regression analysis. A
variable named ��ℎ1 was created to correspond with values ranging from 1-
25, ��ℎ2 corresponds with values ranging from 26-50, ��ℎ3 corresponds
to values from 51-75, and ��ℎ4 indicates responses between 76- 100. The
encoding of data for Hyperbolic Discounting resulted in the creation of another
dummy variable, where variable ℎ� = 1 if the respondent answered $15 today
and ℎ�= 0 if the response was $20 tomorrow. About 74.5% of consumers
responded with a preference for receiving $20 tomorrow, indicating that there
was very little desire for a smaller, sooner reward. Instead, there was a strong
willingness to delay receiving money if it entailed pocketing a later but larger
reward. Though the numeric value is relatively small, waiting another day to
receive an extra $5 was hypothetically more enticing. Only an underwhelming
25.4% of participants appeared to be hyperbolic discounters. Graph 1: Actual
Spending vs. Perceived Spending of Sample Population
raph 1 shows the relationship between Actual Spending and Perceived
Spending. Actual Spending was measured by students who self-reported
average monthly spending, recording responses through five different options
including a) less than $100 b) $100-199 c) $200-399 d) $400-599 or e) $600+.
These five options correspond to the y-axis scale ranging from 1-5. Perceived
spending was measured by asking students their observed changes in
spending over their time at Skidmore. The question was posed in a likert-scale
manner where students could select an option on a range of 0-100, where 0
denoted no change at all and 100 denoted lots of change. The graph shows
that students have false perceptions of their spending habits, as there is no
trend present from these findings. After processing the raw data, I decided to
test not only demographics and theory alone, but also various combinations of
demographic characteristics as well as economic theories of consumption. To
my knowledge, previous literature has not accounted for testing a model that
incorporates both demographics and theory altogether. Model 1 tests ethnicity,
Model 2 tests for variances in gender, Model 3 observes class year, Model 4
accounts for the Permanent Income Hypothesis, Model 5 analyzes Hyperbolic
Discounting, Model 6 is a combination of demographic factors with Hyperbolic
Discounting, Model 7 combines demographics with the Permanent Income
Hypothesis, and Model 8 is a full aggregation testing all independent variables
against the dependent variable, average spending. The full model equation is
as follows: Pr (���_�����.) = �(�3�ℎ���. + �9�����. +
�:�����. + �K������.+ �N����ℎ���. +
�O���ℎ�����. + �P������. + �S��ℎ1. + �T��ℎ2. +
�3U��ℎ3. + �33ℎ�. + e. ) 31 Previous literature tends to utilize an Ordinary
Least Squares (OLS) regression model, such as the study of Hayhoe et al
(2000). The downfall to an OLS model in the context of my study is that it
assumes a linear probability, and in some cases, the predicted probabilities
may lie outside the boundaries of 0 and 1. This poses as an issue, as
probabilities need to be within the range of 0 and 1. A probit regression model
corrects for this issue by imposing a normal distribution assumption on the error
term. Probit models can thus bind the probability between the threshold of 0
and 1, to ensure that the results are applicable within the context of a probability
model. IV. RESULTS For this study, a probit regression analysis was employed
where ���_����� is the dependent variable. The marginal effects were
reported to interpret each independent variable using their sub-categorical
assignments. Results from all eight models are provided in the table below.

Ethnicity As discussed in the methodology section, the variable for ethnicity

was grouped into four categories consisting of students who are White, Black
or African American, Asian or Other. To compare spending habits cross-
culturally, the regression was run using the ��ℎ�� category as the baseline.
Results from the regression analysis reveal that Whites, Blacks, and 33 Asians
all spend more than the Other ethnicity group. The coefficients for �ℎ���
and ����� are significant at the 10% and 1% levels, respectively. Results
indicate that Whites spend 29.4% more, Blacks spend 31.8% more and Asians
spend 57.8% more than those grouped in the Other category. Permanent
Income Hypothesis The survey question pertaining to the Permanent Income
Hypothesis was designed using a scale bar, such that those who recorded a
response between 0-49 are deemed to be in support of the hypothesis while
those between 50-100 were not. When observing the data, it was found that
52.7% of survey respondents fell within the threshold of 0-49 on the scale, and
47.3% of survey respondents fell within the threshold of 50-100 on the scale.
More than half of the participants err on the side of agreement with the
statement, warranting support for the Permanent Income Hypothesis. To run
the regression analysis, the upper hand threshold of ��ℎ4 was used as the
baseline. Results showed that there is statistical significance at the 5% level
for the threshold of values in ��ℎ4. Compared to respondents who are
currently less willing to spend now despite the possibility of greater future
income, those who are more willing to spend now, will spend on average 38.2%
more. Class Year Collecting data from students in various class years allowed
the results to be representative of individuals at different ages and phases of
their college careers. This allows for the analysis of age and its relationship
with personal finance. When comparing average 34 spending amounts
between those in the Class of 2017 and those in the Class of 2020, the data
revealed that 66.7% of fourth year students indicated spending anywhere
between $100-399 while 62.5% of first year students indicated spending
anywhere between $400-600+ on an average monthly basis. In addition, 66.7%
of fourth year students noticed significant changes in their spending habits over
their time at Skidmore. When running the regression model, the Class of 2017,
or senior students, was used as the baseline for comparative analysis. Results
show that freshman are 6.8% more like to spend than seniors. On the other
hand, sophomores are 8% less likely to spend and juniors are also less likely
to spend at a rate of 9.8%. These results show that there is reason to believe
that first year students spend significantly more within this demographic
framework, as predicted. Although these results were insignificant, the signs of
the coefficients are as expected, where freshman students were positively
correlated and sophomore and juniors were negatively correlated with average
spending. Gender Data for gender consist of a split between 25 male
participants and 30 female participants. These numbers for the gender variable
are aggregated across all class years and ethnicities. Between males and
females, data supports the idea that males are more likely to spend more in a
given month than females. More precisely, results reveal that females are 4.1%
less likely to spend than the average male student, a finding that was expected.
Again, the sample population, consisting of 54.5% female and 45.5% male, is
representative of the more populous female demographic of Skidmore College.
35 Hyperbolic Discounting The idea that consumers are hyperbolic discounters
and make decisions based on time preferences does not seem to be an initial
trend within the scope of personal finance literature among college students.
When running the regression model, it was found that for every incremental
increase of hyperbolic discounting, average spending increases by 9.5%.
However, this finding proved to be insignificant. There may also be a limitation
to the design of this particular survey question in that the time lag of one day is
too small. Given this short time frame, the value of $5 is arguably too high. Both
of these factors may have contributed to a poor question design, ultimately not
able to capture the intended theoretical nature of the study. Demographic and
Theoretical Models Models 6, 7, and 8 all combine demographic characteristics
with theoretical framework. These models test for the significance when
accounting for both demographics and Economic theory, which to my
knowledge, has not been previously tested in related literature. When looking
strictly at demographics and Hyperbolic Discounting in Model 6, there is
significance among ethnicity and class year. The coefficient for �ℎ��� is
32.4%, significant at the 10% level, and for ����� is 72.8%, significant at
the 1% level. For the analysis on ethnicity, the variable group ��ℎ�� was
used as the baseline. Interpreting these results, �ℎ��� has increasing
probability of average spending by 32.4% while ����� has increasing
probability of average spending by 72.8%. The result for ������ is also
significant at the 10% level but with a 36 decreasing probability by 23.7% in
relation to average spending. Hyperbolic Discounting results were insignificant,
but positively correlated with average spending, as expected. Model 7
examines demographics and the Permanent Income Hypothesis. The
coefficients for ����� and ������ are positive and significant at the
1% level while the Permanent Income Hypothesis also shows positive
significance at both the first and second thresholds. The variable �����
has increasing probability of spending by 81.2% and ������ has
decreased probability of spending by 25.9%. ��ℎ1 is significant at the 10%
level while ��ℎ2 at the 1% level. Those in the threshold of ��ℎ1 have
increased probability to spend by 55.% and those in the threshold of ��ℎ2
have increased probability of spending at a rate of 45.5%. The full model,
aggregating all 11 independent variables, shows significance among the same
variables discussed in Model 7. Again, ����� and ������ are
positive significant, this time with increased probability to spend at 80.6% and
decreased probability to spend at 27%, respectively. However, in the full model,
������ is positively significant at the 5% level while ����� still holds
significance at 1%. ��ℎ1 and ��ℎ2 show positive coefficients at the level of
5% significance, both with increased probabilities of 57% and 44.3%. An
interesting finding to note is that in each model, the variables ����� and
��ℎ2 remain significant, confirming that these variables are strongly
correlated with average spending.

Savings Similar regressions were also run to observe any patterns in savings
that may be present among the sample population. While there was still
significance present among certain variables, changes in saving habits did not
seem to be as present as that of spending habits. The dependent variable,
���_����, was measured through the question that asked about 37
perceived changes in habits over time at Skidmore. Students were asked to
report their perceptions of changes in both spending and saving on a scale of
0 to 100, with 0 indicating no change at all and 100 indicating lots of change.
Results from the regression analysis are shown in table 5 below.

The same set of specifications was utilized to study savings behavior. Model 1,
focusing on ethnicity, shows that the coefficient for ����� is positive and
significant at the 10% level. Specifically, black students are 30.5% more likely
to save than other races. Model 3, observing differences among class years,
finds that the coefficient for ������ is positive and significant at the 10%
level, with juniors 25.5% more likely to save than other class years. Model 4
investigates average savings versus the Permanent Income Hypothesis, with
both ��ℎ1 and ��ℎ2 resulting in positive and significant coefficients at the
5% level. In Model 5, the result for Hyperbolic Discounting is positive and
significant at the 10% level. Model 6, looking at demographics and Hyperbolic
Discounting, find significance for ����� at the 5% level with a coefficient
of 34.1%. This indicates blacks are 34.1% more likely to save in comparison to
 other ethnicities. Model 7 analyzes demographics and the Permanent Income
Hypothesis, where ��ℎ1 and ��ℎ2 are both positive and significant at the
5% level. Students in the threshold of ��ℎ1 are 32.1% more likely to save
while those in the threshold of ��ℎ2 are 30.4% more likely to save. Finally,
the full model results show that Blacks are more likely to save at a rate of
29.3%, a finding that is significant at the 10% level. ��ℎ1 is also positive
significant at the 5% level with a coefficient of 32.1%. When comparing the
results of spending and saving, there are more statistically significant results
within the spending models versus the savings models. This is in line with the
students’ responses, where 25.5% of students observed lots of change in
spending while only 16.4% of students observed lots of changes in saving
habits over their time at Skidmore. Some patterns have emerged when
interpreting the regression results. The spending models display significance
of the variable ������, which is negative and significant in each model it
is used. 39 This is indicative of the fact that juniors spend the least in
comparison to other class years. This can be confirmed when looking at the
savings regression results, where ������ is positively correlated with
savings in model 3. The finding can suggest that juniors spend less on average
because they have observed more implementation of savings mechanisms.
Similarly, the variable ����� is also significant in all spending regressions
that it is utilized, each with a significance at the 1% level. This demographic is
strongly correlated with average monthly spending. Another interesting result
to note is that ����� is significant in models 1, 6, and 8 of the savings
models. This supports the idea that Blacks are more likely to save as opposed
to students of other ethnicities. There is also no significance of the variable
����� within the 8 models of spending, which further confirms the findings.

V. DISCUSSION Much of the previous literature focuses on different variables that

may influence the spending and saving habits of a college student, but few
researchers develop a model that accounts for both demographic and
theoretical values altogether. While age, gender and ethnicity have all shown
to largely influence financial behaviors of a college student, little has been done
to examine the role of certain theoretical frameworks including the theories of
Hyperbolic Discounting and the Permanent Income Hypothesis. This study
examines spending and saving behaviors among college students, taking into
account variables of class year, gender and ethnicity in one model and the
Permanent Income Hypothesis and Hyperbolic Discounting in a separate
model. 40 In terms of planning ahead for future spending, females were more
likely to plan for spending. When observing the number of responses that
reported planning for spending often, 16.4% of females selected that option as
compared to 7.3% of males. On average, females in this data set spend 4.1%
less than males. This finding is in line with other literature. Sabri and
MacDonald (2010) conclude that females employ more saving mechanisms,
which included planning spending budgets. This may relate to the socialization
and upbringing of females in comparison to males. In some societies, males
are given the freedom to begin working at a younger age whereas females are
held back until a certain age (Furnham, 1999). As males begin to earn income
earlier, they have more disposable income to spend as opposed to women who
are reliant on financial support from parents or guardians. Chen and Volpe
(1998) find that women, particularly those who are ranked lower among their
class and with little work experience, are less knowledgeable on personal
finance and therefore tend to develop wrong opinions and execute incorrect
financial decisions. However, this statement seems to be a cultural bias. To
claim that an individual has “wrong opinions” is a judgment against certain
cultural norms and standards that may not hold true in every societal structure.
Perhaps many of these individuals were enculturated under different
mannerisms, where spending is revered. Furnham’s (1999) study reviews
money pathology, which shows that males report greater confidence,
independence, risk taking and gambling with money matters. This may apply
to findings from this study, explaining why males were found to spend more on
average than females. From an ethnicity standpoint, Cummins et al. (2009)
claim that American students have cherished the use of credit more. The credit-
dependent society is often discussed as a growing 41 problem within the realm
of personal finances. Relying on credit usage can lead to financial issues such
as incurring large amounts of debt (Cummins et al. 2009). Many college
students tend to be present oriented in that they are not concerned with
covering the costs of credit card spending, mostly due to the fact that at this
age, students are dependent on parents and guardians to pay off these
balances (Hayhoe et al., 2000). The regression results of class year from this
study find that freshman and senior students exhibit higher spending behaviors
while sophomores and juniors exhibit less spending. This finding may be in line
with the fact that first year students are younger and therefore do not know how
to handle finances well (Chen & Volpe, 1998). There is a learning curve that
exists when making the transition from being completely financially dependent
to slowly becoming financially independent. Chen and Volpe (1998) believe that
participants with less work experience, many of whom are younger, also are
unable to manage finances. An interesting finding from this study is the fact
that fourth year students are also spending more on average. The transition
from college to post graduation may also probe more spending in preparation
and anticipation of a higher income. Within the context of the Permanent
Income Hypothesis, research has found that younger individuals seem to be
more optimistic about their future financial earnings, which can be a good
indication of the reason why they are able to take on more debt now, expecting
to pay it off later (Norvilitis et al., 2006). Similarly, Roberts and Jones (2001)
find compelling evidence for the Permanent Income Hypothesis through their
review of the UCLA/American Council on Education Annual Survey. Three out
of four students said that one of the main reasons for going to college was to
make more money (Roberts & Jones, 2001). 42 In an analysis of the Hyperbolic
Discounting Function, David Laibson (1998) suggests that age, income, and
wealth are all correlated with various levels of patience. In other words, at
different stages of the life cycle, there is a certain preference of present
orientation or future orientation. Accounting for age, income and wealth at the
university level, Hyperbolic Discounting may not be as applicable. This study
finds that, in general, most people do not display habits that are reflective of
Hyperbolic Discounting. About 74.5% of participants opted for the choice to
receive $20 tomorrow over $15 today. The population tended towards a later,
higher reward amount, indicating future oriented thinking. The study comes with
inherent limitations that can be addressed in future studies. First, there is the
issue of endogeneity that has surfaced after conducting this study. It cannot be
completely determined whether the independent variables are the variables
with the confounding effects on the dependent variable. The issue, also known
as reverse causality, indicates that there is a constant feedback loop to indicate
if the independent variable is impacting the dependent variable, or if this
relationship exists in the opposing direction as well. For instance, there is no
way to completely determine whether the relationship strictly exists in the sense
that average spending is affected by class year, gender, ethnicity, PIH and HD,
or if the opposite could happen. It is possible that average spending can result
in changes in certain variables such as the Permanent Income Hypothesis and
Hyperbolic Discounting. The theoretical frameworks can have a confounding or
unexpected effect based on changes in average spending. To correct for this
issue, an instrumental variable can be introduced, which does not correlate with
the error term but instead correlates with the independent variables. 43
Secondly, the sample size is small, with a total of 55 responses. As
convenience sampling was used and no monetary incentive was provided, this
may have decreased the probability of gaining more participants. However,
while only 55 students participated, it was ensured that an equal representation
of class years was present and that there were representations of gender and
ethnicity reflective of the Skidmore College population. Thirdly, there may have
been some selection bias that inevitably played into this study’s design.
Participants were carefully selected via convenience sampling at common
spaces around campus, however, students were asked to participate in the
study based on demographics. Since demographic characteristics are a large
focus in this study, it was important that survey participants came from a variety
of combinations in age, gender, and ethnicity. As such, students were first
asked their class year and ethnicity prior to recruiting them to participate in the
proposed study. Though selection bias was present, it was necessary to
account for demographics, as it was a large focus in the context of this study.
VI. CONCLUSION The spending and saving habits of college students provide
an insight into the financial mechanisms that are utilized by young adults.
Results from this study show that there are clear patterns that have emerged,
which are in conjunction with findings captured by other researchers.
Conclusive evidence is present of the fact that ethnic background is a strong
determinant of certain spending patterns. As deemed by other researchers,
namely Chen and Volpe (1998) and Pritchard et al. (1989), students who are
White tend to spend more than other demographics. Not only is this further
exemplified in my study, but it is also found that 44 Asian students also spend
a significant amount more compared to other ethnic backgrounds. In an effort
to find a connection to existing economic theories of spending, there was
significance behind the data collected for the Permanent Income Hypothesis. If
this attitude towards spending holds true and continues on an upward trend,
students may start to overestimate future earnings, resulting in more financial
issues to deal with at a later time. The results of this study provide various
implications and policy suggestions that can contribute to the literature of the
spending and saving habits of college students. As it stands today, the breadth
and depth of studies can be extended to further analyze other variables that
may have significant effects on the financial habits of college students.
Demographic factors such as age, gender and ethnicity seem to be most
commonly studied while many theoretical frameworks of consumption and
savings have not. The findings pertaining to existing economic models of
Hyperbolic Discounting and the Permanent Income Hypothesis can be
extended over longer periods of time. For instance, if a study was able to follow
a population of students precollege and post-college, this may give a better
understanding of the changes that occur within the time frame of university
education. This would come with observed perceptions prior to college that may
influence financial habits as well as practices that were developed during this
period. The scope of this literature also fails to take into consideration habitual
spenders, and how these individuals may affect the results. Future studies
should account for categorization of types of spenders in order to compare
findings and draw conclusions about financial practices among different
spenders. As more focus is being drawn towards studying the financial habits
of young adults, there is increasing desire to understand the issue and the main
45 driving forces that lie behind the development of financial habits. It would be
interesting to note the impact of formal education on the spending and saving
habits of college students. Very minimal research has been conducted in this
particular branch of the topic, and doing so could shed light on methods that
allow students to develop good financial habits. Most young adults have their
first sense of financial independence during their college years, and having no
prior knowledge of experience may have adverse effects in the future.
References Angeletos, G., Laibson D., Repetto, A., Tobacman J., & Weinberg, S. (2001). The
Hyperbolic Consumption Model: Calibration, Simulation, and Empirical Evaluation. Journal of
Economic Perspectives, 15(3), 47-68. Chen, H., & Volpe, R. (1998). An Analysis of Personal
Financial Literacy Among College Students. Financial Services Review, 7(2), 107-128. Furnham,
A. (1999). The Saving and Spending Habits of Young People. Journal of Economic Psychology,
20(6), 677-697. Gutter, M. S., Garrison, S., & Copur, Z. (2010). Social Learning Opportunities
and the Financial Behaviors of College Students. Family and Consumer Sciences Research
Journal, 38(4), 387-404. Hayashi, F. (1982). The Permanent Income Hypothesis: Estimation
and Testing by Instrumental Variables. Journal of Political Economy, 90(5), 895-916. Hayashi,
F. (1984). The Permanent Income Hypothesis and Consumption Durability: Analysis Based on
Japanese Panel Data, C(4), 1083-1113. Hayhoe, C. R., Leach, L. J., Turner, P. R., Bruin, M. J., &
Lawrence, F. C. (2000). Differences in Spending Habits and Credit Use of College Students.
Journal of Consumer Affairs, 34(1), 113-133. Laibson, D. (1998). Life-Cycle Consumption and
Hyperbolic Discount Functions. European Economic Review, 42(3-5), 861-871. 47 Norvilitis, J.
M., Merwin, M. M., Osberg, T. M., Roehling, P. V., Young, P., & Kamas, M. M. (2006).
Personality Factors, Money Attitudes, Financial Knowledge, and Credit-Card Debt in College
Students. Journal of Applied Social Psychology, 36(6), 1395-1413. Palan, K. M., Morrow, P. C.,
Trapp, A., & Blackburn, V. (2011). Compulsive Buying Behavior in College Students: The
Mediating Role of Credit Card Misuse. The Journal of Marketing Theory and Practice, 19(1),
81-96. Phau, I., & Woo, C. (2008). Understanding Compulsive Buying Tendencies Among
Young Australians. Marketing Intelligence & Planning, 26(5), 441-458. Pritchard, M. E., Myers,
B. K., & Cassidy, D. J. (1989). Factors Associated with Adolescent Saving and Spending
Patterns. Adolescence, 24(95), 711-723. Roberts, J. (2000). Consuming in a Consumer Culture:
College Students, Materialism, Status Consumption and Compulsive Buying. The Marketing
Management Journal, 10(2), 76- 91. Roberts, J. A., & Jones, E. (2001). Money Attitudes, Credit
Card Use, and Compulsive Buying among American College Students. Journal of Consumer
Affairs, 35(2), 213-240. Sabri, M. F., & MacDonald, M. (2010). Savings Behavior and Financial
Problems Among College Students: The Role of Financial Literacy in Malaysia. Cross-Cultural
Communication, 6(3), 103-110. Skidmore College – CollegeData College Profile.
http://www.collegedata.com/cs/data/college/college_pg01_tmpl.jhtml?schoolId=57