Content uploaded by Nasima Khatun
Author content
All content in this area was uploaded by Nasima Khatun on Mar 22, 2023
Content may be subject to copyright.
Am J Econ Sociol. 2022;00:1–17.
|
1
wileyonlinelibrary.com/journal/ajes
INTRODUCTION
Financial well- being (FWB), often known as financial freedom, is the ability to make decisions
that allow individuals to live a fulfilling life while still being able to satisfy their current and future
financial commitments. Personal well- being includes FWB, which can be improved regardless of
a person's income. Through a wide range of strategies and initiatives, hundreds of financial pro-
fessionals, including counselors, educators, coaches, planners, and others, assist customers every
day in navigating their financial possibilities and challenges. There is a growing understanding
that the main objective of these initiatives is to raise the beneficiaries' financial security.
Received: 16 October 2022
|
Accepted: 1 November 2022
DOI: 10.1111/ajes.12489
ORIGINAL ARTICLE
What are the determinants of financial
well- being? A Bayesian LASSO approach
Donald J.Lacombe
|
NasimaKhatun
© 2022 American Journal of Economics and Sociology, Inc.
School of Financial Planning, Texas Tech
University, Lubbock, Texas, USA
Correspondence
Donald J. Lacombe, School of Financial
Planning, Texas Tech University,
Lubbock, TX 79409, USA.
Email: donald.lacombe@ttu.edu
Abstract
The financial well- being (FWB) of individuals is a
topic that is becoming increasingly important across a
multitude of disciplines. In this study, we use the 2016
National Financial Well- Being Survey administered by
the Consumer Financial Protection Bureau to assess the
determinants of an individual's FWB. We identify 144
potential covariates that could explain variation in the
FWB score of individuals. The statistical methodology
of choice is the Bayesian LASSO, which is a covariate
selection algorithm that also allows for the importance
ranking of covariates. Out of the 144 potential covari-
ates, we find that 26 have 95% credible intervals that
do not contain zero. Broadly speaking, the results show
that objective measures of financial competency and
psychological and sociological factors contribute the
bulk of the explanatory power that help explain an indi-
vidual's FWB score.
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
This article provides empirical evidence of the determinants of the FWB score of individuals
using a Bayesian LASSO approach. The 2016 National Financial Well- Being Survey (NFWBS)
data administered and released by the Consumer Financial Protection Bureau (CFPB) are used to
measure the determinants of an individual's FWB score. According to the CFPB, enhancing FWB
is the ultimate goal of financial capability policies, programs, and interventions. Yet, despite the
critical importance of this outcome for assessing the effectiveness of financial literacy and ed-
ucation efforts, the fields of consumer finance and financial capability have operated without
an accepted definition or measure of FWB. For the first time, the CFPB research team explicitly
defined FWB from the consumer perspective as, “a state of being wherein you have control over
day- to- day, month- to- month finances; have the capacity to absorb financial shock; are on track
to meet your financial goals and have the financial freedom to make the choices that allow you
to enjoy life” (CFPB Report,2015). Using this definition, the CFPB developed a scale with a set
of 10 questions to measure the FWB score which we use as the dependent variable in our study.
The literature review section covers prior research related to the subjective and objective con-
cepts of FWB. Many studies define FWB using subjective and objective measures, oftentimes
using specialized surveys covering different target audiences, e.g., white- collar clerical workers
or young individuals. The differing samples and definitions of FWB create ambiguity in the re-
sults of these studies and drawing overall conclusions from the literature can be challenging. The
next section discusses the CFPB dataset, which is a cross- sectional dataset with 6394 observa-
tions and 217 explanatory variables. After cleaning the dataset, the remaining 144 variables are
used in the empirical exercise to measure the determinants of the FWB score.
Given the large number of covariates and the lack of a clear theoretical motivation for covari-
ate inclusion, the next section introduces the Bayesian LASSO, which is widely used to perform
variable selection when the candidate set of explanatory variables is large. The results from the
Bayesian LASSO technique are shown in the subsequent section. We find that 26 out of 144 po-
tential covariates in determining an individual's FWB score have 95% credible intervals that do
not contain zero, and thus have explanatory power in terms of the FWB score. Additionally, we
are able to rank the importance of the covariates and the results broadly fit into two categories,
one being objective financial measures and the other being psychological/sociological factors.
The final section offers conclusions and policy recommendations.
LITERATURE REVIEW
In recent times, FWB has received rising attention as a topic for individuals, households, socie-
ties, and countries. Previously, FWB has been studied in multiple disciplines, including econom-
ics, financial planning and counseling, consumer decision- making, amongst others. However,
no universally agreed- upon definition or measurement of FWB has yet been conceptualized that
might fit in all academic fields (Brüggen et al.,2017).
“Financial well- being is about effectively managing your economic life. People with high fi-
nancial well- being manage their personal finances well and spend their money wisely” (Rath
et al.,2010, p. 47). Porter and Garman(1992) defined FWB as objective and subjective aspects
of the financial situation evaluated against comparison standards to form a person's opinion of
his/her financial situation. Similarly, Vosloo et al.(2014) defined FWB as an objective and sub-
jective concept that contributes to a person's assessment of his/her current financial situation.
By surveying 9057 employees from different sectors in South Africa, the study measured the
relationship that subjective measures of financial efficacy and satisfaction with remuneration
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
3
DETERMINANTS OF FINANCIAL WELL BEING
significantly impact personal FWB. The study found that personal financial efficacy and satisfac-
tion with remuneration strongly correlate with personal FWB. The study also revealed that the
relationship between satisfaction with remuneration and FWB significantly impacts people with
higher personal financial efficacy.
Prior research viewed FWB as a composite concept with both objective and subjective compo-
nents. At the same time, other studies redesigned factors for these objective and subjective aspects
of FWB. For example, Shim et al.(2009) examined young people's FWB (i.e., debt, its relation to
financial satisfaction, financial worries, and coping). Here, FWB included the young people's
level of debt as a measure of their objective well- being and their satisfaction with financial status
as a subjective measure. Via an online survey, 781 observations were collected from a random
sample of students at a large state university in the southwestern United States. Initially, the mea-
surement model was developed by conducting a confirmatory factor analysis of the multi- item
scale. After that, they evaluated the validity of the construct. The authors also employed stochas-
tic regression imputation to deal with missing values. The results suggested that self- actualizing
personal values, financial education at home, and formal financial education at school can play
important roles in forming attitudes and behavioral intentions of young adults. Moreover, these
financial domains, along with parental normative expectations and young adults' perceived be-
havioral control, are related to their FWB. Likewise, Porter and Garman(1992) included finan-
cial variables such as income level as the objective measure of FWB and perceived satisfaction
with standard of living as the more subjective measure.
Other studies included income and other financial indicators as objective measures of FWB
without considering subjective variables. Joo and Grable(2004) utilized an exploratory frame-
work of the determinants of financial satisfaction. This study used a random sample of 220
white- collar clerical workers from a community in west Texas. The data were used to test the
robustness of the proposed framework of the determinants of financial satisfaction. A one- item
10- point stair- step question was used to measure respondents’ FWB. This self- anchoring scale
has its origins in research conducted by Cantril(1965). Path analysis (Pedhazur,1982) was used
to analyze the data. It was found that FWB is directly and indirectly related with diverse factors
including financial behaviors, financial stress levels, income, financial knowledge, financial sol-
vency, risk tolerance, and education. In another study, Kahneman and Deaton(2010) raised one
of the most debatable questions of whether money buys happiness, separately for two aspects of
well- being: emotion and life evaluation. They reported the findings based on an analysis of more
than 450,000 responses to the Gallup- Healthways Well- Being Index (GHWBI), a daily survey
of 1000 US residents conducted by the Gallup Organization (Harter & Gurley,2008). The study
aimed to examine possible differences between the correlates of emotional well- being and life
evaluation, focusing on the relationship between these measures and household income with
the help of a multiple regression model. They concluded that high income buys life satisfaction
but not happiness and that low income is associated with low life evaluation and low emotional
well- being.
Previously, other studies incorporated financial information, financial ratios, and benchmarks
as objective measures of FWB (Greninger et al.,1996), while others confirmed that a household's
ability to increase and manage liquidity, maintaining a balance between risk and return through
investment and ability to grow the value of household wealth can be used to determine its FWB
(Aggarwal,2014).
The review of the literature reveals that research on this topic remains somewhat ambigu-
ous. Even though researchers and practitioners characterized FWB in various ways, the term
FWB has not considered consumers’ perspectives on their financial lives. Given the financial
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
markets’ complexity, improving individuals' FWB is being prioritized as a salient goal of financial
education consistent with the US National Strategy for Financial Literacy (Financial Literacy
Education Commission,2016).
Recently, the CFPB‘s research team designed a set of research activities to understand and
formally define FWB. According to the CFPB, FWB is defined as “… a state of being wherein you
have control over day- to- day, month- to month finances; have the capacity to absorb financial
shock; are on track to meet your financial goals; and have the financial freedom to make the
choices that allow you to enjoy life” (CFPB Report,2015). As the definition reflects an inherently
subjective concept of FWB, the CFPB initially developed a scale with 10 set questionnaires. CFPB
administered and released the 2016 NFWBS data to measure FWB.
The literature measuring FWB with the help of the CFPB developed scale is quite large.
However, we have observed that limited research has been conducted using CFPB NFWBS data.
Fan and Henager(2021) developed a structural determinant framework to provide a broad
understanding of FWB. They used the 2018 National Financial Capability Study and structural
equation modeling methods to examine FWB determinants. The FWB variable was constructed
following the CFPB scale. The structural framework incorporated financial perception and
knowledge factors, financial stress, positive financial behaviors (comprised of positive short- and
long- term financial behavior), and financial satisfaction. The study's findings indicate that finan-
cial satisfaction, short- term financial behavior, and perceived financial capability showed posi-
tive and direct associations with FWB, whereas financial stress and long- term financial behavior
were negatively and directly associated with FWB.
In a recent study, Lee et al.(2020) investigated the relationship between financial knowledge
and FWB and added the moderating role of propensity to plan. This study utilized an Ordinary
Least Squares (OLS) regression model to analyze the relationship between financial knowledge,
the propensity to plan, and FWB while controlling for various socio- economic characteristics,
including age, education, race/ethnicity, employment, employment status, and marital status.
Finally, the authors found that having a better knowledge of fundamental financial concepts and
their application increased the level of FWB. Subsequently, the moderating effect of propensity
to plan increased the positive effect of financial knowledge on the level of FWB. Even though
the 2016 NFWBS pointed out that financial knowledge and a propensity to plan have a positive
association with FWB, this study confirmed that the propensity to plan plays a moderating role
in enhancing the positive association between financial knowledge and FWB.
In another study, Michael Collins and Urban(2020) demonstrated how this broader mea-
sure of FWB can offer insights beyond traditional measures and how subjective FWB can
potentially deepen our understanding of households' financial health. To understand how
this FWB scale score differs from other common measures of financial status, the study used
a cross- sectional dataset collected by the CFPB NFWBS. An OLS regression model was de-
veloped to measure the FWB scale score. The authors predicted that as people age, they have
established consumption and earnings patterns that will lead to higher levels of FWB scores.
Essentially, they wanted to analyze the FWB scale by age cohort in order to compare these pat-
terns to what they might predict based on standard life cycle explanations in household finan-
cial behaviors. Michael Collins and Urban(2020) were also interested in estimating financial
literacy, subjective well- being, and other demographic characteristics of the respondents, in-
cluding gender, race, the highest education level in the household, the highest education level
of the respondent's parents, the number of children under age 18, marital status, employment
status including gender, race, the highest education level in the household, the highest edu-
cation level of the respondent's parents, the number of children under age 18, marital status,
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
5
DETERMINANTS OF FINANCIAL WELL BEING
employment status (i.e., employed full or part time, unemployed, or retired from labor mar-
kets), and prior or current military service. The study concluded that FWB increases with
age, which is consistent with financial development as people's human capital improves and
savings accrues. All income levels showed a fairly normal distribution, with lower incomes
exhibiting lower levels of the FWB scale. Additionally, they have found that subjective well-
being and FWB are positively correlated.
Likewise, in another study, Patel and Wolfe(2019) investigated whether subjective well-
being is associated with FWB for self- employed individuals, and whether those with financial
skills are better able to leverage subjective well- being to realize higher FWB. They used a sam-
ple of 332 self- employed individuals in the CFPB's NFWBS. The empirical strategy used an
OLS regression model to test the proposed hypotheses that one, subjective well- being is posi-
tively associated with FWB for self- employed individuals, and two, financial skills strengthen
the positive association between FWB and subjective well- being, such that subjective well-
being will be more positively associated with FWB for individuals with high, rather than
low, levels of financial skills. The study showed a positive association between subjective
well- being and FWB, which is strengthened among those with higher financial skills. These
findings highlight the need for assessing the association between different types of well- being
from self- employment.
A joint research study (Nagypal & Tobacman,2019) between the CFPB and Credit Karma
examined how consumers’ subjective FWB relates to objective measures of consumers’ fi-
nancial health, specifically, consumers’ credit report characteristics. It was the first study of
its size to examine the relationship between an individual's FWB and credit score and other
credit report variables. The report also attempted to relate consumers’ subjective FWB to con-
sumers’ engagement with financial information through educational tools. For this project,
Credit Karma administered a voluntary survey to some of its members when they logged out
of their Credit Karma session. The survey consisted of the full 10- question version of the
CFPB's Financial Well- Being Scale. Of the 4067 completed survey responses, Credit Karma
provided matched credit report and engagement data for 2966 respondents, and the summary
statistical analysis was based on those 2966 matched observations. The study provided a better
understanding of these relationships and helped uncover the factors that work together to
determine consumers’ FWB and inform the CFPB's long- term strategy for improving financial
capability.
Hermann et al.(2020) examined the relationship of FWB with both the incidence of mort-
gage debt and housing cost burdens based on the CFPB's FWB Scale and, therefore, to as-
sesses the association between the use of mortgage debt among older adults and measures of
financial skill, knowledge, planning, and saving habits. This paper is a descriptive outcome
study employing the CFPB NFWBS, a public use dataset surveying over 6000 respondents.
Both descriptive and multivariate approaches have been used to analyze the survey data. To
estimate the relationship among the variables, an OLS model was utilized. The results indi-
cated that both higher monthly housing costs and the presence of high levels of mortgage
debt are associated with lower FWB, which is associated with a greater likelihood of material
hardship. The results also implied that older adults with higher financial skill scores and a
greater propensity to save are less likely to have higher levels of mortgage debt. The findings
suggest that reductions in mortgage debt among older homeowners can improve FWB and
that efforts to improve financial skills and encourage savings could potentially help bring
about such reductions.
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
DATA AND SAMPLE SELECTION
This study uses the 2016 NFWBS data, which were administered and released by the CFPB.
Previously, many studies have been conducted to understand the concept, components, and fac-
tors affecting FWB. Until recent work by the CFPB, FWB had not been explicitly defined from
the consumer's perspective. Therefore, the CFPB, a U.S. government agency, developed a defini-
tion of FWB and provided initial validation of a survey scale to measure individuals’ well- being
under this definition. Their definition of FWB reflects an inherently subjective concept, so it
cannot be observed directly.
The NFWBS is a US- nationally representative dataset consisting of 6394 respondents and 217
variables. Besides fielding the CFPB FWB questions, this survey also fielded questions of three
other financial scores: the Lusardi and Mitchell (2011) financial knowledge scale score, the Knoll
and Houts (2012) financial knowledge scale score, and the CFPB financial skill scale score(2018).
Some other features included in the dataset are based on selected individuals and households.
These include federal poverty level and demographic characteristics such as race/ethnicity, age,
household size, household income, and survey items belonging to financial behaviors, financial
attitudes, and personal experiences.
Out of 217 variables in the public use dataset, 10 questions were used to measure an individual's
FWB score, which were deleted from the list of possible covariates used in this study. Likewise,
another 10 set questions for scaling Financial Skill (FS) scores were in the dataset. These questions,
including summed up FS score, was deleted as our primary focus is on FWB. Moreover, three un-
labeled variables i.e., Public Use File ID, Sample ID for uniquely identifying each respondent, and
Final Weight, were deleted as these variables are unrelated to FWB measurement.
The dataset contains seven missing values represented by numeric codes. They are coded as
(1) if respondents refused to answer, (2) question not asked because respondent not in item base,
(3) invalid response, (4) response not written to the database, (5) the county is not known, 98; I
don't know, 99; Prefer not to say. We deleted 48 subjective variables from this dataset that con-
tains more than 50 missing values. Then we explored if there were any observations with missing
values. We applied the listwise deletion method for handling these missing values. Therefore,
437 out of 6394 responses (7%) were excluded from the dataset resulting in 5957 observations. In
summary, after cleaning up 72 variables in total, we obtained 144 potential explanatory variables
for analysis.
Econometric technique
As mentioned earlier, the CFPB survey contains a total of 217 possible explanatory variables,
which was reduced to 144 variables after the data cleaning process. Given the large number of
covariates, the empirical technique that we chose is the Bayesian LASSO, which is a technique
specifically designed to allow for covariate selection.
Tibshirani (1996) introduced a new technique for estimation in linear models called the
LASSO, which is an acronym for “least absolute shrinkage and selection operator.” The LASSO
technique minimizes the sum of squared residuals, subject to the constraint that the absolute
value of the sum of the coefficients is less than some arbitrary constant. The technique shrinks
some coefficients to zero and thus acts as a variable selection technique. The LASSO is a general
technique that can be utilized in many different types of econometric models, e.g., the normal
linear model and the logit model.
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
7
DETERMINANTS OF FINANCIAL WELL BEING
Park and Casella(2008), building upon the work of (Tibshirani,1996), developed a Bayesian
variant of the LASSO technique. The Bayesian LASSO places priors on the regression coefficients
which is shown to be equivalent to the frequentist version of the LASSO estimator. In a recent
paper, (Korobilis & Shimizu,2021) illustrate Bayesian approaches to shrinkage and sparse esti-
mation. This paper surveys modern shrinkage and variable selection algorithms and methodolo-
gies and compares different types of priors and their associated costs and benefits.
Another important feature of the Bayesian LASSO is that it also provides a mechanism to en-
gage in statistical inference using the posterior marginal distributions for the regression parame-
ters, which is not available in the frequentist version of the LASSO. For example, the 95% highest
posterior density intervals can be calculated to determine which marginal distributions for the
regression coefficients do not contain zero, and thus can be said to explain variation in the depen-
dent variable. Another feature of the Bayesian LASSO technique is the ability to rank the covari-
ates according to their importance via a methodology developed in Makalic and Schmidt(2011).
All Bayesian inference begins with the application of Bayes’ Theorem, which can be mathe-
matically expressed as follows:
where
𝜋(𝜃|y)
is the joint posterior distribution of the parameters,
𝜋(y|𝜃)
is the familiar likelihood
function,
𝜋(𝜃)
is the prior distribution for the parameters, and
𝜋(y)
is the marginal likelihood, which
is a normalizing constant that ensure that the posterior distribution integrates to one. Usually, we
work with the following form of Bayes’ Theorem:
where all the quantities are the same as before, but now it represents a proportional relationship.
Bayesian inference proceeds by examining the marginal distributions for each of our parameters
(e.g., the regression coefficients) by integrating the joint posterior distribution over the parameters
that we are not interested in. Performing this integration for each of our regression parameters re-
sults in a marginal distribution that we can use to perform inference.
In terms of the Bayesian LASSO methodology of Park and Casella(2008), our point of depar-
ture is the standard normal linear model, where we use their notation for convenience:
where y is an n × 1 vector of observations on the dependent variable, μ is the intercept term, X is
an n × k matrix of explanatory variables, β is a k × 1 vector of regression coefficients, and ε is a n × 1
vector if i.i.d. normally distributed errors, with mean 0 and variance σ2. The LASSO estimator can
be expressed as follows:
where
y=y−y1n
is the mean centered dependent variable. The LASSO procedure minimizes
the residual sum of squares as in the standard OLS algorithm; however, there is now a penalty
(1)
𝜋
(𝜃
|
y)=
𝜋(y|𝜃)𝜋(𝜃)
𝜋(y)
(2)
𝜋(𝜃|y)∝𝜋(y|𝜃)𝜋(𝜃)
(3)
y=𝜇1n+X𝛽+𝜀
(4)
min
𝛽(
y−X𝛽)T(
y−X𝛽)+𝜆
k
∑
j=1|
|
𝛽k
|
|
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
term that is added representing the absolute value of the sum of the regression coefficients. The
last term in Equation(4) is what gives the LASSO the ability to engage in variable selection be-
cause the minimization now has to consider the penalty term, with larger absolute value sums of
the coefficients being penalized.
Although the LASSO can be estimated in a non- Bayesian fashion, Park and Casella(2008)
show that the LASSO can be expressed in a Bayesian context by placing a conditional Laplace
prior on the regression coefficients of the following form:
After deriving the joint posterior distribution, Park and Casella(2008) note that the joint posterior
has no closed form solution. The solution is to obtain the set of full condition distributions for the
parameters of interest and use a Gibbs sampling routine to obtain the marginal distributions for all
of the parameters. Interested readers are encouraged to read Park and Casella(2008) for further
mathematical details regarding the derivation of the Gibbs sampling algorithm.
In this study, we utilized the R/MATLAB package “bayesreg” as developed by Makalic and
Schmidt(2016) to perform the empirical analysis. Their algorithm utilizes a standard Gibbs sam-
pling algorithm to estimate a posterior marginal distribution for each of the model parameters
using a Laplace prior distribution for each of the coefficients. The mode of the resulting marginal
posterior distribution for the regression coefficients is equivalent to the standard LASSO esti-
mator. Additional details regarding the “bayesreg” package and instructions on how to use the
package is available in Makalic and Schmidt(2016).
The dependent variable used in this study is the FWB scale score developed by the CFPB. The
scale is mainly based on four elements: (1) control over daily and monthly finances, (2) capacity
to absorb a financial shock, (3) on track to meet financial goals, and (4) the financial freedom
to make choices that allow enjoyment of life. The scale incorporates consumers’ perceptions of
FWB to deliver a single FWB score that captures the four elements of FWB. Individuals with a
high level of FWB feel that she or they can meet current and future financial obligations, have a
secure financial future, and can make choices that allow enjoyment of life. FWB was operation-
alized by asking 10 questions on a Likert- type scale that were incorporated to create a single score
that ranged from 0 to 100 with a mean of 54.3.
Explanatory variables
As mentioned earlier, our final dataset contains 144 potential covariates that would help to ex-
plain variation in FWB. Instead of specifying a model beforehand, we opted to let the LASSO
algorithm select the variables. Out of the 144 potential candidate variables, 26 variables were
found to have 95% credible intervals that did not contain zero, and thus help to explain variation
in the dependent variable.
Table1 provides summary statistics of the 26 variables. Central tendency (mean, median)
and dispersion (standard deviation, range) were measured based on the response values of each
of the variables. We observed that the difference between the mean and median of the response
values is small. We can infer that the data are symmetrically distributed. The smaller standard de-
viation for almost all the variables indicates that the response values are mostly centered around
(5)
𝜋
𝛽
𝜎2
=
k
j=1
𝜆
2𝜎
2
e−𝜆
𝛽k
∕𝜎2
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
9
DETERMINANTS OF FINANCIAL WELL BEING
TABLE Summary statistics of the variables selected by LASSO.
Parameter Mean Median SD Min Max
Financial well- being (F WB) score- dependent
variable
56.1561 56 0.1819 14 95
Difficulty in covering monthly expenses 1.4569 10.6306 1 3
Follow- through set financial goals 3.6211 40.8 917 1 5
Confidence in achieving financial goals 3.2375 30.7227 1 4
Satisfied with own life 5.3910 61.4248 1 7
Lots of stress in respondent's life 3.1509 3 1.0859 1 5
Household income 5.5630 62.6537 1 9
Savings habit 4.3923 51.4566 1 6
Age categories 4.4442 42.1071 1 8
Consult own budget to perceive the balance 3.6757 41.0325 1 5
Paid off monthly credit card balance 3.5535 41.5722 1 5
Couldn't afford medical costs 1.1936 10.4757 1 3
Confidence in ability to raise $2000 in
30 days
3.5387 41.4416 1 8
Frugality- being economical with things 5.2325 50.8678 1 6
Retired 0.2941 00.4557 0 1
Stayed within own budget 3.8169 41.0673 1 5
Ability to work toward long- term goals 3.0470 30.6806 1 4
Financial products currently have: Non-
retirement investments
0.3196 00.4664 0 1
Belief that ability to manage money is
unchangeable
3.7505 41.5861 1 7
Frequency felt not respected or mistreated
w/financial services
1.8088 2 0.7919 1 4
Consider the steps needed to stick to own
budget
3.6302 40.934 4 1 5
Respondent did not select any item in
SHOCKS bank
0.4924 00.5000 0 1
Poverty status 2.6718 30.6460 1 3
Checking/savings account at a bank/credit
union
0.8623 10.3446 0 1
Do research before making monetary
decisions
3.7230 41.0169 1 5
Received a large sum of money beyond
normal income
0.0660 00.2483 0 1
More than 18- year- old kids being supported
financially by respondent
0.2013 00.5240 0 2
N=5957
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
the mean. Min and max values represent the lower and upper limits, respectively, of the response
values of the variables.
RESULTS
The results from the Bayesian LASSO routine are contained in Table2. We ran the Gibbs sam-
pler for 100,000 draws and the first 1000 draws were eliminated to allow for “burn in” of the
sampler. The sampler also used “thinning” which only retains every 5th draw to eliminate cor-
relation among the draws. Our results in Table2 are based on the 19,800 retained draws. The
first column of Table2 is the variable description, followed by the mean and standard deviation
of the posterior marginal distributions for each variable. The next two columns in Table2 show
the lower and upper 95% credible intervals, indicating which marginal distributions that do not
contain zero and therefore contribute to explaining the dependent variable. The last column in
Table2 is the ranking of the importance of the explanatory variable as measured by the algorithm
as developed in Makalic and Schmidt(2011), with lower numbers indicating more important
explanatory variables.
Out of the 144 potential covariates, we find that there are 26 covariates that have 95% credible
intervals that do not contain zero, which indicates that the LASSO algorithm has successfully
performed covariate selection. For the sake of brevity, we discuss the five most important vari-
ables in determining an individual's FWB score.
According to the ranking algorithm, difficulty in covering monthly expenses is the most im-
portant variable in explaining FWB. The results indicate that as individuals move from situations
where monthly expenses are not at all difficult to cover, to a situation where covering monthly
expenses becomes more and more difficult results in a decrease in the FWB score of 5.5571,
which is almost a 6- point decrease in the FWB score. Given that the standard deviation of the
FWB variable is 0.1819, this represents an extremely large decrease in the FWB score. This result
is in line with previous research that shows that a consumer's ability to cover monthly expenses
can have a negative effect on the FWB score. The difference is that we are able to rank this as the
most important variable in determining the FWB score, a feature absent from previous studies.
The next most important variable, as calculated by the ranking algorithm, is following through
on setting financial goals. The point estimate is 2.1399 indicating that as individuals more ac-
tively follow through on setting financial goals, their FWB score increases. Our results provide
evidence that an individual's ability to follow through on goals is the second most important
factor in determining their overall FWB.
The third most important variable is an individual's confidence in achieving financial goals.
As individuals become more confident in their ability to achieve their financial goals, their FWB
score increases by 2.5381, a rather large increase. It appears that the second and third most im-
portant variables that determine an individual's FWB are related to the psychology of the individ-
ual and their ability to set goals and to follow through on these goals.
The fourth most important variable is determining an individual's FWB score is their satis-
faction with their life circumstances. As individual's life satisfaction increases, their FWB score
increases by 1.1703. As in the previous two variables, it appears that an individual's “state of
mind” is an important factor in determining their FWB. Our results bolster the argument that
non- financial stressors in a consumer's life can also affect their FWB. It appears that overall psy-
chological health is an important factor in a consumer's ability to manage their FWB.
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
11
DETERMINANTS OF FINANCIAL WELL BEING
Finally, the fifth most important variable according to the ranking algorithm is how much
stress an individual is experiencing. When the stress that an individual is experiencing increases,
their FWB score decreases by a value of 1.4344, a fairly substantial decrease. As mentioned
TABLE Bayesian LASSO results.
Variable Mean SD
Lower
95%
Upper
95% Rank
Difficulty in covering monthly expenses −5.5571 0.2384 −6.0258 −5.0906 1**
Follow- through set financial goals 2.1399 0 .1763 1.7923 2.4861 2**
Confidence in achieving financial goals 2.5381 0.2026 2.1441 2.9376 3**
Satisfied with own life 1.17 03 0.0893 0.9942 1.3448 4 **
Lots of stress in respondent's life −1.434 4 0.114 8 −1. 6618 −1.2107 5**
Household income 0.5502 0.0717 0.4106 0.6909 6**
Savings habit 0.8133 0.0970 0.6237 1.0046 7 **
Age categories 0.50 01 0.1498 0.2232 0.8132 8**
Consult own budget to perceive the
balance
−0.8523 0.134 4 −1.1152 −0.5900 9**
Paid off monthly credit card balance 0.5326 0.0893 0.3581 0.7079 10 **
Couldn't afford medical costs −1.4752 0.3072 −2.0739 −0.8707 11 **
Confidence in ability to raise $2000 in
30 days
0.4212 0.0811 0.2629 0.5803 12 **
Frugality- being economical with things −0.6837 0.1321 −0.9427 −0.4256 13 **
Employment status: Retired 1.4178 0.4386 0.5492 2.2739 14 **
Stayed within own budget 0.5348 0.1342 0.2712 0.7968 15 **
Ability to work toward long- term goals 0.7447 0.1937 0.3651 1.1234 16 **
Financial products currently have: Non-
retirement investments
1.0311 0.2661 0.5101 1.5538 17 **
Belief that ability to manage money is
unchangeable
−0.2888 0.0707 −0.4271 −0.1505 18 **
Frequency of feeling not respected or
mistreated w/financial services
−0.574 8 0.1423 −0.8544 −0.2955 19 **
Consider the steps needed to stick to own
budget
−0.4399 0.1663 − 0.7674 −0.1154 20 **
Respondent did not select any item in
SHOCKS bank
0.8290 0.3098 0.2220 1.4427 21 **
Poverty status −0.6529 0.2693 −1.1820 −0.1279 22 **
Checking/savings account at a bank/
credit union
1.1132 0.3622 0.4037 1.8252 23 **
Do research before making monetary
decisions
−0 .3175 0.1216 −0.5561 −0.0803 24 **
Received a large sum of money beyond
normal income
0.9362 0.4381 0.0924 1.8038 25 **
More than 18- year- old kids being
supported financially by respondent
−0.4442 0.2114 −0.8631 −0.0355 26 **
**Indicates variables with a 95% credible interval that does not contain zero. N=5957.
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
earlier, additional stressors in an individual's life aside from specific financial concerns can also
affect an individual's FWB.
The remaining 21 variables that the LASSO algorithm found to affect an individual's FWB
are a combination of objective financial factors (e.g., the ability to raise $2000 in 30 days) and
psychological factors (e.g., belief that ability to manage money is unchangeable), with additional
variables that could be describes as more demographic and sociological, such as poverty status
and age.
In summary, our results broadly indicate two main categories that affect an individual's FWB
score. The first are objective measures of FWB, such as the ability to cover monthly expenses,
paying off credit card debt, and the ability to pay off medical debt. The second main category of
variables that affect an individual's FWB are psychological in nature. For example, the ability to
set and follow through on financial goals is an important indicator as to whether an individual
will experience an increase or decrease in FWB. Other variables that could broadly be described
as psychological are the ability to work toward long- term goals, the inability of which can reduce
FWB.
CONCLUSION
In this study, we utilized a Bayesian LASSO method to determine what factors affect an in-
dividual's FWB score. The Bayesian LASSO is a covariate selection technique designed to
capture the most important explanatory variables, especially when there are many covari-
ates to choose from for analysis and there is a clear theoretical reason to include particular
explanatory variables.
Using data from the 2016 NFWBS administered by the CFPB, we find that out of the 144 po-
tential explanatory variables in the dataset, 26 of those variables have a 95% credible interval that
does not contain zero, indicating that the Bayesian LASSO technique was successful in selecting
covariates. Additionally, we also are able to rank the covariates by their importance.
The results also indicate that the most important explanatory variable fall into two broad
categories: the first being objective measures of FWB and the second being psychological
factors.
According to the Council for Economic Education, nine states require a standalone personal
finance course to graduate high school, and 14 others integrate personal finance topics in an
integrated fashion in existing courses.1 Given that several factors that affect FWB are related to
objective knowledge in personal finance, it may behoove school districts to increase the number
of students who study personal finance issues. Additionally, psychological factor such as the abil-
ity to set goals is also an important factor that helps explain FWB. Programs that are designed to
help consumers increase their FWB should cover both of these import aspects of the education
process.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon
reasonable request.
ORCID
Donald J. Lacombe https://orcid.org/0000-0003-0668-4047
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
13
DETERMINANTS OF FINANCIAL WELL BEING
ENDNOTE
1 https://www.counc ilfor econed.org/wp- conte nt/uploa ds/2022/03/2022- SURVE Y- OF- THE- STATES.pdf.
REFERENCES
Aggarwal, S. (2014). Developing an index for measuring financial well- being in a geography. IFMR Finance
Foundation.
Brüggen, E. C., Hogreve, J., Holmlund, M., Kabadayi, S., & Löfgren, M. (2017). Financial well- being: A concep-
tualization and research agenda. Journal of Business Research, 79, 228– 237. https://doi.org/10.1016/j.jbusr
es.2017.03.013
Cantril, H. (1965). The pattern of human concerns (Vol. 4). Rutgers University Press.
Consumer Financial Protection Bureau. (2015, January). Financial well- being: The goal of financial education.
https://www.consu merfi nance.gov/data- resea rch/resea rch- repor ts/finan cial- well- being/
Consumer Financial Protection Bureau. (2018, September). Measuring financial skill. https://www.consu merfi
nance.gov/data- resea rch/resea rch- repor ts/measu ring- finan cial- skill/
Fan, L., & Henager, R. (2021). A structural determinants framework for financial well- being. Journal of Family
and Economic Issues, 43, 415– 428. https://doi.org/10.1007/s1083 4- 021- 09798 - w
Financial Literacy Education Commission. (2016). National strategy for financial literacy 2016 update. The
Department of the Treasury.
Greninger, S. A., Hampton, V. L., Kitt, K. A., & Achacoso, J. A. (1996). Ratios and benchmarks for measuring the fi-
nancial well- being of families and individuals. Financial Services Review, 5(1), 57– 70. https://doi.org/10.1016/
S1057 - 0810(96)90027 - X
Harter, J. K., & Gurley, V. F. (2008). Measuring well- being in the United States. APS Observer, 21. https://www.
psych ologi calsc ience.org/obser ver/measu ring- well- being - in- the- unite d- states
Hermann, A., Herbert, C., & Molinsky, J. H. (2020). The association between high mortgage debt and financial
well- being in old age: Implications for the financial education field (p. 51). Joint Center for Housing Studies of
Harvard University.
Joo, S., & Grable, J. E. (2004). An exploratory framework of the determinants of financial satisfaction. Journal of
Family and Economic Issues, 25(1), 25– 50. https://doi.org/10.1023/B:JEEI.00000 16722.37994.9f
Kahneman, D., & Deaton, A. (2010). High income improves evaluation of life but not emotional well- being.
Proceedings of the National Academy of Sciences of the United States of America, 107(38), 16489– 16493.
https://doi.org/10.1073/pnas.10114 92107
Knoll, M. A. Z., & Houts, C. R. (2012). The financial knowledge scale: An application of item response theory to
the assessment of financial literacy. Journal of Consumer Affairs, 46(3), 381– 410.
Korobilis, D., & Shimizu, K. (2021). Bayesian approaches to shrinkage and sparse estimation (SSRN Scholarly
Paper No. 3971832). https://doi.org/10.2139/ssrn.3971832
Lee, J. M., Lee, J., & Kim, K. T. (2020). Consumer financial well- being: Knowledge is not enough. Journal of Family
and Economic Issues, 41(2), 218– 228. https://doi.org/10.1007/s1083 4- 019- 09649 - 9
Lusardi, A., & Mitchell, O. S. (2011). Financial literacy around the world: An overview. Journal of Pension
Economics & Finance, 10(4), 497– 508.
Makalic, E., & Schmidt, D. F. (2011, December). A simple Bayesian algorithm for feature ranking in high di-
mensional regression problems. In Australasian Joint Conference on Artificial Intelligence (pp. 223– 230).
Springer.
Makalic, E., & Schmidt, D. F. (2016). High- dimensional Bayesian regularised regression with the BayesReg package
(arXiv:1611.06649). arXiv. https://doi.org/10.48550/ arXiv.1611.06649
Michael Collins, J., & Urban, C. (2020). Measuring financial well- being over the lifecourse. The European Journal
of Finance, 26(4– 5), 341– 359. https://doi.org/10.1080/13518 47X.2019.1682631
Nagypal, E., & Tobacman, J. (2019). Credit characteristics, credit engagement tools, and financial well- being
(SSRN Scholarly Paper ID 3455574). Social Science Research Network. https://papers.ssrn.com/abstr
act=3455574
Park, T., & Casella, G. (2008). The Bayesian LASSO. Journal of the American Statistical Association, 103(482), 681–
686. https://doi.org/10.1198/01621 45080 00000337
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
14
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
Patel, P. C., & Wolfe, M. T. (2019). Money might not make you happy, but can happiness make you money? The
value of leveraging subjective well- being to enhance financial well- being in self- employment. Journal of
Business Venturing Insights, 12, e00134. https://doi.org/10.1016/j.jbvi.2019.e00134
Pedhazur, E. J. (1982). Introduction to linear structural relations (LISREL). In Multiple regression in behavioral
research. Harcourt Brace College Publishers.
Porter, N. M., & Garman, E. T. (1992). Money as part of a measure of financial well- being. American Behavioral
Scientist, 35(6), 820– 826. https://doi.org/10.1177/00027 64292 03500 6016
Rath, T., Harter, J. K., & Harter, J. (2010). Wellbeing: The five essential elements. Simon and Schuster.
Shim, S., Xiao, J. J., Barber, B. L., & Lyons, A. C. (2009). Pathways to life success: A conceptual model of finan-
cial well- being for young adults. Journal of Applied Developmental Psychology, 30(6), 708– 723. https://doi.
org/10.1016/j.appdev.2009.02.003
Tibshirani, R. (1996). Regression shrinkage and selection via the LASSO. Journal of the Royal Statistical Society.
Series B (Methodological), 58(1), 267– 288.
Vosloo, W., Fouche, J., & Barnard, J. (2014). The relationship between financial efficacy, satisfaction with remuner-
ation and personal financial well- being. International Business & Economics Research Journal (IBER), 13(6),
1455. 10.19030/ iber.v13i6.8934
How to cite this article: Lacombe, D. J., & Khatun, N. (2022). What are the
determinants of financial well- being? A Bayesian LASSO approach. American Journal of
Economics and Sociology, 00, 1–17. https://doi.org/10.1111/ajes.12489
APPENDIX A
VARIABLE SUMMARIES
Variable label/parameter Response value with response label
1. Difficulty in covering monthly expenses 1. Not at all difficult
2. Somewhat difficult
3. Very difficult
2. Follow- through set financial goals 1. Not at all
2. Very little
3. Somewhat
4. Very well
5. Completely
3. Confidence in achieving financial goals 1. Not at all confident
2. Not very conf ident
3. Somewhat confident
4. Very confident
4. Satisfied with own life 1. Strongly disagree
2. Disagree
3. Somewhat disagree
4. Neither agree nor disagree
5. Somewhat agree
6. Agree
7. Strongly agree
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
15
DETERMINANTS OF FINANCIAL WELL BEING
Variable label/parameter Response value with response label
5. Lots of stress in respondent's life 1. Strongly disagree
2. Disagree
3. Neither agree nor disagree
4. Agree
5. Strongly agree
6. Household income 1. Less than $20,0 00
2. $20,000 to $29,999
3. $30,000 to $39,999
4. $40,000 to $49,999
5. $50,000 to $59,999
6. $60,000 to $74,999
7. $75,000 to $99,999
8. $100,000 to $149,999
9. $150,000 or more
7. Savings habit 1. Strongly disagree
2. Disagree
3. Disagree slightly
4. Agree slightly
5. Agree
6. Strongly agree
8. Age categories 1. 18– 24
2. 25 – 34
3. 35– 44
4. 45– 54
5. 55– 61
6. 62– 69
7. 70 – 74
8. 75+
9. Consult own budget to perceive the balance 1. Strongly disagree
2. Disagree
3. Neither agree nor disagree
4. Agree
5. Strongly agree
10. Paid off monthly credit card balance 1. Not applicable or never
2. Seldom
3. Sometimes
4. Often
5. Always
11. Couldn't afford medical costs 1. Never
2. Sometimes
3. Often
12. Confidence in ability to raise $2000 in 30 days 1. I am certain I could not come up with $2000
2. I could probably not come up with $2000
3. I could probably come up with $2000
4. I am certain I could come up with the full
8. I don't know
APPENDIX A (Continued)
(Continues)
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
16
|
AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY
Variable label/parameter Response value with response label
13. Frugality- being economical with things 1. Strongly disagree
2. Disagree
3. Disagree slightly
4. Agree slightly
5. Agree
6. Strongly agree
14. Reti red 0. No
1. Yes
15. Stayed within own budget 1. Not applicable or never
2. Seldom
3. Sometimes
4. Often
5. Always
16. Ability to work toward long- term goals 1. Not at all
2. Not very well
3. Very well
4. Completely well
17. Financial products currently have: Non-
Retirement Investments
0. No
1. Yes
18. Belief that ability to manage money is
unchangeable
1. Strongly disagree
2. Disagree
3. Somewhat disagree
4. Neither agree nor disagree
5. Somewhat agree
6. Agree
7. Strongly agree
19. Frequency felt not respected or mistreated w/
financial services
1. Never
2. Rarely
3. Sometimes
4. Often
20. Consider the steps needed to stick to own budget 1. Strongly disagree
2. Disagree
3. Neither agree nor disagree
4. Agree
5. Strongly agree
21. Respondent did not select any item in SHOCKS
bank
0. No
1. Yes
22. Poverty status 1. <100% FPL
2. 100%– 199% FPL
3. 200%+ FPL
23. Checking/savings account at a bank/credit
union
0. No
1. Yes
24. Do research before making monetar y decisions 1. Never
2. Seldom
3. Sometimes
4. Often
5. Always
APPENDIX A (Continued)
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|
17
DETERMINANTS OF FINANCIAL WELL BEING
Variable label/parameter Response value with response label
25. Received a large sum of money beyond normal
income
0. No
1. Yes
26. More than 18- year- old kids being supported
financially by respondent
0
1
2+
APPENDIX A (Continued)
15367150, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ajes.12489 by Texas Tech University Libraries, Wiley Online Library on [12/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
