INTERNATIONAL BLACK SEA UNIVERSITY

FACULTY OF BUSINESS MANAGEMENT

DOCTORATE PROGRAM IN BUSINESS ADMINISTRATION PROGRAM

THE IMPACT OF FINANCIAL LITERACY ON THE FINANCIAL DECISIONS OF

CONSUMERS IN LOW-INCOME COUNTRIES (ARMENIA CASE)

Artur Pokrikyan

Doctoral Dissertation in Business Administration

Tbilisi, 2016
Scientific
Supervisor:
(full name)

(academic title)

____________________________________________
(supervisor’s signature)

I confirm that the work corresponds to the field, is characterized by novelty, scientific and
practical value and is presented by the format defined by International Black Sea University.

Experts (full name & academic title):

1.
signature

2.
signature

3.
(if any)

Opponents (full name & academic title):

1.
signature

2.
signature

3.
(if any)

I acknowledge that this is my own work, which is presented in the format defined by
International Black Sea University and is attached by the publications relevant to the
dissertation.

ii
CONTENTS

ACKNOWLEDGEMENTS ............................................................................................................v

ABSTRACT .................................................................................................................................. vi

LIST OF TABLES ....................................................................................................................... vii

LIST OF FIGURES ....................................................................................................................... ix

ABBREVIATIONS ....................................................................................................................... xi

INTRODUCTION ...........................................................................................................................1

CHAPTER 1. LITERATURE REVIEW .........................................................................................7

1.1 The concept of financial literacy ......................................................................................7

1.2 The importance of financial literacy ...............................................................................13

1.3 National Strategies for Financial Education ...................................................................18

1.4 The measurement approaches of financial literacy.........................................................23

1.5 Financial literacy, socio-demographics and financial decisions .....................................30

CHAPTER 2: METHODOLOGY .................................................................................................35

2.1 Primary research methodology and data .........................................................................35

2.1.1. Sample and Survey Questionnaire ...............................................................................36

2.1.2. Binary Logistic Regression ..........................................................................................41

2.1.3. Financial literacy workshop design ..............................................................................46

2.1.4. Pre-post design method ................................................................................................47

2.1.5. Quasi-Experimental design method .............................................................................48

2.2 Secondary research methodology and data .....................................................................55

2.2.1 Methodological Principles and Specificities .................................................................55

2.2.2 Data and analysis ...........................................................................................................57

2.3 Methodological limitations ..................................................................................................58

CHAPTER 3: RESULTS ..............................................................................................................61

3.1 Evaluation of the level of financial literacy in Armenia ......................................................61

3.2 Relationship between financial literacy and financial preferences ......................................64

iii
 3.3 Effectiveness of financial literacy intervention ...................................................................73

3.4 Relationship between financial literacy and financial decisions of consumers ...................91

3.5 Variations of results by socio-demographics .......................................................................99

CONCLUSION AND RECOMMENDATIONS ........................................................................106

REFERENCES ............................................................................................................................112

APPENDIX 1: RESEARCH QUESTIONAIRES .......................................................................120

APPENDIX 2: TREATMENT AND CONTROL GROUP CHARACTERISTICS ..................128

APPENDIX 3: WORKSHOP DESIGN ......................................................................................130

iv
ACKNOWLEDGEMENTS

This research is elaborated by Artur Pokrikyan, a PhD student at International Black Sea
University, the Faculty of Business Administration, Finance program. In conducting these
research, the author wants express deep gratitude towards his scientific supervisor Prof. Dr.
Ketevan Lapachi for huge support and guidance during the entire period.

For the invaluable backing during the whole PhD studies and research the author would
like to thank also Goga Buchashvili and Tinatin Kublashvili. Special thanks to Metin Mercan,
Tatiana Papiashvili, Salavat Sayfullin, Tea Kbiltsetskhlashvili, Sophie Khundadze for the
valuable insights, guidance and comments during the discussions.

Superior appreciations to the Central Bank of Armenia for providing the data and support
in conducting the research, particularly Armenak Darbinyan (Board Member of CBA) and
Armenuhi Mkrtchyan (Head of Consumer Rights Protection and Financial Education Center).

v
ABSTRACT

The importance of financial literacy has especially been underlined after the Global
Financial Crisis in 2008. Nowadays many international organizations and policy setting bodies
such as OECD, the World Bank and G20 as well as countries themselves implement strategic
actions to increase the level of financial literacy. The latter is an important prerequisite also for
financial stability, financial inclusion and consumers’ rights protection. However, to understand
the effectiveness of these actions the impact of financial literacy on financial decisions of
consumers need to be evaluated. Particular, it is important to understand those relationships in
low-income countries, as there is little evidence on such academic research in such countries,
especially the ones from post-Soviet space.

In frame of this dissertation, the relationship of financial literacy, financial preferences

and financial decisions of consumers regarding saving and debt management is analyzed and
evaluated by using binary logistic regression and quasi-experimental research on primary data
and descriptive analysis of secondary data.

The results, outline that financial literacy significantly affects the financial decisions of
consumers regarding saving. The impact of financial literacy on debt management decisions of
consumers is not significant. Financial literacy intervention in form of workshop proved to be
efficient in the short-run, while its long-term effectiveness is controversial. The relationship
between financial literacy and financial decisions is different across socio-demographics. Thus,
policymakers and all stakeholders should consider the research findings to implement more
targeted, differentiated, efficient policies towards respective target groups.

vi
LIST OF TABLES

Table 1.1 The conceptual definitions of financial literacy given by various authors …...… 7

Table 1.2 The nature of financial education mandates in several countries ……………… 20

Table 1.3 The status of national strategies in different countries in 2015 ……………...… 22

Table 1.4 Surveys used by different countries to measure the level of financial literacy ... 24

Table 1.5 Comparison of international approaches of financial literacy measurement …... 28

Table 2.1 The number of parameters of binary logistic regression based on sample ….…. 38

Table 2.2 Independent Two-Sample T-Test for Treatment and Control Groups ………… 39

Table 2.3 Summary Statistics and Pearson chi-square test for financial preferences related to
saving, taking debt and seeking for information …………………………….… 43

Table 2.4 Spearman’s rank correlation between the variables ………………………….... 44

Table 2.5 Test on statistically significant difference of pretest across treatment and control
groups ……………………………………………………………………….….. 53

Table 2.6 Homogeneity of regression in case of saving and debt management behavior ... 54

Table 2.7 The scoring model of Financial Capability Barometer methodology ………...... 57

Table 3.1 Binary Logistic regression output for saving preferences model …………….... 65

Table 3.2 Specification link test for savings preferences logit model ……………….…… 66

Table 3.3 Odds ratio for saving preferences model …………………………………….… 67

Table 3.4 Binary Logistic regression output for debt preferences model ………………… 68

Table 3.5 Specification link test for debt preferences logit model ……………………….. 69

Table 3.6 Odds ratio for debt preferences model ………………………………………..... 70

Table 3.7 Specification link test for information preferences probit model …………….... 70

Table 3.8 Binary Probit regression output for information preferences model …………... 71

Table 3.9 Number of participants to financial literacy workshops in treatment villages … 74

Table 3.10 Shapiro-Wilk test of normality ……………………………………………….... 78

Table 3.11 The paired t-test results for Financial Literacy Score ………………………….. 79

Table 3.12 The paired t-test results for Saving Preferences ……………………………..… 84
vii
Table 3.13 The paired t-test results for Debt Preferences ………………………………….. 85

Table 3.14 The output of the Wilcoxon signed rank sum test for information preferences .. 87

Table 3.15 The paired t-test results for the level of Trust ………………………………..… 88

Table 3.16 ANCOVA output for the SBS in case of Treatment and Control Groups …....... 91

Table 3.17 ANCOVA output for the DMS in case of Treatment and Control Groups ...….. 95

viii
LIST OF FIGURES

Figure 1.1 The conceptual framework of financial literacy (capability) ………………...… 12

Figure 2.1 The framework of primary research …………………………………………… 36

Figure 2.2 Non-experimental Pre-Post Design Framework …………………….…………. 48

Figure 2.3 Nonequivalent Groups Design Framework ……………………………….…… 49

Figure 2.4 Descriptors of people with high SBS and DMS values …..……..….………….. 51

Figure 2.5 Linearity between dependent variable and covariate across all groups (Treatment
– green, Control – blue) …………………………………………….………….. 52

Figure 3.1 The share of financial services and tools people use in Armenia ……………… 62

Figure 3.2 The use of financial services in treatment group ………………………………. 77

Figure 3.3 The impact of financial literacy workshop on the concepts related to inflation .. 80

Figure 3.4 The impact of financial literacy workshop on interest rate related concepts …... 80

Figure 3.5 The impact of financial literacy workshop on the concepts of diversification and
understanding of deposit guarantee scheme ………………………………...…. 81

Figure 3.6 The impact of financial literacy workshop on understanding APRC ………….. 82

Figure 3.7 The impact of financial literacy workshop on understanding financial frauds and
scams ………………………………………………………………………….... 82

Figure 3.8 The impact of financial literacy workshop on capabilities regarding personal
rights protection ……………………………………………………………..…. 83

Figure 3.9 The shifts in Saving Preferences based on different ways of saving …………... 84

Figure 3.10 The shifts in Debt Preferences based on different ways of taking debt ………... 86

Figure 3.11 The shifts in Information Preferences based on sources of information ………. 88

Figure 3.12 The changes in the level of Trust towards different financial institutions …...… 89

Figure 3.13 The impact of financial literacy intervention on the share of savings in Treatment
and Control Groups after 6 months ………………………………………….…. 94

Figure 3.14 The impact of financial literacy intervention on the share of deposits in Treatment
and Control Groups after 6 months …………………………………………..… 95

ix
Figure 3.15 The impact of financial literacy intervention on the share of credit in Treatment
and Control Groups after 6 months …………………………………………….. 97

Figure 3.16 The impact of financial literacy intervention on the share of debt in Treatment and
Control Groups after 6 months ………………………………………………… 98

Figure 3.17 The impact of financial literacy intervention on the share of cosigners in
Treatment and Control Groups after 6 months ……………………………….... 99

Figure 3.18 The differences between individuals’ financial literacy, trust, saving and debt
management behavior by gender ……………………..…………………...….. 100

Figure 3.19 The change in saving and debt management behavior after financial literacy
intervention by gender …………………………………………..……………. 101

Figure 3.20 The differences between individuals’ financial literacy, trust, saving and debt
management behavior by age groups …………………………………....……. 101

Figure 3.21 The change in saving and debt management behavior after financial literacy
intervention by age groups …………………………………………………..... 102

Figure 3.22 The differences between individuals’ financial literacy, trust, saving and debt
management behavior by family income levels …………………………….... 103

Figure 3.23 The change in saving and debt management behavior after financial literacy
intervention by family income levels …………………………………….…… 104

x
ABBREVIATIONS

ANCOVA – Analysis of Covariance

APRC – Annual percentage rate of charge

ASIC – Australian Securities and Investments Commission

CBA – Central Bank of Armenia

CMVM – Portuguese Securities Market Commission (Comissão do Mercado de Valores

Mobiliários)

DMS – Debt Management Score (Index)

FCAC – Financial Consumer Agency of Canada

FCB – Financial Capability Barometer

FCMC – Financial and Capital Market Commission of Latvia

FLS – Financial Literacy score

FSA – Financial Services Authority

G20 – Group of Twenty Major Economies

GDP – Gross Domestic Product

GLM – Generalized Linear Model

HMT – Her Majesty's Treasury of UK

INFE – International Network on Financial Education

NCEE – National Council for Economic Education

NSFE – National Strategy for Financial Education

OECD – Organization for Economic Co-operation and Development

PACFL – President's Advisory Council on Financial Literacy

PISA – Programme for International Student Assessment

RTF – Russia Trust Fund

SBS – Saving Behavior Score (Index)

WB – World Bank

xi
INTRODUCTION

The developments in the area of finance and technology during recent years made
financial markets, financial products and services increasingly accessible for people having
different socioeconomic background (Lusardi & Mitchell, 2013). Global Financial Crisis in
2008, which is rooted in the housing price bubbles, raised many serious financial difficulties for
the consumers who previously benefited from the simple crediting procedures of the banks
without considering their financial capability. Although irresponsible financial decision-making
is not the only reason for these difficulties, it is argued that consumers will have less financial
problems if they are financial literate. For understanding financial literacy level among
population, it is necessary to assess consumers’ capacity to make sound financial decisions,
however the academy has not given much attention to the ways of measuring financial literacy
(Huston, 2010), meanwhile the interest towards financial literacy measurement by different
countries and international organizations have been increased gradually during last years. The
latter is especially important for analyzing the behavioral patterns of the consumers related to
their decisions, while managing their personal finances.

Financial literacy has received increased attention since the Global Financial Crisis in
2008 underlined the importance of basic financial knowledge and skills for consumers. The
value of financial literacy, as one of the main pillars of financial inclusion, has been increasingly
emphasized during the last decade. The knowledge, skills, attitude and behavior incorporated in
financial literacy are important for people to make responsible and informed decisions, manage
potential risks and improve their financial well-being. Standard setting bodies, international
forums and organizations, such as World Bank, G20, Global Partnership for Financial Inclusion
and others has also recognized the significance of financial literacy for long-term financial
stability and consumers well-being (OECD, 2013a). Thus, improving financial literacy has
emerged as a strategic policy objective that complements governments’ financial inclusion and
consumer protection agendas of increasing number of countries.

The financial burden for individuals with limited opportunities can be especially
devastating resulting issues related to poverty trap. Given the current financial developments, it
is easier for people living in low-income countries to become victims of financial fraud and
scams because of low level of adequate financial literacy (Lyons & Scherpf, 2004). In addition,
the diversification of ways to take debt in form of credit, overdraft, in place loans at shops etc.
from different sources such as banks, credit organizations, pawn shops, informal sources require
better debt management knowledge, skills and abilities. Considering this context one of the one
1
of the main reasons for unmanageable debt for consumers become the number of sources of
credit (Widdowson & Hailwood, 2007).

Problem Statement

Given the development trends globally around financial literacy, financial education and
financial inclusion, exploring the behavioral patterns of consumers in relation to financial
literacy is especially important in developing countries considering different social and economic
issues, such as poverty, unemployment, low personal income, financial exclusion, not well
developed financial institutions, poor developed stock markets, dependence from developed
countries etc. There is little evidence on the relationship of financial behaviors and literacy in
general (Consumer Financial Protection Bureau, 2015), however research lacks to answer this
topic related questions in developing countries especially from post-Soviet space even more.

Thus, for effective policy making, especially in the areas of financial stability, financial
inclusion and consumers rights protection, it is needed to explore the interdependence of
financial literacy and consumers’ decisions and create simple heuristics to explain those
connections. It is also valuable to analyze how demographic factors such as gender, age,
education, income affect those connections.

Goal Statement and Research Questions

This research aims to explore how financial literacy influences consumers’ financial
decisions related to savings and debt management in developing countries especially in post-
Soviet space by the example of Armenia. The research also aims to analyze the short-term and
long-term effects of financial literacy intervention on the behavior of individuals related to
savings and debt management by highlighting the variations among different socio-demographic
groups. In frame of this study, 6 months is considered to be the minimum effective period of
time to evaluate the long-term effects. Given the mentioned challenges and goals of the research,
the following research questions for this study are formulated:

1. To what extent increase in financial literacy of consumers living in Armenia impact

their financial decisions related to saving and debt management?
2. To what extent financial literacy is related to financial preferences of consumers?
3. To what extent financial literacy intervention in form of workshop is an efficient
method to impact financial literacy, preferences and behavior in short-term and long-
term (after 6 months)?

2
 4. To what extent socio-demographic differences affect the relationship between
financial literacy and financial decisions of consumers?

In line with the main research, additional research questions are raised and studied from
the perspective of financial literacy and financial decision making of consumers as well as from
the perspective of policymakers in the area of financial stability, financial inclusion and
consumers rights protection. Some of these additional insights include reflections related to the
ways of financial literacy measurement and development of financial education activities.

Novelty and Actuality

The novelty of the research lies in the fact, that the research concentrates on the issues in
low-income countries, there the consumers’ decisions are usually affected by many objective and
subjective factors typically relevant for emerging markets. Especially important factor to
consider is little evidence on similar research initiatives in developing countries, especially
including post-Soviet space. Though each low-income country has its own characteristics and
development patterns, there are some major economic similarities among them, which
objectively and/or subjectively impact financial situation of households. Thus, another important
aspect related to the novelty of the research is the analysis and evaluation how financial
decisions of people regarding saving and debt management change depending on different socio-
demographic factors such as gender, age, education, income etc. It is valuable also to examine
the preferences of consumers toward formal and informal sources of saving, taking debt and
seeking form information about financial services in relation to financial literacy.

The actuality of research is highly confirmed by international organizations such as

OECD/INFE, the World, Bank, G20 and other policy setting bodies. As there is little evidence of
such research initiatives from little-income and post-Soviet countries it is important to
understand whether there is any link between financial literacy and financial decisions of
consumers and if yes, how this relationship might be described. Considering vast amount of
resources spent on financial literacy initiatives and policies this research will help policymakers
to understand the possible outcomes of these efforts in terms of impact on financial behavior of
consumers. In addition, many countries are in the process of developing nationwide strategies
and strategic actions to strengthen consumers in terms of financial literacy and achieve healthy
financial behaviors among the population. Thus, this research results will help them to advance
their policies and strategic actions.

3
Significance of the problem

Financial illiteracy can hinder to the welfare and prosperity. Financial difficulties of
individuals and families can have a negative impact on both micro and macro levels of the
economy in a country. In line with the global technological and financial developments there is a
shift of responsibilities for financial well-being and decision making from the Government and
private sector to households and individuals. Thus, the value of financial literacy and financial
education become more vivid (OECD/INFE, 2013). Especially after the Global Financial Crisis
in 2008 financial literacy has been widely recognized as a key skill for the life as well as
locomotive for economic development and financial sustainability (OECD/INFE, 2009). Given
the current economic developments, when the debt burden on households has a tendency to
increase, while wealth and income decreases, the issue of financial literacy and effective
financial decision-making become significantly bigger.

Practical and Theoretical Value

The most significant theoretical value of this research will incorporate findings related to
the consumers’ behavioral patterns based on the level of financial literacy. The latter is important
to understand not only from the perspective of the Economics discipline, but also from the
perspective of Psychology. The research findings might bring new theoretical dilemmas related
to the behavioral patterns of people with limited opportunities and contribute to the development
theories.

The practical value of the work will be the actual use of findings by the policymakers.
This is extremely important for developing effective financial literacy interventions.
Understanding behavioral patterns of consumers will help policymakers on national and
international level to implement more informed macroeconomic policies and build development
interventions. The variations by social-demographic factors will also ensure proper targeting of
the policy interventions.

Research Methods

To answer the research questions aimed in this work mainly quantitative research
methods are applied, particularly the research comprises of primary and secondary data analysis.
The primary research aims to identify the relationship between financial literacy and financial
decisions of consumers. It also highlights the variations across different target groups. The
secondary research supports the primary research findings by providing descriptive analysis and

4
additional evidence on findings. The primary research focuses on three main dimensions
presented below:

1. The relationship between financial literacy and financial preferences,

2. The short-term impact of financial literacy intervention on financial knowledge
and financial preferences of consumers.
3. The long-term impact of financial literacy on financial decisions of consumers
related to saving and debt management, understanding variations by socio-
demographics.

The primary data is collected using a survey, which is done in 24 villages from 6 regions
of Armenia (Shirak, Gegharkunik, Lori, Tavush, Ararat, Kotayq). The villages are included in
frame of financial literacy project implemented by CBA, thus the choice of villages is not
random. In total 504 respondents participated in the survey by filling the special questionnaire
designed in frame of this research. Based on these data the binary logistic regression models are
built to describe the relationship between the level of financial literacy and formal or informal
financial preferences of people regarding saving, debt management and shopping around while
choosing financial services (binary variables).

By the second step a quasi-experimental research study is designed to address the next
two objectives of primary research and bring out the answers for research questions in frame of
this study. For this purpose, the 24 villages are randomly assigned to the Treatment and Control
groups. For the Treatment villages a special financial literacy workshop is developed and
implemented, while in control villages no financial literacy intervention is done. To evaluate the
short-term impact of financial literacy intervention on financial knowledge and financial
preferences of consumers, non-experimental pre-post design are used in treatment villages. This
method suggests doing pretest before and posttest after the financial literacy workshop and based
on the data analysis evaluate the effectiveness of financial literacy intervention in terms of
change in the level of financial literacy and financial preferences among participants. Given the
fact that the same people filled in both pretest and posttest, for the data analysis and evaluation
the paired t-test method is mainly used, which enables you to compare two groups when the
observations are related.

In order to understand the impact of financial literacy on financial decisions of consumer

in long-term, after 6 months of workshops a phone survey are done among the respondents from
both Treatment and Control groups. Considering the time limitations of the study, 6 months’

5
period is assumed to be the minimum effective period to evaluate a change in consumer’s
decisions. Based on the data from the initial survey and phone survey the impact of financial
literacy on financial decisions of consumers regarding saving and debt management are
evaluated. Considering the fact, that the villages are not assigned randomly for the analysis and
evaluations Nonequivalent Groups Design approach are used and the differences of treatment
and control groups are examined by using Analysis of Covariance (ANCOVA) statistical
method. In addition, based on this data the socio-demographic variations are analyzed by using
descriptive statistics.

Secondary research is done based on the data from financial capability assessment study
done by CBA in 2014. The data are used to analyze, evaluate and demonstrate the level of
financial literacy in Armenia by using descriptive statistics.

Structure of the Dissertation

Dissertation consists of 131 pages, includes 3 chapters, 18 sub chapters and conclusions
and recommendations. It contains 29 tables, 29 figures and 3 appendixes. Author has published
three scientific articles devoted to the various aspects of study. All of them are relevant to the
dissertation․

The structure of the dissertation include literature review on the important aspects of
financial literacy and financial decisions of consumers. Particularly, the concept of financial
literacy, its importance and measurement is discussed, in addition the experience of different
countries in terms of policies related to financial literacy and financial education is presented. In
the last part of literature review various studies are discussed, which focus on the relationship of
financial literacy and financial behaviors of consumers mainly related to saving and debt
management, as well as socio-demographics. In the second chapter of the dissertation the
methodology of the research is described, particularly including various aspects of primary and
secondary research. The methodological limitations are also addressed in this chapter. The last
chapter includes the results of primary and secondary analysis and evaluation of data to answer
research questions. At the end the conclusions and recommendations are derived and research
questions are answered. The dissertation also incorporates the lists of tables, figures and
abbreviations as well as appendixes presented separately.

6
CHAPTER 1. LITERATURE REVIEW

1.1 The concept of financial literacy

In order to ensure comparability as well as steadiness across this work, it is important to

clearly define the core concepts from the beginning. The concept of financial literacy is
relatively new in the academy and its definition varies from study to study. According to Lee
(2012) the earliest definition of financial literacy is given by Noctor et al. in 1992 in the report
for the National Foundation for Educational Research commissioned by NatWest Bank. Another
of the early pioneers of financial literacy Jump$tart Coalition for Personal Financial Literacy
defines it as “the ability to use knowledge and skills to manage one’s financial resources
effectively for a lifetime of financial security” (2015, p. 1). Similar definition is used by the
President’s Advisory Council on Financial Literacy of USA (PACFL), which offers a consensus
approach, however it is unclear, if it is widely accepted (Hung, Parker, & Yoong, 2009). In their
study on household financial management Hilgert, Hogarth, & Beverly (2003) associate financial
literacy mainly with financial knowledge, thus explaining the connection between financial
literacy and financial behavior.

Table 1.1 represents some of the conceptual definitions of financial literacy used by
various organizations and researchers during last two decades. The table presented by Hung et al.
(2009, p. 6) are modified and includes conceptual definitions of financial literacy, which are
taken from different studies and placed in chronological order. The most common idea across the
given definitions is financial knowledge and understanding and ability to use it. Some of them
are emphasizing on the aspect of financial literacy as an important prerequisite for making
decisions.

Table 1.1 The conceptual definitions of financial literacy given by various authors

Authors Definition

“The ability to make informed judgements and to take

Noctor, Stoney and Stradling, 1992 effective decisions regarding the use and management of
money”
“An individual’s ability to obtain, understand and evaluate
Mason and Wilson, 2000 the relevant information necessary to make decisions with
an awareness of the likely financial consequences”

7
 “The ability to read, analyze, manage and communicate
about the personal financial conditions that affect material
well-being. It includes the ability to discern financial
Siegenthaler, Anderson, Lyter,
choices, discuss money and financial issues without (or
Kent, and Ward, 2000
despite) discomfort, plan for the future, and respond
competently to life events that affect everyday financial
decisions, including events in the general economy”
Hilgert, Hogarth, and Beverly,
Financial knowledge
2003
“Individuals are considered financially literate if they are
competent and can demonstrate they have used knowledge
Moore, 2003
they have learned. Literacy is obtained through practical
experience and active integration of knowledge”
“Familiarity with basic economic principles, knowledge
The National Council on Economic
about the U.S. economy, and understanding of some key
Education (NCEE), 2005
economic terms”
“The ability to evaluate the new and complex financial
instruments and make informed judgments in both choices
Mandell, 2008
of instruments and extent of use that would be in their own
best long-run interests”
“The ability to use knowledge and skills to manage
President’s Advisory Council on
financial resources effectively for a lifetime of financial
Financial Literacy (PACFL), 2008
well-being”
“The ability to make simple decisions regarding debt
contracts, in particular how one applies basic knowledge
Lusardi and Tufano, 2009
about interest compounding, measured in the context of
everyday financial choices”
“Financial literacy is a measure of the degree to which one
understands key financial concepts and possesses the
ability and confidence to manage personal finances
Remund, 2010
through appropriate, short-term decision-making and
sound, long-range financial planning, while mindful of life
events and changing economic conditions”

8
 “Knowledge of basic financial concepts, such as the
working of interest compounding, the difference between
Lusardi, 2008a, 2008b
nominal and real values, and the basics of risk
diversification”
“Financial literacy is a combination of awareness,
knowledge, skill, attitude and behavior necessary to make
Atkinson and Messy, 2012
sound financial decisions and ultimately achieve individual
financial well-being”
“Financial literacy is knowledge of financial concepts and
risks, and the skills, motivation and confidence to apply
such knowledge and understanding in order to make
OECD, 2013b
effective decisions across a range of financial contexts, to
improve the financial well-being of individuals and
society, and to enable participation in economic life”

Even though, it is difficult to find out which financial literacy definition is widely
accepted from the wide range of definitions discussed above, it can be noticed the transformation
of the financial literacy concept throughout these years. The presented definitions demonstrate
that prior to the Global Financial Crisis in 2008 the concept of financial literacy is based on such
concepts as knowledge, awareness, familiarity and people being informed. At the same time,
professionals did not concentrate more on the practical aspects and actual application of financial
literacy. However, the definition provided by PACFL is notable by the fact that it views financial
literacy not simply as reserve of knowledge and information, but as an instrument that allows
people to achieve financial well-being. Nowadays, financial literacy has become much wider
concept including skills dedicated to actual usage of that knowledge, as well as proper attitude
and motivation to be financially included and appropriate financial behavior.

Developments and policies in financial literacy during these years led researchers, policy
makers and different organizations to start using other financial literacy related concepts, such as
“financial education”, “financially capable person” and “financial capability”. Due to the
necessity of having clear distinction between the mentioned concepts, from the academic point
of view it will be helpful to discuss these concepts as well.

OECD has been actively involved in global policy aimed at increasing the level of
financial literacy of population. In its publication related to the issues and policies in the sphere

9
of financial literacy, the concept of financial education has been defined as follows: “the process
by which financial consumers/investors improve their understanding of financial products and
concepts and, through information, instruction and/or objective advice, develop the skills and
confidence to become more aware of financial risks and opportunities, to make informed
choices, to know where to go for help, and to take other effective actions to improve their
financial well-being” (OECD, 2005, p. 26).

Many researchers (Hung, Parker, & Yoong, 2009; Lusardi & Mitchelli, 2007; O’Connell,
2009; Orton, 2007; Pahnke & Honekamp, 2010) also used this definition as a basis for their
reviews on financial literacy. Moreover, it served as a basis for the definition of financial literacy
given by Atkinson & Messy (2012). However, the latter is not used within the framework of
Program for International Student Assessment (PISA) also conducted by OECD and the
subsequent report. Both definitions are presented in the Table 1.1.

In mid-2000s, UK has suggested the concept of financial capability, which is broader

than the earlier traditional concept of financial literacy and includes not only financial
knowledge and skills, but also attitude and behavior. Particularly, FSA of UK (Kempson,
Collard, & Moore, 2005) defined financially capable people as ones who “are able to make
informed financial decisions. They are numerate and can budget and manage money effectively.
They understand how to manage credit and debt. They are able to assess needs for insurance and
protection. They can assess the different risks and returns involved in different saving and
investment options. They have an understanding of the wider ethical, social, political and
environmental dimensions of finances” (p. 13).

This definition focuses not only on the aspect of knowledge and understanding, but also
considers other important dimensions as well. The latter better explains another definition given
by the HMT of UK (2007), which states that “financial capability is a broad concept,
encompassing people’s knowledge and skills to understand their own financial circumstances,
along with the motivation to take action. Financially capable consumers plan ahead, find and use
information, know when to seek advice and can understand and act on this advice, leading to
greater participation in the financial services market” (p. 19).

The World Bank (WB) is another important institution providing support to global
policies on strengthening financial capability of population. Therefore, it is actively involved in
promoting responsible way of personal finance management.

10
 One of the major projects dedicated to the objective mentioned above is financial
capability measurement project initiated in 2013. In this study and further publications, WB
manifests rather streamlined approach in terms of financial capability definition. Here, financial
literacy is mostly associated with financial knowledge, and it is specified that financial capability
has a broader meaning. Particularly, WB (2013, p. 1) in its study refers to financial capability as
“internal capacity to act in person’s best financial interest, given socio-economic environmental
conditions. It therefore encompasses the knowledge attitudes, skills, and behaviors of consumers
with regard to managing their resources and understanding, selecting, and making use of
financial services that fit their needs”.

In various studies aimed at the measuring financial capability or financial literacy, the
authors always indicate what they understand under the terms “financial literacy” or “financial
capability” in the context of their study. Orton and others (Orton, 2007) also discuss that the
terms financial capability and financial literacy might usually be used conversely. Financial
capability is the term commonly used in the United Kingdom and Canada, then financial literacy
is usually used in Australia and the United States. As a result, the following two conclusions can
be made based on the definitions of these concepts:

 even though several researchers and organizations such as OECD continue to use the
formulation of financial literacy, nowadays there is a consensus among researchers,
international organizations and policy makers that financial literacy is much broader
concept than before and is based on behavioral concept.
 various conceptual definitions of financial literacy (or financial capability) put the
emphasis on the content rather than the way definition is formulated (difference in
wording does not imply difference of concept), which clearly denotes the essence of
financial literacy (or financial capability).

Following the same principals, many policymakers of various countries developed

National Strategies dedicated to improve the level of financial literacy among population.
National Strategy for Financial Education (NSFE) including awareness-raising activities of
policymakers is discussed in one of the upcoming parts of this study. However, it is important to
mention that policymakers define financial literacy in frame of NSFE. In the study, it is aimed to
explore financial literacy related issues in Armenia. Particularly, the definition given by CBA in
the NSFE in Armenia is used. The latter reads that “financially literate (capable) person has
knowledge, skills and culture, which enable him/her to be informed on financial matters, make

11
reasonable decisions about personal finances and take act respectively depending on a situation”
(Central Bank of Armenia, 2014, p. 3).

In their paper Hung, Parker, & Yoong (2009) provide a conceptual framework for
financial literacy, which is mainly based on the definition given by PACFL (2008), as well as
other theoretical and operational definitions by various researchers. However, based on the
developments in the area this conceptual model should be revised considering the broader
meaning of financial literacy/capability. In this work an adjusted conceptual model of financial
literacy/capability will be used which is based on the framework provided by Hung, Parker, &
Yoong (2009) and recent approaches taken by international organizations, policy-makers and
researchers.

Figure 1.1 The conceptual framework of financial literacy (capability)

Financial
knowledge

Attitude Skills

Financial
behavior

Given the discussions above on the concept of financial literacy (Figure 1.1) can be
drawn, which represents the logical relationship among the interrelated components of financial
literacy. Financial knowledge is the base and early form of financial literacy, as it might be
noticed from Table 1.1. Financial skills show the ability of person to use the financial knowledge
while making decisions. There as financial attitude is boarder component and includes feelings,

12
motivation and experience of person with financial system. All of these three components are
related to each other and are affecting financial behavior of consumers as well.

For further discussions in frame of dissertation, “financial literacy” and “financial

capability” concepts will be used interchangeably, considering the points and conceptual
developments discussed above.

1.2 The importance of financial literacy

While discussing the definitions for financial literacy, it is already noticeable one of the
major reasons why financial literacy is important, because it impacts financial well-being of
consumers and households in general. The link between financial literacy, financial decisions
and financial well-being is reflected in the definition of financial well-being given by the
Consumer Financial Protection Bureau, which states that it is “a state of being wherein a person
can fully meet current and ongoing financial obligations, can feel secure in their financial future,
and is able to make choices that allow enjoyment of life” (2015, p. 18).

Recent Global Financial Crisis in 2008 has also proved how important is the
circumstance that individuals and families have the needed information, knowledge and skills for
undertaking justified financial decisions. Financial illiteracy can hinder to the welfare and
prosperity. Financial difficulties of individuals and families can have a negative impact on both
micro and macro levels of the economy in a country. In line with the global technological and
financial developments there is a shift of responsibilities for financial well-being and decision
making from the Government and private sector to households and individuals. Thus, the
importance of financial literacy and financial education become more vivid (OECD/INFE,
2013). Especially after the Global Financial Crisis in 2008 financial literacy has been globally
recognized as a key life skill and important engine for economic development and financial
sustainability (OECD/INFE, 2009).

The role of financial literacy cannot be underestimated. It is substantial for both

customers of financial institutions and financial system as well as the economy of a country. It
has an impact on individuals’ decisions about sparing, acquiring, contributing and dealing with
monetary undertakings, affecting people’s enrichment and having noteworthy ramification
every-day life. Financial literacy also has a noteworthy impact on decisions of financial
institutions (banks, credit organizations, insurance companies etc.) in managing financial risk
and developing new products. Moreover, besides the fact financial literacy has a significant

13
impact on people’s financial choices, including risk/return tradeoffs, it also affects economy
making the allocation of resources more reasonable. Thus, financial literacy is important in terms
of economic growth and development as well as economic stability of a country as also
discussed by Widdowson and Hailwood (2007).

People face many financial decisions in every-day life, comprehending an extensive

variety of financial matters, including decisions on budgeting, spending as well as on saving and
investments. People also should manage their financial risks and decide, how much obligation
they may need to support their consumption, and what frame that obligation ought to take.
These decisions vary in complicatedness; however, all require no less than a fundamental level
of financial literacy.

From the beginning of the 20th century the growing relevance of financial literacy as a
fundamental ability has been identified by various professionals (Morris, 2001). Due to fact that
new technologies have altered the conception of distribution of the financial services around the
globe, there is a constantly expanding variety of financial products and services (obligations,
venture openings) accessible to the general population. While this gives expanded advantages, it
likewise involves more intricate risks, including ones that are not generally promptly obvious to
the unwary. Henceforth, the range and complexity of the financial decisions seriously affects the
people’s every-day life, as currently people need to be aware of the positive and negative aspects
of an extensive variety of financial products and services.

In order to have a net gain in using financial services and increase standard of living, the
ability to make well-informed financial decisions are must. The results of well-made financial
decisions guarantee individual’s financial security as well. An individual with a decent level of
financial literacy is likely to be in better position than somebody without those aptitudes and
information to deal with their financial choices judiciously. These include issues on budgeting,
making investment plans, managing indebtedness level, etc. Particularly, absence of
comprehension of financial matters can cause various negative results, such as unnecessary
financial expenses and high level of indebtedness.

The relevance of financial literacy in terms of the household ability to manage family
budget, savings and debts in low-income countries is even higher. Particularly in Armenia during
last decade the developments increased the access of population to much broader range of
financial services. Especially the taking debt has become easier. As a result, after the Global
Financial Crisis in 2008 the volume of consumer loans in Armenia increased almost three times

14
reaching to 653.5 mln dollars. The financial stability report for the first half of 2016, states that
the lending to households remains the most active area for financial institutions. Given the
weaker creditworthiness of households, as well as the fast growth of loans in the previous years,
the credit risk for the banks remain high, which is not affecting the solvency of financial
institutions, but affect their profitability (Central Bank of Armenia, 2016). The share of non-
performing loans increased as well during last years from 2.9% in the beginning of 2008 to
10.1% in the mid-2016. This is mainly connected to the difficulties of households pay back the
consumer loans taken in foreign currencies.

In addition, the diversification of debt acquiring ways (credit, overdraft, in place loans at
shops etc.) and sources (banks, credit organizations, pawn shops, informal sources) require better
debt management knowledge, skills and abilities. Widdowson and Hailwood (2007) in their
research related to the role of financial literacy in promoting a sound financial system mention
that one of the main reasons for unmanageable debt is the number of sources of credit.

At the same time as issues with debts appear, there is a decline in net current private
transfers (12.9%) and outbound workers' compensation (22.6%), conditioned by the impact of
global economic developments such as economic decline in Russia and Europe. In contrast, the
household deposits increased by 4.2% and the nominal wage also increased reaching to $360
(Central Bank of Armenia, 2016). Given the described situation the increase in financial literacy
is assumed to impact how households make their financial decisions related to managing their
debts, savings and budget.

Financial literacy is positively affecting consumers’ financial welfare. A study by

Bernheim, Garrett, and Maki (2001) showed that financially literate consumers who received
mandatory financial education are more likely to have higher savings rates and welfare. This
relation held due to the fact that financially literate consumers will be better at managing their
debts as well as will know the economic benefit of saving. This will help them stay informed and
easily identify financial deals which might be risky and dangerous. Financially literate
consumers are expected to protect their rights in case of an issue with any financial institution.

Increase in financial literacy is also an important precondition for financial inclusion and
financial sustainability. The evidence from the OECD/INFE Financial Literacy Measurement
Survey provides a unique opportunity to study the financial inclusion issues in various countries,
describe the extent to which individuals are financially excluded and provide possibility
analyzing whether financial inclusion is related to financial literacy. Many countries support

15
policy initiatives to close the gap between financially included and excluded people, by fostering
the provision of financial services targeted to people, who are financially excluded. In this regard
the policy-makers also value the role of financial literacy as an important driving force
(OECD/INFE, 2013).

The stability and effectiveness of financial system greatly depends on the financial
literacy of the public, as financially educated people could be expected to have beneficial effects
on the soundness and productivity of the monetary framework. This can take place in various
ways:

 Credit risk management of financial institutions can be facilitated by the extent

financial literacy and more effective management of household indebtedness level,
 Financially literate population will have a decent choice of savings and investments,
which in turn is likely to make financial institutions react creatively to increasing
demand prompting to an all the more progressively efficient financial system
 Financial literacy of the population could be relied upon as it will make them study
the behavior of financial institutions thoroughly, by practicing more prominent
analysis tools as well as by applying more advanced researches over the risk level of
specific organizations and their services. Thusly, more grounded market orders are
probably going to energize more judicious administration of financial risks within
organizations and a higher standard of service conveyance.
 Efficient investment decisions made by the individuals with high level of financial
literacy will result in a more efficient resource allocation within the country as well as
in reducing of risk hazard in the whole financial market. Thus, this ought to increase
higher potential development rate, and conceivably reduce the uncertainty level and
provide long-term advantages of financial stability.

Fundamentals of financial literacy are needed for effective implementation of disclosures

and credit ratings, as people should comprehend how they work and what advantages they
provide.

Likewise, the regulation of investment funds and insurance agencies depends vigorously
on the behavior of the general population and their attitude towards implementation of
disclosures and credit ratings. Henceforth, the recommendations for regulatory changes in this
area certainly depend on a sensible level of financial literacy among people. The same is true for
some different players in financial market, such as superannuation and managed funds.

16
 Moreover, the high share of financially literate people, among total customers of
financial institutions, the more viable the market train channels on latter will be and the more
stable and productive the overall financial system is probably going to turn into. By supporting
more grounded market disciplines financial literacy empowers some degree less concentrated
way to deal with the regulation and management of financial institutions. Hence, it has an impact
in diminishing the consistence costs and administrative bends that can emerge under a more
escalated type of monetary segment regulation and management.

Financial literacy also has some implications in the context of financial system
efficiency. The latter relates to its role in risk and resource allocation throughout the economy
(allocative efficiency), the economic costs related to the financial service provision (productive
efficiency), and its aptitude of innovation to meet consumer demand (dynamic efficiency)
(Hunter, Orr and White 2005). As a result, financial effectiveness can be disintegrated by
variables that make it troublesome for customers to look for and estimate the effectiveness of
financial products and services. It can be disintegrated by product complexity, or by many-sided
quality in divulgences or execution appraisals – or by inadequate levels of financial literacy.

The outcome can be imperfect and can prompt to some financial losses. Individuals can
neglect to comprehend risk they face and may purchase improper financial services – for
instance purchasing either with excessed risk or too little gain. On the other hand, they may just
neglect to buy financial services where it is valuable to do as such. While financial advice is
accessible, it can be costly or attached to a specific financial product provider.

The benefits of competition may be reduced if cannot make effective financial decisions.
More competent purchasers are better put to contribute towards competition and thus more
grounded and more effective market deals are struck. According to John Tiner (CEO of the UK
Financial Services Authority and Financial Capability Steering Group Chair) “If people know
what they want, and how to get it, the market for financial services becomes less one-sided and a
lot more efficient. Consumers will demand better, cheaper and more appropriate products and
services” (Financial Services Authority, 2004, p. 1).

Notwithstanding being vital for people and the financial system, monetary education
likewise have essential ramifications for the economy as a whole. Assets allocation is
influencing the potential development rate of the economy in the long-term. All else being
equivalent, the lower risk hazard for financial assets, the more extended term development in the
economy could be relied upon.

17
 Due to the fact that, financial literacy has a significant impact on the allocation of
resources in the economy, it can be concluded that financially literate investors, are more likely
to build a more constructive way to deal with their venture procedures, giving careful
consideration to the risk of option venture openings and the tradeoffs amongst risk and return.
They will probably try to be better set to boost the risk adjusted rates of profit for their ventures.
Thus, this is probably going to bring about assets streaming to their most gainful uses with
respect to risk hazard, prompting to a higher longer-term development rate in the economy and,
conceivably, a lower risk of cyclical volatility in the economy.

1.3 National Strategies for Financial Education

Given the importance of financial literacy nowadays many countries including Armenia
are already taking measures to increase the financial literacy among the population. In order to
achieve this goal, they develop and implement National Strategies for Financial Education
(NSFE). The latter is a high level policy document accepted by the Government, which is setting
the frame for the development of financial literacy in the long-term. NSFE gives the policy
directions, principles, methods for financial education by engaging and coordinating financial
literacy related initiatives by stakeholders in a centralized manner. According to the OECD/INFE
High-level Principles on National Strategies for Financial Education (2012, p. 7), national
strategy related to financial education is defined as:

a nationally coordinated approach to financial education that consists of an adapted

framework or programme, which

 Recognizes the importance of financial education – including possibly through

legislation – and defines its meaning and scope at the national level in relation to
identified national needs and gaps;
 Involves the co-operation of different stakeholders as well as the identification of
a national leader or coordinating body/council;
 Establishes a roadmap to achieve specific and predetermined objectives within a
set period of time; and
 Provides guidance to be applied by individual programmes in order to efficiently
and appropriately contribute to the National strategy.

The experience from G20 countries show that there is no one-size-fits-all model which
will be accepted and suitable for all countries. In order to be effective it is highly recommended

18
for NSFE to be tailored to the needs and culture of people as well as country circumstances
including the level of advancement of the financial system, regulatory and market conduct
framework, general development of infrastructure and institutions etc. This is why in some
countries the main objective of NSFE might be supporting financial inclusion (e.g. India,
Indonesia, Mexico etc.), while in others it can be generally focused on empowering consumers to
make informed and responsible financial decisions (OECD, 2013a). In addition, depending on
country it is usually important to define key target groups (e.g. women, low-income citizens,
youth etc.) of NSFE, even though it is aimed at increasing financial literacy generally for the
public. The key target groups can be separated by life-cycle principle as well, depending on
teachable moments (e.g. marriage, retirement, first job etc.). Defining key target groups might
help to build more targeted and efficient policy framework.

The national strategy should also include respective procedures regarding the governance
bodies responsible for the implementation of the policy. Depending on country specific situation
the implementation of NSFE might be under the responsibility of a certain governmental
authority (e.g. Ministry of Finance, Central Bank) or a committee formed especially for these
purpose. Most of the countries adopt the second approach, which allows to benefit from the
multiplayer effect by joining the efforts of stakeholders engaged in the committee. Here comes
the importance of the nature of mandates for financial education. Typically, when a mandate
exists, it helps to achieve efficiency in terms of governance (OECD/INFE, 2015). Particularly, it
ensures:

• Sustainability - public institutions are able to easily design and implement financial
education related projects, which are sustainable over the time;
• Coordination – the responsibilities of public authorities are established in a clear
way, which helps to keep public accountability as well;
• Optimization – the resources can be used more efficiently between the shareholders,
as well as there will be a separate budget within the institutions specifically devoted
to the implementation of financial education projects, which will positively affect
planning as well;
• Multiplication – proper communication and planning will ensure a multiplicative
effect of joint efforts between the engaged stakeholders. In addition, a legal basis will
be formed, which will help to promote financial education for the private sector as
well as in frame of other legal acts related to education, social security etc.

19
 The policy handbook developed by the OECD/INFE (2015, p. 38) highlights the role of
mandate and separates some countries by their nature of financial education mandates, which is
show in the Table 1.2 below.

Table 1.2 The nature of financial education mandates in several countries

Source Economy and Institution

Australia - Australian Securities and Investments
Commission (ASIC)
Explicitly enshrined in the
Canada - Financial Consumer Agency of Canada (FCAC)
institution’s founding act or its
Estonia - Financial Services Authority
revisions
Indonesia - Financial Services Authority
(“financial
Japan - Financial Services Agency and Central Council for
education/empowerment” and/or
Financial Services Information
“financial literacy/capability/etc.”
South Africa - Financial Services Board
are mentioned)
Turkey - Capital Markets Board
United Kingdom – Money Advice Service
Armenia – Central Bank
Latvia - Financial and Capital Market Commission
(FCMC), (in accordance with Law on the FCMC)
Lebanon - The Institut des Finances Basil Fuleihan
Indonesia – Bank Indonesia
Implied from other
Malaysia – Central Bank (Bank Negara Malaysia)
responsibilities
New Zealand - Commission for Financial Capability (from
(for example: provision of
the Superannuation and Retirement Income Act)
information to consumers, ensuring
Peru - Superintendence of Banking, Insurance and Private
a smooth functioning of financial
Pension Funds (from financial inclusion)
markets, financial consumer
Portugal - Central Bank (Banco de Portugal), Portuguese
protection mandates, etc.)
Securities Market Commission and Portuguese Insurance
and Pension Funds Supervisory Authority
Spain – Securities Market Authority (from protection of
investors) and Central Bank
Thailand - Bank of Thailand (raise people’s standard of

20
 living and ensure financial stability)
Czech Republic - Ministry of Education
Granted through primary Hong Kong, China - Investor Education Centre
legislation (amendment of the Securities and Futures Ordinance,
(an act from parliament, legislative 2012)
assembly, etc., giving your Mexico - Committee on Financial Education (Law to
institution a clear mandate for regulate financial groups)
financial education in the context United States - Financial Literacy and Education
of a specific policy or reform) Commission (Fair and Accurate Credit Transactions Act of
2003)
Given through secondary
Brazil - The National Committee for Financial Education
legislation
(Presidential Decree)
(Financial markets regulation,
Russian Federation - Ministry of Finance
government decree, National
Spain - Ministry of Economy (Royal Decree)
Development Plan, etc.)

As mentioned in the table in Armenia the responsibility to carry financial education

related projects is mainly given to CBA with the mandate to implement financial consumer
rights protection. Starting from 2007, CBA highlighted the importance of financial education as
prerequisite for financial stability, financial inclusion and economic prosperity in the country. In
order to ensure effectiveness in 2012, a Steering Committee for the development and
implementation of National Strategy of Financial Education is established with the involvement
of CBA (leading), the Government, around 20 private and social sector representatives. Later
during the same year, the National Strategy of Financial Education in Armenia and 2015-2019
action plan are developed by this Steering Committee. Within 2 years a major change in the
landscape of national strategies across the globe might be noticed.

In the policy handbook, OECD/INFE also discuss the situation is several selected
countries regarding the status of implementation of NSFE, particularly the Table 1.3 below
(2015, p. 11) demonstrates the situation in 2015 in different countries.

21
 Table 1.3 The status of national strategies in different countries in 2015

Status of the national strategy Number Countries and territories

Australia, Czech Republic, Japan,
A NSFE is being revised or a second Malaysia, Netherlands, New Zealand,
11
version is being implemented Singapore, Slovak Republic, Spain,
United Kingdom, United States
Armenia, Belgium, Brazil, Canada,
Croatia, Denmark, Estonia, Ghana, Hong
Kong, China, India, Indonesia, Ireland,
A (first) NSFE is being implemented 23
Israel, Korea, Latvia, Morocco, Nigeria,
Portugal, Russia, Slovenia, South Africa,
Sweden, Turkey
Argentina, Chile, China, Colombia,
Costa Rica, El Salvador, France,
Guatemala, Kenya, Kyrgyzstan,
A NSFE is being actively designed 25 Lebanon, Malawi, Mexico, Pakistan,
Paraguay, Peru, Poland, Romania, Saudi
Arabia, Serbia, Tanzania, Thailand,
Uganda, Uruguay, Zambia
Austria, Macedonia, Philippines,
A NSFE is being planned 5
Ukraine, Zimbabwe

It is very rewarding to outline a group of 11 countries, which in 2015 have been in the
process of implementing a new national strategy based on outcomes of the first national strategy
or reviewing their initial strategy to ensure effective implementation. In opposite to these
countries there are many countries, which are in the process of planning and designing national
strategies for financial education.

An important role in NSFE plays evaluation. The proper measurement of financial

literacy is important to understand the effectiveness of implemented policies as well as revise the
projects which have less impact. For this reason, it is important to mention that Armenia
developed quite comprehensive evaluation scheme, which includes baseline evaluations within 5
year periods as well as impact evaluation for separate projects implemented in frame of NSFE.
The baseline evaluations aim to measure the level of financial literacy in the country and are an

22
important prerequisite for revising the strategy for the next 5-year period, in contrast the impact
evaluations are important to assess the effectiveness of projects developed and implemented in
frame of NSFE for separate target groups.

To conclude, countries and institutions might design and implement financial literacy
related interventions, however in order to be effective there is a need to have systematic, holistic
approach to the issue. The latter will ensure that each institution in the country, which is
interested to increase the level of financial literacy in the country will join efforts and under
centralized governance will deliver better outcomes.

1.4 The measurement approaches of financial literacy

During the G20 Presidency meeting in 2012, the importance of financial literacy and
education is highlighted on a global scale. Moreover, improving financial literacy has emerged
as a strategic policy objective that complements governments’ financial inclusion and consumer
protection agendas. Especially, more attention is paid to the role of NSFE as an important policy
insight for the improvement of financial literacy among population. According to the OECD at
least 36 countries have established or are in the process of designing NSFE. Moreover, 80
percent of these countries have used a survey as a diagnostic method to identify the key priorities
for their national strategies (OECD, 2013a), as it allows respective public authorities to set
specific and realistic target measures, which will serve as a base for measuring the success. Most
of the countries, which conducted financial literacy measurement, used national financial literacy
surveys relying either methodologies developed by the international organizations and/or
elaborated national methodologies on financial literacy assessment (OECD/INFE, 2015).

Even though globally there is a consensus related to the importance of financial literacy
and many countries have policies on national level dedicated to this issue, there is not a common
generally accepted approach for measuring the level of financial literacy, which also might be
noticed from the Table 1.4 (OECD/INFE, 2015, p. 23).

23
 Table 1.4 Surveys used by different countries to measure the level of financial literacy

Measurement
Number Economies
Tool
Albania (2011), Armenia (2010), Bolivia (2013),
British Virgin Islands (2010), Colombia (2013),
Czech Republic (2010), Ecuador (2013), Estonia
(2010), Finland (2014), France (2014), Germany
(2010), Hungary (2010), Iceland (2011), Indonesia
OECD/INFE
31 (2012), Jamaica (2012), Japan (2011), Korea (2013),
survey
Ireland (2010), Latvia (2014), Lithuania (2012),
Malaysia (2010), New Zealand (2013), Norway
International
(2010), Peru (2010, 2013), Poland (2010), Saudi
surveys
Arabia (2014), Serbia (2012), South Africa (2010,
2012), Thailand (2013), United Kingdom (2010)
Armenia (2012), Azerbaijan (2009), Bosnia and
Herzegovina (2011), Colombia (2012), Lebanon
WB/RTF (2012), Mexico (2012), Mongolia (2012), Russian
12
survey Federation (2008, 2012, 2015) Tajikistan (2012),
Turkey (2012), Uruguay (2012), West Bank and
Gaza (2011)
Australia (2003, 2005, 2008, 2011, 2014), Armenia
(2014), Austria (2014), Brazil (2008), Canada (2009,
2014), China (2013), Czech Republic (2007), France
Financial
(2014), Hong Kong, China (2012, 2014), Indonesia
literacy 18
(2013), Israel (2012), Japan (2011), Netherlands
National surveys
(2013), New Zealand (2005, 2009, 2013), Portugal
surveys
(2010), Singapore (2005, 2013), United Kingdom
(2006, 2013), United States (2009, 2012)
Argentina, Chile, Italy (2008, 2010, 2012),
Household
6 Philippines (2009), Russia (2013), Spain (2002,
surveys only
2005, 2008)

24
 To conduct a nationwide survey, the countries mostly used international surveys, which
are developed during decade. As outlined in Table 4 international organizations such as the
Organization for Economic Co-operation and Development (OECD) with its International
Network on Financial Education (INFE) and World Bank (WB) in cooperation with Russian
Trust Fund (RTF) have elaborated financial literacy measurement methodologies, which
respectively are used in 31 and 12 countries. One of the key objectives of these methodologies is
to create a unified approach, which will help to get globally comparable data related to the level
of financial literacy across different countries. Even though these approaches helped many
countries, including Armenia, to set up a baseline measure concerning the level of financial
literacy, several issues arose and remain unsolved for policymakers depending on various aspects
related to national context. In order to understand how financial measurement should be
conducted in low-income countries considering national context, one of the main objectives of
this part of the research is dedicated to the introduction and comparison of the international
methodologies of financial literacy measurement developed by OECD/INFE and WB/RTF. It
will also be highlighted the positive and negative aspects of each approach in terms of
policymaking. Particularly, the definition of financial literacy, key components and scoring
system of each methodology will be discussed.

To discuss the OECD/INFE approach, it is important to present main principals of

financial literacy measurement methodology. For the measurement of financial literacy,
OECD/INFE has developed a standardized questionnaire, which is discussed with OECD
member countries as well. Mainly questions related to money management, planning and
choosing financial services are used in the questionnaire. In terms of financial literacy
measurement, OECD/INFE considered three main components: Knowledge, Attitude and
Behavior. Based on question results related to these components a separate Knowledge, Attitude
and Behavior scores are calculated. Interesting to notice that even though OECD mentions
“skills” separately in the definition of financial literacy, there is no separate score for this
component, and some consistency issues may arise. OECD/INFE also highlighted the
importance of calculating single score of financial literacy based on the sum of Knowledge,
Attitude and Behavior scores.

For the measurement of financial literacy, OECD/INFE used a universal approach to be

able to get internationally comparable data as a primary objective. As a result, OECD/INFE
approach may not fully reflect the specific characteristics of a country and priorities of
policymaker’s in the field of financial education. As it is already mentioned, one of the primary

25
objectives of OECD/INFE approach is the comparison of different countries in terms of financial
literacy. For this reason, the questionnaire used during the surveys is also developed according to
universal standards and may not include some important policy related or country-specific
questions. On the other hand, some questions used by OECD/INFE in different countries may
lead to bias due to cultural differences. Thus, in order to elaborate and implement national
strategy of financial education efficiently, it is important to consider policy related and country-
specific aspects of financial literacy. Though countries, which used this approach, could add
some additional questions to the survey questionnaire, this is limited and may not cover the
mentioned issue. In addition, the thematic areas included in the questionnaire are limited as well,
and do not include financial literacy related topics such as personal rights protection. The latter is
important topics as in practice many people are not aware of their rights and responsibilities
while using financial services.

As mentioned before, according to OECD/INFE approach there is not any separation

between “Knowledge” and “Skill” components of financial literacy, which makes it difficult to
understand needs of target groups more specifically and implement a targeted policy as well as
evaluate its efficiency in regard of these components. In order to design and implement financial
education related projects more efficiently it is important fully understanding the needs of target
groups. For example, usually people may have theoretical knowledge about economy as a whole,
but are not able to project its effects on the management of personal finances. It is important to
separate these components for policymakers to be able to see the cumulative result of financial
education related initiatives. Unlike OECD, FSA separated “Knowledge” and “Skill”
components of financial literacy during the exploratory study in UK (Kempson et al., 2005).
World Bank also used the separated approach.

Calculating certain scores for thematic dimensions of financial literacy is important in

terms of its assessment, but the OECD/INFE approach does not contain such a set of tools. The
financial knowledge, behavior and attitude scores calculated in accordance with OECD/INFE
methodology do not allow the financial education policymakers to comprehensively identify the
needs of the population, make clear targeting for and measure the effectiveness of the financial
education policy.

The OECD/INFE approach does not have equal approach towards “attitude” component
in the context of financial literacy measurement. The overall Financial Literacy Score as
provided for by OECD/INFE is calculated using the amount of the knowledge, attitude and
behavior scores and may take any value in the range 1 to 22. The financial knowledge and

26
behavior scores have greater weight in the overall score calculation, and the financial attitude is,
therefore, less important according to the OECD/INFE approach. Although forming an
appropriate behavior is more important in the financial education strategy point of view, the
remaining components of financial literacy should not be underestimated. Specifically, a
negative attitude can be a serious obstacle to a successful and effective financial education
projects. For example, in Armenia after the collapse of the Soviet Union, many people lost their
deposits held with Soviet banks, which resulted in citizens having built little trust in the banking
system. As a result, it is assumed that implementing financial education programs become less
effective for people who demonstrate similar attitude. Moreover, there is little evidence showing
the importance of financial literacy components in terms of their impact of financial literacy.

In case of WB/RTF, the approach to measure financial literacy is different. Here at the
beginning focus groups are used to identify main manifestations of financially capable people in
relation to knowledge, behaviors, skills and attitudes, which are then grouped into four major
thematic dimensions such as managing money, planning ahead, making choices and getting help.
As a result, the financial literacy assessment questionnaire included questions relating to the
main pattern of financial literacy reflecting these thematic areas. The WB/RTF approach is used
to identify 12 scores in four major thematic dimensions of financial literacy.

In case of this approach as well, some challenges may arise in terms of policymaking.
Thematic dimensions of financial literacy pointed out by WB/RTF are general in nature and may
not reflect the priorities set by domestic policymakers in financial education. As told earlier, the
WB/RTF Financial Literacy Assessment questionnaire is based on the feedback from the focus
groups. However, the information may not entirely cover the priorities set by policymakers in
financial education, so questions that are essential to the policy may have stayed out of the
questionnaire.

The WB/RTF approach to financial literacy assessment does not consider such factors as
education and income level. The financial literacy assessment questionnaire includes questions,
which are mostly neutral with regard to education and income levels. Whereas, reviewing
different factors, including the effect of education and income levels on financial literacy, and
considering peculiarities of the target groups, is important in terms of developing effective
education programs. Moreover, the WB/RTF approach considers calculating various financial
literacy scores only by thematic dimension, which does not enable financial education
policymakers to fully identify the needs of population and evaluate the effectiveness of the
financial education strategy. The 12 scores mentioned above do not reflect the financial literacy

27
components – knowledge, skills, attitude and behavior. It enables to understand the financial
literacy only in four major thematic fields. Moreover, the scoring model is not simple enough in
terms of public accountability, so the financial education policymakers, therefore, may face
difficulty in describing the impact of their programs to the public. The WB/RTF methodology
also does not allow policymakers to develop and implement an effective and targeted NSFE.

The brief comparison table of discussed international methodologies developed by

OECD/INFE and WB/RTF is presented below. It includes several important aspects of both
methodologies.

Table 1.5 Comparison of international approaches of financial literacy measurement

OECD Pilot Survey in 2011 World Bank Survey in 2012

A combination of awareness,
Financial capability constitutes the
knowledge, skill, attitude and
Definition of financial behavior demonstrated by
behavior necessary to make sound
financial an individual that is considered by
financial decisions and ultimately
capability/literacy his peers to be desirable as it leads to
achieve individual financial well-
good financial outcomes.
being.
Budgeting and money management,
Topics and Managing money, planning ahead,
short and long-term financial plans,
thematic areas making choices, and getting help
and financial product choice
Financial
Knowledge, Skills, Attitude,
capability Knowledge, Attitude, Behavior
Behavior
components
The methodology provides 3 types The methodology provides, 12
of scores. Financial knowledge score robust scores. 10 scores are applied
is created by counting the number of to the whole population: budgeting,
correct responses given by each not overspending, living within
Methodology respondent, financial behavior score one’s means and not borrowing for
counts positive behaviors exhibited essentials, keeping track of money,
and the score of attitude tells us saving money when possible,
whether the respondent tends covering unexpected expenses,
towards short-term gratification, or gathering information and three

28
 long-term security. In order to assess more general motivations—not
overall level of financial literacy, the being impulsive, not focusing on the
three scores for knowledge, behavior present, and being achievement-
and attitudes are summed. The score oriented. The remaining 2 scores are
can take a minimum value of 1, and applied to only part of the
a maximum value of 22. As the three population: planning for one’s old
scores have different maximum age (only people aged under 60), and
values, the combined score is choosing appropriate financial
implicitly weighted. 9 point out of products, which encompasses
22 is formed from the score of shopping around and checking the
behavior, 8 from knowledge and 5 features of the product bought, and
from attitude. only applied to people who are
active product purchasers.
3 scores for knowledge, attitude and
10 main scores and 2 specific scores
Scoring model behavior as well as financial literacy
calculated for specific target
overall score

Whenever there is a need to assess the level of financial literacy in the country, it should
be considered that both international approaches of financial literacy measurement developed by
OECD/INFE and WB/RTF contain some challenges, which impact policymaking process. This
is why, for the countries there is a need to develop more comprehensive tool for financial literacy
measurement, which will be based on national context.

In addition, it is important to consider also the role of qualitative and quantitative

methodologies. Findings of qualitative and quantitative are important for policymakers to
develop and implement evidence-based and targeted financial literacy related strategies.
Particularly these kind of research initiatives can be used to:

 develop certain priority policy areas, and design targeted strategic action in frame of
each area,
 make prioritization by identifying key target groups and/or priority themes,
considering the limited resources
 ensure proper public accountability and communication with media,
 identify important aspects for the financial service providers related to various aspects
of financial services. For example, qualitative research initiatives might reveal that
29
 designed credit related services available in the market do not meet the needs of
consumers.

The above-mentioned advantages brought by conducting quantitative and qualitative

research in the topic of financial literacy and financial behavior of consumers confirm that
proper monitoring and evaluation scheme is vital for effective NSFE implementation. The
financial literacy measurement methodology should be consistent with the international
methodologies discussed in this section, as no matter the challenges, the existing approaches
have created a strong base for financial literacy measurement. In addition, it should be designed
in a way to fully reflect the domestic characteristics and priorities established by the
policymakers, allow flexibility and thus international comparison between different countries,
use simple set of indicators for both financial literacy components and thematic areas. The latter
will allow determining the weak and strong aspects of financial literacy in the country and based
on further analysis develop more targeted and effective financial education initiatives.

It is important to capture also that NSFE development and implementation experience

leads to better understanding of financial literacy components and decision-making process of
consumers. Thus, gradually the methodologies and scope of survey instruments aimed at
measuring financial literacy should change in a way to reflect the transformation of financial
knowledge, skills, attitudes and behavior.

To conclude, it is important to underline that financial literacy measurement initiatives

among population require big amount of resources. Thus, simple and cheaper schemes should be
developed by the policymakers to keep track of key financial literacy indicators. This type of
interventions will also help policymakers to check the validity of findings and gather more
insights by triangulating data from different sources (OECD/INFE, 2015). The discussed
recommendations will be helpful for the policymakers to ensure proper measurement of financial
literacy and effective implementation of national strategies.

1.5 Financial literacy, socio-demographics and financial decisions

The relevance to study the impact of financial literacy on different economic aspects such
as financial decision-making thereby financial stability, financial inclusion and other
macroeconomic measures has been emphasized after Global Financial Crisis in 2008. Thus, in
order to be able to develop and implement an effective financial education policy the
policymakers need very comprehensive data related to the level of financial literacy in the

30
country and people behavior in terms of decision-making and based on the data understand the
relationships existing among different aspects of financial literacy and financial behavior of
consumers.

The role of financial literacy has been especially highlighted in terms of financial
decisions of individuals related to savings, retirement planning and household wealth (Lusardi,
2008b; Rooij, Lusardi, & Alessie, 2011a), choice of financial services (Lusardi, 2008a) and
borrowing (Lusardi & Tufano, 2009). Lusardi (2008b) discusses how well-equipped people are
to make saving decisions, especially related to the retirement planning. The author analysis if
individuals are financially literate, informed about existing saving plans and whether they plan
for retirement. Low level of financial literacy and lack of information impact saving behavior
and question secure retirement. The latter is especially widespread across some specific
demographic groups, such as women, African-Americans, and Hispanics, as well as those with
low education. Moreover, the amount of older workers, who do not know their type of pension
plan is about 50%. In their study, Rooij et al. (2011a) also show that the increase in financial
literacy increases the likelihood of investing in stock markets. The authors confirm the positive
link between saving behavior and financial literacy concluding that overall, financial literacy
positively affects family wealth accumulation both directly and indirectly. Thus, financial
education programs might help improve financial decision-making and saving practices,
however much should be done to ensure the effectiveness of these programs.

Related to the impact of financial literacy on consumer choices, Lusardi (2008a)

discusses the importance of financial literacy to make informed decisions regarding saving,
retirement, stock market participation, borrowing considering the complex financial instruments.
The issue of over indebtedness is a serious threat to the overall wealth of people, which is
usually connected to inefficient and not informed borrowing practice. The latter served as a base
for emerging of the concept of responsible borrowing. Lusardi & Tufano (2009) analyzed the
situation in US related to financial literacy, particularly debt literacy, and indebtedness of people,
which is a self-reported measure. The authors show that the level of debt literacy is low, and
people with lower levels of debt literacy make decisions which are costlier by incurring higher
fees and using high-cost borrowing. These people also report that they are heavily indebted or
that they cannot evaluate their situation.

In recent years, low-income people were focused by other empirical research (Atkinson
et al., 2007; Buckland, 2010). Atkinson et al. (2007) conducted a survey directed to about 5,000
UK adults, from 2006 to 2007. As a result, a financial literacy index has been put forward, which

31
has shown that through the globe, the respondents which have lower income also have middle or
low level of financial literacy.

Another survey has been conducted by Buckland (2010) among low-income Canadian
adults. The outcomes indicate proof that respondents are generally financially literate. They
figured out how to adapt to strict spending plans, carried about increasing their income, obliged
their credit, and are genuinely proficient about enrolling in governmental programs and in
buying financial services. As a result, it has been presumed that financially literacy mainly
depends not on socioeconomic status, but on personal experience. Whether Atkinson et al.
(2007) showed that individuals with lower income have middle or low level of financial literacy,
in the findings of Buckland (2010) there is evidence of financial literacy among respondents with
low-income.

Many authors have concentrated on studying the relationship between investment

decisions of adults and financial literacy (Al-Tamimi and Kalli, 2009; Van Rooij et al., 2011c).
CMVM conducted a special survey in Portugal to assess the level of financial literacy (2009). An
overview of around 15,000 families dwelling in Portugal, in the period amongst January and
April of 2009, inferring that in generally the financial literacy level in Portugal is quite low. In
particular, it has been revealed that population of Portugal does not consider securities as an
investment opportunity. However, some interesting relationship between investment decisions
and socioeconomic factors have been identified, revealing that people with higher income,
education level and higher-skilled occupations are more likely to invest in securities. The
behavioral analysis of United Arab Emirates population has been conducted by Al-Tamimi and
Kalli (2009). They have found that financial literacy level of men is much higher than those of
women in UAE. Moreover, some other aspects of social economic life have an impact on
financial decisions such as religion, firms’ reputation, firms’ ethics and diversification.

The relationship between financial literacy of the population and their investment
decisions in Netherlands has been studied (Van Rooij et al., 2011c). The result of the study
suggested that people with low financial literacy are not intended to invest in securities. The
research by Calvet et al. (2007) further strengthened the notion that financial decisions in stock
market are highly correlated with financial literacy. These results are also steady with the results
of Van Rooij et al. (2011c) and Christelis et al. (2010), suggesting that more financially literate
investors tend to diversify their investments thus reducing the risk hazard. To summarize, there
is a positive relationship between financial literacy and investment decisions as well as between
financial knowledge and portfolios diversification.

32
 The study by Klapper et al. (2012) measures the impact of higher financial literacy on the
use of financial products and services among Russian population. As a result, they state that
financial literacy is positively related to the volume of investments in security market and
negatively related to the personal debt taking process.

Carpentier and Suret (2012) conducted a review of 1,814 investors from Canada who
deal with their own particular stock portfolios, so as to gauge the level of speculators'
information and soundness, reasoning that their financial literacy level is quite low. The
researchers find that speculators are for the most part unconscious of holes in their financial
knowledge. Notwithstanding, they in a roundabout way perceive these crevices by hoping to
accomplish returns underneath than or equivalent to those of the market index.

Moreover, some researches show that investors with higher financial literacy tend to
invest in lower cost funds (Hastings et al., 2010; Hastings and Mitchell, 2011) and there is a
negative relationship between investment performance and the investor’s age (Korniotis and
Kumar, 2011). Meanwhile, several studies show that financial literacy and retirement programs
are positively related (Fornero and Monticone, 2011; Van Rooij et al., 2011b; Bucher-Koenen
and Lusardi, 2011; Lusardi and Mitchell, 2011a, 2011b). However, the results are not
consensual, as Crossan et al. (2011) and Almenberg and Säve-Söderbergh (2011) found no
significant relation between financial literacy and retirement planning.

There is a low level of financial literacy among young people as well, which limits their
ability to make informed financial decisions (Chen and Volpe, 2002; Hoare, 2003; Allen et al.,
2007; Jorgensen and Savla, 2010; Altintas, 2011). The lack of financial and retirement literacy in
the curriculum highlights the significance of including these topics as a requirement at colleges
and universities (Power et al., 2011). It is also found that people are inclined to overestimate
their financial knowledge (Banco de Portugal, 2010), which brings to overconfidence
phenomenon (DeBondt, 1998).

Socioeconomic conditions influence the financial literacy, and affect the financial
decisions. The worldwide empirical studies imply that there is a negative relationship between
financial knowledge and variables such as gender, age, marital status, employment, income,
education (Henriques, 2010; Almenberg and Säve-Söderbergh, 2011; Fornero and Monticone,
2011; Bucher-Koenen and Lusardi, 2011; Lusardi and Mitchell, 2011b).

When talking about the relationship between financial literacy and gender, the outcomes
are not consistent. Though a few researchers find that female population are for the most part

33
less financially literate, (Chen and Volpe, 2002; Al-Tamimi and Kalli, 2009; Almenberg and
Säve-Söderbergh, 2011; Lusardi and Mitchell, 2011a, 2011b), others determine no relationship
between these factors (Wagland and Taylor, 2009; Jorgensen and Savla, 2010; Altintas, 2011;
Bucher-Koenen and Lusardi, 2011).

In addition, although parents might influence the financial attitude of young adults, there
is no effect on financial knowledge (Jorgensen and Savla, 2010). Clarke et al. (2005) highlight
that many parents do not communicate basic financial knowledge and skills for their children.

34
CHAPTER 2: METHODOLOGY

In this research on the impact of financial literacy on financial decisions of consumers,

the core methodology is based on the principals of quantitative research. However, some
qualitative research principals have been incorporated as well in order to ensure the analysis of
the research problem and research question from different angles. Considering the specific
aspects of the topic, the impact of financial literacy on financial decisions of consumers has been
studied by using primary and secondary data sources. On the level of individuals, the research
aims to explore the impact of financial literacy on financial decisions of consumers by using
primary data sources. Particularly, the primary research has the intention to study the relationship
between financial literacy and financial preferences of consumers regarding saving, taking debt
and seeking for information related to financial products, as well as analyze the short-term and
long-term effectiveness of financial literacy intervention (workshops) regarding the relationship
between financial literacy and financial behavior of consumers. For this purpose, binary logistic
regression, pre-post and quasi-experimental evaluation methods are used. In addition to the
primary research, secondary research is done to support the goals and objectives of this study.
The secondary research data is taken from financial literacy assessment study implemented by
CBA in 2014. The descriptive analysis and cross tabulation of the secondary data help
understand the interdependence between different aspects of financial literacy and financial
decisions of consumers. All the methodological aspects of primary and secondary research,
which has been undertaken to study the research question in frame of this thesis, are presented in
details in the next sections.

2.1 Primary research methodology and data

As already mentioned the primary research will be focusing on studying the relationship
between financial literacy and financial decisions of consumers related to saving and debt
management. The primary research in the frame of this work is comprised of two major steps
presented in Figure 2. The second step of experimental research builds on the findings of first
step and answers to more specific questions connected with the research question. Respectively
during each step different research methods and tools are used.

35
 Figure 2.1 The framework of primary research

Step 1 - Evaluating the relationship

between financial literacy and financial
preferences.

Step 2 - Evaluating the impact of financial

literacy on financial decisions of
consumers.

The first step is dedicated to understand the relationship between financial literacy and
financial preferences of consumers regarding saving, taking debt and seeking for information
about financial services by using a regression analysis. Considering the findings of regression
analyses, by the second step an experimental study is designed to test the short-term
effectiveness of financial literacy intervention (workshops) in terms of knowledge accumulation
and its impact on financial decisions of consumers in the long-term. For the short-term
evaluation of workshops pre-post evaluation method will be used. The latter will help to
understand how financial literacy intervention impacted both financial knowledge and skills
accumulation and financial preferences of consumers in the short term. For the long-term
evaluation quasi-experimental research method will be used. In frame of this study, 6 months is
considered to be the minimum effective period of time to evaluate the long-term effects. The
latter is mainly connected to the average duration of using main financial services such as
deposit and credit, as well as time limitations to complete this study.

2.1.1. Sample and Survey Questionnaire

Primary research is based on the data received through surveys in 24 villages from 6
regions of Armenia (Shirak, Gegharkunik, Lori, Tavush, Ararat, Kotayq). The villages are
chosen as a part of the project initiated and implemented by CBA, which aims to develop
financial literacy in rural areas in frame of NSFE in Armenia. Even though the villages are not

36
chosen randomly, they are relatively diverse considering factors such as region, local population,
distance from the big cities etc. Overall in 24 villages 504 respondents participated in the survey
by filling the special questionnaire designed in frame of this research (Appendix 1).

The survey questionnaire included questions related to the preferences of respondents

regarding saving, taking debt and seeking for information as well as different questions revealing
respondents’ knowledge and skills on financial matters, trust towards the financial system as
well as use of financial services. There are questions to find out the descriptive information
about respondents as well. The detailed questionnaires used in frame of this research are
presented in the Appendix 1.

Based on the number of correctly answered questions related to financial knowledge and
skills a combined score for financial literacy is calculated. Financial Literacy Score reveals
respondents’ financial knowledge and skills. It is the average number of correctly answered
questions out of 9 questions and may take values from 1 to 5 according to (1). For example, if
the respondent correctly answered only 4 questions out of 9 the FLS will be equal to 44%, thus 3
based on the values presented below. Below the qualitative descriptors for the values of the score
are presented as well. The score is used as the main indicator for financial literacy in frame of
primary research.

�� < %, − � � �
%< �� < %, − � � �
%< �� < %, − � � � (1)
%< �� < %, − ℎ� ℎ � � �
{ �� > %, − ℎ� ℎ � � �

The first step of primary research has a goal to estimate a model of qualitative nature by
using Binary Logistic regression method. These kinds of analysis are typical in the area of
financial literacy and financial inclusion, when there is a need to understand the relationship
between binary dependent variable with different explanatory variables. This approach is used to
identify the factors impacting the three dimensions of financial preferences of consumers
regarding saving, taking debt and seeking for information about financial services. According to
Hosmer et al. (2013) in order to perform binary logistic regression, the number of sample should
be about 10 times the number of parameters (p). The later might be presented by the following
formula:

37
 ≥ + , ℎ = min , (2)

In logistic regression, since the values of variables are binary and take only values 0 or 1,
there obviously will be a great deal of stacking up of the values. Thus, how many explanatory
variables can be fitted also involves how “spread out” the values of dependent variables are. This
is why in (2) the sample size is determined by the minimum number of 0’s and 1’s.

Following the rule described above, the Table 2.1 is constructed, which shows the
optimal maximum number of parameters to be included in the binary logistic regression models.

Table 2.1 The number of parameters of binary logistic regression based on sample

Obs Missing
Saving
508 378 130 130 12 7
preferences
Debt
503 301 202 202 20 12
preferences
Information
494 451 43 43 4 21
preferences

Based on the information presented in the table above the binary logistic regression
models for the first step of primary research can have maximum 12, 20 and 4 explanatory
variables respectively according the three dimensions of financial preferences of consumers
discussed in this work.

In addition, as mentioned an experimental study is conducted in frame of this research.

One of the conceptual aspects of experimental study, is the separation between treatment and
control groups. In general, while comparing two groups the null hypothesis suggest that there is
no difference between the population means of the treatment and control groups, the alternative
hypothesis states vice versa. When comparing two groups, the statistical power is the probability
of rejecting the null hypothesis and therefore stating that there is a difference between the
population means of two groups. Thus, statistical power gives an opportunity to make correct
decisions while comparing two groups. For example, if the statistical power is 0.90, a difference
between the groups can be correctly identified with the probability of 0.90. Statistical power

38
greater than or equal to 0.80 is often recognized by the research community to be sufficient,
though some researchers seek higher level of statistical power. The significance level is the
probability of rejecting the null hypothesis when it is true.

The sampling frame and sample for the experimental study designed in frame of primary
research include the same 24 villages, however in this case 12 of the villages have been assigned
as treatment group and the other 12 are assigned as control group. The form of financial literacy
intervention is workshop, which is expected to have short-term and long-term impact on the
participants. Considering the importance to ensure having significance level of 0.05 and the
statistical power level of 0.85, the sample size strongly depends on the possible effect of the
experiment. For this study a scenario when possible effect is equal to 0.2 have been chosen.
Thus, the minimum number of participants for financial literacy workshop should be 224, which
is met as on average there are 21 participants per village in the treatment group of 12 villages
(overall 251 participants). The sample size are calculated by using online sample size calculator
(2016), which is based on respective methods provided by Rosner (2010, p. 301).

In order to understand, which experimental design method should be applied in this case,
it is important to compare the samples of treatment and control groups based on the main
characteristics. For this purpose, independent two-sample T-test assuming unequal variances is
applied. The results from comparison in terms of the three dimensions of financial preferences
and financial literacy are presented in the Table 2.2.

Table 2.2 Independent Two-Sample T-Test for Treatment and Control Groups

Financial Literacy Score

Group Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
Treatment 244 2.57 0.06 0.99 2.44 2.69
Control 237 2.77 0.06 0.92 2.66 2.89
combined 481 2.67 0.04 0.96 2.58 2.76
Diff -0.20 0.09 -0.37 -0.03
diff = mean(Treatment) - mean(Control), Ho: diff = 0
t-statistic -2.3258
Satterthwaite's degrees of freedom 477.674
Ha: diff < 0 Ha: diff ! = 0 Ha: diff > 0
Pr (T < t) 0.0102 Pr (|T| > |t|) 0.0204 Pr (T > t) 0.9898

39
 Saving Preferences
Group Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
Treatment 259 0.25 0.03 0.43 0.20 0.30
Control 249 0.26 0.03 0.44 0.21 0.32
combined 508 0.26 0.02 0.44 0.22 0.29
diff -0.01 0.04 -0.09 0.07
diff = mean(Treatment) - mean(Control), Ho: diff = 0
t-statistic -0.2597
Satterthwaite's degrees of freedom 504.612
Ha: diff < 0 Ha: diff ! = 0 Ha: diff > 0
Pr (T < t) 0.3976 Pr (|T| > |t|) 0.7952 Pr (T > t) 0.6024
Debt preferences
Group Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
Treatment 255 0.36 0.03 0.48 0.31 0.42
Control 248 0.44 0.03 0.50 0.38 0.50
combined 503 0.40 0.02 0.49 0.36 0.44
diff -0.07 0.04 -0.16 0.01
diff = mean(Treatment) - mean(Control), Ho: diff = 0
t-statistic -1.712
Satterthwaite's degrees of freedom 499.287
Ha: diff < 0 Ha: diff ! = 0 Ha: diff > 0
Pr (T < t) 0.0438 Pr (|T| > |t|) 0.0875 Pr (T > t) 0.9562
Information preferences
Group Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
Treatment 248 0.07 0.02 0.26 0.04 0.11
Control 246 0.10 0.02 0.30 0.06 0.14
combined 494 0.09 0.01 0.28 0.06 0.11
diff -0.03 0.03 -0.08 0.02
diff = mean(Treatment) - mean(Control), Ho: diff = 0
t-statistic -1.1435
Satterthwaite's degrees of freedom 479.83
Ha: diff < 0 Ha: diff ! = 0 Ha: diff > 0
Pr (T < t) 0.1267 Pr (|T| > |t|) 0.2534 Pr (T > t) 0.8733

40
 In the Table 2.2, the value for Pr (|T| > |t|) shows the two-tailed p-value computed using
the t distribution. It is the probability of observing a greater absolute value of t under the
hypothesis that the difference between means of treatment and control groups is equal to zero. If
p-value is less than 0.05 it can be concluded that the difference in means is statistically
significantly different from zero. In this case p-value is below 0.05 only in case of Financial
Literacy Score. According to the analysis the Treatment and Control groups have statistically no
significant differences in terms of financial preferences however, they are statistically
significantly different in terms of Financial Literacy Score and descriptive characteristics such as
gender, age, income etc. The summary results of independent two-sample T-tests assuming
unequal variances for descriptive variables are presented in Appendix 2.

2.1.2. Binary Logistic Regression

The aim of this section is to present the constructions of Binary Logistic Models which
will be used to predict the probability of relationship between financial literacy and respondents’
preferences related to savings, taking debt and seeking for information. This method is chosen to
be used considering the binary nature of dependent variables. It is also important to mention that
for this stage of analysis the data is not split into treatment and control.

The saving preference in terms formal (e.g. deposit) and informal (e.g. keeping at home)
ways illustrated by �� = �( = | ≈ ). Similar approach is used in case of debt and
information preferences as well. Meanwhile, since the regression is binomial the method of
Generalized Linear Model (GLM) introduced by Nelder and Wedderburn (1972) is used. The
scores are calculated in a first step by � + � � that represents a linear combination of
explanatory variables. In the second step, the saving preference will be estimated by means of
the function G, which is a logistic function that takes values between 0 and 1.

�� = �( = | ≈ )= � + �� ) (3)

Variance function and link function are given by the following way:

41
 � = ∗ − / (4)

= ln / − (5)

= exp / + exp (6)

In this study the preference of respondents to save in form of deposits serves as a

dependent variable, which takes only values 1, if the responded mentioned deposits as a
preferable way of saving and 0, if any other option (save at home etc.). Respectively, if the
respondent prefers to take debt or ask for information related to financial services from formal
sources, dependent variable takes value equal to 1 and 0 vice versa, when the respondent prefers
to take debt or get informed from friends and relatives. As mentioned before the goal is to find
out how the preferences of respondents are impacted from different factors. Thus, as explanatory
variables serve the descriptive demographic variables such as gender, age, education, whether
respondent has finance related education or not, marital status, employment, family income,
proportion of personal income in family income, as well as trust towards financial system and
financial literacy score (FLS).

In order to understand which explanatory variables should be included in the model to

have best fit, the summary statistics of the variables are studied as well as Pearson chi-square test
is performed, which shows the relationship of each independent variable with the dependent one.
Table 2.3 presents the summary statistics of variables as well as results from the Chi-square test.

In the P-value columns of Table 2.3, the Pearson chi-square test results for explanatory
variables are presented respectively preferences for saving, taking debt and seeking for
information. The ones, which are going to be included in the models, are chosen based on the
significance. In order to make sure that no important variable is excluded from the model, the
variables with P-value less than 0.1, which are mentioned in bold, are included in the model.

42
Table 2.3 Summary Statistics and Pearson chi-square test for financial preferences related
to saving, taking debt and seeking for information

P- P- P-
Std.
Variable # Possible values Obs Mean value value value
Dev.
(save) (debt) (info)

Gender, =1, if the respondent is

Female 2 511 0.58 0.49 0.003* 0.045* 0.003*
female, and =0 otherwise.

Age, =1, if the respondent is

Age 4 <18, =2, if 18-30, =3, if 31-60, 473 2.76 0.65 0.715 0.076* 0.716
=4, if 60>.
Education, =1, if no education,
=2, if primary, =3, if secondary,
Edu 5 500 4.13 0.98 0.793 0.749 0.02*
=4, if vocational, and =5, if
higher education.

Education related to finance, =1,

Fedu 2 480 0.22 0.41 0.079* 0.979 0.356
if Yes, and =0 otherwise.

Marital Status, =1, if married,

Ms 3 =2, if not married, and =3, if 473 1.27 0.56 0.746 0.081* 0.715
widow.
Employment of respondents, =1,
if employee, =2, if self-
employed, =3, if seasonal
Emp 8 worker, =4, if unemployed, =5, 494 2.18 1.92 0.032* 0.496 0.054*
if student, =6, if retired, =7, if
businessman and =8, if unable to
work.
Family income, =1, if <$100,
=2, if $101-$200, =3, if $201-
Inc1 5 427 2.71 1.10 0.233 0.492 0.005*
$400, =4, if $401-$700, and =5,
if $700>.

1 In the survey income are mentioned in Armenian Drams, rounded 1 USD = 500 AMD

43
 Proportion of personal income
in family income, =1, if <25%,
Pinc 5 =2, if 26%-50%, =3, if 51%- 411 2.47 1.48 0.318 0.018* 0.864
75%, =4, if 76%-99%, and =5, if
100%.
Trust towards financial system,
=1, if very low, =2, if low, =3, if
Trust 5 187 2.94 1.24 0.016* 0.73 0.363
average, =4, if high, and =5, if
very high.
Financial literacy score, =1, if
very low, =2, if low, =3, if
FLS 5 481 2.67 0.96 0.000* 0.726 0.018*
average, =4, if high, and =5, if
very high.
* the significance level is p-value<0.1

Table 2.4 Spearman’s rank correlation between the variables

female age edu fedu ms emp inc Pinc trust FLS

female 1.00 - - - - - - - - -

age -0.2* 1.00 - - - - - - - -

edu 0.18 -0.12 1.00 - - - - - - -

fedu -0.08 0.00 0.21* 1.00 - - - - - -

ms 0.07 -0.21* -0.10 0.12 1.00 - - - - -

emp -0.23* -0.11 -0.44* 0.02 0.12 1.00 - - - -

inc -0.06 -0.08 0.35* 0.18 0.08 -0.25* 1.00 - - -

pinc -0.13 0.09 -0.06 0.05 0.12 -0.07 0.06 1.00 - -

trust -0.01 0.05 -0.05 0.05 -0.16 0.07 -0.02 -0.07 1.00 -

FLS -0.19* -0.01 0.16 -0.04 -0.08 -0.16 0.13 0.01 -0.03 1.00

* the significance level is p-value<0.05

44
 Before applying logistic regression, it is also important to test the variables, which are
chosen for the model, for mutual collinearity. As some of variables are ordinal and are not
normally distributed Spearman’s rank correlation is used. The analysis results for collinearity are
presented in Table 2.4.

There is no big correlation between any of the variables included in the model, even
though only some of the values are significant. The positive Spearman's coefficient shows
positive relationship between the variables, while the negative sign shows negative relationship,
as in case of some variables. Relatively high correlation has “edu” in relation towards “emp” and
“inc”, respectively 0.44 and 0.35 in absolute terms. As there is no high collinearity between any
of the variables, it is justified to include them in the models depending on the Pearson chi-square
test results.

As a result of these statistical tests, the binary logistic models might be built describing
the relationship between different explanatory variables (including financial literacy score) and
three identified financial preferences of consumers related to savings, taking debt and seeking for
information.

Considering (3), (4), (5) the model for saving preferences has the following expression:

_ � = � +� +� +� +� +� (7)

The significance of the model for saving preferences Prob > chi2 is less than 0.05 and
Chi2 statistic is equal to 36.37. Respectively, (8) and (9), demonstrate the models for debt
preferences and seeking for information. With the global significance of Prob > chi2 less than
0.05 and with a Chi2 statistic of 26.4 and 144.02 respectively:

_ =� +� +� +� � +� (8)

_� =� +� +� +� +� � +� (9)

When a logistic regression model is built, it is assumed that the logit of the outcome
variable is a linear combination of the independent variables. It comprises two features, as both
sides of logistic regression equation are considered. First, the link function on the left-hand side

45
of the equation is assumed to be correctly represented by the logit function (in this case (5)).
Secondly, on the right-hand side of the equation, all the relevant variables should have been
included and the logit function is a linear combination of the factors. A specification error exists,
in case if the logit function is not correct as a link function or there is no linear relationship
between the logit of outcome variable and the independent variables. In practice, it is important
to understand whether the model includes all the relevant factors and if the linear combination of
them is sufficient. For this purpose, for each model a specification link test should be used to
make sure that both link function and model variables are properly specified (including all
relevant variables, and excluding irrelevant variables). The probit (based on the normal
distribution) might be used to replace logit as a link function.

The Hosmer and Lemeshow's goodness-of-fit test is another test, which is commonly
used to check the model fit. The idea behind this test is that the predicted frequency and
observed frequency should match. The more closely they match, the better the fit. The Hosmer-
Lemeshow goodness-of-fit statistic is computed as the Pearson chi-square from the contingency
table of observed frequencies and expected frequencies. Like a test of association of a two-way
table, a good fit will yield a large p-value. When there are continuous predictors in the model,
there will be many cells defined by the predictor variables, making a very large contingency
table, which would yield significant result more than often. So, a common practice is to combine
the patterns formed by the predictor variables into 4 or 10 groups and form a contingency table.

2.1.3. Financial literacy workshop design

After understanding the relationship between financial literacy and financial preferences
of consumers in terms of primary research it is important to understand the impact of financial
literacy intervention. One of the popular interventions applied by the policymakers in frame of
National Strategies for Financial Education is workshop. The latter is an intensive practical
intervention, which aims to provide basic financial knowledge and skills to the participants in
relatively short period of time. In frame of primary research, the short-term and long-term
effectiveness of financial literacy workshops is evaluated.

The workshop is designed by CBA in cooperation with Savings Banks Foundation for
International Cooperation, which later is adapted according to the needs of this research. The
content is based on the international experience as well as local focus groups among target
population. Thus, a two-day workshop is designed, which covers essential topics in financial

46
literacy. The optimal duration for the workshop is 5-6 hours in total, which may vary depending
on participants.

The workshop covers several important topics such as budgeting, saving and long-term
planning, managing debt, protecting personal rights, avoiding financial frauds and scams.
Particularly it is important that it focuses on basic financial services (bank account, payment
cards, deposit, credit, insurance) as well, discussing their benefits and risks by comparing with
alternative informal methods. Considering that people in rural areas mostly use credit services,
the workshop is structured in a way that questions related to managing debt and credit are
discussed on the second day to create an additional motivation for participants to attend the
second day of the workshop as well. The detailed content of the workshop is presented in
Appendix 3.

2.1.4. Pre-post design method

In order to evaluate the short-term effectiveness of financial literacy workshops in 12

treatment villages the pre-post evaluation methodology is applied. The main purpose in this case
is to be rigor in methods as possible while at the same time design and use an evaluation method
that interrupts workshop delivery as little as possible. The rigorous evaluation design and
methods ensure that the output and outcomes of a program are accurate and credible as much as
possible, however if the evaluation is complex it might harm the dynamic flow of the program
(Colosi & Dunifon, 2006).

One of the widely-used evaluation methods is pre-post evaluation, where program

participants are given a simple questionnaire to fill both at the beginning of a program (pre-test)
and then again at the program’s completion (post-test). This method is usually used to measure
changes in participant knowledge, skills and attitudes regarding whatever the program content is
(in this case financial literacy). In case of short-term evaluation, as only treatment group villages
are studied, non-experimental pre-post design is applied, which can be presented by using the
diagram in Figure 2.2.

In this case as a pre-test serves the survey performed in both treatment and control
villages. In treatment villages in addition, right after the workshop participants filled similar
survey, which served as a post-test. Thus, the sample both for pretest and posttest is comprised of
the workshop participants. By analyzing the data received from the 12 treatment villages through
pre-post surveys the short-term effectiveness of financial literacy workshops are evaluated in

47
terms of financial knowledge and skills accumulation and change in three dimensions of
financial preferences.

Figure 2.2 Non-experimental Pre-Post Design Framework

In this case the data from the pre-test and post-test is collected from the same sample in
12 treatment villages, thus the observations are not independent of one another. To compare the
results of pre-post evaluation a paired (or "dependent") t-test will be used.

2.1.5. Quasi-Experimental design method

For the purpose of research question of this work, it is important to assess the impact of
financial literacy on financial decisions of consumers. To put the latter in other words, it is
important to analyze how the financial behavior of consumers change over the period of time if
they become financially more literate. To answer this question in addition to the non-
experimental pre-post evaluation there is a need to conduct an experimental study in order to
reveal the long-term impact of the financial literacy intervention in the form of workshops.

Cook (1979) suggested quasi-experimental designs, as an alternative to randomized

controlled trial experiments, which are difficult to organize in real life. Quasi-experimental
studies are not truly experimental, where research units are randomly assigned to treatment and
control groups, but might provide natural and generalizable setting that better establishes
effectiveness. For this research, as mentioned before, the 24 villages included in the sample for
control and treatment groups are not randomly assigned and considering sample characteristics

48
the Nonequivalent Groups Design method will be used. This method is one of the frequently
used quasi-experimental research approaches.

During Nonequivalent Groups Design method, the data is collected before and after the
intervention. However, assigning subjects to the intervention and comparison groups is non-
random. Thus, evaluators cannot assume equivalence between the two groups. Instead, they must
assess the differences at baseline and account for any demographic or behavioral differences in
the analysis. As already discussed there are some significant differences between treatment and
control groups. The exact differences are discussed in the in sampling and Appendix 2.
Graphically the process of the experiment can be presented in form of Figure 2.3.

Figure 2.3 Nonequivalent Groups Design Framework

The data received from the surveys among 504 respondents from 24 villages will be used
as the pre-intervention data, after six months another survey is conducted in treatment and
control villages to get post-intervention data. If the two groups are similar in their pre-test scores
prior to treatment but differ in their post-test scores following treatment, it can be confidently
claimed about the effect of the financial literacy intervention. The data for post-intervention is
collected by using phone surveys. The questionnaire used in this case is mostly focused on
financial behavior of consumers and included some additional questions regarding the details of
financial service use of respondents. In the frame of this research it is expected to reveal
consumers’ behavioral changes, which will be due to the financial literacy intervention. The
change in the financial behavior driven by financial decisions will show if consumers start

49
preferring deposit versus saving money at home, using formal financial sources to take debt or
ask for information related to financial services.

The control groups in the quasi-experimental design can be identified through matching
— a process of identifying individuals that are similar to the participants in the intervention
group on all relevant characteristics, such as age, gender, education and other factors associated
with the program exposure. As it is difficult to control the matching process, there might arise
significant differences between control and treatment groups, as in case of this research.
Therefore, it is important to interpret the findings of this research in quasi-experimental design if
any threats to internal validity are present. In case of applying nonequivalent control group
design, the threats to internal validity should be considered. These threats might include additive
effects with selection, differential regression, observer bias, contamination, and Hawthorne
effects. The groups might be comparable on a pretest. However, this does not ensure that they
are comparable in all possible ways that are relevant to the study outcome.

Threats to internal validity discussed above, which may occur due to history, maturation,
testing, instrumentation, and regression, can be eliminated in a nonequivalent control group
design. In frame of this research the internal validity will be considered and adjusted by using
Analysis of Covariance (ANCOVA), which is a statistical method that allows you to variate the
pretest measurement with the outcome measurement. ANCOVA is basically a linear regression
model that allows you to adjust for the pretest measure. In other words, the effect of the pretest
measure is removed so that the difference between the posttest measurements between the
treatment and control groups can be studied. In this case a single covariate is considered, which
is the pretest situation in terms participants’ financial behaviors. The model might be presented
by the following equation:

=� +� +� + (10)

A dummy variable Zi is used to represent the treatment and control groups (1 = treatment,
0 = control). The beta values (β) are the parameters to be estimated. The value β0 represents the
intercept, ei is the residual for the ith unit. The dependent variable (Yi) in this case is the predicted
posttest value for a given Xi value, which is the covariate.

The discussed statistical approach will be used to understand the financial behavior of the
participants. Separate scores are formed describing the financial behavior of participants related
50
to savings and debt management. Thus, the long-term impact of financial literacy intervention on
saving and debt management behavior of participants, namely change in financial decisions of
participants related to saving or debt management are assessed through composite scores.

In case of the saving behavior the following aspects are considered:

 Change in saving preferences of participants,

 Change in actual savings of the participants,
 Self-reporting saving behavior,
 Change in the proportion of deposits among participants.

Based on the mentioned aspects the Saving Behavior Score/Index (SBS), are formed,
which shows whether participants make decisions towards formal ways of saving or informal.
Similarly, the following aspects are considered for building the Debt Management Score/Index
(DMS):

 Change in debt preferences of participants,

 Change in actual debts of the participants,
 Change in the proportion of credit among participants,
 Shopping around while choosing credit
 Change in the proportion of participants, who guarantee someone to take credit.

Both SBS and DMS can take values from 1 to 5 indicating the worst and best decision-
making practice by the participants. To make the composite scores more understandable the
following descriptors can be used for participants with higher SBS and DMS:

Figure 2.4 Descriptors of people with high SBS and DMS values

High Saving • Person who is tend to make regular savings and

Behavior Score prefers formal ways of saving.

High Debt • Person, who manages his/her financies without

Management aquiring debts and prefers formal ways of
Score covering personal budget deficit.

51
 After building the needed variables for the analysis, the last part related to methodology,
is related to the assumptions of ANCOVA model. Particularly assumptions related to linearity
between dependent variable and covariate across all groups, whether pretest is statistically
significantly different across treatment and control groups and homogeneity of regression should
be tested. The mentioned test results are presented in Figure 2.5.

Figure 2.5 Linearity between dependent variable and covariate across all groups
(Treatment – green, Control – blue)

Figure 2.5 shows that both in case of saving and debt management scores the dependent
variable (posttest) and covariate (pretest) are linearly increasing across both treatment and
control groups.

52
Table 2.5 Test on statistically significant difference of pretest across treatment and control
groups

Dependent Variable: SBS_pre

Type III Sum of
Source Squares df Mean Square F Sig.
Corrected Model .236a 1 .236 .236 .627
Intercept 711.098 1 711.098 710.981 .000
Group .236 1 .236 .236 .627
Error 272.045 272 1.000
Total 995.000 274
Corrected Total 272.281 273
a. R Squared = .001 (Adjusted R Squared = -.003)
Dependent Variable: DMS_pre
Type III Sum of
Source Squares df Mean Square F Sig.
Corrected Model 3.206a 1 3.206 1.591 .208
Intercept 2787.294 1 2787.294 1383.635 .000
Group 3.206 1 3.206 1.591 .208
Error 547.936 272 2.014
Total 3371.000 274
Corrected Total 551.142 273
a. R Squared = .006 (Adjusted R Squared = .002)

Table 2.5 shows that both in case of saving and debt management scores the pretest is not
statistically significant in the 95% significance level across both treatment and control groups, as
the p-values are equal to 0.627 and 0.208 respectively.

The summary tables below both for saving and debt management behavior for the
ANCOVA using only the interaction terms, show that the homogeneity of regression test is also
met, as the p-values of the covariates by dependent variable interaction (Group*SBS_pre,
Group*DMS_pre) are respectively equal to 0.365 and 0.351 considering the significance level of
95%.

53
 Table 2.6 Homogeneity of regression in case of saving and debt management
behavior

Dependent Variable: DMS_post

Type III Sum of
Source Squares df Mean Square F Sig.
Corrected Model 26.388a 3 8.796 4.574 .004
Intercept 375.961 1 375.961 195.520 .000
Group .127 1 .127 .066 .797
DMS_pre 20.061 1 20.061 10.433 .001
Group * DMS_pre 1.679 1 1.679 .873 .351
Error 519.178 270 1.923
Total 4015.000 274
Corrected Total 545.566 273
a. R Squared = .048 (Adjusted R Squared = .038)
Dependent Variable: SBS_post
Type III Sum of
Source Squares df Mean Square F Sig.
Corrected Model 16.616a 3 5.539 5.789 .001
Intercept 210.835 1 210.835 220.366 .000
Group 6.729 1 6.729 7.033 .008
SBS_pre 3.133 1 3.133 3.275 .071
Group * SBS_pre .789 1 .789 .825 .365
Error 258.322 270 .957
Total 1191.000 274
Corrected Total 274.938 273
a. R Squared = .060 (Adjusted R Squared = .050)

To conclude, from the perspective of the research questions it is important to understand

whether the financial literacy workshops, which are implemented in treatment group villages,
positively affected participants’ saving and debt management behavior after 6 months in
comparison with the results in control group villages given the setup of quasi-experimental study
with nonequivalent groups design.

54
2.2 Secondary research methodology and data

In the developing countries, it is difficult to find data on financial literacy because of

many reasons. On one hand, some of the countries has not yet started to implement any policies
related to financial literacy and capability (e.g. Macedonia, Ukraine, Zimbabwe), which is why
there is no research done on the topic. On the other hand, there are countries, which participated
only in studies done by international organizations such as OECD/INFE and/or WB/RTF
(OECD/INFE, 2015). In this case, it is hard to get the row data for research purposes.

As mentioned before in Armenia financial literacy related policies are implemented since
2007 and by now there have been several studies initiated in Armenia related to financial
literacy. Particularly, Armenia is included in both studies implemented by OECD/INFE in 2010
and WB/RTF in 2012. Apart these studies CBA initiated a national survey of financial capability
to support the further development of NSFE in Armenia in 2014. This national survey is done
based on the methodology developed by CBA called Financial Capability Barometer (FCB). The
data from this study is taken to be used for secondary research. However, due to the lack of time
series it is not possible to do valuable regression analysis, so the data will be used for descriptive
analysis in contrast with primary research results.

2.2.1 Methodological Principles and Specificities

For the secondary research purpose, it is important to understand and explore the
methodological aspects of the study carried out by CBA. After reviewing the methodology of
CBA the following basic principles can be highlighted:

1. Inheritance: the existing international methodologies are considered for FCB so that
continuity and inheritance of preceding works are not obsolete. FCB approach uses the
main concepts of WB/RTF and OECD/INFE approaches regarding financial
capability, its components, and thematic areas, survey and questionnaire design.
However, FCB suggests a number of new approaches that allow to effectively
determine the level of financial capabilities by taking into account country-specific
characteristics and priorities established by the policymakers. It is more about
‘repackaging’ and ‘tuning’ the previous methodologies by providing more
comprehensive assessment system.

55
 2. Flexibility: Despites the FCB being country-specific, its toolkit allows any country
that would like to use this approach to easily tune it into its country specifications and
policymakers’ priorities.
3. Simplicity: The methodology provides a comprehensive financial literacy assessment
system, which is based on simple mathematical principles. FCB scoring approach does
not use weights while measuring financial capability across thematic areas and
financial capability components in order to get an objective and realistic picture of the
country.

FCB scoring model gives also a dynamic perspective to financial capability through
measuring it by many dimensions: thematic areas, financial capability components (knowledge,
skills, attitude, behavior), as well as gender, geography, education, age, and income. This
approach allows detailed multilateral mapping of financial capabilities and thus more targeted
and effective financial education initiatives.

FCB adopted the WB approach, which is based on 4 main financial capability

components including knowledge, skills, attitude, and behavior. In terms of thematic areas, seven
dimensions are selected, which play a primary role in implementing NSFE in Armenia and are
based on the country characteristics. The thematic dimensions are mentioned below:

1. Economy (main financial and economic concepts and their impact on personal
finances),
2. Budget Management (effective management of personal and family budget),
3. Savings and Long-term Planning (short-term and long-term savings based on
goals),
4. Debt Management (effective management of debts),
5. Shopping Around (information collection and comparison of financial products
and services for making informed decisions),
6. Rights Protection (protection of personal rights in case of any issue),
7. Safety (understanding of safe usage of financial instruments and avoidance from
scams, deception and fraud).

These thematic dimensions in extent cover topics incorporated in the international

methodologies developed by OECD/INFE and WB/RTF. Based on the financial capability
components and thematic dimensions FCB scoring model is elaborated which consists of overall

56
Financial Capability Index and 39 different sub-indexes, which are expressed in percentages.
The table below demonstrates the scoring model of FCB methodology.

Table 2.7 The scoring model of Financial Capability Barometer methodology

Components and
Knowledge Skills Attitude Behavior Overall
thematic areas

Economy AA AB AC AD A_

Budget Management BA BB BC BD B_

Savings and Long-Term

CA CB CC CD C_
Planning

Debt management DA DB DC DD D_

Shopping around EA EB EC ED E_

Rights protection FA FB FC FD F_

Safety GA GB GC GD G_

Overall _A _B _C _D FCI

Sub-indexes of the scoring model of FCB approach include 28 indexes across 7 thematic
areas and financial capability components (knowledge, skills, attitude, and behavior), 7
aggregated indexes for each thematic area, 4 aggregated indexes for each financial capability
component. Each index and sub-indexes are calculated based on the questions in the survey by
using simple average without applying any weighting.

2.2.2 Data and analysis

The financial capability assessment done by CBA is designed to be based on data

collected through direct face to face interviews using standardized questionnaire, which includes
many questions (118 questions). Questionnaire consists of 8 sections – respondent’s information
and the 7 thematic areas fully reflecting the presented scoring model. Respondent’s information
allows assessing capability by region, age, gender, income, and education.

The target for the study is adult population (18+). Individuals rather than households are
chosen as target considering that studying the needs of individuals rather than households would
57
be more appropriate and effective from financial education perspective. Geographical coverage
includes entire territory of the country including both urban and rural population of all regions.

Financial capability measurement study conducted by CBA allows country wide

representativeness with the confidence level of 95% and confidence interval of 2.5%. As a result,
the sample included 1535 respondents. In terms of secondary research, the statistical results for
separate groups of populations, with given sample, are not deemed representative but are enough
indicative for demonstrating common patterns.

For the analysis in case of secondary research descriptive approach and cross tabulation
are used to present the overall situation in the country related to the financial literacy of adult
population and their financial behaviors and decision making patterns.

2.3 Methodological limitations

While studying the relationship between financial literacy and financial decisions of
consumers, there are some objective methodological limitations, which should be addressed in
the research. Particularly, in frame of this research, it is important to mention the universal
research limitations and then go into details and discuss the specific issues associated specifically
with the primary and secondary research.

The main universal challenge related to this research is data finding related to financial
literacy in countries such as Armenia. As mentioned before financial literacy is relatively new
concept and there is not much research done in developing countries such as Armenia.
Particularly, it is difficult to find secondary data related to financial literacy and financial
decisions of consumers during any time period. Due to this issue in order to find the answer to
the research question there is a need to collect primary data regarding financial literacy, financial
decisions and behavior of people. In frame of NSFE in Armenia the Central Bank is
implementing financial literacy project in rural areas of Armenia. Thanks to this project it is
possible to collect primary data from 24 villages in 6 regions of Armenia, where the project is
being implemented. Particularly, insights from 504 individuals are studied during this research.
Otherwise it would be impossible to invest such big amount of time and financial resources in
order to collect necessary data for research. In addition, CBA provided some data from Financial
Capability Barometer 2014 survey, which served as a secondary data in this research.

Another not less important universal issue is the trust towards financial system. It may be
argued that due to historical evidence, when many people lost their deposits in the Soviet banks,
58
still have low level of trust towards financial system. The latter is an important prerequisite for
the research. To evaluate the relationship of trust and other variables in the research, a separate
question related to the trust is added to all the questionnaires used in frame of this research.
Later, based on the data collected from this question a separate variable “trust” is added to the
research analysis.

In terms of primary research there are the following challenges, which are addressed in
the research by using certain statistical approaches:

1. Sample
2. Randomization
3. External validity

The experimental research would have more statistical power and possible effect if it is
possible to engage more locality in the research. In this case the control and treatment groups
would have been bigger thus research findings would be more valid. In addition, the bigger
number of respondents, would positively affect binary logistic regression analysis as well.
Bigger sample size would also positively support the analysis for a longer-term (1 year and
more) effects of financial literacy intervention of the decisions of consumers. However, the
sampling frame of this research is more than sufficient for the analysis in frame of this research
and 6 months is enough to notice any change in the decisions of consumers regarding saving and
debt management.

Secondly, as already discussed, the primary research would have been more valid, if
treatment and control groups are assigned randomly and significantly not different by their
population means. In this case, the possible effect of financial literacy intervention would be
more vivid at the end of research. However, perfect randomization is difficult to arrange,
especially out of lab, while dealing with real environment. In this case, it is also challenging to
gather participants for the financial literacy workshops, as many people think that the
opportunity cost of workshop participation is very high, so they prefer to spend time on other
things. In some villages, it is efficient to gather participants with the help of village councils in
others better to gather them in local schools. The low level of motivation to participate is such
workshop might be assumed to relate to low level of trust towards financial system.

The issues related to sampling and randomization discussed above harm the internal
validity of the research. In order to overcome this issue, as already mentioned previously,
ANCOVA regression model is applied. The latter helps to reduce within-group error variance.

59
 Similar to internal validity, the external validity of primary research findings must be
critically examined. The best evidence for the external validity of research findings is replication
with different populations, settings, and times. Unfortunately, with given resources it is not
possible to ensure replication of this research findings in other setting.

In terms of the secondary research the following two challenges are identified:

1. Data continuity
2. Representativeness

The secondary data related to financial literacy is difficult to find. Until now only three
financial literacy related researches have been implemented in Armenia. As discussed in the
assessment of financial literacy, the first research has been done in frame of OECD/INFE pilot
study in 2011, after which another study is initiated by the World Bank. The latest research is
implemented by CBA in 2014. Even though these studies are very valuable, they are completely
different in terms of methodology, which is a serious constrain in terms of data comparison and
regression analysis. Moreover, there is no access to the raw data of first two studies, however,
CBA provided data for the Financial Capability Barometer. Considering these issues only
descriptive analysis and cross-tabulation are done in frame of this research based on secondary
data to support the primary research findings.

In addition, the sample size for Financial Capability Barometer is 1,536 with 95%
confidence level and 2.5% confidence interval, which is aimed at conducting survey on a
national level and the results are indicative, but not representative for separate groups of
population by socio-demographics. Although in the study some analysis for separate population
targets based on some characteristics are performed to support primary research findings.

In conclusion, it is important to mention that the objective limitations of the research

have been addressed at maximum by using different statistical and regression tools. The latter are
already presented separately and in detail under the respective headings.

60
CHAPTER 3: RESULTS

3.1 Evaluation of the level of financial literacy in Armenia

According to the Financial Capability Barometer the overall level of financial capability
in Armenia is 44.5%, which corresponds to the lower middle level of financial capability
according to the methodology. The Index estimated based on the first three characteristics
(knowledge, skills and attitude), and not including the behavior is 44.9%, which again
corresponds to an average level of financial capability. Among the 7 topics of financial
capabilities savings and long-term planning has the lowest sub-index (33.7%), which
corresponds to a low level of financial literacy. The weakest characteristics in savings and long-
term planning is Knowledge (27.5%). Protection of personal rights has the highest sub-index
among the topics (57.8%). Among 4 characteristics of financial decision making Knowledge has
the lowest sub-index (29.3%). At the same time, attitude has the highest score (56.7%). Hence,
lack of financial knowledge of population is compensated by more developed skills and attitudes
(Central Bank of Armenia, 2015).

Looking more in details into the data of the study done by CBA, it might be noticed that
the banks are the most well-known financial organization among Armenian population: almost
80% of respondents mentioned banks as organizations providing financial services and 70%
recalled them in the first place. The next most frequently mentioned financial organizations are
the credit organizations and pawnshops, mentioned by 30% and 27% of respondents
respectively. The most often mentioned financial service is credit (35% of respondents use credit
services). The second most popular financial service is bank account (mentioned by 30% of
households). Compared to Yerevan (the capital of Armenia), rural and regional population use
credit more often. The situation is the opposite in case of bank account, as a higher share of
urban population has bank accounts. 35% of respondents indicated that they do not use any
financial services at all. This indicator is a little higher in Yerevan and is nearly the same across
urban and rural population. The main reason of not using financial services is explained by the
absence of any need. Interestingly, in case of saving by deposits, one third of respondents
claimed that deposits “are not accessible” to them, which is assumed to be explained by a
behavioral pattern, that people need to have big amount of savings in order to put in a bank.
Figure 3.1 below shows the use of financial services and tools by the population according to the
data.

61
 Figure 3.1 The share of financial services and tools people use in Armenia

Credit/Credit Line /Overdraft 35.4%

Bank Account 27.9%
POS terminal/ATM 20.4%
Payment Cards 18.6%
Money Transfer 16.8%
Insurance 16.2%
Currency Exchange 11.3%
POS terminal/ATM 7.7%
Deposit 4.9%
Guarantee 4.1%
Pension Fund Account 2.4%
Borrowing 2.2%
E-Commerce 1.6%
On-line (internet) banking 1.4%
Other 1.2%
None of them 34.4%

Population considers inflation and currency exchange rate as the most important
economic and financial indicators impacting family finances, although 40% of respondents failed
to name any concrete financial risk affecting family budget. Furthermore, only one third of
respondents correctly defined the concept of inflation, then the share of people understanding the
actual impact of inflation in everyday life is more than twice higher. The similar pattern might be
noticed in case of currency exchange rate. The fact that respondents consider inflation and
exchange rate important and have rather developed understanding and skills to calculate the
impact of these factors on their finances indicate that people often deal with these issues in their
everyday life. Particularly, as explained by the World Bank data on migration and remittances,
Armenia is classified among the top emigration and remittance-receiving countries. In 2014
Armenia, together with countries like Tajikistan, Kyrgyzstan, Nepal, Tonga, Moldova, is among
the top 10 remittance-receiving countries with 17.9% as a share of GDP (Ratha, Eigen-Zucchi, &
Plaza, 2016).

Financial literacy assessment data also shows that most of respondents correctly
understand the types of income and expenses and clearly differentiate variable and fixed
expenses. Due to agricultural production, seasonal income is more common among rural
population. 55% of rural respondents mentioned they had seasonal income, whereas in Yerevan
only 3% of population earns seasonal income. 10% of respondents evenly distribute seasonal
income throughout the year. One third of respondents do not distribute seasonal income evenly
throughout the year another 20% spend the entire seasonal income immediately after receiving it.

62
While most respondents acknowledge the importance of personal budget planning, more than
one third never plan and about 45% of respondents do not register and do not keep track of their
income and expenses at all. In terms of financial behaviors, it is interesting to see, that 60% of
those who have payment card read their statements.

The share of people who reported to have saving among the respondents is 13.6%. One
third of respondents find that it is impossible to save when expenses exceed income, and more
than half of them indicate they are not able to save because all income is spent on current needs.
Population in regions and rural areas is more inclined to save. The majority of those who save do
it for a specific purpose. Big future expenses, unexpected expenses and expenses on children are
the primary reasons for saving. More than half of those who have savings keep them entirely at
home, because “it is more secure” or “the amount is small”. The latter explains also the low level
of deposits (4.9%). Only 40% of respondents heard about the deposit guarantee scheme in
Armenia, and the majority does not know the actual size of deposit recompense. The issue of
long-term savings is more vivid, only 40% of respondents give importance to long-term savings
and the vast majority of them does not consider cumulative pension account as means for saving.

Related to the debt management, the shares of those who prefer formal versus informal
sources of borrowing are nearly equal. Around 22% borrow exclusively from informal sources.
The majority of respondents think that the borrowed amount should be in line with income and
expenditures. Half of those who borrowed said that they borrowed as much as they could afford,
while 40% borrowed as much as they needed. While most disapprove repaying the debt with a
new debt, one third either always or occasionally use that practice. The majority of respondents
think that scheduling the debt repayment is important although only half of them actually do so.
Almost two thirds of the respondents claim that they always repay their debts on time.

The most frequently mentioned source of receiving information about financial services
is television. Offices of financial institutions are in the second place. To gather information about
financial system population, prefer to watch TV and listen to the radio (mentioned by 30% of
respondents). The clear majority of respondents prefer to use formal sources for obtaining
information about financial services. While the majority of respondents consider professional
financial consultation important, only one third of respondents are ready to pay for it. Younger
people are more inclined to pay for professional advice. The majority of respondents did not
apply for financial advice over the last year. Related to shopping around, the majority of
respondents consider that the comparison of financial information is important before making

63
decisions related to financial services, however, less than one fifth of respondents do such
comparison.

3.2 Relationship between financial literacy and financial preferences

As already mentioned one of the main objectives of the primary research are to find out
the relationship between financial literacy and financial preferences of consumers. Particularly,
the latter is focused on the preferences related to saving, taking debt and seeking for information
related to financial services. For this purpose, a binary logistic regression analysis is used to
construct separate models for each financial preference discussed above.

In order to test the quality of models, which are represented by the expressions (7), (8)
and (9), Specification link test and Hosmer and Lemeshow's goodness-of-fit test are applied.
Based on the results of the tests, the models have been transformed to better describe the
relationship between dependent and explanatory variables. Particularly, the following changes
are done to ensure that the models become more practical and do not have specification issues:

1. The constant term is suppressed from all the three models, as it is greatly affecting the
specification of the models.
2. Even though the Pearson chi-square value are not significant, for research purposes
the explanatory variable “edu” are included in the model for saving preferences and
“FLS” are included in the model of debt preferences. This step had no negative
impact of the specification of the models and goodness of fit.
3. In case of debt preferences, the marital status is significant explanatory variable due
to Pearson chi-square test, however in case of including it the model has specification
issues and therefore this variable is removed from the model. The variable “edu” in
case of information preferences model is also removed due to the same issue.
4. In case of information preferences model there is a specification issue with the link
function, as a result, instead of logit link function probit is used, which has better
specification and goodness of fit.

Given the changes in the models, below the results of regression analysis are presented
separately for each model to reveal the relationship between financial preferences and
explanatory variables including financial literacy.

At first the model related to financial preferences for saving is discussed. After
transformation, the model for saving preferences now has the following mathematical form:
64
 _ � =� +� +� +� +� +� (11)

The binary variable in this case is the saving preference of consumers to invest their
savings in deposits or no. The results of the estimation for the model related to saving
preferences are shown in the Table 3.1. According to the output, 3 out of 6 variables are
statistically significant to the threshold of 5%. The significance of “edu” is close to the level.

Table 3.1 Binary Logistic regression output for saving preferences model

Logistic regression Number of obs = 163

Wald chi2(6) = 38.32
Log likelihood = -83.365691 Prob > chi2 = 0.0000
Q2_deposit Coef. Std. Err. Z P>|z| [95% Conf. Interval]
female -0.78 0.39 -2.01 0.044* -1.54 -0.02
edu -0.29 0.15 -1.9 0.057 -0.59 0.01
fedu 0.44 0.45 0.96 0.336 -0.45 1.32
emp 0.11 0.09 1.23 0.22 -0.07 0.29
trust -0.47 0.15 -3.04 0.002* -0.77 -0.17
FLS 0.59 0.2 3.02 0.002* 0.21 0.98
* the significance level is p-value<0.05

The values presented in column Coef. show the relationship between the explanatory
variables and the dependent variable, where the dependent variable is on the logit scale. These
estimates tell the amount of increase in the predicted log odds of Q2_deposit = 1 that would be
predicted by a one-unit increase in the predictor, holding all other predictors constant. For the
variables “edu”, “fedu” and “emp”, which are not significant, the coefficients are not
significantly different from 0. In case of the variable “female” one-unit increase in the predictor,
in this case going from male to female, it is expected a 0.78 decrease in the log-odds of
dependent variable, holding all other explanatory variables constant. On other hand, from cross
tabulation with financial literacy score it is obvious that women have significantly low level of
financial literacy than men by 4 percentage points, p=0.021.

65
 Similar relationship is observed between trust and respondents saving preferences as one-
unit increase in the trust towards financial system results to the 0.47 decrease ceteris paribus.
From Table 3.1, it can be inferred that the significance level for this variable is very high. The
importance of trust is assumed to be explained by the historical evidence, which shaped the
negative attitude of people towards financial system. In the beginning of 1990s many people lost
their savings in the banks. It is also assumed to explain the fact that family income has no
significant impact on the preferences of respondents regarding saving in deposits.

Table 3.1 indicates that financial literacy score has very high significance and one-unit
increase leads to 0.59 increase in the log-odds of the dependent variable, holding all other
explanatory variables constant.

In order to check the model validity specification link test is used to make sure it is
properly specified (including all relevant variables, and excluding irrelevant variables). The
results from the test are presented in Table 3.2.

Table 3.2 Specification link test for savings preferences logit model

Number of obs = 163

Logistic regression
LR chi2(2) = 27.61
Prob > chi2 = 0.0000
Log likelihood = -83.186355
Pseudo R2 = 0.1423
Q2_deposit Coef. Std. Err. Z P>|z| [95% Conf. Interval]
_hat 1.19 0.39 3.09 0.002* 0.43 1.94
_hatsq 0.1 0.18 0.6 0.548 -0.24 0.45
_cons 0.01 0.26 0.03 0.976 -0.5 0.52
* the significance level is p-value<0.05

The link test suggests that if a model is properly specified, no additional independent
variables should be found that are significant except by chance. One kind of specification error is
called a link error. In terms of regression, this means that the dependent variable needs a
transformation or “link” function should be properly related to the independent variables. The
idea of a link test is to add an independent variable to the equation that is especially likely to be
significant if there is a link error (Pregibon, 1980). Particularly, after the regression command (in
this case, logit or logistic), link test uses the linear predicted value (_hat) and linear predicted
66
value squared (_hatsq) as the predictors to rebuild the model. The variable “_hat” should be a
statistically significant predictor, since it is the predicted value from the model. This will be the
case unless the model is completely misspecified. On the other hand, if the model is properly
specified, variable _hatsq should not have much predictive power except by chance. Therefore, if
_hatsq is significant, then the linktest is significant. This usually means that either relevant
variable(s) have been omitted or the link function is not correctly specified.

The results from specification test show that there is not any specification error in case of
this model related to saving preferences and financial literacy. The variable “_hat” is significant
by the level of 0.002, then the variable “_hatsq” is not, which means that there is no any
specification error in the model.

As coefficients in are in log-odds units, they are often difficult to interpret, so it is better
to use odds ratios. Odds ratio shows the ratio of probability of success and probability of failure.
So given equation (6) and changing it according to the model parameters, the equation for odds
ratio will be:

= exp � +� +� +� +� +� / +

exp � +� +� +� +� +� (12)

Table 3.3 Odds ratio for saving preferences model

Q2_deposit Odds Ratio Std. Err.

female 0.46 0.18
edu 0.75 0.11
fedu 1.55 0.7
emp 1.12 0.1
trust 0.63 0.1
FLS 1.81 0.36

Table 3.3 presents the odds ratios for explanatory variables of the study. The odds ratio
analysis shows that the probability of success in case of financial literacy score is high and equal
to 1.81, which means that one-unit increase of financial literacy score, will change the saving
preference towards deposits with the probability of 64.5% based on the form in (12). Similarly,

67
one-unit increase in trust will lead to change in saving preferences towards deposits with the
probability of 38.6%.

The second model describes the relationship between the preference of consumers to take
debt and different explanatory variables. As already discussed the model is transformed and is
represented by the following mathematical form:

_ =� +� +� � +� (13)

Unlike the previous model, in this model the variable “trust” is not included in the model,
as it does not have significant Pearson chi-square value.

Table 3.4 Binary Logistic regression output for debt preferences model

Logistic regression Number of obs = 366

Wald chi2(4) = 26.4
Log likelihood = -239.40008 Prob > chi2 = 0.0000
Q9_debt Coef. Std. Err. Z P>|z| [95% Conf. Interval]
female 0.39 0.21 1.86 0.064 -0.02 0.80
age 0.35 0.11 3.11 0.002* 0.13 0.57
pinc -0.20 0.07 -2.91 0.004* -0.34 -0.07
FLS -0.11 0.10 -1.11 0.265 -0.31 0.09
* the significance level is p-value<0.05

Table 3.4 show the binary logistic regression output for this model. According to the
output only “age” and “pinc” variables are significant with respectively 0.002 and 0.004 P>|z|
value. For the variables “female” and “FLS”, which are not significant, the coefficients are not
significantly different from 0. In case of the variable “age” one-unit increase in the predictor, in
this case going from one age group to another, it is expected a 0.35 increase in the log-odds of
dependent variable, holding all other explanatory variables constant. However, in case of
explanatory variable “pinc”, which shows the proportion of personal income in family income, it
can be noticed a negative relationship.

68
 Based on the link test results, which are shown in Table 3.5, it can be affirmed that there
is no specification issue for the model of debt preferences as well.

Table 3.5 Specification link test for debt preferences logit model

Number of obs = 366

Logistic regression
LR chi2(2) = 16.74
Prob > chi2 = 0.0002
Log likelihood = -239.33589
Pseudo R2 = 0.0338
Q9_debt Coef. Std. Err. z P>|z| [95% Conf. Interval]
_hat 1.06 0.33 3.25 0.001* 0.42 1.70
_hatsq -0.15 0.44 -0.35 0.725 -1.01 0.70
_cons 0.03 0.15 0.19 0.846 -0.27 0.33
* the significance level is p-value<0.05

However, as mentioned before, the coefficients are in log-odds units and often difficult to
interpret, so it is better to use odds ratios. Odds ratio for the model of debt preferences are
presented below in Table 3.5. In addition, changing the given equation (6) according to the
model parameters, the equation for odds ratio will be:

= exp � +� +� � +� / +

exp � +� +� � +� (14)

The odds ratio analysis shows that the probability of success in case of age group is high
and equal to 1.41, which means that one-unit increase of age group, will change the debt
preference towards formal sources such as credit from banks or credit institutions with the
probability of 58.5% based on the expression in (14). Similarly, one-unit decrease in personal
contribution in family budget will lead to change in debt preferences towards formal sources
with the probability of 45%. In this model financial literacy score is not significant, which
basically means that there is no significant relationship between financial literacy and consumers
preferences of taking debt from informal or formal sources.

69
 Table 3.6 Odds ratio for debt preferences model

Q9_debt Odds Ratio Std. Err.

female 1.47 0.31
age 1.41 0.16
pinc 0.82 0.06
FLS 0.89 0.09

The third model describes the relationship between the preference of consumers related
to the sources of information about financial services and different explanatory variables. This
model is also transformed and gender, employment, family income and financial literacy score
are used as explanatory variables in the model. The rule of choosing explanatory variables for
this model are the same as for previous two, so the variables, which do not have significant
Pearson chi-square value, are excluded from the model. However, in this case the link test shows
that there is a specification error, which might be related either to the link function or
explanatory variables included in the model. As in this case the explanatory variables are already
chosen based on the Pearson chi-square and there is not multicollinearity problem, it is assumed
that the specification issue is related to the link function, which is why as already mentioned
probit link function (14) is used to slightly correct the specification of the model.

�0 +� � �
= ∫
� −∞
− (15)
√

Table 3.7 Specification link test for information preferences probit model

Number of obs = 388

Probit regression
LR chi2(2) = 19.24
Prob > chi2 = 0.0001
Log likelihood = -93.484281
Pseudo R2 = 0.0933
Q12_info Coef. Std. Err. Z P>|z| [95% Conf. Interval]
_hat 2.06 1.06 1.94 0.053 -0.02 4.14
_hatsq -0.40 0.36 -1.11 0.268 -1.11 0.31
_cons -0.61 0.74 -0.83 0.409 -2.07 0.84
* the significance level is p-value<0.05

70
 The link test results for the probit regression are slightly better and are presented in Table
3.7. Based on the link test results, it can be noticed that “_hat” variable is on the margin of
significance level of 0.05, unlike the logit regression model for information preferences, where
P>|z|=0.083>0.05. Thus for this case the probit link function better specifies the relationship of
information preferences with explanatory variables.

Table 3.8 Binary Probit regression output for information preferences model

Probit regression Number of obs = 388

Wald chi2(4) = 201.33
Log likelihood = -94.110226 Prob > chi2 = 0.0000
Q12_info Coef. Std. Err. Z P>|z| [95% Conf. Interval]
female 0.43 0.18 2.40 0.017* 0.08 0.79
emp -0.03 0.05 -0.71 0.478 -0.12 0.06
Inc 0.29 0.09 3.30 0.001* 0.12 0.47
FLS 0.21 0.09 2.35 0.019* 0.04 0.39
* the significance level is p-value<0.05

As it is already clear the model, which is used to describe the relationship of information
preferences and explanatory variables, it is important to look through the output of binary probit
regression, which is presented in Table 3.8 and take out the main conclusions.

Table 3.8 shows the binary probit regression output for this model. According to the
output the variables “female”, “inc” and “FLS” are significant with respectively 0.017, 0.001 and
0.019 P>|z| value. For the variable “emp”, which is not significant, the coefficients are not
significantly different from 0.

The regression coefficients can be used to predict the probability of information

preference. For this model, the predicted probability of information preference is described by
the cumulative distribution function by using the respective coefficients. However, interpretation
of the coefficients in probit regression is not as straightforward as the interpretations of
coefficients in linear regression or logit regression. The increase in probability attributed to a
one-unit increase in each predictor is dependent both on the values of the other predictors and the
starting value of the given predictors. For example, if “female”, “inc” and “emp” explanatory

71
variables are held constant at zero, the one-unit increase in financial literacy score from 2 to 3
has a different effect than the one-unit increase from 3 to 4. In addition, the interpretation of the
individual regression coefficients is limited as well. A positive coefficient means that an
increase in the predictor leads to an increase in the predicted probability. A negative coefficient
means that an increase in the predictor leads to a decrease in the predicted probability (Hosmer et
al., 2013). Particularly, in this model all three variables which have coefficients significantly
different from zero, have positive impact on the predicted probability.

As a result of the regression analysis, it is important to summarize and see the

relationship of financial preferences with different explanatory variables, especially the
relationship of financial literacy and financial preferences is crucial. As the models show
financial literacy has a significant impact only while discussing the preference of consumers to
save and seek for information. Although it is difficult to compare the impact of financial literacy
on these preferences due to different regression analysis, which are used, it can be noticed that
the relationship is positive. Thus, any increase in financial literacy will lead to more formal
savings and formal communication between financial service providers and consumers.
Particularly, in case of saving preferences one-unit increase of financial literacy score will lead
to the change towards deposits with the probability of 64.5%; on the other hand, it will the
increase the z-score of information preference prediction by 0.21. Unlike these two, financial
literacy has no significant impact on consumers’ debt preferences, which basically mean that no
matter how financially literate the person is the preference to take debt from formal sources such
as credit or informal sources such as friends is not affected. In this case, one-unit increase of age
group, will change the debt preference towards formal sources with the probability of 58.5%, but
one-unit increase in personal contribution in family budget will lead to the preferences towards
informal sources with the probability of 45%. Such results regarding debt preferences of
consumers can be assumed to be connected with the understanding of responsibility among
consumers. The consumption and saving preferences change parallel to the age. According to
Stöver (2012) the consumption pattern has a bell shape and people tend to consume more
between the ages of 25-70, thus the preference towards formal sources, as the loans are usually
targeted and when for example a consumer wants to buy a house it is way more difficult to find
such amount of money through informal borrowing, than taking a mortgage. As for financial
literacy, it can be assumed that whenever person needs money, the access plays a bigger role and
person makes a decision based on the access of needed money rather than efficiency and
personal benefit. For example, in low-income countries the health insurance system is not well

72
developed and because of this people always have to borrow big amount of money when there is
any health issue in the family and in this case, they try to find money as soon as possible without
considering how effective the terms and conditions of credit are.

While discussing other important variables affecting saving and information preferences
of people, it is valuable to mention the role of trust in forming the saving preferences of
consumers. As mentioned before this is probably connected with the bad experience with the
financial system before. Many consumers in countries like Armenia lost their savings in the
banks after the collapse of the Soviet Union, which formed financially insecure environment. In
their research Fredholm & Taghavi-Awal (2006, p. 26) mention that “in many developing
countries there has been a historical shortage of institutions that promote transparency and
openness. In others, such institutions have been installed but the public trust and confidence in
them are often low because of mismanagement or corruption. By installing or reinstalling
institutions, and providing them with a setting (economic, legal and informational) that supports
rather than impedes them, an institutional side-effect that enhances trust and confidence for the
new system in whole can be reaped”. The issue of trust in financial system became actual due to
Global Financial Crisis in 2008 as well. In one of its reports World Economic Forum (2015)
highlights the important role of trust as a driving engine towards using more technically
advanced and innovative financial services.

The gender also has a significant role in case of saving and information preferences.
Particularly, according to the logistic regression analysis in case of saving men are more likely to
prefer formal saving than women, however in terms of information preference regarding
financial services the relationship is vice versa. The probability that women will look for more
reliable information related to financial services is higher.

3.3 Effectiveness of financial literacy intervention

As already discussed in the methodology section the second step of primary research is
the evaluation of short-term impacts of financial literacy intervention (workshops) on the level of
financial literacy and financial preferences of consumers such as saving, taking debt and seeking
for information, as well as the evaluation of long-term (after 6 months) impact on the decisions
of consumers and their financial behavior regarding the above mentioned aspects. In this section,
the results from non-experimental pre-post evaluation and quasi-experimental evaluation with
nonequivalent groups design are discussed.

73
 Before introducing the pre-post evaluation results, it is also important to discuss several
challenges, which are faced during conducting workshops in villages. One of the main
challenges is related to participation. Gathering participants in villages is difficult considering
the communication and attitude related issues. People do not value financial literacy workshops
as they tend to overestimate their knowledge and skills in this area and this is visible from their
reaction and attitude. Despite this issue the required number of participants are gathered by the
help of local authorities based on the sampling frame. The number of participants per village on
the first and second days of the workshop is separately presented in the Table 3.9.

Table 3.9 Number of participants to financial literacy workshops in treatment villages

Number of participants Number of participants

Village Region
(pre) (post)
Sevqar Tavush 34 23
Lusadzor Tavush 23 23
Darpas Lori 28 16
Bazum Lori 20 13
Mrgavan Ararat 9 6
Marmarashen Ararat 22 10
Bjni Kotayk 15 7
Lernanist Kotayk 30 17
Gandzak Gegharkunik 13 11
Noratus Gegharkunik 17 13
Basen Shirak 28 18
Beniamin Shirak 12 7
TOTAL 251 164

Based on the numbers presented in the table above, it can be noticed that in some villages
the workshops are quite popular and around 30 participants are gathered, on the other hand some
had too few participants. However, this had little impact on the overall sample, as 251 people
participated and on average each village had 21 participants, which is satisfactory considering
the logistical limitations.

74
 In addition, it is also challenging to ensure the presence of the same participants for the
second day of the workshop. The issue is reflected in the number of participants, who filled the
pre-survey and number of participants, who filled the post-survey. Overall the retention rate for
the financial literacy workshop in all 12 villages are on average around 65%. In this case also
some villages had higher retention rate comparing to others. In this case it is important to
mention that no motivating tricks are used to ensure high retention rate, except building the
workshop content in a way that debt management topic, which is the most relevant for the target
group, are scheduled on the second day. The exact reasons why some of participants did not
appear on the second day remain controversial and the participation would be much higher if
some incentives (e.g. lottery, presents, remuneration scheme, certificates) are integrated into the
workshop.

While conducting the workshops another significant challenge is negative predisposition

of participants and low level of trust towards financial system, which is only around 60% only on
average. However, the negative predisposition is not only conditioned by low level of trust
towards financial system, but also low level of trust towards the government and low standard of
living in general. These reflections are mainly based on the experience during implementing
workshops and there is a need to conduct deeper behavioral research to find out the casual
relationships and factors affecting such attitude.

The participants are mainly active community members (teachers, employees of

administration, farmers etc.). The majority of participants are women around 68%. Similar
situation can be noticed on the country level, as Armenia has relatively high number of seasonal
work migration to other countries, which creates a gender imbalance in the country. In terms of
age groups, the participants are mainly middle aged, 47% of participants are between the ages of
25 to 44 years old, another major part (29%) are between 45 and 64 years old. The overall
literacy level in Armenia is quite high compering to the other low-income and lower middle
income countries (Education Policy and Data Center, 2014). This is reflected in this research as
well, as almost 92% of the participants have at least secondary education; moreover, around 22%
of them has higher education related to finance or economics. Regarding the employment of the
participants as mentioned before mainly the active community members joined the workshops, in
other words the majority about 72% are employees and the rest of participants are almost equally
distributed among other categories. The second major part of participants (10%) are seasonal
workers who live and work in other countries (mainly in Russia) most of the year.

75
 While discussing financial literacy and financial decision making, it is important to
compare different variables based on categories related to personal and family income. For this
reason, two questions are included in the questionnaire, first of which aims to reveal the level of
family income per month and the second identifies the proportion of participant’s income in the
family income. The latter is also an important indicator, which shows the decision making power
of a participant in the family. Based on the pre-survey it turned out that around 27% of the
respondents have family income between $100-$200 and 23% of participants have family
income between $201-$400. The average household in rural areas of Armenia comprised of 5
family members on average (Petrosyan & Magluchants, 2000). Regarding the proportion of
personal income in the family in general the level is low, around 44% of participants had less
than 50% share of the family budget, which also might be connected with the gender imbalance
deviated towards women. In this case it is assumed that a lot of men work abroad and send
remittances back to their families in Armenia and, although women in the families have
proportionally less income, they are managing the family budget. While talking about income
related data, it is important to mention that many participants preferred to leave these questions
empty, this is why in case of family income the missing values in the data are about 25% and in
case of proportion of personal income respectively 27%.

Before discussing the short-term pre-post evaluation results, it is important to discuss

some financial behavior related aspects, which will be discussed more in details in frame of
evaluation of financial literacy workshop in long-term. Particularly, at the moment of workshops
almost half (48%) of the participants had debt mainly from formal sources. Only 28% of the
participants had debt from informal sources only. Almost 14% had savings and only 4% has
deposits. These results also highlight the issues associated with saving preferences and behavior
of consumers. The binary logistic regression analysis confirmed not only the significant
relationship between financial literacy and saving preferences but also relationship of trust with
saving preferences. More than 25% reported that they have neither debt, nor saving. However,
while crosschecking the date it can be noticed that 15% of these respondents mentioned that they
have a credit and 3% mentioned that they have a deposit, which enables to assume that some of
the respondents tried to conceal true information by artificially deteriorating or improving their
financial state. The use of financial services by the participants of the workshops are presented in
Figure 3.2.

76
 Figure 3.2 The use of financial services in treatment group

Bank account 27.2%

Payment card 47.2%
Credit 47.9%
Deposit 4.2%
Money transfer 15.1%
Credit guarantee/ Co-signer 29.4%
Other 2.3%
None 13.2%

From Figure 3.2 it can be noticed that the most popular financial services among
participants of the workshops are loans and payment cards. More than 47% of the participants
have payment cards, though only 27% of them mentioned to have a bank account, which is a
lack of understanding related to the statement that one cannot have a payment card without
having a bank account. The number of people co-signing loan agreement is also relatively high
(29%); moreover, 34% of the cosigners do not have any loan, and 41% of participants, who have
loan, are also cosigners, which is increasing the risk of over indebtedness. Surprisingly only 15%
of participants reported to use money transfer services. This number should have been higher due
to high number of seasonal labor migration. Around 13% of participants do not use any financial
service at all, which might have different reasons, such as trust towards financial service
providers, access to the branches or ATMs, need of financial services, access to information as
well as financial literacy.

As mentioned in the methodology for the sake of this part of primary research Paired t-
test will be used to compare the population means of the same group before and after the
financial literacy intervention. A paired (or "dependent") t-test is used when the observations are
not independent of one another. In this case, the same participants filled both the pre-test and the
post-test. Hence, it is expected that there is a relationship between the financial literacy score and
other variables before participation to the workshop and after it. The paired t-test accounts for
this. For each participant, it is essentially being considered the differences in the values of the
two variables and testing if the mean of these differences is equal to zero. In order to perform

77
Paired t-test, the variables for financial literacy score, three dimensions of financial preferences
and trust towards financial system are grouped by Villages based on the average. As a result, the
impact of financial literacy workshops in each village is calculated, having in mind that for the
pre-test on average there are 21 participants in each village, and for the post-test there are 14
participants.

One of the main assumptions before using paired t-test states that the distribution of the
differences should be approximately normally distributed. The paired t-test only requiring
approximately normally distributed data because it is quite "robust" to violations of normality in
case of real life primary research, meaning that the assumption can be a little violated and still
provide valid results. In order to check the validity of the data on village level, the Shapiro-Wilk
test of normality is used. The table below presents the output of the test.

Table 3.10 Shapiro-Wilk test of normality

Variable Obs W V z Prob>z

sp_pre 12 0.96 0.67 -0.77 0.780
sp_post 12 0.89 1.80 1.14 0.126
dp_pre 12 0.95 0.77 -0.51 0.694
dp_post 11 0.90 1.61 0.89 0.186
ip_pre 12 0.73 4.45 2.91 0.002
ip_post 11 0.82 2.91 2.11 0.017
fls_pre 12 0.85 2.50 1.79 0.037
fls_post 12 0.95 0.80 -0.44 0.669
trust_pre 11 0.94 0.92 -0.14 0.557
trust_post 12 0.88 1.95 1.30 0.097

The null-hypothesis of this test is that there is a normally distributed data. Thus, if the p-
value is less than 0.05, then the null hypothesis is rejected and there is evidence that the data
tested is not normal. From the Table 3.10 it is visible that variables ip_pre, ip_post and fls_pre
are not normal. In order to evaluate the pre-post results for the information preferences Wilcoxon
matched-pairs signed-rank test is used, which is a nonparametric alternative for Paired t-test.

78
 Table 3.11 The paired t-test results for Financial Literacy Score

Variable Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
FLS_pre 12 0.49 0.02 0.08 0.45 0.55
FLS_post 12 0.67 0.04 0.14 0.58 0.76
Diff -0.17 0.05 0.18 -0.29 -0.05
mean(diff) = mean(FLS_pre - FLS_post), Ho: mean(diff) = 0
t-statistic -3.2192
degrees of freedom 11
Ha: mean(diff) < 0 Ha: mean(diff)! = 0 Ha: mean(diff) > 0
Pr (T < t) 0.0041 Pr (|T| > |t|) 0.0082 Pr (T > t) 0.9959

Despite the mentioned challenges, the results of pre-post evaluation with the Paired t-test
analysis show that in terms of short-term financial knowledge and skills accumulation
workshops have a substantial impact. There is a distinctive change in the financial literacy score
of in villages regarding the key concepts outlined in the workshop. The results of paired t-test are
shown below in the Table 3.11.

Based on the values presented in Table 3.11, it can be noticed that the paired t-test is run
for the 12 villages included in the primary research to determine whether there is a statistically
significant mean difference between the financial literacy score before and after the workshops.
The mean of financial literacy score before the workshop is 49% then after the workshop it
increased by 37% becoming 67%. Based on the methodology the null hypothesis is rejected as
there is significant difference between the two groups. A substantial progress can be noticed in
knowledge and skills related to different specific concepts of financial literacy. This is why, in
order to understand how the knowledge and skills changed after the workshop, it is important to
look into each concept separately.

Particularly, in Figure 3.3 below, it can be noticed that the understanding of inflation and
its impact on personal or family budget has increased related to as a result of workshop
respectively 15 and 20 percentage points.

79
 Figure 3.3 The impact of financial literacy workshop on the concepts related to inflation

80.0% 74.6%
70.0%
60.0% 54.2%
50.0%
39.5%
40.0% Treatment_pre
30.0% Treatment_post
20.0% 14.6%
10.0%
0.0%
Inflation definition Inflation calculation

Positive impact can be noticeable also in terms of the skills of participants to calculate
simple and compound percentage which is one of the key concepts in financial literacy. The
impact of the workshop, especially on the knowledge and skills of calculating compound
percentage is promising. The proportion of correct answers after the workshop for the compound
percentage, as can be seen in Figure 3.4 below, increased from 43% to 69%.

Figure 3.4 The impact of financial literacy workshop on interest rate related concepts

80.0%
69.5%
70.0%
60.0% 53.8%
50.0% 45.8% 43.5%
40.0% Treatment_pre
30.0% Treatment_post

20.0%
10.0%
0.0%
Simple interest Compound interest

80
 The workshop is especially helpful in terms of giving information about important
institutions such as Financial System Mediator and Deposit Guarantee Fund2. The latter plays a
crucial role in building the trust towards financial system. Figure 3.5 shows that the short-term
impact of the workshop on the knowledge and understanding related to deposit guarantee scheme
in Armenia increased from 26% to 66%, which is a big shift. In addition, another important
concept such as understanding the principles of diversification is also impacted positively. The
concepts such as diversification, inflation and compound interest play a key role in terms of
evaluating the level of financial literacy as discussed in the study by Lusardi and Mitchell
(2011).

Figure 3.5 The impact of financial literacy workshop on the concepts of diversification and
understanding of deposit guarantee scheme

90.0%
79.6%
80.0%
70.0% 65.7%
60.0%
48.2%
50.0%
Treatment_pre
40.0%
Treatment_post
30.0% 25.7%
20.0%
10.0%
0.0%
Diversification Deposit Gurantee

One of the key measures, which is used to understand the total cost associated with
borrowing is annual percentage rate of charge (APRC). The regulation in Armenia is requiring
banks and credit organizations to disseminate all charges and fees associated with a loan and
include APRC both in marketing materials and contract. APRC is especially important while
comparing consumer loans. This is why it is important to know APRC. As it can be noticed from
Figure 3.6 below, the understanding of APRC among participants increased almost twice, which
is assumed to lead to more responsible borrowing in the future.

2 More about these institutions can be found at the official webpages www.fsm.am and www.adgf.am

81
 Figure 3.6 The impact of financial literacy workshop on understanding APRC

90.0%
80.0%
80.0%
70.0%
60.0%
50.0% 45.0%
Treatment_pre
40.0%
Treatment_post
30.0%
20.0%
10.0%
0.0%
APRC

The financial knowledge and skills related to identifying frauds and scams are crucial
nowadays in regards to the development of financial services and information technologies.
During the workshop some practical advices are given to the participants how to identify
financial fraud and scams and ensuring personal security while using various financial services
(Figure 3.7).

Figure 3.7 The impact of financial literacy workshop on understanding financial frauds
and scams

80.0% 75.4%
70.0%
60.0%
50.0% 44.4%
40.0% Treatment_pre
30.0% Treatment_post

20.0%
10.0%
0.0%
Fraud & scams

82
 In addition, the details are discussed related to personal rights protection principals. As a
result, after the workshop the knowledge and skills related to financial fraud and scams increased
from 44% to 75% and the understanding on personal rights protection increased from 27% to
32% (Figure 3.8).

Figure 3.8 The impact of financial literacy workshop on capabilities regarding personal
rights protection

33.0%
32.1%
32.0%
31.0%
30.0%
29.0%
Treatment_pre
28.0%
26.9% Treatment_post
27.0%
26.0%
25.0%
24.0%
Rights protection

To conclude, overall there is a distinguishing positive impact due to the workshop in

terms of the knowledge and skills of participants related to different financial services and
concepts. It is especially important to notice the increase in understanding of such concepts as
inflation, diversification, deposit guarantee scheme, APRC and financial security. These are the
main concepts, which is expected to drive responsible financial decision making in the future.
The long-term impact of the workshop is discussed in the next section.

There is a distinctive change in preferences of participants as well regarding ways of

saving, taking debt and seeking for information. The results of paired t-test for saving
preferences are shown below in Table 3.12.

83
 Table 3.12 The paired t-test results for Saving Preferences

Variable Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
sp_pre 12 0.23 0.03 0.12 0.15 0.31
sp_post 12 0.77 0.06 0.20 0.64 0.90
Diff -0.54 0.08 0.27 -0.71 -0.37
mean(diff) = mean(sp_pre - sp_post), Ho: mean(diff) = 0
t-statistic -6.9108
degrees of freedom 11
Ha: mean(diff) < 0 Ha: mean(diff)! = 0 Ha: mean(diff) > 0
Pr (T < t) 0.0000 Pr (|T| > |t|) 0.0000 Pr (T > t) 1.0000

The mean for pre-test and post-test for the saving preference of participants towards
formal ways of saving differ significantly. Before the workshop only 23% of participants
mentioned deposit or government bonds as a preferred way of saving, while after the workshop
the number expressively increased to 77%. In this case the corresponding two-tailed p-value is
0.000, which is smaller than 0.05. Thus, it can be concluded that the mean difference of pre-test
and post-test is significantly different from 0, moreover the mean for post-test is bigger than the
pre-test with 54 percentage points. Due to the workshop the shift in saving preferences from
informal ways to formal can also be seen in Figure 3.9 below.

Figure 3.9 The shifts in Saving Preferences based on different ways of saving

80.0% 75.7%
70.0%
60.0%
50.0%
40.0% 31.7%
31.3%
30.0% 25.1% 24.3% 21.2%
20.0% 16.8%
11.0% 8.1%
10.0% 4.2% 7.0%
1.2%
0.0%
Home Deposit Gold Property Gov. bond Other
Treatment_pre Treatment_post

84
 Based on the data it can be noticed that before the workshop only 25% of participants
mentioned deposit as a preferred way to save. However, after the workshop the share of
participants increased to 75%, in contrast to it the preferences towards saving money at home
decreased around 20 percentage points. It is expected to be the result of increased knowledge and
understanding about inflation and other risks associated with saving money at home, as well as
increased knowledge about deposit guarantee fund in Armenia, which is insuring the deposits of
consumers at the commercial banks by the amount ~$20k for the deposits in national currency
and ~$10k in foreign currencies. Small increase can be noticed also in terms of government
bonds. Still preferable ways of saving stay investment in form of gold or in property, even
though proportionally small amount of participants mentioned these ways after workshop
(respectively 24% and 17%). Apart from the given options some participants mentioned
investing money in the business as a preferred way of saving, which is included in the data under
the category “Other” and has a very small proportion compering with other options both in case
of pretest and posttest. It is worth mentioning that the financial literacy workshop is not designed
to increase business related knowledge and skills of participants.

The preference has been changed in terms of taking debt as well. After the workshop in
this area it also can be noticed a positive change in preferences towards formal sources of debts.
The paired t-test output for debt preference are presented in the Table 3.13 below:

Table 3.13 The paired t-test results for Debt Preferences

Variable Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
dp_pre 11 0.59 0.04 0.15 0.49 0.69
dp_post 11 0.80 0.05 0.15 0.70 0.91
Diff -0.21 0.05 0.17 -0.32 -0.10
mean(diff) = mean(dp_pre - dp_post), Ho: mean(diff) = 0
t-statistic -4.2506
degrees of freedom 10
Ha: mean(diff) < 0 Ha: mean(diff)! = 0 Ha: mean(diff) > 0
Pr (T < t) 0.0008 Pr (|T| > |t|) 0.0017 Pr (T > t) 0.9992

The mean for pre-test and post-test for the debt preference of participants towards formal
ways of taking debt also differ significantly. Before the workshop only 59% of participants

85
mentioned bank or credit organization as a preferred way of saving, while after the workshop the
number expressively increased to 80%. In this case the corresponding two-tailed p-value is
0.0017, which is smaller than 0.05. Thus, it can be concluded that the mean difference of pre-test
and post-test is significantly different from 0, moreover the mean for post-test is bigger than the
pre-test with 21 percentage points. The reason, why it is important to promote formal borrowing
is mainly related to the fact that people frequently get involved into informal financial relations
and it is not possible to protect their rights, whenever there is an issue. However, the workshop is
built on the principle that people should focus on effective management of the debt no matter is
formal or informal. As the over indebtedness might cause a serious harm to personal welfare.
From Figure 3.10, it can be noticed that the pattern of debt preference before and after the
workshop is the same, participants are preferring formal borrowing from banks or credit
organizations more than informal borrowing from acquaintances or local shops. However, after
the workshop the preferences are changed towards formal sources.

Figure 3.10 The shifts in Debt Preferences based on different ways of taking debt

80.0%
67.9%
70.0%
60.0%
50.0% 46.3%
40.0% 36.5%
30.0%
17.9% 16.7%
20.0% 14.2%
10.0% 5.9% 4.7% 2.6%
0.0%
0.0%
Bank Credit Friends Local shop Other
organization
Treatment_pre Treatment_post

Particularly, from Figure 3.10, it is visible that the preference to take debt from friends
and relatives decreased from 36% to 17%. It is also interesting to notice, that due to the
workshop, people stopped preferring to borrow products from local shops and paying back later
at all. In contrast, the preference to take loans from banks and credit organizations increased
respectively by 22 and 4 percentage points. Small preference change towards credit
organizations is expected to be connected with increased understanding on the real cost of loan

86
and comparison that even though it is easier to take a loan from credit organizations, but it is
more expensive.

The preferences regarding seeking for information about financial services before and
after the workshop are not changed a lot. As mentioned before in order to compare pre-test and
post-test results for information preferences Wilcoxon matched-pairs signed-rank test is used,
which is a nonparametric alternative for Paired t-test. The Wilcoxon signed rank sum test is used
when the difference between the two variables is not normally distributed (but it is assumed to be
ordinal). The output of the test is shown below in the Table 3.14.

Table 3.14 The output of the Wilcoxon signed rank sum test for information preferences

Sign Obs Sum ranks Expected

Positive 2 14 25.5
Negative 4 37 25.5
Zero 5 15 15
All 11 66 66
Ho: ip_pre = ip_post unadjusted variance 126.50
z = -1.084 adjustment for ties -0.13
Prob > |z| = 0.2785 adjustment for zeros -13.75
adjusted variance 112.63

Based on the output of Wilcoxon signed-rank test, it can be said that the null hypothesis,
which is that both distributions of pre-test and post-test are the same, can be rejected at any level
above 27.8%. Particularly, the financial literacy workshop did not cause a statistically significant
change in information preferences of participants (z = -1.084, p-value = 0.2785 > 0.05). Both
before and after the workshop the participants’ preferences are directed to formal sources of
information such as official webpages or branch offices of the financial service providers. From
Figure 3.11 below small differences in information preferences might be noticed.

The data in Figure 3.11 shows that in terms of getting information about financial
services there is a slight change in preferences from TV, radio towards more professional sources
such as branches and official web pages of financial institutions (respectively 19 and 7
percentage point increase). In terms of using informal sources of information (friends, relatives),
there is no distinguishable change.

87
 Figure 3.11 The shifts in Information Preferences based on sources of information

60.0%
52.6% 52.6%
50.0% 48.0%

40.0%
32.9%
30.0%

20.0% 14.1%
10.0% 6.5% 7.3%
2.6% 3.6% 2.6%
0.0%
TV, radio Web page Branch Friends Other
Treatment_pre Treatment_post

As a key variable summarizing the overall attitude towards using financial services and
financial system as a whole, it is important to evaluate the impact of financial literacy workshop
on the trust of participants as well. The paired t-test output for trust is presented below:

Table 3.15 The paired t-test results for the level of Trust

Variable Obs Mean Std. Err. Std. Dev. [95% Conf. Interval]
trust_pre 11 0.67 0.05 0.17 0.55 0.78
trust_post 11 0.77 0.04 0.14 0.68 0.87
Diff -0.11 0.04 0.13 -0.20 -0.02
mean(diff) = mean(trust_pre - trust_post), Ho: mean(diff) = 0
t-statistic -2.6357
degrees of freedom 10
Ha: mean(diff) < 0 Ha: mean(diff)! = 0 Ha: mean(diff) > 0
Pr (T < t) 0.0125 Pr (|T| > |t|) 0.0249 Pr (T > t) 0.9875

As can be noticed from the output in Table 3.15, the mean for pre-test and post-test for
the trust of participants towards financial system as a whole changed significantly. Before the
workshop the average level of trust towards financial system is 67%, while after the workshop
the number increased by 11 percentage points by becoming 77%. In this case the corresponding

88
two-tailed p-value is 0.0249, which is smaller than 0.05. Thus, it can be concluded that the mean
difference of pre-test and post-test is significantly different from 0. The positive impact of the
workshop on trust is important as it is expected to affect the long-term shift in behavior of the
participants.

Figure 3.12 The changes in the level of Trust towards different financial institutions

100.0% 86.8% 92.1%

76.3%
80.0% 69.7% 68.3% 69.9%
60.2% 57.9% 60.0% 57.5%
60.0% 54.4% 51.4%
40.0%
20.0%
0.0%
CBA Banks Credit Pension Insurance FSM
organization funds companies
Treatment_pre Treatment_post

In Figure 3.12 about it is noticed that relatively the lowest trust is expressed towards
credit organizations and pension funds. The latter is assumed to be connected with latest pension
reforms in the country, which brought huge wave of protests. It is gratifying to see that the trust
towards Financial System Mediator has increased from 70% to 92%, which is essential for
personal rights protection.

To conclude, based on non-experimental pre-post evaluation results, in the short-run after

the workshop the following main results can be brought out:

1. There is on average 37% increase in knowledge and skills of participants (49%-

>67%). A substantial progress can be noticed in knowledge and skills related to
the concepts of inflation, diversification, deposit guarantee and APRC (about
101%). Considering the general high level of literacy of participants, sizable
effects have not been noticed regarding participants’ understanding concepts
related to personal rights protection and calculation of simple and compound
interests (more than 32%).

89
 2. Saving preferences has been changed greatly from informal ways of saving to
formal. Before the workshop only 23% of participants preferred formal ways of
saving, however after the workshop the share of participants increased to 77%, in
contrast the preferences towards saving money at home decreased around 20
percentage points. This may be the result of increased knowledge about inflation
and other risks associated with saving money at home as well as increased
knowledge about deposit guarantee scheme in Armenia.
3. After the workshop in terms of preferences regarding taking debt also a positive
change happened towards formal sources of debts. Based on pre-post evaluation
results the preference to take debt from friends and relatives decreased from 36%
to 17%, in contrast taking debt from banks and credit organizations increased
respectively by 22 and 4 percentage points.
4. The preferences regarding shopping around before and after the workshop are not
changed significantly. People prefer to get information from web pages and
branches of financial service providers. In terms of getting information about
financial services there is a slight change in preferences from TV, radio towards
more official sources such as branches and web pages of financial institutions
(respectively 5 and 20 percentage point increase). In terms of using informal
sources of information (friends, relatives) while shopping around, there is no
perceptible change.
5. Trust towards the financial system increased by 11 percentage points on average.
Particularly, the trust toward Financial System Mediator increase by 21
percentage points, which is almost twice higher than the trust increase towards
other financial institutions.

Thus, the overall impact of financial literacy workshops in 12 villages is positive in terms
of both knowledge and skills accumulation and desired change in preferences. However, the
short-term positive impact might be neutralized in the long-term. This is why in order to evaluate
the relationship between financial literacy and financial decisions of consumers, it is important to
discuss the results of long-term quasi-experimental evaluation and see how the increase in
financial literacy affected financial behavior of financial literacy workshop participants after 6
months.

90
3.4 Relationship between financial literacy and financial decisions of consumers

The last and most important part of the research in frame of this study is evaluation of
long-term impact of financial literacy increase on financial decisions of consumers. As already
discussed in the methodology section in order to evaluate the long-term (after 6 months) impact
of financial literacy on the decisions of consumers and their financial behavior quasi-
experimental research method is used by applying nonequivalent groups design.

As discussed in the methodology in order to assess the relationship between financial

literacy and financial decisions of consumers related to savings and debt management the
Analysis of Covariance (ANCOVA) approach is used.

From the ANCOVA output for Savings Behavior Score, it can be noticed that the mean
for Treatment is greater than the mean of Control group, which means that after 6 month
participants of financial literacy workshops demonstrate better saving behavior than individuals
from the control villages (Table 3.16a). The next Table 3.16b of Levene's Test show that
significance value is equal to 0.197, which is greater than 0.05, so the assumption of equality of
error variances is not violated.

Table 3.16 ANCOVA output for the SBS in case of Treatment and Control Groups

a) Descriptive Statistics
Dependent Variable: SBS_post
Group Mean Std. Deviation N
Control 1.643 .8698 157
Treatment 2.077 1.1154 117
Total 1.828 1.0035 274

b) Levene's Test of Equality of Error Variancesa

Dependent Variable: SBS_post
F df1 df2 Sig.
1.674 1 272 .197
Tests the null hypothesis that the error variance of the
dependent variable is equal across groups.
a. Design: Intercept + SBS_pre + Group

91
 c) Tests of Between-Subjects Effects
Dependent Variable: SBS_post
Type III Partial
Sum of Mean Eta Noncent. Observed
Source Squares df Square F Sig. Squared Parameter Powerb
Corrected
15.827a 2 7.913 8.276 .000 .058 16.553 .961
Model
Intercept 210.153 1 210.153 219.796 .000 .448 219.796 1.000
SBS_pre 3.222 1 3.222 3.370 .068 .012 3.370 .448
Group 12.221 1 12.221 12.782 .000 .045 12.782 .945
Error 259.111 271 .956
Total 1191.000 274
Corrected
274.938 273
Total
a. R Squared = .058 (Adjusted R Squared = .051)
b. Computed using alpha = .05

d) Group
Dependent Variable: SBS_post
95% Confidence Interval
Group Mean Std. Error Lower Bound Upper Bound
Control 1.646a .078 1.492 1.800
Treatment 2.073a .090 1.895 2.251
a. Covariates appearing in the model are evaluated at the following values: SBS_pre = 1.624.

e) Pairwise Comparisons
Dependent Variable: SBS_post

Mean 95% Confidence Interval for

Difference (I- Differenceb
(I) Group (J) Group J) Std. Error Sig.b Lower Bound Upper Bound
Control Treatment -.427* .119 .000 -.662 -.192
Treatment Control .427* .119 .000 .192 .662

92
Based on estimated marginal means
*. The mean difference is significant at the .05 level.
b. Adjustment for multiple comparisons: Bonferroni.

The third Table 3.16c of tests of Between-Subjects Effects show whether the difference
between Treatment and Control group is significant. From the table it can be noticed that the F-
value for Group variable is equal to 12.782, which is significant by the level of 0.05. This means
that groups are significantly different from one another. Particularly, the means for Treatment
and Control groups are respectively equal to 2.073 and 1.646. The latter correspond to the level
of 41.5% and 32.9% respectively in percentage terms. It is important also to look at the effect
size, which gives an indication about the magnitude of the effect, so it shows that around 5% of
any variance of SBS_post is explained by the group of individual, in other words changing from
control to treatment would possibly impact SBS of individual by 5%, which is low. From this
table it is also noticed that the covariate SBS_pre is not significant, the effect size of pretest on
posttest is also miserable. The next table to look at, is the estimated marginal means by each
group, which shows that based on the influence of pretest the means are adjusted. As the Group
variable is significant it is interesting to do a post hoc tests to study some details related to each
group (Table 3.16d).

The Table 3.16e above shows the Bonferroni post hoc by presenting pairwise
comparisons. Particularly, it shows that the statistically significant difference between the means
of Treatment and control groups is equal to 0.427. Thus, by controlling the pretest based on the
analysis it can be concluded that financial literacy intervention in form of workshops had
significant effect on the saving decisions of the participants resulting more positive saving
behavior among individuals from the Treatment group comparing to the individuals from the
Control group.

In terms of savings it is important to mention some details related to behavioral

transformation appeared based on the data analysis. Figure 3.13 below shows the change in the
savings of individuals in treatment and control groups.

93
 Figure 3.13 The impact of financial literacy intervention on the share of savings in
Treatment and Control Groups after 6 months

30.0%
26.8%
25.0%
21.0%
20.0% 17.0% 16.4%
15.0%

10.0%

5.0%

0.0%
Treatment Control
Pre Post

From Figure 3.13, it is obvious that the relative number of people, who had savings after
6 months of the workshop increased from 21% to 26.8% in the Treatment group, when in
Control group the level of saving almost stayed the same, not considering a small decrease. In
addition, the issue of saving in form of deposits is still actual, before 6.8% of individuals
participated in financial literacy workshops mentioned to use deposits, however the percentage
dropped to 4.3% after 6 months. In contrast, the rate of deposits in control villages dropped much
more comparing to the treatment group. Respectively the percentage change in treatment and
control villages are 37.5% and 85.7% decrease in both cases. It is challenging to interpret this
situation, as there are no major economic or social event, which could potentially serve as a
reason for such changes. Figure 3.14 illustrates the situation in case of deposits.

The mentioned behavioral aspects of saving also confirm the results from ANCOVA,
which stated that Treatment group shows significantly positive saving behavior than control
group (Table 3.17).

94
 Figure 3.14 The impact of financial literacy intervention on the share of deposits in
Treatment and Control Groups after 6 months

8.0%
6.8%
6.0%
4.3% 4.5%
4.0%

2.0%
0.6%
0.0%
Treatment Control
Pre Post

Table 3.17 ANCOVA output for the DMS in case of Treatment and Control Groups

a) Descriptive Statistics
Dependent Variable: DMS_post
Group Mean Std. Deviation N
Control 3.433 1.4379 157
Treatment 3.726 1.3685 117
Total 3.558 1.4137 274

b) Levene's Test of Equality of Error Variancesa

Dependent Variable: DMS_post
F df1 df2 Sig.
.098 1 272 .754
Tests the null hypothesis that the error variance of
the dependent variable is equal across groups.
a. Design: Intercept + DMS_pre + Group

c) Tests of Between-Subjects Effects

Dependent Variable: DMS_post

95
 Type III Partial
Sum of Mean Eta Noncent. Observed
Source Squares df Square F Sig. Squared Parameter Powerb
Corrected
24.709a 2 12.355 6.428 .002 .045 12.856 .901
Model
Intercept 391.339 1 391.339 203.612 .000 .429 203.612 1.000
DMS_pre 18.939 1 18.939 9.854 .002 .035 9.854 .879
Group 4.257 1 4.257 2.215 .138 .008 2.215 .317
Error 520.856 271 1.922
Total 4015.000 274
Corrected
545.566 273
Total
a. R Squared = .045 (Adjusted R Squared = .038)
b. Computed using alpha = .05

d) Estimates
Dependent Variable: DMS_post
95% Confidence Interval
Group Mean Std. Error Lower Bound Upper Bound
Control 3.450a .111 3.232 3.669
Treatment 3.703a .128 3.450 3.956
a. Covariates appearing in the model are evaluated at the following values: DMS_pre = 3.208.

e) Pairwise Comparisons
Dependent Variable: DMS_post

Mean 95% Confidence Interval for

Difference (I- Differencea
(I) Group (J) Group J) Std. Error Sig.a Lower Bound Upper Bound
Control Treatment -.253 .170 .138 -.587 .082
Treatment Control .253 .170 .138 -.082 .587
Based on estimated marginal means
a. Adjustment for multiple comparisons: Bonferroni.

96
 In case of debt management behavior from the descriptive statistics of the output it can be
noticed that the mean for Treatment and Control groups is respectively equal to 3.726 and 3.433,
which shows that the Treatment group poses relatively better debt management behavior,
however the difference is not statistically significant, as the Group variable in the model is
significant at the level of 0.138, which is greater than 0.05. Unlike the case of SBS, the pretest
value for DMS in this case is significant at the level of 0.02 and has a partial eta square equal to
0.035. In this case as well the significance level of Levene's Test shows that the assumption of
Equality of Error Variances are not violated.

The Bonferroni post hoc shows that the statistically significant difference between the
means of Treatment and control groups is equal to 0.253 after the adjustment of group means for
the Treatment and Control (3.703 and 3.45 out of 5). Thus, by controlling the pretest based on
the analysis it can be concluded that financial literacy intervention in form of workshops had
some positive effect on debt management behavior of individuals from the villages in Treatment
group in comparison to the individuals from Control group villages. However, it is important to
mention again that the difference between the groups is not significant.

While reviewing the behavioral aspects related to debt management, it can be noticed
from the data that in Treatment group the share of people, who took credit increased by 3.4
percentage points, unlike the control group, where the share of people owning a credit decreased
from 57.3% to 43.9%. Figure 3.15, demonstrates the situation in case of both treatment and
control groups.

Figure 3.15 The impact of financial literacy intervention on the share of credit in
Treatment and Control Groups after 6 months

70.0%
57.3%
60.0%
45.3% 48.7%
50.0% 43.9%
40.0%
30.0%
20.0%
10.0%
0.0%
Treatment Control
Pre Post

97
 However, it is unclear the causes for such situation related to the share of credit among
individuals in treatment and control group, one of the reasons might be that after financial
literacy intervention people feel more confident about their skills of managing credit obligations.
As well as they might count on their knowledge of personal rights protection. The described
situation can be noticed from the data of people, who reported to have debts.

Figure 3.16 The impact of financial literacy intervention on the share of debt in Treatment
and Control Groups after 6 months

74.0% 73.0%
72.0%
70.0%
68.0% 66.4%
65.8%
66.0%
64.0%
62.0% 61.1%
60.0%
58.0%
56.0%
54.0%
Treatment Control
Pre Post

Figure 3.16 shows the similar pattern in case of debt of consumers, the share of people
having debts increased in treatment group, while it decreased in the control group, respectively
by 4.7 and 6.6 percentage points.

It is interesting to show the change in behavior related to the credit guarantee. After 6
months, the data shows that the obligations of consumers as loan cosigner decreased on average
by almost 24% in both groups. Even though there is no big difference between treatment and
control groups, it is important to mention that after 6 months the share of people having cosigner
obligations decreased and respectively is 16.2% and 21.7% in treatment and control groups
(Figure 3.17).

98
 Figure 3.17 The impact of financial literacy intervention on the share of cosigners in
Treatment and Control Groups after 6 months

40.0% 36.5%
35.0% 32.5%
30.0%
25.0% 21.7%
20.0% 16.2%
15.0%
10.0%
5.0%
0.0%
Treatment Control
Pre Post

Finally, from the nonequivalent group design quasi-experimental study analysis it can be
concluded that financial literacy intervention in 12 villages of Armenia significantly affected the
saving behavior of participants, in case of debt management the positive change is noticeable,
but it is not significant.

3.5 Variations of results by socio-demographics

The variations by socio-demographics are presented based on the pretest and posttest data
both from treatment and control groups. The main categories, which are analyzed, are gender,
age, education, the level of family income and the share of personal income. The latter is
assumed to indicate the power to affect the financial decisions in the family. While analyzing the
variations by the mentioned characteristics the following aspects are considered:

1. How financial literacy and financial behaviors of individuals both in treatment and
control groups differ based on their socio-demographic characteristics.
2. How differently financial behavior are affected by the financial literacy intervention
among the participants from the treatment group of based on their socio-demographic
characteristics.

Based on the survey results from 274 individuals in 24 villages of Armenia, it turned out
that there are some differences between female and male individuals. Particularly, Figure 3.18

99
shows the differences in terms of the level of financial literacy, trust, saving and debt
management behavior by gender. Men are showing better results in saving behavior and
financial literacy, when women are better at debt management. The level of trust among both
men and women is similar around 60%. The difference in saving behavior is not big either, the
SBS for men is 34.5% and for women 31%. However, the difference between financial literacy
by gender is equal to 6% in favor of male respondents. The major difference can be noticed in
case of the behavior related to debt management. Women demonstrate 8% higher DMS in
comparison to men.

Figure 3.18 The differences between individuals’ financial literacy, trust, saving and debt
management behavior by gender

4
67.5%
3.5 59.8% 62.2% 60%
3 56%
49.9%
2.5
2 31% 34.5%
1.5
1
0.5
0
Female Male Female Male Female Male Female Male
Saving Behaviour Debt Management Trust Financial Literacy
Index Index Score

The long-term impact of financial literacy intervention in terms of saving and debt
management behavior are different by gender. Figure 3.19 demonstrates the change in saving
and debt management behavior among men and women after 6 months of financial literacy
workshops. It can be noticed that in terms of debt management, the intervention impacted male
more than female, particularly the change in DMS for male and female are respectively 17.3%
and 9.8%. In contrast, the opposite pattern can be noticed in case of SBS. The financial literacy
impacted the saving behavior of women by increasing SBS by 35.7%, unlike women SBS for
men increase only by 4.6%. As a result, after the financial literacy workshops after 6 months the
SBS both for male and female is equal to 41%. Such behavioral shift might relate to higher
awareness about benefits of formal saving and Deposit Guarantee Fund as well as relatively
higher trust towards financial system.
100
 Figure 3.19 The change in saving and debt management behavior after financial literacy
intervention by gender

40.0% 35.7%

30.0%

20.0% 17.3%
9.8%
10.0% 4.6%

0.0%
Female Male
Change in Debt Management Index Change in Saving Behaviour Index

Related to the age groups of the research it is interesting to notice different trends among
different age groups related to financial literacy, trust, saving and debt management behavior.
Figure 3.20, shows the results for each age group.

Figure 3.20 The differences between individuals’ financial literacy, trust, saving and debt
management behavior by age groups

80.0% 72.3%
68.6%
70.0% 64.5% 64.2% 62.9% 63.3%
60.0%
60.0% 56.7%
53.3% 52.0% 51.4%
48.9%
50.0%
35.6% 37.1%
40.0% 31.4% 32.2%
30.0%
20.0%
10.0%
0.0%
18-24 25-44 45-64 65> 18-24 25-44 45-64 65> 18-24 25-44 45-64 65> 18-24 25-44 45-64 65>
Saving Behaviour Index Debt Management Index Trust Financial Literacy Score

In case of Saving Behavior, it might be noticed the decreasing trend at the beginning then
after middle ages the SBS is starting to grow, which also matches with the consumption patter of
life cycle of individuals. However, the result goes counter the principle of economic models as it
is more optimal for people to start saving once a person has a sustainable income source. The

101
SBS for 18-24 age group is 35.6%, which decreases on average by 3-4 percentage points for the
age groups of 25-44 and 45-64, then reaches to 37.1% in case of people older than 65. The DMS
is increasing parallel to the age from 60% to 68.6%, which suggests that people get more
experienced over time and manage their debts better. Trust is relatively high among younger
people, which is assumed to be connected with the absence of negative experience related to
financial system in the country. The lowest level of trust showed people between the ages of 45
to 64. In terms of financial literacy there is no big difference among the groups, the data suggests
that people between the ages of 25-44 have higher financial literacy 53.3%, which might be
connected with the education and experience they had after the independence of Armenia in
1991.

Figure 3.21 The change in saving and debt management behavior after financial literacy
intervention by age groups

60.0% 56.2%
50.0%
38.5%
40.0% 34.4%
30.0% 23.3%
20.0% 16.2%
10.0% 6.9%

0.0%
18-24 25-44 45-64
Change in Debt Management Index Change in Saving Behaviour Index

When the impact of financial literacy workshops on age groups is discussed (Figure 20),
it can be noticed that the big change on saving and debt management behavior experienced the
age group 18-24, which is the young generation and they are more interested in learning and
applying new financial instruments and using banking services. The SBS and DMS for the
participants between the ages of 18-24, are respectively changed by 38.5% and 56.2%, unlike
this age group relatively low results showed participants between the ages of 25-44 for SBS and
DMS respectively 23.3% and 6.9%. The age groups of people below 18 and older than 65, is not
discussed, as the number of these people in treatment group is not representative.

102
 In terms of education, the data analysis shows that there is no clear trend related to the
saving and debt management behavior of people depending on the level of education they
received; however, in case of trust and financial literacy the data shows that there is a positive
relationship. Parallel to the level of education, the financial literacy score increases from 40% for
people with no education to 55% for people with higher education, similarly the level of trust
increases from 40% to 62.4%.

In case of the level of family income the pretest data shows that there is a positive
relationship across some measures; however, this relationship is not obvious, as can be noticed in
Figure 3.22. The saving behavior changes from 30.3% till 35.6% for the people, who reported to
have family income respectively $101-$200 and $401-$700. The people with higher family
income than $700 a month are not discussed, due to the low frequency in the sample.

Figure 3.22 The differences between individuals’ financial literacy, trust, saving and debt
management behavior by family income levels

80.0%
70.0% 64.0% 65.9% 62.9% 67.8% 65.2%
58.1% 58.8% 57.8%
60.0% 52.8% 54.7% 56.1%
50.0% 41.3%
40.0% 32.0% 30.3% 32.2% 35.6%
30.0%
20.0%
10.0%
0.0%
< $100

< $100

< $100

< $100
$101-$200

$201-$400

$401-$700

$101-$200

$201-$400

$401-$700

$101-$200

$201-$400

$401-$700

$101-$200

$201-$400

$401-$700
Saving Behaviour Index Debt Management Index Trust Financial Literacy Score

Related to the debt management behavior, the data does not show any distinct differences
in the line of changes in family income. The similar results are presented in case of trust.
Financial literacy score similar to SBS, shows positive relationship. For the people, who reported
to have low level of family income the financial literacy score is only 41.3%, while for the
people with higher family income the score is 56.1%. It case of financial literacy it is interesting
to notice the sharp change between first and second income groups by 11.5 percentage points,
while afterwards the level of score is not changing sharply. The impact of financial literacy
workshops on people with different levels of family income is presented in Figure 3.23. The

103
results are shown for the groups which have family income up to $400, as a small number of
participants to the workshops reported higher family income.

The positive changes in both saving and debt management behavior of the participants
can be reported for all three groups discussed. Particularly, the SBS changed by 40.3% for the
participants from the group with $101-$200 monthly family income, which is bigger than the
change in other two groups. Related to the debt management, the bigger change is reported in the
lowest income group (17.1%).

Figure 3.23 The change in saving and debt management behavior after financial literacy
intervention by family income levels

45.0% 40.3%
40.0%
35.0%
30.0% 25.0%
25.0% 20.3%
20.0% 17.1%
15.0%
9.1%
10.0% 6.2%
5.0%
0.0%
< $100 $101-$200 $201-$400
Change in Debt Management Index Change in Saving Behaviour Index

In case of the personal income proportion in the family income, there is no distinct
difference in terms of saving behavior across individuals who reported to have minimal and
maximal contribution to the family budget, as the SBS is around 30% for all levels of personal
contribution to the family income.

In case of Debt management, the situation is interesting, the people who reported a low
level of contribution to the family budget turned having higher DMS, particularly the people who
reported that their contribution is between 1%-25% in the family budget, the DMS is 73.8%,
when people who reported that the whole family income is earned by them, DMS is 54%. This
phenomenon basically shows that people who have greater impact on financial decisions in the
family poses relatively low level of responsible debt management behavior. It is also interesting

104
to notice that the level of trust has a positive relationship with the level of personal contribution
to the family income.

Thus, in conclusion, the research shows that financial literacy impacts the financial
decisions of individuals differently depending on their social and demographic characteristics.

105
CONCLUSION AND RECOMMENDATIONS

The analysis and evaluations done in frame of this study are mainly addressing the
relationship between financial literacy and financial decisions of consumers. Particularly, the
following research questions are aimed to be addressed in frame of this research:

1. To what extent increase in financial literacy of consumers living in Armenia impact

their financial decisions related to saving and debt management?
2. To what extent financial literacy is related to financial preferences of consumers?
3. To what extent financial literacy intervention in form of workshop is efficient method
to impact financial knowledge accumulation in the short-term and financial behavior
in the long-term (after 6 months)?
4. To what extent socio-demographic differences affect the relationship between
financial literacy and financial decisions of consumers?

For this purpose, both primary and secondary research are implemented by using various
statistical and methodological approaches. To address the main research question of the
dissertation a quasi-experimental study is conducted. In addition, non-experimental pre-post
design is used to address the short-term effectiveness of financial literacy intervention in form of
workshops, descriptive analysis is done across socio-demographic characteristics and binary
logistic regression model is built to describe the relationship between financial literacy and
financial preferences of consumers. From the analysis and evaluation carried out in frame of this
research, the following conclusions can be stated:

 Financial literacy significantly affects the financial decisions of consumers regarding

saving. The statistical analysis done in frame of quasi-experimental study confirms that
there is statistically significant difference between the saving behavior of individuals,
who received financial literacy workshop. The Analysis of Covariance shows that the
mean of SBS in treatment group is equal to 41.5%, which is greater than the mean of
control group by 8.6 percentage points. This suggests that individuals, who participated
in financial literacy workshop, started making financial decisions in favor of saving.
Particularly, the data shows that after 6 months the saving rate among these people
increased from 21% to 26.8%, while in the control group the saving rate stayed almost
unchanged. In addition, the number of people, who reported to have a deposit in the bank
decreased during these months, however the share of deposits in treatment group
decreased only by 37.5%, while in control group it decreased by 85.7%. The significant

106
 impact of financial literacy on saving decisions of consumers might also be explained by
the strong relationship between financial literacy and saving preferences of consumers.
The binary logistic regression analysis shows that financial literacy is one of the
important determinants of saving preference of individuals. Particularly, the odds ratio
for financial literacy score shows that the probability of success is equal to 1.81, which
means that one-unit increase of financial literacy score, will change the individual’s
preference towards formal ways of saving with the probability of 64.5%.
 The impact of financial literacy on debt management decisions of consumers is not
significant. The data analysis from treatment and control groups show that there is no
statistically significant difference between the individuals of study. The Analysis of
Covariance show that the mean of DMS between the groups differ by 5% only in favor to
the treatment group. However, then separate debt management related aspects are
analyzed and some interesting patters can be brought out. Particularly, the data shows
that after 6 months the share of individuals, who reported to have a debt and credit
increased in the treatment group respectively by 7.7% and 7.5%, while in the control
group the opposite happened, the share of people having debt and credit decreased by 9%
and 23.4%. This pattern might be assumed to be connected with increased confidence
regarding debt management among the individuals from the treatment group, but this
assumption should be tested. Regarding decisions of consumers related to acquiring debt
related obligations in form of co-singing there is a positive change among individuals
from the treatment group. There is a sharp decrease in the share of individuals, who
reported to be a guarantor for somebody to take a credit, both in treatment and control
groups, while in treatment group the decrease is bigger by 10 percentage points. The
binary logistic regression analysis aimed to explore the relationship of financial literacy
and debt preferences also brought out non-significant results, outlining that there is no
link between financial literacy and debt management.
 Financial literacy intervention in form of workshop proved to be efficient in the short-
run. Based on non-experimental pre-post evaluation results, it is confirmed that there is a
distinctive change in financial literacy as well as financial preferences of participants
towards formal ways of saving and taking debt. Particularly, there is on average 37%
increase in financial knowledge and skills of participants (49%->67%). A substantial
progress can be noticed in knowledge and skills related to the concepts of inflation,
diversification, deposit guarantee and APRC. Saving preferences has been changed from
informal ways of saving to formal by 54 percentage points. Regarding debt preferences,
107
 the pre-post evaluation results shows that the preference to take debt from friends and
relatives decreased from 36% to 17%.
 The impact of financial literacy intervention in form of workshop on the long-term is
controversial. The impact of financial literacy intervention after 6 months, showed a
significant impact on saving behavior and not significant impact on debt management
behavior. In addition, the considering the methodological limitations and the issues
associated with nonequivalent group design, it is expected that there are more efficient
ways to impact financial decisions of consumers in the low-income countries considering
also the fact that different interventions should be designed for promoting different
financial behaviors and among different target groups. One size does not fit all.
 Trust towards financial institutions is an important predetermining factor affecting the
relationship of financial literacy and saving preferences of consumers. In this aspect,
the binary logistic analysis shows that trust is significantly affecting the preference of
consumers towards formal ways of saving. One unit change in the level of trust will lead
to change in saving preferences towards deposits with the probability of 38.6%. The data
analysis shows also that in the short-term the trust towards the financial system increased
by 11 percentage points on average among the participants of the workshop. Particularly,
the trust toward Financial System Mediator increase by 21 percentage points. In the long-
term, it is interesting to see that trust decreased among the individuals from control group
by 6.2%, while in treatment group the level of trust stayed almost the same.
 Male and Female demonstrate different decision making patterns regarding savings
and debt management in relation to financial literacy. The descriptive data analysis
show that men tend to have better saving behavior than women, though the difference is
only 3.5%. In addition, in terms of debt management, the intervention impacted male
more than female, particularly the change in DMS for male and female are respectively
17.3% and 9.8%. In contrast, the opposite patter can be noticed in case of saving behavior
of women by increasing SBS by 35.7%. SBS for men increased only by 4.6%.
 Different age groups display different patters related to financial literacy, financial
behavior and trust towards financial system. In case of saving behavior, the score
decreases at the beginning and starts growing after middle ages. The SBS for 18-24 age
group is 35.6%, which decreases on average by 3-4 percentage points for the age groups
of 25-44 and 45-64, then reaches to 37.1% in case of people older than 65. The DMS is
increasing parallel to the age from 60% to 68.6%, which suggests that people get more

108
 experienced during the life and manage their debts better. Trust is relatively high among
younger people, which is assumed to be connected with the absence of negative
experience related to financial system in the country. The lowest level of trust showed
people between the ages of 45 to 64. In terms of financial literacy there is no big
difference among the groups, the data suggests that people between the ages of 25-44
have higher financial literacy 53.3%.
 The influence of financial literacy intervention on saving and debt management
behavior of individuals by the level of family income is refrained. SBS changes from
30.3% till 35.6% for the people, who reported to have family income respectively $101-
$200 and $401-$700. For the people, who reported to have low level of family income
the financial literacy score is only 41.3%, while for the people with higher family income
the score is 56.1%. Related to the debt management behavior, the data does not show any
distinct differences in the line of changes in family income. The similar results are
presented in case of trust.

Exploring the behavioral patterns of consumers in relation to financial literacy is

especially important in low-income countries considering different social and economic issues,
such as low personal income, financial exclusion, not well developed financial institutions, poor
developed stock markets, dependence from developed countries. Thus, considering the main
concluding points presented above, there is a need to suggest a comprehensive set of
recommendations, which will help policymakers on national and international level to build
effective and targeted policies in low-income countries.

Particularly, considering the relationships and connections brought out in frame of this
research, the policymakers put more efforts in developing and designing targeted strategic
actions across different target groups. In frame of these strategic activities much effort should be
used on designing the channels and methods of financial education, which will ensure significant
impact on financial behavior of consumers, thus increasing the quality of financial decision
making process. Financial literacy intervention in form of workshop proved to have positive
impact on saving decisions of consumers and no significant impact on debt management
behavior. In addition, the impact is different across various socio-demographic groups. The latter
confirms the issue that it is barely possible to design a universal approach of financial literacy
intervention, which will positively impact all dimensions of financial decisions across all target
groups. In contrast, separate interventions should be designed for each target group focusing on

109
concrete financial behaviors. It might be efficient also develop a core intervention and modify it
based on different characteristics.

Traditionally, financial service providers and policymakers tried to foster healthy

financial behaviors among population by providing in-depth financial education classes or
seminars that cover complex financial management and accounting concepts. Similarly, national
financial education strategies have focused efforts on broad-based, consumer education with a
particular bias towards general information and knowledge improvements. Unfortunately, there
is little evidence that indicated these interventions to be successful in improving knowledge and
causing behavioral change, which is somewhat showed in this research too. Fernandes et al.
(2014) studied the raised issue in deep manner and concluded with some discussion on financial
education and regulation practices to help consumer financial decision making.

In addition to the development of behaviorally informed financial education

interventions, the policymakers should also consider regulatory and supervisory aspects, which
will support to the development of environment. Particularly, proper regulations on consumer
rights protection should be developed related to information disclosure, advertisement and
publications by financial institutions about consumer loan conditions and other financial
services. Apart from process related regulations, major reforms in the area of insurance
(especially health and agricultural) and retirement planning should be initiated by the
policymakers, which might be an important long-term perspective to increase the well-being of
people.

In frame of this study it is important to highlight that there are some methodological
limitations, which if solved could bring to more significant results. The experimental research
would have more statistical power and possible effect if it is possible to engage more locality in
the research. In this case the control and treatment groups would have been bigger thus research
findings would be more valid. In addition, the bigger number of respondents, would positively
affect binary logistic regression analysis as well. In addition, as already discussed in
methodological limitations, the primary research would have been more valid, if treatment and
control groups are assigned randomly and significantly not different by their population means.
In this case, the possible effect of financial literacy intervention would be more vivid at the end
of research. However, perfect randomization is really difficult to arrange, especially out of lab,
while dealing with real environment. In this case it is also challenging to gather participants for
the financial literacy workshops, as many people think that the opportunity cost of workshop
participation is very high, so they prefer to spend time on other things. Even though the

110
mentioned limitations are addressed in the study while applying the statistical methods,
policymakers might invest financial and human resources and receive better indications by
replicating this study.

111
REFERENCES

Al-Tamimi, H. A. H., & Kalli, A. A. B. (2009). Financial literacy and investment decisions of
UAE investors. The Journal of Risk Finance, 10(5), 500-516.

Allen, M. W., Edwards, R., Hayhoe, C. R., & Leach, L. (2007). Imagined interaction, attitudes
towards money and credit, and family coalitions. Journal of Family and Economic Issues,
28, 3-22.

Almenberg, J., & Säve-Söderbergh, J. (2011). Financial literacy and retirement planning in
Sweden. Journal of Pension Economics & Finance, 10(4), 585-598.

Altintas, K. M. (2011). The dynamics of financial literacy within the framework of personal
finance: An analysis among Turkish university students. African Journal of Business
Management, 5(26), 10483-10491.

Atkinson, A., & Messy, F.-A. (2012). Measuring Financial Literacy (OECD Working Papers on
Finance, Insurance and Private Pensions). Paris: Organization for Economic Co-operation
and Development. Retrieved from http://www.oecd-
ilibrary.org/content/workingpaper/5k9csfs90fr4-en

Atkinson, A., McKay, S., Collard, S., & Kempson, E. (2007). Levels of financial capability in
the UK. Public Money and Management, 27(1), 29-36.

Banco de Portugal (2010). Survey on the financial literacy on the Portuguese population.
Retrieved: March 15, 2012, from http://clientebancario.bportugal.pt.

Bernheim, B. D., Garrett, D. M., & Maki, D. M. (2001). Education and saving: The long-term
effects of high school financial curriculum mandates. Journal of Public Economics, 80(3),
435–465. https://doi.org/10.1016/S0047-2727(00)00120-1

Bucher-Koenen, T., & Lusardi, A. (2011). Financial literacy and retirement planning in
Germany. Journal of Pension Economics & Finance, 10(4), 565-584.

Buckland, J. (2010). Are low-income Canadians financially literate? Financial literacy in the
context of personal and structural constrains. Adult Education Quarterly, 60(4), 357-376.

Calvet, L., Campbell, J., & Sodini, P. (2007). Down or out: assessing the welfare costs of
household investment mistakes. Journal of Political Economy, 115, 707-747.

Carpentier, C. & Suret, J. (2012). “Financial knowledge and rationality of Canadian investors”.
Working Paper from http://ssrn.com/abstract=2038930.

112
Central Bank of Armenia. (2014, November). National Strategy of Financial Education in
Armenia. Central Bank of Armenia.

Central Bank of Armenia. (2015). Financial Capability Barometer 2014. Yerevan, Armenia.

Central Bank of Armenia. (2016). Financial Stability Report. The Central Bank of the Republic
of Armenia. Retrieved from https://www.cba.am/EN/pperiodicals/Fin_stab_1_16_eng.pdf

Chen, H., & Volpe, R. P. (2002). Gender differences in personal financial literacy among college
students. Financial Services Review, 11(3), 289-307.

Christelis, D., Jappelli, T., & Padula, M. (2010). Cognitive abilities and portfolio choice.
European Economic Review, 54, 18-38.

Clarke, M. D., Heaton, M. B., Israelsen, C. L., & Eggett, D. L. (2005). The acquisition of family
financial roles and responsibilities. Family and Consumer Sciences Research Journal, 33,
321- 340.

Colosi, L., & Dunifon, R. (2006). What’s the difference: “Post then Pre” & “Pre then Post.”
Retrieved on August, 15, 2009.

Comissão do Mercado de Valores Mobiliários – CMVM (2009). “O perfil do investidor

particular Português”, Estudos CMVM, 3, p. 1-19. Retrieved: March 24, 2012 from
http://www.cmvm.pt/
CMVM/Estudos/Em%20Arquivo/Documents/EstudoCMVM0320093.pdf.

Consumer Financial Protection Bureau. (2015). Financial well-being: The goal of financial
education. Retrieved from
http://files.consumerfinance.gov/f/201501_cfpb_report_financial-well-being.pdf

Cook, T. D. (1979). Quasi-Experimentation: Design & Analysis Issues for Field Settings (New
Ed edition). Boston: Houghton Mifflin.

Crossan, D., Feslier, D., & Hurnard, R. (2011). Financial literacy and retirement planning in
New Zealand. Journal of Pension Economics & Finance, 10(4), 619-635.

DeBondt, W. (1998). A portrait of individual investor. European Economic Review, 42(3-5),

831-844.

Education Policy and Data Center. (2014). National Education Profile: 2014 update. Retrieved
November 4, 2016, from
http://www.epdc.org/sites/default/files/documents/EPDC%20NEP_Armenia.pdf

113
Fernandes, D., Lynch Jr, J. G., & Netemeyer, R. G. (2014). Financial literacy, financial
education, and downstream financial behaviors. Management Science, 60(8), 1861–1883.

Financial Services Authority (2004), Building financial capability in the UK, London: Financial
Services Authority.

Fornero, E., & Monticone, C. (2011). Financial literacy and pension plan participation in Italy.
Journal of Pension Economics & Finance, 10(4), 547-564.

Fredholm, J., & Taghavi-Awal, B. (2006). Capital markets in developing countries: A model for
capital market diagnostics, with a field study implementation in Georgia. Retrieved from
http://www.diva-portal.org/smash/record.jsf?pid=diva2:196476

Hastings, J., & Mitchell, O. S. (2011). “How financial literacy and impatience shape retirement
wealth and investment behaviors”. NBER Working Paper No. 16740.

Hastings, J., Mitchell, O. S., & Chyn, E. (2010). “Fees, framing, and financial literacy in the
choice of pension managers”. Pension Research Council Working Paper, Wharton
School.

Henriques, S. M. (2010). “Aspetos da literacia financeira dos Portugueses: um estudo empírico”,

Master Thesis, Aveiro, DC: University of Aveiro, Portugal.

Hilgert, M. A., Hogarth, J. M., & Beverly, S. G. (2003). Household Financial Management: The
Connection between Knowledge and Behavior. Federal Reserve Bulletin, 89(7), 309-322.

HM Treasury. (2007). Financial capability: The Government’s long-term approach. London: HM

Treasury.

Hoare, S. (2003). Making it all. Financial World, 32-37.

Hosmer, D. W. J., Lemeshow, S., & Sturdivant, R. X. (2013). Applied Logistic Regression (3
edition). Hoboken, New Jersey: Wiley.

Hung, A., Parker, A. M., & Yoong, J. (2009). Defining and Measuring Financial Literacy.
RAND Working Paper Series WR-708. Retrieved from
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1498674

Hunter, L., Orr, A., & White, B. (2006). Towards a framework for promoting financial stability
in New Zealand. Reserve Bank of New Zealand Bulletin, 69(1), 5-17.

Huston, S. J. (2010). Measuring Financial Literacy. Journal of Consumer Affairs, 44(2), 296–
316. https://doi.org/10.1111/j.1745-6606.2010.01170.x

114
Jorgensen, B. L., & Savla, J. (2010). Financial literacy of young adults: The importance of
parental socialization. Family Relations, 59, 465-478.

Jump$tart Coalition for Personal Financial Literacy. (2015). National Standards in K-12 Personal
Finance Education. Washington, DC.

Kempson, E., Collard, S., & Moore, N. (2005, June). Measuring financial capability: an
exploratory study. Financial Services Authority. Retrieved from
http://www.bristol.ac.uk/geography/research/pfrc/themes/fincap/measuring-fincap.html

Klapper, L. F., Lusardi, A. & Panos, G. A. (2012). “Financial literacy and the financial crisis”.
Working Paper from http://ssrn.com/abstract=2013864.

Korniotis, G. M. & Kumar, A. (2011). Do older investors make better investment decisions? The
Review of Economics and Statistics, 93(1), 244-265.

Lee, N. (2012). What is financial literacy, and does financial literacy education achieve its
objectives? Evidence from banks, government agencies and financial literacy educators in
England. Academy of Financial Services Annual Conference. San Antonio.

Lusardi, A. (2008a). Financial Literacy: An Essential Tool for Informed Consumer Choice?
(Working Paper No. 14084). National Bureau of Economic Research. Retrieved from
http://www.nber.org/papers/w14084

Lusardi, A. (2008b). Household Saving Behavior: The Role of Financial Literacy, Information,
and Financial Education Programs (Working Paper No. 13824). National Bureau of
Economic Research. Retrieved from http://www.nber.org/papers/w13824

Lusardi, A., & Mitchell, O. S. (2011a). Financial Literacy around the World: An Overview
(Working Paper No. 17107). National Bureau of Economic Research. Retrieved from
http://www.nber.org/papers/w17107

Lusardi, A., & Mitchell, O. S. (2011b). Financial literacy and retirement planning in the United
States. Journal of Pension Economics & Finance, 10(4), 509-525.

Lusardi, A., & Mitchell, O. S. (2013). The Economic Importance of Financial Literacy: Theory
and Evidence (Working Paper No. 18952). National Bureau of Economic Research.
Retrieved from http://www.nber.org/papers/w18952

Lusardi, A., & Mitchelli, O. (2007). Financial literacy and retirement preparedness: Evidence
and implications for financial education. Business Economics, 42(1), 35–44.

115
Lusardi, A., & Tufano, P. (2009). Debt literacy, financial experiences, and overindebtedness.
National Bureau of Economic Research. Retrieved from
http://www.nber.org/papers/w14808

Lyons, A. C., & Scherpf, E. (2004). Moving from unbanked to banked: Evidence from the
Money Smart program. Financial Services Review, 3(13), 215–231.

Mandell, L. (2008). Financial Literacy of High School Students. In J. J. Xiao (Ed.), Handbook of
Consumer Finance Research (pp. 163–183). Springer New York. Retrieved from
http://link.springer.com/chapter/10.1007/978-0-387-75734-6_10

Mason, C. L., & Wilson, R. (2000). Conceptualising financial literacy. © Loughborough

University.

Moore, D. L. (2003). Survey of financial literacy in Washington State: Knowledge, behavior,

attitudes, and experiences. Washington State Department of Financial Institutions.

Morris, L. (2001). An investigation into the financial knowledge levels of New Zealand senior
secondary school students. Wellington: Enterprise New Zealand Trust.

Nelder, J. A., & Wedderburn, R. M. (1972). Generalized linear models. Journal of the Royal
Statistical Society, Series A 135(3), 370-384.

Noctor, Michael, Sheila Stoney, and Robert Stradling. "Financial literacy: A discussion of
concepts and competences of financial literacy and opportunities for its introduction into
young people’s learning." National Foundation for Educational Research (1992).

O’Connell, A. (2009). Evaluating the Effectiveness of Financial Education Programmes. OECD

Journal: General Papers, 2008(3), 9–51. https://doi.org/10.1787/gen_papers-v2008-art17-
en

OECD. (2005). Improving Financial Literacy Analysis of Issues and Policies: Analysis of Issues
and Policies. OECD Publishing.

OECD. (2012). Measuring Financial Literacy: Results of the OECD / International Network on
Financial Education (INFE) Pilot Study. OECD Working Papers on Finance, Insurance
and Private Pensions No. 15.

OECD. (2013a). Advancing National Strategies for Financial Education: A Joint Publication by
Russia’s G20 Presidency and the OECD. Retrieved from

116
 https://www.oecd.org/finance/financial-
education/G20_OECD_NSFinancialEducation.pdf

OECD. (2013b). PISA 2012 Assessment and Analytical Framework. OECD Publishing.
Retrieved from http://www.oecd-ilibrary.org/education/pisa-2012-assessment-and-
analytical-framework_9789264190511-en

OECD/INFE. (2009). Financial Education and the Crisis: Policy Paper and Guidance.

OECD/INFE. (2012). OECD/INFE High-Level Principles on National Strategies for Financial

Education. Retrieved from http://www.oecd.org/finance/financial-
education/OECD_INFE_High_Level_Principles_National_Strategies_Financial_Educati
on_APEC.pdf

OECD/INFE. (2013). Current status of national strategies for financial education: OECD/INFE
Comparative Analysis and Relevant Practices.

OECD/INFE. (2013). The role of financial education in financial inclusion. OECD/INFE

evidence, policies and illustrative case studies. Retrieved from
https://www.microfinancegateway.org/sites/default/files/mfg-en-paper-the-role-of-
financial-education-in-financial-inclusion-oecdinfe-evidence-policies-and-illustrative-
case-studies-jun-2013.pdf

OECD/INFE. (2015). National Strategies for Financial Education: OECD/INFE Policy

Handbook. Retrieved from http://www.oecd.org/daf/fin/financial-education/National-
Strategies-Financial-Education-Policy-Handbook.pdf

Orton, L. (2007). Financial literacy: Lessons from international experience. Canadian Policy
Research Networks, Incorporated. Retrieved from
http://rcrpp.ca/documents/48647_EN.pdf

Pahnke, L., & Honekamp, I. (2010). Different effects of financial literacy and financial education
in Germany. Retrieved from http://mpra.ub.uni-muenchen.de/22900/

Petrosyan, & Magluchants. (2000). Household Population and Housing Characteristics, 11–23.

Power, M. L., Hobbs, J. M., & Ober, A. (2011). An empirical analysis of the effect of financial
education business students’ perceptions on graduating business students’ perceptions of
their retirement planning familiarity, motivation, and preparedness. Risk Management
and Insurance Review, 14(1), 89-105.

117
Pregibon, D. (1980). Goodness of Link Tests for Generalized Linear Models. Journal of the
Royal Statistical Society. Series C (Applied Statistics), 29(1), 15–14.
https://doi.org/10.2307/2346405

President’s Advisory Council on Financial Literacy (PACFL). (2008). 2008 Annual Report to
the President. Retrieved December 14, 2016, from
http://jumpstart.org/assets/files/PACFL_ANNUAL_REPORT_1-16-09.pdf

Ratha, D., Eigen-Zucchi, C., & Plaza, S. (2016). Migration and Remittances Factbook 2016.
World Bank Publications.

Remund, D. L. (2010). Financial literacy explicated: The case for a clearer definition in an
increasingly complex economy. Journal of Consumer Affairs, 44(2), 276–295.

Rosner, B. (2010). Fundamentals of Biostatistics (7 edition). Boston: Duxbury Press.

Sample size for before-after study (Paired T-test). (2016, January 24). Retrieved from
http://www.sample-size.net/sample-size-study-paired-t-test/

Siegenthaler, J. K., Anderson, C., Lyter, D. M., Kent, J., & Ward, J. (2000). Personal Finance
and the Rush to Competence: Financial Literacy Education in the US. Child & Youth
Finance International.

Stöver, B. (2012). The influence of age on consumption. EcoMod. Retrieved from

http://ecomod.net/system/files/Discussion%20Paper_Britta%20Stoever.pdf

The National Council on Economic Education (NCEE). (2005). What American Teens & Adults
Know About Economics. The National Council on Economic Education (NCEE)).
Retrieved from
http://www.councilforeconed.org/cel/WhatAmericansKnowAboutEconomics_042605-
3.pdf

The World Bank. (2013, August). Why Financial Capability is important and how surveys can
help.

Van Rooij, M. van, Lusardi, A., & Alessie, R. J. (2011a). Financial Literacy, Retirement
Planning, and Household Wealth (Working Paper No. 17339). National Bureau of
Economic Research. Retrieved from http://www.nber.org/papers/w17339

Van Rooij, M. C. J., Lusardi, A., & Alessie, R. J. M. (2011b). Financial literacy and retirement
planning in the Netherlands. Journal of Economic Psychology, 32, 593-608.

118
Van Rooij, M. C. J., Lusardi, A., & Alessie, R. J. M. (2011c). Financial literacy and stock market
participation. Journal of Financial Economics, 101, 449-472.

Wagland, S. P., & Taylor, S. (2009). When it comes to financial literacy, is gender really an
issue? Australasian Accounting Business and Finance Journal, 3(1), 13-25.

Widdowson, D., & Hailwood, K. (2007). Financial literacy and its role in promoting a sound
financial system. Reserve Bank of New Zealand: Bulletin, 70(2), 37–46.

World Bank. (2013, August). Financial Capability Surveys Around the World: Why Financial
Capability is important and how surveys can help. Retrieved from
www.worldbank.org/fpd

World Economic Forum. (2015, December). The Global Financial System: Policy
Recommendations for the Future. Retrieved November 4, 2016, from
http://www3.weforum.org/docs/IP/2016/FS/WEF_AM16_FGFS_TaskForce_PolicyRecs.
pdf

119
APPENDIX 1: RESEARCH QUESTIONAIRES

For the pretest and posttest, the following questions are used as a part of questionnaires.
The pretest questionnaire is filled by the respondents in both Treatment and Control villages
before financial literacy intervention, posttest questionnaire is filled only by the participants of
financial literacy workshop right after it in Treatment villages and it had the same structure as
pretest, however the order of answers and numbers in calculation type questions are changed:

Pretest/Posttest Questionnaire

1. Please mention which of the below mentioned financial services you use now.
1) Bank Account 5) Money transfer
2) Payment Card 6) I am a guarantor/cosigner
3) Loan 7) Other _______________________________
4) Deposit 8) None of them

2. Please mention which is your preferable way to save.

1) Keep money at home
2) Deposit in a bank
3) Buy gold/jewelry
4) Buy a capital (e.g. apartment)
5) Buy governmental bonds
6) Other _____________________________________

3. In your opinion how inflation is calculated?

1) Inflation is calculated based on exchange rate (USD, EUR).
2) Inflation is calculated based on changes of international prices.
3) Inflation is calculated based only on the price changes of flour, sugar and cooking oil.
4) Inflation is calculated based on the changes of prices of widely consumed products
99) I don’t know

120
4. If your yearly income is 1,000,000 dram and inflation yearly rate is 5%, how much you
should earn after a year to keep your living standards unchanged?
1) 1,000,000 drams
2) 1,050,000 drams
3) 950,000 drams
99) I don’t know

5. If you deposit 100,000 drams in a bank for a year with 8% interest rate, how much
money you will have after a year taking into account 10% income tax?
1) More than 108,000 drams
2) 108,000 drams
3) Less than 108,000 drams
99) I don’t know

6. You deposit 100,000 drams in a bank for 5 years with 5% interest rate. How much
money you will have after 5 years if you capitalize interests?
1) 125,000 drams
2) Less than 125,000 drams
3) More than 125,000 drams
99) I don’t know

7. If you have some solid amount of money, will you prefer to save it in one bank or in
several banks?
1) In one bank
2) In several banks
99) I don’t know

8. If you have 8,000,000 drams in one of the banks in Armenia, please mention which part
of it is insured by the Deposit Guarantee Fund?
1) 6,000,000 drams
2) 4,000,000 drams

121
3) 8,000,000 drams
99) I don’t know

9. Please mention which source you will prefer the most if you need to borrow some
money?
1) Banks
2) Credit organization
3) Friends, relatives, neighborhood
4) Local store
5) Other ________________________________

10. How the interest rate which includes all the costs (nominal interest rates, commissions,
service fees etc.) related to the loan is called?
1) Nominal interest rate
2) Annual percentage rate of charge (APRC)
99) I don’t know

11. Imagine you have seen an advertisement by TV which offers to deposit some money
now and receive twice more after three months. What you will do?
1) I will use that opportunity and will advise it to my friends too
2) I will not pay any attention
3) I will check it by depositing only a small amount of money
99) I don’t know

12. For receiving information about financial services which source is more preferable for
you?
1) TV, radio, newspapers
2) Official webpages of financial organizations
3) Branches of financial organizations
4) Friends, relatives, neighborhood
5) Other _____________________________________

122
13. In case of a problem with a financial organization, where will you approach first to
solve it?
1) Financial organization
2) Central Bank of Armenia
3) Financial System Mediator
4) Court
5) Other _____________________________________

14. Please mention how much you trust below mentioned entities?
(1 – not at all, 5 – very much)

1) Central Bank of Armenia 1 2 3 4 5

2) Commercial banks 1 2 3 4 5
3) Credit organizations 1 2 3 4 5
4) Pension funds 1 2 3 4 5
5) Insurance companies 1 2 3 4 5
6) Financial System Mediator 1 2 3 4 5

In addition to these questions, socio-demographic questions are asked as well.

Particularly, the respondents are asked to mention their gender, age, education level,
employment, family income and share of personal income in family income. For the phone
survey the following questionnaire are used after 6 months:

Phone survey questionnaire

1. Please mention which source you will prefer the most if you need to borrow some
money?
1) Banks
2) Credit organization
3) Friends, relatives, neighborhood
4) Local Store

123
5) Other _____________________________________

2. Please mention which of the below mentioned financial services you use now.
1) Bank Account 5) Money transfer
2) Payment Card 6) I am a guarantor/cosigner
3) Loan 7) Other ________________________________
4) Deposit 8) None of them

If in the Question 2 the “Loan” is not mentioned, ask only from Question 3 to 5 and
continue from Question 9.

3. Did you have loans before?

1) Yes
2) No (go to Question 9)
99) I don’t know

4. How many loan offers did you review before taking the loan?
1) I didn’t review any offer 3) Three and more
2) Two 99) I don’t know

5. Please mention the source of the loan you took.

1) Bank 4) Friends, relatives, neighborhood
2) Credit organization 5) Other ______________________
3) Pawn shop 99) I don’t know

If in the Question 2 the “Loan” is mentioned, continue from Question 4.

4. How many loan offers did you review before taking the loan?
1) I didn’t review any offer 3) Three and more
2) Two 99) I don’t know

124
5. Please mention the source of the loan you took.
1) Bank 4) Friends, relatives, neighborhood
2) Credit organization 5) Other ______________________
3) Pawn shop 99) I don’t know

6. Did you have loans before?

1) Yes
2) No (go to Question 9)
99) I don’t know

7. If yes, please evaluate your new loan terms compared to the previous one.
1) Loan rate is more is less is unchanged

2) Loan amount is more is less is unchanged

3) Loan duration is more is less is unchanged

Before it is _______ currency, now is

4) Loan currency is unchanged
_______.

8. Did you use new loan to repay the old one?

1) Yes
2) No
99) I don’t know

9. Did you take any in place loan from a shop during last 6 months?
1) Yes
2) No
99) I don’t know

10. Please mention which is your preferable way to save.

1) Keep money at home
2) Deposit in a bank
3) Buy gold/jewelry

125
4) Buy a capital (e.g. apartment)
5) Buy governmental bonds
6) Other _____________________________________

11. You deposit 100,000 drams in a bank for 5 years with 5% interest rate. How much
money you will have after 5 years if you capitalize interests?
1) 125,000 drams
2) Less than 125,000 drams
3) More than 125,000 drams
99) I don’t know

12. If you have a loan in USD and exchange rate will change from 1 USD = 500 AMD to 1
USD = 400 AMD, how your monthly payments will be affected?
1) Will grow 3) Will not change
2) Will decrease 99) I don’t know

13. Imagine the yearly interest rate of your saving account is 1% and yearly inflation rate
is 2%. After a year how much will you be able to buy with money in your account?
1) More than today 3) Less than today
2) The same 99) I don’t know

14. Please mention which one of the mentioned descriptors fits you the best?
1) I allocate some money for saving, then I spend other part of my income.
2) I spend for current needs, then I save the rest part.
3) I spend all my income for current needs without saving.
4) I spend for everything without planning.
99) I don’t know

15. For receiving information about financial services which source is more preferable for
you?
1) TV, radio, newspapers

126
2) Official webpages of financial organizations
3) Branches of financial organizations
4) Friends, relatives, neighborhood
5) Other _____________________________________

16. In case of a problem with a financial organization, where will you approach first to
solve it?
1) Financial organization
2) Central Bank of Armenia
3) Financial System Mediator
4) Court
5) Other _____________________________________

17. Please mention how much you trust below mentioned entities?
(1 – not at all, 5 – very much)

1) Central Bank of Armenia 1 2 3 4 5

2) Commercial banks 1 2 3 4 5
3) Credit organizations 1 2 3 4 5
4) Pension funds 1 2 3 4 5
5) Insurance companies 1 2 3 4 5
6) Financial System Mediator 1 2 3 4 5

In this case too socio-demographic questions are asked to the respondents. Particularly,
the respondents are asked to mention their gender, age, education level, employment, family
income and share of personal income in family income.

127
APPENDIX 2: TREATMENT AND CONTROL GROUP CHARACTERISTICS

The summary results of independent two-sample T-tests assuming unequal variances for
descriptive variables are presented below:

Treatment Control
Women 67.9% 45.5%
Sex Men 30.6% 54.5%
N/A 1.5% -
< 18 5.7% 1.6%
18-30 23.4% 20.2%
Age 31-60 52.8% 65.6%
60 > 4.9% 9.9%
N/A 13.2% 2.8%
No Edu 0.8% 3.2%
Secondary 1.9% 3.2%
High School 18.5% 25.7%
Vocational 20.0% 31.2%
Education
University 52.8% 36.0%
PhD 0.4% 0.8%
N/A 5.7% -
Financial Edu 21.5% 19.4%
Married 67.9% 77.5%
No married 15.5% 12.3%
Marital Status
Widow 6.0% 4.7%
N/A 10.6% 5.5%
Average number of children under
Number of children 1 1
18
Employee 63.8% 59.7%
Self-employed 0.8% 12.3%
Employment Seasonal worker 9.4% 5.5%
Unemployed 5.3% 6.3%
Student 7.5% 2.8%

128
 Retired 1.9% 4.0%
Household business owner 4.5% 5.9%
Unable to work 0.8% 1.6%
N/A 6.0% 2.0%
< 50000 9.1% 13.4%
50001 – 100000 26.4% 26.9%
100001 – 200000 23.0% 29.2%
Family income
200001 – 350000 10.9% 15.8%
350001 > 5.7% 5.9%
N/A 24.9% 8.7%
< 25% 24.2% 24.1%
26% - 50% 19.2% 29.6%
51% - 75% 15.5% 18.6%
Level of personal income
76% - 99% 3.0% 5.1%
towards family income
100% 2.3% 5.9%
No personal income 8.3% 4.3%
N/A 27.5% 12.3%

129
APPENDIX 3: WORKSHOP DESIGN

The two-day workshop developed as a financial literacy intervention and implemented in

12 treatment villages had the content presented below:

DAY 1
1. Budget management
 The structure of household/personal budget,
 Differences between household/personal and business budget,
 Prioritizing and monitoring budget items,
 Identifying risks affecting household/personal budget and taking respective measures.
2. Bank account & payment cards
 Bank account and payment cards benefits and risks,
 Types of bank account and payment cards,
 Choosing right payment card depending on needs,
 Important aspects to consider, while using payment cards (including online purchases).
3. Savings and long-term planning
 Importance and ways of saving,
 Options of investing the savings, comparison based on benefits and risks,
 Evaluating investment options in regards of access, profitability and security,
 Choosing and compering deposits (APY, diversification, exchange rate risk etc.),
 Deposit Guarantee Fund in Armenia,
 Retirement planning.
DAY 2
4. Debt management
 Consumption versus saving,
 Options of taking debt, comparison based on benefits and risks,
 Rights and responsibilities while signing a credit agreement,
 Choosing and compering credits (APRC, exchange rate risk etc.),
 Credit History and Credit score.
5. Insurance
 Principles and types of insurance,
 Benefits and risks associated with insurance contract,

130
  Important aspects to consider while signing an insurance agreement.
6. Personal Rights Protection
 Rights and responsibilities of consumers,
 Steps to take for personal rights protection,
 Financial System Mediator in Armenia.
7. Financial Security
 Types of financial scams and frauds,
 Identifying potential scams, frauds and taking respective measures,
 Actions to take in case of being harmed by scams and frauds.

131