Essay:

Essay details:

  • Subject area(s): Marketing
  • Price: Free download
  • Published on: 14th September 2019
  • File format: Text
  • Number of pages: 2

Text preview of this essay:

This page is a preview - download the full version of this essay above.

ntroduction

Financial markets become more and more accessible to all society levels due to the extensive growth of financial services and products. The importance of financial literacy has shown a constant increase at various levels. It influences directly individual's welfare through the efficient management of financial assets.  (Satchell et al, 2010) Also financial literacy has the significant effect on economic stability. Overall, the behavior of institutions influences the relocation of recourses that, in the long run, can predict the economical condition.  (Hastings et al, 2013) That's why financial literacy is a topic of concern of many economists and politicians all over the world.

Everyone deals with financial decisions every single day.  For instance, how much to spend and how much to save, where to invest and how to effectively distribute expenditures and debts in order to avoid bankruptcy.

The aim of this work is to provide an appropriate evidence to

 As we can see these decisions vary from the easiest to the more complicated ones, which require some special knowledge from individuals. This knowledge can be acquired either from the process of socialization or from educational institutions where individual interacts during lifetime.

Financial literacy is an important topic not only because it affects individual's life standards but also it influences efficiency of financial system. First of all, in cautious literate society banks and other financial institutions operate under the policies of reduced risk. And as a result more stable investment environment for consumers appears. More innovative and effective financial system is likely to develop in order to satisfy consumer's demand. (банк новой зеландии)

Financial educational programs are directly related to the level of financial literacy and the decisions financial institutions make. Various researches have proven that the overall financial literacy of the earth population is still low. This thesis will not only benefit the current financial institutions, but also will provide a better and more exact numbers to the policymakers nowadays. Such updated information will also help banks in their decision making targeted on specific social classes of population, therefore, creating a better economic environment for the upcoming generations.

Basic concepts about financial literacy and World Financial Crisis

The understanding of financial literacy helps to comprehend the significance of its application in society.  The first step in promoting financial education was made in the United States between 1950s and 1960s, when government began authorizing the inclusion of economics, personal finance, household budgeting and consumer rights and responsibilities disciplines in the 12th grade educational curriculum (Bernheim et al. 2001). However, the evaluation of such policies was only made in 90s also in the United States when researchers didn't aim to measure the financial literacy.  The lack of financial knowledge and skills problem takes its origins from practical-oriented topics such as retirement savings of working class and student's financial behavior (Kuzina, 2015).

There are many definitions of this concept. According to OECD definition, ‘financial literacy is a combination of awareness, knowledge, skill, attitude and behavior, necessary to sound financial decision-making and ultimately achieving individual financial well being' (OECD/INFE, 2016). Two main aspects can be outlined from this statement. First aspect is that personal knowledge of finance and its application in real life. And second one is the relationship between financial knowledge and informed judgment decisions (Samy et al., 2008).

When talking about financial literacy, it is necessary to mention more expansive term- financial capability. The main difference between them is that “financial literacy refers to the knowledge and skills that individuals use to make an appropriate financial decision while financial capability include also behavioral factors, social influences, and emotions” (Jump$tart Coalition, 2009). This statement follows a logical question: why don't we study financial capability instead of financial literacy, as this term is more comprehensive. Financial capability relates to practical goal-reaching actions. To measure someone's financial capability, not only knowledge should be examined but also individuals'actions in their everyday life. At the same time to evaluate financial literacy, individuals' level of knowledge should be distinguished by conducting relevant survey in form of tests or quastionneres. Indeed, individuals' behavior examination causes two types of problems. Firstly, financial capability questions built according to respondents' self-assessments, but their words do not always correspond to their actions in real life. Secondly, it is hard to take the gauge of right and wrong answers. For instance, a respondent can keep a record of his expenses not fully and not on a regular basis or not effectively follow prior made financial plans.

Financial literacy, as it has already mentioned, affects individual's well being through their financial decisions. The behavior of financial companies will be also changed causing allocation of recourses and potential growth of the economy. The rational financial decisions bring to the raise of financial security and hence the standards of living. (Satchell et al, 2010) Due to the lack of financial education individuals can be under threat of bankruptcy or financial fraud. The situation in Russia in mid-90s can be considered as a relevant illustration. Many people invested all their savings into financial pyramids such as “MMM”, “Hoper-Invest”, “Chara Bank” etc. because of highly influencing advertising and marketing campaign. After few profitable outcomes, millions of people lost all their money.

Consequantly, financial literate households tend to manage their debts without any negative consequences. The world financial crisis that happened in 2008 can be a good example of irrational financial behavior and low financial literacy level that were one of the factors that caused a stock market crash and a fall of the economy all over the world.

On the grounds of low interest rates introduced by Fed, banks and investment funds, more loans (mostly with floating interest rate) and risky mortgage products became more frequently approved. Consumer lending became more and more popular and the share of debt to the disposable personal income grew to 143% mostly because of mortgage lending. In the middle of 2004 interest rates on mortgage loans with a floating interest rates started to increase making it difficult for borrowers to refinance their loans. Most worldwide securities reinforced with subprime mortgages dramatically leading to the collapse of trade markets in US and Europe. As a result world economy experienced “the worst financial crisis since the Great Depression of 1930s”. Most people did not understand the conditions of their subprime mortgage contacts. They were obtruded the opinion that their income would grow over the time and they could easily pay off their mortgage debts. Also, financial brokers did not consider the long-run outcome after their loan approval decision, mostly working on the quantity of loans given. It is a good example of how financial illiteracy effects on the world economy that still makes millions of people to suffer.  

Previous studies

The issue of lack of financial literacy was not studied until last 20 years, when researches realized the importance of it, therefore leading to the various studies. The purpose of this review is not only to  

Before analyzing the facts, which affect the financial literacy, it is important to discuss the researches, which have already shown an impact in reaserch of financial literacy. There are a lot of surveys that prove that low level of financial literacy all over the world such as one made by Anna Maria Lusardi and Olivia Mitchell (2011b). They find out that women in the U.S. are significantly less likely to answer questions correctly, and more likely to say that they don't know the answer. However, there were no sex differences in financial literacy in Russia and in Germany – both genders are equally illiterate. According to this research, the strong correlation between higher education and financial literacy was proven. At the same time even respondents with higher education showed low level of financial literacy, but still it is higher among those who are working and in some countries among those who are self-employed contrary to those who does not work. It can be explained by existence of trainings and educational programs for workers offered in United States. Lusardi and Mitchell (2011b) also concluded that people who tend to be more financial literate are more likely to plan retirement contrary to those whose financial literacy indicators were slightly lower. Researchers generalize that financial literacy should not to be taken as granted even in countries with developed financial markets as being strongly correlated with retirement planning in the long-run financial literacy can influence the allocation of resources in the economy.   

Klapper and Panos (2011) considered a case of Russia to discover relationship between financial literacy and retirement planning. Russia has a pervasive public provision system. Consumer borrowing is increasing rapidly there, but it was found that only 36.3 percent of respondents in the sample, which were used, knew about the working of interest compounding and only a half could answer a simple question about inflation. Country with relatively large regional disparities and rapidly emerging markets shows a positive correlation between financial literacy and retirement planning using private planning schemes and funds. At the same time residents from rural areas would less likely invest in private schemes and savings and more likely would rely on the public provision, due to the mass financial illiteracy.

Example of the World Financial Crisis, which was described above, assumes that financial literacy affects debt and mortgage outcomes for individuals. Gerardi et al. (2010) showed a significant and quantitatively large association between one aspect of financial literacy, numerical ability, and mortgage delinquency. Micro data on subprime mortgage terms was combined with individual's entire stream of payments and survey data from telephone interview with a sample of subprime borrowers. It was found that correlation is robust at some level of mortgage delinquency and it includes a lot of socio-economic and demographic control variables. This work focuses on many aspects of financial literacy, but the correlation is high to numerical ability (one of the financial literacy aspects). However, the researcher himself remarked that survey with controlled experiment has always a probability of omitted variable existence due to disability to randomize. Therefore, other factors can be corresponded to financial literacy measure and can cause the variation in mortgage repayment behavior.

Data

This analysis mainly bases on two data sources: financial literacy survey data provided Laboratory for Studies in Economic Sociology (LSES) and the Russian banking system statistic provided by Central Bank of Russia.

Financial literacy survey dated on November 2015 and covered 1,600 respondents from 42 Russian regions. The questions set consist of several parts, such as:

Savings;

Credits and loans;

Banking services;

Insurance, and

Other.

A special part is focused on financial literacy and covers both subjective and objective factors through different types of questions. Respondents were asked about their financial background and further they tried to solve a set of problems. We base on this data in defining the financial literacy factor.

To measure the level of development of a banking sector in a particular region we relied on the Central Bank of Russia official statistic available on cbr.ru.

Methodology

Before we run a regression model, we need to introduce formal definitions for financial literacy and the level of banking sector development.

A term of “financial literacy” consists of two concepts:

The definition of “literacy”, and

Its particular qualities in terms of financial sphere.

According to Zarcadoolas, Pleasant, Greer (2009) , “literacy” means (in general terms) the ability of an individual to read and wright. Using the definition provided by Kirsch, Irwin, Yamamoto, Norris, et al. (2001) , we can summarize that the literacy refers to understanding and use of knowledge to pose (written), document (graphical or tabular) and quantitative (numerical) information.

According to Huston (2010) , this concept of general literacy expanded to some specific skill sets. Callingham and Watson (2004)  used this concept to define statistical literacy: Wecker, Kohnle and Fischer (2007)  – to define computer literacy.

It is important to note that the applied instrument for measuring financial literacy should cover both: knowledge dimension and application dimension. At the same time, application is a consequence of knowledge. Following Huston (2010), we suppose that a successful financial literacy measure should allow researcher to distinguish in case of deficiency in financial literacy is responsible for bad (in terms of welfare) financial decisions and also should allow educator to identify the education level to achieve a desired result. To be financially literate, a particular person has to demonstrate skills and knowledge required to take a decision within a financial environment that all people face regardless of their own characteristics.

After we introduce a formal definition for financial literacy into the model, we should choose an instrument that will be used to measure the level of financial literacy of a particular person. Recent researches show that the optimal instrument is survey question related to financial knowledge and experience, this instrument was applied by Chen and Volpe (1998) , Lusardi and Mitchell (2008) , Lusardi and Mitchell (2011) .

Basing on previous researches, we measured financial literacy through the set of questions related to financial calculations and operations. Full text of questions provided in the Annex A.

Then we calculate financial literacy “score” for each respondent according to the following formula:

FL=(# of correct answers)/7 (3.1)

This assumption arrives to the following financial literacy distribution among respondent.

Figure 3.1: Financial literacy distribution among respondent

Source: LSES, author's calculations

As we can see from the chart above, 95 respondents correctly answered on all questions, at the same time there are 111 respondents, who does not answer on any question. The distribution seams normal and symmetric except first extreme column. Thus, we can treat applied measure for financial literacy as acceptable.

However, the distribution of correct answers varies among questions since some of them are relatively easy and the others are more difficult. Thus we decided to weight each question on values equal to the inverse share of correct answers to solve the problem that easy and complicated questions have the same impact on the financial literacy determination. This approach assumes that if relatively less people answer correctly a particular question than this question is relatively more difficult. The main advantage of this approach can be shown through the example: let us pretend a situation when one respondent answered correctly only question X and another respondent – only question Y. In addition, 99% of respondent answered correctly the question X and only 1% – the question Y. Under standard approach these two respondents will have a similar FL level. However, we can state that question X is relatively less difficult than the question Y, thus the second respondent should have higher FL level that is true under the weighted approach. In that environment, respondents who correctly answered on relatively more sophisticated question will be considered as more financial educated. This arrives to the following distribution of financial literacy level among respondents:

 Figure 3.2: Weighted financial literacy distribution among respondent

Source: LSES, author's calculations

The above distribution is more smoothed and flexible, however it depends of exogenously introduced weights. As a result we obtained two measures of financial literacy level that will be used in further analysis.

Concerning banking system, we used three different variables to measure the development of banking sector in a particular region:

Total number of banks;

Average size of banks;

Number of branches.

When choosing these variables we rely on three key points: quality, flexibility and availability. Quality thesis means that these items are used by researches as a measure for financial system development. For example, Boyd and De Nicolo (2005) showed the relationship between number of banks and total deposit market and the risk of default. Casu and Molyneux (2003) claimed that the average size of banks is substantial in determining the difference between European countries' banking systems.

Flexibility means that applied data can be divided by regions and by applicable time periods, thus we decided to use CBR database as a source for banking system parameters. Speaking in terms of availability, we consider only these variables because other parameters from CBR official statistic (that can be potentially used in our analysis) contain many miss values across all regions.   

All three variables correspond to banking sector size; in addition, average size of bank provides some qualitative information about banking system. We consider two time points: October 2015 and October 2016.

As mentioned above the main aim of this study is to determine how banking system development affects the financial literacy level. However, there is an endogeniety problem, that there are many other factors that affect both financial literacy and banking system. Taking into account defined variables, two approaches are applicable to test the relationship between financial literacy and banking system parameters.

First, we operates with differences when construction an independent variable that is we calculate the change in total number of banks/branches or average size change. This allows us to take into account the scale of a particular region by adding a lag of dependent variable (see Anatoliev (2003) for details). Also, when we use differences we exclude individual effects , which are not interesting in our case.

Second approach directly assumes that financial literacy in a particular region depends on many factors other than banking sector development, such as population, regional GDP etc. Under this approach we directly add these factors into the model, instead including these factors through previous values of independent variable.

Both described methods have their pros and cons , thus we are going to apply them as compliments.

The Model

After we introduce the applied methodology let's proceeds to the models description. At the first stage, we analyze the relationship between average financial literacy in region and the banking sector variables. The figure below illustrates this relationship and also provides information about data scatter.

Figure 3.3: Financial literacy impact

 

Source: LSES, CBR, author's calculations

As we can see, there are many regions which experienced no change in the total number of banks in 2016; however there are some outliers such as Moscow and St. Petersburg. It should be noted that the Figure 3.3 does not take into account the number of respondents in each region and shows a simple average financial literacy level among respondents in a region.

After visual analysis, we construct a correlation matrix to check weather financial literacy significantly affected by these banking items. We treat each respondent as a single observation. This allows us to take into account the number of respondents in each region, since we are interesting in financial literacy and banking system relationship across all Russia. We give a higher weight to regions that are represented by larger number of respondents, this allow us adjusting results toward the whole Russian level.

We denote “diff” operator as y_2016-y_2015, where y is one of the banking sector indicators listed above. The results of estimations presented below.

Table 3.1: Correlation coefficients, using the observations 1 – 1305

diff_bank diff_branch diff_size FL

1.0000 0.8712 -0.2690 -0.0824 diff_bank

1.0000 -0.0754 -0.0809 diff_branch

1.0000 0.1396 diff_size

1.0000 FL

5% critical value (two-tailed) = 0.0543 for n = 1305

Source: author's calculations

Table 3.1 shows that relatively more financial literal regions  have experienced higher reduction in total number of banks and branches relative to less literal ones. Contrariwise, financial literacy level has positive impact on average size of banks. This result is in line with Levine (1999) , who showed that countries with large value of banks do not necessarily have a developed banking system.

However, the situation changes if we will consider relative differences in financial sector parameters. Relative differences do not include the scale of a region and we can see that under this assumption the total number of banks and branches positively influence financial literacy (see Table 3.2). Nevertheless the accuracy of calculation is relatively low , thus we cannot present a statistically significant result and proof or reject any hypothesis. So, for convenience let us denote “rel_diff” operator as  (y_2016-y_2015)/y_2015 , where y is one of the banking sector indicators listed above. The results of estimations presented below.

Table 3.2: Correlation coefficients, using the observations 1 - 1305

rel_diff_bank rel_diff_branch rel_diff_size FL

1.0000 0.2375 0.1059 0.0845 rel_diff_bank

1.0000 0.0486 0.0637 rel_diff_branch

1.0000 0.1118 rel_diff_size

1.0000 FL

5% critical value (two-tailed) = 0.0543 for n = 1305

Source: author's calculations

According to provided calculations, the sanitation of banking system through decreasing the number of relatively small banks arrives to an increase in financial literacy. The one of the possible explanations of such relationship is that nowadays the Russian Central Bank systematically liquidate “bad” banks and people begin to pay more attention to investing in incomprehensible banks. Thus, people start thinking about their financial activities more deeply: compare options, immerse themselves into banking products, etc. As a result, the financial literacy of people in average increases.

However, there might be another explanation of this relationship. Under the current Russian CB policy of tightening regulations many relatively small banks cannot withstand the regulator's pressure, thus number of small banks reduces. At the same time, the Russian government pursuing a policy of financial literacy improvement that arrived to the increase of financial literacy level. These changes are simultaneous and collinear, but might be independent.

Let us assume that financial literacy can be described by the equation below:

FL=β*X+γ*у+ε (3.1)

Where X corresponds to all independent variables that affect financial literacy level in a particular region, and y represents banking sector component. However, as we noted above, financial literacy also directly depends on the level of banking sector development and can be represented by the following way:

FL=f(Z,Y(y)) (3.2)

Where Z represents all factors that define financial literacy level, but do not belong to the banking sector (e.g. age, sex, income etc.). By the other hand, Y(y) corresponds to the direct influence of banking system on financial literacy level. Taking into account the equation (3.1) we cannot directly put y into the equation (3.2), because there would be endogeneity problem during an estimation process. Thus we used a proxy variables Y(y), the list of them is presented below.

Table 3.3: Variables description

Variable name Description

FS1 Consumer credit (excluding credit cards)

FS2 Mortgage

FS3 Car loan

FS4 Plastic сard for salaries, pensions, fellowship payments

FS5 Debit card ordered by yourself)

FS6 Credit card ordered by yourself)

FS7 Term deposit

FS8 Current accounts, deposits \"on demand\"

FS9 Investment services (mutual funds, fund management, asset management)

FS10 Voluntary insurance

FS11 Services of private pension funds

FS12 Deposit box

Source: constructed by the author

Now we should provide some qualities of the function f(.) from the equation (3.2). Intuitively, this function has to fit the following properties:

∀ Z,Y  f∈[0;1]

lim┬(Z→+∞)⁡1;  lim┬(Z→-∞)⁡0; f_Z^\'≥0

lim┬(Y→+∞)⁡1;  lim┬(Y→-∞)⁡0; f_Y^\'≥0

Additionally, to check the assumption stated above, we should proceed to the additional stage of analysis: we are going to divide the total financial literacy of a respondent by two parts:

Obtained literacy – application dimension (see Fig. 3.4);

“Default” literacy – knowledge dimension (see Fig. 3.4).

As we stated above, financial literacy includes knowledge dimension and application dimension.

Figure 3.4: The structure of financial literacy

Source: Huston (2010)

This separation helps to take it into account through respondent's experience in banking sector. We assume that people improve their financial knowledge when they use banking services. At this stage, we clearly introduce a variable Y that equals:

the average number of services that respondent has ever used in banking sector ;

zero, if respondent has never used banking services.

Taking into account all listed above, we will introduce an additional restriction for the function f(.) type, that will make the further analysis more comfortable. So, function f(.) has to fit the following property.

FL=f(Z,Y)=f_1 (Z)+f_2 (Y)=FL_obt+FL_def

To separate two types of literacy we used the following logit-type function:

FL=β*e^(γ_1 Y)/(1+e^(γ_1 Y) )+(1-β)*e^(γ_2 Z)/(1+e^(γ_2 Z) )+ε (3.3)

Where:

Y – the set of parameters that belongs to obtained literacy;

Z – the set of parameters that belongs to “default”  literacy;

β – the weight of obtained literacy in the model that bounded by 0 below and by 1 above;

ε – normally distributed random variable N(0,σ^2 ).

It allows estimating the share of financial literacy that can be explained by previous experience in banking sector. Here we cannot use simple GLS model since we wanted to get financial literacy separation each part of which is greater than 0 and bounded above .

We fit the model by the maximum likelihood estimation (MLE). The MLE is such method of estimating the parameters that finds the parameter values that maximize the likelihood of making the observations given the historical values of parameters.

Due to inability to use simple statistical packages to obtain results by the MLE method, in this case we use the Python programing language that is a widely used and high-level by its nature for general-purpose programming. In particular, we constructed the log-likelihood function and applied the COBYLA minimizing method that minimizes a function of one or more variables using the Constrained Optimization BY Linear Approximation (COBYLA) algorithm.

As a main result of the model, we show that people who have used credits, deposits and other banking products are more literate by circa 40% than the others are. By the other hand, there is an endogeniety problem, thus we cannot identify the direction of the effect at this stage. However, if we consider FS4 variable (“Plastic сard for salaries, pensions, fellowship payments”) as a proxy we may address this issue, since plastic card for salaries, pensions and fellowship are usually given out to owners forcibly – it does not depend on their level of financial literacy.

Table 3.4': Obtained and default literacy estimation

Variable fin_lit

Const 0.41***

(0.01)

FS2 0.05*

(0.03)

FS3 0.07***

(0.02)

FS4 0.08***

(0.01)

FS5 0.07***

(0.02)

FS7 0.06***

(0.02)

FS8 0.04*

(0.02)

FS10 0.15***

(0.02)

FS12 0.10*

(0.05)

Model GLS

# of observation 1,600

Adjusted R-squared 0.087

Source: author's calculations

From the table above, we can see that the usage of banking services positively affect the FL level of respondents. However this corresponds to a direct impact of banking system on the financial literacy, while we are interesting in indirect impact component.

Also at that stage, we introduce modern approach to identify the most important factor in the model. In particular, for that purpose we use Random Forest ensemble learning method that we use for classification the financial literacy parameter. This method constructs a multitude of decision trees and give us the class that is the mode of the classes of the individual trees or the mean prediction (regression) depending on the type of main parameter (in our case depending on choose of FL or FLw). In addition, one of the important features is that random decision forests correct for habit of decision trees of overfitting to their training set.

To do such estimation we also use Python programing language framework with sklearn package that allows us to use all variety of possible modern instruments.

At first, we tried to use random forest regression framework in assessing parameters that we are interested in. The regression framework, as was said before, works with the floating parameters such as FLw, but unfortunately such method leads to poor results:

by testing our model by cross validation technique that used for getting the information how the results of our analysis will generalize to an independent data set and for estimating how accurately a predictive model will perform in practice, we obtained a score of the model that equals to 0.01 – it means that our model provides the similar result as a constant model that predict for all parameters zero value;

moreover, we compute the R^2 parameter and get the value of 0.08 on the test dataset, which is higher by around 8 times than simple linear regression model, but not so high as we will see further.

After the random forest regression model we apply the random forest classifier technique for the discrete parameter FL that belongs to the set A={i/7 } where i∈[0,…,7] . This model gives us the following results:

By the same testing by cross validation technique we obtained a score of the model that equals to 0.2. It means that our model provides much more applicable results;

Moreover, we compute the R^2 parameter and get the value of 0.91 on the overall dataset, which is higher by around 90 times than simple linear regression model and is really close to the maximum possible value of 1.

At the graph below, we demonstrate the separation of FL variable by obtained and “default” literacy parameters:

Figure 3.5: Financial literacy separation results

Source: author's calculations

As we can see from the graph, the obtained literacy has a stable effect on overall financial literacy, whereas the initial literacy is more volatile parameter. It can be explained by the fact that the overall financial literacy depends more on the initial factors such as sex, age and income and not highly correlated with the banking services experience.

Finally, the Random Forest Classifier model gives us the importance of features on a classification task. Below we introduce the assessing importance of all used variables:

Table 3.4: The assessing importance of variables

Feature name Importance

age 0.38

income 0.37

fin_services 0.11

sex 0.05

city 0.04

educ 0.04

Source: author's calculations

As we can see from the table above the most important features are age and income – “default” literacy features. However, the feature “financial services” that represents obtained financial literacy is also highly important in overall financial literacy assessing.

After clarifying the mechanism of separation, we run regressions to identify the indirect impact of banking sector on the level of default literacy.

FL_def=const+β*((y_2016-y_2015 ))/y_2015   +ε (3.4)

Where:

FL_def – are fitted values of expression (1-β)  e^(γ_2 Z)/(1+e^(γ_2 Z) ) from the previous model;

ε – normally distributed random variable N(0,σ^2 ).

The results of estimations presented below.

Table 3.5: 2 step GLS model estimation (the Model 3.4)

Variable rel_diff_bank rel_diff_size rel_diff_branch

Const -0.30*** 0.30*** 0.30***

(0.00) (0.00) (0.01)

FL_def -0.01 -0.01 -0.04

(0.01) (0.01) (0.05)

Model 2 step GLS 2 step GLS 2 step GLS

# of observations 1,256 1,256 1,305

Adjusted R-squared -0.001 -0.001 0.000

Source: author's calculations

As we can see from the table above, financial literacy has lost its importance in number of banks and branches. However, obtained financial literacy has significant positive impact on the average size of bank. At the same time, default financial literacy level has no statistically significant effect on any variable that describe banking sector development. The explanation of this result can be the following: financially literate people do not use services of incomprehensible small banks. Hence, the number of banks will decrease if people become more financially literate. But this result also leads to an increase in the size of other more good banks.

It should be noted, that the model 3.4 has negligible prediction power and low quality results. R^2 adjusted is still significantly low (below 1%). It can be explained by several reasons:

First, the number of observation is artificially increased by considering each respondent as an observation instead of using regions as observation.

Second, there are other factors that influence default financial literacy in a particular region besides banking sector.

Third, taking into account, the economy concentration in Russia we can predict the existence of some outliers in the sample (such as Moscow, St. Petersburg). This inference supported by the summary statistic presented below.

Table 3.6: Summary statistics for banking sector variables

Variable Mean Median Minimum Maximum

bank2015 15.10 4.00 0.00 407.00

bank2016 12.52 4.00 0.00 329.00

branch2015 646.07 487.50 44.00 3,725.00

branch2016 594.38 468.00 36.00 3,340.00

size2015 883.92 396.25 0.00 5,500.00

size2016 898.19 400.00 0.00 5,500.00

Variable Std. Dev. C.V. Skewness Ex. kurtosis

bank2015 62.27 4.13 6.15 36.19

bank2016 50.34 4.02 6.13 36.03

branch2015 640.03 0.99 2.93 10.99

branch2016 573.63 0.97 2.86 10.68

size2015 1,094.23 1.24 2.22 5.88

size2016 1,144.25 1.27 2.08 4.72

Source: author's calculations

From the table above we can see that used data is very volatile:

Mean differs from median by significant value;

C. V. for all variables is almost 1 or above;

There are enormous differences between minimum and maximum values.

In further part of this work, we are going to solve all three problems with estimation. The following three steps will allow us to escape from mentioned drawbacks and improve the quality of analysis.

First, we will consider each region as a unique observation (instead considering any respondents). This arrives to significant reduction in the number of observation. To get the financial literacy level in each region we will use a simple average. In addition, we are not going to exclude calculation by respondent from the analysis to access possible non-representativity issues.

Second, as mentioned above, default financial literacy in a particular region depends on many factors other than number of banks or the average size of bank. We used official statistic from Federal State Statistic Service of Russian Federation  to get an additional data. The applied dataset include the following variables, which can be considered as factors influenced default financial literacy level:

Pop – number of people in a region as of 1 January 2016 (thsd ppl).

Empl – average number of employed people for the year ended 1 January 2016 (thsd ppl).

Income – average monthly income per head in a region (RUB).

Expenses – average monthly consumer expenses per head in a region (RUB).

Salary – average monthly salary in a region (RUB).

GDP – gross regional product (mln RUB).

Trade – retail trade turnover in a region (mln RUB).

Profit – financial result of all companies in a region (profit or losses) (mln RUB).

Invest – total investments into equity capital in a region (mln RUB).

...(download the rest of the essay above)

About this essay:

This essay was submitted to us by a student in order to help you with your studies.

If you use part of this page in your own work, you need to provide a citation, as follows:

Essay Sauce, . Available from:< https://www.essaysauce.com/essays/marketing/2017-6-12-1497287959.php > [Accessed 16.10.19].