Findings, Data Presentation, Analysis and Interpretation

This chapter presents the findings, data analysis, and interpretation of the study. It includes empirical studies, questionnaires return rate, results from the pilot, demographic information, and tax fairness and compliance from a general viewpoint.

CHAPTER FOUR: FINDINGS, DATA PRESENTATION, ANALYSIS AND
INTERPRETATION
4.1 Introduction
The researcher has effectively presented the data analysis, findings and discussions for all
the objectives in this chapter. Data was analyzed using several methods. The study used
descriptive methods (means and standard deviations), Bartlett’s sphericity test, KMO
measure and the factor analysis.
4.1.1 Empirical studies
In an Australian cross-cultural study by Gilligan and Richardson (Gilligan and Richardson,
2005), it was found that perceptions of tax fairness is very crucial for tax complying among
the citizens. The respondents were business students. Fairness and legitimacy were found to
be related and legitimacy perception was seen as a significant influence in relation to
compliance to tax responsibilities. The current study sought the relationship of fairness
perceptions on compliance to tax obligations.
In a comparative cross-cultural study of Malaysia and New Zealand, Natrah (2012) studied
the tax compliance as affected by tax fairness perceptions. Results from partial least squares
and t-tests suggested that New Zealand tax payers were more tax compliant than the
Malaysian citizens. This was despite the fact that the Malaysian respondents had more
significantly better tax fairness perceptions regarding their tax systems. The researcher
found that tax compliance was affected by attitudes that individuals had regarding
compliance but not the fairness itself.
1

4.2 Questionnaires that were Returned
The return rate is shown in table 4.1 below.
Table 4.1 Percentage of Questionnaire Return Rate
Category Issued
questionnaires
Returned
questionnaires
Return rate
Respondents 350 332 94.57%
Total 350 332 94.57%
The Table 4.1 shows that from the 350 respondents, all the respondents did not comply. This
reduced the questionnaires by 19 since the returned questionnaires were 332. Consequently,
the return rate was 94.57. Although not a perfect return rate, this was a good return rate that
would ensure reliable and adequate information from the filled questioners.
4.3 Results from the Pilot
The researcher conducted a pilot study so as to ascertain the how reliable the research
instruments were. The researcher sought to identify and address any ambiguities that the
questionnaires might present. Any questions in the questioners that were hard to be
understood were modified so as to ensure that the respondents were responding to items that
were clear to them. Other items that were removed from the questionnaires are the items that
were irrelevant. The Table 4.2 has the results of the pilot study.
Table 4.2 Reliability Test Results
Scale Cronbachs alpha Items
Respondents .791 33
The resulting reliability coefficient (reliability index) was above 0.78 (0.791) which is
universally accepted as reliable.
2

4.4 Demographic Information
The study gathered the respondents’ demographic information. It was believed that the
demographic factors had a role that they played in the perception in regard to tax compliance
in the area of study. The table 4.3 has the statistics for the demographic information.
Table 4.3: Statistics on Demographic Information
gender age level of education Occupation marital status
N
Valid 332 332 332 332 332
Missing 0 0 0 0 0
Mean 0 1.00 2.30 2.20 2.00
Std. Deviation .483 1.155 1.337 1.317 .943
Sum 332 332 332 332 332
The table 4.3 shows the demographic variables with their means and their standard
deviations. For the gender, 0 was taken to represent female and 1 male. It therefore shows
that there were more females than men who participated in this study. The standard
deviation was 0.483. On age, the age bracket 20-29 was assigned 0, age bracket 30-39 was
assigned 1, age bracket 40-49 was assigned 2 while those who were 50 years and above
were assigned 3. Since the mean for age was 1, it means that those who were in the age
bracket 30- 39 represented the highest number. The standard deviation was not large since it
did not deviate much from the threshold one (1.15), meaning there were no outliers and the
data had a normal distribution. Therefore, the respondents were mainly people who were in
their middle age who are generally settled down by that age. On the other hand, the
researcher assigned values to responses for the level of education. Those who were
secondary professionals were assigned 0, primary professionals were assigned 1, post
graduate were assigned 2, graduate were assigned 3, and doctorate were assigned 4. Level of
3

education had a 2.3 mean together with a standard deviation of 1.3. The mean shows that
most of the people were post-graduates. Since the standard deviation was 1.3, there is no
likelihood of outliers in the data. This is a normal distribution. Occupation was also sought
and the results are indiative that the mean was 2.2 and a standard deviation of 1.3, which
was low. The students were assigned the value 0, pensioner 1, self-employed was assigned
2, employed was assigned 3, and unemployed was assigned 4. This shows most of the
people were self employed. Lastly, the researcher assigned values for marital status. Single
was assigned 0, single with kids was assigned 1, married was assigned 2, while married with
kids was assigned 3. The mean was 2 with a standard deviation of 0.9. This shows that most
of the respondents were married. The standard deviation shows that a large part of the
study’s data was most of the data was gathered together near the mean.
4.5. Tax Fairness and Tax Compliance from a General View point
4.5.1 General fairness
Descriptive statistics were run to test the general perception of tax fairness. Five lickert scale
responses on ratings for the perception on tax based on statements (0 was taken as a
representative of strongly disagree, 1 was taken as a representative of disagree, 2 was taken
as a representative of neutral, 3 was taken as a representative of agree and 4 was taken as a
4

representative of strongly agree) gives the results. The scores of agreeing and strongly
agreeing were taken to represent a mean of 3.0 to 4. On the other hand, the scores of neutral
were taken to represent a mean of 2.0 to 2.9. Lastly, the score of strongly disagreeing and
disagreeing were taken to represent a mean of 0 to 1.9. The results are tabulated in table 4.4
below.
Table 4.4 Statistics on Descriptives
N Minimum Maximum Sum Mean Std. Deviation
The tax burden cannot
be referred as fairly
distributed
332 0 4 607 1.91 1.308
The tax law that is
prevailing in this
country has a
requirement that I pay
more than more in taxes
that fair share
332 0 4 777 2.47 1.215
There are big breaks
that are provided by the
tax and this covers
people who do not
deserve to be covered
332 0 4 785 2.47 1.267
It is only a few who are
under the tax provisions
law
332 0 4 806 2.56 1.251
Some legal tax that can
be described as being
perfectly deduction is
unfair since apart from
the wealthy people, not
many others can use
them
332 0 4 866 2.75 1.175
5

Those who earn high
income have a highly
greater ability to meet
their tax obligations
332 0 4 849 2.70 1.280
It would be best if those
who earn much paid
more in taxes
332 0 4 782 2.50 1.345
A fair tax is the one that
does not discriminate 332 0 4 587 1.86 1.424
Tax that middle
incomes earners pay is
obviously too high
332 0 4 795 2.54 1.235
A comparison with
other tax payers, I
clearly pay less
332 0 4 565 1.79 1.233
In comparison with
what wealthier people
pay, I clearly pay more
332 0 4 661 2.10 1.140
The tax amount that I
pay is too high in an
unreasonable manner
when it is gauged
against the benefits that
are provided
332 0 4 827 2.62 1.125
The results show that on the statement touching on the fair distribution of the burden of tax,
the mean was 1.91 with a small standard deviation. This means that a greater percentage of
those who responded to the items in the questionnaires disagreed with the posed statement.
Most respondents were neutral (M=2.47) on the statement that touched on “the tax law that
is prevailing in this country has a requirement that I pay more than more in taxes that fair
share”. Again, most respondents were neutral (M=2.47) on the statement “The tax system
6

provides big breaks for people that do not deserve”. Still, most respondents were neutral
(M=2.56) on the statement “There are big breaks that are provided by the tax and this covers
people who do not deserve to be covered”. Still, most respondents were neutral (M=2.75) on
the stamen “It is only a few who are under the tax provisions law”. On the statement “It
would be best if those who earn much paid more in taxes”, most respondents were neutral
(M=2.7). on the statement “It is fair that high income earners pay more tax”, most
respondents were neutral (M=2.5). On the statement “A fair tax is the one that does not
discriminate”, most respondents disagree (M=1.86). Still, on the statement “Tax that middle
incomes earners pay is obviously too high” most respondents were neutral (M=2.54). On the
statement “Compared to other taxpayers, I pay less”, most respondents disagreed (M=1.79).
On the statement “In comparison with what wealthier people pay, I clearly pay more” most
respondents were neutral (M=2.1). Lastly, on the statement “The tax amount that I pay is too
high in an unreasonable manner when it is gauged against the benefits that are provided
(M=2.62).
In order to be able to reduce the large number of variables into a few manageable variables,
it was needful to run factor analysis. This process identified latent constructs. Table 4.5
shows the KMO and the Bartletts Test.
Table 4.5: KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure for Adequacy of Sampling .634
Bartlett's Test of
Sphericity
Approx. Chi-Square 532.495
df 91
Sig. .000
7

As seen in table 4.5, the KMO value is 0.634, an indication that the data is suited for use in
factor analysis. The sampling is adequate. Researchers accept KMO that is greater than 0.5,
although Kaiser referred to 0.6 as mediocre. Bartlett's test of sphericity results show that the
sig value was .000 with 91 degrees of freedom. Since the value is more than 0.50, the factor
analysis results will be useful. Since the purpose of Bartlett's test of sphericity is meant to
check on the null hypothesis, the variables in this study are related. This means that the
variables fit structure detection. The null hypothesis that perceived tax fairness does not
affect tax compliance is rejected. Following is the scree plot.
Figure 4.1 Scree plot
8

From the scree plot, the elbow is around 6. This means that the maximum number of
variables that can be taken for analysis is 6 and there would be no value added in taking any
more than that number. Following is the communalities.
Table 4.6 Communalities
Initial Extraction
there is fair tax
distribution 1.000 .911
tax law requires more 1.000 .817
undeserved big tax
breaks 1.000 .784
tax provision applies to
few 1.000 .764
unfair tax deduction 1.000 .789
high income earners can
pay more tax 1.000 .381
The communalities table explains what variance of the variables can be attributed to
influence of the factors. In the extraction column, five of the factors have values greater
than 0.5 meaning that they are high, hence good extractions.
Below is a table on the total variance that was explained.
Table 4.7 Total Variance Explained
Component Initial Eigenvalues Extraction Sums of Squared Loadings
9

Total % of
Variance
Cumulative
%
Total % of
Variance
Cumulative
%
1 3.033 50.545 50.545 3.033 50.545 50.545
2 1.413 23.558 74.103 1.413 23.558 74.103
3 .923 15.376 89.480
4 .333 5.555 95.035
5 .282 4.696 99.731
6 .016 .269 100.000
The table 4.7 shows that 74.103% of the variance that is in the model and is explained by
the variables. This is a high percentage.
The findings are congruent with the literature that perceived tax fairness affects tax
compliance. As explained by the slippery slope in the literature, the relationship of tax
payers in Lithuania with the government is indeed affected by the perception that the
taxpayer has of the government. Kirchler (2008) posited that fairness is an important
determinant of tax compliance and these results have established that this is the case of
Lithuania. The results are also congruent with the argument posited by Jha (1957) in the
literature that governments of yesteryears considered the welfare of the people who would
bear the burden of the taxes before imposing the taxes on them.
4.5.2 Self interest effects in exchange with the government on tax compliance
Descriptive statistics were run to test the effects that self interest has in exchange with the
government on tax compliance. Five lickert scale responses on ratings for self interest in
exchange with the government on tax based on statements (0 representing strongly disagree,
1 representing disagree, 2 representing neutral, 3 representing agree and 4 representing
10

strongly agree) gives the results. The scores of agreeing and strongly agreeing were taken to
represent a mean of 3.0 to 4. On the other hand, the scores for borderline (neutral) were
taken to represent a 2.0 mean to a 2.9 mean. Lastly, the score of strongly disagreeing and
disagreeing were taken to represent a mean of 0 to 1.9. Descriptive results are tabulated in
table 4.8 below.
Table 4.8 Descriptive Statistics
N Minimum Maximum Sum Mean Std. Deviation
Exchanging or trading
services as well as
goods with a neighbor
of a friend and then
failing to report in your
tax returns is not bad at
all
332 0 4 550 1.71 1.001
Income reporting
without adequately
including any particular
extra small amounts of
extra income is okay
332 0 4 607 1.47 1.301
When someone gets
paid for a job and then
they fail to report in
their tax report, is
apparently okay
332 0 4 709 1.00 1.002
Failure to report some
income emanating from
investment or interest is
okay if the government
will to be able to
realize or discover
332 0 4 780 1.97 1.000
11

Making some small
additions than actual
spending is okay and
never a great deal
332 0 4 566 1.93 1.116
Non-taxed purchases
from shadow traders to
their possible limits to
reduce costs is okay
332 0 4 849 1.30 1.110
Making an little
underpayment is not big
deal
332 0 4 502 0.02 1.307
The rates of tax are very
high and therefore we
are justified to cheat as
it cannot be regarded as
cheating as it makes
you pay less
332 0 4 699 2.86 1.424
It obviously makes a
reasonably good sense
to reduce your tax when
a chance avails itself
332 0 4 805 2.38 1.322
Ensuring that you cut
some corners when
paying our tax is okay
332 0 4 765 1.45 1.203
It is indeed right to hold
back some little tax 332 0 4 761 0.10 1.150
If you find a way t
make a deduction, it is
okay since the
government will never
allow you to make it
332 0 4 927 1.62 1.105
Table 4.8 shows the means for all the variables in this test together with their standard
deviations. All the standard deviations are not high, suggesting that this is a normal
12

distribution that does not have any outliers. On the statement “Exchanging or trading
services as well as goods with a neighbor of a friend and then failing to report in your tax
returns is not bad at all”, the respondents disagreed (M=1.71). On the statement “Income
reporting without adequately including any particular extra small amounts of extra income is
okay”, the respondents disagreed (M=1.47). On the statement, “Being paid cash for a job
and then not reporting it in tax report seems fine”, the respondents disagreed (M=1.0). on the
statement. On the statement “Making some small additions than actual spending is okay and
never a great deal”, the respondents disagreed (M=1.30). On the statement “Making an little
underpayment is not big deal” the respondents strongly disagreed (M=0.02). On the
statement “The rates of tax are very high and therefore we are justified to cheat as it cannot
be regarded as cheating as it makes you pay less” the respondents were neutral (M=2.86).
On the statement “Failure to report some income emanating from investment or interest is
okay if the government will to be able to realize or discover” the mean was 1.97 meaning
they disagreed. On the statement “It obviously makes a reasonably good sense to reduce
your tax when a chance avails itself” the respondents were neutral (M=2.38). On the
statement “Ensuring that you cut some corners when paying our tax is okay” the respondents
disagreed (M=1.45). On the statement “It is indeed right to hold back some little tax” the
respondents strongly disagreed (M=0.10). On the statement “If you find a way t make a
deduction, it is okay since the government will never allow you to make it” the respondents
disagreed (M=1.62). Table 4.9 shows the a logistic regression output.
Table 4.9 Information on model fitting
Model Model fitting
criteria
Likelihood ratio tests
Intercept only -2 likelihood Chi-square Df Sig.
365.736 49.21 6 .000
13

332.641
The table 4.9 shows that chi-square value for the likelihood ratio is 49.21. it has a p-value
<0.0001. the significance of this is that the model fits and it cannot be like an empty model
since it has predictors.
Table 4.10 Pseudo r square
Cox and snell .153
Nagelkerke .174
McFaden .079
From table 4.11, it is seen that the model fits well in the data.
B Std. Error Wald df Sig.
I pay taxes
Intercept
Trading services
Inadequate
reporting
Making additions
Little underpayment
Cheating
Failure to report
Reducing tax
Cutting corners
Holding back some tax
Make deductions
1.29 1.00 1.32 10 .003
0.58 0.21 1.89 10 .005
0.67 1.32 1.50 10 .000
0.98 1.00 2.86 10 .000
1.34 1.38 1.40 10 .005
1.65 1.23 1.98 10 .055
0.95 1.72 1.10 10 .001
0.86 1.11 1.39 10 .004
0.92 1.20 1.60 10 .005
0.42 1.07 1.11 10 .000
From the table 4.11, a one-unit increase in the variable Exchanging or trading services as
well as goods with a neighbor of a friend and then failing to report in your tax returns is not
bad at all is associated with a 0.58 increase in the relative log odds of paying taxes.
14

From the table 4.11, a one-unit increase in the variable income reporting without adequately
including any particular extra small amounts of extra income is okay is associated with a
0.67 increase in the relative log odds of paying taxes.
From the table 4.11, a one-unit increase in the variable being paid cash for a job and then not
reporting it in tax report seems fine is associated with a 0.98 increase in the relative log odds
of paying taxes.
From the table 4.11, a one-unit increase in the variable making some small additions than
actual spending is okay and never a great deal is associated with a 1.34 increase in the
relative log odds of paying taxes.
From the table 4.11, a one-unit increase in the variable making an little underpayment is not
big deal is associated with a 1.65 increase in the relative log odds of paying taxes.
From the table 4.11, a one-unit increase in the variable the rates of tax are very high and
therefore we are justified to cheat as it cannot be regarded as cheating as it makes you pay
less is associated with a 0.95 increase in the relative log odds of paying taxes.
From the table 4.11, a one-unit increase in the variable failure to report some income
emanating from investment or interest is okay if the government will not be able to realize
or discover is associated with a 0.86 increase in the relative log odds of paying taxes.
15

From the table 4.11, a one-unit increase in the variable obviously makes a reasonably good
sense to reduce your tax when a chance avails itself is associated with a 0.92 increase in the
relative log odds of paying taxes.
Lastly, From the table 4.11, a one-unit increase in the variable making deductions is
associated with a 0.42 increase in the relative log odds of paying taxes.
Further, it was needful to run factor analysis. This process identified latent constructs. Table
4.10 shows the KMO and the Bartletts Test.
Table 4.10: KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .894
Bartlett's Test of
Sphericity
Approx. Chi-Square 543.502
Df 330
Sig. .000
As seen in table 4.10, the KMO value is 0.894, indicating that the data is suited for factor
analysis. The sampling is adequate. Since the KMO that is greater than 0.5 and very close to
1.0, it is very useful. Results of Bartlett's test of sphericity show that the sig value was .000
with 330 degrees of freedom. Therefore, it is justified to reject the null hypothesis.
The results rhyme with the explanation on Adam Smith’s canons of taxation in the literature.
Since the respondents trusted the government, they felt it was their responsibility to pay
taxes and didn’t look for every available opportunity to dodge tax remittance.
16

17

CHAPTER FIVE: SUMMARY OF FINDINGS CONCLUSION AND
RECOMMENDATIONS
5.1 Introduction
This chapter presents the summary of findings, discussions, conclusions, recommendations
of the study and areas for further research.
5.2 Summary of the Findings
5.2.1 General fairness
The study established that the relationship of tax payers in Lithuania with the government is
indeed affected by the perception that the taxpayer has of the government. Fairness
perception is a very critical element in tax compliance in this study. The findings are
consistent with other studies conducted in Australia. If the citizenry perceived the tax system
as oppressing and discriminatory, they were more unlikely to comply to their tax
obligations. On the other hand, if the respondents perceived the tax system as being fair,
there were increased chances of tax compliance.
5.2.2 Effects of self interest in exchange with the government on tax compliance
The study established that all the independent variables taken together significantly predict
tax compliance. As a consequence, the null hypothesis is rejected. Since the respondents
trusted the government, they felt it was their responsibility to pay taxes and didn’t look for
every available opportunity to dodge tax remittance.
5.3 Conclusion
From the current study, there is a statistically significant relationship between tax
compliance and the perception of tax fairness. Lithuanians are more likely to pay their taxes
18

if their perception of tax fairness in the country are positive. Therefore, the government
should strive to create a rapport with its citizens so as to ensure that they collect more taxes
at minimal cost. Perception of tax fairness is a critical part of tax compliance and the
Lithuanian government should treat it as so. With an improved positive perception of the tax
system, the Lithuanian government will collect more taxes and offer better services to the
people.
5.4 Recommendations of the Study
The following recommendations were made from the study:
The county government of Lithuania should ensure that it takes note of the perceptions of
the people as it navigates the tax options so as to improve on the relationship and get more
tax remittance form the people.
The study also recommends that since the people of Lithuania know that they need to pay
taxes, the government should not take advantage of their willingness to pay taxes and
impose unreasonable tax levies. It is important that the government understands that the
people are willing t trust the government based on the benefits that they receive from the
government in the form of services paid for by the taxes.
5.5 Areas for Further Research
The following are recommendations for further research:
In a bid to ensure that there is a better understanding on the effect of perception of fairness
on complying to tax responsibility , another study needs to be conducted in the country. A
larger sample should be used and a different methodology employed. When this is done it
will be possible to see of the results can be replicated. After a replication of the results,
stronger conclusions can be drawn.
19

References
Jha, D.N. (1975). Revenue system in Post-Maurya and Gupta Times. Doi: 10.2307/3517941
Gilligan, G. & Richardson, G. (2005) "Perceptions of tax fairness and tax compliance in
Australia and Hong Kong ‐ a preliminary study", Journal of Financial Crime, Vol.
12 Issue: 4, pp.331-343, https://doi.org/10.1108/13590790510624783
Kircher, E., Hoelzl, E., and Wahl, I. (2008). Enforced versus voluntary tax compliance: The
slippery slope” framework. Journal of Economic Psychology, 29 (2), 210-225.
Natrah, N. (2012). Perceptions of tax fairness and tax compliance behaviour: A comparative
study. Jurnal Pengurusan. 36. 89-100.
20