See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/360055445

Assessment on Factors that Affect Tax Collection in Tepi Town Merchant’s

Yeki Woreda, Sheka Zone, South West Ethiopia

Article · August 2021

CITATIONS READS

0 952

1 author:

Alemu Bekele
Mizan-Tepi University
10 PUBLICATIONS   13 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Research methodology and Statistical software View project

All content following this page was uploaded by Alemu Bekele on 19 April 2022.

The user has requested enhancement of the downloaded file.

 ur
nal
of Acc
ou OPEN ACCESS Freely available online
International Journal of
o

nt
tional J

ing searc
Re Accounting Research
rna
te

ISSN: 2472-114X
h
In

Research Article

Assessment on Factors that Affect Tax Collection in Tepi Town Merchant’s Yeki
Woreda, Sheka Zone, South West Ethiopia
Alemu Bekele Eticha*, Dawit Wendimu
Department of Statistics, College of Natural and Computational Science, Mizan-Tepi University, Tepi, Ethiopia

ABSTRACT
Background: Taxation is a system of raising money to finance government. Without taxes funds and government
services could not exist. In most developing countries, like Ethiopia, the revenue generated by the government is
quite less than the expenditures spent. This low revenue yield of taxation can only be attributed to the fact that tax
provisions are not properly enforced. Thus the main objective of this study is to investigate factors of tax problems
in taxpayers in Tepi town.
Methods: The community based cross-sectional study design was conducted on 118 merchants that randomly selected
from Tepi Town. All the vital information was gathered from selected tax payers by structured questionnaire. The
data was organized and analyzed by IBM SPSS 20. Moreover, a Binary logistic regression was used to identify the
potential determinants of tax collection problem.
Results: Out of 118 tax payers 14.4% were said that they had not faced tax collection problems and 85.6% were said
that they have tax collection problems. From the total, 36 female tax payers 5(4.2%) of them were reported that they
had not tax payment problem while 31(26.3%) of them were had faced tax payment problem. Similarly, from the
total 82 male tax payers 12 (10.2%) of them were said they had not faced tax collection problem but 70(59.3%) of
them were said they had faced tax collection problem. The result of Chi-square of independence revealed, service
satisfaction, cash register machine use, pay amount, and duration of time liability had p-value 0.001, 0.045, 0.001
and 0.000 respectively which is less than level of significance 0.05. So this indicated the significant relationship that
found between these variables and tax gathering system. The logistic regression revealed that service satisfaction, pay
amount and time liability were the major factor that affects tax collection system in Tepi town.
Conclusions: Based on results, about 85.6% of merchants in Tepi town were had complain on their taxation. The
identified significant factors of these merchants are service satisfaction, pay amount estimate and time liable to pay
the taxes. The study suggested to Tepi town Economic developmental office must adopt a comprehensive strategy
that satisfy the tax payers and adequate pay estimate that depend on profit of the merchants.
Keywords: Tax payment; Tepi town; Logistic regression; Factors of taxation; Problem of taxation

Introduction are an increasing government spending and deficit financing. The

major revenue of the government to do infrastructure is tax and
In most developing countries, it is a common phenomenon to if the problem in tax collection was solved the life standard of the
notice serious problems in developing adequate tax systems that people is improved.
permits a government to sufficiently finance its expenditures.
Ethiopia, like any other developing countries, faces difficulty in A tax is defined as a "a compulsory contribution payable by an
raising revenue to the level required for the promotion of economic economic unit to a government without expectation of direct
growth. Taxation has increased in importance not only as a tool of and equivalent return from the government for the contribution
raising revenue for the traditional roles but also for accelerating made” [1]. Tax administration refers to the identification of
the economic growth and ensuring social justice. Along with the tax liability based on the existing tax law, the assessment of this
growth in the overall Ethiopian economy, it observes that there liability, and the collection, prosecution and penalties imposed

*Correspondence to: Alemu Bekele Eticha, Department of Statistics, College of Natural and Computational Science, Mizan-Tepi University, Tepi,
Ethiopia, E-mail: alemubekele1@gmail.com
Received: January 24, 2021; Accepted: August 10, 2021; Published: August 17, 2021
Citation: Eticha AB, Wendimu D (2021) Assessment on Factors that Affect Tax Collection in Tepi Town Merchant’s Yeki Woreda, Sheka Zone, South
West Ethiopia . Int J Account Res 9:220.
Copyright: © 2021 Eticha AB, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Int J Account Res, Vol.9 Iss. 8 No: 220 1

Dicicco JM, et al. OPEN ACCESS Freely available online

on recalcitrant taxpayers. Tax administration, covers a wide area adopt a comprehensive strategy, and minimize the observed tax
of study, encompassing aspects such as registration of taxpayers, administration problems to increase tax revenue. Similar approach
assessments, returns processing, collection, and audits [2]. Tax can be replicate in identifying the problems in tax assessment and
administration therefore, should aim at improving on laws collection procedures of revenue administrations in other towns.
regarding the registration, assessment, collection revenue, and
exploiting fully taxation potential of a country [3]. MATERIALS AND METHODS
In most developing countries, like Ethiopia, the revenue generated
Description of the study area
by the government is quite less than the expenditures spent. This
low revenue yield of taxation can only be attributed to the fact The study was conducted in Tepi town which is located in the Yeki
that tax provisions are not properly enforced either on account Woreda, Sheka Zone, Southern Nation Nationality and People’s
of the inability of administration or on account of straight (SNNPR) Region at 611 km distance from Addis Ababa (the capital
forward collusion between the tax administration and taxpayers. city of Ethiopia) toward south west direction of the country and
In Ethiopia, tax is administered at federal or central and regional this study will conduct to tax payer merchants in case of Tepi town.
levels. The constitution of Federal Democratic Republic of
Ethiopia (FRDE) has separated the tax revenue to be collected by Study population: In this study the target population is merchants
federal government, state or regional government and jointly by those who have financial account in the in Tepi town revenue
the federal and state government (Proclamation No33/1992). office. The total populations are 1706 and the sample was selected
from these taxpayers.
The revenue of the Ethiopian government comes from different
sources such as tax public borrowing, sales of public assets, and Sampling techniques: The cross-sectional study design conducted
transfer payments. The main purpose of generating revenue from on 118 tax payer merchants that selected with simple random
these various sources is to finance government expenditure. These sampling from 1706 sample frame. Because each individual is
public expenditures are meant for public goods and services that chosen entirely by chance and each member of the populations
are very essential for the development and wellbeing of the society. have an equal chance of being included in the sample.
Ethiopian Revenue and Customs Authority (ERCA) focuses on Study variables: The interest variable is the tax collection problem
those people and vehicles that may involve in the act of bringing that categorized as the tax payer faced problem or did not faced
into or taking out of goods, which customs duty and taxes are not any problem on tax collection. The independent variables were:
paid and whose importation or exportation are prohibited by law. education level of merchant, income of merchant, satisfaction level
Taxes levied by central and regional government consist of direct of the merchant, tax use awareness, willing to pay tax, Cash register
and indirect taxes. Direct taxes are taxes including employment machine use, business type, tax category, pay amount annually, and
income taxes, business income tax while indirect taxes are mainly duration to pay tax.
composed of value added tax (VAT) and excise taxes. The law has
classified the business income tax payers on business profit on to METHODS OF DATA ANALYSIS
three major categories with respect to their legal personality and
annual turnover as category A, B and C. Category "A", category "B" Descriptive statistics
and category "C" taxpayers are classified as follows [5]. Category “A”
taxpayers are composed of two groups. The first group comprises Descriptive statistics is a part of statistics that deal with methods
of those taxpayers whose annual turnover for a single tax year is and techniques of organizing, summarizing, presenting, reporting
500,000 or more. In addition, any company incorporated under and arranging the data without making generalization beyond the
the laws of Ethiopia is a category “A” taxpayer irrespective of data. It summarizes mass of the numerical data in to meaningful
their annual turnover. Category “B” taxpayers are those taxpayers form by using various statistical techniques such as tables, charts,
with annual turnover greater than 100,000 but less than 500,000 graphs, and so on.
Ethiopian. All taxpayers with annual turnover income less than
100,000 Ethiopian Birr are grouped as category C taxpayers Birr. Chi-square Test

According to Mekonnen, A., Deneke, Z. and Reda E, tax The test is used to test the association between two variables. Test of
assessment and collection problems on taxpayers were problem association we apply to know the relationship between dependent
for system connection due to this many taxpayers does not get and independent variables [6]. The Chi-square is given by:

(O − E ij )
the service as they came to the tax office and there is lack of tax 2
n m
χ = ∑∑
2 ij
knowledge by taxpayers. Due to this, delay in tax payment, do (1)
not pay the proper amount they should pay and negligence are i =1 j E ij
taken by taxpayers as solution to escape from payment of taxes.
Hence, proper assessment and collection of tax is one of the factors Where: m = number of column
that enable the government to achieve its goals programs. Besides, O ij = number of occurrence in the cell
it reduces the country’s dependability on the foreign loan and
E ij = expected value of i th =row and j th =column
donations. This study is focus on tax collection problem in Tepi
Town. n = number of row
The findings of this study will contribute in enhancing the X2 cal = chi-square test of statistics
taxation or revenue of the town and the government at large by
P-value is the smallest level of the test for which the null hypothesis
uncovering the core problems on the tax assessment and collection
(Ho) is rejected. That is when p-value greater than the significance
activities so as to enabling them to put their effort to trump over
level, Ho is not rejected.
the observed problems. Thus, the government would be able to

Int J Account Res, Vol.9 Iss. 8 No: 220 2

Dicicco JM, et al. OPEN ACCESS Freely available online

The test X2cal with X2α(c-1) (r-1)…. tabulated value. commonly used 0.05. That is the large p-value indicates good fit.

Binary logistic regression Maximum Likelihood:

The overall significance of the binary logistic regression model is
Logistic regression describes the relationship between a categorical
checked by using maximum likelihood test. This is given by:
response variable and any set of explanatory variables. Binary
logistic regressions were used to perform logistic regression on a L 
G 2 = -2log  0  = -2log ( L0 - L1 )
binary response variable. A binary variable has only two possible  L1 
values, such as presence or absence of a particular event. Models
Hypothesis:
with one or more predictors we fitted using an iterative-reweighed
least squares algorithm to obtain maximum likelihood estimates of Ho: B0=B1=…..= Bk=0 vs.
the parameters. Binary logistic regression has also used to classify
observations into one of two categories, and it may give fewer H1: At least one of the coefficients is different from the other.
classification errors than discriminates analysis for some cases [4]. Decision: Reject H0 if p-value is less than α-value commonly used
The response variable in the study is tax problem faced or not 0.05
faced.
Wald Test
1 − if thetax payer faced atleast any problem 
Y =  The significance of the coefficient of individual variable in the
 0 − if thetax payer has not raise any problem 
binary logistic regression model is tested by using Wald test statistic.
As stated the model binary logistic regression as Which is given by:-
exp ( β 0 + β1 X 1i + β 2 X 2 i +  + β k X ki )  2 β − β0 2 
π ( x) = (2) Z = ( ) 
1 + exp ( β 0 + β1 X 1i + β 2 X 2 i +  + β k X ki )  SE ( β ) 

Where, X ki =independent variables in the model Hypothesis Test: Ho: [ β j = 0] vs. H1: β j ≠ 0

π = probability of success; 1–π = the probability of failure Decision: Reject Ho if Z 2 >= Z2α (v) or the corresponding P-value of
each variables less than significance level α.
β0 = intercept; βk= the coefficients of independent variable
The general model is simplified to: RESULTS
 π 
Logit (π i ) = log   = β 0 + β1 X 1i + β 2 X 2 i +  + β k X ki Descriptive statistics Results
1− π 
Where, π i = y = 1 is the probability of success and 1–πi = probability The main purpose of this section is to interpret the results on tax
of failure collection problem of tax payers. The descriptive statistics was
employed to describe the tax collection problem of tax payers.
Odds Ratio: Logistic regressions work with odds so it is necessary to
πi From Table 1, out of 36 female respondent 5(4.2%) of them were
define both odds and odds ratio. The odds are simply odd = reported that they had no tax payment problem while 31(26.3%)
1 − πi
π1 of them were had faced tax collection problem. Similarly, from the
In 2 × 2 tables, within row 1 the odds of success are Odds1 = , and
1 − π1 total 82 male respondents 12 (10.2%) of them were said they had
π2 no faced tax collection problem but 70(59.3%) of them were said
Within row 2 the odds of success equals Odds 2 = they had faced tax collection problem.
1 − π2
π1 Out of 8 total number age group less than 20 respondent 1(0.9%)
odds1 1 − π1
The ratio of the odds from the two rows,=
Πi = of them were said they had not faced tax collection problem and
oods 2 π2
7(5.9%) of them were said that they had faced tax collection
1 − π2
problem, from 15 total number of age group between 21-30
Пi (Odds ratio) is the ratio of two odds; whereas the relative risk is respondents 3(2.5%) of them were said that they had not faced
a ratio of two probabilities tax collection problem and 12(10.2%) of them were said that they
Model adequacy check had faced tax collection problem, from 38 total number of age
group between 31-40 respondents 7(5.9%) of them were said that
The Hosmer-Lemeshow goodness of fit statistic measures the they had not faced tax collection problem and 31(26.3%) of them
correspondence between the actual and predicted values of were said that they had faced tax collection problem, from 29 total
the dependent variable [4]. It also commonly used to assess the number of age group between 41-50 respondents 2(1.7%) of them
goodness of fit of a model and allows for any number of explanatory were said they had not faced tax collection problem and 27(22.9%)
variables, which may be continuous or categorical. It is given by: of them were said that they had faced tax collection problem and
n
(Oi − Ei ) 2 out of the total 28 age group above 50 respondents 4(3.4%)
Cˆ = ∑
i =1 mi ∗ πˆ i (1 − πˆ i )
of them were said they had not faced tax collection problem and
24(20.3%) of them were said that they had faced tax collection
Where, Oi =observed value; Ei =Expected value; mi = number of problem.
observation and π= average predicted risk for ith group
Out of 16 total number of illiterate education level respondents
Hypothesis: H0: The model is good fit Or H1: The model is poor fit 3(2.6%) of them were said they had not faced tax collection
Decision rule: do not reject H0 if p-value is greater than α-value problem and 13(11%) of them were said that they had faced tax

Int J Account Res, Vol.9 Iss. 8 No: 220 3

Dicicco JM, et al. OPEN ACCESS Freely available online

collection problem, from 25 total number of elementary school respondents 4(3.4%) of them were said that they had not faced tax
respondents 5(4.2%) of them were said they had not faced tax collection problem, and 14(11.9%) of them were said that they had
collection problem and 20(17%) of them were said that they had faced tax collection problem. The remaining 100 respondents of
faced tax collection problem, from 23 total number of high school tax awareness are yes group and from those 13(11%) of them were
completed respondents 3(2.5%) of them were said they had not said that they had not faced tax collection problem and 87(73.7%)
faced tax collection problem and 20(17%) of them were said that of them were said that they had faced tax collection problem.
they had faced tax collection problem, from 20 total number of
Out of total 118 cash register machine respondents of 16(13.6%)
vocational certificate and Diploma respondents 3(2.5%) of them
of them are no group and 1(0.9%) of yes group were said that they
were said they had not faced tax collection problem and 17(14.4%)
had not faced tax collection problem. The remaining respondents
of them were said that they had faced tax collection problem, and
of cash register machine 72(61%) of no group and 29(24.5%) of
out of 34 degree and above respondents 3(2.5%) of them were said
them were said that they had faced tax collection problem.
they had not faced tax collection and 31(26.3%) of them were said
that they had faced tax collection problem. The percentage of get adequate training about the machine
respondents of 0% and 0.9% were said that they had not faced tax
out of 29 total number of average income level less than 50,000
collection problem. The remaining of get adequate training about
respondents 5(4.2%) of them were said they had not faced tax
the machine respondents 10.2% and 14.4% were said that they
collection problem and 24(20.4%) of them were said that they
had faced tax collection problem.
had faced tax collection problem, from 41 total number of average
income level between 50,000-250,000 respondents 5(4.2%) of The percentage of signature to tax identification respondents of
them were said they had not faced tax collection problem and 2.5% and 11.9% of them were said that they had not faced tax
36(30.5%) of them were said that they had faced tax collection collection problem. The remaining respondents of signature to tax
problem, from 17 total number of income level between 250,000- identification 4.2% and 81.3% of them were said that they had
500,000 respondents 4(3.4%) of them were said they had not faced faced tax collection problem.
tax collection problem and 13(11%) of them were said that they
Out of 118 total respondents for their category 25.4% of them
had faced tax collection problem, from 22 total number of income
are category A, 26.3% of them are category B and 48.3% of them
level between 500,000-1,000,000 respondents 1(0.9%) of them
are category C tax payers and from those categories 2.5%, 3.4%
were said they had not faced tax collection problem and 21(17.7%)
and 8.2% of them were said that they had not faced tax collection
of them were said that they had faced tax collection problem and
problem respectively. The remaining respondents of tax categories
out of 9 total number of income level above million respondents
such as category A, category B and category C of tax payers 22.9%,
2(1.7%) of them were said they had not faced tax collection
22.9% and 39.8% of them were said that they had faced tax
problem and 7(5.9%) of them were said that they had faced tax
collection problem respectively.
collection problem.
The percentage of payment amount they should respondents of
From total of 15 not pay tax annually respondents 4(3.4%) of them
59.3% of them were not pay and 40.7% were says pay amount
were said they had not faced tax collection problem and 11(9.3%)
they should, from not pay amount they should group 3.4% and
of them were said that they had faced tax collection problem and
from those pay amount they should group 11% of them were said
out of 103 total pay tax annually 13(11%) of them were said they
that they had not faced tax problem. The remaining respondents
had not faced tax collection problem and 90(76.3%) of them were
of not pay amount they should 55.9% and pay amount they should
said that they had faced tax collection problem.
group 29.7% of them were said that they had faced tax collection
From the above table 4.1 business types such as hotel and problem.
restaurant, clinic, trader, games and others is 1.7%, 0%, 5.9%,
The percentage of duration of time liability respondents of 4.2%
1.7% and 5.1% of them were said they had not faced tax collection
and 10.2% of them were said that they had not faced tax collection
problem respectively. The remaining respondents of business type
problem. The remaining respondents of duration of time liability
such as hotel and restaurant, clinic, trader, games and others is
for 66.9% and 18.6% of them were said that they had faced tax
5.8%, 9.3%, 26.3%, 14.4% and 28.8% of them were said that they
collection problem (Table 1).
had faced tax collection problem respectively.
Out of 118 tax payer respondents 93(78.8%) of them were not Chi square test of association for tax payers
satisfied to tax collection service, from those 93 respondents
From Table 2, we seen the independent variables (service
8(6.8%) of them were said they had not faced tax collection problem
satisfaction, cash register machine, pay amount you should, and
and 85(72%) of them were said that they had faced tax collection
duration of time liability) have p-value (0.001, 0.045, 0.001 and
problem. The remaining, 25 respondents were says satisfied to the
0.000) respectively which is less than level of significance (0.05
service, from those 25 respondents 9(7.6%) of them were said they
respectively). So we conclude that there is strong relationship
had not faced tax collection problem and 16(13.6%) of them were
between the independent variables and tax collection problem of
said that they had faced tax collection problem.
tax payers (dependent variable) (Table 2).
The percentage of level of service satisfaction such as very satisfied,
moderately satisfied and low satisfaction categories is 2.5%, 1.7% Binary Logistic Regression
and 3.4% of them were said they had not faced tax collection
From the Table 3, the sig value=0.00 is less than the test statistics
problem respectively. The remaining respondents level of service
α=0.05 we reject the null hypothesis and conclude as the overall
satisfaction such as very satisfied, moderately satisfied and low
model is significantly different from zero. This means that at least
is 1.7%, 7.6% and 4.2% of them were said they had faced tax
not one of the explanatory variable are excluded from the model,
collection problem respectively. Out of 18 tax awareness no group
the response variable is equal to 0.000 which is constant (Table 3).

Int J Account Res, Vol.9 Iss. 8 No: 220 4

Dicicco JM, et al. OPEN ACCESS Freely available online

Table 1: Table of descriptive statistics on tax payment data, 2019.

Variable Categories Status Tax payers on Tax collection problems Total
No Yes
Gender Female 5(4.2%) 31(26.3%) 36(30.5%)
Male 12(10.2%) 70(59.3%) 82(69.5%)
Age category Less than 20 1(0.9%) 7(5.9%) 8(6.8%)
21-30 3(2.5%) 12(10.2%0 15(12.7%)
31-40 7(5.9%) 31(26.3%) 38(32.2%)
41-50 2(1.7%) 27(22.9%) 29(24.6%)
Above 50 4(3.4%) 24(20.3%) 28(23.7%)
Education Level Illiterate 3(2.5%) 13(11.1%) 16(13.6%)
elementary school 5(4.2%) 20(17%) 25(21.2%)
High school 3(2.5%) 20(17%) 23(19.5%)
Certificate & Diploma 3(2.5%) 17(14.4%) 20(16.9%)
Degree & above 3(2.5%) 31(26.3%) 34(28.8%)
Average annual income Less than 50,000 5(4.2%) 24(20.4%) 29(24.6%)
50,000-250,000 5(4.2%) 36(30.5%) 41(34.7%)
250,000-500,000 4(3.4%) 13(11%) 17(14.4%)
500,000-1,000,000 1(0.9%) 21(17.7%) 22(18.6%)
More 1 million 2(1.7%) 7(5.9%) 9(7.6%)
Pay tax annually No 4(3.4) 11(9.3) 15(12.7%)
Yes 13(11.1%) 90(76.2) 103(87.3%)
Business type Hotel and restaurant 2(1.7%) 7(5.8%) 10(8.5%)
Clinic 0(0%) 11(9.3%) 11(9.3%)
Trader 7(5.9%) 31(26.3%) 38(32.2%)
Games 2(1.7%) 17(14.4%) 19(16.1%)
Others 6(5.1%) 34(28.8%) 40(33.9%)
Satisfaction No 8(6.8%) 85(72%) 93(78.8%)
Yes 9(7.6%) 16(13.6%) 25(21.2%)
Level of satisfaction Very satisfied 3(2.5%) 2(1.7%) 5(4.2%)
Moderately satisfied 4(3.4%) 9(7.6%) 13(11%)
Low satisfied 2(1.7%) 5(4.2%) 7(5.9%)
Tax awareness No 4(3.4% 14(11.9%) 18(15.3%)
Yes 13(11%) 87(73.7%) 100(84.7%)
Cash register machine No 16(13.6%) 72(61%) 88(74.6%)
Yes 1(0.9%) 29(24.5%) 30(25.4%)
Adequate training No 0(0%) 12(10.2%) 12(10.2%)
yes 1(0.9%) 17(14.4%) 18(15.3%)
Signature to tax No 3(2.5%) 5(4.3%) 8(6.8%)
identification Yes 14(11.9%) 96(81.3%) 110(93.2%)
Category of tax payers Category A 3(2.5%) 27(22.9%) 30(25.4%)
Category B 4(3.4%) 27(22.9%) 31(26.3%)
Category C 10(8.5%) 47(39.8%) 57(48.3%)
Pay Amount No 4(3.4%) 66(55.9%) 70(59.3%)
Yes 13(11%) 35(29.7%) 48(40.7%)
Duration time liability No 5(4.3%) 79(66.9%) 84(71.2%)
Yes 12(10.2%) 22(18.6%) 34(28.8%)

From Table 4 , the result of Omnibus tests of model confidents From table 5, Cox & Snell R2 indicated that 38.3% of the
had a chi-square value of 57.029 at 26 degrees of freedom, which effectiveness of tax collection problem tax payers is explained by
was highly significant at alpha =0.05. Meaning, model coefficients the independent variables. In addition to this, Nagelkerke’s R2
of chi-square test is highly significant, which shows that the indicated that 68.2% of the dependent variable is explained by
independent variables predict the dependent variables well and the the independent variables (Table 5). From the above table of 6,
model was a good fit (Table 4). Hosmer and Lemeshow Test p value 0.599 is greater than 0.05,
therefore we concluded that the model is good fit the data (Table 6).

Int J Account Res, Vol.9 Iss. 8 No: 220 5

Dicicco JM, et al. OPEN ACCESS Freely available online

Table 2: Chi square test of association between the problem and explanatory variables, 2019.
Variables Chi square Value Degree of freedom Asymp.sig(2-sided)
Level of education 1.779 4 .776
Average annual income 3.680 4 .451
Pay tax annually 2.095 1 .148
Business type 2.845 4 .584
Service satisfaction 11.994 1 .001
Level of satisfaction 1.572 2 .456
Tax awareness 1.052 1 .305
Cash register machine 4.000 1 .045
Adequate training .690 1 .406
Signature to tax Identification 3.711 1 .054
Category of tax payers .984 2 .611
Pay amount you 10.544 1 .001
Duration of time liability 16.898 1 .000

Table 3: Variables in the equation for tax payers, 2019.

Model B S.E. Wald df Sig. Exp(B)
Constant -1.782 .262 46.202 1 .000 .168

Table 4: Omnibus Tests of Model Coefficients.

Model Chi-square df Sig.
Step 1 Step 57.029 26 .000
Block 57.029 26 .000
Model 57.029 26 .000

Table 5: General model summary of regression model for tax payers.

Step -2 Log likelihood Cox & Snell R Square Nagelkerke R Square

1 40.269a .383 .682

Table 6: Hosmer and Lemeshow test of binary logistic regression model for tax payers.
Step Chi-square df Sig.
1 6.435 8 .599

Table7: Classification of response variable in the data, 2019.

Observed Predicted
Problem Faced Percentage Correct
yes no
Step 1 problem Faced yes 98 3 97.0
no 7 10 58.8
Overall Percentage 91.5

From Table 7, showed that allows to correctly classifying 58.8% statistics, df, level of SIG, odds & CI for individual odds. It is
of the subjects where the predicted event (tax collection problem) important to test the significance of individual variables by using
was observed. This is known as the sensitivity of prediction, the P Wald’s statistics and p-value (0.05) correspondently. The level of
(correct/ event did occur), that is, the percentage of occurrences education, satisfaction level, machine use, type of tax and amount
correctly predicted. We also see that this rule allows us to correctly of tax pay are significant factors (Table 8).
classify 97.0% of the subjects where the predicted event was not
observed. This is known as the specificity of prediction, the P Interpretations
(correct event did not occur). Overall Percentage- This gives the
The estimated odds of tax service satisfaction (satisfied group) e β
overall percent of cases that are correctly predicted by the full
=2.032 times more likely than the estimated odds of tax satisfaction
model. Overall, our predictions were correct for an overall success
(not satisfied) for tax collection problem on tax payers.
rate of 91.5 (Table 7).
The estimated odds of pay amount you should (pay amount they
The Table 8, showed, the coefficient, standard error, Wald’s

Int J Account Res, Vol.9 Iss. 8 No: 220 6

Dicicco JM, et al. OPEN ACCESS Freely available online

Table 8: Results from Binary Logistic Regression Analysis for tax payer’s data, 2019.

Variables B S.E. Wald df Sig. Exp(B) 95% C.I.for EXP(B)

Lower Upper
gender(1) 1.334 1.378 .937 1 .333 3.796 .255 56.509
Age(<20) 1.782 4 .776
age(21-30) 2.347 3.527 .443 1 .506 10.451 .010 10498.292
age(31-40) 2.184 2.226 .962 1 .327 8.880 .113 697.487
age(41-50) .459 1.907 .058 1 .810 1.582 .038 66.447
age(>50) .535 2.172 .061 1 .805 1.708 .024 120.538
Edu. (illiterate) 3.074 4 .546
Edu(elementary) -1.652 2.110 .613 1 .434 .192 .003 11.980
Edu(high school) 1.007 1.345 .560 1 .454 2.736 .196 38.185
Edu(certificate & diploma) -1.513 2.048 .546 1 .460 .220 .004 12.196
Edu (degree & above) -2.198 2.293 .919 1 .338 .111 .001 9.935
Avarageicome(<50000) 1.741 4 .783
avanincome(50000-250000) -4.998 41054.571 .000 1 1.000 .007 .000 .
avanincome(250000-500000) -2.606 41054.571 .000 1 1.000 .074 .000 .
avanincome(500000-1000000) 16.157 40193.020 .000 1 1.000 1.310 .000 .
avanincome(above 1000000) -.031 2.992 .000 1 .992 .970 .003 341.419
Besinestype(hotel&restorant) 2.191 4 .701
besinestype(clinic) 2.465 2.499 .972 1 .324 11.758 .088 1576.661
besinestype(trader) -19.348 8531.841 .000 1 .998 .000 .000 .
besinestype(games) .713 1.152 .383 1 .536 2.041 .213 19.519
besinestype(others) -.658 1.772 .138 1 .710 .518 .016 16.694
paytaxannualy(1) 3.218 1.866 2.973 1 .085 24.975 .644 968.249
Category tax(catA) .000 2 1.000
category tax(catB) -4.424 41054.571 .000 1 1.000 .012 .000 .
categorytax(catC) -17.821 8366.519 .000 1 .998 .000 .000 .
satisfaction(1) 4.956 1.692 8.578 1 .003 2.032 5.153 3915.093
awareness tax(1) -3.135 1.996 2.465 1 .116 .044 .001 2.178
cashregmachine(1) -4.591 2.654 2.992 1 .084 .010 .000 1.842
signtident(1) -3.534 2.191 2.602 1 .107 .029 .000 2.138
payamountyoushould(1) 2.777 1.789 2.409 1 .121 16.074 .482 536.210
timeliabilityadequate(1) 4.378 1.544 8.038 1 .005 9.650 3.863 1642.429
Constant -1.769 41054.571 .000 1 1.000 .171

should group) is e β = 16.074 times more likely than the estimated DISCUSSIONS
odds of pay amount you should (not pay amount they should) for
tax collection problem on tax payers. Under this section the discussion is made. According to Tumescent,
factors that affect tax collection problem on tax payers in Debre
The estimated odds of time liability adequate is e β =9.650 times Tabor town were; level education, business sector, tax knowledge,
more likely than the estimated odds of not time liability adequate cash register machine, level of support from tax officers, sufficient
for tax collection problem on tax payers. number and qualified tax officer personnel, any training relevant
The estimated odds using cash register machine is e β =0.010 to duty and responsibility of tax officer and tax officers provide
times less likely than the estimated odds of not using cash register regular information to tax payers. In this research paper the service
machine for tax collection problem on tax payers. satisfaction, pay amount, cash register machine and time liability
adequate were the significant factors to tax collection challenge.
The general binary logistic regression model is. The education level is similar to the study by Tumescent.
 π 
Log it   =β 1x1 + β 2 x 2 + β 3 x3 + β 4 x 4 Another study done by Mekonnen, A., Deneke, Z. and Reda E,
1− π  on tax assessment and collection problems on taxpayers in Addis
Where x1, x2, x3 and x4 are the fitted independent variables such Ababa city were point out that, system connection due to this many
as (service satisfaction, pay amount they should and time-liability taxpayers does not get the service as they came to the tax office
adequate respectively) and there is lack of tax knowledge by taxpayers. Due to this, delay
π in tax payment, do not pay the proper amount they should pay
Logit ( = 4.956( service satisfaction) + 2.777( pay amount they should ) +
1−π and negligence are taken by taxpayers as solution to escape from
4.374(timeliability adequate respectively ) − 4.591(cash register machine). payment of taxes. The paper is not confirmed this result because

Int J Account Res, Vol.9 Iss. 8 No: 220 7

Dicicco JM, et al. OPEN ACCESS Freely available online

knowledge is not factor in the paper. ACKNOWLEDGMENT

CONCLUSION The authors also greatly indebted to employees of Tepi town
Economic Development Bureau, and merchants those participated
The main objectives of this study have been to investigate factors during data collection. The great thanks to Mizan-Tepi University
that affect tax collection problems in respect to taxpayers found in department of statistics for their data entry lab offer.
Tepi town.
About 85.6% of the tax payers in the town were had faced tax References
collection problem and it is necessary to identify the sources of 1. Bhatia .Definition of tax Blackwell Publishers, Oxford, UK and
problems. The significant source of problems were: lower level Cambridge,UK 2003;2:741-768.
education in tax payer, cash machine use, service satisfaction of
2. Kangave. Improving Tax Administration: A Case Study of the Uganda
revenue office, estimate large pay amount and time liability to pay
Revenue Authority. Afr Law J. 2005;2:145–176.
tax are the main factors that lead tax payment problem. Another
variables such as average annual income, gender, age, business type, 3. World Bank .Lessons of Tax Reform, the World Bank, Washington,
get adequate training about the cash register machine, signature to D.C, USA 1991.
tax identification, level of satisfaction, tax awareness and category 4. Agresti A. Categorical data Analysis. Florida : John Wiley & sons inc
of tax payers indicate that they are not significant to tax collection 2002, USA.
problem. Thus this study suggested Tepi town revenue office to 5. Council of ministers income tax proclamation no. 78/2002, Nayarit
facilitate its service and give adequate time on payment of tax. Gazette, FDRE 2002.
Additionally the tax payer must increase their education and
knowledge on Cash register machine use. 6. Bluman AG. Elementary statistics step by step Approach . New york,
USA.1997.

Int J Account Res, Vol.9 Iss. 8 No: 220 8

View publication stats