Determinants of Tax Revenue: A Cointegration and Causality Analysis for Ethiopia, 1992–2022

Abstract

Tax revenue stands as the lifeblood of any economy, necessary for financing government functions, public services, and developmental projects. To uncover the factors affecting tax revenue and develop strategies for improving or modifying it, conducting quantitative research on taxation is beneficial. Although this area has been extensively studied, existing research often focuses on developed countries and frequently overlooks the impact of variables such as corruption perception and shadow economy on a country’s tax revenue. To this end, the study empirically examines factors influencing tax revenue in Ethiopia, using data from 1992 to 2021. A variety of statistics such as unit root test, ARDL bound test, and ARDL-ECM were employed in the analysis. Post estimation analyses such as stability, normality and autocorrelation tests were also performed. The ARDL-ECM result shows trade openness (LNOPPS), inflation rate (LNFLN), share of manufacturing (LNMAF), corruption perception (LNCP) and (LNGDP) have positive and significant effect on tax revenue both in the long run and short run. The result also indicates a negative and significant relationship between foreign direct investment (LNFDI) and shadow economy (LNSE) with tax revenue. Furthermore, the past value of tax revenue negatively influences the current tax revenue. The findings suggest that to enhance tax revenue, the government should broaden the tax base, curtail overly generous tax holidays, bolster tax enforcement to curb evasion, and promote the development of industrial parks and special economic zones. The findings will have practical implications for policy makers, tax regulators and taxpayers in that it provides useful insight into the potential drivers to domestic revenue.

Full text

Journal of Tax Reform. 2024;10(3):493–509
493
eISSN 2414-9497
© Kebede T.N., Erena O.T., Bawiso E.P., 2024
Original Paper
https://doi.org/10.15826/jtr.2024.10.3.180
Determinants of Tax Revenue:
A Cointegration and Causality Analysis for Ethiopia,
1992–2022
Tekalign Negash Kebede1
, Obsa Teferi Erena1
, Elias Pawulos Bawiso2
1 Hawassa University, Hawassa, Ethiopia
2 Sidama Regional State Revenue Ofces, Leku, Ethiopia
 tekanega@gmail.com
ABSTRACT
Tax revenue stands as the lifeblood of any economy, necessary for nancing
government functions, public services, and developmental projects. To uncover the
factors affecting tax revenue and develop strategies for improving or modifying it,
conducting quantitative research on taxation is benecial. Although this area has
been extensively studied, existing research often focuses on developed countries
and frequently overlooks the impact of variables such as corruption perception and
shadow economy on a country’s tax revenue. To this end, the study empirically
examines factors inuencing tax revenue in Ethiopia, using data from 1992 to 2021.
A variety of statistics such as unit root test, ARDL bound test, and ARDL-ECM
were employed in the analysis. Post estimation analyses such as stability, normality
and autocorrelation tests were also performed. The ARDL-ECM result shows trade
openness (LNOPPS), ination rate (LNFLN), share of manufacturing (LNMAF),
corruption perception (LNCP) and (LNGDP) have positive and signicant effect on
tax revenue both in the long run and short run. The result also indicates a negative
and signicant relationship between foreign direct investment (LNFDI) and shadow
economy (LNSE) with tax revenue. Furthermore, the past value of tax revenue
negatively inuences the current tax revenue. The ndings suggest that to enhance
tax revenue, the government should broaden the tax base, curtail overly generous
tax holidays, bolster tax enforcement to curb evasion, and promote the development
of industrial parks and special economic zones. The ndings will have practical
implications for policy makers, tax regulators and taxpayers in that it provides useful
insight into the potential drivers to domestic revenue.
KEYWORDS
ARDL, co-integrations, determinants, ECM, Ethiopia, tax revenue
JEL C22, C32, H20, H71
УДК 336.22
Детерминанты налоговых поступлений:
анализ коинтеграции и причинно-следственных связей
для Эфиопии в 1992–2022 гг.
Текалигн Негаш Кебеде1
, Обса Тефери Эрeна1
, Элиас Павлос Бависо2
1 Университет Хавасса, г. Хавасса, Эфиопия
2 Региональные управления государственных доходов Сидамы, г. Леку, Эфиопия
 tekanega@gmail.com
АННОТАЦИЯ
Налоговые поступления являются источником жизненной силы любой эконо-
мики, необходимым для финансирования государственных функций, государ-
ственных услуг и проектов развития. Для выявления факторов, влияющих на
налоговые поступления, и разработки стратегий по их улучшению или моди-

Journal of Tax Reform. 2024;10(3):493–509
494
eISSN 2414-9497
фикации актуально проводить количественные исследования в области нало-
гообложения. Несмотря на то, что эта область широко изучена, существующие
исследования в основном сосредоточены на развитых странах и часто упускают
из виду влияние таких переменных как восприятие коррупции и теневой эко-
номики на налоговые поступления страны. С этой целью в исследовании эмпи-
рически изучаются факторы, влияющие на налоговые поступления в Эфиопии,
используя данные с 1992 по 2021 г. В анализе использовались различные ста-
тистические методы, такие как тест единичного корня, граничный тест ARDL
и модель исправления ошибок ARDL-ECM. Также были проведены постоце-
ночные анализы, такие как тесты стабильности, нормальности и автокорре-
ляции. Результаты ARDL-ECM показывают, что открытость торговли, уровень
инфляции, доля производства, восприятие коррупции показывают положи-
тельное и значительное влияние на налоговые поступления в долгосрочной
и краткосрочной перспективах. Результаты также показывают отрицательную
и значительную связь между прямыми иностранными инвестициями и тене-
вой экономикой с налоговыми поступлениями. Кроме того, прошлый объем
налоговых поступлений негативно влияет на текущие налоговые поступления.
Полученные данные свидетельствуют о том, что для увеличения налоговых по-
ступлений правительство должно расширять налоговую базу, сокращать чрез-
мерно щедрые налоговые каникулы, усиливать налоговое правоприменение
для борьбы с уклонением от уплаты налогов, а также содействовать развитию
индустриальных парков и особых экономических зон. Полученные результаты
будут иметь практическое значение для директивных органов и налоговых ре-
гуляторов, поскольку они дают полезную информацию о потенциальных фак-
торах, влияющих на внутренние доходы.
КЛЮЧЕВЫЕ СЛОВА
ARDL, коинтеграция, детерминанты, модель исправления ошибок, Эфиопия,
налоговые поступления
1. Introduction
Globally governments rely heavily
on tax revenue to support their ability
to maintain economic stability, provide
basic public services, and promote sus-
tainable growth. It is the main way that
governments may make investments in
healthcare, education, and infrastructure
to keep the economy and society running
smoothly. Taxes contribute to the creation
of a robust economic framework and the
achievement of long-term growth goals
by providing a consistent ow of revenue.
Any economy depends on tax revenue to
pay public services, government opera-
tions, and development initiatives (Kessy
& Sukartini [1], Balasoiu et al. [2], Cavalli
et al. [3]).
Despite the crucial role of tax reve-
nue mobilization in poverty reduction,
infrastructure development, and service
delivery in developing countries, where
tax-to-GDP ratios range from 10% to 20%
(Fig. 1), compared to 30% to 40% in OECD
countries, there remains a signicant tax
gap that presents a substantial opportuni-
ty for revenue growth, particularly in na-
tions like Ethiopia, necessitating increased
efforts to enhance tax revenue in low-in-
come and Sub-Saharan African coun-
tries (Ajeigbe et al. [4], Ajeigbe et al. [4],
Mascagni & Mengistu [5]).
Since most empirical studies conduc-
ted to study the driver of the amounts of
tax revenue primarily rely on cross-sec-
tional and panel data sets, see for exam-
ple: Chelliah [6], Chelliah et al. [7], Mah-
davi [8], Baunsgaard & Keen [9], Profeta &
Scabrosetti [10], Castro & Camarillo [11],
Delessa & Teera [12], Rodríguez [13], Pi-
ancastelli & Thirlwall [14], Raouf [15],
Chettri et al. [16], Todorovi´c et al. [17].
It is rare to nd country-specic time se-
ries studies on the subject (Gupta [18], Ka-
wadia & Suryawanshi [19]). Country-level
time series analysis is benecial for iden-
tifying country-specic tax revenue deter-
minants and understanding the various
factors affecting tax revenue across diffe-
rent countries.

Journal of Tax Reform. 2024;10(3):493–509
495
eISSN 2414-9497
Although some studies have explored
the variables inuencing tax revenue, re-
search remains limited in the Ethiopian
context. Aynew [20] using the Johansen
maximum likelihood co-integration me-
thod, an empirical analysis was conducted
from 1975 to 2013 to determine the main
factors inuencing tax income in Ethio-
pia. The results showed that real GDP per
capita, foreign aid, and the GDP’s indus-
trial value-added share all have a positive
and considerable long-term impact on tax
revenue. On the other hand, ination has
a negative and substantial impact. A study
utilizing data from 1999/00 to 2015/16
was conducted on the factors inuencing
tax income in Ethiopia.
Gobachew et al. [21] based on the
OLS approach, the results show that
trade openness, per capita income, and
the industrial sector’s percentage of GDP
all have a major impact on tax revenue,
whereas the yearly rate of ination and
the percentage of GDP that is made up of
agriculture have a negative impact.
Desta et al. [22] investigated the de-
terminants of tax revenue in Ethiopia
using the Autoregressive Distributed Lag
(ARDL) approach with time series data
from 1996 to 2020. The ndings indicate
that ination has a positive relationship
with tax revenue; however, agricultural
GDP negatively impacts tax revenue in
the short run during the study period. In
contrast, political stability, service sector
contribution to GDP, and ination posi-
tively affect tax revenue in the long run,
while corruption has a negative impact.
However, there are contradictions in
the ndings of different researchers as
well as in previous investigations. Certain
variables are shown to be relevant in some
research but not in others. The resear-
chers were drawn to this discrepancy and
decided to look at it more. Furthermore,
no research has been done in Ethiopia to
look at the size of the factors affecting tax
revenue using a wider range of variables,
like the shadow economy as a percentage
of GDP and corruption perception, which
is not previously investigated.
This study makes a signicant contri-
bution to tax collection research in Ethiopia
by harnessing reliable empirical data and
advanced econometric methodologies. It
delves into the intricate relationships be-
tween economic structure – encompassing
the shares of agriculture and services in
GDP, as well as the informal and shadow
economies – and key economic factors like
foreign direct investment, nominal GDP,
ination, and trade openness.
Additionally, it explores political and
institutional inuences, including politi-
cal corruption, rule of law, and control of
corruption, all in relation to tax revenue,
providing crucial insights. The ndings
are poised to inform strategic policy de-
cisions that will enhance tax revenue mo-
bilization and ensure scal sustainability
in Ethiopia. By offering a robust analytical
framework for understanding the drivers
of tax income, this study aims to deepen
our comprehension of the scal policy
landscape and lay the groundwork for fu-
ture research in this critical domain.
8.81 8.35 8.09 7.60 7.52 6.66 6.20 5.32 4.51
10.90 10.80 11.12 11.60 11.68 12.24 11.39 12.46 12.55
15.19 14.80 14.97 15.05 14.36 15.10 14.30 13.26
14.56
2014 2015 2016 2017 2018 2019 2020 2021 2022
Ethiopia Kenya Uganda
Figure 1. Tax burden as a percentage of GDP in Ethiopia, Kenya and Uganda
Source: Authors own construction (2024)

Journal of Tax Reform. 2024;10(3):493–509
496
eISSN 2414-9497
The primary objective of the current
study is to investigate the determinants of
tax revenue in Ethiopia from 1992 to 2022
by applying a cointegration and causality
analysis for Ethiopia.
The following research hypotheses were
formulated.
H1: FDI positively and signicantly
impacts tax revenue.
H2: Nominal GDP statistically and
positively affects tax revenue.
H3: The share of agriculture in GDP
has a statistically positive effect on reve-
nue from tax.
H4: The share of manufacturing in
GDP statistically and positively affects the
tax revenue.
H5: Ination statistically and positive-
ly affects the tax revenue.
H6: Trade openness statistically and
positively affects the tax revenue.
H7: Shadow economy statistically and
negatively affects the amount of tax re-
venue.
H8: Corruption Perception statistically
and negatively affects the amount of tax
revenue.
Article structure: Section 2 reviews
the literature on tax revenue drivers.
Section 3 details data sources, variables,
and methodologies. Section 4 presents
empirical results on causality and coin-
tegration, and Section 5 summarizes con-
clusions, policy implications, and future
research directions.
2. Literature Review
The study of tax revenue is guided
by three main theories: the cost-of-service
theory, the benet theory, and social-po-
litical theories. The cost-of-service theory
suggests that taxes should correspond to
the direct cost of services received, akin
to paying for a product, but it is criticized
for neglecting welfare services and be-
ing impractical in calculating individual
costs (Ojong et al. [23]). The benet theory
evolved as a response, proposing that ta-
xes should reect the benets individuals
gain from government services, though
this is challenged by the difculty in
quantifying such benets. Social-political
theories argue that taxation should serve
broader social and political goals rather
than individual benets, aiming to ad-
dress societal issues comprehensively by
(Ojong et al. [23]).
Variables that affect the amount of tax
revenue identied by previous research
are classied under various facets, inclu-
ding economic factors (FDI, nominal GDP,
ination, trade openness), political and in-
stitutional factors (corruption perception),
and sectoral composition (share of Agri-
culture to GDP and share of Manufactu-
ring to GDP, and Shadow Economy).
ALshubiri [24] found a favorable
correlation between tax revenue and fo-
reign direct investment. In the same vein,
Mmbulaheni et al. [25] depicted that FDI
affects signicantly and positively the
South African tax revenue. One possible
explanation for this might be in develo-
ping countries, FDI enhances investments
and employment opportunities while in-
creasing tax revenue. Contrary to this n-
ding, Serin & Demir [26] found in OECD
countries that FDI hurts the amount of tax
revenue. This is mainly due to FDI, which
entails substantial tax incentives and con-
cessions that can reduce the tax base and
overall revenue for local governments
Countries with sustained GDP growth
typically experience a corresponding in-
crease in tax revenues, reecting broader
economic expansion. Supporting this no-
tion, numerous researchers have found
a signicant positive correlation between
GDP and tax revenue, as evidenced by
a study on the BRICS nations (Rahman &
Islam [27]).
Nominal GDP can negatively affect
tax revenue, as Mirović et al. [28] repor-
ted in the case of Baltic countries, in which
case ination outpaces economic growth,
leading to higher nominal income without
a corresponding increase in real purcha-
sing power, which may reduce tax com-
pliance and limit government revenue
from real economic activity.
Ination, through the phenomenon
called “bracket creep”, can lead to in-
creased nominal income and higher tax
brackets, resulting in increased govern-
ment revenue and sales tax receipts, as evi-
denced by research conducted in Ethiopia

Journal of Tax Reform. 2024;10(3):493–509
497
eISSN 2414-9497
Desta et al. [29], despite remaining con-
stant tax rates. In response to ination-re-
lated “tax increases”, taxpayers engage in
tax evasion, shadow economies, and infor-
mal economic activity Negash [30].
On the opposite, Caballé & Pana-
dés [31] reported a negative association
between tax revenue and ination. One
argument put forth is that people or rms
have an incentive to underreport income
to preserve their purchasing power, as in-
ation lowers real taxpayment’s.
Concurring with the research con-
ducted by Todorovi´c et al. [17] in Viseg-
rad Group countries’ ndings, trade open-
ness can positively affect tax revenue by
expanding the tax base through increased
economic activity, leading to higher im-
port duties and sales taxes. But this does
not always persist. In some situations,
trade openness can negatively impact rev-
enue by intensifying competition, which
may reduce domestic rms’ protability
and limit corporate tax collections. Addi-
tionally, countries may offer tax incentives
to attract foreign investment, potentially
diminishing overall tax revenues Shrestha
et al. [32].
The share of manufacturing in GDP is
a powerful engine for economic growth,
often driving substantial tax revenue that
fuels public services and infrastructure.
As established through countless previous
investigations, the proportion of Manu-
facturing in GDP has a favorable relation-
ship Wulandari & Wijaya [33]. However,
as industries evolve, over-reliance on
manufacturing can lead to vulnerability,
especially during economic downturns or
shifts toward automation. This delicate
balance presents both opportunities and
challenges: a robust manufacturing sector
can enhance scal health, yet uctuations
can strain revenue systems.
Agriculture positively impacts tax
revenue by generating employment, sup-
porting exports, and fostering related in-
dustries, enhancing overall economic ac-
tivity and tax contributions, as evidenced
by Wulandari & Wijaya [33] research
in low- and middle-income countries.
However, it can also reduce tax revenue
through subsidies and tax breaks, lead to
environmental remediation costs, and cre-
ate income volatility that affects tax sta-
bility. This negative association between
the agricultural share and tax revenue is
supported by research conducted in India
by Kawadia & Suryawanshi [19] which
produced similar results.
The shadow economy, a double-edged
sword, thrives in the cracks of regulation,
siphoning off tax revenue while simul-
taneously fueling entrepreneurship and
resilience in tough times. While it under-
mines government resources meaning
there is a negative relationship between
shadow economy and tax revenues which
is supported by previous research espe-
cially conduced in case of developing
countries such as Dokas et al. [34], Gnang-
non [35], but this idea also opposed by
researches conducted using unbalanced
panel of 125 countries over 1990–2011 by
Vlachaki [36] found a positive association
among them this can be transformed in to
shadow economy pave the way for inno-
vation and eventual formalization.
There is a broad consensus that pub-
lic ofcial corruption are social issues
that can greatly diminish tax revenue and
severely impact economic growth and
development Ajaz & Ahmad [37]. Yet, it
often fuels a culture of innovation, as ci-
tizens nd creative ways to navigate and
thrive within a awed system. This dua-
lity creates a perplexing landscape where
mistrust and inefciency coexist with re-
silience and resourcefulness Gupta [18].
3. Data and Methods
3.1. Data and Variable Measurement
The dataset was obtained from the
popular database of the World Develop-
ment Indicator and Ethiopian Statistical
Service Annual Survey Report. Detailed
variables denition and measurement are
presented in Table 1.
3.2. Stationary Test
The natural start in a time series or
panel data analysis is testing the statio-
nary of the interested variables. Statio-
nary refers to the stability of a variable
over time in terms of its mean, variance,

Journal of Tax Reform. 2024;10(3):493–509
498
eISSN 2414-9497
and autoregressive. There have been
several conventional statistical tools for
stationary test in literature, but we have
used two methods: Augmented Dic-
key-Fuller (ADF) and Phillips – Perron
(PP) – Fisher Chi-square. Each test pro-
vides different statistics and probability
values with different assumptions: For
instance, the ADF test helps to capture
a series of autoregressive orders with
higher lag variables. PP – Fisher Chi-
square is a nonparametric test that pro-
vides more robust results in addressing
autocorrelation heteroscedasticity in er-
ror term distribution.
Given the specic assumptions in
each test method, the general equation of
unit root can be expressed as follows:
11
1
.
k
t i i t it i t t
i
xa t x x
−−
=
Δ = +δ +γ +ρ + αΔ +ε
∑
(1)
The parameters δ and γ represent
the two xed effects. There are two ba-
sic assumptions; (1) specic xed effect
is homogenous and (2) the parameter of
the lagged dependent variable is homo-
genous over time. The optimal number of
lag difference variables is determined by
lag variable determination criteria.
3.3. Optimal lag Length
The choice of lag time signicantly im-
pacts the inference of the ADF test in time
series analysis, particularly for autoregres-
sive models ([Abdou & Atya [38], Agiak-
loglou & Newbold [39], Schwert [40]).
The optimal lag number helps to esti-
mate the parameters properly because too
few large or large numbers of lags would
be misled by reducing the power of the mo-
del or making type I or II errors concerning
null hypothesis rejection or acceptance.
Various statistical methods, including
Sequential modied test statistics (LR),
Final Prediction Error (FPE), Akaike In-
formation Criterion (AIC), Schwarz infor-
mation criterion (SC), and Hannan-Quinn
information Criterion (HQ) are commonly
used to determine optimal lag variable
length. In the time-series most scholars
use AIC because it provides robust results.
Moreover, we would use most results to
x the lag number.
3.4. ARDL-ECM
The section of the parameter estima-
tion model is highly dependent on the
stationary property and co-integration. If
all variables are stationary at level I (0),
Table 1. Variable denition and measurement
Symbol Variables Measurements Formula Source
TR Tax revenue (% of GPD) Tax revenue (% of GDP) ESS
OPPS Trade Openness
GDP
Import Export+
=WDI**
AGRI Share of Agriculture As %
of GDP Agriculture value added (% of GDP) ESS*
INFLN Ination
GDP
100%
GDP
Nominal
Real
= × WDI**
LNGDP Nominal GDP = C + I + G + (X – M) WDI**
LNMAF Share of Manufacture as
% of GDP In terms of revenue, production volume,
% workforces that are employed in the
manufacturing sector, % of the GDP
ESS
LNFDI Foreign Direct Investment Data- obtained from WDI WDI**
LNCP Corruption Perception Corruption Perceptions Index ranges
between 1 to 100 Transparency
International****
LNSE Shadow Economy Shadow economy as percent of total
annual GDP The global
economy***
Notes: * http://www.statsethiopia.gov.et/; ** https://databank.worldbank.org/source/global-nancial-
development/Series/GFDD.EI.10#; *** https://www.theglobaleconomy.com/Ethiopia/shadow_
economy/; **** https://www.transparency.org/en/countries/ethiopia

Journal of Tax Reform. 2024;10(3):493–509
499
eISSN 2414-9497
the OLS model would be the appropriate
estimation method. Alternatively, if all
variables are non-stationary at level but
stationary at rst order: I (1) the Autore-
gressive Disturbance Lag Model provides
the optimal estimation.
ARDL-ECM provides the short-run
dynamics or relationship between the ex-
planatory and dependent variables. Mean-
while, ARDL bound test examines the
long run cointegration among variables in
the context of ARDL. The ARDL long-run
estimate models can be transformed into
short-run dynamic ECM models, with
a negative error correction term, indica-
ting how deviations from long-run equi-
librium are corrected over time.
12
34
56
78
01
1
11
11
11
11
1 12 1
3 14
t it
i
j tj k tk
jk
l t l m tm
lm
n tn q tq
nq
s ts r t r
sr
tt
t
LNTR LNTR
LNOPPS LNGDP
LNFLN LNFDI
LNMAF LNAGRI
LNCP LNSE
LNTR LNOPPS
LNGDP
ρ
−
=
ρρ
−−
= =
ρρ
−−
= =
ρρ
−−
= =
ρρ
−−
= =
−−
−
=α+ α +
+θ +β +
+γ +λ +
+ +Ω +
+ϒ +π +
+δ δ
δ
Φ
+
+ +δ
∑
∑∑
∑∑
∑∑
∑∑
1
5 16 1
7 18 1
91
,
t
tt
tt
tt
LNFLN
LNFDI LNMAF
LNAGRI LNCP
LNSE
−
−−
−−
−
+
+δ +δ +
+δ +δ +
+δ +ε
(2)
where t = 1, ..., T, t – i – denotes the number
of lag variables; ε – error term; LNTR – na-
tural logarithm of total revenue; LNOPPS –
natural logarithm of trade openness;
LNGDP – natural logarithm of the gross
domestic product; LNFLN – the natural
logarithm of ination rate; LNFDI – the na-
tural logarithm of net foreign direct invest-
ment; LNMAF – natural logarithm of share
of manufacturing sector of GDP; LNAGRI –
natural logarithms of share of agriculture
sector of GDP; LNCP – natural logarithms
of corruption perception; LNSE – natural
logarithms of shadow economy.
The parameters (α, β, θ, γ, λ, Φ, ϒ and π)
measure the short-run dynamics or re-
lationship between the explanatory va-
riables and the dependent variables, total
revenue (TR).
The null hypothesis that there is
no short run cointegration can be ex-
pressed (α = β = θ = γ = λ = Φ = 0). In
the equation, the long-run relation-
ship is represented by δ1, δ2, δ3, δ4,
δ5, δ6 and δ7 with the null hypothesis
δ1 = δ2 = δ3 = δ4 = δ5 = δ6 = δ7 = δ8 = δ9 = 0.
3.5. Diagnostic Test
The classical regression model as-
sumptions test is essential for ensuring
the robustness of coefcient estimation, in
addition to assessing the stability of varia-
bles and model sections. The protuberant
assumptions tested in the current study
involve serial correlation, heteroskedas-
ticity, normality, and model stability.
Serial correlation refers to the inde-
pendence of error terms throughout the
study Wooldridge [41]. The study em-
ploys the Breusch-Godfrey Serial Correla-
tion LM test, a popular technique for ana-
lyzing the heteroskedasticity of constant
or homogenous variance over time.
We examined the assumption using
multi-statistics such as Breusch-Pagan
Godfrey, Harvey, Glesjer, and ARCH. The
model stability test was conducted using
the Cumulative Sum of Recursive Resi-
duals (CUSUM) and Cumulative Sum of
Square Recursive Residuals (CUSAR), as-
suming normality and bell-shaped error
terms distribution.
4. Results
4.1. Descriptive Analysis
In Table 2 descriptive results of the
variables with the original data are pro-
vided. Ethiopia performs poorly in terms
of tax revenue when compared to other
Sub-Saharan nations, as seen by the mean
value of Ln of tax revenue ETB (10.33).
The gure would not be surprising be-
cause Ethiopia is characterized by limited
tax bases, weak tax administration which
result in high noncompliance costs, and
a large degree of tax evasion (Adedeji
et al. [42], Alabede [43], Okunogbe &
Santoro [44]). Trade openness (LNOPPS)

Journal of Tax Reform. 2024;10(3):493–509
500
eISSN 2414-9497
has a maximum score of 3.91 and a mean
value of 3.53, showing that the country’s
trade volume has expanded signicantly
over the observation periods.
In terms of sectorial contribution to
GDP, Agriculture (LNAGRI) has been
leading with a mean value of (3.74) and
maximum value (4.15) during 1992 to
2022. Like other African developing coun-
tries, Ethiopia still uses a traditional agri-
cultural system which is likely to have less
contribution to the government tax reve-
nue Gebissa [45].
4.2. Unity Root Test
Stationary is the main attribute of time
series analysis which deals with stability
of the interested variables over time. In
this study, we tested it using the popu-
lar statistical methods: the ADF test and
PP test with two assumptions trend and
trend & intercept. To do it rst, all varia-
bles were examined at level I (0) with the
two test methods.
The result in Table 3 indicates that
three variables (LNFDI, LNFLN, and
(LNOPPS) appeared stationary at 5%
critical value when ADF is assumed but
only two variables (LNFDI and LNFLN)
are stationary under PP test at 5% critical
value.
It is worth noting that some macro
variables might not be stationary at I (0)
so that rst-order I (1) or second order can
be assumed. Our test result reveals a com-
bination of stationary stages with three
Table 2. Descriptive Analysis
LNTR LNAGRI LNFDI LNMAF LNFLN LNOPPS LNGDP LNCP LNSE
Mean 10.332 3.747 0.267 1.562 2.154 3.537 18.005 3.364 3.464
Maximum 13.293 4.156 1.719 1.988 4.011 3.910 22.889 3.664 3.696
Minimum 7.428 3.441 –6.425 1.136 –2.260 3.154 10.158 3.045 3.145
Observations 31.000 31.000 31.000 31.000 31.000 31.000 31.000 31.000 31.000
Source: Author computations and EViews output (2024)
Table 3. Unit Root Test
Variables
ADF PP
Level 1st Difference Level 1st Difference
Intercept Trend &
Intercept Intercept Trend &
Intercept Intercept Trend &
Intercept Intercept Trend &
Intercept
LNTR 0.556
(0.985) –3.535
(0.054) –2.219
(0.204) –2.264
(0.438) –0.091
(0.941) –1.673
(0.738) –3.175
(0.032) –3.118
(0.120)
LNAGRII –2.571
(0.110) –2.639
(0.267) –4.017
(0.004) –4.309
(0.013) –2.570
(0.110) –2.082
(0.5340) –3.926
(0.005) –4.087
(0.016)
LNFDI –5.727
(0.000) –5.038
(0.001) –5.747
(0.001) –5.803
(0.003) –5.747
(0.000) –5.803
(0.000) –6.284
(0.000) –6.315
(0.000)
LNMAF –2.004
(0.283) –2.112
(0.517) –5.315
(0.000) –5.350
(0.000) –2.763
(0.075) –2.742
(0.228) –5.312
(0.000) –5.350
(0.000)
LNFLN –4.553
(0.001) –5.903
(0.000) –5.606
(0.000) –5.668
(0.000) –4.620
(0.000) –6.109
(0.000) –26.67
(0.000) –26.99
(0.000)
LNOPPS –4.275
(0.003) –3.248
(0.100) –3.610
(0.011) –4.000
(0.020) –1.125
(0.692) –0.467
(0.979) –3.610
(0.011) –3.586
(0.488)
LNNGDP –1.792
(0.376) –1.889
(0.634) –4.011
(0.006) –4.172
(0.014) –1.596
(0.471) –1.883
(0.637) –5.660
(0.000) –6.427
(0.000)
LNCP –1.922
(0.317) –2.661
(0.258) –5.865
(0.000) –5.756
(0.003) –1.844
(0.352) –2.687
(0.248) –10.590
(0.000) –10.450
(0.000)
LNSE –0.175
(0.930) –1.643
(0.749) –7.158
(0.000) –7.052
(0.000) –0.454
(0.886) –2,94
(0.164) –6.175
(0.000) –6.035
(0.000)
Source: Authors Computations and EViews (2024)

Journal of Tax Reform. 2024;10(3):493–509
501
eISSN 2414-9497
4.4. ARDL Bound Test
ARDL bounds test examines the exi-
stence of long run cointegration among
the study variables from. Table 5 indicates
F-statistics along with lower bound and
upper bound. If the value of F-statistics is
less than the lower bound there is no coin-
tegration but if it is greater than the upper
bound there exists a long run relationship
among variables. The test result reveals
F-statistics 4.299343 which is higher than
the upper bound at 10%, 5% & 1%, sug-
gesting the existence co-integration.
4.5. ARDL-ECM Results
Upon conrming the long term coin-
tegration, we examine the short run and
long run relationship between tax revenue
and explanatory variables. Table 6 pre-
sents short run coefcients and long run
coefcients.
5. Discussion
The short run result indicates that the
current value of tax revenue is negatively
and signicantly affected by its lag value.
Even though the result is contradicting
with the general view that previous tax
revenue collected helps to mobilize more
tax revenues in the succeeding periods,
it is convincing where there is a weak tax
system, higher tax administration costs,
extensive tax evasion or compliance cost.
If the prior years’ tax revenue collected is
renanced to conduct out comprehensive
variables at level and the remaining six
variables at rst order I (1). There has been
a common view in the extent literature, if
all relevant variables are stationary at the
same level ordinary least squares would
be the appropriate estimation method.
Nevertheless, if the variables are statio-
nary at different order positions OLS mo-
del would be misleading. Alternatively,
the Autoregressive Disturbance Lag Mo-
del (ARDL) could provide reliable estima-
tion Ghouse et al. [46].
4.3. Lag Length Criterion Test
Lag variables are useful in time series
analysis as they aid in accurately predic-
ting both the long-term co-integration and
the short-term dynamism parameters. The
critical argument is how many lag varia-
bles should be included in the analysis.
There have been certain criteria for
deciding on the optimal lag length. Spe-
cically, the Akaike information criteria
(AIC), the Schwarz information criterion
(SC), the Hannan-Quinn information cri-
terion (HQ), the sequential modied LR
test statistic (LR), and the nal prediction
error (FPE). Table 4 summarizes the test
result.
Even though each criterion would
suggest a different lag length, one could
take the majority of test results with the
same level of length. As indicated in AIC,
SC, HQ, FPE, and LR criteria, two lag va-
riables were used in the analysis.
Table 4. Lag Length Criterion Test
Lag LogL LR FPE AIC SC HQ
0–44.7561 NA 3.30e-10 3.707316 4.131649 3.840212
1 191.0702 309.0137 9.90e-15 –6.970358 –2.727026 –5.641399
2361.9734 117.8643* 1.58e-16* –13.17058* –5.108247* –10.64556*
Source: EViews (2024)
Table 5. ARDL – Bound Test
Test Statistic Value Signif., % I (0) I (1)
F-statistic 5.682208*** 10 1.66 2.79
5 1.91 3.11
2.5 2.15 3.40
1 2.45 3.79
Note: *** indicates a signicant level at 1%.

Journal of Tax Reform. 2024;10(3):493–509
502
eISSN 2414-9497
Table 6. ARDL – ECM Results
Variable Coefcient Std. Error t-Statistic Prob.
Short Term Result
D(LNTR(–1)) –0.644*** 0.143 –4.485 0.004
D(LNOPPS) 1.357*** 0.180 7.535 0.0003
D(LNGDP) 0.006 0.007 0.903 0.4012
D(LNFLN) 0.064*** 0.008 7.863 0.0002
D(LNFDI) –0.029* 0.012 –2.388 0.0542
D(LNFDI(–1)) 0.107*** 0.013 7.905 0.0002
D(LNAGRI) 0.198 0.186 1.063 0.3286
D(LNAGRI(–1)) 1.067*** 0.214 4.971 0.0025
D(LNMAF) 0.354** 0.100 3.507 0.0127
D(LNMAF(–1)) –0.364*** 0.092 –3.915 0.0078
D(LNCP) 0.572*** 0.104 5.456 0.0016
D(LNCP(–1)) –0.518*** 0.110 –4.689 0.0034
D(LNSE) –0.734*** 0.109 –6.703 0.0005
D(LNSE(–1)) 0.448*** 0.101 4.423 0.0045
CointEq(–1)* –0.290*** 0.026 –10.923 0.000
Long Term Result
LNOPPS 2.821** 0.559 5.044 0.0023
LNGDP 0.108* 0.051 2.084 0.082
LNFLN 0.620*** 0.114 5.444 0.0016
LNFDI –0.789** 0.230 –3.430 0.014
LNAGRI 0.254 1.245 0.204 0.844
LNMAF 3.417*** 0.902 3.787 0.0091
LNCP 4.684*** 0.703 6.654 0.0006
LNSE –6.819*** 1.076563 –6.3349 0.0007
tax reforms and structure, there will be
a positive impact in the long run.
Complicated tax laws, on the other
hand, may promote tax evasion and raise
the proportion of noncompliant taxpayers
by imposing unfair taxes. Taxpayers may
become disinclined to decline tax return
reports and may have trouble in paying
taxes if the tax authority makes aggres-
sive measures to boost collection within
a given timeframe. Ultimately, this results
in a decrease in tax revenue. The recent
study by Jemiluyi & Jeke [47] provides
similar result that recent past tax revenue
might not have signicant effect on cur-
rent tax revenue because mobilization of
tax revenue requires long term efforts.
The other noteworthy nding is that
trade openness signicantly increases
Ethiopia’s tax revenue both in long run and
long run. Based on our ndings, we con-
dently accept the formulated hypothe-
ses, which indicate that trade openness
signicantly and positively affects tax re-
venue. It indicates a 1% increase in exports
and imports, raises tax revenue by 1.357%
and 2.821% in the short run and long run,
respectively.
This could occur in two ways: rst,
an increase in import goods from foreign
countries associated with several taxes
structure and other charges at different
transaction stages. For instance, import
goods are subject to exercise and other

Journal of Tax Reform. 2024;10(3):493–509
503
eISSN 2414-9497
charges at port level and when gets the
goods into the destination country, the
same goods are subject to sales taxes such
as value added tax and turn over tax.
Second, an increase in export goods and
services is basically proceeding by a large
scale of domestic production which are
taxed at production level and export level.
Given these facts, Ethiopia may great-
ly benet from promoting trade openness,
which will provide vital funds for public
services and development. This empha-
sizes the importance of trade liberalizing
policies A large number of prior studies
suggest that trade openness substantially
benet tax revenue in the long run Piancas-
telli & Thirlwall [14], Todorovi´c et al. [17],
Gnangnon & Brun [48], Tien et al. [49].
The long run coefcient in GDP indi-
cates a positive signicant effect (at 10%) on
tax revenue. It shows a 1% growth in GDP
contributes to a 0.108% in tax revenue in
the long run while all other factors remain
constant. Drawing from our ndings, we
afrm the formulated hypotheses, which
indicate that nominal GDP signicantly
and positively inuences tax revenue. In
practice, tax revenue potential depends
on the economic status of the country. As
economy growths, the tax structure or bas-
es will broaden, and efcient tax adminis-
tration be achieved. Prior studies like Afon-
so et al. [50], Alinaghi & Reed [51], Arnold
et al. [52] conrms the results. The short
run coefcient is insignicant.
Table 6 also shows that the nominal
ination rate (LNFLN) has a positive sig-
nicant effect both in the short run and
long run coefcients, meaning that a one
percent increase in the CPI might result
in a 0.620% in the long run and 0.064%
in the short run. There is an argument in
macroeconomics that an increase in con-
sumer price index would negatively affect
the income of taxpayers which in turn re-
duce taxable income. On the other hand,
as the economy gets more ination there
are likely more transactions which are
subject to tax and consequently tax rev-
enue will increase. Our ndings are in
line with earlier research that established
a positive correlation among tax revenue
and the rate of ination Desta [29], Bouk-
bech et al. [53]. In light of our ndings, we
substantiate the formulated hypotheses,
suggesting that ination has a signicant
and positive impact on tax revenue.
In contrast to our proposed hypothe-
ses, our ndings revealed unexpected re-
sults that reveal a negative and signicant
relationship between FDI and tax revenue
in the long run and short run dynamics at
lag zero. However, the coefcient of FDI
at lag 1 is positive in the short run. The
negative result is not shocking given that
most developing nations offer foreign in-
vestors and exporters a specic set of tax
incentives. For instance, in Ethiopia any
foreign direct investment has three to ve
years corporate income tax incentive and
export free duty. As a result, we reject the
proposed hypotheses with regard to FDI
and tax revenue.
In addition, import goods by foreign
investors (machinery, cars, building ma-
terials, medical supplies) for investment
purpose are assumed excise tax and oth-
er charge free in Ethiopia. As a result, the
amount of tax that would have been paid
is declining as the volume of foreign direct
investment has occasionally increased.
In fact, the primary purpose of inviting
foreign direct investment in developing
countries particularly in Africa is to re-
solve their hard currency decit at the
expense of collecting tax from the invest-
ment. Based on our ndings, we reject our
proposed hypothesis. On the other hand,
the connection would eventually reverse
as a nation developed and the tax benets
were maintained.
Our ndings is similar with the stu-
dies by ALshribi [24] and contradict to
other ndings who evident positive re-
sult Camara [54], Ha et al. [55], Musah
et al. [56], Odabas [57], Pratomo [58].
The other important nding of the
present study is the share manufacturing
sector signicantly contributes to domes-
tic tax revenue in Ethiopia. As the share
of the sector to GDP rises by 1% there
will be increase in tax revenue by 3.417%
in the long run and 0.354% in short run.
Based on our ndings, we conclude that
the share of manufacturing in GDP signi-
cantly and positively impacts tax revenue,

Journal of Tax Reform. 2024;10(3):493–509
504
eISSN 2414-9497
and we accept the hypothesis accordingly.
It is notable that the more industrialized
a country the higher it has diversied in-
cluding direct and indirect tax revenues
(e.g. corporate tax, sales taxes, excise tax,
export duty, etc.).
However, since the sector is very sen-
sitive to global issues such as war, pan-
demic, climate change, political turmoil
which might collectively decline the sec-
tor’s contribution to tax revenue. Based
on our ndings, we accept our proposed
hypothesis. This result aligns with the
existing literature that documents the
manufacturing sector’s pivotal role in mo-
bilizing domestic revenue (Wulandari &
Wijaya [33]).
It is also clearly indicated in Table 6
that corruption perception (LNCP) has
a positive and strongly signicant re-
lationship with tax revenue born in the
short run and long run dynamic, which
contrasts with our hypotheses, leading us
to reject them. Indeed, a strong corruption
control system helps to reduce the mi-
suse of public resources and the number
of non-compliance taxpayers as well as
unlawful practices in an economy. On the
other hand, a country with a good corrup-
tion control would better provide public
services effectively and efciently which
foster trust between the government and
taxpayers and subsequently increase tax
revenues. In addition, effective corruption
perception improves tax administration
system and increases tax revenue. Based
on our ndings, we accept our proposed
hypothesis This view is consistent with
a dozen of prior studies like (Bogeti´c &
Naeher [59]).
One of the important ndings of the
present study is that Shadow economy
(LNSE), which is a proxy for tax evasion,
has a negative and statistically signicant
effect on tax revenue in the long run and
short run. Based on long term coefcient,
a 1% increase in shadow economy reduces
tax revenue by –6.819%. Based on this fact,
we have accepted the formulated hypo-
theses. Notably shadow economy erodes
public resources and leave shortage in
public nancing. It also expands pover-
ty among community by limiting public
services and investment in core economic
foundations.
More broadly, in least developing
countries shadow economy broadens dis-
trust between government and resource
providers which cause political unrest,
internal collision and war. It also creates
unfair income distribution and resource
allocation among societies. Based on our
ndings, we accept our proposed hypo-
thesis and this phenomena supported by
a number of studies like Dokas et al [34],
Gnangnon & Brun [48].
6. Conclusion
The current research mainly focuses
on examining the various factors inuen-
cing tax revenue in Ethiopia with dataset
ranging from 1992–2021. In this respect,
the present research deploys time series
regression analysis using ARDL-ECM. The
essential properties of the time series: sta-
tionery and co-integration tests conrmed.
The Unit Root test is done using the popu-
lar statistical approaches ADF and PP and
the result conrms the existence of statio-
nery for the analyzed variables but at dif-
ferent stages. The ARDL-ECM provides
long-term and short coefcients.
The result shows trade openness, in-
ation rate, GDP, share of manufacturing
sector and corruption perception found es-
sential factors in fostering tax revenue both
in the long run and short run dynamics.
Conversely, foreign direct investment
and shadow economy appeared to have
a negative relationship in the long run
and short run. The study found an insig-
nicant relationship between share of ag-
riculture in the long run, but a negative
signicant coefcient is found at lag 1.
Indicating past value of agriculture could
reduce the current tax revenue. One of the
possible reasons would be that the imme-
diate last value of agriculture would not
have signicant effect on tax revenue be-
cause the agricultural sector more contri-
butes to domestic revenue in the long run
than it does in the short run. With regards
to post analysis tests, the residuals of the
model are normally distributed, thus con-
rming the reliability of the model in pre-
dicting tax revenue.

Journal of Tax Reform. 2024;10(3):493–509
505
eISSN 2414-9497
We also tested serial autocorrelation
heteroscedastic problems in the model
and the result afrms the absence of the
assumptions. The test result of CUSUM
and CUSUM squares showed stability of
the model overtime, thereby further re-
inforcing the dependability of prediction
by the model. An error correction term
is negative and highly signicant, thus
afrming a long-run relationship among
variables.
Based on the empirical conclusion, the
research has inferred the following policy
implications in the context of recommen-
dation. The FDRE Ministry of revenue
should realign its tax policies and expand
its base to comprehend more agricultural
activities, especially those currently un-
taxed or undertaxed yet liable for the tax-
payers of the relevant commodity. A pro-
gressive tax system calls for large earners,
including big farming businesses, to pay
a greater share of their income. Corre-
spondingly, increasing the enforcement
and monitoring will reduce tax evasions
by strengthening the measures to reduce
tax evasion within the agricultural region.
By trimming very generous tax holi-
days and offering only performance-based
incentives, such as those on employment
or technology transfer or local sourcing,
and by investing in capacity-building that
should better monitor and enforce its tax
laws, in particular with respect to multi-
national corporations, much of the impact
on FDI tax revenue can be better managed
by Ethiopia to ensure that it indeed adds
positively to the economic development of
the country.
It has also provided incentives to en-
courage industrial park and Special Eco-
nomic Zones (SEZ) development, offering
attractive regulatory and scal incentives
in expectation of attracting manufacturing
investments to enhance the contribution
of the manufacturing sector to GDP and
the tax base of at least Ethiopia, as well
as promoting export diversication into
higher-value goods.
The research concludes that there
exists a capacity for other researchers to
extend this using further time periods and
alternative econometric models. Although
the model was very promising, investi-
gating it for a longer period and using
an alternative methodology, it placed too
much value in models such as the Vector
Auto-Regression (VAR). By so doing, re-
searchers would be better placed to un-
derstand the relationship between agri-
culture and tax revenue, which could lead
to efcient ways of making tax generation
effective and aid national development.
In this regard, future researchers would
be better placed to carry out inter-country
or groups of countries’ research based on
their income level since these focuses on
one country’s context.
Moreover, because this is a research
report that does not take every factor that
affects tax revenue simultaneously, fu-
ture research will be special in those fac-
tors that create obstacles to the nation’s
tax revenue. These include the scale of
aid services, urbanization, government
expenditures, government stability, good
governance indicators, institutional qua-
lity indexes, etcetera.
References
1. Kessy M., Sukartini N. Impacts of Taxation on Economic Growth in Africa in 2008–2018 –
Panel Data Analysis. Journal of Developing Economies. 2023;8:244–260. https://doi.org/10.20473/
jde.v8i2.43290
2. Balasoiu N., Chifu I., Oancea M. Impact of Direct Taxation on Economic Growth:
Empirical Evidence Based on Panel Data Regression Analysis at the Level of Eu Countries.
Sustainability. 2023;15(9):7146. https://doi.org/10.3390/su15097146
3. Cavalli F., Mainini A., Visetti D. The role of taxation in an integrated economic-envi-
ronmental model: a dynamical analysis. Decisions in Economics and Finance. 2024. https://doi.
org/10.1007/s10203-024-00449-x
4. Ajeigbe K.B, Ganda F., Enowkenwa R.O. Impact of sustainable tax revenue and
expenditure on the achievement of sustainable development goals in some selected African
countries. Environment, Development and Sustainability. 2024;26(10):26287–26311. https://doi.
org/10.1007/s10668-023-03730-y

Journal of Tax Reform. 2024;10(3):493–509
506
eISSN 2414-9497
5. Mascagni G., Mengistu A. Effective tax rates and rm size in Ethiopia. Development
Policy Review. 2019;37(S2):O248–O273. https://doi.org/10.1111/dpr.12400
6. Chelliah R.J. Trends in Taxation in Developing Countries. IMF Economic Review.
1971;18(2):254–331. https://doi.org/10.2307/3866272
7. Chelliah R.J., Baas H.J., Kelly M.R. Tax Ratios and Tax Effort in Developing Countries,
1969–71. IMF Economic Review. 1975;22(1):187–205. https://doi.org/10.2307/3866592
8. Mahdavi S. The level and composition of tax revenue in developing countries: Evidence
from unbalanced panel data. International Review of Economics & Finance. 2008;17(4):607–661.
https://doi.org/10.1016/j.iref.2008.01.001
9. Baunsgaard T., Keen M. Tax revenue and (or?) trade liberalization. Journal of Public
Economics. 2010;94(9-10):563–577. https://doi.org/10.1016/j.jpubeco.2009.11.007
10. Profeta P., Scabrosetti S. The Political Economy of Taxation: Lessons from Developing Countries.
Cheltenham, UK: Edward Elgar Publishing; 2010. 208 p. https://doi.org/10.4337/9781849805490
11. Castro G.Á., Camarillo D.B.R. Determinants of tax revenue in OECD countries over
the period 2001–2011. Contaduría y Administración. 2014;59(3):35–59. https://doi.org/10.1016/
S0186-1042(14)71265-3
12. Delessa K., Teera J. Determinants of tax revenue in East African countries: An appli-
cation of multivariate panel data cointegration analysis. Journal of Economics and International
Finance. 2018:10(11):134–155. https://doi.org/10.5897/JEIF2018.0924
13. Castañeda Rodríguez V.M. Tax determinants revisited. An unbalanced data panel
analysis. Journal of Applied Economics. 2018;21(1):1–24. https://doi.org/10.1080/15140326.
2018.1526867
14. Piancastelli M., Thirlwall A.P. The Determinants of Tax Revenue and Tax Effort in
Developed and Developing Countries: Theory and New Evidence 1996–2015. Nova Economia.
2020;30:871–892. https://doi.org/10.1590/0103-6351/5788
15. Raouf E. The impact of nancial inclusion on tax revenue in EMEA countries:
A threshold regression approach. Borsa Istanbul Review. 2022;22(6):1158–1164. https://doi.
org/10.1016/j.bir.2022.08.003
16. Chettri K.K., Bhattarai J.K., Gautam R. Determinants of Tax Revenue in South Asian
Countries. Global Business Review. 2023. https://doi.org/10.1177/09721509231177784
17. Todorovi´c J.Ð., Ðordevi´ M., Mirovi´c V., Kalaš B., Pavlovi´c N. Modeling Tax Revenue
Determinants: The Case of Visegrad Group Countries. Economies. 2024;12(6):131. https://doi.
org/10.3390/economies12060131
18. Gupta A.S. Determinants of Tax Revenue Efforts in Developing Countries. IMF Working
Papers No. 07/184. 2007. 39 p. https://doi.org/10.5089/9781451867480.001
19. Kawadia G., Suryawanshi A.K. Tax Effort of the Indian States from 2001–2002 to
2016–2017: A Stochastic Frontier Approach. Millennial Asia. 2023;14(1):85–101. https://doi.
org/10.1177/09763996211027053
20. Ayenew W. Determinants of Tax Revenue in Ethiopia (Johansen Co-Integration Ap-
proach). International Journal of Business, Economics and Management. 2016;3(6):69–84. https://
doi.org/10.18488/journal.62/2016.3.6/62.6.69.84
21. Gobachew N., Debela K.L., Shibiru W. Determinants of Tax Revenue in Ethiopia.
Economics. 2017;6(6):58–64. https://doi.org/10.11648/j.eco.20170606.11
22. Desta T., Reta M., Girma B. Factors Affecting Tax Revenue in Ethiopia: Autoregressive
Distributed Lag Approach. Preprint (Version 1). Research Square; 4 April 2022. 23 p. https://doi.
org/10.21203/rs.3.rs-1484380/v1
23. Ojong C., Ogar A., Oka, Arikpo O. The Impact of Tax Revenue on Economic Growth:
Evidence from Nigeria. IOSR Journal of Economics and Finance (IOSR-JEF). 2016;7:32–38.
https://doi.org/10.9790/5933-07113238
24. ALshubiri F. Do foreign direct investment inows affect tax revenue in developed
and developing countries? Asian Review of Accounting. 2024. https://doi.org/10.1108/ARA-
06-2023-0155
25. Mmbulaheni A, Nyamazunzu Z, Dagume M. Determinants of tax revenue performance
in South Africa for the period 1990–2018. International Journal of Applied Economics, Finance and
Accounting. 2024;19(1):82–96. https://doi.org/10.33094/ijaefa.v19i1.1515
26. Serin S, Demir M. The Effect of Foreign Direct Investments on Corporate Tax Reve-
nues: An Empirical Analysis for OECD Countries. Kafkas Üniversitesi İktisadi ve İdari Bilimler
Fakültesi Dergisi. 2023;14(27):223–248. https://doi.org/10.36543/kauiibfd.2023.009
27. Rahman Md.M., Islam M.E. Does Trade Openness Affect Taxation? Evidence from
BRICS Countries. Millennial Asia. 2023. https://doi.org/10.1177/09763996231199310

Journal of Tax Reform. 2024;10(3):493–509
507
eISSN 2414-9497
28. Mirović V., Kalaš B., Milenkovic N., Andrašić J. Different modelling approaches of tax
revenue performance: The case of Baltic countries. Ekonomie a Management. 2023;26(3):20–32.
https://doi.org/10.15240/tul/001/2023-3-002
29. Desta T., Reta M., Girma B. Determinants of Tax Revenue in Ethiopia: Autoregressive
Distributed Lag Approach. Ethiopian Journal of Business and Social Science. 2023;6(2):1–30. https://
doi.org/10.59122/154F59lk
30. Negash T. Determinants of Tax Evasion: Evidence from Hawassa City Taxpayers. Pre-
print (Version 1). Research Square; 19 September 2023. 15 p. https://doi.org/10.21203/
rs.3.rs-3331474/v1
31. Caballé J., Panadés J. Ination, tax evasion, and the distribution of consumption. Journal
of Macroeconomics. 2004;26(4):567–595. https://doi.org/10.1016/j.jmacro.2003.06.001
32. Shrestha S., Kotani K., Kakinaka M. The relationship between trade openness and go-
vernment resource revenue in resource-dependent countries. Resources Policy. 2021;74:102332.
https://doi.org/10.1016/j.resourpol.2021.102332
33. Wulandari D., Wijaya S. Agriculture and Manufacturing for Lower Middle-Income
Countries: Exchange rate as a Moderating Variable. Ilomata International Journal of Tax and
Accounting. 2023;4(3):508–523. https://doi.org/10.52728/ijtc.v4i3.798
34. Dokas I., Panagiotidis M., Papadamou S., Spyromitros E. The impact of the shadow
economy on the direct-indirect tax mix: Can central banks’ independence mitigate the effect?
Journal of Policy Modeling. 2024;46(3):475–493. https://doi.org/10.1016/j.jpolmod.2024.03.001
35. Gnangnon S.K. Effect of the Shadow Economy on Tax Reform in Developing Countries.
Economies. 2023;11(3):96. https://doi.org/10.3390/economies11030096
36. Vlachaki M. The Impact of the Shadow Economy on Indirect Tax Revenues. Economics &
Politics. 2015;27(2):234–265. https://doi.org/10.1111/ecpo.12054
37. Ajaz T., Ahmad E. The Effect of Corruption and Governance on Tax Revenues. The
Pakistan Development Review. 2020;49(4II):405–417. https://doi.org/10.30541/v49i4IIpp.405-417
38. Abdou D., Atya E. Investigating the energy-environmental Kuznets curve: Evi-
dence from Egypt. International Journal of Green Economics. 2013;7(2):103–115. https://doi.
org/10.1504/IJGE.2013.057436
39. Agiakloglou C., Newbold P. Empirical evidence on dickey-fuller-type tests. Journal of
Time Series Analysis. 1992;13(6);471–483. https://doi.org/10.1111/j.1467-9892.1992.tb00121.x
40. Schwert G.W. Why Does Stock Market Volatility Change Over Time? The Journal of
Finance. 1989;44(5):1115–1153. https://doi.org/10.1111/j.1540-6261.1989.tb02647.x
41. Wooldridge J. Introductory Econometrics: A Modern Approach. 5th ed. South-Western,
Cengage Learning; 2013. 881 p. Available at: https://cbpbu.ac.in/userles/le/2020/STUDY_
MAT/ECO/2.pdf (accessed: 03.07.2024).
42. Adedeji A.A., Oyinlola M.A., Adeniyi O. Public debt, tax and economic growth in
Sub-Saharan African countries. Journal of Social and Economic Development. 2023. https://doi.
org/10.1007/s40847-023-00295-4
43. Alabede J.O. Economic freedom and tax revenue performance in sub-Saharan Africa.
Journal of Financial Reporting and Accounting. 2018;16(4):610–638. https://doi.org/10.1108/
JFRA-04-2017-0024
44. Okunogbe O., Santoro F. Increasing Tax Collection in African Countries: The Role
of nformation Technology. Journal of African Economies. 2023;32(Suppl. 1):i57–i83. http://doi.
org/10.1093/jae/ejac036
45. Gebissa W.Y. The challenges and prospects of Ethiopian agriculture. Cogent Food &
Agriculture. 2021;7(1):1923619. https://doi.org/10.1080/23311932.2021.1923619
46. Ghouse G., Khan S.A., Rehman A.U., Bhatti M.I. ARDL as an Elixir Approach to Cure
for Spurious Regression in Nonstationary Time Series. Mathematics. 2021;9(22):2839. https://
doi.org/10.3390/math9222839
47. Jemiluyi O.O., Jeke L. Tax revenue mobilization effort in Southern African Development
Community (SADC) bloc: Does ICT matter? Cogent Economics & Finance. 2023;11(1):2172810.
https://doi.org/10.1080/23322039.2023.2172810
48. Gnangnon S.K, Brun J.-F. Trade Openness, Tax Reform and Tax Revenue in Developing
Countries. The World Economy. 2019;42(12):3515–3536. https://doi.org/10.1111/twec.12858
49. Tien H., Tran Xuan H., Nguyen Q.K. Tax revenue-economic growth relationship
and the role of trade openness in developing countries. Cogent Business & Management.
2023;10(2):2213959. https://doi.org/10.1080/23311975.2023.2213959
50. Afonso A., Jalles J.T., Venâncio A. Structural Tax Reforms and Public Spending Efciency.
Open Economies Review. 2021;32(5):1017–1061. https://doi.org/10.1007/s11079-021-09644-4

Journal of Tax Reform. 2024;10(3):493–509
508
eISSN 2414-9497
51. Alinaghi N., Reed W.R. Taxes and Economic Growth in OECD Countries: A Meta-
analysis. Public Finance Review. 2021;49(1):3–40. https://doi.org/10.1177/1091142120961775
52. Arnold J.M., Brys B., Heady C., Johansson Å., Schwellnus C., Vartia L. Tax Policy for
Economic Recovery and Growth. The Economic Journal. 2011;121(550):F59–F80. https://doi.
org/10.1111/j.1468-0297.2010.02415.x
53. Boukbech R., Bousselhami A., Ezzahid E. Determinants of Tax Revenues: Evidence From
a Sample of Lower Middle Income Countries. Applied Economics and Finance. 2019;6(1):11–20.
https://doi.org/10.11114/aef.v6i1.3280
54. Camara A. The Effect of Foreign Direct Investment on Tax Revenue. Comparative Eco-
nomic Studies. 2023;65(1):168–190. https://doi.org/10.1057/s41294-022-00195-2
55. Minh Ha N., Tan Minh P., Binh Q.M.Q. The determinants of tax revenue: A study of
Southeast Asia. Cogent Economics & Finance. 2022;10(1):2026660. https://doi.org/10.1080/2332
2039.2022.2026660
56. Musah A., Kodjie P.K., Abdulai M. Short-term and long-term effects of foreign direct
investment on tax revenue: empirical evidence from an emerging economy. Journal of Humani-
ties and Applied Social Sciences. 2024;6(4):331–344. https://doi.org/10.1108/JHASS-08-2023-0099
57. Odabas H. Foreign direct investment inows on tax revenues in the transition
economies of European Union. New Trends and Issues Proceedings on Humanities and Social
Sciences. 2016;2(2):17–22. https://doi.org/10.18844/gjhss.v2i2.408
58. Pratomo A.W. The effect of foreign direct investment on tax revenue in developing
countries. Jurnal BPPK Badan Pendidikan dan Pelatihan Keuangan. 2020;13(1):83–95. https://doi.
org/10.48108/jurnalbppk.v13i1.484
59. Bogetić Ž., Naeher D. Corruption and government revenue: Evidence of a non-linear
relationship driven by crises. Journal of Applied Economics. 2024;27(1):2295733. https://doi.org/
10.1080/15140326.2023.2295733
Acknowledgments
The authors would like to thank Prof. Igor A. Mayburov (Editor-in-Chief) and anonymous
reviewers for their time and effort devoted to critical review and giving helpful as well as
constructive comments. The authors received no nancial support for the research.
Information about the authors
Tekalign Negash Kebede – MSc. Lecturer, Department of Accounting and Finance, College of Busi-
ness and Economics, Hawassa University, Hawassa, Ethiopia (P.O.Box 05, Hawassa, Sidama
Region, Ethiopia); https://orcid.org/0000-0001-5600-5546; e-mail: tekanega@gmail.com
Obsa Teferi Erena – MSc, Assistant Professor. Department of Accounting and Finance, College
of Business and Economics, Hawassa University, Hawassa, Ethiopia (P.O.Box 05, Hawassa,
Sidama Region, Ethiopia); https://orcid.org/0000-0003-4304-5359; e-mail: obteferi@gmail.com
Elias Pawulos Bawiso – MSC, Senior Tax Expert, Sidama Regional State Revenue Ofces, Leku
Branch, Leku, Ethiopia (Shebedino Woreda, Leku town, Sidama Ethiopia); https://orcid.
org/0009-0000-0219-0969; e-mail: elapaulos14@gmail.com
For citation
Kebede T.N., Erena O.T., Bawiso E.P. Determinants of Tax Revenue: A Cointegration and
Causality Analysis for Ethiopia, 1992–2022. Journal of Tax Reform. 2024;10(3):493–509. https://
doi.org/10.15826/jtr.2024.10.3.180
Article info
Received July 31, 2024; Revised October 2, 2024; Accepted October 15, 2024
Благодарности
Авторы благодарят проф. И.А. Майбурова (главного редактора) и анонимных рецензен-
тов за их время и усилия, потраченные на критический обзор и предоставление полезных
и конструктивных комментариев. Авторы не получали финансовой поддержки для ис-
следования.

Journal of Tax Reform. 2024;10(3):493–509
509
eISSN 2414-9497
Информация об авторах
Текалигн Негаш Кебеде – магистр, преподаватель кафедры бухгалтерского учета и финан-
сов, Колледж бизнеса и экономики Университета Хавасса, г. Хавасса, Эфиопия (P.O.Box 05,
Hawassa, Sidama Region, Ethiopia); https://orcid.org/0000-0001-5600-5546; e-mail: tekanega@
gmail.com
Обса Тефери Эрена – магистр, доцент, факультет бухгалтерского учета и финансов, Кол-
ледж бизнеса и экономики, Университет Хавасса, г. Хавасса, Эфиопия (P.O.Box 05, Hawassa,
Sidama Region, Ethiopia); https://orcid.org/0000-0003-4304-5359; e-mail: obteferi@gmail.com
Элиас Павлос Бависо – MSC, старший налоговый эксперт, Региональное управление государ-
ственных доходов Сидамы, филиал Леку, г. Леку, Эфиопия (Shebedino Woreda, Leku town,
Sidama Ethiopia); https://orcid.org/0009-0000-0219-0969; e-mail: elapaulos14@gmail.com
Для цитирования
Kebede T.N., Erena O.T., Bawiso E.P. Determinants of Tax Revenue: A Cointegration and
Causality Analysis for Ethiopia, 1992–2022. Journal of Tax Reform. 2024;10(3):493–509. https://
doi.org/10.15826/jtr.2024.10.3.180
Информация о статье
Дата поступления 31 июля 2024 г.; дата поступления после рецензирования 2 октября
2024 г.; дата принятия к печати 15 октября 2024 г.