3. 1. INTRODUCTION
• Kassim and Mansour (2016) define tax expenditures as
the amount of revenue forgone through the application of
special tax regimes or tax provisions that are intended to
promote and encourage a particular sector, activity or
taxpayer.
• Include: exemptions, deductions, credits, rate reliefs,
deferrals.
4. 2. MOTIVATION AND OBJECTIVE OF THE
RESEARCH
• Tax to GDP ratio in Rwanda within the last five fiscal
years was 13.41%, 13.88%, 14.83%, 15.26% and
16.07% respectively.
• The contribution of tax to Government budget was
42.1%, 45.5%, 48.7%, 54.5% and 55.6% respectively.
• The ratio of tax expenditure to GDP was 8.73%, 8.80%,
8.35%, 7.76% and 7.86% respectively.
• A report from Tax Justice Network-Africa and ActionAid
criticizes the many tax incentives and exemptions that
governments in East Africa provide.
5. 2. MOTIVATION AND OBJECTIVE OF THE
RESEARCH CONT’D
• The report states that the money lost from tax
expenditures can be mind-boggling.
• The objective of this research was to assess the impact
of tax expenditure on tax revenue performance in
Rwanda.
6. • The study occupied a causal research with the aim of
examining the responsiveness of tax revenue
performance with respect to changes in tax
expenditures.
• According to Sekaran and Bougie (2010) a causal study
is done when it is necessary to establish a definite cause
and effect relationship.
• Zikmund et al,.(2010) observed that causal research
allows causal inferences to be made and that causal
inference is a conclusion that when one thing happens,
another specific thing will follow.
7. • The study involved six variables:
Tax revenues,
VAT exemptions (Customs taxes),
VAT exemptions (Domestic taxes),
Import duty exemptions,
Excise duty exemptions; and
Investment allowances.
• The study used secondary data on monthly basis from
2008 to 2015.
• The research applied econometric analytical methods.
8. 3. METHODOLOGY CONT’D
• Unit Root Test
• If one wishes to use hypothesis tests, either singly or
jointly, to examine the statistical significance of the
coefficients, then it is essential that all of the components
in the VAR are stationary, Chris Brooks (2014).
• If a time series is non-stationary, the regression analysis
done in a traditional way will produce spurious results,
Shrestha et Al (2005).
• This paper applied Augmented Dickey-Fuller Unit Root
Test.
9. 3. METHODOLOGY CONT’D
• Autoregressive Distributed Lag (ADRL) Model
• According to Dritsakis the ARDL approach has the
advantage that it does not require all variables to be I(1)
as the Johansen framework and it is still applicable if we
have I(0) and I(1) variables in our set.
• Furthermore, it can distinguish dependent and
explanatory variables, and allows to test for the
existence of relationship between the variables.
• Finally, with the ARDL it is possible that different
variables have differing optimal number of lags.
11. • Unit root test
• Four variables were found stationary at levels:
VAT exemptions (Customs taxes),
Import duty exemptions,
Excise duty exemptions; and
Investment allowances.
• Two variables were found stationary at first difference:
Tax revenues; and
VAT exemptions (Domestic taxes).
12. 4. EMPIRICAL RESULTS CONT’D
• Test for serial correlation and heteroscedasticity
Breusch-GodfreySerialCorrelationLMTest:
F-statistic 0.104628 Prob.F(2,82) 0.9008
Obs*R-squared 0.236722 Prob.Chi-Square(2) 0.8884
HeteroskedasticityTest:White
F-statistic 1.082205 Prob.F(44,48) 0.3933
Obs*R-squared 46.31374 Prob.Chi-Square(44) 0.3770
ScaledexplainedSS 50.96799 Prob.Chi-Square(44) 0.2186
13. 4. EMPIRICAL RESULTS CONT’D
• ARDL Bounds Test for cointegration and stability of the
model
-30
-20
-10
0
10
20
30
40 45 50 55 60 65 70 75 80 85 90
CUSUM 5% Significance
F-Bounds Test Null Hypothesis: No levels relationship
Test Statistic Value Signif. I(0) I(1)
F-statistic 16.89957 10% 2.08 3
k 5 5% 2.39 3.38
2.5% 2.7 3.73
1% 3.06 4.15
15. 4. EMPIRICAL RESULTS CONT’D
• Granger causality
Null Hypothesis: Obs F-Statistic Prob.
LOGVATEXEMPTIONSDTD does not Granger Cause LOGTAXREVENUE 90 4.96337 0.0032
LOGTAXREVENUE does not Granger Cause LOGVATEXEMPTIONSDTD 4.97453 0.0032
Null Hypothesis: Obs F-Statistic Prob.
LOGEXCISEDUTYEXEMPTIONS does not Granger Cause LOGTAXREVENUE 90 0.77525 0.5111
LOGTAXREVENUE does not Granger Cause LOGEXCISEDUTYEXEMPTIONS 0.72067 0.5424
16. Conclusion
• Three variables were found statistically significant:
VAT exemptions on domestic taxes (1% : -0.36%),
excise duty exemptions; and
investment allowances (1% : -0.12%).
• Two variables were found not statistically significant:
VAT exemptions in customs; and
import duty exemptions.
17. 5. CONCLUSION AND POLICY IMPLICATION
CONT’D
Policy implication
• In order to raise more revenue for public services such as
health, education and infrastructure the Government can
reduce VAT exemptions on domestic taxes and investment
allowances.