Transcript Policy Coordination to Strengthen Regional Integration

THIRD ANNUAL CONFERENCE ON REGIOAL INTEGRATION IN
AFRICA (ACRIA 3)
Hotel Fleur de Lys, Almadies, Dakar, Senegal
Theme: Policy Coordination to Strengthen
Regional Integration
July 4-5, 2012
TAX EFFORT IN ECOWAS COUNTRIES
M. Ben Umar Ndiaye & Robert D. Korsu
(West African Monetary Agency)
1
OUTLINE OF THE PRESENTATION
1. INTRODUCTION
2. LITERATURE REVIEW
3. METHODOLOGY
4. EMPIRICAL RESULTS
5. RECOMMMENDATIONS
2
1. INTRODUCTION
Governments provide goods and services:
- public goods, merit goods and social
services
(Education, Health, Physical
infrastructure, Electricity, Water etc.)
- Hence, it makes expenditure: Investment
and consumption
- It therefore requires revenue: Domestic
and Foreign
Domestic Resource Mobilisation: Tax and
3
Non-Tax Revenue
1.INTRODUCTION
• EMCP: The Macroeconomic Convergence
Criteria ( 4 Primary and 6 Secondary Criteria)
• Macroeconomic Convergence Criterion
Relating to Tax Revenue: Tax Revenue to be
at least 20 % of GDP
• Satisfaction of the Criterion from 2001 to 2010
: Gambia (2004), Cape Verde (2005 to 2010).
Liberia (2009,2010). Ghana– good
performance under the old GDP but poor
performance with the new GDP
• Satisfaction of the Criterion: Challenging
4
1.INTRODUCTION
• Fiscal Deficit, Tax Revenue and Tax Effort: The Link
- A country operating above tax capacity : Requires
Expenditure Rationalisation for deficit reduction
- A country operating below tax capacity: Requires Tax
Reforms to increase revenue
How do we know whether a country is above or below
the taxable capacity?
Through Estimation of Tax Effort ( The extent to which a
country can exploit its tax potential)
5
1.INTRODUCTION
•
Objectives
i.
To investigate the determinants of tax
revenue in the ECOWAS Countries
ii. To determine the tax effort of the
ECOWAS Countries
6
2. LITERATURE REVIEW
THEORETICAL CONCEPT
• Tax Effort was introduced by Lotz and Morss
(1967), Bahl (1971).
When:
• Tax Effort > 100 %, the country is above tax
potential. In which case, there is difficulty in
mobilising additional resources using tax policy
• Tax Effort < 100 %, the country is below tax
potential In which case, it is not difficulty in
mobilising additional resources using tax policy
• Tax Effort =100 %, the country is at its tax potential
7
2. LITERATURE REVIEW
THEORETICAL CONCEPT
• Tax Effort is a model based concept
• It is actual tax ratio ( Tax-GDP ratio)
divided by the predicted tax ratio ( from a
model of tax ratio).
• Hence, it depends on the model applied
8
2. LITERATURE REVIEW
EMPIRICAL LITERATURE
Empirical Approach
 Focus estimation of a tax share function and determination
of tax efforts using actual tax ratio and the estimated model (
For example, Teera (2002, Piancastelli,2001, Bahl, 2003,
Stotsky and WoldeMariam 1997, Ahsan and Wu, 2005, IMF
(2011)
 Common Data Type: cross-section and time series data
 Common Estimation Method: Panel Data TechniquePooled, Fixed and Random Effects
 Recent but Uncommon Method: Stochastic Frontier
(Pessino and Fenochietto (2010) and IMF (2011)
9
2. LITERATURE REVIEW
Departure of this study from previous
studies
 By focusing on the ECOWAS countries but considers some nonECOWAS countries in the estimation.

o
o
o
o
o
Moreover tax effort is considered for the various tax categories:
Direct Tax
Indirect Tax
Trade Tax
Total tax with natural resource related tax
Total tax without natural resource related tax
 Use of the Stochastic Frontier instead of Panel Data Regression
10
3. METHODOLOGY
• A two-step Procedure
(i) Estimation of Tax Revenue Function
(ii) Construction of Tax Effort Index from (i)
• (i) Estimation of Tax Revenue Function
Application of the stochastic frontier production functiondeveloped by Aigner, Lovell and Schmidt (1977).
Initially developed for measuring inefficiency in production
and cost.
Recently extended to tax function (Pessino and
Fenochietto (2010) and IMF (2011)
11
3. METHODOLOGY
The Stochastic Frontier function as applied to production is given as
Yit  exp( 0   Xit  Vit  Uit )
i = 1,2,3…………. N:
t = 1,2,3………T
Where:
Y is the output variable
X is the vector of input variables
 is a vector of parameters to be estimated
V is the stochastic disturbance term with zero mean, constant
variance and is normally distribute
U follows a truncated normal distribution with constant mean and
variance. U is a measure of technical inefficiency.
12
3. METHODOLOGY
• In the context of tax function: Y is tax ratio (tax-GDP ratio)
and X is a vector of determinants of tax revenue.
In accordance with the Tax Literature ( for example, Pessino and
Fenochietto (2010), IMF (2011), Teera (2004), Begun (1987), Tanzi
(1978) and Bahl (1971), the components of X are: GDPPC, OPNM,
AGRICS, URB, M2/GDP,LIR,INF
• Other variables maybe included in X but empirical studies
do not include all the possible determinants (e.g Gini
Coefficient which is a measure of income inequality, a
measure corruption, a measure of debt overhang etc.
13
3. METHODOLOGY
Tax Revenue Function is calculated for 5 tax
categories: direct, indirect, trade, total tax
(with and without natural resource related
taxes)
(ii) Construction of Tax Effort Index (It ranges
between 0 and 1 (i.e. 0 % and 100 %)
• Tax Effort of country i at time t is given as the ratio
of actual tax ratio to predicted tax ratio. That is:
TAXEFFit 
Yit
exp(  0  Xit Vit )

exp(  0  Xit Vit Uit )
exp(  0  Xit Vit )
 exp(Uit )
14
3. METHODOLOGY
• The tax functions were estimated using data on
all the 15 ECOWAS countries and 5 nonECOWAS sub-Sahara African countries(SSA).
These are: Botswana, Kenya, Namibia, South
Africa and Zambia) over the period 2000 to
2010.
• The choice of the 5 non-ECOWAS -SSA
countries was predicated on best tax
performance over the period 1991-2006 based
on IMF (2011)
• Data Sources: WDI, ADI, AEO and IFS
15
3. METHODOLOGY
• ESTIMATION TECHNIQUE
• General to specific method: dropping of insignificant
variables to arrive at the parsimonious model
• Both Battese and Coellie half normal (BCHN) and
general truncated normal (BCTN) versions were
estimated
• Use of Log-likelihood to decide the BCHN vs BCTN
• Use of LR test for presence of inefficiency- in the BCHN
frontiers
• Use of Normal distribution (Z) test for presence of
inefficiency- in the BCTN frontiers
16
4. EMPIRICAL RESULTS
1.The Stochastic Frontier Tax Function: Direct Tax
Variables
Constant
Battese Coelli
Half
Coefficient
P-Value
Battese Coelli
Truncated
Coefficient
P-Value
0.129
0.593
0.098
0.674
-
-
-
-
- 0.203
0.000
- 0.234
0.000
Ln ( GDPPC)
-
-
-
-
Ln (URB)
-
-
-
-
Ln ( OPN)
-
-
-
-
0.626
0.000
0.632
0.000
Ln( M2/GDP)
Ln (AGS)
Ln(LIR)
Log Likelihood = - 82.55
Log Likelihood = - 76.56
Likelihood-ratio test for no
inefficiency
Chi squared = 42.16 (0.000)
Z test for no inefficiency
Z= - 5.73 (0.000)
17
4. EMPIRICAL RESULTS
2.The Stochastic Frontier Tax Function: Indirect Tax
Battese Coelli
Half
Variables
Battese Coelli
Truncated
Coefficient
P-Value
Coefficient
Constant
1.941
0.000
1.945
0.000
Ln( M2/GDP)
0.118
0.000
0.115
0.000
Ln(AGS)
-
-
-0.002
0.000
Ln ( GDPPC)
-
-
-
-
Ln(URB)
-
-
-
-
Ln ( OPN)
-
-
-
-
0.143
0.000
0.142
Ln(LIR)
P-Value
0.000
Log Likelihood = - 190.174
Log Likelihood = - 187.310
Likelihood-ratio test for no inefficiency
chi-squared: 34.88 (0.000)
Z test for no inefficiency:
Z= -2.05 (0.020)
18
4. EMPIRICAL RESULTS
3.The Stochastic Frontier Tax Function: Trade Tax
Variables
Constant
Battese Coelli
Half
Coefficient
P-Value
Battese Coelli
Truncated
Coefficient
P-Value
- 0. 162
0.000
-0.162
0.000
0.158
0.000
0.157
0.000
Ln(AGS)
-
-
-
-
Ln(URB)
-
-
-
-
Ln ( GDPPC)
0.059
0.000
0.059
0.000
Ln ( OPN)
0.324
0.000
0.324
0.000
Ln(LIR)
0.316
0.000
0.315
0.000
Ln( M2/GDP)
Log Likelihood = - 219.26
Log Likelihood = - 216.26
Likelihood-ratio test for no
inefficiency
chi squared=88.70 (0.000)
Z test for no inefficiency:
Z=-14.71 (0.000)
19
4. EMPIRICAL RESULTS
4.The Stochastic Frontier Tax Function: Total Tax Including
Natural Resource Related Taxes
Battese Coelli
Half
Coefficient
P-Value
Battese Coelli
Truncated
Coefficient
P-Value
Constant
1.365
0.000
1.994
0.000
Ln( M2/GDP)
0.187
0.000
0.201
0.000
Ln (AGS)
-
-
-
-
Ln ( GDPPC)
-
-
-
-
Ln (URB)
0.102
Ln ( OPN)
Ln(LIR)
0.076
-
0.388
0.118
-
0.000
0.042
-
0.368
0.000
Log Likelihood = - 14.581
Log Likelihood = - 19.29
Likelihood-ratio test for no
inefficiency:
Chi-squared = 9.94 (0.001)
Z test for no inefficiency:
Z= -5.71(0.000)
20
4. EMPIRICAL RESULTS
5.The Stochastic Frontier Tax Function: Total Tax
Excluding Natural Resource Related Taxes
Battese Coelli
Half
Coefficient
Constant
Battese Coelli
Truncated
P-Value
- 0.095
Coefficient
0.000
P-Value
0.462
0.000
Ln( M2/GDP)
0.012
0.000
0.061
0.000
-
-
-
-
Ln ( GDPPC)
0.079
0.000
0.049
0.000
Ln (URB)
0.034
0.000
-
-
Ln ( OPN)
0.149
0.000
0.054
0.000
Ln(LIR)
0.529
0.000
0.569
0.000
Ln (AGS)
Log Likelihood = - 57.15
Log Likelihood = - 56.74
Likelihood-ratio test for no inefficiency: Chi- Z test for no inefficiency: Z= -11.478 (0.000)
squared
= 64.86 (0.000)
21
4. EMPIRICAL RESULTS
6. Tax Efforts of the Countries: 2000-2010
Direct Tax
0.82
0.78
0.73
Indirect Tax
0.33
0.67
0.32
Trade Tax
0.86
0.30
0.43
Total Tax
Excluding
Natural
Resource
Related Tax
0.81
0.78
0.54
Guinea Bissao
Mali
Niger
Senegal
Togo
Liberia
Nigeria
Gambia
Ghana
Guinea
Sierra Leone
Cape Verde
Botswana
0.53
0.79
0.85
0.77
0.63
0.77
0.41
0.87
0.82
0.83
0.78
0.71
0.28
0.13
0.24
0.28
0.92
0.11
0.36
0.15
0.38
0.86
0.55
0.25
0.58
0.25
0.31
0.81
0.78
0.27
0.51
0.70
0.17
0.58
0.36
0.31
0.58
0.33
0.46
0.35
0.80
0.86
0.84
0.47
0.68
0.24
0.79
0.75
0.72
0.62
0.61
0.41
0.78
0.86
0.81
0.82
0.78
0.86
0.82
0.85
0.78
0.81
0.82
0.80
0.84
Kenya
Namibia
South Africa
Zambia
0.84
0.90
0.88
0.87
0.75
0.56
0.77
0.21
0.15
0.75
0.06
0.66
0.59
0.88
0.70
0.65
0.80
0.82
0.80
0.82
Country
Benin
Burkina Faso
Cote D’Ivoire
UEMOA
WAMZ
CAPE VERDE
NON-ECOWAS
Total Tax
Including
Natural
Resource
Related Tax
0.84
0.77
0.85
22
4. EMPIRICAL RESULTS
7. Tax Efforts of the Countries: 2010
UEMOA
WAMZ
CAPE VERDE
NON-ECOWAS
Direct Tax
Indirect
Tax
Trade Tax
Total Tax
Excluding
Natural
Resource
Related Tax
Total Tax
Including
Natural
Resource
Related Tax
Benin
0.88
0.38
1.00
0.97
0.88
Burkina Faso
0.80
0.71
0.33
0.80
0.81
Cote D’Ivoire
0.77
0.36
0.45
0.58
0.83
Guinea Bissao
0.21
-
0.04
0.07
0.80
Mali
0.82
0.24
0.58
0.65
0.81
Niger
0.91
0.52
0.60
1.00
0.95
Senegal
0.90
0.82
0.23
0.84
0.82
Togo
0.50
0.06
0.49
0.41
0.82
Liberia
0.91
0.42
1.00
0.86
0.85
Nigeria
0.55
0.20
0.09
0.25
0.81
Gambia
0.89
0.93
0.02
0.71
0.88
Ghana
0.95
1.00
0.48
1.00
0.83
Guinea
0.94
0.67
0.40
0.97
0.83
Sierra Leone
0.80
0.26
0.31
0.49
0.86
Cape Verde
0.66
0.90
0.23
0.57
0.82
Botswana
0.21
0.28
0.45
0.39
0.89
Kenya
0.94
0.92
0.14
0.75
0.69
Namibia
0.88
0.48
0.81
0.84
0.82
South Africa
0.90
0.75
0.07
0.73
0.91
Zambia
0.87
0.23
0.65
0.64
0.82
23
5. RECOMMENDATIONS
 Guinea Bissau, Togo and Nigeria: to enhance direct tax
revenue mobilisation through improved tax administration as
they have low tax effort on direct tax
 ECOWAS countries: Require revision of procedure for
mobilising indirect tax. A review of VAT administration
procedure is important as VAT takes a chunk of indirect tax.
VAT administration in Ghana and Senegal could be used as
reference as these two countries have high tax efforts
on indirect tax
 The ECOWAS countries require efforts to improve trade tax
mobilisation. This is more important in Burkina Faso,
Guinea Bissau, Ghana, Guinea, Nigeria and Cape Verde
as they are below potential by 70 %. However, Benin,
Mali, Niger and Liberia are not candidates for this policy
24
consideration since they have high trade tax efforts
5. RECOMMENDATIONS
 Guinea Bissau, Togo and Nigeria require shifting tax
administration from emphasis on natural resource related
taxes to other tax types
 Continued encouragement of the banking sector for making
payments by the private sector as financial depth has
positive effect on indirect and trade tax revenue
 Emphasis on policies that would encourage the development
of the agricultural sector so that it becomes an easy-to-tax
sector- for example, developing the
transformation of
agricultural products to industrial products. This flows from
the fact that agricultural share of GDP has a negative
effect on direct and indirect tax
25
5. RECOMMENDATIONS
 As ECOWAS countries impose taxes on imports, a policy of
no non-tariff barriers should be encouraged and maintained,
except for health, social and security reasons. This flows
from the fact that openness of the ECOWAS economies
has a positive effect on trade tax revenue
 Effort at improving literacy rates should be sustained as
literacy rate has positive effect on all the tax types
considered
 Strengthening of supply-side policies to improve economic
growth. This includes: investment in physical capital ( roads,
electricity etc.) and health. This derives from the positive
effect of per capita GDP on tax revenue
26
THE END
• THANK YOU FOR YOUR
ATTENTION
27