Transcript Tax Potential vs. Tax Effort

Tax Potential vs. Tax Effort:
Factors Behind the Stubbornly Low
Tax Collection in Armenia
Hamid R. Davoodi and David A. Grigorian
Armenia: Challenges of Sustainable Development
Conference in Honor of Professor Armen Alchian
UCLA, May 6, 2006
Outline
I. Overview
II. Armenia’s Tax Performance
• Evolution over time and vis-à-vis other CIS
• Fund-supported programs, 1996-99 and 2001-04
• Current PRGF program 2005-08
III. Cross-Country Regression Analysis
• Methodology and Data
• Regression Results
• Robustness
IV. Conclusion
I. Overview of Economic Performance
 Generally strong macroeconomic performance
 Yet, weak structural foundations:
–
–
–
–
Large scale emigration
Poverty
Corruption
Weak fiscal package/incentives
 Need to improve tax collection because:
– Long-term growth potential
– Poverty reduction goals (as outlined in the PRSP)
– Expected decline in concessional borrowing
Figure 1. Armenia: Central Government Tax Revenues
(In percent of GDP)
18
17
16
15
14
13
12
11
10
9
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Figure 1a. Armenia: Main Components of Tax Revenues
(In percent of GDP)
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
VAT
Excises
2000
CIT
2001
PIT
2004
Customs
Figure 1c. Direct Taxes
(In percent of total taxes)
70
60
50
40
30
20
10
0
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
jan
n
m
en
ia
Ar
Re
pu
bli
c
Ge
or
gi
a
er
ba
i
sta
6
an
a
ia
ru
s
Ru
ss
Be
la
an
ai n
e
Uk
r
ak
hs
t
old
ov
11
Uz
be
ki s
t
Ka
z
M
CI
S
CI
S
km
en
i st
an
Tu
r
gy
z
Ky
r
Az
Ta
ji k
i
Figure 1f. CIS: Tax Revenue, 1998-2004
(In percent of GDP)
30
25
20
15
10
5
0
Tu
z
sia
Ka
za
kh
st
an
Uz
be
ki
st
an
Uk
ra
in
e
Ru
s
Be
la
ru
s
CI
S1
1
n
Az
er
ba
ija
n
st
a
CI
S6
Ar
m
en
ia
Ge
or
gi
a
M
ol
do
va
lic
ta
n
Re
pu
b
is
rk
m
en
i
Ky
rg
y
Ta
j ik
Figure 1g: CIS, Direct Taxes, 1998-2004
(In percent of GDP)
10
8
6
4
2
0
Figure 2. CIS: Total Tax Revenues vs. Direct Taxes, 1998-2004
30
Total tax revenue-GDP ratio (%)
Russia
25
Moldova
20
15
10
Armenia
Azerbaijan
Kyrgyz Republic
Tajikistan
5
0
5
10
15
20
25
30
Share of direct taxes in total tax revenue (%)
35
40
45
lic
Ru
s
sia
CI
S1
1
CI
S6
ep
ub
e
Mo
l
a
us
do
v
Be
lar
n
Ar
m
en
ia
zR
ra
in
Uk
Uz
be
kis
ta
Ky
rg
y
Ge
or
gi
Tu
a
rk
m
en
ist
an
n
sta
n
an
Az
er
ba
ija
Ka
za
kh
is t
Ta
j ik
Figure 2a: CIS, VAT Revenues, 1998-2004
(In percent of GDP)
10
8
6
4
2
0
Figure 3. CIS . Tax Revenues vs. VAT Productivity
30
Russia
Tax-GDP ratio
25
20
15
Armenia
T ajikistan
10
Azerbaijan
5
0.10
0.15
0.20
0.25
0.30
0.35
VAT productivity
0.40
0.45
0.50
Figure 3a. CIS: Tax Revenue Buoyancy, 1998-2004
Kazakhstan
Georgia
Azerbaijan
Tajikistan
Russia
CIS6
CIS11
Kyrgyz Republic
Turkmenistan
Belarus
Moldova
Armenia
Uzbekistan
Ukraine
0.6
0.8
1.0
1.2
1.4
Figure 3b: CIS, VAT Revenue Buoyancy, 1998-2004
Tajikistan
Georgia
Azerbaijan
Kyrgyz Republic
CIS6
CIS11
Russia
Kazakhstan
Turkmenistan
Moldova
Belarus
Armenia
Uzbekistan
Ukraine
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
Figure 3c: CIS, Direct Tax Buoyancy, 1998-2004
Kazakhstan
Georgia
Azerbaijan
Russia
CIS11
Kyrgyz Republic
CIS6
Moldova
Ukraine
Turkmenistan
Belarus
Uzbekistan
Tajikistan
Armenia
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
Figure 4.
Tax Revenues in Armenia: Performance Under ESAF and PRGF
(Percent of GDP)
17
16
15
14
13
12
11
10
9
1996
1997
1998
Program
1999
2001
Outcome
2002
2003
2004
What’s wrong with this picture?
 Tax collection low compared to other CIS countries
 Tax collection low compared to its potential
 Decline in share of direct taxes → Poor balance
between equity and efficiency
 Lowest buoyancy of (especially direct) taxes in CIS
 Low VAT productivity, reflecting problems in
administration and exemptions
 Eighteen revenue measures in 2005-08 PRGF
– Six tax policy measures
– Seven tax administration measures
– Five customs administration
III. Regression Analysis
Following the literature, determinants of tax-GDP ratio
across countries and time are:
 Per capita GDP
 Institutional quality
 Inflation
 Share of agriculture
 Openness of the economy
 Shadow economic activity
 Fuel share of merchandize exports
 Share of Urban population
Data Sources
 Publicly available data on up to 113 countries
 For the period of 1990-2004
 Indicators of institutional quality from the
International Country Risk Guide (ICRG)
 Estimates of Shadow economy from
Schneider (2005)
 VIOLENCE, constructed based on the dataset
provided by Center for Systemic Peace
Determinants of Shadow Economy
 All coefficients have expected signs:
 Bureaucratic quality reduces shadow
economic activity, but at a declining rate
 Corporate tax rate is not significant
 Higher per capita income is associated with
lower shadow economy
 More political/social violence leads to more
shadow economic activity
Time trend
PPP-adjusted income p.c.
Composite institutional quality
Shadow economy-1
-0.033
-1.25
7.0E-05
(2.20)**
0.045
(3.94)***
-0.025
(1.83)*
Dependent Variable:
Total Tax Revenue as a Share of GDP
-0.034
-0.027
-0.029
-0.032
-1.29
-1.00
-1.04
-0.92
7.0E-05
9.0E-05
9.0E-05
8.0E-05
(2.24)**
(2.55)** (2.59)***
(1.76)*
0.043
0.041
0.040
0.038
(3.75)*** (3.61)*** (3.41)*** (3.04)***
-0.026
-0.026
-0.025
-0.031
(1.88)*
(1.88)*
(1.77)*
(2.12)**
-0.012
-0.012
-0.011
-0.021
(1.72)*
(1.72)*
-1.6
-1.08
-0.009
-0.011
-0.005
(1.76)*
(1.98)**
-0.77
-0.024
-0.001
-0.93
-0.02
0.081
(4.83)***
79.11
-1.52
81.15
-1.56
67.18
-1.28
73.41
-1.32
76.54
-1.12
-0.082
(1.93)*
1.1E-04
(2.33)**
0.038
(3.11)***
-0.029
(2.00)**
-0.021
-1.1
-0.005
-0.86
-0.001
-0.01
0.084
(5.03)***
0.104
(2.00)**
170.72
(2.06)**
898
113
48.43(8)
0.03
898
113
51.59(9)
0.03
896
112
52.90(10)
0.04
863
111
56.24(11)
0.04
775
107
53.37(12)
0.06
775
107
53.34(13)
0.07
Inflation
Openness
Share of agriculture in GDP
Share of fuel in exports
Urban population
Constant
Number of observations
Number of countries
Hausman test 1/
R-squared
Main Results
 Positive effect of per capita GDP on tax revenue level
 Positive and strong impact of institutional quality
 Negative and significant impact of the shadow
economic activity
 Negative but insignificant impact of agriculture
 Negative sign on inflation
 Negative correlation between openness and taxes
 Finally, OIL and URBAN are both positive and
significant
Table 11. Summary Statistics for Actual vs. Predicted Total Tax Revenue 1/
Mean
No. of Obs.
Std. Dev.
Min
Max
Armenia
Actual Tax Revenue
Predicted Tax Revenue
Ratio of Actual to Predicted
13.89
15.25
0.91
9
9
9
1.50
0.18
0.10
10.64
15.04
0.69
16.08
15.53
1.05
CIS 9 2/
Actual Tax Revenue
Predicted Tax Revenue
Ratio
16.25
14.94
1.10
44
44
44
4.27
1.97
0.30
11.24
10.97
0.71
26.20
18.79
1.85
Low-Income Countries
Actual Tax Revenue
Predicted Tax Revenue
Ratio
14.33
13.84
1.05
180
180
180
5.38
1.01
0.41
2.00
10.97
0.15
26.20
16.92
2.05
Lower Middle-Income Countries
Actual Tax Revenue
Predicted Tax Revenue
Ratio
15.93
15.55
1.03
254
254
254
6.18
1.11
0.41
2.56
12.48
0.17
37.15
18.79
2.28
Full Sample
Actual Tax Revenue
Predicted Tax Revenue
Ratio
16.46
16.46
1.01
851
851
851
6.52
2.36
0.40
0.09
10.97
0.00
37.15
22.38
2.28
16.16
85
4.20
11.24
27.90
15.95
31
1.92
13.22
21.08
Post-estimation Analysis
Memorandum items:
All CIS 11 obs., including those not in the
regressions 3/
Baltic countries, including obs. not in the
regressions 4/
1/ Unweighted averages over the values of the indicators used in the regressions, unless specified otherwise.
2/ Includes Azerbaijan, Belarus, Moldova, Kazakhstan, Kyrgyz Republic, Russia, Tajikistan, Ukraine, and
Uzbekistan.
3/ In addition to CIS 9, includes Georgia and Turkmenistan.
4/ Includes Estonia, Latvia, and Lithuania.
Figure 6. Armenia: Tax Potential vs. Tax Effort, 1996-2004
18.0
17.0
1.10
Post-estimation Analysis
1.05
1.00
16.0
0.95
15.0
0.90
14.0
0.85
0.80
13.0
0.75
12.0
0.70
11.0
0.65
10.0
0.60
1996
1997
1998
Actual tax revenue (Left scale)
1999
2000
2001
2002
Predicted tax revenue (Left scale)
2003
2004
Ratio (Right scale)
Table 13. Armenia: Short-term Impact on Tax Revenue of Improvements
in Institutional Quality
Resulting improvement in
potential tax revenue, based on:
Benchmark
value of the
Required
indicator improvement
Overall Institutional quality at the level of
Baltic countries
Shadow economy at the level of Baltic countries
of which,
through improvements in Bureaucratic
quality to the level of Baltic countries
Total impact (of improving institutional quality
to the level of Baltic countries)
Baseline
regressions
Truncated
regressions
75.6
32.9
15.3
18.3
0.58
0.51
0.74
0.33
2.5
1.5
0.17
0.11
0.75
0.85
Conclusions
 Armenia demonstrates low (particularly direct) tax buoyancy
 While finding oil may not be a feasible alternative, other tools
are available
 Better institutions appear to cause higher tax-GDP potential
both directly and indirectly (through shadow economy)
 Improvements in tax policy and administration would help so
long as they address the fundamental incentive issues
 Therefore, for any attempt to increase the tax performance in
Armenia to be successful, measures to improve governance,
rule of law, and bureaucratic quality—all in the realm of the
political will—should be in its core