Policy Studies Institute Working Paper 028 Tariff policy, Fiscal Cost and Efficiency in Ethiopia * Firew Bekele Woldeyes Biniam Egu Bedasso Andualem Telaye Mengistu * This research work has been funded by the Agence Française de Développement (AFD) under the Structural Transformation and Industrial Policy in Ethiopia (STIP) program.
28
Embed
Tariff policy, Fiscal Cost and Efficiency in Ethiopia
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Policy Studies Institute Working Paper 028
Tariff policy, Fiscal Cost and Efficiency in
Ethiopia*
Firew Bekele Woldeyes
Biniam Egu Bedasso
Andualem Telaye Mengistu
* This research work has been funded by the Agence Française de Développement (AFD) under the
Structural Transformation and Industrial Policy in Ethiopia (STIP) program.
THE POLICY STUDIES INSTITUTE WORKING PAPERS
ABOUT PSI
Policy Studies Institute (PSI) is a policy think tank established in November 2018 by the Ethiopian
government. It was founded by merging two state-owned think tanks in Ethiopia, Ethiopian Development
Research Institute (EDRI) and Policy Study and Research Centre (PSRC), whose establishment dates back to
1999 and 2014 respectively.
About STIP
STIP is a comprehensive research and policy support program on structural transformation and industrial
policy in Ethiopia. The program is designed to serve as a knowledge backstop for policymakers and other
key stakeholders. The STIP program is funded by the Agence Française de Développement (AFD).
ABOUT THESE WORKING PAPERS
The PSI Working Papers contain peer reviewed material from PSI and/or its partners. They are circulated in
order to stimulate discussion and critical comment. The opinions are those of the authors and do not
necessarily reflect those of their home institutions or supporting organizations. EDRI takes no responsibility
for any errors or omissions in, or for the correctness of, the information contained in papers.
Paper citation: Firew Bekele Woldeyes; Biniam Egu Bedasso and Andualem Telaye Mengistu (2019). Tariff
policy, Fiscal Cost and Efficiency in Ethiopia. PSI Working Paper 028. Addis Ababa: Policy Studies Institute.
About the Author(s):
(1) Firew Bekele Woldeyes is a Lead Researcher at the Policy Studies Institute (email: [email protected])
(2) Biniam Egu Bedasso a Lead Researcher at the Policy Studies Institute (email: [email protected])
(3) Andualem Telaye Mengistu a Lead Researcher at the Policy Studies Institute (email:
There is a voluminous literature regarding the gain in productivity and economic growth
as a result of reduction in tariff in general and input tariff in particular. However, this
literature does not deal with the impact the reduction in tariff has on the loss of
government revenue. In addition, if the tariff reduction is not uniform, it pauses
significant challenge for administration. When the reduced tariff is dependent on
fulfilling certain criteria, beneficiaries may differ from the ones intended in the policy.
We find that the second schedule tariff scheme in Ethiopian leads to substantial loss in
government revenue and the loss has not been captured by an increase in other domestic
taxes. In addition, although the scheme is intended for wider sectors and firm distribution,
we find that bigger firms and firms in basic metal and engineering sector benefit more
from the scheme than others in the economy.
1 Research Fellow at the Policy Studies Institute (PSI). He can be reached at [email protected] 2 Research Fellow at the Policy Studies Institute (PSI). He can be reached at [email protected] 3 Research Fellow at the Policy Studies Institute (PSI). He can be reached at [email protected]
2
1 Introduction
The benefit of free trade is one area where there is little disagreement among economists
around the world provided that the right policies are in place (see Greenaway et al. 2002;
Wacziarg & Welch 2008). In addition, there is a voluminous literature regarding the gain
in productivity as a result of reduction in tariff. This is especially the case when the tariff
reduction is aimed at inputs.
The growth literature (Estevadeordal & Taylor 2013; Williamson & Clemens 2004) deal
with the overall impact of trade liberalization on growth rather than the fiscal cost and its
associated efficiency implications. In addition, a quick review of the literature on firm
performance, shows that firm productivity increases with the use of imported
intermediate inputs4.
One policy tool that encourages the use of intermediate inputs is targeted trade
liberalization. Trade liberalization measures such as lowering tariffs on inputs have been
found to be important instruments in increasing productivity and introducing new
products into the market (Goldberg et al. 2010). Several studies have documented a
similar result using different methodologies and data(Amiti & Konings 2007; Bigsten et
al. 2016; Topalova & Khandelwal 2011). In light of such findings, there is a steady
reduction in the average tariff rates in both high and low income countries for the last
three decades (Turnovsky & Rojas-Vallejos 2018).
In most countries, tariff reductions tend to be uniform for a broad set of commodities
with similar characteristics to reduce the incentive for tax evasion (Afontsev 2012). At
the extreme end, almost uniform tariff rate of 10% has been adopted in Chile in 19795
(Corbo 1997; Ffrench-Davis 2010). On the other hand, Ethiopia still relies on
differentiated tariff rates for identical items depending on the importer as an industrial
policy tool. Differentiated tariff for identical items are likely going to create a suitable
4 (Foster-McGregor et al. 2016)(19 sub- Saharan African countries), (Goldberg et al. 2010) (for India),
(Kasahara & Rodrigue 2008) (for Chile), and (Aristei et al. 2013)(for 27 eastern European and central Asian countries), (Edwards et al. 2018)(for South Africa), and (Gebreeyesus 2008)(for Ethiopia) 5 In Chile, 99% of imports face the same 10% tariff rate in 1979.
3
ground for tax evasion and pauses significant challenge on administrative capacity of the
Customs Authority (Fisman & Wei 2004) and Mengistu, et al. (forthcoming)6.
In low-income countries, revenue from taxes on international trade is still a significant
part of the total government revenue. (Baunsgaard & Keen 2010). The significance of
trade tax in government revenue in low income countries, the recent move towards
increasing domestic resource mobilization, and the limited revenue mobilization potential
of developing countries, makes the need to take in account the fiscal cost of trade
liberalization of paramount importance. In addition, the differential tariff rate approach to
trade liberalization as is the case in Ethiopia will also lead to tax evasion and additional
revenue loss.
However, the literature on the impact of tariff reduction on firm performance has dealt
with partial equilibrium scenarios. It doesn’t address how the reduction of tariff affects
revenue and how the resulting lack of fiscal space affects the performance of firms and
the overall economy.
Few papers have addressed the issue of revenue loss due to tariff reduction and how other
sources of revenue respond in the aftermath (Agbeyegbe et al. 2004; Baunsgaard & Keen
5 Simple average tariff rate 79.1 35 28.8 24.3 24.3 20 20 20 20 6 Weighted average tariff
rate 41.6 29.6 24.6 23.6 21.5 19.5 17.5 17.5 17.5
7 Tariff dispersion 225 75 55 45 45 35 30 30 30 Source: Ministry of Finance and Economic Development, Government of Ethiopia (2014)
As shown in Table 1, the total number of applicable tariff lines increased from 1,821 to
5,332 with the first reform and has not changed significantly ever since. Before the first
reform, the highest tariff rate was 230% (and was mainly levied on selected luxury
consumer goods). This rate declined to 80% with the first round amendment and
gradually reached 35% in 2003. Besides, the tariff bands, the average tariff rates, and the
tariff dispersions have all been continually declining with each round of amendment. The
number of tariff bands declined from 23 to 9 with the first round amendment and
gradually reached 6 by 2003. Similarly, the simple average tariff rate was 79.1% before
1993 which declined to 35% with the first reform and finally reached 20% by 1998. The
6
weighted average tariff rate was around 42% before 1993, declined to about 30% in 1993,
and gradually became 17.5% in 1998.
Although Ethiopia has recently been attempting to reduce its dependence on international
trade taxes by expanding the domestic tax base, import duties and taxes still account for a
significant proportion of its total tax revenue. The share of import duties and taxes in total
tax revenue was around 36 percent in 2012/13, higher than any single source of domestic
tax revenue. Yet it is also worth mentioning that the share of tax revenue generated from
customs duties alone has been declining continuously in the last decade. For instance,
while customs duties accounted for about 21% of the total tax revenue in 2005/06, this
figure has eventually shrunk to hover just above 12% in 2012/13 (Mengistu et al. 2015).
In general, the Ethiopian government uses customs tariff reductions and exemptions as a
means of promoting investment and exports. The role of tariff policy for industrial
development seems to have gained further official recognition to the extent that the most
recent customs proclamation No. 859/2014 explicitly acknowledges the imperative of
manufacturing development as one of the motivations behind the law. Among the tariff
concession schemes employed by the government to stimulate the development process,
the second schedule tariff scheme is specifically meant to promote the competitiveness of
the manufacturing sector in the domestic market. According to the customs valuations
and program development directorate of the ERCA, with every round of customs tariff
reforms, the government has been revising the second schedule tariff scheme with the
aim of supporting the manufacturing sector.
The implementation of the second schedule scheme has taken the form of iterative
enactments of a series of directives which are aimed to improve effectiveness, close
loopholes and minimize government revenue loss. Although it is difficult to clearly trace
the content and timeline of all interpretative and procedural directives, it appears that
successive directives may have restricted the privileges of some firms while they extend
or consolidate tariff incentives given to others.7 This is because many of the recent
7 See (Lencho 2012)for an extensive discussion of the challenges of drawing timelines and tracing the evolution of such provisions as tax exemptions in the Ethiopian tax system.
7
directives focus on regulating what used to be a more subjective and ad hoc
administration of tariff privileges. For the purpose of this report, we focus on directives
issued over the past five years with regard to second schedule tariff incentives. We draw
on content analysis and key informant interviews to set the policy background which is
applicable for the period we have chosen for analysis.
During the last five years, the government has issued two directives with the intention to
systematically regulate the use of tariffs as an incentive for import substitution. The first
of this directive, No. 35/2013, requires manufacturers to have a minimum of 30 percent
domestic value addition on top of imported raw materials to qualify for tariff reduction
privileges. It also stipulates that manufacturers should obtain a certificate from the
Ethiopian Revenue and Customs Authority before they import raw materials under the
scheme. Almost all public sector and private sector stakeholders we interviewed pointed
out that the most problematic aspect of the directive was that it put a flat value addition
threshold of 30 percent which most manufacturers in many sectors could not meet. The
value addition threshold was benchmarked against the 35 percent domestic value addition
threshold stipulated in COMESA’s rules of origin protocol. In that sense, the directive
left a lot to be desired in terms of taking the right context in to account.
The second directive, No. 45/2016, was issued by the MOFED to rectify the above
problem with differentiated value addition requirements for various industrial subsectors.
It also shifts the mandate of issuing a certificate from ERCA to MoI. By changing the
value addition requirement from a flat rate of 30 percent to a range from 5 to 40 percent,
the present directive appears to have taken the current stage of development of different
industries into consideration. However, interviews with government officials reveal that
the issue of determining value addition requirements remain a major challenge. As a first
strategy to determine value addition thresholds, officials tried to use the administrative
structure of sectoral institutes to gather data from relevant firms operating in import
substituting sectors. However, the data collection process was ravaged by delays.
Moreover, the committee in charge of administering the scheme deemed the value
addition figures reported by firms through the institutes unreliable. Therefore, as a second
strategy, the committee sought the help of the CSA to determine average value addition
8
for a selected set of industrial subsectors based on firm census data. Although the value
addition thresholds are officially set to be revised every 3 years, officials reported that
they were forced to issue amended figures for some sectors only a few months after the
directive took effect due to unrealistic thresholds.
Apart from the determination of value addition requirements, the other problems with the
present directive that are perceived both by the implementing agencies and the private
sector alike are a lengthy process of certification and slow rectification of problems
arising during implementation. This is partly because executive decisions on the scheme
are taken by a committee comprised of several governmental bodies with sometimes
conflicting objectives and priorities. Some of the measures taken to minimize discretion –
and therefore possible abuse – such as promulgating a predetermined list of inputs seem
to have created inflexibility and red tape. Some key informants conveyed their concern
that the predetermined list of inputs might leave little room for the innovative use of new
inputs.
Like any other policy that involves significant government intervention, the design and
implementation of the tariff scheme requires sizable capacity on the part of the
government. The effectiveness of the policy measure is likely to depend, among other
things, on the institutional and human resource capacity of the policy-making and
implementing agencies. Our interviews revealed the lack of confidence many officials
have in the capacities of their own departments to help implement such an
administratively demanding incentive scheme. Given the lack of motivation and financial
incentives a previous study has identified among the staff of the same set of agencies that
are tasked with administering this scheme, the skeptical assessments of the interviewees
regarding implementing capacity seem justified.8 We also found that there is virtually no
monitoring and evaluation mechanism set up for the scheme except for the annual reports
that the directive requires ERCA to produce.
8 The problems of incentives and motivation in selected public institutions are documented in the following report: “Industrial Policy Governance and Bureaucratic Capacity in Ethiopia”, Policy Report, STIP, EDRI, 2016.
9
3 The impact of input tariff reduction on Government
Revenue
We will start with a detailed description of Ethiopia’s tariff lines and duty rates and their
progression over time to provide the context for the second schedule tariff rates and lines.
Table 2 presents the changes in the number of tariff lines under each duty rates between
2007 and 2015 from the overall tariff book. During this period, the number of tariff lines
that fall under the zero duty rate has increased by slightly more than 1 percentage point.
Accordingly, over 5 percent of the tariff lines are zero rated. Consistent with the above
changes, the number of tariff lines under the maximum 35% duty rate has declined by
about 1 percentage point. This confirms the general trend of falling duty rates and these
changes have concentrated around 2008 and 2015.
Table 2 Tariff Lines by Duty Rates between 2007 and 2015
Next, we present the revenue loss associated with applying the reduced duty rates instead
of the general duty rates. Figure 1 presents the duty revenue collected, duty reduction due
11
to second schedule and duty reduction due to other reasons. The duty collected for 2016
is 21 billion birr compared to duty reduction of 16 billion due to the second schedule and
duty reduction of 37 billion birr due to other duty reductions (such as investment
incentives)910.
The result presented here does not take into account behavioral changes that would result
in lesser import (and lower duty collection) if the exemptions were not available.
Figure 1 Duty paid and Duty that would have been paid between 2010 and 2017 in millions of birr)
9 Note that we don’t exactly have the data for duty deduction because of the second schedule. Therefore,
we have estimated it by using the difference between the first schedule duty rate and second schedule duty rate for those who paid the same rate as the second schedule rate. This approach might overstate the second schedule duty deduction in cases where other duty incentives are exactly the same as the second duty rate.
10 We have only managed to get data for the first six months for 2017, and hence both the duty reductions
and duty collection appear lower than the previous year. One can multiply the values for 2017 by a factor of two to get a rough idea about the revenue loss involved.
12
Source: Author’s computation based on ERCA data11
The birr amount of duty reduction has increased sharply in line with increasing tariff lines
under the second schedule scheme and increasing imports. There are a number other
schemes in addition to the second schedule scheme that will contribute to the reduction of
duty collection such as investment incentives – which is significantly larger than the
second schedule scheme reduction (dark blue).
Figure 2 break downs the tariff reduction received by sector and year and shows that the
basic metal and engineering sector has been using the preferential treatment the most.
Other sectors which have benefited from the second schedule scheme are the chemical
and construction inputs, the leather and pharmaceutical industries. The leather industry
started using the scheme after 2012 probably as a result of the addition of new tariff lines
into the second schedule tariff lines.
11
Notice that the data for 2017 covers the first six months only.
13
Figure 2 Second Schedule Duty Reduction by Sector and Year
Source: Author’s computation based on ERCA data
In the next section, we will present a firm level analysis on how the domestic tax
payment of firms has been impacted as a result of getting the preferential treatment of the
second schedule duty rates.
4 Efficiency of Tariff Policy Implementation
Despite the tariff reductions and associated positive impact on firm productivity (Bigsten
et al. 2016), the contribution to improved domestic tax collection as a result of better
profitability of those firms benefiting from the scheme is an unanswered question.
Studies that have compared countries have found that low income countries have failed to
recover the revenue loss, arising from trade liberalization through improved domestic tax
collection. This raises the question, have the firms that benefited from the scheme paid
more in profit taxes? This is one of the questions we want to investigate in this section.
14
Another concern raised is that despite the improved productivity of the firms as a result
of trade liberalization, the manufacturing sector remains weakly integrated into global
supply chains. A recent study commissioned by EDRI finds that “restrictive trade
policies, as well as poor trade facilitation, make it difficult for producers and exporters to
access inputs at world market prices”(Shepherd 2017). The results are likely to be mixed
because policy pronouncements such as the ones that have been issued to stimulate
import substitution do not turn into intended outcomes automatically.
In this regard, the bureaucratic apparatus plays a key role in implementing such policies
to fruition. An exploratory study conducted by EDRI to map out the institutional
background for industrial policy implementation has identified the Ethiopian Revenue
and Customs Authority (ERCA) as the most influential government agency playing a role
in the operation of export oriented manufacturing industries (Tefera et al. 2018). This
finding points to the crucial role of fiscal policy formulation and implementation for
industrial development in Ethiopia. In this regard, we provide qualitative and quantitative
evidence on the efficiency with which the policy of preferential tariff schedule has been
implemented.
4.1 Regression analysis
In order to investigate the impact on profit tax payment of those firms that have benefited
from the second schedule system, we have used data from Ethiopian Revenue and
Customs Authority (ERCA) on customs paid from the ASYCUDA++ database and
merged it with the database on firm characteristics and income statement. The firms in
our dataset are those firms that have imported goods between 2011 and 2014. We have
excluded government bodies and cooperative societies. The period of the analysis is
between 2011 and 2014. We have a panel of approximately 52,000 business. The
number of owners is much less than the number of businesses as one taxpayer can
operate two businesses in different sectors. As the different businesses face different duty
rates we treat them separately.
15
Figure 3 plots the log of profit tax payable (y-axis) against the percentage of the values of
goods imported through the second schedule system (x-axis). While zero (on the x-axis)
indicates importing all goods through the regular (first schedule system), 1 indicates
importing all goods for that year through the second schedule system. The result shows
that benefiting from the second schedule is not correlated in any way with the profit tax
payment during the period 2011 to 2014.
Figure 3 Log of profit tax payable and second schedule benefit
One explanation for the unimproved profit tax reported for those firms that have
benefited from the second schedule could be that these firms would have paid less in the
absence of the second schedule scheme. In other words, those firms that are benefiting
from the second schedule operate in a less profitable sector or these businesses tend to be
small. Therefore, we have implemented a fixed effect regression which controls for time-