Working Paper 38

Taxing the Urban Boom:

Property Taxation and Land Leasing in Kigali
and Addis Ababa

Tom Goodfellow
July 2015

www.ictd.ac
ICTD Working Paper 38

Taxing the Urban Boom: Property Taxation

and Land Leasing in Kigali and Addis
Ababa

Tom Goodfellow

July 2015

1
Taxing the Urban Boom: Property Taxation and Land Leasing in Kigali and Addis Ababa
Tom Goodfellow
ICTD Working Paper 39
First published by the Institute of Development Studies in July 2015
© Institute of Development Studies 2015
ISBN: 978-1-78118-241-3

A catalogue record for this publication is available from the British Library.
All rights reserved. Reproduction, copy, transmission, or translation of any part of this publication may
be made only under the following conditions:
- with the prior permission of the publisher; or
- with a licence from the Copyright Licensing Agency Ltd., 90 Tottenham Court Road, London W1P 9HE, UK,
or from another national licensing agency; or
- under the terms set out below.

This publication is copyright, but may be reproduced by any method without fee for teaching or nonprofit purposes, but not for
resale. Formal permission is required for all such uses, but normally will be granted immediately. For copying in any other
circumstances, or for reuse in other publications, or for translation or adaptation, prior written permission must be obtained from
the publisher and a fee may be payable.

Available from:
The International Centre for Tax and Development
at the Institute of Development Studies,
Brighton BN1 9RE, UK
Tel: +44 (0) 1273 606261 Fax: +44 (0) 1273 621202
E-mail: [email protected]
Web: www.ictd/en/publications

IDS is a charitable company limited by guarantee and registered in England (No. 877338)

2
Taxing the Urban Boom: Property Taxation and Land Leasing in Kigali
and Addis Ababa

Tom Goodfellow

Summary

Much contemporary economic growth in Africa is driven by urban service sectors including
construction and real estate. This manifests in rapidly transforming landscapes and the
proliferation of valuable property in the continent’s booming large cities, often accompanied
by growing socio-economic inequality. In this context, improving systems for property
taxation is an urgent and growing need – something that national and international
policymakers increasingly recognise. Despite this, even in states considered particularly
‘developmental’ and committed to increasing tax revenue, property taxation has fallen by the
wayside. This paper argues that, in addition to the usual technical and political difficulties
associated with property taxation, it is crucial in reform-oriented developing countries to
understand the nature of land tenure systems and how they are changing, as well as
historical legacies relating to land and housing. Moves to introduce or improve property
taxation need also to be considered in relation to how contemporary development strategies
constrain or facilitate investment in real estate, and the political economy underpinning this
sector. Through an analysis of existing property tax and land leasing systems in Kigali and
Addis Ababa, as well as failed, stalled and ongoing reform attempts in each, the paper
identifies key themes for the study of property taxation systems and their potential to
succeed in rapidly transforming cities in the global South.

Keywords: property tax; land leasing; land value capture; urban land reform; real estate;
urban development; Africa; Rwanda; Ethiopia.

Tom Goodfellow is Lecturer in the Department of Urban Studies and Planning at the
University of Sheffield. His research concerns the political economy of urban development,
the politics of urban governance and planning, the informal economy, land and taxation in
sub-Saharan Africa. He has acted as an advisor to Oxfam on urban development issues, and
provided policy analysis for the Department for International Development and UK
Parliamentary Select Committee on International Development, as well as UN-HABITAT and
government authorities in Rwanda and Uganda. His work has been published in a range of
academic journals, and he is co-author of the book Cities and Development (Routledge
forthcoming).

3
Contents
Summary 3
Acknowledgements 5
Acronyms 5
Introduction 6

1 Property tax in context 7

1.1 Fiscal decentralisation and property tax 8
1.2 Accounting for failures of property taxation in the developing
world 9
1.3 Property tax, land and the real estate sector 10

2 Kigali 11
2.1 Property tax and the framework for fiscal decentralisation
(and recentralisation) 11
2.2 Taxation and land tenure regularisation: ‘there is huge confusion’ 16

3 Addis Ababa 20
3.1 Property tax and the framework for fiscal federalism (and
recentralisation) 20
3.2 Taxation and land leasing: ‘people are already feeling taxed’ 24

4 Reform: opportunities and obstacles in comparative perspective 27

4.1 Rwanda 27
4.2 Ethiopia 29
4.3 Cross-cutting issues 31

5 Conclusion 36

References 37

Tables
Table 1 Summary statistics of plots of land in the ten sub-cities in
Addis Ababa 25
Table 2 Comparison of existing systems of property taxation in Rwanda
and Ethiopia 32

Figures
Figure 1 The composition of local revenue in Kigali’s three districts
2008-9 13
Figure 2 Local revenue for all districts in Rwanda combined, 2012 14
Figure 3 Local revenue for Kigali districts, 2013 15
Figure 4a Addis Ababa, total revenue, 2013/14 22
Figure 4b Addis Ababa revenue minus income and profit tax, VAT,
non-tax revenue and loans, 2013/14 23
Figure 4c Addis Ababa municipal revenue only 23
Figure 5 Addis Ababa selected revenue over time, 2006-2014 26

4
Acknowledgements
I would like to extend thanks to Mick Moore, Wilson Prichard, Wollela Yesegat, Odd-Helge
Fjeldstad, Philip Kargbo, Oliver Morrissey and Attiya Waris for taking time to read and
comment on the initial research proposal and helping me to frame the project. I would also
like to thank two anonymous reviewers who commented on the draft paper, as well as the
many interviewees in Rwanda and Ethiopia who were so generous with their time and
assistance. All errors are my own.

Acronyms
AACA Addis Ababa City Administration
DFID Department for International Development, UK
EPRDF Ethiopian People’s Revolutionary Democratic Front
ERCA Ethiopian Revenue and Customs Authority
GDP Gross Domestic Product
GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit
ICTD International Centre for Tax and Development
IMF International Monetary Fund
LTRP Land Tenure Regularisation Programme
NGO Non-governmental organisation
RDF Rwandan Patriotic Front
RRA Rwandan Revenue Authority
RWF Rwandan Franc
VAT Value Added Tax

5
Introduction
Gradually, but insistently, property taxation is emerging on the international development
agenda, amid growing recognition that it is widely under-utilised as a source of revenue
mobilisation, local state-building and urban redistribution. Recent decades have seen
widespread recognition of the importance of tax in general as a tool of state-building and
development, both in terms of its role in fostering a fiscal contract between government and
the governed, and its capacity to increase self-reliance and self-determination in highly aid-
dependent countries (Levi 1988; Lieberman 2002; Moore 2004; Bräutigam et al. 2008). Many
of the tax reforms taking place across the developing world since the late twentieth century
have, however, been primarily concerned with indirect taxes, and in particular the
simplification and reduction of customs duties, as well as the introduction of Value Added
Tax (VAT). Even where direct taxes are concerned, the primary focus has been on the
creation of effective national revenue systems and the simplification of income taxation. This
is understandable, as the majority of revenue-raising everywhere happens at the national
level, especially in developing countries with colonial and postcolonial legacies of
centralisation. Yet the renewed focus on tax reform has been occurring concurrently with the
surge of decentralisation reforms across the developing world, through which more and more
responsibilities have been shifted to local authorities that generally lack the fiscal capacity to
execute them.

Local government revenue generation in the developing world often constitutes a bewildering
array of taxes, administrative fees, payments for permits and licences, fines, user fees for
specific services, and varying forms of land use fees and rents. Amid all these lies a sleeping
giant: property taxation. In much of Africa – a primarily rural continent that has long been
dependent on agriculture – an urban revolution is underway (Parnell and Pieterse 2014), with
many of the continent’s cities booming economically, physically and demographically.
Against a background of economic growth in many African states, cities on the continent
have been termed the final frontier for international property development (Watson 2014), as
well as being spaces for capital accumulation by domestic elites. The rapid growth of many
economies in Africa and elsewhere in the developing world is in large measure driven by
construction and other urban service sectors, especially given the difficulties of pursuing the
more conventional forms of economic catch-up through industrialisation. Such property-
driven development trajectories imbue property taxation with new urgency: high-end real
estate is sucking up vast amounts of capital in many very poor countries, with a questionable
developmental effect. Yet despite considerable attention to property taxation by scholars and
the development community in recent years,1 very few lower-income countries have a
property tax system that is fit for purpose (Monkam and Moore 2015).

Governments and international donors alike are waking up to this urgency. Yet designing and
implementing more effective property taxation systems is fraught with challenges and
obstacles that are highly context-specific, and require case-based research as well as
comparative analysis in order to distinguish generalities from specificities and to facilitate
cross-national learning. Among the most pressing concerns is the question of how states,
which are radically overhauling their mechanisms for land governance and land tenure in an
effort to promote rapid development, can combine new urban land institutions with effective
and progressive property taxation. Failure to align urban land reform with urban tax reform
not only represents a missed opportunity, but could actively encourage modalities of urban
development that are divisive and exclusionary, as well as economically and environmentally
unsustainable.

1
See, for example, Smoke (2001); Bird and Slack (2004); Fjeldstadt (2006); Bahl and Bird (2008); Franzsen and
McCluskey (2013); McCluskey and Franzsen (2013a, 2013b).

6
This paper explores the challenges, opportunities and pitfalls of property taxation through the
two cases of Rwanda and Ethiopia, with particular attention to their respective capital cities,
Kigali and Addis Ababa. These two states are at the forefront of the debate on successful
economic growth experiences (and dubious human and political rights experiences) in Africa
– sometimes being compared in their orientation and ambition to the developmental states of
East Asia in the late twentieth century. While they exhibit similarities to one another, there
are also critical differences with respect to their development strategies and the specific
challenges they face regarding property taxation. The varying approaches towards the real
estate sector, different legacies of land tenure and land reform, and particular relationships
between property taxation and other taxes, present unique bundles of challenges and
opportunities. Through an analysis of progress towards (and regress from) effective property
taxation in these cities, the paper examines the obstacles to and potential for progressive
reform in comparative perspective. In so doing, it also highlights critical elements necessary
for context-specific political economy analysis of property tax regimes and their potential for
reform in any given case.

The research is based on a project funded by the International Centre for Taxation and
Development (ICTD), and involved field research in Kigali and Addis Ababa in June,
September and October 2014. With respect to Rwanda, the paper also draws on a longer
history of field research between 2009 and 2015. The research involved semi-structured
interviews with a diversity of stakeholders including local and national government officials,
investors, property developers and estate agents, urban planners, architects, taxpayers,
international advisers, researchers, lawyers and financial institutions. Tax data was also
collected from local and national authorities, along with legal documents, reports and other
policy documentation.

The paper begins with a section that contextualises the property tax challenge in relation to
decentralisation reforms, urban politics and the real estate sector. It then turns to Rwanda,
examining the framework for fiscal decentralisation and recent moves to recentralise tax
collection, before exploring the relationship between property taxation and land reform in the
context of the capital city, Kigali. Following this I turn to Ethiopia, considering the framework
for fiscal federalism before again examining the relationship between property taxation and
land reforms in Addis Ababa, with particular attention to the recent significance of land
leasing. The paper then considers efforts to reform the property tax system in each case in
comparative perspective, assessing the obstacles and opportunities facing each city. This
section concludes with discussion of some cross-cutting issues that emerge from the
comparison, which may hold significance for reform efforts across Africa and the developing
world more broadly.

1 Property tax in context

No tax exists in isolation, and to understand the effectiveness of a given revenue stream it is
crucial to contextualise it in a number of ways. This section places property taxation in
context in three respects. First, it situates it in relation to the fiscal decentralisation reforms
that have been taking place across the developing world in recent decades. Second, it
considers broad international trends with respect to property taxation, and why its potential is
so rarely realised in the developing world. Finally it considers the importance of
contextualising property taxation in relation to land reform, and especially the nature of the
urban real estate sector.

7
1.1 Fiscal decentralisation and property tax

Even enthusiastic exponents of fiscal decentralisation agree that not all taxes are suitable for
local government. Taxes affecting the government’s stabilisation functions, such as trade
tariffs and VAT, should remain centralised, as they are cyclical and local governments
require stable revenue inflows for core services (Smoke 2001; Fjeldstad 2001, 2006). On the
other hand, decentralising distributive taxes such as corporate income tax could result in the
departure of the wealthy and influx of the poor into certain areas, exacerbating regional
inequalities (Bahl 1999; Smoke 2001; Fjeldstad 2001). A range of other taxes are, however,
considered suitable for local administration. This includes distributive ones applied to
immobile assets, minimising the problem of assets being shifted between localities
(Musgrave 2000; Smoke 2001; Bahl and Bird 2008; Fjeldstad and Heggstad 2011). Land and
property are the ultimate immobile assets (Fjeldstad 2006; Di John 2011). Various other
forms of tax, including local business taxes, are also thought appropriate for decentralising
(Bahl 1999; Fjeldstad 2006; Bahl and Bird 2008).

In reality, many sources of local revenue are not actually taxes at all, but various forms of
fees. These can include everything from fees for advertising billboards and parking, to
payments for local security services, planning applications and other administrative services,
as well as fines for an array of urban offences. Unlike these sources of government revenue,
taxes are not linked to the ‘provision or promise of any specific good or service in return for
payment’ (Lieberman 2002: 91). Taxes are thus usually considered to be qualitatively
different from fees, playing broader roles that feed into accountability, state-building and
social contracts (Bräutigam et al. 2008; Moore 2004, 2008; Von Stein [1885] 1964). Local
authorities in developing countries are often highly dependent on fees, but generally seek to
increase the ratio of local taxes to local fees over time. Regardless of the balance between
the two, most local authorities, even under conditions of substantial fiscal decentralisation,
are usually highly dependent on transfers from central government.

A further and relatively neglected question relates to which sorts of tax are most suitable for
decentralising to urban vis-à-vis rural local authorities. Bahl (1999) acknowledges the two
require different systems, but focuses primarily on the greater capability of urban
governments to raise taxes, rather than the vastly greater mandates for service delivery
delegated to them. Awareness that city governments are often severely underfunded despite
their comparatively lucrative tax base, has, however, been growing (Devas 2003; Fjeldstad
2006; Monkam and Moore 2015). Many local authorities employ a pot-pourri of over-
complicated revenue instruments to address this (Brosio 2000), often with perverse
consequences (Smoke 2001), but the most critical sources of urban revenue are usually
property taxes and business/trading licences, in addition to large numbers of the kinds of
fees described above (Fjeldstad 2006; Fjeldstad and Heggstad 2011).

The scattergun approach to local taxation prevalent in the developing world reflects an
assumption that no single source of local taxation holds the potential to raise much money.
This is understandable with regard to rural areas and the small towns that characterised
much of Africa until relatively recently, which commonly hold limited potential for local
revenue generation. In major cities, however, real property or real estate – defined as an
area of land that has been legally delineated and improved or altered through human effort,
including through the construction of buildings – is a central feature of capital accumulation.
In such areas there is great potential to generate the majority of local income through taxing
this in a way that is also progressive and redistributive. Indeed, consensus has recently
grown that urban property tax is not only among the most underutilised, but potentially the
most progressive, forms of taxation in less developed countries (Fjeldstad 2006; Di John
2011; Fjeldstad and Heggstad 2011; Jibao and Prichard 2013; Monkam and Moore 2015).

8
1.2 Accounting for failures of property taxation in the developing world

While in developed countries property tax can account for up to 100 per cent of local revenue
and 4 per cent of GDP (Norregaard 2013), in most African countries it contributes far less. In
Kenya, for example, it amounted to just 22 per cent of local revenue and 0.3 per cent of GDP
at the start of the millennium (Kelly 2000a). One study found that property tax as a
percentage of GDP in the developing world averaged 0.6 per cent in the 2000s, though this
conceals great variation (Norregaard 2013).2 The tax is underutilised partly because of
reluctance to pay by residents dissatisfied with services (Smoke 2001; Bahl and Bird 2008;
Monkam and Moore 2015). It is also extremely challenging administratively, especially where
informal construction and absent street names render cadastral systems unworkable
(Fjeldstad 2006), and can be inelastic because old valuations are not updated and new
properties not identified (Brosio 2000).

There is a wide range of approaches to the valuation of property (Bahl and Linn 1992; Kelly
2000a; Franzsen 2010; Franzsen and McCluskey 2013; McCluskey and Franzsen 2013b). In
the most straightforward formulation one can distinguish between rates based on plot area,
capital value and rental value (sometimes referred to as annual value) (Monkam and Moore
2015). Within these categories there are potential variations relating to factors such as
whether land and property are both taxed (and whether they are taxed together or
separately), whether plot areas are calibrated using other factors (such as land use and age
of buildings), and whether square metres of floor space are valued alongside land plot area
(McCluskey 1999; Franzsen 2010; Franzsen and McCluskey 2013; McCluskey and Franzsen
2013b). Regardless of method, it has been said that valuation ‘is an art, not a science’,
leaving ‘much room for discretion’ (Bahl and Bird 2008: 15). Most methods are also
administratively intense. Yet valuation is only part of the property tax problem. Kelly’s (2000a,
2000b) model involves five key variables affecting property taxation: i) the tax base, which
involves decisions about what to tax (e.g. land, property or both, and what to exempt); ii) the
tax rate, which is the specified amount per value of the property to be paid; iii) the coverage
ratio, which refers to the amount of taxable property actually captured in the fiscal cadastre;
iv) the valuation ratio, which is the value of the properties as specified on valuation rolls
relative to the actual market value; and v) the collection ratio, which is the tax revenue
actually collected relative to the total tax liability billed for that year.

Among these critical factors there is ample scope for inefficiency, bad policy, poor
coordination and loss of information. Consequently, it is often argued that the primary
problem of taxation is poor administration (Bird and Casanegra de Jantscher 1992; Kelly
2000b). The differences in tax take even between low-capacity states indicate, however, that
administrative capacity is not the only requirement. Taxation – especially where it is direct,
and where it hits the visible assets of powerful groups – is the stuff of politics, and its relative
effectiveness cannot be understood without attention to the dynamics of political economy.
Levi’s focus on the state’s ‘relative bargaining power’ as it impacts on ‘quasi-voluntary
compliance’ (Levi 1988: 13) illustrates how much more there is to effective taxation than
administrative capacity. This bargaining power depends heavily on the buy-in of different
groups to particular taxation regimes and how this affects tax morale – the willingness to pay
taxes – including among elites (Torgler 2007; Lago-Peñas and Lago-Peñas 2010). In
developing countries where relatively small groups of elites often comprise the vast majority
of the formal tax base, elite tax morale is central to the broader relationship between tax
authorities and society and the consequent tax culture (Nerré 2008). Indeed, it has been said
that ‘elite attitudes shape every aspect of the tax system’ (Everest-Phillips 2009: 5).

When it comes to property, one of the most fundamental barriers to effective taxation is
therefore the sustained resistance it faces from property-owning elites, who form a powerful

2
This is also based on a sample of only 29 developing countries, very few of which are in Africa.

9
lobby that can block both policy reform and effective implementation (Bird and Slack 2004).
This is a particular problem in large capital cities where resources are concentrated, and
political and economic elites tend to be especially closely bound together (Jibao and Prichard
2013). Indeed, in such cities, where bureaucratic capacity to overcome some of the
administrative challenges is likely to be higher than elsewhere, elite resistance may even be
the primary obstacle. This resistance can affect not only standard taxes on immovable assets
but also other property-related taxation, such as taxes on the transfer of property (including
capital gains tax) and taxes on rental income.

1.3 Property tax, land and the real estate sector

The property tax regimes in place in a given context are shaped not only by the broader
range of taxes and elite attitudes to tax in general, but also by the nature of the real estate
sector. Issues such as the pace of urban growth, government approaches to managing the
urban land question, and whether there are particular strategies for developing the real
estate sector, all feed into the property taxation problematic. Both property tax morale and
public comprehension of the role of property tax are significantly affected by the existence of
other fees and payments relating to land, and by incentives governments may have put in
place that affect real estate development and its profitability. Moreover, efforts to reform the
way land is distributed, used and valued can provide crucial windows of opportunity for
property tax enhancement and reform, but if not handled carefully can also generate
obstacles to effective property taxation, with long-term repercussions.

This paper will explore these challenges and opportunities through the two strategically-
selected cases of Rwanda and Ethiopia. Rwanda has not only experienced some of the
world’s fastest rates of urban growth in recent years – with an average of 18 per cent per
year in the immediate post-genocide 1995-2000 period marking a particularly acute episode
– but has seen the rapid, government-supported proliferation of high-end real estate
development in its capital, Kigali. This is part of a strategy of economic development focused
on urban service sectors, finance and business tourism, as well as the strategic positioning
of the city as a hub of security within a troubled region (Gatsinzi and Donaldson 2010;
Goodfellow and Smith 2013).

Like Rwanda under the Rwandan Patriotic Front (RPF), Ethiopia under the Ethiopian
People’s Revolutionary Democratic Front (EPRDF) since 1991 has been associated with a
highly ambitious developmental vision, consistent growth and substantial reduction in
poverty, as well a long history of statehood and authoritarian governance (Clapham 2009;
Markakis 2011; Lefort 2012; Gagliardone 2014).3 The transformation of the capital city has
likewise elicited comparison with Kigali. In her work on African urban fantasies, which
highlights the East Asian and Gulf city-state-influenced masterplans and design visions
proliferating in contemporary Africa, Watson (2014) singles out plans for the Rwandan and
Ethiopian capitals as envisioning especially radical transformation. In Addis Ababa the real
estate sector grew by an annual average of 14.1 per cent in the second half of the 2000s –
significantly above the overall GDP growth rate (Access Capital 2010) – and the construction
boom in the city is even more visible than that in Kigali. Both states have also been
undergoing land reform programmes with significant implications for urban development, with
the majority of land being state-owned under previous regimes and a concerted effort to free
up the economic potential of land in recent decades. These historical legacies and current
policies affecting urban land differ substantially, however, as will be explored below.

3
For a discussion of Rwanda as ‘developmental patriominal’ regime, see Booth and Golooba Mutebi (2012). For an
exploration of a range of development issues in Rwanda under the RPF regime, see the special issue of Journal of
Eastern African Studies edited by P. Clark and J. Mosley (Journal of Eastern African Studies 8(2) 2014).

10
In comparing the property tax regimes and efforts to reform them in these two cities, this
paper will explore why there has been very little growth in property taxes in contrast with
many other revenue sources, and despite the proliferation of urban real estate in both cases.
The selection of two states in which the government has demonstrated clear commitment to
economic growth, as well as increasing self-reliance through domestic revenue mobilisation,
throws the relative failure of property taxation into particularly sharp relief. This enables a
comparative analysis with important implications across Africa and beyond.

2 Kigali
2.1 Property tax and the framework for fiscal decentralisation (and
recentralisation)

Rwanda’s decentralisation policy in 2000 heralded major governance reforms through which
substantial resources and responsibilities were devolved to local governments (Chemouni
2014). Law No. 17/2002 established the basic principles for fiscal decentralisation, which
institutionalised various types of central-local transfer and decentralised three taxes
previously collected by the central government: property tax, trading licence tax and rental
income tax (Article 3).4 These were supposed to form the bedrock of local revenue for
districts, alongside a wide range of fees for specific services and rent paid on plots of
leasehold land (discussed below). All property owners with a freehold title were due to pay
property tax, barring a significant number of exemptions (Article 14). Some of the city’s older
properties were already on government registers, though often the valuations were from as
long ago as the 1960s and 1970s (with owners annually paying 0.2 per cent of this value in
tax).5 New properties were supposed to be valued for taxation through a system of self-
assessment based on square metres of land and buildings. Provision was made for
adjustment of the rate at district level, with Kigali having the highest rates. Rwanda’s system
for the valuation of new properties was therefore what Franzsen (2010) terms a calibrated
area-based approach.

Some significant changes to the property taxation framework were introduced by a new local
taxation law at the end of 2011 (Law No. 59/2011). 6 Large numbers of exemptions were
retained (with houses valued at under RWF3 million (around US$4,350) now also exempt),
though there are no longer exemptions for buildings over three storeys. The most important
change brought by the 2011 law, however, was that, in contrast to the previous approach, it
aimed to introduce a tax based on market (capital) value. The market value is defined as ‘the
amount of money for which a property should be sold on the date of its valuation in the open
market by a willing buyer’ (Article 2). The law specifies that market values should be
determined on the basis of self-assessment: every taxpayer must file a tax declaration to the
district in which the asset is located, using an official self-assessment form.7

The law outlines only in very general terms how market values should be assessed in these
declarations. Where the asset has recently been purchased, the purchase price is used to

4
The decision to decentralise rental income tax was highly unusual, and also controversial. See Goodfellow (2012) for a
discussion.
5
Interview with government decentralisation official, 4 February 2010.
6
In this law, property tax is referred to as ‘Fixed Asset Tax’. I will, however, retain the use of the term property tax
throughout this paper for consistency and to avoid confusion.
7
Note that self-assessment is not the same as self-declaration. The latter, where people have to provide basic
information on the house in order for the government to then use it in mass appraisal, is probably more common.
Rwanda’s current approach seems to be more along the lines of that developed in Bogota in the 1990s, in which
property-owners have to declare what they actually believe to be the market value of the property (see Norregaard
2013). In the words of one local tax expert, to base a system on self-reported market values is ‘asking for trouble’
(personal communication, 5 November 2014).

11
denote the value. In other cases, it simply states that the ‘aggregate value of the land, the
building and improvements constitute the market value of such fixed asset’ (Article 17), with
little guidance on valuation procedure. In 2010 Rwanda established an Institute of Real
Property Valuers, and part of the job of the estimated ninety valuers now in the country,8 is to
provide certified appraisals of property values in cases where there has been no sale to base
the value on.9 The district is supposed to review self-assessments within six months of them
being submitted (Article 13), though district data suggests that this very rarely happens.10
Despite these measures, the nascent character of the property valuation profession in
Rwanda and the lack of capacity in districts renders the self-assessment system highly
problematic. Arguably even more problematic, however, is the fact that in the new law the tax
rate was set at 0.1 per cent of the assessed value for all properties, whether commercial,
residential or industrial. This is lower than the 0.2 per cent levied previously, which was
already a remarkably low rate by any international standards.11

The performance of property tax as a revenue instrument has been extremely weak, both
before and after these changes. As demonstrated in Figure 1, in the late 2000s property tax
was at most 4 per cent of locally-collected revenue, with the majority of local government
income instead being composed of various forms of fees.12 In 2012, property tax brought in
less than a third of the amount generated by either market fees or fees for public cleaning
services (see Figure 2). The heavy dependence on fees is problematic. Although user fees
have advantages in terms of providing a direct link between demand and supply of public
services and promoting efficiency in resource use, some authors argue that they do not fulfil
the same functions in terms of accountability, the fiscal contract and state-building as proper
taxes (Lieberman 2002; Moore 2004, 2008; Von Stein [1885] 1964). Moreover, in the
Rwandan context, the fees in question are mostly not user fees of the kind often associated
with services such as water and electricity provision, but compulsory fees paid by all for
neighbourhood services such as cleaning and night-time security. The relatively strong
performance of fees in Kigali reflects a broader problem in Rwanda, where in 2014 fees
amounted to 65.5 per cent of local revenue collection, with taxes only comprising 35.5 per
cent – a ratio that the Rwanda Revenue Authority (RRA) is keen to reverse.13 Improving
property taxation will be central to achieving this, as it is by far the most under-performing of
the three local taxes – trading licence tax and rental income tax have generally combined to
amount to around 25-30 per cent of local revenue in recent years.

8
Interview with property valuer, 5 June 2014
9
It is worth noting, however, that many of these property valuers have been trained in the ‘quantity survey method’ in
order to obtain replacement costs, primarily for the purposes of expropriation. The Kigali Master Plan unveiled in 2007
has led to widespread expropriations in the city, necessitating compensation to large numbers of households, and it is
largely this that has kept many of Rwanda’s valuers busy (interviews with property valuers, 2009 and 2014).
10
GIZ have provided a property valuation manual for district officials, which aims to enable district officials ‘of any
educational background’ to conduct their own approximate valuations to compare with the self-declarations submitted by
property owners. This is based on categorising properties into low, medium and high standard, and then calculating a
per-square-metre value based on professional valuations of such houses in Kigali. Although expressed in
straightforward terms, the procedure is fairly demanding, involving up to 18 steps and numerous calculations.
11
Interviews with various taxation experts, 2009-2015.
12
Note that in Figure 1, the share of local revenue labelled ‘rent on plots’ refers to the land lease fee, discussed in the
following section.
13
Comment made by Richard Mushabe, Commissioner General of Rwanda Revenue Authority, 4 June 2014.

12
Figure 1 The composition of local revenue in Kigali’s three districts, 2008-9

Rental tax
Property tax
7%
2% Trading licence
9%

Kicukiro District
(Jan-Sep 2009) Fees and other
income Rent on plots
44% 21%
Buiding
permits
Nyarugenge District 3%
Rent on plots
(2008) (arrears)
Notary charges
Stadium 3%
1% 10%
Property Rental Property Tax
Building Mkt Admin Rental
fees Tax Income Tax 4%
Notary permits fees income tax
0% 3% 6% 3% 12%
Gasabo District 12%
charges
9% (2008)
Billboard
tax Admin and Trading
1% other fees Licence
31% 11%
Rent on Trading
plots licence
4% 18%
Market Fees Rent on
Rent on Other
Fines and 5% Plots
shopping revenues
Penalities 33%
centres 42% Notary
1% 1%
Charges billboard tax
3% 1%

Source: data collected from district offices, 2009-10.

13
Figure 2 Local revenue for all districts in Rwanda combined, 2012

Source: produced using data from Cyan et al. (2013)

Data on property tax is extremely patchy and problematic,14 but in the most recent local tax
data available for Kigali at the time of this research (2012/13), there were some indications
that property tax may be increasing its share (see Figure 3).15 Meanwhile, 2012 data for the
whole country (see Figure 2) indicates that 4 per cent of local revenue countrywide
comprised property tax; this suggests that the proportion would be significantly higher in
Kigali because virtually all of the country’s property tax is collected in the capital – indeed, 96
per cent of land and property-related taxes and fees are collected there.16 Many of the
districts outside Kigali have literally zero registered property tax payers; even some districts
containing reasonably large towns have as few as three.17 Nevertheless, even if we assume
that property taxation has recently been closer to 10 per cent of Kigali’s locally-collected
revenue, this is still a far cry from what might be expected of the tax in a capital city, even by
regional standards. In neighbouring Uganda, which has enormous local taxation problems of
its own, property tax amounts to around a third of local revenue in Kampala.18

14
Despite Rwanda’s relatively intensive local governance structures, property tax data can be difficult to interpret for a
number of reasons. First, it was until very recently collected at the district level, and districts often had different reporting
systems and used different names for certain taxes and fees. Second, the major land tenure regularisation programme
over the past decade has meant that ‘old’ and ‘new’ property titles often do not refer to exactly the same land plots and
have different identification numbers, which has created substantial problems for reconciling data. Third, the shift from
French to English as the main official language (alongside Kinyarwanda) means that the names of different taxes and
fees relating to land can be difficult to distinguish in district-level data (this is not aided by the fact that the word used for
fee in French is taxe).
15
This data is especially problematic because it is not clear whether the revenue coded 113104 is actually property tax
(fixed asset tax) or in fact land lease fees, probably due to the issues discussed in footnote 12 above.
16
Interview with taxation adviser, 8 June 2014.
17
Interview with researcher, 9 June 2014.
18
See Goodfellow (2012) for a discussion.

14
Figure 3 Local revenue for Kigali districts, 2013

Source: Integrated Financial Management Information Systems data from Kigali’s three districts, collected in 2014.

In short, to quote a recent report by a group of local revenue specialists, Rwanda’s fixed
asset tax is severely underutilised (Cyan et al. 2013). It is also not increasing much at all;
despite the problems involved in interpreting property tax revenue over time, evidence
suggests that property tax grew by 0 per cent in 2011/12 (Cyan et al. 2013), and even
declined in the year after the new law came into force as the tax rate was set so low.19 The
missed potential is further highlighted by the fact that the number of taxable items eligible for
property tax – the tax base – increased by 15 per cent in 2011/12, and this has apparently
not resulted in revenue increases (Cyan et al. 2013). This 15 per cent also only includes
those properties that are officially eligible for property tax – freehold properties – which does
not reflect the reality of property development in the city. Meanwhile, local revenue growth
has primarily been in fees (for example for public cleaning), and where taxes have grown it
has primarily been in the less progressive instruments such as the trading licence tax (Cyan
et al. 2013).

The relatively poor performance of local taxes led to a controversial decision in 2014 to
involve the RRA centrally in the collection of all local taxes. Without attempting to alter the
2011 fiscal decentralisation law, the government signed a one-year extendable
Memorandum of Understanding with all districts in March 2014, enabling the RRA to take
over the management of local tax collection on their behalf.20 Opinions differ as to the likely
efficacy of this move in terms of long-term improvements in local tax collection. The present
research was conducted when data was not yet available to assess whether property tax
intake had significantly increased as a result of this change, though the expertise and

19
Interview with researcher, 9 June 2014.
20
Interview with finance official, 3 June 2014.

15
perceived weight of the RRA is likely to have improved efficiency in administration and
bolstered compliance. Some observers felt, however, that even if tax administration improves
under this system, the amount collected might not actually increase: while local governments
were previously desperate to secure their own funds so worked hard to collect taxes, now
there would be ‘less incentive and commitment’.21 There might be a risk that with the RRA at
the helm ‘citizens will give less credit to districts’,22 and the public might be less willing to pay
local taxes to an authority that many people already feel is extracting too much from them.23

By January 2015, there were rumours that the RRA was not likely to meet its targets for
property taxation under the first year of this system.24 Meanwhile, there were moves to
entrench the RRA’s role as collectors of district taxes as a more permanent arrangement,
bolstering fears about decentralisation going into reverse.25 In any case, the centralisation of
local tax collection, while it is likely to streamline administration, reconciliation and reporting,
and may (or may not) improve enforcement and compliance, does not address some of the
most fundamental problems affecting Rwanda’s existing property tax regime. These relate
not only to the fundamentals of the tax policy – including the extremely low tax rate and the
dependence on self-assessment of market values without robust valuation procedures – but
to the way in which property taxation interacts with land lease fees, and the government’s
broader strategy for managing urban land. It is to this critical question that we now turn.

2.2 Taxation and land tenure regularisation: ‘there is huge confusion’

The ongoing effort to reform the property tax system since the early 2000s26 – in which the
2011 law was merely a punctuation mark – cannot be viewed in isolation from the major
reforms to land taking place in Rwanda over the same period, especially as they pertain to
Kigali. It is critical to situate the property tax problematic in relation to the land lease system
that was introduced in the context of Rwanda’s major land reform in 2004-5. The 2005
Organic Land Law introduced a system whereby investors in urban land who meet certain
specifications are entitled to freehold titles, while all other land holders are granted an
‘emphyteutic lease’ of variable length (most land having been officially owned by the state
since 1976 (Sagashya and English 2009). This grants them leasehold rights in exchange for
paying a lease contract fee and developing the land in accordance with their lease
conditions. These conditions relate to its classification (as commercial, residential, etc.) and
any zoning rules, such as those accompanying the Kigali Master Plan. The critical point here
is that those with leasehold titles pay land lease fees, while holders of freehold properties –
which constitute only 3 per cent of land parcels in the country (IMF 2014) – pay property
tax.27 The one replaces the other; nobody should have to pay both.

This system has a certain logic, but it is one relating to Rwanda’s land scarcity and
development needs rather than to fiscal considerations. As specified in the Organic Land
Law and rolled out through the subsequent Land Tenure Regularisation Programme (LTRP),
financed primarily by the UK’s Department for International Development (DFID), all those

21
Interview with former taxation official, 4 June 2014.
22
Interview with taxation adviser, 8 June 2014.
23
The RRA takeover of local tax collection has also resulted in considerable anger and disappointment among those who
had fought hard to institutionalise fiscal decentralisation in the first place, with one commenting that ‘it was supposed to
be building local capacity! … We wanted local government to raise their shoulders and say “we can do this”’. There had
been a major battle to bring local taxation powers to districts in the first place, and a shift in the balance of power at the
ministerial level in recent years has turned the tide back against fiscal decentralisation. In short, ‘those people who were
against fiscal decentralisation, now they are in positions to make new decisions’ (Interview with former taxation official, 4
June 2014).
24
Interview with taxation adviser, 20 January 2015.
25
Interview with taxation adviser, 20 January 2015; interview with former taxation official, 4 June 2014.
26
See Goodfellow (2012) for a discussion of this.
27
In Kigali there are 2,905 properties held in freehold and therefore eligible to pay property tax, compared with 258,724
held under leasehold.

16
who prior to 2005 held land through customary tenure (which was most of the population)
were given leases of twenty to ninety-nine years depending on land use. Most residential and
commercial plots were granted leases of twenty to thirty years depending on location, with
those in the city centre mostly just twenty years.28 To convert the leases into a full freehold
title, landholders have to pay a fixed land lease fee in full (determined by the size of their plot
and its location), as well as using or developing their land in accordance with approved
plans. As most people cannot afford to pay this entire fee upfront, they may pay it in
instalments over ten years. After paying this off, as long as its usage harmonises with district
or city plans, they are eligible for a freehold title and must pay property tax.29 This was
intended as a way of spurring planned development of every plot, while also ensuring a
steady flow of revenue from land based on simple square-metre calculations.

The whole idea rests, however, on the perception of freehold ownership as being something
desirable. The assumption is that people will be keen to convert their lease into a full freehold
title for the security this provides them over their land – an assumption that it is difficult to
disassociate from the influence that de Soto’s (2000) ideas on formal property titling have
had among donors – and this will spur both development and the broadening of the property
tax base. The experience so far in Kigali is not, however, bearing this out. Officials
concerned with land and housing in the city affirm that even those eligible to pick up a
freehold title are not doing so, primarily because if they did they would have to pay property
tax.30 This was affirmed by other sources including real estate brokers and a lawyer. 31

The result was therefore a fiscal situation in which two identical properties could be paying
very different amounts of tax/fees simply based on whether or not they have picked up a
freehold title, with the latter paying a lease contract fee that effectively exempts structures on
the land from any taxation whatsoever. Calculations suggest that many property owners with
leasehold titles are likely paying around a fifth of what a freeholder would pay in property tax
on an identical plot.32 The former could be receiving exactly the same services as the latter;
as Cyan et al. (2013) notes, ‘Land tenure choices do not affect the amount of district services
received by occupants of structures in a given year’ (Cyan et al. 2013: 149). The existing
situation was therefore one in which, in the words of one observer, ‘more or less only the
people who are stupid enough to pay, pay’.33 The supposed benefits of freehold ownership
were not sufficient to incentivise most people to upgrade to freehold in spite of the tax
implications.

The reasons why people should aspire to freehold ownership are also not being clearly
articulated in a way that resonates locally. For example, the argument made by one official
that freehold is clearly desirable because it means the land ‘is yours forever’,34 is problematic
in a context where sources suggest that people are used to thinking that they could use their
land permanently anyway.35 Meanwhile, somewhat paradoxically, other sources suggest that
people don’t really understand the idea of freehold ownership because they would never
expect to own land outright.36 Part of the problem is therefore that both the concepts of
freehold and leasehold – and the supposedly significant distinction between them – are
essentially foreign to the Rwandan context and contribute to the lack of public understanding
about the differences between the local taxes and fees they are now expected to pay. One
long-term advisor on decentralised revenue confirmed that ‘there is huge confusion over

28
Interviews with various city officials, June 2014.
29
Interview with government decentralisation official, 4 February 2010; interview with land official, 9 December 1011.
30
Interview with official, 4 June 2014.
31
Interview with lawyer, 2 June 2014; interview with property broker, 5 June 2014.
32
Interview with RRA officials, 4 June 2014.
33
Interview with taxation adviser, 8 June 2014.
34
Interview with city official, 6 June 2014.
35
Interview with land and housing official, 8 June 2014; Interview with researcher, 9 June 2014.
36
Interview with estate agent, 6 June 2014.

17
what is property tax and what is Location Parcelle [the land lease fee]’.37 This is further
compounded by the fact that there is also great confusion over the distinction between
property tax and rental income tax. The (highly unusual) decentralisation of the latter tax
means that when property-owners who rent out their property are asked by districts to pay
both, ‘they think it’s double taxation, when it isn’t’.38

It is not only among the general public that the distinction between land lease fees and
property taxation is poorly understood. There are also widely diverging understandings of
property tax among those who are supposed to administer it; in the words of one observer,
‘not even the districts can tell the difference between property tax and [land lease fees] most
of the time’.39 Other government officials believed that the 2011 local tax law suggested that
everyone should pay property tax, whether holding a leasehold or freehold title.40 Meanwhile,
people working in urban real estate brokering and sales were likewise apparently confused
by the distinction, suggesting that if you own a house and have freehold title but live in the
property yourself (or it is unoccupied) then you pay nothing at all on that house in tax.41 This
reflects a common belief within the city’s property market that leaseholders pay land lease
fees, and freeholders who rent out should pay rental income tax, but there is nothing in-
between for those who hold freehold titles but don’t rent out the property. In other words, as
far as many actors are concerned, property tax does not exist.

As a result of this widespread divergence in interpretation between different actors, it is not

just the understanding but the reality of taxation that differs from the official rules regarding
who should be paying what. Thus some people do pay property tax on leasehold properties
even though they are not required to. One estate agent and valuer insisted that he had just
valued a property for exactly this purpose, at the district’s behest, despite the owner only
holding a leasehold title.42 Meanwhile, a source with detailed knowledge of local taxes
pointed out that with an undervalued property and given the very low tax rate in place, some
people were paying less in property tax than they would have been if they were paying the
land lease fee.43 Hence some leaseholders opted to pay property tax. This was reflected in
considerable disparity between the city-level taxation data and land registry figures on the
number of freehold relative to leasehold titles.44 Essentially, the lack of understanding about
the distinction provided opportunities to ‘game’ the system: regardless of whether they
actually held freehold or leasehold titles, some people were able to choose between paying
land lease fees or property tax based on which cost them the least.

In addition to these ongoing problems, there were missed opportunities with regard to how
the LTRP was conducted. Starting in earnest in 2009, the LTRP process provided a window
of opportunity to register properties for taxation purposes: regularising plots and allocating
land titles could be combined with registration of properties on the land and perhaps even the
gathering of data for the purposes of valuation.45 The government, however, was concerned
that people might be discouraged from participating in the LTRP if they knew it could lead to

37
Interview with former taxation official, 4 June 2014.
38
Interview with former taxation official, 4 June 2014.
39
Interview with researcher, 9 June 2014.
40
Comment by government official, 4 June 2014. In some respects this is understandable, as the text of the 2011 tax law
does not actually say anything explicit about the distinction between freehold and leasehold. In defining what constitutes
an asset eligible for property taxation, the law simply says that the tax should be levied on all land and buildings ‘for
which the owner has obtained a title deed’ (Article 6), which could be interpreted as a leasehold title.
41
Interview with property broker, 5 June 2014.
42
Interview with estate agent, 6 June 2014.
43
Interview with taxation adviser, 8 June 2014.
44
Interview with taxation adviser, 8 June 2014.
45
Interview with taxation adviser A, 26 November 2009. The LTR progress was taking place with impressive speed; while
in February 2010 only 18,000 titles (mainly leasehold) had been issued on Rwanda’s 10 million plots, by November
2011 a million such titles had been issued and the process was ongoing (Interview with land official, 9 December 2011).

18
their properties being registered for taxation,46 and the pilot phase confirmed that survey
respondents were worried higher rates of tax might emerge from the process (Sagashya and
English 2009).

Consequently, despite a very impressive land regularisation programme, the land registry
contains no information on what is built on the land; there was no cadastre of real property
developed through the process that would facilitate ‘switching on’ a property tax at a later
stage.47 Unsurprisingly, some long-term observers have suggested that behind all this lay a
deeper political resistance to developing the tax.48 The lack of integration of property taxation
concerns into the LTRP goes some way to explaining why the system introduced in 2011
was entirely dependent on self-assessment. Moreover, although by 2014 the official
regularisation of land parcels in Rwanda was said to be complete, the parcelling of land is
only half of the equation; the corresponding titling of land was progressing more slowly and
this, of course, was the more relevant process for fiscal purposes.49

A further complication arose through a legal change in a 2013 land law (Law No, 43/2013),
which updated and abrogated the 2005 land law. This introduced a prohibition on foreigners
owning freehold land, except in the Free Trade Zone on the outskirts of Kigali.50 In theory,
this meant that regardless of the nature of their development, foreign investors would only be
granted leases of up to forty-nine years, and by definition therefore not be subject to property
taxation. The implication was that they would continue paying land lease fees, but the
mechanisms for this were unclear and a source of further confusion. Several sources
indicated that the ability of foreigners to get freehold titles was still a grey area.51 This legal
change, made on the basis of rules elsewhere in the East African community (and in some
people’s view introduced to make expropriation of land easier),52 was clearly enormously
problematic from a fiscal perspective under Rwanda’s system, pulling against efforts to
increase property taxes. It exemplifies the way in which policies around land and around
taxation were not really speaking to one another – a problem that is often a particular feature
of heavily donor-dependent countries, where different donors working on specific policy
issues in ‘silos’ can result in systems with little effective coordination and symbiosis between
them.53

As of mid-2014 the rules were in the process of changing again, to the effect that foreign
investors would be allowed to get freehold titles as long as they had fulfilled planning
obligations up to the point of securing an occupation permit. Moreover, under the emerging
rules – already being informally implemented by the city council but not yet codified in
legislation – if any investor (foreign or domestic) has built their property to the final stage of
occupancy, they would be granted a freehold title free of charge, without paying their ten
years’ worth of lease fees.54 These latest changes suggest a new momentum with regard to
increasing the number of properties held under freehold, reflecting the government’s
awareness of the need to both attract investment and substantially broaden the property tax
base.

46
Interview with government advisers, 23 November 2009; interview with officials at the National Land Centre, 3
December 2009.
47
Interview with city official, 5 June 2014.
48
Interviews with revenue advisers, November 2009.
49
Interview with taxation adviser, 8 June 2014
50
Interview with lawyer, 2 June 2014.
51
Various interviews, June 2014.
52
Interview with taxation adviser, 8 June 2014.
53
Interview with international consultant, 19 January 2015.
54
Interview with senior city planner, 6 June 2014.

19
3 Addis Ababa
3.1 Property tax and the framework for fiscal federalism (and recentralisation)

Since the EPRDF took power in 1991, Ethiopia has instituted a radical system of ‘ethnic
federalism’, whereby the country is divided into nine federal states and two separately-
administered chartered cities, one of which is Addis Ababa. This federal system has
important fiscal implications, the basic parameters of which are set out in the 1994
Constitution. The governments of federal regions and cities possess substantial powers to
tax the incomes of employees within their regional state, as well as the profits of enterprises
located within them and various other revenue sources usually collected at the national level.
The Addis Ababa City Administration (AACA) has powers to mobilise substantial revenue of
its own from income (including rental income),55 business profits, and VAT on certain
products and services. Thus unlike cities in decentralised systems – or indeed most urban
areas in Ethiopia, which acquire 70 per cent of their resources in transfers from the regional
states in which they are located – there are no transfers of revenue to Addis Ababa.

In addition to the state taxes outlined above, there are a series of municipal revenues
collected locally. While state taxes are collected by regional branches of the Ethiopian
Revenue and Customs Authority (ERCA), municipal revenue is collected by the very lowest
rungs of local government (woredas or kebeles, depending on the part of the country) as well
as, in urban areas, the sub-city government offices that sit above these. In the case of Addis
Ababa, all the taxes and fees collected at woreda or sub-city level end up at the Bureau of
Finance and Economic Development at the AACA for budgeting and redistribution, along
with the state taxes and other revenue such as external loans.

Among the municipal taxes sits property taxation. Currently there is nothing actually called
property tax, though there is a municipal tax known variously as ‘roof tax’ and ‘city house
tax’.56 Although today there is no single revenue instrument based on the combined value of
land and buildings, there was in the past. The idea of property taxation in Ethiopia dates back
to 1937, under the Italian occupation, and was from the very start based on annual market
value rather than capital value.57 After the socialist revolution in 1974, when land was
nationalised, a new law split property tax into roof tax and land rent – land itself no longer
being taxable property but something rented from the state. In theory, the roof tax was still
based on annual rental value even after the socialist revolution, with rates ranging from 1 per
cent to 4.5 per cent of annual value depending on a range of criteria.58 This tax was to be
paid once annually.

Soon after the EPRDF displaced the socialist Derg regime in 1991, it became apparent that
the property values being used for the purposes of taxation were extremely out of date and
only captured a fraction of the city’s fast-growing housing stock. Consequently, a major
census of properties in Addis Ababa was undertaken in 1996, using a computerised system
and a team of 3,000 enumerators. Through this exercise all properties (including informal
structures) were captured. A valuation method was adopted whereby characteristics of the
building and locality were assessed to calculate a capital value, which was then translated
into an imputed annual value through a formula. This resulted in a new tax bill being
presented to all property owners. It did not go down well, because it indicated that the
property tax they were currently paying was only around 25 per cent of what it should be

55
Details of the division of taxation of different kinds on income between the federal and regional governments can be
found in the Income Tax Proclamation (No.286/2002).
56
Hereafter where the term property tax is used, this refers to the roof tax/city house tax.
57
Interview with property taxation expert, 26 September 2014.
58
Interview with property taxation expert, 26 September 2014.

20
according to the recalculated value of their property. There was such an outcry about this
that the city government instructed the valuation team to slash the values to 25 per cent of
what had been painstakingly calculated, rendering the whole exercise essentially pointless
for properties already on the register. The fact that vastly more buildings had now been
captured meant there was, however, a step change in property tax take, which leapt instantly
from Birr 4 million to Birr 50 million in 1999 – though without the artificial reduction in property
values, this would have been Birr 200 million.59

Since 1996, there has been no revaluation of properties in the city.60 Meanwhile the
population of Addis Ababa has almost doubled, from 1.8 million in 1990 to 3.2 million in
2014.61 Even for those houses that the city authority does have on its register, severe
undervaluation is the norm. One 2006 study of a particular sub-city of Addis Ababa (Lideta)
examined cases where properties listed on government valuation rolls had also been valued
by banks, and found that on average the banks’ valuations were 300 per cent higher. For
some properties, the banks’ valuations differed by thousands of per cent, with one being
7,685 per cent higher (Sisay 2006, cited in Yusuf et al. 2009: 60). This disparity is likely to be
even higher today. When new structures are built, property owners can – in the words of an
expert – get someone to come and value the property ‘if they want to pay the tax’.62 Evidently
(and unsurprisingly) very few do. The general consensus was therefore that most owners of
privately-owned houses ‘pay nothing, because the government does not know them’.63

This point about private house ownership serves to distinguish such houses from those
owned by the state, which constitute a substantial proportion of the housing stock due to the
nationalisation of housing as well as land under the Derg regime. Many houses dating from
this period are still government-owned, and some low- and middle-income city-dwellers live
in these and pay a small amount of annual rent for them – often as little as Birr 40
(approximately US$2) per month for the most low-end houses (known as kebele houses).
Meanwhile, plots of land that have not been formally leased under the new leasing system
(discussed below), and are therefore still fully owned by the state, are subject to small
amounts of land rent paid annually. When these rent streams were introduced, neither bore
any relationship to market value (Yusuf et al. 2009) – and they certainly do not today, having
barely increased over time.64

Each of these three sources of local revenue relating to property – roof tax, rent on houses
and rent on plots of land – constitute a miniscule proportion of Addis Ababa’s budget.65 As
demonstrated in Figure 4a, as a proportion of total city revenue (including all state revenue
and municipal revenue) the contribution of each of them is well under 1 per cent. Even when
state taxes (which constitute the vast majority of Addis Ababa’s revenue, being around Birr
60 billion annually as compared with Birr 3 billion from municipal sources) are removed from
the picture, none of the three payments noted above constitutes more than 2 per cent (see
Figure 4c). In any case, only one of these is an actual tax – termed ‘city house tax’ in the
revenue data – with the other two being rents or usage fees. This taxation stream constitutes
a tiny proportion of city revenue by any standards, even with some of the largest state taxes
such as income tax and VAT removed (see Figure 4b), or all state revenue removed (Figure
4c). To put this in perspective, Addis Ababa collects more revenue from ‘engineering and
asset estimation service fees’ or ‘fees for medical examinations’ than it does from property

59
Interview with property taxation expert, 26 September 2014.
60
A pilot programme to reform property taxation is currently underway (discussed in Section 4), but this will be
implemented only in three secondary cities in Ethiopia, excluding Addis Ababa in the first instance.
61
UN population division data, available at:
<http://esa.un.org/unpd/wup/wallcharts/WUP_2014%20Urban%20Agglomerations%20Wallchart.pdf>.
62
Interview with property taxation expert, 26 September 2014.
63
Interview with local researcher, 26 September 2014.
64
Interview with tenants, 2 October 2014.
65
Land lease fees are a very different story, discussed separately in the subsection below.

21
taxation. The situation in the capital is also significantly worse than in some other federal
regions, where regional governments have used their devolved powers to substantially
increased roof tax rates and also apply taxes to land. One source pointed out that ‘I pay four
times more property tax in Adama for a small house than I pay for my big house in Addis
Ababa’.66

Figure 4a Addis Ababa, total revenue 2013/14

66
Interview with official in international agency, 24 September 2014.

22
Figure 4b Addis Ababa revenue minus income and profit tax, VAT, non-tax revenue
and loans, 2013/14

Figure 4c Addis Ababa municipal revenue only, 2013/14

Source: Figures 4a-4c are compiled from data collected at the Addis Ababa City Administration in 2014.

23
In short, despite the thorough efforts made in the 1990s to capture and revalue properties,
property tax in Addis Ababa has been rendered ‘almost non-existent’67 by interventions to
deliberately undervalue properties, lack of revaluation over time and the failure to capture
virtually all new properties constructed in the past two decades. As well as privately-built
houses, this failure to capture properties on the register includes the approximately 100,000
condominiums built by the government over the past decade, which (despite being targeted
at lower-income groups) have actually proved lucrative revenue sources for the owners and
have not been registered, let alone valued, as taxable assets.68 In the words of one
government official, the problem of property taxation in the city ‘is untouched’.69 Meanwhile,
the fact that it constitutes just one of several small payments relating to land and property –
all of which are often paid to the kebele simultaneously70 – also provides scope for confusion.
How all these payments are supposed to relate to each other is unclear to many, causing
potential for misinformation at the local government level as ‘there is no one source of
information so you can’t trace the payments’.71

Interestingly, as in Rwanda, there has been a recent shift in fiscal governance, with the
central branch of the ERCA taking over the administration of taxes for Addis Ababa.72 As in
the Rwandan case, the central revenue authority currently manages the collection and then
remits these taxes directly back to the city authority for budgeting. This is supposedly a
temporary arrangement, based on the perception in central government that the AACA
lacked capacity to collect its own taxes. As of late 2014, however, this system had already
been in place for three years with little sign of a shift back towards collection at the city level.
This does not seem likely to make much difference to property taxation in the short term,
though it may be having an impact on other sources of revenue relating to land and property,
including rental income tax and fees generated by the government’s land leasing
programme. While a detailed discussion of the former is beyond the scope of this paper, the
latter constitutes a vital and rapidly-growing revenue source that requires examination if the
property tax problematic is to be adequately understood.

3.2 Taxation and land leasing: ‘people are already feeling taxed’

Although relinquishing national ownership of land was considered out of the question when
the EPRDF came to power in 1991 (Markakis 2011), the incoming regime found itself faced
with a land use system that neither yielded much revenue nor provided the foundations for
the rapid development it was determined to promote. Existing land use was governed by a
permit system, in which the government would issue permits for people or organisations to
use particular plots of land for indefinite periods in exchange for rent, which (as noted above)
bore no relation to market value.73 Within the formal property market at least, there was
therefore no mechanism for capturing the value of land. The new regime wanted to combine
continued state land ownership with market-driven development, and in pursuit of this
objective soon began investigating the possibilities for a system of land leasing along the
lines of what was happening in other countries transitioning from socialism.74

Based on a series of study visits including several to China, the AACA led the way in
developing a land leasing system designed to unlock the market value of land. The first
expression of this was Proclamation no. 80/1993, which provided for government to transfer
rights over land for specified periods of time while retaining the ultimate title. One of the
67
Interview with local researcher, 23 September 2014.
68
Interview with housing expert, 1 October 2014.
69
Interview with tax official, 23 September 2014.
70
Interview with taxation expert, 22 September 2014.
71
Interview with taxation official, 29 September 2014.
72
Interview with taxation expert, 22 September 2014.
73
Interview with land expert, 4 October 2014.
74
Interview with land expert, 4 October 2014.

24
problems faced in implementing this was the lack of any existing basis for pricing land
leases. Based on experience elsewhere, the government approached this by setting
benchmark prices for land per square metre, based partially on the estimated cost of
infrastructure provision and resettlement expenses for those with pre-existing permits to use
the land (Yusuf et al. 2009). This benchmark price would ordinarily be the minimum possible
cost per square metre for a lease, and would apply to certain kinds of favoured development.
Many commercial developments, on the other hand, would be expected to pay more. Initially,
the modality for establishing prices for most commercial developments was exclusively
through auction, with the benchmark as the starting price. Again following the example of
China, however, a modality of negotiation was introduced in 2002 alongside auctioning,
though the basis for such negotiation was never entirely clear (Yusuf et al. 2009).

The general lack of understanding of public leasehold systems – and the fact that having to
purchase leases resulted in radically increased costs for usage of urban land – caused
considerable resistance to the new system. To ease its passage, the permit system was
maintained alongside the leasing framework, so that people already holding permits could
continue to use their land in exchange for small amounts of rent while the government tried
to conceive a way to establish existing land into the leasehold system. Meanwhile, the
emerging system whereby land lease prices were established by auction for certain kinds of
projects and by negotiation for others (depending on whether the project in question was
linked to government priorities) resulted in highly erratic pricing. Most land in the period
2002-2009 was leased through negotiation, with a study of 1,000 land transactions finding
that 96 per cent were negotiated and the remaining 4 per cent split between allotment at
benchmark prices and auctioning (Yusuf et al. 2009). According to the same report, ‘Little
importance is attached to fair dealing’ (Yusuf et al. 2009: 103), and land lease prices followed
‘neither logic nor theory, neither policy priorities nor grades, neither floor prices nor size of
plots. In most cases, a homogeneous plot of land does not fetch a similar price’ (Yusuf et al.
2009: 105-6). These disparities are clearly illustrated in Table 1. By this time, however, land
leasing was already generating far more revenue for local governments than property tax or
any other taxes relating to property (see Figure 5).

Table 1 Summary statistics of plots of land in the ten sub-cities in Addis Ababa

Sub-city Cases Total land made Average price/m sq Min/max price range
available (m sq) (Birr/m sq)
Arada 19 434,688 931 210-3,984
Akaki/Kalilti 89 471,254 331 126-2,747
Lideta 18 67,943 1,207 137-3,601
Kolfe-Keranyo 157 535,143 231 108-3,140
Kirkos 63 125,092 1,293 118-4,310
Bole 469 1,783,213 322 115-6,269
Yeka 72 855,577 427 92-2,863
Gullele 12 26,509 577 250-1,375
Nefas Silk/Lafto 126 680,090 490 102-3,737
Addis Ketama 10 28,069 1,981 353-4,100
Total 1,035 5,017,578 454

Source: based on Yusuf et al. (2009: 105).

In 2011, a new proclamation took the leasing system a step further through a controversial
measure to try and incorporate all land gradually into the leasing system. The Land Lease
Holding Proclamation No 721/2011 specified that ‘no person may acquire urban land other
than through the lease holding system’ (Article 5). This meant that whenever a plot of land
held under the old permit system was sold, instead of the permit transferring as part of the
sale (as had happened previously) the new landholder was required to take out a lease at
the benchmark price, which is updated every two years.75 At least 10 per cent of the lease

75
Leases in Addis Ababa are: 99 years for residential housing, science and technology, research and study, government
offices, charities, etc; 70 years for industry; 60 years for commerce; and 15 years for urban agriculture (Article 18).

25
price must be paid as a down payment (Article 20), and the rest can be paid either upfront or
in instalments over fifteen years, with interest.76 These measures contributed to the ongoing
increase in amount of local revenue garnered through land leasing, and the notable uptick
from 2012 apparent in Figure 5.

Figure 5 Addis Ababa selected revenue over time, 2006-2014

Source: data collected from Addis Ababa City Administration.77

Probably more significant in terms of this increase, however, was the fact that the 2011
proclamation also abolished the modality of negotiation, making all commercial land leasing
dependent on auction.78 The average price of leases sold through auction was always
significantly higher than those sold through negotiation (Yusuf et al. 2009), so subjecting all
land purchases (other than those eligible for allotting at benchmark prices) to compulsory
public bidding in monthly auctions79 paved the way for prices to escalate rapidly. Indeed, in
contrast with even the highest square metre prices displayed in Table 1, by late 2014 plots of
land were selling for as much as Birr 65,000 in Bole sub-city, and Birr 60,000 around
Lideta.80 Some sources claimed that land prices were higher than in New York or Geneva.81
If that was not true at the time, it soon came to be so: on 5 December 2014, a plot of land in
Addis Ketema sub-city was auctioned for Birr 307,000 per square metre – around
US$15,500, which is higher than the average price for developed real estate in Geneva
($15,250), which is the fifth most expensive city in the world for property.82

These soaring prices will clearly contribute to increased revenue for the city administration,
particularly given the undeniable boom in construction and demand for land across the city.
They do, however, raise all kinds of questions about the stability and sustainability of urban
76
Interview with finance expert, 23 September 2014.
77
Note that there is one year (2007-8) for which I was not able to get this data.
78
This decision was taken partly due to concerns about the complexity of negotiating land prices and the scope it allowed
for corruption, and followed the experience of the Chinese who had similarly abolished their negotiation modality in
2000.
79
Interview with city land official, 29 September 2014.
80
Various interviews, September-October 2014.
81
Interview with official in international organisation, 1 October 2014.
82
<http://asokoinsight.com/news/official-addis-one-expensive-cities-world-ethiopia/>.

26
revenue over time – quite aside from the effect that this kind of land-based financing might
have on displacement of the urban poor, the path to which is eased by new rules on
government powers to ‘clear’ land in Proclamation No 721/2011. As sources of government
revenue, leases are fixed sums of money rather than continuous payments like taxes. They
are also determined at a specific point in time and therefore do not change over time with
market values of the land. In terms of revenue stability, sales may vary hugely from one year
to the next, and urban land is ultimately finite. Moreover, land lease fees will not recapture
any of the value of structures built on the land; for this a real property tax is required.

Figure 5 starkly illustrates again how, despite the (often dramatic) increase in other revenue
linked to property – a clear indication of how much the property market is booming – the
existing City House Tax has remained virtually stagnant under the existing system. There
are, however, important questions regarding the feasibility of introducing an effective
property tax system when people are, in the words of one source, ‘already feeling taxed, due
to the new land lease system’. All people who acquire new property find themselves now
paying lease fees substantially higher than any nominal land rent they had paid previously,
which makes the introduction of an effective property tax highly challenging. Making this all
the more problematic is the fact that when the controversial lease fee system was
introduced, explicit public announcements were made as part of the hard sell that people
would not be taxed on their property once they had paid for the lease. Political battles over
municipal revenue generation are therefore brewing for the future, even as current revenue
continues to rise due to land leasing. Despite this, there are important opportunities arising in
Ethiopia, as in Rwanda. The final section of the paper now turns to a comparative reflection
on the potential for, and obstacles to, property tax reform in each of the two cities.

4 Reform: opportunities and obstacles in

comparative perspective
Given the gaps between potential and actual revenue from property taxation, it is
unsurprising that in both cases international financial organisations and consulting firms have
identified property tax as a critical area for reform. The economic, political and technical
challenges to realising this differ in each case, however. This section discusses the potential
and challenges in Kigali and Addis Ababa in turn, before considering some cross-cutting
issues that highlight what can be learned from comparing the two cases and suggesting key
areas of focus for political economy analysis of property taxation more generally.

4.1 Rwanda

In 2014 the government of Rwanda invited the International Monetary Fund (IMF) to conduct
a study on revenue potential, with the aim of furthering its long-term agenda of reducing aid
dependence. Although the RRA is widely considered a success story, Rwanda has a low tax
effort even compared to other countries in the region.83 Along with taxes on agriculture and
mining, property tax was selected by the Rwandan authorities as an area for exploration by
the IMF, whose study found that ‘the area with the largest potential for collecting tax revenue
with the least amount of economic distortions is property or land taxation’ (IMF 2014: 10).
Other taxation specialists working in the country have long been aware of this untapped
potential; the property development taking place in Kigali over the past two decades led one
such expert to suggest in 2009 that in Rwanda property taxation could, hypothetically

83
Tax effort is the ratio of actual tax collection to GDP relative to the tax capacity of a given country. Tax capacity is
defined as the maximum tax/GDP ratio that is attainable given the economic structure of the country.

27
speaking, amount to as much as 5 per cent of GDP.84 In reality, in 2013 the amount of
property tax collected nationally (the vast majority of which in Kigali) was just 0.018 per
cent.85 This is extraordinarily low even by the standards of developing countries, a selection
of which was found to average at 0.6 per cent (Norregaard 2013), or African countries, for
which the overall average is estimated at 0.5 per cent (Fjeldstad and Heggstad 2011).86

The IMF team concluded that all property in urban areas (and especially Kigali) should be
liable to pay property tax, irrespective of the type of land tenure. They also argued that there
is a lack of capacity at the local level to administer property tax, and that the administration
could be centralised within a central valuation department in the ministry that deals with land
issues (IMF 2014). Similar conclusions were reached by another study, which also
suggested that all properties should be defined as taxable (Cyan et al. 2013). Despite
obvious reasons for issuing such advice given the disparity between land lease fees and
property tax, some sources were concerned that such a change could actually result in
negative impact on revenue generation. The problem of compliance would not necessarily be
improved by putting all properties under real property tax. Moreover, while land lease fees
are based on simple calculations, ‘As soon as you put things under real property tax you
have the problem of how to value the property’,87 and the question of a suitable valuation
system for Rwanda is still far from resolved. Observers have also raised questions about
how the IMF’s advice is being interpreted, particularly with respect to whether to include non-
urbanised land under property tax (in the 2007 Master Plan, a remarkable 90 per cent of land
in Kigali’s boundaries is actually classed as rural). If all land in Kigali were switched to
property tax instead of land lease fees, revenue on many plots might decline because land
lease fees were higher than a fixed asset tax would be on undeveloped land at the low rate
set in the 2011 law. IMF advice was to only apply property tax to urbanised land within the
city boundaries, but in some instances this may not have been clear and provided scope for
misinterpretation.88

Nevertheless, there is no reason why these ambiguities cannot be resolved, and the
technicalities of which plots are eligible for property taxation reviewed, in order to expand the
property tax base. This is a clear priority, and one that the Rwanda Revenue Authority is
aware of.89 The question, however, is whether political constraints will prevent a sensible
redefinition of the property tax base actually being implemented, given the rapid accruing of
real estate wealth in Rwanda that has thus far largely remained untaxed. There are political
benefits to the current system that merit consideration. As well as providing generous scope
for capital accumulation by elites in the form of untaxed property, the system of land lease
fees means that a steady stream of income is garnered from all plots with lease titles, many
of which would actually be exempt under an expanded property tax because they are of little
market value. The Rwandan government has always been unusually effective in raising
revenue from the lower-income groups in society, as the estimated 95 per cent compliance

84
Interview with taxation adviser, 26 November 2009. This is would be even higher than is typical for OECD countries,
which average at around 2.1% of GDP. Some countries do have significantly higher proportions, however, with the UK
ranking the highest at 4.3% (OECD library, data available at: <http://www.oecd-ilibrary.org/taxation/taxes-on-
property_20758510-table7>, accessed 20 April 2015).
85
This is based on the amount of property tax collected nationally in 2013 as a proportion of the 2013 GDP according to
World Bank figures.
86
This refers to a range of levies on land and buildings, which might have been calculated to include land fees/lease fees
in some cases. It is inevitably also skewed by Africa’s largest economies, which are also more effective in property
taxation, such as South Africa and Nigeria.
87
Interview with taxation specialist, 20 January 2015.
88
Interview with taxation specialist, 20 January 2015.
89
Comment made by Richard Mushabe, Commissioner General of Rwanda Revenue Authority, 4 June 2014. As of April
2015, there were positive developments on this front in the form of a law being drafted that aims to overcome the
problematic distinction between taxes on freehold and leasehold properties, defining a new category of ‘immovable
property’ that will be subject to taxation regardless of tenure.

28
rate on trading licence tax indicates.90 Shifting over to a property tax for all urban plots would
be a radically progressive move that partially reverses this, as it would mean high-value real
estate paying far more while many ordinary city-dwellers might find themselves legitimately
paying nothing. In Rwandan political economy, where the elite coalition that has been
fundamental to maintaining security since the late 1990s is extremely fragile, this could be
seen as a very risky move.

This goes some way towards explaining the long history of what has been described as
passive resistance to meaningful property tax in Rwanda, which now dates back almost two
decades.91 One advisor claimed he had first talked to the government about developing a
functional system in 1997, and, despite enthusiasm on the part of ministers, found that by
2003 there had been no progress at all on the issue, and hardly any more by 2009.92 This is
uncharacteristic of Rwanda, which generally implements reforms fast, and the government
‘know they’ve dropped the ball on property tax’.93 While the 2011 law could in theory have
been a turning point because of its shift towards market-based values, it has proved a false
start for three reasons. The first is that instituting a market values-based system in a country
with a nascent real estate profession, currently lacking in transparent transactions that can
even form a basis for valuation, is highly problematic. The second is that the 2011 law did
nothing to resolve the dichotomy between different kinds of payments on freehold and
leasehold properties, which has severely constrained the tax base.

Third, and most glaring, was the slashing of the tax rate to 0.1 per cent. According to close
observers of developments in taxation, the draft version of the 2011 law had set the property
tax rate at 0.5 per cent, which reflects international norms, and had addressed the
freehold/leasehold issue by making leasehold properties subject to property taxation.94 Both
of these controversial clauses were revised by politicians in parliament before the law was
passed, which if anything suggests a shift from passive to more active resistance. Indeed,
the fact that ‘property owners are often lawmakers’ is self-evident.95 It therefore seems clear
that there is an enormous political hurdle to be overcome in Rwanda if the situation with
regard to property taxation is to improve. In many respects this is unsurprising, given that
real estate has been an explicitly favoured sector in Rwanda in connection with the
international business, tourism and construction-led development strategy envisaged for
Kigali and promoted by the Rwandan Development Board.96 Real estate – particularly for
rent to international personnel – has also by default been a major source of investment by
domestic elites given the severe lack of housing after the 1994 genocide, combined with a
sustained influx of international aid and non-governmental organisation (NGO) workers.97
These vested interests mean that even though finance officials state that ‘We are agreed on
the need for further reforms’, they also know only too well that ‘you never know where the
resistance will come’.98

4.2 Ethiopia

In Ethiopia there has been significantly less by way of formal institutional change impacting
on property taxes than in Rwanda, with virtually no developments since the revaluation
90
See Goodfellow (2012) for a discussion, and Goodfellow (2015) for an account of the taxes and fees paid by informal
transport workers.
91
Interview with taxation adviser, 1 December 2009.
92
Interview with taxation adviser, 26 November 2009.
93
Interview, 3 June 2014.
94
Interviews with taxation specialists, 6 and 8 June 2014.
95
Interview, 3 June 2014.
96
For example, there were very generous incentives for construction projects over $1.8 million that were believed to have
substantially boosted the construction sector, until they were removed in 2014 to harmonise with East African
Community rules.
97
See Goodfellow and Smith (2013) for a discussion.
98
Interview with finance official, 3 June 2014.

29
exercise commencing in 1996, and no major legislation since the 1970s. As in Kigali, there
have been recent studies of revenue potential in Addis Ababa undertaken by major
international organisations, including the World Bank and Ernst and Young. These have not
led to much change: the government is said to have been suspicious of the ideological
baggage associated with the World Bank,99 and rejected many of the recommendations in
the Ernst and Young report, of which one city official said ‘you can say it is on the shelf’.100
Despite this, some very recent developments provide grounds for optimism, and awareness
of weakness of the current system now pervades government at many levels: in the words of
one senior tax official, ‘we feel that there is big potential and it’s not being tapped … For
Addis Ababa city, it’s clear that property tax is one of the main issues to be addressed’.101

This interest in developing a new property tax system is coming to fruition in the form of a
major four-year study and pilot project currently under way, which is supported by the Gates
Foundation with initial grant support from the German development corporation, GIZ. While
the initiative began under the aegis of the Ministry of Urban Development, Housing and
Construction, the local experts drafted in to lead the pilot shrewdly insisted on also involving
the Ethiopian Revenue and Customs Authority and Ministry of Finance and Economic
Development. The Memorandum of Understanding that launched the pilot was therefore
signed by all three, indicating significant buy-in across government. Also significant was a
strategic decision not to conduct the pilot in Addis Ababa, where it was felt that there was
‘too much politics, the land leasing system had just come in [and there would be] too much
mayhem’. This decision was mutually agreed between Gates and the ministries
concerned.102

The property tax project being developed is therefore going to be piloted instead in three
secondary cities: Dire Dawa, Bahir Dar and Mekele. The exact nature of the valuation system
to be used is yet to be determined, forming part of the study leading up to the launch of the
pilot – at the time of research it was not yet decided whether this would be similar to the
system used in the 1990s exercise or not.103 What was clear was that a progressive
approach was being adopted towards exemptions in the proposed new system: while low-
income groups living in shacks or rented housing (including kebele housing rented from the
state) would be exempt, major institutional buildings including government and religious
buildings would not, marking a radical departure from many property tax systems.104 A multi-
faceted approach was being adopted to the study itself, with a housing market study taking
place alongside a capacity gap assessment and a public engagement study to investigate
issues likely to affect compliance.105

This study clearly represents a positive development, the results of which will not be evident
for several years. While the thought and strategic planning going into it are evident, the
question for advisors is whether the government will actually ‘pull the trigger’ by rolling the
project out after the pilot exercise, which ultimately means bringing it to Addis Ababa.106 The
obstacles to doing so are going to be substantial, for reasons not dissimilar to those in Kigali.
However, there are important differences regarding how the government views real estate
development, with likely implications for its willingness to follow through with reform. While
there is not scope to discuss this in detail here, unlike in Rwanda there is no favouring of
property development in Ethiopia – at least not at the level of official policy. On the contrary,
property developers have had an increasingly fraught relationship with the EPRDF regime,

99
Interview with international NGO representative, 30 September 2014.
100
Interview with city official, 26 September 2014.
101
Interview with taxation official, 25 September 2014.
102
Interview with project adviser, 30 September 2014.
103
Interview with project official, 26 September 2014.
104
Interview with project adviser, 30 September 2014.
105
Interview with project official, 26 September 2014.
106
Interview with project adviser, 30 September 2014.

30
as they are not allocated land cheaply or even through negotiation, benefit from no tax
incentives, suffer from constrained access to finance and have been subject to new
legislation aimed to prevent land speculation.107 This links to the philosophy of the late Meles
Zenawi, the EPRDF figurehead, for whom real estate development was a form of ‘parasitic’
rather than ‘productive’ investment (De Waal 2013). This is not to say that massive real
estate wealth has not accrued in Addis: it certainly has, which is partly why the pilot exercise
is giving the capital a wide berth. The point is just that the government is not ultimately afraid
or unwilling to take on this sector, and indeed has already demonstrated its willingness to do
so.

There are also challenges relating to the land lease system and how it is unfolding against
Ethiopia’s communist backdrop of state-owned land and housing. While elite resistance is
the most substantial obstacle in Rwanda, in the case of Ethiopia, and especially Addis
Ababa, there is also the question of how the wider population will respond to the imposition
of a real property tax when many are used to paying very small land fees, and in some cases
minimal rents to government for a house. This is likely to impact substantially on the growing
urban middle classes living in housing a tier above the exempt kebele and shack housing,
especially if they have acquired their land under the new lease system and find themselves
hit by both lease fees and property tax. Having specified when the lease system was
introduced that they would not tax property, Ethiopian policymakers are now looking to China
to explore how property tax can be introduced on top of lease fees.108

This is a particular concern because, unlike in Rwanda, courting the urban middle classes
has been an important priority for the EPRDF since the shock in the 2005 election in which
they lost all but one seat in the capital (Abbink 2006). While they managed to recapture these
seats amid controversy in 2010 (Tronvoll 2010), the impetus to build urban support is
commonly believed to be a driving factor behind major infrastructure projects such as the
condominium housing and Chinese-funded Addis Ababa light railway. The difficulty of
building or retaining support among a wide range of urban interests is complicated by the
government’s determination to hold on to state ownership of land at all costs. International
advisors frequently comment on the extreme sensitivity over land issues and the challenges
this can create for developing policies on land, including regarding taxation.109 Some sources
relate this concern to the outsider status of EPRDF cadres, who are primarily associated with
Tigray in Northern Ethiopia and are fearful of relinquishing authority over land in the core of
Ethiopia that they now control. Retaining state ownership means that the incentive of
freehold tenure as a benefit in exchange for taxation is not available to them. A major political
battle therefore awaits if the redesigned property tax system is to be brought to Addis Ababa.

4.3 Cross-cutting issues

A comparison of the property tax challenge in these two countries reveals a range of key
issues that need to be addressed in both cases if reform is to be successful. Table 2, which
is based on Kelly’s (2000a, 2000b) framework for assessing the effectiveness of property
taxation, illustrates the extent to which both countries’ systems are failing on multiple
dimensions – albeit with some different weaknesses in each case. The last three of the
factors listed on the table can be seen as primarily administrative issues, which would need
to be addressed through more robust systems of registration, valuation and administration
that will require years of painstaking capacity-building and awareness-raising activities. The
first two factors, which could in theory be addressed with the stroke of a legislative pen, are
inescapably political.

107
Interviews with real estate developers, September-October 2014.
108
Interview with land adviser, 4 October 2014.
109
Interview with international agency representative, 1 October 2014.

31
Table 2 Comparison of existing systems of property taxation in Rwanda and Ethiopia

Rwanda Ethiopia
Tax base Severely constrained by limiting to freehold Constrained by various exemptions including
properties, which constitute approximately 3% of state-owned housing
registered landholdings
Tax rate Levied on capital value (self-assessed); rate very Levied on annual value; rate relatively low (1-
low (0.1% compared with 0.5-1% even in other 4.5%, compared with e.g. up to 6% in Uganda)
developing countries; around 1% in South Africa)
Coverage ratio Low and dependent on self-declaration; Extremely low, due to lack of census of
discrepancy between those who declare freehold properties since 1990s
for tax purposes and those registered as freehold
in land registry
Valuation ratio Extremely low, with a sample revaluation by one Extremely low, as indicated by bank valuations
Kigali district finding that all 150 houses sampled that differ by an average of 300% and in some
were highly undervalued (though with great cases up to 7,685% from values on register
variation in the extent of undervaluation)
Collection ratio Low - calculated to be below 40% for one of Unknown
Kigali’s three districts110

In this regard there are serious difficulties constraining all five factors. The combination of
factors to tackle in order to improve the system most effectively is inevitably context-specific.
Moreover, notwithstanding the value and utility of this framework, conceiving of property tax
reform in terms of a checklist of technical characteristics can also obscure some of the
important political economy and governance dimensions shaping how property tax systems
are operating (and changing) on the ground. The remainder of this section briefly considers
some of the key cross-cutting challenges, suggesting where weaknesses and strengths in
current approaches lie, and how the two countries might learn from one another.

Whether to target or avoid the capital city

Both Rwanda and Ethiopia exhibit high degrees of urban primacy, with the capital city being
around ten times larger than the next largest city. In their steps towards reform, however, the
two countries are adopting diverging approaches with regard to whether to begin with the
capital. In Rwanda the emphasis has overwhelmingly been on Kigali, mirroring the effort to
ensure that the land regularisation exercise was completed in the capital first.111 This is
unsurprising given the degree to which property development is concentrated in the capital,
and also reflects the government’s broader perspective that Kigali needs to lead by example
in Rwanda’s development (see Goodfellow and Smith 2013). Yet some observers were
concerned by this policy of ‘separating off Kigali from the rest of the country’, especially given
the proliferation of luxury properties elsewhere in the country.112 There is also little doubt that
attempting to introduce an effective property tax in Kigali is likely to meet more powerful
continued resistance than elsewhere. 113 In Ethiopia, by contrast, the decision to exclude the
capital completely from the current project heightens the chances of a successful pilot. The
hope is that the new system might result in such increased municipal revenue in the pilot
cities that the case for imposing it on Addis Ababa becomes overwhelming, providing
government with the fuel it needs to face down vested interests. Whether this gamble pays
off will depend both on the success of the pilot and on political developments.

110
The additional information on Kigali districts in this table came from an interview with a district finance official, 2 June
2014.
111
Interview with land official, 9 December 2011.
112
Interview with taxation specialist, 20 January 2015.
113
Jibao and Prichard’s (2013) study of Sierra Leone provides an interesting direct comparison of more or less
simultaneous property tax reforms in a capital city and various smaller cities. They find that while revenue increased the
most in absolute terms in Freetown – which is unsurprising given its vastly larger tax base – reform efforts were by far
the least successful there in relative terms, with increases in revenue largely due to coercive enforcement of taxes on
the existing (regressive) tax base rather than increased dialogue and building of locally-embedded support.

32
Recentralisation and its dangers

A further common theme across both cases – and indeed more widely in Africa – is the
recentralisation of local revenue collection. It is very difficult to see this being reversed in
either case, given that it is likely to increase revenue and harmonise collection across
district/sub-city regions. While in both cases the involvement of the central revenue authority
at this stage is restricted to collection on behalf of city authorities, it seems inevitable that it
will increase central control over tax matters that have officially been decentralised. This may
affect willingness to pay taxes that are supposed to be linked to local services, as well as
impeding municipal government responsiveness to local needs and the building of local
capacity. A further danger is that the greater involvement of central authorities, particularly if
they become engaged in valuation, reduces the role of local knowledge regarding the reality
of local property markets (Norregaard 2013). It can also get in the way of efforts to streamline
relationships between land, taxation and planning authorities at city level (discussed below).
There a further risk that central involvement in local taxation increases collection ratios
sufficiently to boost overall local revenue (as is likely happening now in Kigali), which could
be used to undermine the case for more fundamental and progressive reform of property
taxation. On the other hand, if major changes are pushed through, the involvement of central
revenue authorities might make the implementation of controversial progressive tax reforms
a more realistic prospect.

Finding the right valuation method

Valuation remains a contested issue in both countries, though Ethiopia at least has
experience to build upon and continuity of personnel between the 1990s valuation exercise
and the pilot project currently taking place. It seems likely that the valuation approach
adopted will be based on mass appraisal of certain property characteristics that may or may
not reflect market values. The risk is that revaluation will not happen with sufficient regularity,
and valuation rolls will quickly become out of date. In Rwanda there is a deeper problem with
the existing approach, because commitment to market values has been enshrined in law
despite the fact there is no history of market-based valuation or computer-assisted mass
appraisal. Both the IMF’s advice and evidence from elsewhere suggest that in contexts
where property markets are immature and it is expensive for governments to buy in valuation
expertise, it may make sense to stick with fairly crude area-based approaches in the short
term. Many countries have used such approaches as a temporary fix (Monkam and Moore
2015), and a calibrated area-based approach proved a relatively effective way of both
increasing local tax revenue and building local government legitimacy in the case of Sierra
Leone (Jibao and Prichard 2013; Monkam and Moore 2015).

Although an area-based approach could undermine the yield of the tax over time and
dampen its potential progressivity, for the same reasons such an approach might make
reform more politically appealing in Rwanda. Moreover, in a context where ‘nobody has any
clue about market values’,114 there is enormous scope for widespread undervaluation if self-
assessment is combined with an aspiration to market valuation. If a system based on self-
assessment is to succeed in Rwanda – and this could be beneficial in that it shifts the onus
to the taxpayer and removes the need for costly and complex periodic revaluation exercises
– it will require clear and simple valuation guidelines for property owners, and acceptance
that market valuation remains a distant goal.

Linking property tax to the planning system

A further issue that has been rather neglected in much of the discussion around property tax
reform, but comes through strongly in these cases, is the importance of streamlining relations

114
Interview with construction sector specialist, 15 January 2015.

33
between tax authorities and the institutions for urban planning. Rwanda, in particular, has
channelled enormous efforts into building its systems for city planning and land use
regulation in Kigali,115 including through the development of a ‘one-stop shop’ for investors
wishing to acquire and develop land. The planning application system has been reformed
such that planning permission can be applied for and processed online, with a team of city
officials vigorously attempting to harmonise all development in line with the Kigali Master
Plan. There has, however, been much less commitment to applying equivalent technological
input into property taxation, or even to developing systems for information-sharing that would
enable tax authorities to become automatically aware of new developments eligible for
taxation. Thus while there is a detailed registry of land parcels and planning applications,
there is nothing linking the completion of new developments to a property register, let alone a
database of property values.

A key tool with potential to make this link is the occupation permit (sometimes called use
certificate), which is the last stage in the building control process, issued before new
buildings can legitimately be occupied and used. A tax declaration could be made a
precondition for acquiring such a permit, thereby bringing new properties onto the tax register
and incentivising self-declaration, but currently there is no such system or any attempt to
create one.116 Occupation permits have recently also been introduced in Addis Ababa, but
very few people are even aware of them yet and again there is no link to taxation, though
observers agreed this would be a very useful instrument.117

The land question revisited

It is very clear from this study that the interaction between property taxation and land reform
has not received adequate attention in the literature on taxation in developing countries, and
is a crucial and growing issue. Even within a region (such as Eastern Africa) in which
countries share certain important similarities, the political economy of land varies enormously
due to factors such as differing precolonial, colonial and postcolonial heritage, widely varying
degrees of ethnic fractionalisation, and different population densities. Layered onto these
systems are differing policy priorities and reform initiatives. Where land leasing systems have
been adopted, as in Rwanda and Ethiopia, these too can diverge in critical ways, for example
with regard to the use of auctioning and ability to acquire freehold status incrementally.
These context-specific land institutions interact in crucial ways with property tax systems.
Such interactions require analysis, as they can affect a wide range of issues including the
following: tax compliance; public understanding of different payments made to local
government; the progressivity of the overall local revenue system; the stability and
sustainability of local revenue; the ability for governments to maintain legitimate control over
urban land use; and the sense that payments to government for land use are part of a fiscal
contract rather than part of a government-driven process of acquisition and leasing for profit.

In some respects, Rwanda avoided taking steps that would have facilitated progressive
property tax reform in order to ease the path of its land reform. Ethiopia, meanwhile, has
forged ahead with a major land leasing programme in such a way as to actively impede the
prospects for property tax reform, at least in the capital city. Supporters of property tax
reform in Rwanda have been battling forces of political resistance steadily over decades.
With each successive problematic move towards reform, they may be coming closer to
wearing down resistance and finding an accommodation between land policy and property
tax policy, as suggested by a draft law at the time of writing that could radically expand the
tax base. Given that problems are not restricted to the tax base, however, a workable system
still seems a long way off.

115
See Goodfellow (2013) for a discussion.
116
Interview with tax adviser, 8 June 2014.
117
Interview with tax adviser, 29 September 2014.

34
In Ethiopia, by contrast, the political battle over taxing property in Addis Ababa has been
displaced to the future. This may not be such a bad thing: in the meantime, the property tax
project outside Addis is likely to develop a more detailed and workable property tax system
than exists in Rwanda. The question is whether it will ever reach the capital. Supporters of
progressive reform in Ethiopia might comfort themselves that by the time the battle comes to
Addis Ababa they will be better armed than proponents of tax reform in Kigali because a
system will have been developed and piloted. The fact that Ethiopia’s government does not
look especially favourably on the real estate sector (or, arguably, the diaspora groups that
fund much of this) might help. Yet there is also a particular importance attached to finding a
way to bring the system to Addis before too long: instituting real property taxation needs to
happen before land for leasing (and the revenue that accompanies this) starts to run out,
leading government to expropriate more land in potentially very socially regressive ways.

Time frame, sequencing and public relations

Given that many developing countries are pursuing multiple reforms to land and taxation
systems in tandem, important questions arise about short- and long-term strategies, as well
as appropriate ways to sequence reform and sell it to the public. Refining systems of
valuation to reflect market values could simply be unachievable in the short term, and there
may be good reasons for working instead to improve collection and coverage ratios, as well
as easing up the tax rate, before attempting to capture values more accurately.

Awareness-raising and public engagement are clearly critical for compliance, but care needs
to be taken to raise the right kind of awareness. Part of the problem is that the more that
consultants and international advisors expound the virtues of property taxation as a
progressive form of revenue generation, the greater the threat to wealthy property-owning
elites and the more there will be new manifestations of resistance. While most advice on
property taxation has come as part of broader efforts to raise tax as a proportion of GDP,
from a political perspective there is a need to make different kinds of arguments in order to
construct a public relations narrative that focuses on the benefits of property taxation beyond
mere revenue-raising. This will probably need to be harmonised with the strategy for, and
narrative concerning, land use in countries where many groups may consider that their rights
to land are under threat.

Rebranding property tax as a sort of benefit tax that is linked explicitly to certain services and
benefits is likely to be important for reforms to succeed (Monkam and Moore 2015). Indeed,
in this connection there is scope to introduce forms of tax in addition to recurrent property
taxes – for example, lump-sum payments or ‘betterment levies’ extracted by local
governments from either developers or homeowners as a charge for public service
improvements such as road paving, drainage and street lights. Given that certain parts of
Kigali have benefited disproportionately from this kind of investment, betterment levies in
such areas could add a progressive edge to a property tax system that cannot currently hope
to reflect market values (Cyan et al. 2013). For Ethiopia there is an additional public relations
hurdle created by the announcement that property would not be taxed in addition to land
lease fees. There needs therefore to be a concerted effort to link property taxation to broader
efforts to provide services in the interests of the majority, distancing it from a land leasing
system associated with displacement of low-income communities, and urban renewal based
on selling land rights to the highest bidder.

35
5 Conclusion
‘You can’t just wake up one day and decide to increase property tax’.118

The above quotation is from an estate agent in Kigali, concerned about possible increases in
property taxation that were being discussed in policy circles. His worry was that such
increases might have unexpected impacts on the housing market, resulting in higher rents in
an already very expensive rental market and doing little to address the enduring problem of
housing undersupply in Rwanda. The question of how property taxation might interact with
other taxes and incentives to affect housing provision is certainly an important question,
though not one that has been explored here. His statement, however, holds true in a broader
sense: no matter how much we might expound the virtues of property taxation, you cannot
simply decide to increase it and expect everything to fall into place. There will be continuous
technical, administrative and above all political obstacles that need to be engaged with, at all
levels of policy design and implementation. The prospects for success are considerably
higher if attention to the property tax system itself is combined with rigorous analysis of the
political-economic context.

This paper has explored the property tax challenge in all its complexity in the contexts of
Rwanda and Ethiopia, with particular attention to the capital cities, where it will ultimately
matter most. Despite many important similarities, both cases exhibit specific challenges and
opportunities. In Rwanda, the political battle to overcome entrenched elite interests in Kigali’s
property sector is starting to tip in favour of those who seriously want to reform the property
tax system. However, even if the tax base is substantially increased, the extremely low tax
rate will be difficult to raise politically, and the administrative challenges remain
overwhelming. In Ethiopia there is a more solid technical base to build upon, but government
ownership of land and the public relations challenge of introducing the tax on top of
controversial lease fees remain formidable obstacles in Addis Ababa. While on one level the
pilot programme under way demonstrates more progress than is evident in Rwanda, the
most serious political obstacles for Ethiopia have been effectively kicked into the long grass
by excluding the capital from the reforms.

These differences illustrate that there is no way of deriving generalisable advice on the best
process and content for improved property taxation in developing countries. Nevertheless,
there are critical factors that should be considered in every case, in order to develop a rich
understanding of the prospect for reforms to be effective. These relate not only to the relative
strengths and weaknesses in the tax base, tax rate, coverage ratio, valuation ratio and
collection ratio. Each of these should be considered in detail, but to understand fully why tax
systems succeed or fail on these counts it is necessary to look to history, political economy
and broader governance structures. Key factors for analysis in this regard include land
tenure systems and the politics of land reform, links between taxation and planning systems,
the roles (and agendas) of national revenue authorities in local taxation, the political-
economic role of the capital city (or other particular cities), and how broader national
development strategies seek to cultivate or constrain real estate development.

118
Interview with estate agent, 9 June 2014.

36
References
Abbink, J. (2006) ‘Discomfiture of democracy? The 2005 election crisis in Ethiopia and its
aftermath’. African Affairs 105(419): 173-199

Access Capital (2010) Sector Report - Real Estate, Addis Ababa: Access Capital

Bahl, R. (1999) Implementation Rules for Fiscal Decentralization, Working Paper 99-1,
Atlanta: Georgia State University

—— and Bird, R. (2008) ‘Subnational Taxes in Developing Countries: The Way Forward’,
Public Budgeting and Finance 28(4): 1-25

—— and Linn, J. (1992) Urban Public Finance in Developing Countries, Oxford: Oxford
University Press

Bird, R. and Casanegra de Jantscher, M. (eds) (1992) Improving Tax Administration in

Developing Countries, Washington DC: International Monetary Fund

—— and Slack, N. (eds) (2004) International handbook of land and property

taxation, Cheltenham: Edward Elgar Publishing

Booth, D. and Golooba-Mutebi, F. (2012) ‘Developmental patrimonialism? The case of

Rwanda’, African Affairs 111(444): 379-403

Bräutigam, D., Fjeldstad, O-H. and Moore, M. (eds) (2008) Taxation and State Building in
Developing Countries, Cambridge: Cambridge University Press

Brosio, G. (2000) Decentralization in Africa, Washington DC: International Monetary Fund

Chemouni, B. (2014) ‘Explaining the design of the Rwandan decentralization: elite

vulnerability and the territorial repartition of power’, Journal of Eastern African
Studies 8(2): 246-262

Clapham, C. (2009) ‘Post-war Ethiopia: the trajectories of crisis’, Review of African Political
Economy 36(120): 181-192

Cyan, M., Karuranga, C., Vaillancourt, F. (2013) Local government revenue potential in
Rwanda, Georgia State University, Andrew Young School of Policy Studies

De Soto, H. (2000) The mystery of capital: Why capitalism triumphs in the West and fails
everywhere else, New York: Basic Books

De Waal, A. (2013) ‘The theory and practice of Meles Zenawi’. African Affairs 112(446): 148-
155

Devas, N. (2003) ‘Can City Governments in the South Deliver for the Poor? A Municipal
Finance Perspective’, International Development Planning Review 25(1): 1-29

Di John, J. (2011) Taxation, Resource Mobilisation and State Performance, Crisis States
Research Centre Working Paper No. 84.2, London: London School of Economics and
Political Science

37
Everest-Phillips, M. (2009) When Do Elites Pay Taxes? Tax Morale and State-Building in
Developing Countries, conference paper for the WIDER Conference on The Role of
Elites in Economic Development, Helsinki

Fjeldstad, O-H. (2006) Local Revenue Mobilization in Urban Settings in Africa, Bergen Chr.
Michelsen Institute, WP 2006: 15

—— (2001) Intergovernmental Fiscal Relations in Developing Countries. A Review of

Issues, Bergen: Chr. Michelsen Institute, WP 2002

—— and Heggstad, K. (2011) Local Government Revenue Mobilisation in Anglophone

Africa, Working Paper EP1, Brighton: Institute for Development Studies

Franzsen, R. (2010) Fiscal Decentralization in Africa and South East Asia: The Role for the
Property Tax, presentation at GTZ 2nd Conference on Local Government Financing,
San Fernando, La Union, Philippines

—— and McCluskey W. (2013) ‘Value-based Approaches to Property Taxation’, in W.

McCluskey, G. Cornia and L. Walters (eds), A Primer on Property Tax: Administration
and Policy, Wiley-Blackwell

Gagliardone, I. (2014) ‘New media and the developmental state in Ethiopia’, African Affairs
113(451): 279-299

Gatsinzi, J. and Donaldson, R. (2010) ‘Investment challenges in the hotel industry in Kigali,
Rwanda: Hotel managers' perspectives’, Development Southern Africa 27(2): 225-240

Goodfellow, T. (2015) 'Taming the “rogue” sector: Studying state effectiveness in Africa
through informal transport politics', Comparative Politics 47(2): 127-147

—— (2013) ‘Planning and development regulation amid rapid urban growth: explaining
divergent trajectories in Africa’, Geoforum 48: 83-93

—— (2012) State Effectiveness and the Politics of Urban Development in East Africa,
London: London School of Economics and Political Science

—— and Smith, A. (2013) ‘From urban catastrophe to “model” city? Politics, security and
development in post-conflict Kigali’, Urban Studies 50(15): 3185-3202

IMF (2014) First Review under the Policy Support Instrument - Staff Report; and Press
Release, Country Report No. 14/185, Washington DC: International Monetary Fund

Jibao, S. and Prichard, W. (2013) Rebuilding Local Government Finance After Conflict: The
Political Economy of Property Taxation in Sierra Leone, ICTD Working Paper 12,
Brighton: International Centre for Tax and Development

Kelly, R. (2000a) ‘Property taxation in East Africa: the tale of three reforms’, in Proceedings.
Annual Conference on Taxation and Minutes of the Annual Meeting of the National Tax
Association, National Tax Association

—— (2000b) ‘Designing a Property Tax Reform Strategy for Sub-Saharan Africa: An

Analytical Framework Applied to Kenya’, Public Budgeting and Finance 20(4): 36-51

38
Lago-Peñas, I. and Lago-Peñas, S. (2010) ‘The determinants of tax morale in comparative
perspective: Evidence from European countries’, European Journal of Political
Economy 26(4): 441-453

Lefort, R. (2012) ‘Free market economy,”developmental state” and party-state hegemony in

Ethiopia: the case of the “model farmers”’, The Journal of Modern African
Studies 50(04): 681-706

Levi, M. (1988) Of Rule and Revenue, Berkeley: University of California Press

Lieberman, E. (2002) ‘Taxation Data as Indicators of State-Society Relations: Possibilities

and Pitfalls in Cross-National Research’, Studies in Comparative International
Development 36(4): 89-115

Markakis, J. (2011) Ethiopia: The last two frontiers, James Currey, Rochester

McCluskey, W. (1999) Property Tax: An International Comparative Study, Ashgate

Publishing Limited

—— and Franzsen R. (2013a) ‘Property Taxes in Metropolitan Cities’, in R. Bahl, J. Linn and
D. Wetzel (eds), Metropolitan Government Finance in Developing Countries, Lincoln
Institute of Land Policy

—— and Franzsen R. (2013b) ‘Non-market Value and Hybrid Approaches to Property

Taxation’, in W. McCluskey, G. Cornia and L. Walters (eds), A Primer on Property Tax:
Administration and Policy, Wiley-Blackwell

Monkam, N. and Moore, M. (2015) How Property Tax Would Benefit Africa, London: Africa
Research Institute

Moore, M. (2008) ‘Between Coercion and Contract: Competing Narratives around Taxation
and Governance’, in D. Bräutigam, O-H. Fjeldstad and M. Moore (eds), Introduction:
Taxation and State Building in Developing Countries, Cambridge: Cambridge
University Press

—— (2004) ‘Revenues, State Formation, and the Quality of Governance in Developing

Countries’, International Political Science Review 25(3): 297-319

Musgrave, R. (2000) Public Finance in a Democratic Society. Volume III: The Foundations of
Taxation and Expenditure, Cheltenham: Edward Elgar

Nerré, B. (2008) ‘Tax culture: A basic concept for tax politics’, Economic Analysis and
Policy 38(1): 153-167

Norregaard, J. (2013) Taxing Immovable Property: Revenue Potential and Implementation

Challenges, Washington DC: International Monetary Fund

Parnell, S. and Pieterse, E. (2014) Africa’s Urban Revolution, London: Zed Books

Sagashya, D. and English, C. (2009) Designing and Establishing a Land Administration

System for Rwanda: Technical and Economic Analysis, Kigali: National Land Centre,
Ministry of Natural Resources (MINIRENA)

39
Smoke, P. (2001) Fiscal Decentralization in Developing Countries: A Review of Current
Concepts and Practice, Democracy, Governance and Human Rights Programme
Paper No. 2, United Nations Research Institute for Social Development

Torgler, B. (2007) Tax compliance and tax morale: a theoretical and empirical analysis,
Edward Elgar Publishing

Tronvoll, K. (2010) ‘The Ethiopian 2010 federal and regional elections: re-establishing the
one-party state’, African Affairs 110(438): 121-136

Von Stein, L. [1885] 1964 ‘On Taxation’, in R. Musgrave and A. Peacock (eds),Classics in
the Theory of Public Finance, London: Macmillan and Co

Watson, V. (2014) ‘African urban fantasies: dreams or nightmares?’, Environment and

Urbanization 26(1): 215-231

Yusuf, B., Tefera, S. and Zerihun, A. (2009) Land Lease Policy in Addis Ababa, Private
Sector Development Hub/Addis Ababa Chamber of Commerce

40
International Centre for Tax and Development
at the Institute of Development Studies
Brighton BN1 9RE, UK
T: +44 (0)1273 606261
F: +44 (0)1273 621202
E: [email protected]
www.ictd.ac