Professional Documents
Culture Documents
Housing Studies
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/chos20
To cite this article: Romano Del Mistro & David A. Hensher (2009) Upgrading Informal Settlements in South Africa: Policy,
Rhetoric and what Residents really Value, Housing Studies, 24:3, 333-354, DOI: 10.1080/02673030902869279
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
Housing Studies,
Vol. 24, No. 3, 333–354, May 2009
ABSTRACT A billion persons live in slums worldwide, and 2 million households live in informal
housing in South Africa. The stated goal of the South African government is to overcome this housing
backlog by 2014; but doubling the budget will achieve this by only 2030. Current policy is to deliver
a choice of housing alternatives; but in practice ‘a house on a fully serviced property with freehold
title’ is seen as the only alternative. This paper describes a model that estimates the value that
residents of an informal settlement place on aspects such as level of municipal engineering services,
location and type of upgrade, and the size of the dwelling. The model was applied to three issues in
the current debate on informal settlement upgrading in South Africa; namely: whether to upgrade by
relocating all residents to a ‘greenfield’ site or upgrading ‘in situ’? Whether to upgrade
incrementally across many settlements or upgrade fully one settlement at a time? Whether to offer
residents more than one upgrade alternative? The stated choice approach provides a method to
develop and test many housing alternatives as part of involving the community in the upgrading of an
informal settlement.
KEY WORDS : Informal settlement upgrading, stated choice, housing choice, housing finance, urban
regeneration
Introduction
It is estimated that 1.1 million households live in informal shelter in the nine major cities
of South Africa (SACN, 2006). These cities account for 40.7 per cent of the population of
South Africa, and 68.7 per cent of the national household income. Twenty-four per cent
of households in these nine cities live below the Minimum Living Level (SACN, 2006).
The housing backlog for the country as a whole is twice that for the nine major cities.
Upgrading informal settlements or slums is not only a challenge for South Africa, but also
for many other countries. It was estimated that 1 billion people in the world are housed in
slums and that this could double in three decades (UN Habitat, 2003).
Correspondence Address: Romano Del Mistro, University of Cape Town, Civil Engineering, Private Bag X03,
Rondebosch, 7701 South Africa. Email: romano.delmistro@uct.ac.za
ISSN 0267-3037 Print/1466-1810 Online/09/030333–22 q 2009 Taylor & Francis
DOI: 10.1080/02673030902869279
334 R. Del Mistro & D. A. Hensher
service’) solution in the late 1980s and early 1990s in response to the sudden influx of
people into the cities after the repeal of the Influx Control Act. It is no wonder that the new
government would consider the only housing option to be a fully serviced house
(bigger than the apartheid house) with freehold title.
The vast majority of subsidised housing developed in South Africa since 1994 has
been project-linked or contractor-driven. This approach encourages individuals to
‘sit back and wait for government to deliver’ . . . The paternalistic development
approach is popular with local politicians, whose support-bases within the
community are defined by the extent to which they are able to broker such
delivery . . . (p. 50)
Land
Land is a major issue in the upgrading process. Essentially, the current policy aims to
allocate freehold title. This is time consuming and costly. The argument for this approach
is that it gives security of tenure and therefore the opportunity to leverage the property
values as a means of poverty alleviation (De Soto, 2000). It must equally be argued that
encouraging a person to risk his/her only asset is not always a desirable livelihood strategy.
Furthermore, it can also be argued that the ability to use one’s property as collateral
depends on the ease with which the lender can collect any outstanding debt. Housing is a
politically emotive issue and this must have a negative effect on the reality of low-income
housing being able to serve this purpose. Furthermore, experience has shown that low-
income houses are being sold without registering change in ownership. This highlights the
informality that exists in an informal settlement, even after upgrading (Huchzermeyer,
2008b); the value to the first beneficiary of the subsidized house; and the inappropriateness
of the complex process to provide freehold title when upgrading. Another aspect that has
arisen with freehold title is the ‘downward raiding’ of these properties by persons with
higher income, thereby defeating the intentions of the subsidy programme.
336 R. Del Mistro & D. A. Hensher
Regulations
Government plays a major role in regulating the upgrading of informal settlements. It sets
the rules and mechanisms which govern funding and approval of proposed upgrading
projects. While these might be appropriate in other circumstances, in a country with a
shortage of technical and administrative skills in both government and the private sector,
complex funding schedules, time-consuming environmental assessments, blurred
boundaries between the functions of the three spheres of government, and frequent
changes in personnel severely delay the upgrading process.
Downloaded by [University of Birmingham] at 21:12 18 November 2014
Community Mistrust
Many residents of informal settlements have been living under these conditions since
before 1994. The unfulfilled promises are now beginning to be replied to through protest
action.1,2,3 In a review of service delivery protests as an indicator of the potential for
revolution, Hough (2008, p. 6) writes that the “most immediate problem is linked to the
frustration of promises not kept regarding service delivery”. Although wider problems
include crime, the growing gap between rich and poor, and deteriorating conditions in
government departments, “ . . . high degrees of poverty or unemployment, perceptions of
corruption or perceptions of relative inequality are important in explaining service
delivery protests”, . . . “poor communication with communities, lack of transparency, poor
governance, ineffective management and political infighting were . . . some of the main
factors contributing to the protests” as well as “the growing tension between the
expectations and the reality, or perceptions about, service delivery” (Hough, 2008, p. 8).
Misselhorn (2008) adds that the
causes of discontent therefore are not only about lack of housing and service
provision, but also in respect of the perception by residents of informal settlements
that the state does not care . . . that they are somehow inferior . . . due to their
informal or shack status. (p. 3)
The government argues that the protests are orchestrated by a ‘third force’ (Hough, 2008,
p. 7), but Hough (2008, p. 11) concludes that “there is significant potential for political
violence in South Africa”. This view is echoed by Misselhorn (2008, p. 3) who notes that
“the current situation constitutes a political and developmental powder keg”.
Recently the government has stepped up its rhetoric on the eradication and eviction of
slums (Huchzermeyer, 2008a), drafted new legislation on ‘slum clearance’ which has been
met with protests4 and moved residents from informal settlements (e.g. from Joe Slovo
(10 km from the centre of Cape Town) to Delft (20 km from the centre)) having promised
them that they would return when it was upgraded.5
Parallel Initiatives
Another aspect of the land issue is the existence of two systems of property. In the years
preceding 2004, properties in the first economy experienced a boom and those of the
second economy experienced a slump. The government has set itself the objective to
integrate the two property economies. As a way of bridging this duality, it has initiated
Upgrading Informal Settlements in South Africa 337
Inadequate Resources
The final reason why the housing objective will not be achieved is the lack of financial and
human resources. Indeed, the Housing Policy of the Western Cape includes a graph that
shows that the housing backlog could only be resolved by 2030 if government funding was
doubled (DPLG, undated, p. 24). Dyantyi (2006), the Minister for Housing and Provincial
Government of the Western Cape, is quoted as saying that “the city’s growing housing
backlog was a ‘moving target’ that needed a multi-faceted approach by all levels of
government”. He said “the province’s housing budget of R600-million would provide only
16 000 houses, while the annual demand was for 27 000 houses”. In her review of
informal settlement policies in Ethekwini (Durban), Vermuelin (2006, p. 67) writes that
“ . . . informal growth from 1996 to 2004 was over 9.2 percent per year . . . 16 000 units per
year would put an end to informal settlements by 2022 . . . even with a massive delivery
(more than 16 000) the local authority task will not be easy . . . ”.
Furthermore, there is underlying thinking that the relocation of residents to ‘greenfield’
sites is the best option. This thinking does not pay enough attention to the fact that this type
of upgrade also requires additional land, which is another scare resource.
There is a need to review the programmes and procedures to improve the conditions of
the people living in informal settlements. In a review of successful upgrading projects,
Mitlin (2001, p. 509) found that diverse strategies will be necessary and that “development
agencies that wish to address the needs of the poor should seek to understand their
individual and collective livelihood struggle”.
The research questions therefore become as follows: first, what aspects do residents of
informal settlements value when their settlement is to be upgraded?; second, how
important are these aspects? and third, how should these values be incorporated into
informal settlement upgrade policy and practice? The remainder of this paper discusses a
study to develop a model of the values that residents of an informal settlements place on
attributes of settlement upgrade and applies the model to discuss how residents value not
having to relocate, incremental upgrading, and a range of housing alternatives.
Graham (2003) proposed that the progressive improvement models cannot solely be the
responsibility of the residents, as proposed by Cotton & Franceys (1988), Choguill et al.
(1993, 1994) and Choguill (1996, 1999), but should be the responsibility of both the residents
and the authority. The South African government’s policy is based on the Bill of Rights of
the Constitution of South Africa which states that “Everyone has the right to have access to
adequate housing” and “The state must take reasonable legislative and other measures,
within its available resources, to achieve the progressive realisation of this right” (RSA,
1996). Kamete (2001, p. 34) describes the housing policy shifts in Zimbabwe, which parallel
those in South Africa and probably many other countries as well, as follows:
Downloaded by [University of Birmingham] at 21:12 18 November 2014
Study Method
Study Area
Sweet Homes is an informal settlement of over 2000 households on an area of 24 ha, located
18 km to the south east of the Central Business District of Cape Town, South Africa. Prior to
becoming an informal settlement it was used as a dumping site for building rubble
(Caleb Consulting, 2003). The settlement has existed on the site for more than 10 years. While
the initial residents were workers from nearby farms, the majority of current residents are
mainly Xhosa speaking persons that have migrated from other informal settlements in
Cape Town and from the Eastern Cape (F. Ndatlane, community leader, personal
communication, 2005). Some areas of the site are subject to flooding during years of high
rainfall. At the time of the study, the city had provided primary level services to the site
(i.e. communal water and sanitation) and was investigating ways to reduce the risk of flooding.
The study area was selected because it was typical of many informal settlements in the
Cape Town area and because the community leadership was known to the author who had
been involved in student projects in the area prior to this study.
Like all analytical methods, the method has strengths and weaknesses. These are
summarized in a growing number of sources (see Louviere et al., 2000 and Hensher, 2008
for overviews). Of particular note is the presence of hypothetical bias, formally defined as
the extent to which individuals might behave inconsistently when they do not have to back
up their choices with real commitments. However, when the focus is on choice
experiments of the type used here (in contrast to contingent valuation methods), the
reliability of the overall utility expressions in terms of willingness to pay is increasingly
being shown to be more compliant with reducing hypothetical bias (see Hensher, 2008).
In the current context, real market data are not available and hence the need to undertake a
form of experiment.
Downloaded by [University of Birmingham] at 21:12 18 November 2014
The stated choice method focuses on understanding the discrete behavioural responses
of individuals to the actions of business, markets and government when faced with two or
more possible outcomes (or choices). Its theoretical underpinnings are derived from
micro-economic theory of consumer behaviour, such as the formal definition of individual
preferences as inputs into a choice setting as determined by the utility maximization of
agents. Given that an analyst has incomplete knowledge on the information inputs of the
agents being studied, they can only explain a choice outcome up to a probability of it
occurring. This is the basis for the theory of random utility (see Louviere et al., 2000 for a
review of literature).7 While random utility theory has developed from economic theories
of consumer behaviour it can be applied to any unit of analysis (e.g. groups, firms,
planners) where the dependent variable is discrete.8
Like any random utility model of the discrete choice family of models, it is assumed that
a sampled individual (q ¼ 1, . . . ,Q) faces a ‘choice’ amongst i ¼ 1,2, . . . ,I alternatives in
each of T occasions. An individual q is assumed to recognize the full set of alternatives on
offer on occasion t and to choose the alternative with the highest utility. The (relative)
utility associated with each alternative i as evaluated by each individual q in occasion t
is represented in a discrete choice model by a utility expression of the following
general form:
Xitq is a vector of explanatory variables that are observed by the analyst (from any source)
and include observed attributes of the alternative outcomes, observed characteristics of the
individual and descriptors of the decision context in occasion t; ßitq and 1itq are not
observed by the analyst and are treated as stochastic influences.
To provide an intuitive explanation of how Equation (1) operates in a choice setting,
think of the task as being one of representing sources of variance that contribute to
explaining a specific outcome. For a specific individual, Equation (1) has variance
potential associated with the coefficient attached to each observed characteristic (i.e. ß), to
each observed characteristic itself (i.e. X) and the unobserved effects term (1). This
equation could be expanded out to reflect these sources of variance for three
characteristics, defining the subscripts ‘0’ as observed and ‘U’ as unobserved, as
(dropping the q and t subscripts):
Ui ¼ ðßo1 Xo1 þ ßu1 Xu1 Þ þ ðßo2 Xo2 þ ßu2 Xu2 Þ þ ðßo3 Xo3 þ ßu3 Xu3 Þ þ 1i ð2Þ
of other influences, giving more depth in the explanation of sources of variance. As the
function is expanded out, deeper parameters to identify are revealed. In the most restrictive
(or simplistic) versions of the utility expression, all the unobserved sources would be
gathered together and replaced (2) with (3):
Ui ¼ ßo1 Xo1 þ ßo2 Xo2 þ ßo3 Xo3 þ ðßu1 Xu1 þ ßu2 Xu2 þ ßu3 Xu3 þ 1i Þ ð3Þ
and would collapse the unobserved influences into a single unknown by assuming that all
unobserved effects cannot be related in any systematic way with the observed effects:
Ui ¼ ßo1 Xo1 þ ßo2 Xo2 þ ßo3 Xo3 þ 1i ð4Þ
Downloaded by [University of Birmingham] at 21:12 18 November 2014
Furthermore, by defining a utility expression of the form in (4) for each alternative
outcome i and imposing a further assumption that the unobserved influences have the
same distribution and are independent across alternatives (IID), the subscript i attached to
1 can be removed. This gives the utility expressions of a multinomial logit (MNL) model,
assumed for illustrative purposes only to be linear additive in the observed
characteristics.9 It is well known that under an extreme value type 1 distribution of the
1, the choice model takes the MNL form (5) estimated using the standard maximum
likelihood method.
Probi ¼ expðUi Þ=1j expðUj Þ ð5Þ
The choice experiment detailed below is an unlabelled binary choice (with generic
parameters) and hence more advanced closed-form models like nested logit are not
appropriate. However, a random parameter (mixed logit) model was investigated, but no
statistical gain was found.
on-site; and if a house had been erected and water was available on the site,
an in-house flushing toilet.
. Solid waste removal offered as a communal bin/skip to which each household
was to take their refuse; or it was to be collected from the kerbside outside
each home.
. Electricity was not offered as an attribute.
. Housing: the existing shack; a 30 m2 house (referred to as a RDP house); and
a 50 m2 house (referred to as a 4-room house).
. Community services: Community services were introduced to determine
whether respondents would consider communal facilities to be more beneficial
Downloaded by [University of Birmingham] at 21:12 18 November 2014
to them than individual facilities. To this end they were offered either open
space (e.g. a play park) or a small community hall.
. Time to delivery: It was felt that reducing the time that respondents would
have to wait for an improvement in living conditions would have a value; 2, 5
or 10 years were offered.
. Upgrading approach: Three upgrading approaches were offered; namely In situ
upgrading, where the infrastructure is constructed around the existing houses,
to minimise the number of households that would need to move or relocate;
Roll-over upgrading, where residents are required to move temporarily from
their current site and return when services had been installed; and Greenfields
where residents are required to move to another area.
Because there was no obvious linear relationship between the levels of each attribute
(i.e. the utility derived from either the communal tap, the on-site tap or the in-house tap
cannot be placed on a numerical continuum, as is possible for a monetary cost) each
attribute was coded as a set of dummy variables (i.e. (1, 0) for the communal tap; (0, 1) for
the on-site tap or (0, 0) for the in-house tap).
Alternative 1 Alternative 2
Time 5 years Time 2 years
Upgrade process In situ Upgrade process Greenfields
Roads Gravel road Roads Paved road
Refuse removal Communal skip Refuse removal Communal skip
Sanitation Own VIP Sanitation Communal VIP
Water In-house water Water In-house water
House 50 m2 House 30 m2
Communal facility Public open space Communal facility Public open space
Downloaded by [University of Birmingham] at 21:12 18 November 2014
Empirical Findings
While software is used to obtain estimates of the parameters of a model, the researcher
needs to decide which attributes to include. For a model to be satisfactory, the signs and
values should be logical (i.e. an improved level of attribute should have a more positive
coefficient), the t-statistic should exceed 1.96 (to indicate that the inclusion of the attribute
has the level of statistically certainty that it is relevant in the model) and the ability of the
model to account for the variance in the data is acceptable (the measure used in MNL
models is r2) (Del Mistro & Arentze, 2002). The overall log-likelihood (LL) is 2 345.83
with an adjusted r2 of 0.2068. The LL value is statistically significant relative
to the base case of no attribute influence for 10 degrees of freedom, based on the
likelihood ratio test: {2 2(LL(0) 2 LL(u)) ¼ 2 2(2 438.025 2 (2 345.832)) ¼ 184.39
(i.e. . (x210 ¼ 18.31))}.
Table 3 shows the coefficients of the attribute levels that were found to be statistically
significant. Three attributes were found not to be statistically significant; namely solid
waste removal, community facility and the method of presentation.
The signs of all the coefficients are intuitively plausible; attributes with a positive
coefficient contribute to increased relative utility and those with a negative coefficient
contribute to a reduced relative utility when compared to the other levels of the same
attribute that are not included in the model. For example, the communal water attribute has
a negative coefficient reflecting a lower utility than an in-yard or in-house water supply.
Similarly, a built house, be it an ‘RDP’ or ‘four roomed’ house, contributes greater relative
utility than a shack, and therefore has a positive coefficient.
Where more than one attribute is included in the model the relative magnitudes of the
coefficients are plausible. For example, the coefficient for dummy variable ‘gravel road’
(2 0.527) is greater than that for tracks (2 0.803) and less than paved road (¼ 0).
The three most statistically significant attributes were ‘in situ’ and ‘roll-over’ upgrading
on site (rather than being relocated to a ‘greenfields’ development off site) and upgrading
from ‘communal’ toilets to ‘in-house’ toilet. The high statistical significance of not being
relocated to another area, and preferably without having to move from the current position
Downloaded by [University of Birmingham] at 21:12 18 November 2014
344
% 55.1 34.8 53.6 11.6 18.8 10.1 40.6 23.2 26.1 78.3
Average 7.1 3.7
Household income (R/month) Age of respondent
Coefficient t-ratio
Communal water tap 20.332 21.965
On site water tap 0
In-house water tap 0
Communal toilet 21.337 25.369
Toilet in the yard 20.853 23.140
In-house toilet 0
Road tracks 20.804 23.839
Gravel roads 20.628 23.863
Downloaded by [University of Birmingham] at 21:12 18 November 2014
Tarred road 0
Shack 0
RDP house 0.818 4.280
Four roomed house 0.891 4.788
Delivery in 2 years 0.446 3.475
Delivery in 5 years 0
Delivery in 10 years 0
Greenfields upgrading 0
In-situ upgrading 1.400 7.967
Roll-over upgrading 1.337 7.129
site, supports the recommendations in the literature and the policy of government,
although not in practice. The statistical significance placed on having an ‘in-house’ toilet is
expected considering the appalling conditions of communal toilets and the danger
experienced of using these at night.
There is a statistically significant positive effect in moving from a ‘shack’ to a
‘four-roomed house’ and from a ‘shack’ to a ‘RDP’ house. In terms of housing policy, it is
important to note that the mean parameter estimates for RDP housing and four roomed
houses are very similar, suggesting little utility in increasing the size of formal housing.
The findings suggest the statistically significant role that the house and its size plays
relative to the other attributes. The findings also suggest the statistical significance of
upgrading from ‘road tracks’ and ‘gravel’ roads to ‘tarred’ roads; of moving away from
‘communal’ toilets to ‘on-site’ toilets; delivering improvement within two years rather
than five or 10 years; and the provision of water taps ‘on site’ or ‘in-house’ instead of
‘communal’ taps.
The model suggests potential benefits can be obtained from incrementally upgrading
sanitation and roads and, as mentioned earlier, housing. This supports the literature, which
proposes an incremental approach.
As part of the interview respondents were asked to rank the importance of the attributes.
Figure 1 shows that respondents considered housing, water, roads upgrade method and
toilets to be the most important and cost and community facilities the least important. This
ranking matches that of the stated choice analysis, except that the stated choice model is
able to estimate the utility of the levels of each attribute as well.
Cost
Com. Fac
Refuse
Time
Toilets
Upgrade
Roads
Downloaded by [University of Birmingham] at 21:12 18 November 2014
Water
Housing
0 1 2 3 4 5 6 7 8 9
Average ranking
different combinations of the attributes. This can assist the housing debate to address the
following questions:
(1) Should a housing authority provide a choice of housing solutions?
(2) Which upgrading interventions are most cost effective?
(3) Should incremental or full upgrading be adopted?
Water Sanitation Roads Upgrade approach Time Housing Utility Capital cost (R) Probability
Pairs of alternatives
Pair 1 Communal Communal Tracks Greenfields 10 years 4-room 21.582 26 000 0.106
In-yard In-yard Paved In situ 10 years Shack 0.547 7500 0.894
Pair 2 Communal Communal Tracks Rollover 2 years 4-room 0.2 26 000 0.369
In-house In-yard Gravel In situ 10 years RDP 0.738 25 500 0.631
Pair 3 In-house In-house Paved Greenfields 10 years 4-room 0.891 38 500 0.491
In-yard In-yard Paved Rollover 2 years Shack 0.929 7500 0.509
Pair 4 In-house In-house Tracks Greenfields 2 years RDP 0.459 24 500 0.431
In-yard In-yard Gravel In situ 10 years RDP 0.738 23 500 0.569
A set of 3 alternatives
Set 1 Communal Communal Tracks Greenfields 10 years 4-room 21.582 26 000 0.065
In-yard In-yard Paved In situ 10 years Shack 0.547 7500 0.548
Communal Communal Tracks In situ 2 years 4-room 0.2 26 000 0.387
Upgrading Informal Settlements in South Africa
347
348 R. Del Mistro & D. A. Hensher
is proposing to upgrade all informal settlements using a ‘traditional’ approach and for
which funds are insufficient; as is the case in South Africa.
The model shows that many upgrade alternatives are possible, and that not all residents
will choose the most costly solution. This opens a way of negotiating with residents when
upgrading informal settlements. For example, many sites on which informal settlements
are located are considered undesirable because of geotechnical or stormwater problems.
If government were to apply its policy of community involvement, for those sites that can
be improved through engineering, it could offer residents the choice of remaining on site
with a smaller superstructure or relocating off site with the larger superstructure.
Downloaded by [University of Birmingham] at 21:12 18 November 2014
From Level 0 to 1 or 2 Water Sanitation Roads Upgrade approach Time Housing Solid waste removal Community facility
Level Marginal capital cost (R)
1 1500 1000 4000 0 0 17 000 500 700
2 3500 4000 5000 0 0 26 000 0 0
Level Incremental utility
1 0.332 0.487 0.177 (1) 0 0.818 n/a n/a
2 0.332 1.34 0.804 (1) 0.445 0.891 n/a n/a
Level Marginal capital cost/Marginal utility
1 4518 2053 22 599 n/a n/a 20 782 n/a n/a
2 10 542 2985 6219 n/a n/a 29 181 n/a n/a
Note: Upgrading approaches are not incremental.
Upgrading Informal Settlements in South Africa
349
350 R. Del Mistro & D. A. Hensher
1
Utility
0
Downloaded by [University of Birmingham] at 21:12 18 November 2014
–3
Cost (R/month)
at least 5 per cent more utility than full upgrading if the time to complete the upgrading
programme is longer than seven years.
However, the real difference between the cumulative utility of the two upgrading
approaches can be seen over the short and medium term. Figure 4 indicates that
incremental upgrading could have a greater effect in stabilizing the growing public protest
due to poor service delivery than the full upgrade approach: the total utility of an
incremental upgrade can be more than twice that of a full upgrade over the first five years
of a 20-year upgrade programme and over the first nine years of a 40-year programme.
Conclusions
Many countries around the world face the need to upgrade informal settlements.
In South Africa, the policy is to upgrade informal settlements by providing fully serviced
homes. While the policy is to provide a choice of housing solutions, the practice is to roll out
one solution, namely a house on a fully serviced site. Other alternatives of social housing and
higher density housing are so expensive that only households with higher incomes can access
them rather than almost all those currently in informal settlements. After the rhetoric, all that
is left is a single solution—a house on a fully serviced site in a greenfield development.
The literature and the application of the stated choice model suggest that a more nuanced
approach is not only possible but could have cost and utility advantages.
The literature and most governments support the view that the process of upgrading
informal settlement should include public participation. Often public participation is about
process and beneficiary rather than permitting a choice between alternative upgrade
options. This paper describes a method whereby the perceived values of these alternatives
could be assessed.
Downloaded by [University of Birmingham] at 21:12 18 November 2014
For an investment
Level of service of R1 000 000:
Added utility Capital cost/
Water Sanitation Roads Housing Utility/unit by upgrading from LOS 0 unit (Rands) Utility/R1000 Households Total utility
Level 0 0 0 0 0 2 2.473 0 0
Full 2 2 2 2 0.891 3.364 38 500 0.0874 26 87
Level 0 0 0 0 0 2 2.473 0 0
Level 1 1 1 0 0 2 1.657 0.816 2500 0.3264 400 326
Level 2 1 1 1 0 2 1.48 0.993 6500 0.1528 154 153
Level 3 1 1 1 1 2 0.662 1.811 23 500 0.0771 43 77
Level 4 2 2 1 1 0.191 2.664 28 500 0.0935 35 93
Full 2 2 2 2 0.891 3.364 38 500 0.0874 26 87
*Note: Excluding utility of upgrade type and time to delivery.
Upgrading Informal Settlements in South Africa
351
352 R. Del Mistro & D. A. Hensher
1.08
1.07
Incremental / full cum. utility (%)
1.06
1.05
1.04
1.03
Downloaded by [University of Birmingham] at 21:12 18 November 2014
1.02
1.01
1.00
0 10 20 30 40
Years to complete delivery
The stated choice model described in this paper provides an insight into what residents
value when their informal settlement is upgraded. The model showed that not all the
attributes have a monetary cost and that preferences are not necessarily directly
proportional to the cost. The model suggests that it is unlikely that all residents would
choose one upgrade solution if alternatives were offered to them. This finding supports the
need for residents to have a choice, as indicated by the literature and government policy.
A stated choice survey can be used to develop a more informed range of alternatives that
are able to better match the objectives and constraints of both residents and government.
4.0
40 years
3.5
Incremental / full cum. utility
20 years
3.0 10 years
2.5
2.0
1.5
1.0
0 10 20 30 40
Year in delivery programme
Figure 4. Ratio of cumulative utility of incremental to full upgrading approaches during the upgrade
programme
Upgrading Informal Settlements in South Africa 353
On the basis of perceived utility, the application of the model found that incremental
upgrading is better than the full upgrading if the time to complete the delivery of a
programme is greater than seven years. However, the fact that incremental upgrading
yields much greater utility among more beneficiaries in the short and medium terms can be
even more important in a political scenario of increasing protest action against poor
service delivery.
It needs to be stressed that it cannot be assumed that residents of other informal
settlements will have the same preferences as described in this paper. However, the
methodology to determine the preferences that residents have for attributes of upgrade
alternatives is transferable and can be used to develop alternative upgrade packages and
Downloaded by [University of Birmingham] at 21:12 18 November 2014
Acknowledgement
The data on which this paper is based were collected by Ms Chintu Lingomba.
Notes
1
Mail and Guardian (21 April 2008) Lenasia-service delivery protest turns violent. Available at http://
www.mg.co.za/article/2008-04-21 (accessed 17 October 2008).
2
The Mercury (28 July 2008) Ntuzuma residents protest against housing. Available at http://www.
themercury.co.za/?fSectionId¼&f ArticleId¼ nw20080728091105971C573977 (accessed 2008).
3
IOL (12 May 2008) Orange Farm residents barricade road. Available at http://www.int.iol.co.za/index.
php?set_id¼ 1&click_id¼13&art_id¼ nw20080512154233670C917910 (accessed 17 October 2008).
4
Outcry over new Slums Bill. Available at http://www.iol.co.za/index.php?set_id¼ 1&click_id¼ 13&-
art_id¼ vn20070513085035657C939324 (accessed 17 October 2008).
5
Except that the upgrading produced social housing which they cannot afford and therefore have not
returned.
6
At the same time, Berrisford et al. (2008) ask for larger land area to accommodate the rural-urban
transition of immigrants.
7
In the theory of discrete choices, an essential departure from traditional micro-economic theory is the
postulate that utility is derived from the properties or characteristics of things, rather than the goods per
se. Discrete choice theory incorporates the work of the standard Lancaster-Rosen model, but modifies
this approach further by assuming that individuals maximize their utility on the basis of their
perceptions of characteristics, rather than the characteristics per se.
8
Random utility theory (RUT) is a very general theory of how the analyst represents the preferences of
agents, where elements of information (known to the agents) are not observed by the analyst. While
RUT has gained particular recognition within discrete choice theory in recent years, it is not restricted to
choice theory and can be implemented in a wide range of possible decision contexts.
9
IID is potentially restrictive in that it does not allow the error or random component of different
alternative outcomes to have different variances (i.e. degrees of unobserved heterogeneity).
References
Arentze, T., Borgers, A., Timmermans, H. & Del Mistro, R. (2003) Transport stated choice responses: task
complexity, presentation format and literacy, Transportation Research Part E, 39, pp. 229–244.
Berrisford, S., de Groot, D., Kihab, M., Morrengene, M., Mhlanza, Z. & van den Brink, R. (2008) In search of
land and housing in the new South Africa, Working Paper 130 (Washington DC: The World Bank).
Caleb Consulting (2003) Informal Settlement Upgrading Study (Cape Town: Metropolitan Council, South Africa).
Choguill, C. (1996) Ten steps to sustainable infrastructure, Habitat International, 20, pp. 389–404.
Choguill, C. (1999) Community infrastructure for low-income cities: the potential for progressive improvement,
Habitat International, 23, pp. 289–301.
354 R. Del Mistro & D. A. Hensher
Choguill, C., Franceys, R. & Cotton, A. (1993) Building community infrastructure in the 1990’s: progressive
improvement, Habitat International, 17, pp. 1–12.
Choguill, C., Franceys, R. & Cotton, A. (1994) Building community infrastructure in the 1990’s: overcoming
constraints, Habitat International, 18, pp. 3 –11.
Cotton, A. & Franceys, R. (1988) Urban infrastructure: needs and the role of aid, Habitat International,
12, pp. 139–147.
Del Mistro, R. & Arenzte, T. (2002) The applicability of stated preference for mode choice studies among less
literate commuters, Journal of the South African Institution of Civil Engineering, 44(4), pp. 16–24.
DOH (Department of Housing) (2004) Breaking New Ground, As approved by MINMEC on September 2004.
Department of Local Government and Housing (DPLG) (no date) Isidima: The Road Map to Dignified
Communities (Cape Town: DPLG of the Western Cape) (circa 2007)).
De Soto, H. (2000) The Mystery of Capital: Why Capitalism Triumphs in the West and Fails Everywhere Else
Downloaded by [University of Birmingham] at 21:12 18 November 2014