Have Urban Compound Houses A Future In SubSaharan Africa?

have urban compound houses a future in sub-saharan africa? observations from ghana ==========================================================

Have urban compound houses a future in Sub-Saharan Africa?
Observations from Ghana
==========================================================
Jørgen Andreasen, Department of Human Settlements (DHS),
Royal Danish Academy of Fine Arts School of Architecture, Copenhagen;
Jørgen Eskemose, Department of Human Settlements (DHS),
Royal Danish Academy of Fine Arts School of Architecture, Copenhagen;
and
Graham Tipple, Global Urban Development Unit, University of Newcastle
upon Tyne.
Introduction
------------
In work based on a 1993 survey of new housing, Tipple et al (1999)
found that about half of the new starts in Kumasi in Ghana, the
traditional capital of the ancient Asante Kingdom, were compound
houses. During a field visit in 2003 to Kumasi by staff and students
of the Human Settlements Department of the Royal Danish Academy of
Fine Arts, School of Architecture, however, the Danish authors noticed
how this seemed to have changed. The villa form now seems to be so
universal that a considerable monetary inducement was offered to their
driver to find a new compound house. Despite the driver’s intimate
knowledge of the city, and the generosity of the promised reward, the
search proved fruitless.1 The compound form seems to have become an
anachronism in major Ghanaian cities such as Kumasi and Accra though
it is still being built in medium sized towns, e.g., Tamale in
Northern Ghana.
For many years, anecdotal evidence, especially from builders and
users, has indicated that the traditional multi-habited house is often
regarded as outdated and unsuited to modern African urban life. This
is well documented in Hansen’s (1997) work on Mtendere Township in
Lusaka and Schlyter’s (2003) research in Chitungwiza near Harare.
Occupants cope and adapt but complain about privacy and express the
wish to live in a nuclear family household. To this end, house forms
closer to the western villa are becoming the presumed and sometimes
actual house type of choice for new middle-and upper-income owners in
many Sub-Saharan African cities.
The affordability context in Sub-Saharan Africa is one in which very
large numbers of households cannot afford to own or rent a ‘decent’ or
“adequate”2 single household dwelling (Tipple, 1994). In attempts to
improve housing conditions overall, housing which is helpful to poor
households should be encouraged, and evidence shows that many of the
characteristics of compound housing are vitally important in this.
In this paper, we discuss the disadvantages and advantages of the
compound form of multi-habited housing in use in Kumasi and
surrounding villages. We demonstrate how, although the western style
nuclear family household is seen by most people as an ideal (modern),
and the multi-habitation compound house as the symbol of the past (backward)
many of the new seemingly-nuclear family dwellings that are built are
designed to accommodate many households. Within this, we will discuss
how far the flexibility of use of the compound and its much more
economic use of land are lost and how far multi-habitation of villas
can sustain the advantages inherent in compound housing. In doing so,
we hold that it is worth reviewing the wisdom of discarding either
multi-habitation or the compound house form. The paper concentrates on
single storey compound houses in Kumasi but brings in other examples
as well.
Background
----------
As the urban population of Sub-Saharan Africa continues to grow at
high rates, formal housing supply remains incapable of providing for
their needs. The governments on the continent are among the poorest in
the world, and housing is not high on their spending agendas. Data
analysed by Tipple (1994) demonstrates that housing supply targets
above 12 dwellings per thousand people would have to be met annually,
consistently over several decades, to catch up with the housing need
predicted for 2000 and 2010 based on 1987 data from UNCHS (1987b).
Such performances are beyond anything recorded in any Sub-Saharan
African country, let alone something to be sustained for years on end
by most of the countries. Recent estimates of urban population in
Sub-Saharan Africa indicate a total of 231 million in 2001.3 With an
annual urban growth rate of 4.6 per cent, a steep rise in the urban
population is inevitable and requires levels of housing investment
that are only fictional in most Sub-Saharan African countries.
In this context of inadequate supply, the majority of households find
access to the very few dwellings built by government and other formal
providers far beyond their ability to pay. In Ghana, for example,
government provision ended years ago and private/formal supply
includes a growing number of private sector developers who target
expatriate Ghanaians used to European or American prices.4 Housing for
the majority in Africa tends to be provided through the informal
sector in informally established settlements for which the owner has
at least traditional occupancy rights5 and the ability to build a
“permanent”6 dwelling. This accommodation is likely to be as
self-built7 dwellings or rented rooms; both are probably unserviced or
reliant on shared supplies of water and sanitation, the latter most
often being a simple pit latrine on the plot
In most African countries, land is informally allocated and there are
specific land delivery systems for lower income groups. The
subdivisions mostly receive some sort of authorisation, although not
from the formal planning system. Many actors are involved in the
process of allocating informal occupancy rights. These range “from
owner occupants to tenants, subsistence landlords to absentee
landlords to petty-capitalist landlords, and developers to rent agents
and protection racketeers” (UN-Habitat, 2003a: 83)
In Mozambique it was recorded that, out of approximately 86,000 new
dwellings being built in the period 1980-1997, approximately 80,000
were supplied through informal mechanisms. Occupants usually paid
bribes to local leaders and institutions (formal and informal) in
order to authorise occupation of land for house construction in the
peri-urban areas of Maputo (Jenkins, 2000a and Jenkins, 2000b).
Money is paid to the plot holder8 based on the market value of the
land. Both of these processes of land transaction are informal
(Eskemose, 2004).
In African countries such as Tanzania, Kenya and Mozambique, the cost
of land in informal subdivisions is increasing and this, together with
increased cost of building materials, is directing the majority of
urban dwellers into the rental market where one room per household is
the norm. The high share of renting observed in Kumasi (57 per cent in
1987 (Tipple and Willis, 1991) and probably even higher today) is
common in much of Sub-Saharan Africa. In the case of Kenya, the
proportion of renters in urban areas was 72 per cent in 1985 (UNCHS,
1987a), higher in Nairobi and in 1985 exceeding 90 per cent in the
town of Thika (Andreasen, 1987). Port Harcourt in Nigeria had close
to 90 per cent renters in 1984 (Oruwari, 2003). Urban Tanzania
accommodates over 70 per cent tenants (Hoek-Smit, 1990) and Kampala,
Uganda, 65 per cent (Makerere University, 2002).
We would argue that low-income households in Ghana are in a
particularly difficult position. In urban Ghana, there is no supply of
small affordable plots, through either formal or informal delivery
systems. The invasion of land by low-income households is very rare
because the local chiefs are vigilant for their land and the sanctions
hanging over those who invade land involve supernatural activity
(Tipple and Korboe, 1998). The presence of “squatters”,9 as found in
a number of Sub-Saharan countries, is virtually absent in present-day
urban Ghana (Konadu-Agyemang, 1991). However, there are some recent
signs, arising from large numbers of newcomers from the north, that
squatting does occur, although most probably remaining a marginal
phenomenon (Devas, 2004). New settlements along the railway line
through Zongo, and very close to Kumasi city centre, are examples of
this. Such settlements are, however, very vulnerable, e.g., 80 wooden
shacks near Asawasi Station, studied in November, 2003 burned down in
April, 2004, but re-erected by June, 2004. These settlers received
informal authorisation from the owner of the land, the railway
authorities.
Figure 1 Housing in the form of one-roomed wooden shacks occur with
‘informal authorisation’ as here along the railway in Kumasi near
Asawasi Station (November 2003).
The land allocation system currently operating in Kumasi
Currently, the land delivery system in Kumasi is leading to a single
form of development in the suburban areas, establishing a virtually
unbroken collar of development round the city. Land-holding chiefs are
subdividing their land, roughly in line with Outline Plans prepared by
the local town planning department (formerly Ashanti Region but now
Kumasi Metropolitan Assembly), and allocating (mostly selling the use
rights of) the large plots to relatively wealthy individuals. For the
allocation of a 99 year lease, granting rights of surface user
(usufruct), a payment known as “drink” money (biri nsa) is made to the
chief of the community to whom the land belongs. The payment is not
strictly a purchase price, but a tribute to the ancestors who watch
over the land on behalf of those yet to be born (Ollennu, 1962).
Today it is, however, a commonplace notion that the chief owns and
sells the land, and that he has the right to dispose of the money. A
share of the money is supposed to be paid to Asantehene, the King of
the Asante, who is resident in Kumasi. Also, part of the chief’s
portion should be passed on to the community for development. There
are, however, substantial variations in how much money flows to
Asantehene and the community from one area to another. In recent years
it has become common practice that the lessee pays an additional one
third of the “drink” money directly to the Asantehene. This arose
because the Asantehene found it increasingly difficult to obtain his
share of the “drink” money. Concerning the local community and their
customary right to one third of the “drink” money received by the
local chief, our studies in settlements such as Apire, Darku and
Mpasatia, near Kumasi, indicate that the community members see little
benefit and are concerned with the transparency of the use of revenue
from land ‘sales’ (Andreasen et al., 2003; Sollien, 2003). The
‘prices’ asked are perceived to be very high. Typically, a 90 x 100
feet ( approx. 27x30 metres) plot on the edge of Kumasi, commanded
“drink” money of 10 million cedis ($900) at the end of 2003. This
corresponds roughly to the annual expenditure of a household in the
lower income group, although annual pay for a low income worker is
likely to be around 4 million cedis a year.10 For comparison, Tipple
et al (1999) found that such plots ‘cost’ 600,000 to 800,000 cedis
($960-1280) in 1993, about half the annual household expenditure of
those who owned them and were building on them at the time. In
addition, the perception of high land cost is driven by its steep rise
in cedi terms as inflation and reduction in international value of the
cedi took hold in the 1980s and 1990s.
Notwithstanding these, land cost is low by international measures.
Strassmann and Blunt (1994) give an international ‘rule of thumb’
that 100 square metres of land on the periphery of a city should cost
about as much as GDP per capita (which in Ghana is $270 - $2,000 at
purchasing power parity (World Bank, 2004: 252, Table 1)). This
would give an urban land price in Ghana of about $2200 (or $16,800 at
purchasing power parity) for 810 square metres. Building costs also
show how low the cost of land is in that the cost of a plot in a high
cost area is likely to be only about five to ten per cent of the cost
of constructing the house upon it. 11 For example, on the outskirts of
Kumasi, a plot costs $1,000 - 3,000 and the estimated cost of building
the house thereon is between $20,000 and $40,000.
As land in itself is not for sale, the ‘price’ does not vary with size
and so smaller, cheaper plots, suitable for low-income purchasers, are
not available. There is no incentive for chiefs to allocate land to
poorer people apart from those entitled to a plot by virtue of
belonging to the lineage who hold the land who can gain rights of
surface use in perpetuity at almost no cost. This might change if
smaller plots could be created and then ‘sold’ for higher prices per
unit area.12.
In general, the spending on housing as a proportion of total household
spending is remarkably low for builders13 of houses, considering their
large size, and very low indeed for renters. Builders of compounds and
villas in 1993 spent about 3.5 and 5.4 times their household annual
expenditure respectively on construction in Kumasi and 3.8 and 4.4
times in Berekum (Tipple et al., 1999). The rent for a room in a
compound house tends to be about 2 per cent of the renter household’s
total spending at the mean (Tipple et al., 1999) but rent advances
of several years’ rent mean that even prospective renters must have
some capital, making it difficult for new households to enter the
rental housing market.
While many plot holders in Tamale still build compounds,14 everybody
building in Kumasi now seems to develop large villas on the plots15.
It is, however, not unusual to leave plots empty, or with foundations
or walls partly installed, for several years awaiting resources to
build.16
Figure 2 A typical section of Zongo, Kumasi, as shown in the Town Maps
of 1924. The Asante urban, rectangular, compound was known in Kumasi
around 1800, and the grid iron plan was adopted in town planning by
the British colonial government.
Figure 3 Tamale, 2001. The grid iron layout with compound houses
became standard in almost all urban subdivisions made by the
government as well as by chiefs. It is still being used new areas in
Tamale.
Figure 4 Plan of typical compound house. Mpasatia, a traditional town
30 kilometres from Kumasi. 2001.
Figure 5 Map of Kumasi showing the City Council boundaries up to the
1980s and currently (source: Adarkwa, 2001 based on Tipple 1987) with
the outer dotted line showing the approximate limit of development by
2004. The built up area has increased about 10-fold over the 20 years
with an estimated 300,000 plots added.
Figure 6 Apire, near Kumasi, 2001. A ring of land subdivided by chiefs
into large plots surrounds Kumasi up to a distance of more than 20
kilometres. Although the number of plots per hectare is only slightly
lower than areas subdivided for compounds, not a single compound house
is evident.
Plot sizes in newly developing areas in Kumasi tend to be around 800
square metres, leading to a gross density of up to nine plots per
hectare. When these are developed as single household dwellings, the
resulting gross population density is very low, about 50 people per
hectare. Only if these dwellings are multi-habited, or if plots are
developed with multi-storey apartments, would densities approach those
of the compound housing areas. The development of multi-storey
apartment blocks is, however, very rare in peri-urban Kumasi. While
Tipple et al (1999) found that 10 per cent of plots being developed
in suburban Kumasi in 1993 were having two- and three-storey blocks of
apartments built on them, new ones are very uncommon.
In contrast, there are parts of Kumasi, especially in the
northerner-dominated areas of Zongo and Aboabo, where there are over
1,000 persons per hectare in single-storey buildings and where every
available courtyard- and street-space is being used for new rooms. We
would argue that, as a result of the lack of new peripheral land being
developed for housing suitable for low-income households, the
occupancy rates, population densities, and floor space indexes17 in
some central areas are now rising to alarming levels. Two studies of
housing in Kumasi in the 1980s found very high but reasonably stable
mean city-wide occupancy of 3.3 people per room (Tipple, 1984; Tipple
and Willis, 1991) with quite low density of development and plenty of
open spaces. A recent study in Zongo documents occupancy rates at 3.6
persons per room in some of the most crowded sections with
approximately 1,640 persons per hectare (Lange, 2002). The study
further concludes that extensions in the adjacent streets and alleys
contribute one third of the total area of the house. There are parts
of neighbouring Aboabo with similar conditions.
Characteristics of a compound house
-----------------------------------
For the purposes of this paper, and in the context of urban Ghana, we
define a compound house as one which has an open to sky courtyard
leading to rooms, either directly or through a porch or veranda. The
traditional compound form in the forest belt is flexible in having
many similarly-sized (often nearly square) rooms of nine or ten square
metres (figures 4 and 2). A multi-storey compound has gallery access
to the upper floors around the courtyard, and is a feature of some
central neighbourhoods of Kumasi (Ashanti New Town, Asafo, Dadiesoaba,
Amakom, Bantama). In 1986, an estimated 41 per cent of houses in
Kumasi were single-storey compounds and 16 per cent were multi-storey
compounds (Malpezzi et al., 1990).18 These figures have probably
declined considerably since 1986 as the construction of detached
villas has taken over all building activity in Kumasi and
surroundings.
The compound or courtyard house is well known from ancient times;
courtyard houses have been built all over the Mediterranean and the
Middle East (Barnow, 2001) and both take account of and build
cultures which can be severely damaged if the house form is replaced
by another in the name of modernisation (Al-Naim, 2002).
Compound houses are widespread in West Africa and have been for
centuries (Oliver, 1971; Schwerdtfeger, 1982). In Ghana, the form is
known to have been used by kings and ‘captains’ in “Coomassie” two
hundred years ago (Bowdich, 1819). Other Sub-Saharan African
countries also have them; in some of them, rooms open off a corridor
or large communal room rather than a courtyard. The ‘face-me-face-you’
form from Southern Nigeria and the Swahili house common in coastal
East Africa both have a 4-6 roomed structure (or larger where there is
more than one storey), with an internal corridor, on the front of the
plot as well as a courtyard flanked by utility rooms in the rear. This
courtyard is the domain of the women for cooking, clothes washing and
home-based enterprises. Furthermore, latrines and bathroom facilities
are to be found here (Lekule, 2004).
Originally, compound houses were intended for occupation by a single
family group, either as a single large household or as several
households who cook for themselves but share many communal activities.
They now represent the traditional forms of multi-habitation in their
local cultures.
With changes in urban lifestyles and the passage of time since their
completion, most compound houses have moved beyond being occupied only
by an owner’s close family. The standard household definition is
difficult to use in Ghana (Tipple et al., 1994; Hanson, 2004). Often
a wife sleeps with the minor children in her parents’ house, or in a
room rented for her by her husband, but stays and cooks during the day
in her husband’s house. Compound houses now accommodate renters in
addition to family members, and occupants may be ethnically mixed even
though major ethnic and religious divisions are still spatially
clearly expressed in the city of Kumasi. In the following section, we
articulate some of the disadvantages and advantages of compound houses
for urban form and current lifestyles in urban Ghana and Kumasi in
particular.
Disadvantages of compound houses
--------------------------------
Privacy
Privacy can be a problem in the traditional compound. English speaking
residents (who may, however, not be representative) articulated this
time and again to the members of the Danish study team during their
sojourn in compound houses in and around Kumasi. Although in-room
visual privacy can be good, sound travels easily among the rooms.
There is also much overlooking of the courtyard and no-one can have a
visitor without all residents knowing and comments being made.
Household activity which must take place in the courtyard is open to
view. Many things are done in the open because of lack of space in the
room or the need for good air movement, particularly for cooking and
pounding starchy staples into the traditional fufu.19 The acts of
cooking and eating are potentially fraught with difficulties in the
communal arena. Anyone with poorer food than others will be open to
comment and anyone with better food will generate duties to share.
Image
In contemporary Ghanaian cities, there is a common desire to appear to
be modern and progressive, especially among the young who are a major
cohort of the urban population following years of high fertility.
Compounds are regarded as old-fashioned (‘bush’) by the young
generation, so they are not seen to bestow high status on the
occupants.20 The members of the Danish study team were often told by
the inhabitants of the houses in which they stayed about the
annoyances and difficulties of living so close to each other. They
want to be “modern”, to live a “modern comfortable life”.
Communal life
Many newer households express a desire to avoid the communal lifestyle
involved in sharing courtyard and services in which there is commonly
a need to mediate the use of and access to courtyard space. Each
household traditionally has a cooking hearth (bokyea),21 on which they
cook, located in the “kitchen”, in the courtyard, or on the veranda
(Korboe, 1993; Hanson, 2004). The “kitchen”, an open room with
several fireplaces, is a common good and, thus, a shared facility in
which the users regularly sweep and maintain the fireplaces. In many
houses, not all residents will have access to the kitchen, at the
owner’s discretion, and must use the courtyard. It may be quite
constricted, especially near the courtyard corners and where there are
many households in the house. Instead of the fixed bokyea, some
households now use portable charcoal stoves placed wherever
convenient. It is not uncommon for a compound house to accommodate
many households that each need space for their cooking. The use of the
courtyard for cooking makes life dangerous for small children
navigating between the fireplaces and stoves.
Fuel is a strictly private item. As the use of charcoal is becoming
more common the dependency on the fixed fireplaces in the kitchen area
designed for firewood is becoming less important. Better-off
households use charcoal and are likely to be willing to spend more on
fuel than those who use wood which is still free for those prepared to
travel around the city to gather it.
Figure 7 Crowded courtyard of a compound house with rental rooms
(Aboabo, 2003)
Figure 8 A public toilet in Kumasi
The obligation to share services, such as water supply and toilet, is
a disadvantage of most multi-habited houses including compounds. Many
compounds are without these facilities within the house and must use
public water taps and toilets in the neighbourhood. . Urban Ghana is
one of the few contexts in Sub-Saharan Africa in which public toilets
are widespread (UN-Habitat, 2003b) (figure 8). A bathroom is mostly
provided as a simple enclosure, often open to the sky, where water
must be provided in a bucket. In quite a few cases toilets within the
compound are not shared but exclusively for the use of the owner and
his/her relatives; renters and others have to use neighbourhood public
toilet facilities. In cases where the toilet is a bucket latrine,
somebody from outside the compound is paid to empty the bucket.22
Usually, this highly unpleasant and poorly paid job is executed by a
particular group of northerners. This also applies to the cleaning and
emptying of the public toilet blocks which generally are in a very
poor state and from which a constant overflow of excreta is more the
rule than the exception. The associated health implications of such a
sanitation system are serious (Frantzen and Post, 2001). Water is
mostly provided in wells or taps outside the compound. The owner’s
household stores its own water to which even households of relatives
may not have access.
Building maintenance is a major problem in urban Ghana. There is not a
lack of technical skill, but a resident owner will find him or herself
responsible for maintenance despite often having only a little more
household income than the other resident households and less income
per capita (Tipple and Willis, 1991). Once the original owner dies,23
inherited ownership (as a ‘family house’) is often so diffuse that
responsibility is easy to avoid. Amole et al (1993) describe how
inherited ownership-in-common leads to a common responsibility for
maintenance among all who have ownership rights. However, in reality,
any resident of substance tends to have to take on the expense of
maintenance through the default of other residents who become
‘free-riders’. Thus, anyone who can tends to move out so that they are
not saddled with the expense.24 No doubt, the remainder hope that
someone of substance remains to allow them to continue free-riding! In
this way, family houses form a vitally important social safety net,
providing housing for the least able in society (the elderly, the
young, the disadvantaged, those with poorly paying employment).
Although some properties achieve an increasing marketable value, most
accommodation is of no value, because it cannot be sold out of the
ownership-in-common of the family, but is precious to those who need
its succour. In consequence, many family compound houses tend to
deteriorate gradually, sometimes simply collapsing over the heads of
the inhabitants who are helpless spectators (Ryslinge, 2003; Razzu,
2005). Today, it is not uncommon for rooms in a compound house to be
maintained separately by the different family members who own the
right to use them. Hence, we may see a well maintained, beautifully
painted façade and sound roof next to a collapsed room. The generally
poor state of maintenance reinforces the compound house form’s poor
reputation, ascribing it with faults which are probably not caused by
the form but rather by the traditions of land and property holding.
Ability to extend
The compound form can be difficult to extend. When all four sides are
fully developed, rooms are sometimes added to the outside, into the
streets and alleys, and even to the inside. This reduces the size and
usefulness of the courtyard space and/or blocks the streets and
alleys. In addition, it spoils what little ventilation is offered by
through draughts available in single-banked ranges of rooms.
Notwithstanding these disadvantages, in areas of Kumasi, such as
Aboabo and Anloga, extensions of this nature are common, reflecting
the desperate need for housing. Surveys made in 2004 by the Danish
team reveal that the majority of extensions are rented rooms, in most
cases erected by the renters themselves. In 2004, it was normal to pay
two years of rent in advance. In the case of renter-built rooms, the
cost of constructing the extension replaces advance rent payment. In
parts of Aboabo and Anloga, streets and alleys are already difficult
to use and outdoor activities such as cooking and petty trading are
under more stress following these extensions (Willems et al., 2004).
Figure 9 Old Zongo is one of the neighbourhoods in Kumasi which are
very densely occupied, perhaps with densities of up to 1,500 persons
per hectare. Source, Lange (2002)
Figure 10 Aerial photograph of Aboabo (2001) showing transformation of
streets and courtyards
Figure 11 Extensions in Aboabo encroach into roads and alleys. The
main street to the north was upgraded in 1999 by a World Bank grant
allowing vehicular access to the area. Recent extensions are
encroaching again. (Source: Willems, Altermark, Pedersen and Larsen,
2004).
Figure 12 Newman’s Palace. A three-storey compound in Kumasi.
(Photograph from 1993)
Vertical extensions may seem logical but are rare unless the building
is originally intended to be extended upwards (which some built in
recent decades are). There are, however, many two- and three-storey
compounds in central areas of Kumasi and there is no reason why
similar edifices should not be built in the twenty-first century. In
the following, we present some of the characteristics of compounds
which might be regarded as advantages in the twenty-first century
city.
Characteristics of compounds that are advantageous in contemporary
cities in Sub-Saharan Africa
------------------------------------------------------------------
Good value for money
We must be careful in referring to value in connection with housing in
Ghana and, indeed, in much of Sub-Saharan Africa, as many owners would
never regard their house as a marketable good (Tipple et al., 1994).
However, we argue that it is reasonable to posit that a house which is
relatively cheap to build and generates generous developmental
benefits through multipliers and linkages (Tipple, 1995), is likely
to be good value. We would also expect that housing that has been
built by builders with relatively low incomes would also fulfil the
good value criteria. We see below that construction costs are low and
most construction is carried out by informal sector contractors who
earn little. Only a few formal sector contractors earn high profits.
Construction cost
Compound houses are relatively inexpensive to build, both per room and
overall, and are affordable by relatively low-income groups. One room
built in a compound house or as a single-roomed ‘caretaker house’25
costs approximately one third the cost of a room in a villa. This is
true in 2004 (from surveys in Kumasi and Accra carried out by the
Danish team) and it was true in 1993 according to Tipple et al
(1999); the simple room cost was C680,000 ($1090) in 1993 and was
C10 million ($1230) in 2004.
Part of the difference in cost arises from the construction method.
The villa, even in its single storey incarnation, tends to be a
reinforced concrete post and slab construction, garnished with
pseudo-classical columns, rococo balusters, ‘hacienda style’ arches,
Versailles-style iron gates, coloured ‘natural’ façade tiles and
expensive bathroom fitting – the cement, steel, tiles and sanitary
fittings are only available to the Ghanaian economy through the use of
foreign exchange. A few compounds intended for upward extension also
tend to be post and slab construction. Single storey compounds, on the
other hand, can be built with cement blocks or, cheaper still, a
soil-based material.26 Their construction is well understood by the
small-scale, informal sector contractors who dominate housing supply
in Kumasi (Tipple et al., 1999) and by their owners. The simple
technology used is better suited to local materials and, therefore,
generates high income-multipliers in their construction. In addition,
backward linkages tend to be high per unit cost as local businesses
gain trade for building materials manufacture and transport
(UNCHS/ILO, 1995). Forward linkages involved in fitting out the
rooms and their maintenance will be relatively small as the former are
sparse and the latter only a low priority, but those from home-based
enterprises will be considerable as many residents in Ghanaian cities
use their rooms in this way (Afrane, 1990; Sinai, 1998; Gough et al.,
2003).
Suitable for inheritance
In Sub-Saharan Africa, housing tends to have an important role as an
inheritance good. In his study of user-initiated of transformations of
government-built housing, Tipple (2000; 2004) found that, in both
Kumasi and Harare (Zimbabwe), quite old owners were still extending
their houses even though their own need for rooms was already
generously fulfilled. The original dwellings are western-style
bungalows with some rooms opening directly off others, especially
living rooms and kitchens, with entry mostly directly into the main
living room. It is obvious that many owners regard providing suitable
rooms for each of their heirs as important both to fulfil customary
duties and to reduce the conflict which is likely if there are fewer
rooms than heirs; the compound form is ideally suited to this as it
tends to be a series of similarly-sized rooms opening off semi-private
circulation spaces. This is, indeed, the form found by Tipple (2000)
in the transformed houses (figure 13), demonstrating that many owners
opted for it (at least up to 1993) when they had a choice of other
more ‘modern’ forms.
Figure 13 In transformation activity in government-built estates in
Kumasi, existing ‘western’ bungalows are converted into compound
houses, this one with two storeys. Source: Tipple (2000)
As Amole et al (1993) point out, an inheritor in a family house may
occupy the room on occasions when the family needs to gather, may
allow some relative or friend to occupy it, or may rent it out to
strangers. In any case, controlling the key to a room to which there
is unrestricted access from semi-public space allows the inheritor to
make best and flexible use of the room even if he or she does not
choose to live there.
Social welfare and control
Compound houses are particularly suited to flexible occupation by an
owner with tenants or family members in some of the rooms. Inherited
in common as a ‘family house’ (Amole et al., 1993) they provide a
good context for supporting the weakest members of society and
offering some social control. The single entrance provides good
security and rooms facing the courtyard provide a good environment for
mutual assistance, especially for those who care for, or are, children
or older people. Andreasen (1987) studied newly built compound
houses for renting in Kenya and found that, in spite of a deliberate
owner strategy of tribal ‘divide and rule’ by renting to people of
mixed ethnic groups, there was a pronounced mutual assistance among
tenants.
The compound provides a busy environment for less mobile or infirm
occupants and is probably more supportive to older people. It is
likely that an elderly owner can derive satisfaction and status from
the role of ‘father’ or ‘mother’ of the house, sitting on a veranda in
the compound, chewing a stick, and being greeted by occupants as they
pass, or watching children play. Reciprocal duties of watchfulness and
care in and around the courtyard create both an affirming and
supportive environment for elders even when they are not owners. This
is much less likely where corridors form the circulation space.
For occupants who need help or are aggrieved, a landlord on the
premises allows easy access to the person responsible for the house.
The problem may not be solved but the line of responsibility is easy
to understand and activate.
Independent living at low cost
The room arrangement in compound houses provides a degree of
independence at minimal cost; households can find very low cost
accommodation without sharing cooking arrangements with relatives. In
a villa, a newly arrived friend or relative would probably have to
share food with the host household. This is likely to result in an
increase in household size which, as larger households tend to have
lower incomes per capita, is likely to reduce the viability of low
income households.
Sharing services with a known group
It is unlikely that low-income households will be able to afford
exclusive access to services such as water supply and sanitation. The
use of public taps and toilets is common in Ghana but, as elsewhere,
is fraught with problems. The responsibility for maintenance and
cleaning, the means of payment, the prevention of equipment theft, and
the need to queue for long periods at peak times are all problems that
dog publicly-provided services. Shared services are customary in
compounds – if they are provided at all27 - and their cost is
manageable among all households. However, the closed system within
which they operate, accessible only to people living in the house,
allows maintenance and cleaning problems to be handled through means
developed and agreed upon among the residents: usually the jobs are
done by the women. In many cases, services are not shared equally.
They may be reserved for the owner’s household so that those without
access to the services in the house must use public latrines and water
taps, or make private arrangements with friends and relatives nearby.
Even open spaces in the house are not always common ground. There are
obviously many points at which conflict can occur and services are
undoubtedly one of the issues influencing households to prefer
individual living. However, though all may not go smoothly, the
chances of well-run shared services are higher in the compound than in
a public setting. They can be invaluable in reducing the cost of
servicing and enabling less-than-individual servicing while
maintaining care of installations.
Compact/ high density, compatible with grid-iron plans.
One of the most important characteristic of the compound for the urban
planner is the compact built form which is generated in areas
developed as compounds. In urban Ghana, where serviced land is
restricted, a compact urban form is valuable and wholly congruent with
grid-iron layouts. Originally, in forest sites, this compactness was
important as it was very hard work clearing huge trees with only
machetes. Thus, even in villages, this compact form is the norm in
Ghana’s forest belt in contrast to traditional compounds in the
northern savannah grasslands which are generally less compact and many
still have round buildings. In the compound form rooms can be linked
or closed off at will; they can also open off the inside communal
space and/or the street with almost no modification.
Figure 14 A village near Tamale on the northern savannah grasslands.
Most structures are still round but rectangular buildings are being
added.
In most areas in Sub-Saharan Africa, the small plots advocated by
international consultants are disliked and it can take years of
discussion before governments unwillingly reduce plot sizes. Recent
experiences of one of the authors in northern Mozambique reinforce
what seems to be a reluctant acceptance of the need for smaller plots
in Sub-Saharan Africa. A new settlement in a provincial town there is
being developed with 800 plots mostly of 312 square metres, instead of
the usual and locally desired 450 square metres. It has, however, been
a long road to reach this and national and (not least) local
resistance has been firm.
Climatic suitability
In the hot and humid climate of Southern Ghana, monolithic building
technologies require some cooling techniques to maintain comfortable
temperatures.28 The compound’s internal courtyard and the through
ventilation available from rows of single-banked rooms provide at
least a semblance of ambient cooling. This is not available in
buildings with access corridors flanked on each side by rooms. The
current trend for extensions on the outside of the compounds spoils
this ventilation and many rooms are left without any through draughts
at all.
Current multi-occupied housing forms
------------------------------------
It is interesting to note how reluctant people in Ghana seem to be to
jettison the characteristics of compounds when they have the
opportunity to develop new housing. In his study of transformations in
Kumasi, Tipple (2000) discovered that many of the European-style
original dwellings have been converted to compound houses (figure 13).
The transformations made by occupants tend to provide a courtyard,
where possible, and suites of rooms opening off open spaces or
corridors. There tends to be a reduction in rooms from which others
open, and they are almost never added except where a ‘chamber and
hall’ arrangement is required for a senior occupant of the house.
Thus, almost all rooms have independent access from semi-public spaces
or access corridors and are suited to occupation by household members,
extended family members living independent lives, or renters.
Figure 15 The ‘modern’ house plans, such as these found in Old Tafo in
2003, attempt to acknowledging the need for the extended family.
Hence, the functions of the common courtyard are often accommodated
under the roof, in the shape of corridors and dining or sitting room.
Figure 16 The villas in their fenced gardens resemble expensive
American suburbs. The gate is the modern face of a family to the
street. (Apire 2003).
Figure 17 Architect’s perspective drawing for a private client in
suburban Kumasi.
The aerial photographs taken in association with the Danish team’s
study (Figure 15) show many new houses in an unfinished state, without
roofs, so that their floor plans are clearly discernible. In addition,
they interviewed some architects and collected plans of their current
projects. It is evident from both sources that, even though buildings
look like villas from the outside, their plans share some important
characteristics with compounds. It is common to find a central room
and/or corridor (or upper-storey gallery) giving access to a number of
similarly-sized rooms. The ability to have a single main entrance,
provide rooms suitable for inheritance and independent living, and
share services within the house, are all available. However, there is
rarely any open-to-the-sky courtyard space providing some cooling and
suitable for cooking and pounding fufu. Thus, those functions have to
be carried out in any open space on the plot. As the house tends to
occupy the middle portion of the plot, the remaining space (especially
at the sides and rear) usually forms a strip between a high wall or
fence and the house. Where there is a main house and a smaller, but
still substantial, ‘caretaker’s quarters’ at the rear, many courtyard
functions will take place in the narrow space between the two
structures. Distribution of open space in narrow strips means that,
when the household members wish to carry out outdoor tasks, or simply
sit and socialise outside, many will have to be some distance from the
door to their room.
In a recent study of the rise of multi-storey buildings in Kumasi
(Hauberg, 2004), it was found that six new six-storey and a limited
number of two- and three-storey structures are under construction or
are already completed in Aboabo (figure 18). However, investment
return is so low in Kumasi that such activity is unlikely to pave the
way for a larger scale of redevelopment as seen elsewhere, e.g., in
Kariakoo in Dar es Salaam, Tanzania (Lupala, 2002).29 These
multi-storey buildings have some of the characteristics of the
compound house. Access from the outside world is normally by a single
shared entrance, the many rooms may be let out singly and the
occupants share services. However, the high rise buildings impose an
alien form in the neighbourhood, relating poorly to the neighbouring
single storey compound houses. The characteristic grid iron module
(approx. 22 x22 metres there) is not respected in some cases when
these structures are built. If they proliferate, they are likely to
create high intensity use that will be virtually impossible to
service. The tendency towards increased car ownership will further put
pressure on the limited street space. There is already a problem with
large trucks and small vehicles parking on the relatively narrow
streets which serve as open public spaces where people gather.
Competition amongst cars, street vendors, pedestrians and playing
children is increasing and clashes between the different users are
frequent.
Figure 18 Multi-storey blocks with 4 flats per floor replace one
storey compounds. Few have been erected so far, but it may represent a
trend (Aboabo, 2003).
Need for compound housing characteristics in Sub-Saharan Africa
---------------------------------------------------------------
Given the massive current shortfalls in urban housing supply in
Sub-Saharan Africa, and in line with the current enabling approach
(UNCHS, 1990), there is an urgent and compelling need to maximise
the number of housing units delivered and to enable supply by as low
income households as possible. In this context, it seems to be
important to introduce house types bearing the same characteristics
and qualities as compound houses in terms of low cost, flexible use,
access for low-income households, high density and a simple well known
building technology with the possibility to be incrementally erected.
In 1993, there was a great difference between the incomes30 (both
household and per capita) of owners of compounds and owners of single
household villas (those with exclusive use of a bathroom) in Kumasi.
The owners of compounds were not much different in income from their
tenants, indeed they had lower per capita incomes because of their
larger households. In the 1990s, the cost of compounds, and their
suitability for earning an income from renting, made them a much more
affordable housing supply mode for low-income prospective owners than
the now almost universal villa. Their simplicity of construction
partly explains the low cost per square metre. Thus, Tipple et al
(1999) found that, in 1993, although the majority of tenants (both
rent-free and rent-paying), were unlikely to be able to afford to own
their own fully serviced dwelling in the city at all, they were much
more likely to be able to afford to own a compound. In the last few
years, construction costs have risen more quickly than incomes so that
even compounds are now less affordable for the majority of tenants
(personal communication, David Korboe, 2005). If the compound has
disappeared as a new housing option, a replacement is needed at the
lower price level if many households are ever to own a house.
In principle, multi-habitated house forms provide owning households
with opportunities to improve their income through renting or
brokerage. The former, however, is little used in Ghana as studies
have suggested that rental income is insignificant in relation to the
cost of providing the accommodation (Malpezzi et al., 1990; Tipple
and Willis, 1992; Korboe, 1993; Tipple et al., 1999) and few would-be
owners regard it as a major reason to construct. From the renters’
point of view, it is important that owner-occupiers continue to supply
rooms in their houses for rent and that this supply increases to
remove some of the upward impetus on advance payments.31 The payments
themselves are unlikely to disappear until rental levels float up and
out of the unrealistically low levels imposed by decades of rent
controls (Malpezzi et al., 1990; Willis et al., 1990) and which
still represented only one third of the amount necessary to pay for
the room’s construction over 25 years.32
The chance of owning a self-contained dwelling could be improved for
some would-be owners by the opportunity to build smaller dwellings on
smaller plots. However, this is blocked by the nature of land
‘pricing’ and holding, not least by a conservative delivery system by
chiefs, and lack of innovative models on the part of surveyors and
planners, who by default make reference to planning regulations. Also
the motives for owning a house in Ghana (Tipple and Willis, 1992),
do not favour this. Smaller plots would still ‘cost’ as much as large
ones unless there is to be a major revolution in the way people regard
land and real property. The attitude to land as a community-owned
entity, and rooted in the lineage and chieftaincy system which provide
a sense of belonging and measures of social control within the city,
was for generations fundamental to the cohesion and character of
Ghanaian society, even in urban areas. Tipple (1984) argued that, if
it did breakdown, the social consequences could be deep-seated and
harmful to Ghanaian society.
There are some signs of a breakdown of this culture, the medium-term
effect of which is unknown as yet. Current trends, as observed in 2004
in Kumasi, and even more so in Accra, suggest that large scale
transfer of and speculation in land is becoming more common and that
the role of chiefs is withering away as they have ‘sold’ all the land
they control.33 A few developers with over 1,000 plots in Accra and
250 in Kumasi are now active in the land market. In one case in
Kumasi, the developer sells on the plots with fully constructed villas
built thereon, for approximately C120 million ($15,000) each (in
October 2003) each. Considering the lending market in Ghana, where it
is virtually impossible to borrow money at affordable rates,34 it is
clear that such a housing market is exclusively for the upper echelons
of the population, expatriates in Ghana, and expatriate Ghanaians.
Although exclusively-used services for each household are the
theoretical ultimate level of provision (Cotton and Franceys, 1991),
it is argued elsewhere that this is unlikely to be feasible in
Sub-Saharan Africa without major improvements in service and dwelling
provision (Tipple, 1994). In addition, many cities do not have the
reservoirs to cope with the level of demand that follows from
exclusive use of services by every household. For example, Kumasi has
inadequate water in the Owabi and Barekese water resources to supply
such demand. Shared services are cheaper to provide than exclusive
services, and encourage manageable demand. However, they are often
extremely difficult to manage, especially in terms of repairs and
cleaning. While not all households in a house have access to the
existing services at present, their location within a multi-habited
house, with known and limited users, renders maintenance and control
much simpler and more direct than public services. If more services
could be provided so that all residents have access to them, this is
an ideal way of providing shared services without most of the problems
of the ‘public’ level but at lower cost and smaller demand than the
exclusive level. This may be the realistic ‘ultimate level’ (Cotton
and Franceys, 1991) of service provision for low-income households in
Ghana.
Multi-habited houses provide an ideal context within which to satisfy
the inheritance function of housing. Whether heirs are decided
according the decrees of 1986 (Tipple, 1986) and include a man’s
children by way of a written will, or whether they are traditionally
decided and include only uterine kin (a man’s full brothers and
sisters, mother’s sisters, and the children of his sisters and
mother’s sisters), the need to provide is deeply felt but the
inheritors may be a disparate group not used to living together. Thus
the independence and flexibility available to each household in
multi-occupied housing continues to be important.
Reflections
-----------
Is the compound form irretrievably backward,35too subfusc to appeal to
urban Ghanaians? Are lifestyles and aspiration so changed that
compounds are now unacceptable to most households? There is no doubt
that urban Ghanaians unanimously aspire to owning a house.
Furthermore, as the style of a house is important to the status of the
owner, it is possible that single storey compounds are seen to be
unacceptably backward. Thus, despite the higher structure cost, the
great majority would opt for a villa rather than a compound.
At the same time, multihabited houses provide the solution among
low-income urban households to the continuing demand for the
following:
*
a few rooms for the owner’s household;
*
rented rooms for the majority who need to rent;
*
rooms let rent-free to family members; and
*
enough rooms in each house to allow each heir to inherit one.
We argue that low income households will never be able to afford a
villa but some could afford a compound incrementally built. A house
form that provides the advantages of the compound house and rooms to
let is essential for many years to come for most low-income urban
households in Ghana.
It is evident that there is a conflict between, on the one hand, the
public interest in higher densities, economical land use and reduced
cost of infrastructure provision, and the possibilities of small
incrementally developed housing units and, on the other hand, the
family/lineage requirement for high status, conspicuous consumption in
impressive initial units and space for future development.
It is important to consider how the discontinued supply of compound
housing or other housing affordable for the majority will affect poor
urban households and at a larger scale the development of Kumasi in
the future. The current trend in which available peripheral land is
ultimately the domain of households who can afford villas appears to
be directly driving up densities and occupancy rates in the already
overcrowded low-income housing stock. It is evident that, in the first
decade of a villa area’s development, there are few inhabitants and
most of them are caretakers and low-income relatives protecting the
villa until the owner decides to move in. But this is a transitional
phenomenon only and, once the owners move in, the lower-income
households will probably constitute a gradually smaller proportion of
the residents.
The lack of low-cost peripheral accommodation is also likely to be
causing many would-be migrants to find accommodation in nearby towns
such as Ejisu, on the Accra road, and commute. As other people choose
to crowd into the central areas to reduce travel time and cost, both
the informal use of all available space for extra rooms and the
crowding of several people in each of the rooms available are causes
for concern. As we saw above, Kumasi has had very high occupancy rates
over the last thirty years, even when pressure for expansion was low.36
If the comparatively low density of the 1980s city (Boapeah and
Tipple, 1983) is superseded by higher density development, the
density resulting not only from many people per room but also many
rooms per hectare becomes an increasing problem with potentially
serious consequences, e.g., outbreaks of diseases such as cholera.37
A healthy supply of compounds, with the cheap rooms they provide,
should reduce the need for the extra rooms being added on to the
outside of current compounds. This phenomenon has become a problem
relatively recently and should be discouraged with vigour as it
destroys the logic of the external space and the internal ventilation
in one operation. As an alternative to outwards expansion, the two- or
three-storey compound, built floor by floor over time, would provide a
supply of additional rooms in popular areas.
Land subdivision designs which assume or encourage villas with
surrounding walls contribute to making compounds inappropriate.
Compounds require neighbourhood plan forms which encourage as much
open space on all sides of the structure as can be achieved for
effective lighting and ventilation to the rooms. This is normally
ensured in high density development by leaving a street or alley
between each house, the land for which is shared between the adjacent
plots. The outer walls of the rooms in the compound constitute the
‘boundary’ wall, keeping outsiders out and providing security. The
traditional single entrance, with a door closed at night, and wooden
shutters on the outer windows reinforce this secure built form. If
neighbours build villas, which then require high walls, this form is
badly disrupted. Even in the areas dominated by compounds, walls are
being built between the buildings , disrupting circulation and leaving
private spaces between the walls and the houses that are of little use
and often filled with litter. This should be actively discouraged by
the planning authorities. Compounds or housing forms with similar
advantages may, therefore, need to be planned for in their own
neighbourhoods.
The evidence that many villas are designed for multihabitation with
corridor access to all rooms seems to suggest that owners are still
willing to sacrifice spatial comfort for rent (Schlyter, 2003) and
the ability to fulfil family and other obligations. It does, however,
raise the issue of whether this form is a viable successor to the
compound or just a poor substitute?
There is little doubt that enclosed corridors or large communal rooms
are not equivalent to rectangular open-to-sky courtyards and likely to
be of less social and functional value. The humid climate of southern
Ghana demands air movement to reduce the discomfort levels in private
and semi-public spaces if they are to be used for cooking, washing and
relaxation, and provide the slight cooling effect available from a
courtyard. Low-income households cannot afford air-conditioning, and
it is certainly poor policy to switch from passive methods of
increasing comfort to ones which consume energy. New housing forms
should also pay attention to the activities carried out in a courtyard
and supply a similar space where possible. Efforts to reproduce
compound space in multi-storey developments, such as the SSNIT
four-storey apartment blocks at Asuoyebua in Kumasi built at huge
expense in the 1980s (Anonymous, 1985), resulted in double storey
voids on the outside of the buildings which need to be structurally
isolated from the rest of the building to cope with the vibration
resulting from the pounding of fufu (personal communication, Sam
Afram, KNUST, Kumasi). In 2003, a short visit to the Asuoyebua blocks
revealed that fufu pounding had been banned apart from on the ground
floor.
Compounds are effective housing for urban Ghana but are evidently too
unpopular for substantial numbers of new ones to be built. The villa
form is filling the niche to some extent by appearing in a form which
adapts some of the characteristics of compounds and packages them in a
‘modern’ form. However, the lack of open space within, and the
additional cost of the large single structure render them much less
effective as low-cost housing than the traditional form. In addition,
the villa form with high boundary walls introduces a suburban
environment alien to the traditional villages they surround. There is
a need for a house form which provides as many of the advantages of
the compound as possible at a cost affordable by the most prospective
owners. Characteristics which should be given priority are low price,
flexibility for family and non-family occupation, suitable space for
the preparation and cooking of local cuisine, internal control of
services, and ambient cooling.
Bibliography
------------
Afrane, S. K. (1990). "Job Creation in Residential Areas: A
Comparative Study of Public and Private Residential Communities in
Kumasi, Ghana". In M. Raj and P. Nientied (ed.) Housing and Income in
Third World urban development. London, Aspect Publishing.
Al-Naim, M. A. (2002). "Cultural continuity: comparing the fereej
system and modern housing development in Hofuf, Saudi Arabia". In S.
Romaya and C. Rakodi (ed.) Building sustainable urban settlements:
approaches and case studies in the developing world. London, ITDG:
152-64.
Amole, B., D. Korboe and A. G. Tipple (1993). "The family house in
West Africa: a forgotten resource for policy makers?" Third World
Planning Review 15 (4): 355-72.
Andreasen, J. (1987). "Rented rooms and rural relations", Unpublished
PhD Dissertation, Royal Academy of Fine Arts, Copenhagen.
Andreasen, J. (1989). "The poor don't squat", Environment and
Urbanization 1 (2).
Andreasen, J., J. Eskemose and A. L. Schmidt (2003). Mpasatia: a Town
in Ghana. Copenhagen, Royal Danish Academy of Fine Arts, School of
Architecture.
Anonymous (1985). "The tragedy of 'low cost' housing", Africa (161,
January): 35-6.
Barnow, F. (2001). The City of the Landlord. Copenhagen, Royal Danish
Academy of Fine Arts, School of Architecture Publishers.
Boapeah, S. N. and A. G. Tipple (1983). "Estimating the housing stock
in a Third World city: a method used in Kumasi", Third World Planning
Review 5 (2): 177-88.
Bowdich, T. E. (1819). Mission from Cape Coast to Ashantee. London, J.
Murray (reprinted by Frank Cass and Co., 1966).
Cotton, A. and R. Franceys (1991). Services for shelter:
infrastructure for urban low income housing. Liverpool, Liverpool
University Press.
Devas, N. (2004). Urban governance, voice and poverty in the
developing world. London, Earthscan.
Diko, J. and A. G. Tipple (1992). "Migrants build at home: long
distance housing development by Ghanaians in London", Cities 9 (4):
288-94.
Eskemose, J. (2004). "Participation in a Post Socialist Society -
cases from Mozambique". In G. Ribeiro, S. Lund and M. Mullins (ed.)
Debating participation. Copenhagen, Royal Danish Academy of Fine Arts,
School of Architecture Publishers.
Frantzen, A. and J. Post (2001). "Public Toilets in Kumasi: Burden or
Boon?" In K. Adarkwa and J. Post (ed.) The Fate of the Tree: planning
and managing the development of Kumasi, Ghana. Amsterdam and Accra,
Thela Thesis and Woeli Publishing Services.
Gough, K. V., A. G. Tipple and M. Napier (2003). "Making a Living in
African Cities: the role of home-based enterprises in Accra and
Pretoria", International Planning Studies 8 (4): 253-77.
Hansen, K. T. (1997). Keeping house in Lusaka. New York, Columbia
University Press.
Hanson, K. (2004). "Rethinking the Akan household: acknowledging the
importance of culturally and linguistically meaningful images", Africa
Today 51 (1, Fall): 27-46.
Hauberg, J. (2004). "Tower houses in Kumasi", Unpublished Masters
Thesis, Royal Danish Academy of Fine Arts, School of Architecture.,
Copenhagen.
Henry, L. and G. Mohan (2003). "Making homes: the Ghanaian diaspora,
institutions and development", Journal of International Development 15:
611-22.
Hill, P. (1966). "Landlords and brokers: a West African trading
system", Cahiers d'Etudes Africaines 6: 349-66.
Hoek-Smit, M. C. (1990). "Urban Housing Sector in Tanzania, Analysis
of the Urban Housing Survey", Dar es Salaam, The Ministry of Regional
Administration and Local Government.
Jenkins, P. (2000a). "City profile: Maputo", Cities 17 (3): 207-18.
Jenkins, P. (2000b). "Urban management, urban poverty and urban
governance :planning and land management in Maputo", Environment and
Urbanisation 12 (1): 137-52.
Konadu-Agyemang, K. O. (1991). "Reflections on the absence of squatter
settlements in West African cities: the case of Kumasi, Ghana", Urban
Studies 28 (1): 139-51.
Korboe, D. T. (1993). "The Low-income Housing System in Kumasi: An
Empirical Examination of Two Neighbourhoods", Unpublished PhD
Dissertation, University of Newcastle upon Tyne,
Lange, U. (2002). "Too Dense? A case study of how people organise
themselves spatially and socially under extreme density conditions in
Kumasi, Ghana", Norwegian University of Science and Technology (NTNU),
Trondheim.
Lekule, C. (2004). "Meanings of urban space to residents in Keko
Magurumbasi informal settlement Dar es Salaam, Tanzania", Royal Danish
Academy of Fine Arts, Copenhagen.
Lupala, J. (2002). "Urban types in Rapidly urbanising Cities. Analysis
of Formal and Informal Settlements in Dar Es Salaam, Tanzania."
Doctoral Thesis, Royal Institute of Technology, Sweden,
Makerere University (2002). "Land tenure and administrative issues in
Kampala City and their effects on urban development." Kampala,
Makerere University.
Malpezzi, S. J., A. G. Tipple and K. G. Willis (1990). "Costs and
Benefits of Rent Control: A Case Study in Kumasi, Ghana", Washington
DC.: The World Bank.
Oliver, P., (ed.) (1971). Shelter in Africa. London, Barrie and
Jenkins.
Ollennu, N. A. (1962). Principles of customary land law in Ghana.
London, Sweet and Maxwell.
Oruwari, Y. (2003). "Conditions of low-income housing, with special
reference to rental stock and rental housing strategies in Nigeria".
In UN-Habitat (ed.) Rental housing: proceedings of an expert group
meeting. Nairobi, UN-Habitat: 32-43.
Razzu, G. (2005). "Urban redevelopment, cultural heritage, poverty and
redistribution: the case of Old Accra and Adawso House", Habitat
International 29: 399-419.
Ryslinge, A. (2003). "Falling Apart". In J. Andreason, J. Eskemose and
A. L. Schmidt (ed.) Mpasatia, a Town in Ghana: tales of architecture
and planning. Copenhagen, Royal Danish Academy of Fine Arts, School of
Architecture: 147-67.
Schlyter, A. (2003). "Multi-habitation: urban housing and everyday
life in Chitungwiza, Zimbabwe. Research Report No. 123", Uppsala,
Nordiska Afrikainstitutet.
Schwerdtfeger, F. W. (1982). Traditional housing in African cities: a
comparative study of houses in Zaria, Ibadan and Marrakech.
Chichester, John Wiley.
Simon, D., D. McGregor and K. Nsiah-Gyabaah (2004). "The changing
urban-rural interface of african cities: definitional issues and
application to Kumasi, Ghana," Environment and Urbanisation 16 (2):
235-47.
Sinai, I. (1998). "Using the home for income-generation: the Case of
Kumasi, Ghana", Cities 15 (6): 417-27.
Sollien, S. (2003). "'How to sell land when the chief is dead and
everybody is a chief' . Negotiating land on the periphery of Kumasi:
The case of Darko village", Royal Danish Academy of Fine Arts,
Copenhagen.
Strassmann, W. P. and A. Blunt (1994). "Land prices and housing in
Manila", Urban Studies 31 (2): 267-85.
Tipple, A. G. (1984). "Towards a Culturally Acceptable Housing
Strategy, The Case of Kumasi, Ghana", PhD dissertation, University of
Newcastle upon Tyne,
Tipple, A. G. (1986). "A revolution in property rights (Ghana)", West
Africa 27th January: 179-80.
Tipple, A. G. (1994). "The need for new urban housing in Sub-Saharan
Africa: problem or opportunity", African Affairs (93): 587-608.
Tipple, A. G. (1995). "Dear Mr President, housing is good for your
economy", Norsk Tidsskrift Geografisk 49 (4): 175-86.
Tipple, A. G. (2000). Extending themselves: user initiated
transformations of government built housing in developing countries.
Liverpool, Liverpool University Press.
Tipple, A. G. (2004). "User-initiated transformations in Ghana and
Zimbabwe", Africa Today 51 (2): 79-105.
Tipple, A. G., B. Amole, D. Korboe and H. Onyeacholem (1994). "House
and dwelling, family and household", Third World Planning Review 16
(4): 429-50.
Tipple, A. G. and D. Korboe (1998). "Housing Policy in Ghana: towards
a supply-oriented approach", Habitat International 22 (3): 245-57.
Tipple, A. G., D. Korboe and G. Garrod (1997). "Income and wealth in
house ownership studies in urban Ghana", Housing Studies 12 (1):
111-26.
Tipple, A. G., D. Korboe, G. Garrod and K. Willis (1999). Housing
supply in Ghana: a study of Accra, Kumasi and Berekum. Progress in
Planning. D. Diamond and B. H. Massam, Pergamon. 51: 324.
Tipple, A. G. and K. G. Willis (1991). "Tenure choice in a West
African city", Third World Planning Review 13 (1): 27-45.
Tipple, A. G. and K. G. Willis (1992). "Why should Ghanaians build
houses in urban areas?" Cities February: 60-74.
UNCHS (1987a). "Case Study of Sites and Services Schemes, Kenya -
lessons from Dandora and Thika", Nairobi, UNCHS (Habitat).
UNCHS (1987b). "Global Report on Human Settlements 1986", Oxford:
Oxford University Press for UNCHS (Habitat).
UNCHS (1990). Shelter for all: the Global Strategy for Shelter to the
Year 2000. UNCHS (Habitat), Nairobi.
UNCHS/ILO (1995). Shelter Provision and Employment Generation. United
Nations Centre for Human Settlements (Habitat), Nairobi; International
Labour Office, Geneva.
UN-Habitat (2003a). Rental housing: an essential option for the urban
poor in developing countries. Nairobi, United Nations Human
Settlements Programme.
UN-Habitat (2003b). "Water and sanitation in the world's cities",
Nairobi, UN-Habitat.
Van Kempen, L. (2004). "Are the poor willing to pay a premium for
designer labels? A field experiment in Bolivia", Oxford Development
Studies 32 (2): 205-22.
Wilks, I. (1975). Asante in the nineteenth century. Cambridge,
Cambridge University Press.
Willems, S., A. Altermark, R. H. Pedersen and L. J. Larsen (2004).
"Transformations in Aboabo", Copenhagen, Unpublished report, Royal
Danish Academy of Fine Arts, School of Architecture.
Willis, K. G., S. J. Malpezzi and A. G. Tipple (1990). "An econometric
and cultural analysis of rent control in Kumasi, Ghana", Urban Studies
27 (2): 241-58.
World Bank (2004). "World Development Report 2004: Making Services
Work For Poor People", Washington DC, The World Bank.

Figure 19 Housing in the form of one-roomed wooden shacks occur with
‘informal authorisation’ as here along the railway in Kumasi near
Asawasi Station (November 2003).

Figure 20 A typical section of Zongo, Kumasi, as shown in the Town
Maps of 1924. The Asante urban, rectangular, compound was known in
Kumasi around 1800, and the grid iron plan was adopted in town
planning by the British colonial government.

Figure 21 Tamale, 2001. The grid iron layout with compound houses
became standard in almost all urban subdivisions made by the
government as well as by chiefs. It is still being used new areas in
Tamale.

Figure 22 Plan of typical compound house. Mpasatia, a traditional town
30 kilometres from Kumasi. 2001.

Figure 23 Map of Kumasi showing the City Council boundaries up to the
1980s and currently (source: Adarkwa, 2001 based on Tipple 1987) with
the outer dotted line showing the approximate limit of development by
2004. The built up area has increased about 10-fold over the 20 years
with an estimated 300,000 plots added.

Figure 24 Apire, near Kumasi, 2001. A ring of land subdivided by
chiefs into large plots surrounds Kumasi up to a distance of more than
20 kilometres. Although the number of plots per hectare is only
slightly lower than areas subdivided for compounds, not a single
compound house is evident.

Figure 25 Crowded courtyard of a compound house with rental rooms
(Aboabo, 2003)

Figure 26 A public toilet in Kumasi

Figure 27 Old Zongo is one of the neighbourhoods in Kumasi which are
very densely occupied, perhaps with densities of up to 1,500 persons
per hectare. Source, Lange (2002)

Figure 28 Aerial photograph of Aboabo (2001) showing transformation of
streets and courtyards

Figure 29 Extensions in Aboabo encroach into roads and alleys. The
main street to the north was upgraded in 1999 by a World Bank grant
allowing vehicular access to the area. Recent extensions are
encroaching again. (Source: Willems, Altermark, Pedersen and Larsen,
2004).

Figure 30 Newman’s Palace. A three-storey compound in Kumasi.
(Photograph from 1993)

Figure 31 In transformation activity in government-built estates in
Kumasi, existing ‘western’ bungalows are converted into compound
houses, this one with two storeys. Source: Tipple (2000)

Figure 32 A village near Tamale on the northern savannah grasslands.
Most structures are still round but rectangular buildings are being
added.

Figure 33 The ‘modern’ house plans, such as these found in Old Tafo in
2003, attempt to acknowledging the need for the extended family.
Hence, the functions of the common courtyard are often accommodated
under the roof, in the shape of corridors and dining or sitting room.

Figure 34 The villas in their fenced gardens resemble expensive
American suburbs. The gate is the modern face of a family to the
street. (Apire 2003).

Figure 35 Architect’s perspective drawing for a private client in
suburban Kumasi.

Figure 36 Multi-storey blocks with 4 flats per floor replace one
storey compounds. Few have been erected so far, but it may represent a
trend (Aboabo, 2003).
1  The payment was made anyway but the driver did not know that this
would be the case.
2“Adequate shelter for all ” was the compromise terminology reached in
the final Istanbul Declaration in 1996 in the UNCHS (Habitat)
conference.
3 http//:www.unhabitat.org/programmes/guo/documents/table4.pdf.
4 They are often overseas specifically in order to build a house in
Ghana (Diko and Tipple, 1992) and are also conscious of their
obligations to the wider family back home (Henry and Mohan, 2003).
Recent research by Department of Human Settlement (DHS) in Apire, a
peri urban Kumasi village now being engulfed by new villa development,
indicates that 42 per cent of the landlords are absent.
5 Usually with some form of written document.
6 The buildings often appear as simple shacks but the owners consider
them to be permanent structures in terms of providing a base despite
their temporary materials and few services.
7 This also includes building by artisans engaged by the occupants
(UNCHS/ILO, 1995).
8 This terminology is used here because of the complex nature of land
market systems in all Sub-Saharan cities. Land ownership in the
Western sense of the word is rare in most informal settlements in
African cities.
9 . Squatters are a well known global phenomenon, the term being
applied to a variety of settlements characterised by poor standards
and/or illegal status. Squatting, in terms of people occupying land to
which they have no rights whatsoever, is quite a rare phenomenon even
in the developing world. There is most often some kind of allocation
or authorisation procedure for land occupation. What is often
considered to be squatting will, after a closer look, usually prove to
be authorised by somebody, often not through the formal system, i.e.,
the government, but through influential groups or individuals, e.g.,
landlords and/or their representatives. Land invasions do occur but
they are normally found only when there is a crisis or some political
instability. Grabbing of government land by people with influence is
another phenomenon which occurs in Kumasi, e.g., in the T.U.C. housing
area to the south of the city (personal communication with David
Korboe, 2005). For further discussion of squatting see Andreasen
(1989) and Konadu-Agyemang (1991).
10 Household expenditure tends to be higher than stated household
income by a factor of more than twice. Tipple’s work, over a span of
many years, found stated household expenditure to be close to a
constant 2.4 times higher than stated household income at the means
(Tipple, 1984; 1991; 1997).
11 In the developed world suburban land cost would typically
constitute 25 per cent or more of the total cost of a new dwelling.
12 See also Devas (2004) who confirms the difficulties for the urban
poor to access land in Kumasi and other cities.
13 In this discussion, the builder of a house is the future owner.
He/she engages contractors to do the construction work. For most
people, this is the only way to own a house.
14 These assertions are based on aerial photographs taken by the
Copenhagen team in late 2003.
15 Simon et al (2004)also found this phenomenon at Esereso (a
village near Kumasi) where there is rapid development of luxury
villa-style housing.
16 Even long periods of hiatus are unlikely to be for speculation but
may be against future hopes of building. Chiefs can reclaim
undeveloped plots after two years has passed. However, although few
have done so, maybe because the money paid in tribute (drink money)
would have to be returned, local people tell of increasing numbers of
plot ‘sales’ revoked even where foundations have been built.
17 Floor space index (FSI) is the quantity of floor space expressed as
an index of the plot area. Thus, a four storey building covering half
the plot would have an FSI of 2.
18 However, because of their size, they constitute a much larger
proportion of the built stock and, because of their multi-occupancy, a
much larger proportion of households occupy them.
19 Fufu is prepared from two of four starchy staples (cassava, yam,
maize and plantain). The chopped staples are first boiled and then
pounded together in a large mortar using a long and heavy pestle. One
person constantly turns the mixture while the other pounds it until it
is glutinous. The rhythmic ‘thump, thump’ is a universal precursor of
mealtimes in Ghana.
20 An interesting insight into image is provided by Van Kempen
(2004) in a study of preference for designer-label clothing among
poor consumers in Bolivia. He found that more than 40 per cent of
those interviewed were willing to pay a premium for such clothing
because of the social identity it brings, even if it is counterfeit
and, therefore, does not possess the quality the label suggests.
21 The bokyea usually consists of three protrusions formed by shaping
the ground or by sinking car-wheel hubs into the ground to make a
robust tripod with spaces to insert fuel.
22 This has been a function of the local authority but, in many areas
of the city, householders make their own private arrangements.
23 This may not be long after construction as most do not achieve
ownership before they are in their mid to late forties (Tipple et
al., 1999).
24 In response to the cost of maintenance conundrum, many family
houses contain a few renter households. In theory, the rental income
should be used for maintenance. However, rents were controlled for
over forty years after 1943 and are still very low in comparison to
the cost of construction (see Tipple et al., 1999). Even if all the
income is faithfully devoted to maintenance, there is unlikely to be
sufficient to keep the whole house in good repair. As a way of
improving the net present value of this income, rents are collected in
advance payments from both new and established tenants.
25 The Asantehene’s Land Office secures a reasonable level of
registration of ownership of land, so that the buyer can be confident
of the authenticity of the seller. Still, and particularly in Accra,
brothers may sell the same land independently. Plots already sold may
be invaded if empty for some length of time. Therefore, the owner may
either build a wall on the plot perimeter, a foundation of the planned
house, or a one-roomed caretaker house which is given free to the
caretaker who will then prevent encroachment of the plot.
26 Not only cheaper but also more comfortable climatically, using less
non-renewable energy to produce, and just as durable if efficiently
built and protected from driving rain.
27 In Kumasi, services in compounds tend to be shared by the resident
households while in settlements at some distance, e.g., Mpasatia 30
kms from Kumasi, it is rare to find shared water supply or toilets.
28 This is particularly important in the forest belt where there is
little wind.
29 The multi-storey blocks built in Kumasi over the last decade appear
not to be developed by investors in the western sense as many of the
buildings have been standing unfinished with no occupants (renters,
owners or dependants) for years. One block that we studied was
occupied solely by immigrants from Mali; a case of ethnic groups
looking after themselves probably inspired by a mixture of
philanthropy and brokerage (Hill, 1966). In the latter case, the
chance to be involved in the businesses of the migrants may be
monetary reward enough.
30 Using expenditure as a proxy.
31 Though anecdotal evidence suggests rent advances of several years’
rent are commonplace, the data on amounts paid in Kumasi in 1993
showed a mean of only 16 months’ rent (C24,000 or $38) when mean rent
was C1,500 ($ 2.50). The maximum advance paid in the Kumasi sample at
that time was C800,000 ($1280) which is more likely to affect
perceptions of advances than is the mean. In 2004, a typical advance
was 24 months rent.
32 At 5 per cent interest and with no maintenance or other on-going
cost. Of course, 25 years amortisation is much too long to make
renting profitable so rents should probably be 6 to 10 times higher to
attract increasing numbers of providers for those who cannot afford to
own their own accommodation.
33 In the nineteenth century, when chiefs in central Asante allocated
all their land holdings, they were given more in the province of Brong
Ahafo so that they could continue to receive revenues (Wilks, 1975).
This is, of course, no longer possible.
34 We have recorded examples of formal loans set up in dollars but
paid for in cedis. Here, the relative deflation of the cedi generates
an ‘interest rate’ equivalent to 30 per cent per year.
35 Universally expressed as ‘bush’ in Ghana.
36 It must be remembered that there can be high occupancy with
comparatively low density when there is plenty of space between
houses. This was the case in 1980.
37 It is cautionary to recall that there were several outbreaks of
plague in Kumasi in the twentieth century. While one of the authors
worked in Kumasi, an outbreak of conjunctivitis spread through the
city with awesome rapidity. It was nicknamed ‘Apollo’ in honour of the
space mission occurring at the time.
52