SA's urban land reform problem

Charles Simkins says we are a more urbanised country than is commonly realised


The major requirement [for dealing with the Cape Town’s housing need] is the availability of large tracts of well-located land.

As it happens, Cape Town is particularly well-endowed with such land. There are five large tracts, all owned by national government (the Department of Defence and Transnet). The sites, [known as Culemborg, Ysterplaat, Youngsfield, Wingfield and Denel] could yield close to half the number of affordable units required to meet the need on the City’s current database.

Ever since I was mayor of Cape Town between 2006 and 2009, I have been trying to secure the release of these sites for this purpose.

- Helen Zille, Daily Maverick, 6 August 2018


The frustration experienced by Helen Zille is at the core of the urban land reform problem. The problem is resolvable and needs to be resolved quickly if disorderly development is to be avoided in Cape Town and in the other large metros. The purpose here is to explain why.

The argument can be summarized as follows:

1. South Africa is more urbanized than is generally realized. Urban land reform has the capacity to reach many more people than rural land reform, although the latter should not be neglected. Urban land reform is capable of improving the lives of many more people than rural land reform.

2. Urban areas can be divided into three categories: the urban parts of the metros[5], urban areas outside traditional areas, and urban areas inside traditional areas. The proportion of the total urban population in the metros has been rising and is expected to continue to do so. The population in urban areas outside traditional areas has been rising more slowly and, in about a third of municipalities it has been dropping in absolute terms in recent years. Rural areas are emptying, both within and outside traditional areas.

3. Average household size has been dropping and it is likely to drop further, to about three people, by 2030.

4. Real income per capita has been dropping since 2014. Absent populist adventures, it can recover and then improve in the coming years, but progress is unlikely to be rapid. Because average household size is falling, real income per household has dropped faster than real income per capita and it will take longer to recover.

5. An assessment of housing need (rather than reference to ‘backlogs’) has the advantage of making clearer what needs to be done.

6. It is fiscally impossible to build Breaking New Ground houses[6] at a rate to meet existing and new need. Currently, production is running at about a third of new need. The government’s human settlements programme needs to be refocused on (a) the rapid release of land, particularly in the five large metros and (b) a more vigorous and incentivized promotion of urban densification.

7. There is nothing to stop the various levels of government from developing a rapid response except government inertia. Municipalities have, or ought to have, integrated development plans and land use management plans to guide land acquisition and the employment of developers to service it.

The justification for each step in the argument is briefly set out below. More detail can be found in the Human Settlements and Urban Land Reform report.

South Africa is more urbanized than is generally realized

Perceptions of urbanization are driven by Statistics South Africa’s definition of ‘urban’. But this definition applies only in areas outside traditional areas (roughly, the old homelands). Traditional areas are reported as a single category. It is clear from publicly available cadastral and land use maps that there are both formal urban areas laid out in demarcated erven and areas of high density informal occupation in traditional areas.

The population in them has to be estimated by indirect methods which, when applied to the 2011 Census and the 2016 Census, indicate that close to 65% of the population in traditional areas is urban. Add this to the 94% of the population outside the traditional areas in urban areas and the national urbanization rate in the country as a whole becomes 84% in 2016, rather than the official 64%. The estimate should be rendered more precise in future censuses and surveys by the use of a standard and appropriate official definition of ‘urban’ across the country.

The urban population is concentrating in the metros, particularly in the large ones

When the results of the 2011 Census and the 2016 Community Survey are placed in a coherent demographic framework, the share of the metro population in the total urban population (metros, urban areas outside traditional areas and urban areas inside traditional areas) rose from 42.5% in 2011 to 46.0%, implying an annual growth rate of 3.3%, and more than 4% in Johannesburg and Tshwane. Growth in other urban areas outside traditional areas was about 1% and in urban areas inside traditional areas was close to zero.

Why was this? There are three main reasons:

1. The employment rate[7] in the metros is higher than in other urban areas.

2. Parts of the platteland are depopulating, with urban populations in about a third of municipalities dropping.

3. There is a slow erosion of displaced urbanization. In the apartheid years, a lot of urban residential development associated with cities and towns outside traditional areas was displaced to traditional areas. The Mbombela municipality, with Nelspruit as its principal town, is a case in point. Just 13% of its population lived in urban areas outside the old Kangwane in 2016. Add the traditional urban areas in, and the urbanization rate rises to 76%. The erosion is slow because households invested where they could and are now reluctant to abandon their dwellings.

Average household size is dropping

Average household size dropped from 3.30 to 3.21 between 2011 and 2016. This was partly a consequence of a change in the age structure of the population. One method of projecting average household size is to hold headship rates by age and sex constant and to apply them to the future structure of the population. On that basis, average household size is projected to drop to 2.99 by 2030.

It follows that new household formation is growing more rapidly than population and that average household income will rise more slowly than per capita income.

Household incomes are under considerable pressure and will be so for nearly a decade longer

April 2018 International Monetary Fund estimates and projections indicate that real per capita gross domestic product has been falling since 2014. It expects the trend to be reversed from2019, but improvement will not be rapid and the level of real per capita income will not exceed the 2014 level even by 2023. Thereafter, if progress can be made on the structural reform needed to promote growth, the growth rate could rise. Assume it rises to 2.8% by 2030, which would exceed population growth in that year by a respectable 2.0%. This would imply a 16% increase in per capita income between 2011 and 2030, but only a 6% increase in income per household, with income per household exceeding the 2011 level for the first time in 2027. Settlement options have to be considered against this backdrop. Slow economic growth and falling household sizes mean that the majority of new households in urban areas are likely to be in the poorest income category (less than R 3 500 per month in 2016 prices) in the next decade.

The dimensions of housing need

Basing analysis on the length of municipal housing waiting lists is a hopeless endeavour. Not every municipality is able to report the statistics, and some households on the lists may have dissolved or ceased to qualify for publicly provided housing. It is much better to start with the fact that every household lives somewhere, and to assess what households have against what they need. Housing may be adequate in the sense that they meet a set of standards. Or it may be inadequate for one of several reasons. It may be overcrowded, it may suffer from major structural defects, or it may lack one or more essential services. If a dwelling which has (a) no more than two people per room (with kitchens, bathrooms and garages excluded), and (b) no major weakness in floor, walls or roof, and (c) a municipal water supply plus a flush toilet or ventilated pit latrine plus refuse removed at least once a week plus access to electricity is regarded as adequate, then 71% of dwellings in urban areas were adequate in 2016. 12% were overcrowded, 8% needed repair and 9% needed to have their services upgraded. Each category of inadequacy requires a different solution.

Government expenditure on human settlements

National government expenditure on human settlements is expected to be R 32.5 billion in 2018/19. Most of this money is transferred to provinces and municipalities. Provinces add about 16% to grants from national government from their equitable share of national revenue. Municipalities, especially metros, may add a little more. Most of the expenditure is on the human settlements development grant, which funds housing development, and the urban settlements development grant, which supports the broader development of the built environment, with a strong focus on upgrading informal settlements.

This is expected to finance 99 000 subsidized houses, 20 000 affordable rental housing units and 131 000 services upgrades in informal settlements. By contrast, the annual increase in the number of new households in urban areas between 2018 and 2024 is expected to be 316 000, 72% of which will be in the metros. To the need to provide for new households should be added an allowance to cover currently unmet housing need. At the present rate of production, queues for BNG housing, already long, can be expected to get longer.

Building costs depend on terrain and wage levels, so they vary across the country. Information about them is also incomplete. The national average cost of a serviced site is of the order of R 45 000 and of a BNG house R 180 000, so that four serviced sites can be produced at the same cost as a BNG house.

It should also be noted that, while densification of existing settlements in urban areas is regarded as desirable by government, there are no fiscal measures to incentivize it. One might, for instance, remit registration fees for subdivision of residential land in urban areas. Or housing developers may be required to construct a certain proportion, say 20%, of their dwellings for rental by low or lower middle income households. Or, redevelopment of run-down urban areas could be supported.

Urban land reform

Urban land reform is about making adequate provision for residential development needed now and in the coming years. Ultimately, as both the demographic and urbanization transitions near their end, the pressure will ease, but a special effort is needed during the coming decade. The necessary components of this effort are as follows:

1. Acceptance of private construction of dwellings, often incrementally, on serviced land in urban areas by government and urban communities. Acceptance has always been the case in urban settlements in traditional areas, but it has long met resistance in urban areas elsewhere. Even so, such development had to be accommodated in the late apartheid years and the pressures for it have been increasing again in recent years. The occupation of new land by households and incremental housing is going to happen, especially in the metros. The only choice is between orderly and disorderly development. Disorderly development will lead to social tension, sub-optimal location and increased cost of servicing in the longer run. Acceptance will require political leadership.

2. The identification and acquisition of land for new residential development, and the installation of both on-site services and connection to bulk infrastructure by contractors from the private sector. Serviced sites can be provided free of charge to the poorest households, at partly subsidized prices to lower middle income households, and at cost to higher income households. This would require some reallocation of funding within the state housing budget, but not necessarily an increase in aggregate human settlements expenditure at the three levels of government.

3. The organization of complementary inputs, such as support of small building contractors, and micro-finance providers, as well as building plans for incremental housing to assist households to build optimally towards consolidation in later years.

4. A more determined approach to densification. It would be rational to incentivize the process up to a unit cost equal to the unit cost of green field development. Densification may lead to locational advantages for new households thus accommodated.

Urban land reform is feasible, and it requires no constitutional change, and little alteration to the existing legal and fiscal framework. It needs to be aligned to urban development plans, and it should be directed towards the entire household income range. It needs to be tackled with speed. Getting to grips with it in the five largest metros would be a very substantial start.

Charles Simkins is a distinguished Economist. He was Vice President and Professor of Economics at St Augustine College and formerly held the Helen Suzman Chair of Political Economy at Wits University. He is a former Rhodes Scholar and is a recipient of the Helen Suzman Chevening Followship, a UK Foreign Office award. He is a currently the Senior Researcher at the Helen Suzman Foundation.

This article first appeared in Focus 83, the journal of the Helen Suzman Foundation.