The Density Dilemma

Sep 12, 2014

Our colonial and apartheid past created a wide set of inequalities, and a fragmented governance system.  The effects of these remain writ large on our cities creating enormous difficulties.  The Group Areas Act, migrant labour system, native reserves and bantustans have created some of the world’s most spatially fragmented, divided and complex cities.

And in all of this, probably the most difficult challenge we must address is the density patterns which have created cities which are inefficient, ineffective, uneconomic and inequitable.  Unless we redress these we will continue to reinforce apartheid spatial patterns.  Housing for the poor and working class people located far from centers of employment, perpetuate cycles of poverty, with many poor households spending a third, sometimes even half of their monthly income on transport.

In most international cities, density patterns are fairly similar, with higher densities near city/suburban centres and places of work and reduced densities further out.  Working class people are usually located closest to work opportunities and are easily able to access and use public transport systems.  This makes it easier (and more efficient) for government to provide services such as water, sanitation, electricity and public transport.  Walking and cycling also become more viable.

By way of contrast in South Africa we have a reverse pattern.  Forced removals and the Group Areas Act meant that black South Africans lived in high density areas quite removed from work opportunities and city centres.  Unsafe public transport systems developed and huge challenges in providing services to informal settlements emerged, including the challenge of providing bulk infrastructure to settlement areas, land ownership challenges.  Dense living in informal settlements made laying service infrastructure very difficult.

In eThekwini and Msunduzi, these patterns play out in ways that continue to reinforce inefficiencies.  For example, on average, eThekwini has some 2300 persons per square kilometre with Msunduzi has some 970 persons per square kilometre.  These are low in comparison with other international cities –  Buenos Aires, for example has an average of over 15 000 persons per square kilometre.  Our average densities are also not located uniformly in space, with Ashburton only having around 100 persons per square kilometre whilst Edendale/Dambuza/Imbali have over 4000 persons per square kilometre.  And the latter areas also have the highest number of informal dwellings.

In eThekwini similar patters are found, with higher densities in Inanda/Ntuzuma/KwaMashu and lower densities in Durban North.

How then do we deal with this problem?   How do we create cities where the poor are not further marginalized, but are able to easily and cheaply access job opportunities and state services?

The first solution is to densify residential housing close to areas of economic opportunity and along major public transport corridors.  Such solutions have been hugely successful in older European and newer Asian cities.   Importantly, densifying these areas must be carefully planned – high or medium density housing must be accompanied by better urban infrastructure.  People living in high density areas need far more parks and recreational facilities than those in the suburbs.  Municipalities need a higher focus on maintaining urban cleanliness and dealing with criminal activity.

Unfortunately though, South Africa’s RDP housing strategy has not been successful in this regard.  Well located land which could be used for higher density settlements is often too expensive and because of the challenges associated with expropriation we have simply not gone this route.  RDP housing has also been predominantly in the form of low density single units, requiring large areas of land.  As a result, we have continued to build housing further from centres of economic opportunity.

Another possibility is to create in-situ densification in central areas.  This requires amending the town planning scheme to allow for higher density development – possibly in some of our more centrally located neighbourhoods.  This concept, however, needs to be carefully and sensitively implemented so that the character of residential areas are not destroyed.

The second route – which will assist those who are already located in outlying areas –  is to create Integrated Transport Systems, using “hub and spoke” approaches, such as having higher speed bus and train systems between residential areas and economic nodes.  This should then be complimented by circulatory public transport systems around those centres.    This is largely the basis on which many cities are developing their Integrated Rapid  Public Transport Systems (IRPTN) – with huge investments being made into transport infrastructure.   The problem here is that for such systems to work billions of Rands in operational subsidies will be required.  Low residential densities make public transport a costly exercise – and this cost clearly cannot be borne entirely by the users of the system. Whilst all our major cities are embarking on such needed schemes, it is not clear how they will be funded going forward.

There is a clear need to knit our cities together to create more efficient and effective spatial patterns.  Densifying some areas and improving public transport to others can achieve this goal.  But, this will need done with focus and care – in identifying suitable central land for housing, creating livable residential environments, looking for acceptable ways to increase the density of our central suburbs and finally to providing better public transport to outer-lying low density suburbs.  None of this is straightforward and will require substantial and ongoing investment.  The consequences of not addressing the issue are however much more costly: land invasions will continue, the poor will be further impoverished by living in poor living environments and having to pay substantial amounts for transport and our cities will not function either effectively nor efficiently.

The images below show the distribution of residential densities in both Durban and Msunduzi.


The image below shows the proposed IRPTN system in eThekwini.