Providing infrastructure before settlement

Under current policies, many developing cities are expanding without co-ordination and lacking supporting infrastructure. As a result of unplanned development, many cities only have 10% of their land allocated to roads. The international benchmark is 30%3. Housing has also been built in areas that lack key infrastructure, and where legal ownership of land is unclear. The lack of policy action today is storing up costly problems for the future; infrastructure and land rights are extremely difficult to retrofit. African cities are set to triple in size by 2050, and South Asian cities are set to more than double. Only if governments plan in advance will they get it right the first time. But for plans to actually get implemented, they need to be realistic and prioritise.

Mathare slum in Nairobi, Kenya. Due to a lack of timely planning, the slum now needs retrofitting. The government lacks the practical authority to do this, and even it did, retrofitting would come at a huge social and economic cost. (Photograph, Claudio Allia, 2009).

Most governments do not have the budget or the expertise to build houses for everyone who comes to the city – but the choice is not between public housing or no plan at all. Instead, governments can provide the core infrastructure around which orderly private development occurs.

In practice, proactive infrastructure provision can be more or less costly depending on the city’s budget:

  • At one end of the spectrum, governments can simply demarcate an arterial road grid on the urban periphery, around which settlement can occur. Estimates using figures from Kigali cost this at roughly $100 per household, for a 1km by 1km grid.
  • At the other end, governments can provide households with serviced plots; this involves creating a neighbourhood street patter, demarcating and titling land plots, and providing water, sanitation and electricity to the unbuilt plots. Estimates in Kigali cost this at $3,500 per 50m2 serviced plot – 8 times cheaper than Kigali’s $30,000 public housing units. 4

The table below uses housing cost estimates from Kigali to provide a rough estimate of relative cost of various policy options to accommodate the next 10 years of population growth5 in Kigali, Lagos and Kampala (costs are in USD).






Proactive infrastructure provision reduces land costs

By connecting the expanding edges of the city, infrastructure ensures a well-connected supply of land on which the private sector can build. Increasing the supply of urban land reduces land costs throughout the city. Providing infrastructure, particularly transport links, to increase land supply was key to solving housing shortages in London and New York as they developed – and may well be even more crucial in developing cities, where land costs can exceed 80% of total construction costs.6

Road grids and serviced plots

One low-cost way of planning for urban expansion is to build an arterial road grid of dirt roads on the periphery before settlement occurs. This was the approach adopted by the City of New York in their 1811 Commissioners Plan. At the start of the 19th century, the population of New York was just under 100,000 but was estimated to increase five-fold in the next 50 years. In response to this, the city devised a bold plan to expand the urban area of the city seven-fold. This plan mapped and demarcated a grid system of roads on undeveloped agricultural land in Manhattan, leaving roughly 30% of land within the expansion area for public infrastructure uses. Initially, the city only obtained a public right of way over the grid, but as the city expanded, land was acquired to build paved roads, with water and sewerage infrastructure underneath. This enabled structured and connected urban development. By 1900, land within the grid had become fully developed, and a new grid was created for further expansion. The same grid systems created by these plans carries New York’s traffic and infrastructure to this day.

The ‘bare bones’ grid-based approach to urban expansion taken in 1811 New York is currently being drawn upon across many cities in Colombia and Ethiopia. With support from the NYU Marron Institute, governments in these countries are predicting future land requirements for urban expansion based on population growth projections, and obtaining the right of way for a 1km by 1km network of arterial roads on the urban periphery.7

This approach can be particularly useful where public trust in government housing policy is low, because it delivers quick, large-scale and visible results. In Valledupar, a city in northern Colombia, the city has planted trees to line the future road grid, providing a visible and popular signal of proactive planning for urban growth.

The grid system laid down by the 1811 Commissioners Plan is still in place in New York to this day. Left photograph: History of Architecture CCA, 2009, Flickr. Right photograph: New York from the air, Getty Images.

With more resources, governments can focus not only on establishing arterial road networks, but also on creating serviced plots for areas soon to be settled.

The World Bank adopted this approach in its ‘Sites and Services’ programmes in the 1970s and 1980s. These were initially scaled back due to high costs and a perception that the programmes had not led to substantial housing developments. However, these programmes are now being re-evaluated as over time they have paved the way for thriving, and often multi-story neighbourhoods. Recent research from Tanzania has shown that areas which received ‘Sites and Services’ programmes 30 years ago have become better planned, and now have land values over five times higher than other areas which received the same amount of investment in the form of restrospective slum upgrading. 8

Funding infrastructure through rising land values

Installing infrastructure greatly increases the value of land nearby; governments can use this land value appreciation to fund the initial infrastructure investment. In 1811 New York, the city was able to acquire a road grid at very limited cost because the cost of providing compensation to landowners was offset by charging the same landowners ‘betterment fees’. These fees were charged to landowners based on the appreciation in value of the rest of their land – which was now connected to the city. Despite initial opposition, landowners benefitted greatly from this arrangement; agricultural land values skyrocketed once land became part of New York’s metropolis, and more than offset initial losses of land.


  • 3 Collier, P. and Venables, A.J. (2016) “Urban infrastructure for development”, Oxford Review of Economic Policy, 32 (3) pp391-409
  • 4 Estimates based on Halusan, B. (2017) “Multi-story Versus Single-Story Residential Construction Cot Analysis” International Growth Centre Draft Policy Brief, February 2017. Estimates include land costs of $20/m2, and density calculations based on 55% of land under residentia use, with housing plots of 50m2. Infrastructure costs are estimated based on official government figures.
  • 5 Population growth figures from Household estimates assume an average household size of five.
  • 6 Woetzel, J., Ram, S., Mischke, J., Garemo, N. and Sankhe, S. (2014) “A blueprint for addressing the global affordable housing challenge” McKinsey Global Institute
  • 7 Lamson-Hall, P., Angel, S , DeGroot, D., Martin, R. and Tafesse, T. (2018) “A New Plan for African Cities: The Ethiopia Urban Expansion Initiative” Journal of Urban Studies.
  • 8 Michaels, G., Nigmatulina, D., Rauch F., Regan, T., Baruah, N. and Dahlstrand-Rudin, A. (2017) “Planning Ahead for Better Neighborhoods: Long Run Evidence from Tanzania” London School of Economics Discussion Paper