Affordable housing will find no buyers unless it comes with high-speed, low-cost transit connections.

Our urban population will expand dramatically, so it should be obvious that city footprints must also expand. “Going vertical” in a city like Mumbai, already one of the world’s densest, is no solution. All it does is multiply congestion and degrade quality of life without affecting property prices – those are determined by market demand and, as we have seen, higher floor space index (which determines the buildable area on a plot) automatically means higher land prices. Only new land, opened up with new rapid transit links, can help attenuate land prices.

Expanding the city’s footprint should not be confused with urban sprawl. That comes with the automobile and low-density living. Instead, we propose expansion into dense settlements efficiently served by public transport. Besides accommodating new arrivals, such settlements could improve living conditions in existing congested areas by attracting away some residents.

Goals for development

We want a zero-subsidy framework, to make such projects easily replicable. Development would then have to be mixed-use, with jobs as well as homes, and mixed-income, so that we can have a cross-subsidy (charging higher prices to one group to subsidise lower prices for another) in land cost and, consequently, affordability across all income groups. Housing of different sizes would be for as wide a range of income groups as possible – on separate plots if need be, but within the same locality. Providing housing for all, including the lowest income groups in proportion to their presence in the population, is the only way to avoid slums. Such mixed-income housing across a locality is mandated in Spain and Ireland, and enabled in many countries, including England, France, the United States and Canada. With shared amenities – schools, parks, medical facilities – this makes for a more egalitarian social mix and, very likely, a more peaceful society than one with ghettos for the poor in one location and gated communities for the rich at another. The mixed-income, mixed-use arrangement also reduces commuting, because many lower level service jobs will be in the households or activity centres of the better off within the same locality.

We suggest the following principles to guide development:

  1. Proportionate representation of every income group across the city in the housing, plus as many jobs as there are resident workers – although, admittedly, many such local jobs may be filled by non-residents and many residents will work outside.
  2. A public transit-oriented development with strict segregation of transit from normal traffic – to guarantee high-speed transit.
  3. Uniform net plot densities for all income groups.
  4. A cross-subsidy in land prices across income groups follows automatically with uniform densities of dwelling units per hectare across income groups, but different limits for buildable area, more for higher-income plots, less for lower income, and exactions of payment for land proportionate to income.

Going the BRTS way

The cheapest and most readily expandable feeder system would be a Bus Rapid Transit System on an exclusive two-lane roadway track, with overtaking lanes at platforms spaced 700 meters apart. For high speed and capacity, it must mimic a railway system – boarding and alighting on platforms that are flush with the floor of the vehicle, wide vehicle doors that allow quick passenger discharge and boarding, and ticketed access to platforms to minimise en-route delays. Such a system is not a regular bus system. As in a railway, the vehicles and the roadway would be exclusive to the Bus Rapid Transit System, never mixed with other traffic.

Assuming such a system serves people residing within 500 meters on either side, a 10-km Bus Rapid Transit System loop would open up about 10 square km (1,000 hectares) of land for new development.

To attract ridership, the headway between Bus Rapid Transit System vehicles should never exceed, say, 10 minutes, and two minutes or less at peak hours. Vehicles may be smaller when the settlement is sparsely inhabited, larger as demand builds. Average speed, including stops, should exceed 25 km per hour.

To achieve this average speed, the system must have exclusive right-of-way without intersections. All other traffic must be grade separated, crossing below or above the ground-level Bus Rapid Transit System. Keeping it at ground level minimises initial investment in the track and platforms. Only at the railway station would the Bus Rapid Transit System be lifted to discharge passengers directly on the footbridge crossing the railway tracks. This would thus, free up the ground immediately outside the station for other vehicular traffic.

A pedestrian underpass at every stop on the Bus Rapid Transit System would provide access to its island platforms. This underpass would be widened at alternate platforms so cars can also cross under the system. Heavier vehicles would use shallow-rise overbridges spaced every few kilometers, just high enough to clear single-decker vehicles below. The road network for cars and other motorised vehicles through the development would thus be completely separated from the ground-level Bus Rapid Transit System.

Apart from being fast and reliable, the Bus Rapid Transit System also needs to be cheap. This almost certainly implies a subsidy, whose source should be a carefully considered progressive tariff independent of the ticket cost to travellers. Those who frown at subsidies should know that bus systems around the world are subsidised, and that this is amply justified by their contribution to reduction in traffic congestion and air pollution.

Financial viability

For replicability, the proposed development must be self-financing, with no call for external subsidies. So we must recover from occupants the cost of land, physical and social infrastructure, and construction of their own homes at a price they can afford. International practice recognises four years’ income for a household as an acceptable standard for housing affordability.

Table 1: Affordable land cost with four years' income from 1,000 households.

Based on the income profile of Mumbai, Column 1 in Table 1 shows ranges of monthly household income, and Column 2 the percentage population in each income range (we exclude the 4% that forms the highest income bracket as irrelevant to the scheme). Column 3 is four years’ income for each category, that is, Column 1 multiplied by 48 months. Column 4 suggests house sizes for each income category, based on what in Mumbai would be generally acceptable. We assume construction costs vary by income category from Rs 18,000 per square meter to Rs 24,000 per square meter, and Rs 20,000 per square meter as the cost for workers’ spaces. Cost of construction is in Column 5. Column 6, the difference between Columns 4 and 5, represents the surplus after construction available per household in each income group to pay for land and infrastructure.

We convert fractions in Column 2 to absolute numbers in each income group for every 1,000 households, and multiplying by the surplus per household in Column 6, Column 7 shows the surplus generated by each income category.

We do not want a purely residential development. Intermingled jobs would be desirable. The employment ratio in Mumbai is 40%. So we provide for a commensurate number of jobs. The resident population itself will generate many lower-income jobs, ideally taken up by residents to minimise transport costs. We should not inhibit the mixing of residences and zero-nuisance businesses in the same building, or be too rigid about proportions of residential to work space – let the market decide that, and keep planning flexible enough to accommodate variations. To calculate area requirements and construction costs, we assume a 9 square meter built-up area per job and a floor area ratio (the ratio of a building’s total floor area to the size of the plot) of 1, which allows for a mix of single- and multi-storeyed construction.

In Table 1, Rows 1 to 6 are for residential construction. Row 7 concerns construction of work places. Row 7, Column 2 shows the percentage of each household that is employed. Assuming Mumbai’s average household size of 4.5 persons, with an employment floor space requirement of 9 square meters per worker, 40 square meters is required for 4.5 workers (Row 7, Column 3). Since we are calculating per household, we might think of this as “workers’ spaces” for computational purposes, with each accommodating 4.5 workers (as in a household). The affordable value for this (Row 7, Column 4) is based on current business property prices per square meter at sites just north of Greater Mumbai.

From our 1,000 households, each paying four years’ household income, after paying for all construction, we have a surplus of Rs 190 crores. The additional 400 workers’ spaces also generate a surplus, the difference between market price and construction cost. The total surplus is about Rs 317 crores for every 1,000 households (with their 400 associated workers’ spaces), available to pay for land and physical and social infrastructure.

Land required

We assume a within-plot density of 270 households per hectare. In the successful 30-year-old Charkop housing project in Mumbai, density in low-income areas is 278 households per hectare, which would still be considered acceptable.

An important assumption here is that the same density applies across income groups, so wealthier residents consuming more floor area will live in taller buildings. Moreover, each household consumes the same land area. Since we are taking four years’ income from every household, the wealthier are paying more per unit of land – a cross-subsidy – and land per se is differentially priced across income groups. This can be justified in many ways, including attractiveness of location or building regulations, say, that might limit the volume of permissible construction or deny motor car access or parking to the lower income locations. Or the charge for land could be uniform per unit of built-up floor space, and with uniform densities this would painlessly deliver the desired cross-subsidy.

Allowing 30% of the total area for roads and rapid transit, 4% for institutions (schools and hospitals), 13% for parks (based on a minimum of 3 square meters per capita, low but acceptable in Mumbai) we are left with 53% for buildable plots, mainly residential but with shops and offices allowed.

Our land area requirement then works out to just over 10 hectares for 1,000 households and 400 workers’ spaces. The gross residential density of the development would be about 100 households per hectare. Assuming Mumbai’s family size of 4.5, this is 450 persons per hectare.

From the surplus, we have in hand Rs 317.5 crores (see Table 1). Deducting the cost of the Bus Rapid Transit System (Rs 0.3 crores per hectare), physical infrastructure (Rs 2.4 crores per hectare), social infrastructure (Rs 2 crores per hectare), and fees, taxes and interest (Rs 2 crores per hectare), we are left with Rs 24.9 crores per hectare to pay for land.

What is striking is how little the transport system costs in relation to all other infrastructure. Providing this kind of dedicated transit service is not as extravagant as one instinctively imagines.

The figures indicate the scheme is financially viable wherever we can find land around Mumbai under about Rs 25 crores per hectare. This sounds eminently within reach at several suburban railway stations around the city.

Shirish B Patel is a civil engineer and urban planner, one of three authors who first suggested the idea of Navi Mumbai and then for five years was in charge of the plan, design and execution of the new city.

Oormi Kapadia and Jasmine Saluja are both architects and urban designers, and recipients of an awarding-winning proposal for a Community Land Trust Model for the redevelopment of Dharavi in Mumbai.

This is a curtailed version of a paper that will be published in a forthcoming issue of Environment and Urbanization.