Principles first, planning follows


back to issue

URBAN planning is as much or as little a part of urban governance as you wish it to be. In India it has long since been turned into an irrelevance. Part of the responsibility lies with politicians. In Maharashtra for example, successive chief ministers have steadily eroded all urban planning functions in every organization, from the municipalities to the regional planning authorities. All such decisions are now centralized in the Urban Development Department, directly under the chief minister. They consist essentially of modifying the Development Control Regulations, the set of rules that regulates development. These are now framed, cast or recast by bureaucrats, with no participation by urban planners in the process. Builders, as it happens, have more influence over the changing of these rules than do urban planners.

But some of the responsibility for their irrelevance surely lies with urban planners themselves. For decades they have not questioned the viability of the urban planning processes as set up in our legislation. That process is roughly as follows: every 20 years or so, prepare a map of existing land use, carry out socio-economic surveys and study population and economic trends. On this basis, project what you think the city will be like 20 years from now, including details of the future mix of economic activities and socio-economic characteristics of the population. Ignore the fact that every such projection in the past has been a spectacular failure, with the reality turning out to be very different from what was projected 20 years earlier.

One can put this down to the incompetence of an earlier generation of planners. Assume that one is far, far smarter, and that in 20 years time the city will turn out to be exactly as predicted. Further, that no new economic activity or new technology (of which one is not now aware) will be invented in the coming decades. On the strength of this assumption, one can decide which parcel of land shall be put to which use, and draw up a master plan which will be sacrosanct and binding on all for the next 20 years, regardless of how the situation may change and the cracks begin to show between what one thought would happen and what is actually happening. The master plan will be accompanied by a set of building control regulations, which may be tinkered with from time to time, but there is no changing the master plan until its next revision 20 years later.

Delhi’s first master plan (1961-1981) was amended and extended for another 20 years (1981-2001), but published in 1990, halfway into the plan period. Recently the Supreme Court to its horror discovered that the city had developed in ways contrary to the provisions of its master plan. Close to over two-thirds of the construction in Delhi was in one way or another illegal. To bring everything within the bounds of legality would require either large-scale demolitions, or recasting the master plan such that it accommodated the existing realities. The new master plan (up to 2021) was formulated accordingly, in a great hurry, and opened up to the public for suggestions and objections. Interestingly, very few questioned the planning process itself. The idiocy of the current process should be self-evident.


Ideally, urban planning should be a three-step process: First, a declaration of principles that will drive the planning process, and serve thereafter as a signpost against which each subsequent policy or project proposal is judged. Typical principles might be the following:

i) Provide municipal services to all income groups in the city.

ii) All localities must be mixed-income and mixed-use localities with a variety of housing sizes to suit different income groups.

iii) Densities should be variable, with the highest densities concentrated around transport nodes (stations or other interchange points where transport mode transfers take place), tapering off to lower and lower densities as one gets further away from the node.

iv) When resettlement is undertaken, and plots or pitches assigned, the period of lease will be not less than 30 years; and when such a lease comes up for renewal it must be based on a fair valuation derived from current market prices, unless the land is needed for some more urgent public purpose.

v) Public transport has priority over private transport: each receives funding in proportion to the number of its users.

vi) Discourage the use of cars; manage traffic with a combination of improved technology, better policing, and pricing policies.

vii) Encourage preservation of the character of the city.

viii) Expand green spaces and make them accessible to all, within walking distance of where they live.

Such principles, and their order of priority, will vary from city to city. They should be established after careful debate that involves the public and reviewed and readjusted every five or ten years.


The second stage of planning should be the formulation of a strategy plan for the city that looks far into the future, including in particular the economic future, and sets out the basic strategies for dealing with the fundamental issues of broad land use, transportation, water supply, sewage treatment, solid waste disposal sites, power supply and the creation and maintenance of a GIS database for the region. The vision for the future should recognize that economic activity may change in unforeseen ways; and the strategies proposed should be flexible enough to cope with such change.


Another essential constituent of the strategy plan would be a parallel companion plan for financing and recovery of investment. Without this no part of the strategy plan proposal should even be considered. This cannot be sufficiently emphasised. All plans for urban development in the past have remained just that: only plans, with no clear direction as to how they are to be implemented. In future we should expect every plan, whether at the strategy or at the more detailed local level, to be accompanied by a detailed, companion financing plan, which covers how capital is to be raised, where the funds for servicing loans will come from, and how borrowed capital is to be repaid. Acceptance of the strategy plan should necessarily require simultaneous acceptance of the financing plan. Only if this is done will the plan proposals be meaningful and implementable. Pure technical plans, even if accompanied by cost implications, in the absence of clear tax and financing arrangements, often remain little more than wishful thinking.

A further, equally essential constituent of the strategy plan would be the action plan, setting out the responsibilities for implementation – who is going to do what – with datelines. Once again, as with the financing plan, the strategy plan is meaningless unless accompanied by a companion action plan.

The strategy plan should also set out the guidelines to be followed for the next level of detailed area planning, including in particular specifying the range of desirable densities and various kinds of possible building controls. Note that transportation and land use planning would go hand-in-hand, with densities of occupation as an integral part of land use planning. Mixed land use could be permitted in any zone, with restrictions if desired, but very few fixed prescriptions at this strategic level of which particular plot of land should be used for what.

Prepared at the level of the regional planning authority, the strategic plan should be reviewed every five years. It would be sent in draft form to the next lower level of planning for comments and suggestions. Differences of opinion would be resolved after giving the public an opportunity to be heard. A mechanism would be in place for time-bound conflict resolution.


The third and final stage would be the preparation of detailed area plans which show detailed land use planning and development controls. The preparation of these local area plans would be the responsibility of the various local authorities, typically acting for a locally elected municipal councillor, working in conjunction with local residents. They would also be reviewed every five years and would cover the following:

Water distribution network; storm drainage network, including rain-water harvesting; sewage collection network; solid waste primary collection, including sorting of garbage; location of bus stops, taxi parking ranks, public parking; power distribution; road and footpath widening; location and demarcation of public open spaces; location and sizing of various public amenities, including public toilets, schools, colleges, hospitals, police stations, fire brigade stations; definition of land uses at the detailed, local level, in consonance with the overall regional guidelines; development of building control regulations, which may vary from one locality to another within the local area, but which are again in consonance with the guidelines spelt out in the regional plan. For both the planning and the implementation of the above, we may involve what might be called citizens’ area sabhas. These would be much like Mumbai’s advanced locality management groups, except that each area sabha should have a footprint coinciding with that of an election booth (between 800 and 1,500 persons).

Here too, as in the case of the strategy plan, it is important to have a companion document that sets out the tax and financing arrangements proposed, and another companion document setting out the action plan, so that one can be assured of successful implementation.


The really vexed question in the matter of detailed land use planning is the determination of the rights of land owners in the locality. There are two problems here: one concerns respecting individual plot boundaries; the other is the decision as to where (on whose plot) a public facility such as a school or a public toilet should be located.

The first issue, of readjusting plot boundaries to provide more public space, should be resolved on the basis that the primary owner of all land in an urban area is the public authority; and that the ownership of individual owners is secondary and subordinate to the first. We have years of experience of town planning schemes, where an entire locality is taken up for development, a layout is prepared, and existing landowners are given plots in the new development which are smaller than their earlier plots, but in proportion to the earlier holding. The same mechanism could apply to a situation of redevelopment where roads need to be widened or common facilities provided which did not exist earlier. Everyone in the locality benefits from the improved public facilities provided, and everyone contributes some fraction of his land ownership to this end.

Where exactly should a particular public facility, such as a school or hospital, be located is normally open to a variety of options. Why one particular plot owner rather than another should be the sacrificial goat – even if he is in principle compensated in some way for giving up his land – is a matter to be resolved. The answer surely lies in making the giving up of land for a public purpose so attractive that the public authority will have multiple offers to choose from, and can do so following any procedure that is open to scrutiny and fair to all.


One instrument whose potential has not been fully explored is the right to construct floor space. Today that instrument is the floor space index, FSI, which defines the ratio of build-able floor space to plot area. It is assumed that this must be some kind of constant number across the whole city, or at least across large swathes of it. Now there is no reason why ownership of land must necessarily carry with it the implicit permission to build large amounts of floor space on it. The right to build could be separate from the ownership of land.

Once this principle is accepted, the right to build could for example arise from the surrender of land for a public purpose. The owner of the right to build is free to sell that right to someone else who owns land but cannot build on it unless he purchases that right to build from someone else. The public authority should of course be careful that it parts with rights to build only in such measure as the locality can sustain.

These rights to build may be issued by the public authority in a variety of ways and for a variety of reasons: in exchange for land to be used for a public purpose; or sold for money to finance infrastructure; or given in exchange for the construction of middle- or low-income housing. The point is to treat the right to build as a tradable commodity in urban areas, subject to such constraints on its use as may be important in particular localities.


There is another vexed question, and this concerns the rights of occupants of lands and buildings who are not the owners but who happen to be in possession, either as legitimate tenants or as squatters. Mumbai still has tenants whose rents are frozen at World War II levels. And it has half its total population living in slums – that is, in unauthorized occupation of land that belongs to someone else. Slum dwellers in Mumbai have acquired a kind of semi-legitimacy which does not exist for example in Delhi, where slum settlements are ruthlessly bulldozed regardless of how long they have been there, and regardless also of whether years ago they were officially put there in the first place by government.

In Mumbai the difference is that slum dwellers throughout the city have been promised that if they can prove they came to live in the city irrespective of where they happen to be before 1 January 1995 (the ‘cut-off’ date, which keeps getting regularly extended), they can continue to live where they are. They will also be provided with free pucca accommodation constructed without cost and effort on their part, at their current location. This is to be achieved by inviting developers to build slum dwellers’ free housing financed through free sale in the open market of an equivalent amount of floor area, both on the same plot. Profits from the free sale are more than enough to pay for the free housing for slum dwellers, and the scheme has been shown to work.

Unfortunately, it can work only for those slum pockets which are in localities where property prices are high, and the high profits realized are sufficient to pay for the free construction. Moreover, if the scheme is expected to cover the entire population of slum dwellers, which is half of Mumbai’s total, it can only work if there is a large enough number of new city residents willing to buy so many new flats, scattered throughout the city where various slums now exist.

This kind of large market for so many new high-value flats sounds unrealistic. But there is another more fundamental problem – there is an upper limit on how high densities can go. So if a slum is already densely occupied, there can be no question of adding further numbers who will occupy the ‘for sale’ component because the densities then become unmanageably high.


This has been realized only recently, when the question came up of further densifying Dharavi beyond its present levels of occupancy. The realization arose from asking the following question: The Government of Maharashtra has specified a minimum built-up floor area of five sq m per capita inside the house (by stipulating a minimum house size, and we assume a family size of five, the norm for Mumbai); but what is the assured ground area per capita outside the house? Here we find that no assurance of any kind has been provided and indeed, all plans are proceeding on the basis that ground area per capita outside the house can be reduced to zero.

This is clearly untenable. People need a minimum amount of space on the ground just to circulate – after all, they live in the city not just to be at home around the clock. They need to get out and about, to do a job, or acquire an education, or visit friends. If we analyse existing localities around the world1 and measure the amount of circulation space people have, we find that in the most crowded places the amount of ground space per capita for roads and footpaths alone cannot be less than three sq m. More would be better. There is also a minimum amount of ground space needed for social amenities, for schools and hospitals, the fire brigade, police stations, electric sub-stations.

While amenities ideally should not be less than 10 or 12 sq m per capita, a rock bottom minimum might be two. Open spaces for recreation are extra and ideally should be around 16 sq m per capita. Let us provide zero on this account, because we are seeking not a comfortable city, but the most densely packed city possible. We find we cannot do with less than 5 sq m per capita of ground area. That is, we need not less than 5 sq m per person inside the house (on any floor) and not less than 5 sq m per person outside the house, on the ground. We call this area outside the house public ground area (PGA).


Now if we assume we have a hectare of land (10,000 sq m), at a PGA of 5 sq m per capita, we cannot have more than 2,000 people on that area. They will need all that ground space for circulation and minimal amenities, with nothing left to build on. If we set aside 40% of the total area for construction (4,000 sq m), with the rest as PGA, we can have 1,200 people (1,200 X 5 = 6,000 sq m of PGA) and on the buildable plot of 4,000 sq m we would house our 1,200 in 6,000 sq m of built-up area (which they need at 5 sq m per capita) for an FSI of 1.5 (6,000 sq m built up on a plot of 4,000 sq m). The critical conclusion is that our gross densities, even with the best management, cannot for practical purposes exceed 1,200 per hectare. And the interesting additional insight is that if you want more crowding you need to reduce the plot areas devoted to construction: the smaller the plot areas in relation to the total, the more people you can have by way of density (but of course you would have to build higher).


While we argue about how densely you can pack the poor, the cities of the developing world seem to be drifting towards a sharp social and economic segregation. There are gated communities of the rich and ghettos of the poor, distinguished not so much by the amount of built-up floor space per capita as by the PGA: the rich, with their golf courses, tennis courts and horse riding trails will have enormous amounts of ‘public’ ground area per capita compared to the poor – the ‘public’ in this case being restricted to members of the club.

Kemer Country,2 a gated community half an hour out of central Istanbul, is spread over 1,200 acres (480 hectares) and is expected to eventually house 5,000 people, that is, a total ground area of about 960 sq m per capita, or if you prefer, a gross density of about 10 persons per hectare. This should be seen in the context of 5 sq m per capita (corresponding to 1,200 persons per hectare) being grudged to the population of Dharavi. There already are gated communities of the rich in Mumbai, with their own swimming pools and clubhouses and jogging tracks and even schools – it would be interesting to know the public ground area per capita within these communities, accessible only to members and their guests, as compared to the PGA in the rest of the city outside these gated walls.

When we have such wide disparities of income, with half the city’s population living in slums, is there an alternative to gated communities of the rich (projects vigorously pursued by developers and the providers of high volume housing finance) and ghettos of the poor (projects mandated by government for one reason or another)? Does all new city growth have to take place only in this way?


As it happens, there exists in north-west Mumbai at Charkop a wonderfully successful example of mixed-income housing, ranging from middle-income apartments to sites-and-services, that is, the provision of plinths with wet points – a water and a sewage connection – on which former slum dwellers were given secure tenure and allowed to build whatever kind of shack they wanted, and improve it over time. Today, 20 years after the scheme was started, Charkop is a thriving settled community with everyone in a pucca house, some single-storeyed, some double. The settlement is organized into multiple societies of 46 families each, with each society built around a common courtyard. Some families are in larger, single-house plots and some people in apartment blocks. The school, playground, public recreation ground and streets are shared by all. The gross density is 850 persons per hectare. The settlement covers 57 hectares and could be similarly extended to any size. We have a real world, working example of successful mixed-income housing. This needs to be replicated instead of leaving it to builders to turn large tracts of urban land over to the very rich for their private use.



1. Shirish B. Patel, Alpa Sheth and Neha Panchal, ‘Urban Layouts, Densities and the Quality of Urban Life’, Economic and Political Weekly, 30 June 2007, 2725-2736.

2. Serife Genis, ‘Producing Elite Localities: The Rise of Gated Communities in Istanbul’, Urban Studies 44(4), April 2007, 771-798.