Why the World Bank is wrong

“Everything should be made as simple as possible, but no simpler” — Albert Einstein

The World Bank has been pressing for the removal of restrictions on Floor Space Index (or at least their considerable relaxation) in the inner parts of Mumbai and Bangalore. The thinking that leads to this recommendation is based on a mathematical model. This particular model, unfortunately, is oversimplified and neglects important relevant parameters. If you take these into account, the recommendation could be, quite possibly, more damaging than helpful.

Assumptions

Called the monocentric-city model, this assumes the city is inhabited by a number of identical residents, each earning the same income, and who all work in the central business district (CBD). The inhabitants commute from their residences to the CBD on a dense radial road network, paying a certain amount per round-trip mile (including the cost of travel time). An implicit assumption is that commuting is by car, not by public transport that might lead to some preferred travel routes, and not by walking which incurs no explicit travel cost. Further, each resident is assumed to be a renter, paying a certain amount per square foot of housing, and occupying a certain amount of square feet of housing. The price naturally falls with distance from the CBD, and the housing area occupied increases.

The mathematical formulation then moves through various complexities to the consideration of FSI, and what happens when FSI is restricted (as it is in most cities). The FSI restriction tends to limit population density in the central part of the city; and so causes the city to spread out.

Studying the condition of the new resident as the city grows, the model suggests that if FSI restrictions in the inner city are removed, this accommodates more people closer to the centre and so their commuting times are reduced. This is claimed to result in a welfare gain for them, which is measured in terms of their reduced commuting costs. So the claim is that lifting FSI restrictions brings about an overall welfare gain.

The first indication that something is flawed in the argument comes from consideration of where the limit lies if we allow unlimited densification. Every increment in density will result in still more saving for commuters. As the model stands, there is no limit to the saving possible, until commuting cost is reduced to zero. We are obviously missing something. There has to be a constraining parameter that would put a cap, sooner or later, on how small the city can get while being both efficient and attractive.

This constraint is congestion and its consequences. The first untenable hidden assumption in the model is that travel speed is constant, and travel time is proportionate to distance and nothing else. In reality, if we retain the model’s assumption that all travel is by car, as densities rise and streets become more crowded there is bound to be a reduction in travel speeds. It is common knowledge in transportation engineering that as demand increases, road speed reduces and the hourly throughput of vehicles falls. So FSI permitted in a locality needs to be judiciously related to transport demand and transport capacity.

The second hidden assumption is that existing residents are not adversely affected by densification. The fact is that they will now have to share their open spaces and other amenities with new residents, and surely this results in some loss of welfare for the older residents.

A third use

In essence, the model assumes that there are only two uses for land in the city: either for transport by car or for residences. In practice, of course, there is a third use, and that is for open spaces, schools and playgrounds, and land needed for other public amenities. These could be considered a fixed proportion of the total and therefore irrelevant to the argument of the model. However, since localities once built up are hard to change, with densification the availability of amenity space per capita will necessarily diminish, resulting in a loss of welfare for existing residents.

This is not to deny that a more compact city would be more efficient in terms of transport costs, provided it was designed ab initio keeping in mind the necessary high transport capacities and providing adequate physical and social infrastructure towards the city centre.

We should also recognise that FSI can be used in two ways: either it can increase the consumption of floor space by individuals without affecting the total population count in a locality; or it can be used to accommodate more individuals within the same locality, thus raising densities.

There can surely be no harm at all in raising FAR if all it does is give each individual more floor space without encroaching on his neighbour’s light and air.

We should also note a number of other limitations of the monocentric-city model:

• Much employment, certainly more than half, is endogenous, such as laundries and restaurants and branches of banks, serving a local clientele, and therefore is distributed throughout the city. This proportionately dilutes any argument based on the monocentricity of employment.

• Such exogenous employment as there is is often concentrated according to specialisation and distributed over multiple centres, not just one CBD. This may or may not invalidate the findings for a monocentric city. But what it certainly does is add trips from one workplace to another, and therefore dilutes the proportion of commuting trips to all trips.

• Transit lines, even regular bus routes, alter the preferred directions of city growth. The dense radial pattern is no longer meaningful.

• People attach a value to living at comfortable densities, neither too high nor too low, and this value is ignored in the model. Individual preferences do vary even within the same income group. Those with families may prefer living at lower densities (this is quite separate from having a larger house or plot).

• Some localities are preferred to others, on grounds that have nothing to do with distance from the city centre. And there is considerable variation in this choice from one individual to another.

A recent study, by Shlomo Angel and others, of a global sample of 120 cities notes that urban densities have been declining in cities around the world. Between 1990 and 2000, mean built-up area densities declined in 75 of 88 cities in the developing world, and all 32 cities in the developed world. The rate of density decline averaged just over 2 per cent per annum.

If, as the World Bank suggests, increasing FSI, with the objective of increasing densification, delivers an overall welfare gain, how come, world-wide, there is a relentless drift the other way?

(Shirish B. Patel is a civil engineer and urban planner. He was one of three original authors of the idea of New Bombay, and for its first five years was in charge of planning, design and execution for the new city)

Related Topics