ArticlePDF Available

Limitations of the anti-floor space index position



Floor space index is only one of the tools in a planning kit that has zoning, growth boundaries, inclusionary housing and the like. This article holds that plans for higher densities in cities such as Mumbai should include urban peripheral nodes and not just the urban core and central business district. Managing land uses and supplying infrastructure is easier in emerging urban nodes, which can encourage a balance between housing and jobs, and help establish an efficient use of expensive transportation infrastructure. To make this a reality, the development focus should equally be on improving institutional capacity.
Economic & Political Weekly EPW july 20, 2013 vol xlviii no 29 123
Limitations of the Anti-Floor
Space Index Position
Manish Shirgaokar
Erick Guerra, a colleague and fellow scholar,
provided critical feedback on this essay.
Manish Shirgaokar (
teaches at the Department of City and Regional
Planning, University of California, Berkeley.
Floor space index is only one of
the tools in a planning kit that
has zoning, growth boundaries,
inclusionary housing and the like.
This article holds that plans for
higher densities in cities such as
Mumbai should include urban
peripheral nodes and not just the
urban core and central business
district. Managing land uses
and supplying infrastructure is
easier in emerging urban nodes,
which can encourage a balance
between housing and jobs, and
help establish an ef cient use
of expensive transportation
infrast ructure. To make this a
reality, the development focus
should equally be on improving
institutional capacity.
In , “Life between Buildings: The Use
and Abuse of FSI”, (EPW, 9 February
2013) Shirish Patel presented an in-
sightful evaluation of why the proposed
changes to fl oor space index (FSI) in
Mumbai might make the urban experience
unbearable (2013: 68-74). His pre mise is
that higher FSI would generate more
built space, leading to higher crowding
(density), and without new infrastruct ure,
this would lead to unworkable urbani-
sation. This premise was based on a cri-
tique of research on height regulations
in Indian cities, a comparison of density
metrics between the islands of Mumbai
and Manhattan, and the author’s exten-
sive knowledge of Mumbai.
Given that the anti-FSI position is
largely popular in India, it needs further
examination, especially since initial evi-
dence on Indian cities shows that (within
limits) increasing FSI could be benefi cial.
Yet, increasing FSI is not enough, because
this planning tool has to work alongside
with zoning of land uses – housing, offi ce,
retail, industrial and so on – and the
supply of adequate infrastructure. Fur-
ther, city-regions such as the Mumbai
metropolitan region are inherently more
complex than central business districts
(CBDs), such as south Mumbai or Man-
hattan, because they have multiple sub-
centres outside the CBD. Therefore, my
argument supporting density is not just
in reference to CBDs, but also for peri-
urban sub-centres. In writing this essay,
my pur pose is to open up Patel ’s arg ument
with an examination of the literature
citied, to add arguments for the benefi ts
of managed and phased densi cation, to
expand on planning beyond FSI, and to
generalise the fi ndings to other cities.
The evidence on increasing FSI through
removing height restrictions is gradually
growing for Indian cities, and this section
examines it. In the works that Patel cites,
including Bertaud (2004), Bertaud and
Brueckner (2004), and Brueckner and
Sridhar (2012), it is not clear whet her these
authors are suggesting that FSI needs to
be “immediately and substantially” raised.
These works combined together argue
that the theoretical and empirical evidence
suggests that removing height restric-
tions can be benefi cial for Indian cities.
Evidence for Increasing FSI
Be rta ud et al (20 04) t hin ks t hat FSI should
increase, claiming that with improvements
in technology and infrastructure, it goes
up in most cities. He argues that increas-
ing FSI will not increase densities and that
most people would end up consuming more
real estate. However, his claim largely
looks at formal housing plus employment,
with benefi ts trickl ing down to the large
section of Indian society that lives in
slums. I contend that conditions for hous-
ing the urban poor will not improve with
higher FSI, unless there are concomitant
requirements for affordable housing and
impact-fees on new development.
Bertaud and Brueckner (2004) present
a “theoretical analysis” with “illustrative
calculations” for Bangalore, while recog-
nising the limits to infrastructure supply
in the Indian context. The authors argue
that as cities expand due to lower FSI,
people end up paying higher commuting
costs and housing prices. Bertaud and
Bruckner (2004) do not use more sophi-
sticated conceptualisations of the city
with multiple sub-centres and movement
of not just housing, but also jobs to
peripheries, but they acknowledge this
limitation in their conclusions. They state
that using Bangalore as a fi t for illustra-
tive purposes makes sense if the theory
of their proposed monocentric city
model is considered. However, they use
para meter values from cities in the
United States (US) for the cost calcula-
tions. Given that the level of infrastruc-
ture supply and institutional capacities
are radically different in the US cities,
their results should be viewed as an
excellent academic endeavour, with
limited direct polic y applications. Further,
their “illustrative welfare-cost calculation”
does not account for (1) societal gains
july 20, 2013 vol xlviii no 29 EPW Econom ic & Political Weekly
from reductions in emissions of green-
house gases and particulate matter from
shorter trips; (2) gains from living in
formal and better housing in the peri-
pheries, especially for the emergent mid-
dle class; and (3) benefi ts from co-location
in urban centres with amenities such as
better educational centres and healthcare,
restaurants, public gardens and so on.
Brueckner and Sridhar (2012) in an
extension of this piece, It follows the
same theoretical trajectory, and there-
fore builds on its strengths, but inherits
its limitations. In both these peer- revie-
wed journal articles, there are ample
qualifi cations that highlight the limits of
the models. Yet, based on this limited
model, Brueckner and Sridhar claim that
“ showing that more-compact cities
closer to international FAR norms are
better for consumers, saving them mon-
ey on housing and commuting costs, this
paper offers an important addition to In-
dian policy debates”. This extension of
theory to applied policy has limits that
policymakers should recognise.
Reframing Planning Strategy
Given the evidence on increasing FSI in
Indian cities, this section examines how
planning in Mumbai can be reframed as
a regional planning proposition. Patel
(2013) informs us that higher building
densities are being proposed for the sub-
urbs of Mumbai. It would be critical to
understand the land use and zoning
plans proposed along with increasing
FSI. Further, he tells us that higher FSI is
being recommended by institutions such
as the World Bank, based on expert
advice. I am not priv y to the World Bank ’s
cou nsel to Indian plan ners. I f i ndeed t he
advice is towards higher FSI, without
evaluating the associated reasons for and
impacts from constrained infrastructure
production, it is extrapolated from limited
evidence, and thus a risky proposition.
My research shows how jobs and
housing are spatially distributed in mul-
tiple sub-centres in the Greater Mumbai
Region (Shirgaokar 2012). In general,
rapidly growing Indian cities are not just
CDBs with decreasing densities towards
the periphery, but also include secondary
nodal developments that create specifi c
changes to the density gradients. There
are planned sub-centres within the mu-
nicipal corporation limits of Greater
Mumbai, such as the Bandra-Kurla Com-
plex, and outside, such as Navi Mumbai.
There are also market-driven agglomer-
ations that have both housing and jobs,
such as Vasai-Virar, Ka lyan-U lhasnaga r,
and Bhiwandi.
I contend that plans for higher densi-
ties should include urban peripheral
nodes, not just urban cores and CBDs.
The potential for managing land uses
and supplying infrastructure is higher in
emerging urban nodes. Such managed
and phased growth, that also encourag-
es a balance between housing and jobs,
can help establish multidirectional fl ows
on expensive transportation infrastruc-
ture (Cervero 1998). Patel claims that
most of the new urban transportation
infrastructure is concentrated within
the municipal limits of Mumbai, and this
does not open up new lands for develop-
ment. I agree that the supply of trans-
portation infrastructure should be a
regional proposition, with a special focus
on growing agglomerations – sometimes
outside municipal boundaries. The supply
of such peri-urban infrastructure should
be taken up in the early phase of a
region’s development, not later after the
die has been cast.
FSI is only one of the tools in a planning
kit that has zoning, growth boundaries,
inclusionary housing, and others. Increas-
ing FSI (say 2 to 4) can work if it is linked
to ratios of the mixes of uses – housing,
offi ce, retail, recreational and so on. A
location with FSI = 4 with only residential
properties may have different crowding
than if the same location is planned to
be a mixed-use development with resi-
dential, offi ce, retail and recreational
space. In a city-region with multiple
dense mixed-use sub-centres, the fl ows
and demands on transportation infra-
structure could make for a lively and
workable city- region.
Readers should be mindful that Patel’s
FSI argument is based on evidence from
island Mumbai, which is a special case.
Bangalore, which tends to be closer in
form to many cities in contemporary
India, has been studied in the literature.
Though it may not be similar in size
or economic potential and growth rates
to other cities, it is similar in the spread
of urban edges, movement of jobs and
housing to secondary nodes, growth of
vehicles, and decrease in standard of
living. However, neither Mumbai nor
Bangalore are similar to other cities
that may have a varying supply of infra-
structure coupled with limited planning
capacity, and different rates of popula-
tion growth, employment mix, and
housing supply.
Unfortunately, with limited data comes
limited understanding of problems. For
example, Brueckner and Sridhar (2012)
rely on aggregate (city/regional level) data
to model the effects of higher FSI. In India,
coming by data is the bigger challenge
compared to r unning the mo dels, and most
rese arch i s wea k due to such data limita-
tions. With Patel’s analysis, crowding is
August 27, 2011
Experimental Economics: A Survey
Sujoy Chakravarty, Daniel Friedman, Gautam Gupta, Neeraj Hatekar, Santanu Mitra, Shyam Sunder
Over the past few decades, experimental methods have given economists access to new sources
of data and enlarged the set of economic propositions that can be validated. This field has
grown exponentially in the past few decades, but is still relatively new to the average Indian
academic. The objective of this survey is to familiarise the Indian audience with some aspects
of experimental economics.
For copies write to:
Circulation Manager,
Economic and Political Weekly,
320-321, A to Z Industrial Estate, Ganpatrao Kadam Marg, Lower Parel, Mumbai 400 013.
Economic & Political Weekly EPW july 20, 2013 vol xlviii no 29 125
an interesting metric to measure effects.
However, better metrics such as square
feet of retail or of ce space per square
kilometre or housing units per square
kilometre would make for richer policy
analyses. Yet, these metrics are diffi cult
to construct because of data limitations.
Patel’s policy arguments for phasing
away the Rent (Control) Act, the need
for inclusionary housing, and the limits
to providing free housing are convinc-
ing, and supported in the wider scholar-
ly literature. Where relevant, these ideas
need to be included in planning and
land development policies. Further,
higher FSI should be linked not just with
minimum quotas for inclusionary hous-
ing per total built area, but also to exac-
tions and impact fees per unit of build
space. Such exactions could ideally
serve social equity questions with re-
gard to housing for all income classes,
and provide city coffers with the re-
sources needed to build infrastructure.
Finally, there is the issue of limited
institutional (and technical) capacity that
haunts planning in emerging economies.
Policy measures such as impact fees for
new development and inclusionary hous-
ing are great in theory, but in practice
can institutions evolve to handle these
new developments? Therefore, I contend
that the development focus should
equally be on improving institutional ca-
pacity, along with propositions such as
managing FSI.
Bertaud, Alain (2004): “Mumbai FSI Conundrum:
The Perfect Storm: T he Four Factors Re strict-
ing the Construc tion of New Floor Space in
Mumbai”, 15 July,
vi ewed on 1 5 Apr il 2013.
Bertaud, Alain, and Bruec kner, Jan (2004): “Ana-
lysing Building-Height Restrictions: Predicted
Impacts and Welfare Costs”, Regional Science
and Urban Economics, 35 (2), p p 109- 25.
Bru eck ne r, Jan, and Sr idh ar, K al a Seeth ara m
(2012): “Measuring Welfare Gains from Relax-
ation of Land-use Restriction s: The Case of
India’s Building-height Limits”, Regional Sci-
ence and Urban Economics, 42 (6), pp 106 1-67.
Cervero, Robert (1998): The Transit Metropoli s: A
Global Inquiry (Washington D C: Island Press).
Patel, Shirish (2013): “Life bet ween Buildings: The
Use and Abuse of FSI”, Economic & Political
Weekly, Vol 48, No 6: pp 68-74.
Shirgaokar, Manish (2012): “The Rapid Rise of
Middle-Class Vehicle Ownership in Mumbai”,
PhD Thesis, University of Cal ifornia, Berkeley,
available at
This paper proposes a new method for urban population density prediction at 30 m resolution. Using data for Bangalore, the paper demonstrates that population within each 30 m residential built-up cell can be modeled as a function of cell-level data on street density and building heights and ward-level data on car ownership. Building-height data were generated from Cartosat-1 stereo imagery using an open-source satellite stereo image processing software. Using this building-height data in conjunction with the other datasets, the paper demonstrates that a 30 m resolution population density surface can be generated such that, when summed to the ward level, the median absolute percentage error between predicted population and known census population at the ward level is 8.29%. The paper also shows that the relationship between population density, street density, building height, and ward level car ownership is spatially non-stationary. A fine-grained understanding of urban population densities, as enabled by the proposed method, can be beneficial to research, policy, and practice related to cities.
Full-text available
FSI (Floor space index-the ratio of the built-up space on a plot to the area of the plot) is a regulation followed in the development control norms of many cities. Cities are known for their complexity and dynamic growth patterns. Often cities have differential growth patterns, density patterns and differential levels of infrastructure. In many Indian cities, the FSI distribution is independent of land availability and densities. In recent news, it was revealed that 33% of Indians occupy living space that is less than that of the prisoners of United States. This was found from the recent National Sample Survey. FSI values of Indian cities are low compared to other cities in the world, which keeps the per capita built space low. The paper analyses and compares the various factors considered by some of the Indian cities in their building regulations with respect to the FSI for residential and related categories. A study of Vijayawada city is undertaken to analyze the impact of plot regulations and to find out various parameters affecting the same. Further the study can carry forward to examine selected planned and unplanned areas with respect to FSI, density and infrastructure.
Full-text available
The World Bank has been complaining that Indian cities are not making optimal use of their land and has been pressing for upward revisions of the floor space index, particularly in Mumbai and Bangalore. However, the bald comparison of FSI across cities that the WB presents is seriously misleading. This paper proposes a new metric, crowding, defined as the number of persons per hectare for a particular urban use. Thus we have indoor crowding, park crowding, and amenity crowding. How the new metric of crowding might be used in planning or replanning urban areas in general and the major policy changes that need to be made if Mumbai's housing shortage is to be seriously addressed is also discussed.
Full-text available
Households in Mumbai consume an average of 2.9 square meter of floor space per person 1 . This is one of the lowest residential floor area per person in the world. Over the last years, however, Mumbai has emerged as an economic regional powerhouse with a sophisticated workforce and a large middle class. Why should an affluent city have one of the worst housing in the world? Four factors are responsible for Mumbaikars' very low floor space consumption – difficult topography, poor land use legislation, muddled property rights and deficient infrastructure. However, the factor with the most negative impact is the draconian reduction of FSI imposed by ill conceived land use legislation. The average FSI imposed on Mumbai's residents is also the lowest in the world for a city of this size. Criticism of Mumbai's FSI practice is certainly not new, but amending the FSI legislation has proven to be the most controversial issue over the last few years. The argument for increasing the FSI is nevertheless quite simple: most of the land accessible from the city center in less than 1 hour is already occupied and is limited by topography; therefore there are only two possible ways to increase the average floor space per person: either by increasing the FSI, – i.e. building more floor space in already developed land – or by forcing a large number of households out of Mumbai to live in one of the faraway suburban municipality or somewhere else in India. For example, one way of doubling the consumption of floor space per person to 5 m2 would require, on average, a doubling of the FSI. The only other way of increasing the current floor consumption to 5 m2 would be to relocate about 6 million people out of the municipal area 2 and to reallocate the floor space thus vacated to the remaining population. Clearly, forcibly evicting millions of people is not an acceptable alternative in a democratic country. The choice therefore is not whether the FSI should be brought in line with other large cities of the world, but how much and where should the FSI be increased and what other measures should be taken to support this increase. 1 Extrapolated from Mumbai Census 1991 based on rooms per household. This figure seems to me terribly low as related to international experience. Let us assume that the 1991 Census underestimated the average floor space by say 50% and the real floor space was about 4.5 m2 per person, this figure will still be extremely low and the argument developed in this paper would remain valid. The 2001 census housing data are not yet available. 2 The population within Mumbai municipal border is evaluated in 2004 at around 12 millions people.
This paper estimates the effect of building-height limits on the spatial sizes of Indian cities. Regression results show that height limits, which are imposed in draconian fashion in India, cause spatial expansion of its cities, as predicted by the theoretical model of Bertaud and Brueckner (2005). The regression coefficients, by yielding the implied reduction in the area of an average city from a marginal increase in its height limit, allow computation of the annual saving in commuting cost for the city's edge household when the limit is relaxed. This cost saving, which is an exact measure of the common welfare gain for each urban household, can be scaled up to yield the aggregate consumer gain in a typical city from relaxation of India's restrictive height limits. For a moderate height-limit increase, this gain equals 106 million rupees.
  • Alain Bertaud
  • Jan Brueckner
Bertaud, Alain, and Brueckner, Jan (2004): "Analysing Building-Height Restrictions: Predicted Impacts and Welfare Costs", Regional Science and Urban Economics, 35 (2), pp 109-25.
The Rapid Rise of Middle-Class Vehicle Ownership in Mumbai
  • Manish Shirgaokar
Shirgaokar, Manish (2012): "The Rapid Rise of Middle-Class Vehicle Ownership in Mumbai", PhD Thesis, University of California, Berkeley, available at