Syndicate content

Success when we deemed it failure? Revisiting sites and services 20 years later

Sumila Gulyani's picture
Between 1977 and 1997, the World Bank supported “sites and services” projects in 27 cities across India
A freshly-minted architect stood staring at a sea of toilets. Row after row of them, on small “housing plots” meant for low-income families who would build their house incrementally as their incomes and savings grew. The neighborhood was “planned” and provided with services—under a World Bank-supported “sites and services” project—to serve as the anti-thesis of and an antidote to the slums that were, at the time, increasingly becoming the only housing option for low-income families.

It was 1980 and the architect, Barjor Mehta, was deeply disappointed. There were no houses, no people and no chance that they would ever come, given the seemingly god-forsaken location—in an area called Arrumbakkam—so far from the city center in Madras (now Chennai). Having just completed his thesis on housing, he wrote a scathing news article in the Times of India denouncing the sites and services approach. Barjor wasn’t alone in his critique, and by the mid-1990s the World Bank had almost entirely abandoned such projects.

In October 2015, Barjor, now Lead Urban Specialist at the Bank, invited me to revisit Arumbakkam and other neighborhoods developed, between 1977 and 1997, under four Bank-supported sites and services projects: With my colleagues Kate Owens and Andrea Rizvi, I visited 15 of the 28 sites developed in Chennai and Mumbai. We also reviewed archival material, analyzed satellite images, and recently presented our preliminary findings (download a peer-reviewed journal article here). Now, Barjor and I agree that previous assessments of failure may have been both premature and erroneous.


Twenty years after completing our last sites and services projects, we found bustling and thriving neighborhoods in all but one of the 15 sites we visited. The neighborhoods are almost fully built out and built up—not only did people come, but they also invested heavily. There are houses on almost all plots; less than 10% of plots are vacant. People have invested to add space, upgrade amenities, and improve construction materials, quality, and appearance. Although there are still a few small single-room and single-story units, most of the houses are 2-3 levels. Families have added bathrooms and kitchens on each floor. They have replaced tin roofs with tiles or concrete, strengthened foundations and the superstructure, and upgraded walls and facades. The idea of “incremental” housing—where people would invest slowly, over time, at a pace that fitted each individual family’s circumstances—has worked.
Incremental Housing in Mumbai: evolution from one-room units with tin roofs (right) to multi-story houses (left)
Four key technical features—two of them innovations at the time—worked exceptionally well.
  1. A key innovation was the introduction of plots that were tiny compared to those “standard” at the time. The smallest plot was 33m2 in Chennai and 21m2 in Mumbai, as compared to minimum plots of about 150-200 m2 in other housing developments in these cities. The small plots were far more affordable and have indeed allowed lower-income households to enter the housing market.

  2. Another innovation was to use spatially-efficient site planning norms that helped lower the unit costs of developed plots while further increasing density. For example, only 34% of land was allocated to streets and open spaces, compared to 50-60% frequently seen in other developments in India at the time. Even so, average road density in these neighborhoods is actually higher than that in their parent city as a whole. Smart planning, thus, lowered the cost of infrastructure provision and individual housing plots, while creating compact, walkable and livable neighborhoods. 

  3. A key design feature was inclusion of a range of plot sizes that would attract different income groups. In Chennai the plot sizes ranged from 33m2 to 223 m2, and in Mumbai from 21m2 to 100 m2. Now, the neighborhoods are indeed mixed-income, with lower-income families occupying smaller plots, and middle-income and high-income families occupying larger plots.

  4. Finally, the design explicitly aimed for mixed use by including commercial areas (shops), amenities (schools, clinics), and, in some cases, plots for light industry. As per plan, the current neighborhoods actually do have all of these types of businesses, services, and amenities. Mixed-use has also translated into vibrant streets.
The third and fourth features—resulting in mixed-income and mixed use communities—are noteworthy because they are used far too rarely. Instead, many government-sponsored affordable housing programs do just the opposite—they explicitly target a narrow band of low-income families, working hard to exclude those with more means. And they focus primarily on delivery of housing units, rather than invest in mixed-use development.

In Mumbai, the sites and services neighborhood of Charkop (left) is characterized by a dense, well-planned, and spatially-economical layout. The private development on a neighboring parcel (right) has fewer roads, larger buildings, and planning that is less efficient.
The experience in Chennai and Mumbai suggests that we have, in our hands, a tool for better managing urban expansion and creating affordable housing. First, city governments can use the sites and services approach and planning norms to shape future urban growth. They can move beyond “putting stakes in the ground” and use it to earmark future neighborhoods.
Second, both governments and private firms can create more affordable housing by scaling up delivery of small housing plots where families can build incrementally. This represents a housing solution that, in terms of cost, lies in between the two classic options—in-situ slum upgrading and public housing—but, potentially, offers quality and livability that is superior to both.

In Chennai and Mumbai many of the families residing in sites and services neighborhoods feel they hit the housing jackpot. Governments should strive to make this option available to many more, not just a lucky few. And in the process, they can build new neighborhoods and cities that are more compact, more inclusive, more vibrant, and more livable.

We see this work as the start of an important conversation and a call for more in-depth research. Do you know of cases that confirm or challenge the findings above? Please share your comments with us.

(Update March 2017) Following the publication of this blog post, we prepared a paper elaborating on these findings. We invite you to download the draft paper and provide your feedback or comments below.

(Update March 2018) Now you can read the peer-reviewed paper! Download it here for free until April 22, 2018. 


Submitted by Genevieve Connors on

So refreshing to see a twist in the tale - failure turned success. We need more historical reviews (time series!) on such interventions that really do transform lives but take a generation. And thank goodness for the foresight of some high density, mixed use, small plot and PLANNED urbanization in India. If you do a future series, could you tell us more about the verdict of S&S in Latin America? Where it all started I believe.

The twists can go both ways – some initiatives that seem successful early on can fail over time and vice versa.  Either way we learn.  So, we really do need to look at results and impact over time. We are hoping that project teams and researchers in other countries and regions will identify older projects that are worthy of a second look and help us learn more about what has worked over time.  Thank you for your comments!

Submitted by MWACHARO L.G. on

As an urban designer I agree 100% with your findings. There was a similar successful scheme in Miritini, Mombasa City, Kenya. Today it is a built up vibrant , mixeduse community where more than 20 years there was only empty serviced land. It was an organic growth that was homegrown and driven by both local and wide consumer demand. Sadly World Bank seems to have forgotten all about site ad services and is now preaching affordable homes as engines for economic growth. Any affordable housing business model should include site and services as one of its tactics of delivery.

Submitted by bernard kilroy on

ANY < GENTRIFICATION > RISK IN FUTURE ? What a relief to see such a past success! Let's hope there will be FUTURE MONITORING too and maybe some POLICY INTERVENTION to ensure that the socio-economic mix stays - so often such neighbourhoods can over time become completely < gentrified > as smart owners or developers see a way to develop plots. Amazing how much gentrification has happened in London to housing originally built by public subsidy or charitable trusts which has now passed to well-heeled buyers who have capital to spend.

Submitted by Slobodan Mitric on

It would be very nice if the author of this blog were to cite the names of the people who pioneered the sites-and-services concept in the Bank and put together the early projects, in Madras/Chennai and elsewhere. Also, it would be useful to cite in-house papers evaluating the sites-and-services experience, in addition to Robert Buckley's book on 30 years of shelter-related activities in the Bank.

References will be included in our paper that will be available shortly.  Regarding the names of the pioneers -- even the early projects were prepared by different teams; the teams changed over time and the project designs also evolved.  Names of team members are available in project documents and links to these are embedded in the blog.  Many thanks for your comments!

Submitted by Matt Nohn on

Dear Sumila,
Thanks for your winderful post. This is Matt Nohn whi has led the concept for the Bank's new core housing program in Indonesia, collaboratively with many others. I am in Chennai at present and would love to visit some of the sites. Is your study available by now or are you able to provide other information on the findings and site locations? The link in the blog does not work...
Thank you,

PS: for everyone following the incremental housing idea, this training may be interesting:

Submitted by Jacqueline Klopp on

Very very interesting blog post. It shows that we need to plan longterm and our horizons need to be longer when we think of urban development "projects". Also, we need to allow incrementalism on housing improvements vs immediate upgrades that always cause downward raiding by middle classes and displacement. It would be interesting to know how many of the inhabitants are the original ones. Look forward to reading more about this case and its lessons. We certainly need some new thinking in this space!

Yes, additional research is needed to find out more about the current residents, figure out how many are original residents, and how their living conditions and lives have changed, over time.  Hopefully, some researchers will take up these questions?  We will also try to follow up with more in-depth studies.  Thanks for your comments. 

Submitted by Gabriella Carolini on

A striking reminder of how orthodox evaluations, which demand rather short-term assessments, can lead to the premature dismissal of promising investments. Planning lacks the luxury of fault-free experimentation in designing urban interventions, and perhaps this is why our evaluations lean toward the conservative notion of short-term measurable results. Thank you for sharing this important story that captures so beautifully why a little more heterodoxy is required in evaluation rationales! It is quite rare yet so critical for development organizations like the Bank to make time for and share real reflections on processes of evaluation. This is a very useful lesson about balancing urgency with patience in urban development.

Submitted by Maniza on

Excellent blog--many lessons including for how and when we measure impact. These type of results probably hold true for some other sectors as well. I think this approach of sites and services for new urban development would be highly feasible in many contexts including for places such as Dadaab and also in other places such as Somalia grappling with returns and IDP issues; and in cities which have grown far beyond their original city centers and are facing both services and housing shortages. I'm sharing this blog.

Supporting displaced communities and those in refugee camps by adopting a sites and services approach?  Thank you for that important suggestion.  Yes, certain elements of the sites and services approach can potentially be modified and used in such situations.

Submitted by Joachim Boko on

Very interesting blog post, indeed.
I think there are a couple of questions that need to be brought up to better explain the success of this experience in India. One of the most important in my opinion is that of the status of the land at the beginning. Did it belong to the city government/state or to individual. I suspect they belonged to the city government which makes things easier. In cases where land belongs to individuals at the beginning implementing this approach will probably be very challenging...
Thanks again for this post!

Land was assembled by the govt for the sites. Your instinct is right, though.  Assembly of land often took a long time and this was one of the key reasons why project implementation periods were long.  It is worth noting that even today many state/local govts in India continue to assemble land and allocate it for different uses, such as infrastructure development and housing for low-income families.  Several questions remain, however, regarding the most effective ways of delivering land or housing -- at scale -- in a manner that can also reach lower-income families. This blog and the (forthcoming) paper on which it is based are one contribution to that discussion. Many thanks for your excellent comments! 

Submitted by Robert Merrill on

As one of the pioneers of Sites and Services (S/S as we called it in the '70s), I was extremely gratified to come across your blog. Having done my dissertation on Chile's national S/S program (Operation Sitio) in the '60s, and subsequently designed and implemented a national WB S/S cum upgrading progam in Tanzania in the '70s, while others such as Richard Martin did so in Zambia and Nairobi, I was happy to see the concept persist in Indonesia (KIP-Kampung Improvement). However, attempts at S/S were overcome in the Philippines due largely to politics and private land ownership in the Marcos era. Along with other pioneers (Jerry Erbach, Roy Brockman (ADB), and others, we're still trying to keep it alive through ger (read yurts) area upgrading in Mongolia...even in -40C weather! FYI, in addition to Bob Buckley, there's some of us still around: Alberto Harth (El Salvador), Alain Bertaud (Bertaud Model), etc. etc....just check out some of the old WB S/S Appraisal documents.
Thanks so much for documenting the work in India!

Fascinating that Mongolia is using it now.  Many thanks also for sharing project locations and the names of your pioneering colleagues on this blog! We know several, but not all, of the people named.  It would be interesting to have a discussion with all of you, regarding your thoughts on the potential of this approach for managing the forthcoming urban propulation growth. 

Submitted by antonio risianto on

We like to restart the Community Based Site and Service program in Indonesia. Anyone knows whether this is still viable and valid? Especially as the Private Developers are so active in the urban fringes . ... and so Gentrification problems.

We wonder how we can compete with the land speculation that is so rampant . How to control this? Policies to enable existing communities to join this Site and Service program thru Land COnsolidation mechanisms , etc. Mixing with rental and partial sales Or even start this as a national housing policy that will not stir political waves ( as the power structure are often Real Estate private Developers who might look at these scheme as a treat for their business? ).

Please comment to Antonio Risianto [email protected] (UC Berkeley / MIT). Practitioner .

Submitted by Sivashanmugam,M on

We fully agree with the findings and views of the Authors of the story.
I work in CMDA and part of Project Managment Group under TNUDP, that covered 10 towns in Tamilnadu including Chennai where sites and services and slum improvement projects were implemented.
Arukambakkam probably most researchers and Bank people cited. Thre are many other sites and services scheme, examples of dynamic and diverse community, compact and mixed developmet, one of the best example of affordable housing thro cross subsidy. More than a lakh plots have been added to the housing stock through these schemes. The percentage of people living in slum is comparitively loe in Chennai when compared to similar cities because of these schemes. It is highly unfortunate that thsee schemes are not replicated or continued.
once again thanks to the Authors

100,000 plots have been added through such programs.  Thank you for sharing that!  It would be excellent to document the broader experience in Tamil Nadu, and assess what happened in cities of different sizes.  But, more importantly, it is terrific that CMDA continued to use this approach -- thereby, significantly broadening the impact of the idea.   

Submitted by Ellen Bassett on

Just to underline some of the comments above, it would be great to extend the analysis to Sites and Services elsewhere, including Africa. I remember being amazed at seeing houses in Dandora in Nairobi advertised for significant prices in the late 1990s; I wonder how those developments have aged and valued over time.

It is also interesting to see the convergence on urban design (small parcels, incremental building, smaller infill units, high walkability)--you could be writing about Portland, Oregon where I am currently sitting. We have an affordable housing crisis--maybe we need sites and services!

Dandora would indeed make for an excellent case study.  The relevance of some of the design strategies for Portland, Oregon, and other US cities is worth exploring.  The work on affordable housing and urban design in developing countries is leveraged far too infrequently to inform strategies in US.  Many thanks for your excellent comments!

Submitted by Meenu Tewari on

Fascinating study!
It reminds me of the power of delving into the "banished histories" of projects that Judith Tendler used to talk about to draw lessons about how something actually worked out and unearth the often hidden processes that contributed to success over time even in the face of apparent failure originally. I was most struck however by the third and fourth points you made -- about mixed use outcomes and the presence of land for light manufacturing/industry and services -- something that is missing in many current projects -- and yet is so crucial to providing jobs in proximity to where people live. I was curious whether this was envisaged in the original design of the settlements or was it something that emerged over time as the inhabitants lived their lives (as happens in so many informal settlements). Did the city then zone these in as mixed use areas? Am curious how the state - or at least the officials on the ground handled this, and whether there was contestation or 'mutual adjustment' over time. An ethnographic look into that process around land use, as well as how things turned out for the original owners or others as plots changed hands (or not) would be fascinating.

Also, the vibrancy that comes through so clearly in your narrative stands is sharp contrast to other housing developments and relocations in Chennai itself that have not (yet) turned out so well. Kannagi Nagar is one such example of a place that is struggling despite banks of brand new apartments (and relatively low rent) that residents can't want to get out of. When I visited that neighborhood a few years ago it was miles away from jobs, schools, affordable (public) hospitals or other basic amenities including reliable transportation. A sea of multistoried homes surrounded by calf-deep water and slush that the residents said were fonts of "nothing but disease."

So this model that you describe and the lessons that you draw from it are especially powerful today for what to do make current investments more viable. Thank you for this wonderful research!

Plots for light industry and services were part of the original designs.  The comparison with public housing in Kannagi Nagar is very interesting.  Time matters, though.  We would need to compare public housing projects completed around the same time as the neighborhoods under the sites and services projects; such research could provide important insights for housing policy and programs.  Many thanks for your insightful and helpful comments!

Submitted by Kristin Little on

Exciting study and findings! Thank you for opening up a new conversation on this topic. I agree with so many of the points already made in the comments, and find your blog post an especially important reminder that change takes time. This is something we hear a lot, but it is rare that we actually circle back to what might have become of long abandoned approaches. Thanks for doing this research, and creating this space for discussion!

Opening a new conversation on an abandoned approach – it would be terrific if this research can actually help do that.  Many thanks for reading and participating in this discussion!   

Submitted by Vidyadhar Phatak on

It is indeed important that a lively debate has been initiated on a significant approach to affordable housing. The reasons for discontinuous of S & S in Mumbai besides loss of Bank support were,
Non availability of public land - in the Bank project Govt. owned wetlands were used. From 1991 such wetlands were not available for development under environmental regulations. Acquiring private lands through eminent domain became impossible.
The visuall impression of S & S gave an impression of underutilisation of scarce land in Mumbai amongst politicians and public officials. This led to lack of any sustained efforts of continuing the approach.
The challenge therefore is to increase the supply of accessible land and promote S & S type development through public private collaboration.

Perception of underutilization of land – that is a very important point.  Our analysis show that, over time, the density achieved at these sites is higher than in other developments.  Underutilization is, therefore, a misperception -- but it is one that is not easy to counter.  We fully agree that access to land is a key issue here and that sites and services types of neighborhoods could be developed through private-public collaboration.  Many thanks for your insightful comments and for sharing your experience, as manager for the Mumbai project, with us!

Submitted by Alain Bertaud on

For us who worked on these projects from 1977 onward, the main objective was to demonstrate that formal housing projects could be affordable to low income households (our target was between the 20th and 35 percentile) if urban minimum regulatory standards were revised. Our hope was that minimum housing standards would be revised after it was shown that they did not result in slums. Everybody agreed at the time that the end result was desirable but the minimum housing standards were never revised, therefore preventing private developers to build sites and service projects. Only the government was allowed to build under these standards. The failure was not the projects themselves, but the inability to obtain a revision of the absurd minimum housing standards.

Some of these ideas were, perhaps, ahead of their time.  At the very least more time was needed for governments to accept them more broadly (beyond the project) and for the private sector to adopt them.  First, we seem to have set a very high bar for “success” – hoping for a complete revision of major city-wide urban planning guidelines and standards.  Second, we expected it to happen in too short a period – over the life of one or two projects in a given city. 
Looking forward, given the projected arrival of an additional 2.5 billion urban residents by 2050, we have the following question: is now a better time to push for reduction of minimum regulatory standards?  We would very much appreciate your thoughts on this question, Alain.  Many thanks for sharing your insights with us and we look forward to continuing this discussion with you! 

Submitted by Judy Jia on

Exciting findings after 30+years! The original planning and design principles have shown desirable results over the time. Early site and service projects in Latin America has now become vibrant communities, too. Incremental housing, as its name indicates, takes time to mature. I am now working in China. The Chinese government has ambitious goal of upgrading all the substandard urban housing by 2020. The promising site and service model has not been applied in cities in China. The model poses challenges in such context: 1) land in cities cannot be owned by individuals; 2) the project may not become a nice neighborhood within 5 years of a mayor's term; and 3) no one wants to see a matrix of toilets.

Vibrant communities in Latin American sites and services neighborhoods – thank you for sharing that with us.  It would be good to record cases in Latin America.  A team of MIT students is doing that in Quito, but more case studies would help.  The latter two points that you raise about why it may not work in China are, probably, valid in other countries as well.  An approach that creates neighborhoods that looks ramshackle for quite a period of time is, indeed, not likely to get support easily from politicians and better-off urban residents.  But the argument that could be made is that: (a) many these neighborhoods do consolidate over time; and (b) the alternatives – e.g. slums or limited public housing (that is often poor quality and becomes segregated and/or ghettoized) – are often much worse.  Thanks for your excellent comments!

Submitted by Aparna Zaveri on

What a fantastic post! and the comments make it all the more relevant.
Incrementalism has always been the 'way of the humans' their settlements as well as in other aspects of their life. Strategies like sites and services take time not just because populating an area is a complex process but also because eventually, a city will naturally expand into those far flung areas where these projects are. For those who might be interested in incrementality, a strategy of the Karnataka Housing Board in India also provides a great case study. They provide sites of different sizes based on a lottery system for residents who have lived in the area for more than 10 years. These projects are increasingly becoming rare in the bigger metro areas but in their heyday were immensely successful in bringing together people from all income groups with a fantastic sense of community and belonging.
The IFC recently launched a good practice tool to enable such inclusive communities to be built, currently aimed at private developers, working in affordable housing. It is called the Affordable and Socially Sustainable Housing Application (ASHA).
Thank you again for bringing to the forefront a strategy that has worked to increase the stock of desperately needed decent housing at the same time serving to build community.

Building inclusive communities is indeed the goal.  Thank you for sharing information with our readers on how IFC is working with private developers on this agenda (via ASHA).  Case studies of earlier work by agencies such as the Karnataka Housing Board would help build the evidence base on features that contribute to this seemingly elusive goal of building inclusive communities.  Thank you for liking the blog and contributing to the discussion!

Submitted by Jasmine Saluja on

Reviving S&S.
We are working on our research to reviveS&S and bring back more pragmatic combinations of the same in the peri urban areas of Mumbai.Your perspective on this subject reassures our hypothesis,I am much keen at sharing our proposal and learning more about your research.The major challenges faced today are the financial models, as the government has taken the role of a facilitator of benefits. Second being poor public transport,which limits the possibility of opening new land parcels within developable boundaries.Thus,one of the added services becomes last mile access to a public transport system.

Your article puts forth relevant arguments to revisit S&S and re-evaluate this model.
Are there more publications which reassess S&S,stating it as a successful model - a departure from its earlier notion of it being a failure?

Transport connectivity and financing were central to the design of the earlier projects in India. For example, they did invest in last mile connectivity.  We now have a more detailed paper that is ready and will send it to you via email.  That paper also has references to other studies on the topics. Hope it helps. Best wishes for your important endeavor! 

Draft paper is avilable. Following the publication of this blog post, we prepared a paper elaborating on these findings. We invite you to download the draft paper (please see link in the blog)and provide your feedback or comments below.

Submitted by Aditya Sawant on

Very interesting study! I am an urban planner and live in Charkop. I was wondering if you could guide me as to where can I get some archival reports which deemed the site and services scheme of Charkop as a failure. Many thanks.

Charkop was not tagged as a failure. In fact, the broader project -- called Bombay Urban Development Project -- included several sites in addition to Charkop, and it was rated "moderately satisfactory" overall, by the Bank, when it closed in 1994. However, there was no follow on project in Mumbai after the project closed.  And the Bank mostly stopped financing new sites and services projects. Please do share your views as a planner residing in Charkop.  What do you see as the positive and negative planning features of your neighborhood?   

Submitted by Vitti Valenzuela on

Hi. I am trying to download the article using the link but it seems not 'free' anymore. Is there another way to access the full article? Currently finishing up a PhD thesis on low-income housing in Naga City, Philippines and the study seems to be relevant literature.

Just a first thought. 'Success' would have to be checked though with the initial intent of sites-and-services projects. The projects showcased the 'low-income housing strategy' of the World Bank in the 1970s and thus was supposed to cater to the poor. SS was supposed to replace public housing which was the default response of states to the poor's demands for housing, other than slum clearance. On this aspect and the time-frame in which the World Bank was expecting benefits to materialize, SS were not.

But in another aspect, your study has a point - on the dynamism over neighborhoods over time. The poor do not stay poor forever, or they may have just been eased out through gentrification or for some other reason. As I have no access to the study, I could not comment further. It would be interesting to know whether the study included case studies of households that have been there from the start and how local politics and relationships shaped the way these SS are right now.

Link to World Development journal paper has been updated and is working.  Thank you for flagging that.  Your points are well taken and the paper addresses some of these.  Additional research, especially socio-economic surveys and longitudinal studies, is needed to help us better understand the dynamics over time.  Best wishes for your research!   

Add new comment