Archive | March, 2013

Toilets in Rural Karnataka – A peer effect?

In my previous post, I had taken a look at how sanitation improved in rural Karnataka and India over the decade of 2001 to 2011. Three broad categories of districts had emerged in Karnataka, and a relationship was visible between a district’s starting position on toilet ownership and how well it improved.

Going by the spirit that drove the previous analysis, I go a level deeper and more granular, and take a look at Karnataka’s taluks (or sub-district units) to see how things are changing in rural sanitation.

Karnataka Rural Sanitation - Improvement in Districts and Taluks

The above graph tries to see just that: where taluks and districts were in 2001, and how much they improved over 2001-2011. It is immediately apparent that taluks follow the districts of Karnataka in their behaviour: there again appears to be a strong link between where a taluk is starting from in terms of toilet ownership, and how much it has improved in the past 10 years.

If we simplify the first graphic by removing the district data points,  the positions of all 176 taluks of Karnataka are more clearly visible.

Karnataka Rural Sanitation - Sanitation Trap

What one can see is that when the starting point of a taluk is below 15% toilet ownership, the improvements are never phenomenal. When the starting point crosses about 20%, many more high performers become visible. It is possible that taluks and districts have to get out of a “poor sanitation trap” before being able to improve significantly.

Complementing the trend, every district that started with 40% toilet ownership or higher, improved by at least 20 percentage points, underscoring the relationship between the two. While the trends are easy enough to visualise and comprehend, the reasons for them may be complex and difficult to be certain of.

One reason for this link between starting position and improvement in sanitation could be the peer effect. The peer effect is where someone’s behaviour is influenced by those around them. It has been well studied in the field of education, where it’s been found that a student’s educational outcome is strongly linked to his or her peer group and the group of friends. ‘Peer pressure’  is a form of this as well, although mostly with negative connotations, where people pick up habits and mannerisms from their peers.

Coming to toilets in Karnataka, imagine the 15-20% mark: it’s where 1 in 6 or 1 in 5 houses have a toilet. At those numbers, most people have a neighbour who owns a toilet and uses it. People who are still defecating in the open can not only imagine, but also see what the benefits and comforts of owning a toilet can do. Possibly, even the transition to using toilets (and cleaning them!) can become easier as people can learn from each other. Toilets also possibly become aspirational objects – in effect nudging more households to avail what subsidies come their way.

An alternate reason could be a lot more mundane: the link could simply be a result of the Nirmal Gram Puraskar award and the programmatic design of the government’s total sanitation campaign. A large focus of the campaign was to take villages that were doing reasonably well on sanitation – and push them to near-complete toilet ownership, making them “Open Defecation Free” (at least in theory). Here, individual toilet subsidies were coupled with a cash award to villages (and their panchayats) which managed to go open-defecation-free. While it is possible for the NGP to have had an effect on the correlation, it is unlikely that it can explain it entirely.

One way to decide between the two (and other!) possible reasons for the pattern of rural sanitation improvement is to go deeper once more look at it at the habitation level – at villages and hamlets and how they improved. Unfortunately, this isn’t possible with Census data. Either way, these results have a significant bearing on how we can improve rural sanitation in the coming decade, where we have to achieve a high, sustained improvement in sanitation but with reasonable public investment.

Data used in this post are available here: Karnataka Districts | Karnataka Taluks.

Comments { 1 }

Sanitation in Rural India and Karnataka – How has the needle moved?

Sanitation is among the most dismal and depressing topics in India, across the country. While sanitation in our cities comes with its own set of problems, rural sanitation in India is stuck a primitive stage where too few people have access toilets.

To promote toilet construction in villages, a ‘Total Sanitation Campaign’ was launched by the Government of India in 1999 where subsidies were given to households to construct individual toilets. By means of the subsidy, cash incentives for village leaders and other communication campaigns, the state has tried to promote toilet construction and the need for adequate sanitation for over a decade now. As one can expect, several problems such as  inadequate subsidies, red tape, corruption, plus a lack of demand for good sanitation have all plagued the campaign.

After the Census results came out last summer, there was a major controversy as the census numbers for toilet ownership violently disagreed with numbers that the government campaign was putting out.

This post takes a look at how the needle has moved on toilets and sanitation – be it because of the government scheme or in spite of it  –  in rural India and in rural Karnataka.

As of 2001, only 21.9% of rural Indian households had toilets. After a decade, the percentage of households increased to 30.7% – an increase of less than 10 percentage points. Below is a graph of how toilet ownership has improved across all states of India, arranged in an ascending order based on how things were in 2011.

Sanitation in Rural India - The Transition State

While Kerala, Manipur, Mizoram and Sikkim are states that are ahead of the rest on rural toilet ownership, it is important to note that Himachal Pradesh, Punjab, Sikkim and Goa have improved the most in toilet ownership between 2001 and 2011. It is also interesting to note that relatively well-off states like Tamil Nadu and Karnataka actually fall below the national average, with Andhra Pradesh barely doing better.

The above performance, however, has been talked of quite a few times by journalists, policymakers and other sectoral experts. To really understand how rural sanitation is improving, one needs to dig deeper and go more granular. This blog makes a preliminary attempt at doing so by looking at all the districts of the southern state of Karnataka.

As shown in the graph above, toilet ownership in rural Karnataka increased from 17.4% in 2001 to 28.4% in 2011. The spread of this growth across districts can be seen in the graph below.

Sanitation in Rural Karnataka - The Transition State

At the district level, a lot of fascinating trends emerge. Firstly, there’s a curious change in the gross “shape” of the graph, compared to the first graph of states. There appear to be three distinct types of districts in 2011: the poor, the middling, and the stellar performers.

The poor performers are the bottom 10-15 districts, that had a low base of toilets to begin with, and improved by only a few percentage points in the last decade. The middling performers are those that had between 10 and 20% toilet ownership in 2001, and all improved by about 10 percentage points since then – similar to the state average. The third type are the stellar performers, who had more than 20% toilets to begin with, and improved significantly over the decade.

However, the most prolific district in Karnataka is undoubtedly Bangalore (Rural) making a phenomenal leap of almost 50 percentage points in toilet ownership. This performance is perhaps attributable, at least in part, to a very proactive civil servant, Manjula Naik, who was the CEO of the district Zilla Panchayat for a while.

One can also posit that how well a district improves is also incumbent on what its starting position is. It is possible that districts with about 20-30% rural toilet ownership have reached a certain stage of socio-economic development, where the prosperity, aspirations, governance quality and cluster effects of some households having toilets spurs the rest in building toilets. If that is so, then the middling districts of Karnataka – Hassan, Mysore, Mandya, Davangere, Ramanagara and Haveri are all ripe for rapid improvements in rural sanitation.

Let us hope that smarter policies and better economic growth result in far greater improvements in the coming decade.

Some of the ideas in this post came about due to several conversations with my colleague, Vijay Krishna

Notes. While this post tracks the percentage of households with individual toilets, a small portion of rural households also has access to community toilets. It is the remainder of households who continue to practice open defecation, along with a subset of the former households who may not be using their toilets.

Data used in this post are available here: States | Karnataka Districts.

Comments { 1 }

Visualising Karnataka’s municipal elections 2013

Urban Karnataka took to the polling booth earlier this month to vote for their corporators and municipal councillors. The elections were to be a four-way contest, with the Indian National Congress (INC), the Bharatiya Janata Party (BJP), Janata Dal (Secular) and Karnataka Janata Party (KJP) being the large political parties in the fray. Congress had a significant victory in the elections, securing 1960 of the 4976 seats contested: over 39% of the total. BJP and JDS were far behind, securing but 905 seats each, about 18% a piece of the total seats. The breakaway former Chief Minister BS Yeddyurappa’s new party the KJP could manage only 274 seats, less than 6% of the total. Independents won a significant portion of the seats as well.

The Karnataka municipal elections are happening at a rather critical time, with state assembly elections due in a few months and with the national parliamentary elections due in 2014. The results here will have a significant impact on both the impending elections. Twitter and other social media are already abuzz with speculations and assertions on what this would mean for the coming elections, on whether the municipal voters represent the entire state’s dispositions, on the signaling value of Congress’s early win, on anti-incumbency as a driver within the state, and much more. This article limits itself to a brief analysis of the municipal elections and their results, with an attempt to understand who will be governing much of urban Karnataka over the next five years.

Out of a total of 61 million residents of Karnataka, 208 cities and towns with a total of 13.9 million residents had local elections this month. At 22.8% of the state’s population, this covered all of urban Karnataka with the exception of Bangalore and a few minor towns. While most news stories are talking about the total number of corporator/councilor seats won by political parties, what matters is who has been able to win a majority of seats in individual local bodies such that they can form the local government there. Like in assembly and parliamentary elections, city mayors are elected indirectly by allowing elected municipal councilors to select their mayor.

The above graphic shows how many cities and towns of different sizes have been won by individual political parties. Overall, the Congress won 79 of the total 208 seats, with JDS and BJP with 25 and 24 each. The KJP managed only 5 seats, but it is interesting to note that there were 12 urban local bodies where all independents put together won a majority of the seats. Also, a good 30% of the local bodies had no clear victors. These 62 cities and towns will be where a lot of negotiations will take place over the next few weeks so that coalitions can be cobbled together to reach a simple majority.

Click on the map to know more about election results in individual towns.

A quick look at the above map tells us about the geographical spread of each party’s victories in Karnataka’s municipal elections. The congress (blue) appears to be well spread out throughout the state. The majority of JDS victories (green) are clustered in Southern Karnataka, conforming with their traditional strength in the Old Mysore region. The BJP (red) appears to have won towns in the eastern half of Karnataka – ranging from Belgaum and Bagalkote all the way down to Dakshina Kannada and Kodagu.

In all, the Congress has won a significant victory in Karnataka – be it in the number of corporator seats won, cities with majorities and their size. Below is a set of major cities and towns won by each of the political parties in the fray.

Notes. In this article, victors have been declared if they have won more than 45% of the seats in any given town or city. This is with the assumption that by taking on board one or two independents, the party can attain a simple majority. As of writing of this article, 24 seats out of the total 4976 are yet to be declared on the NIC website: most of which are seats from Terdal TMC, Bagalkot district. 

Data used in this post is available here.

You can also take a look at my fellow blogger Karthik Shashidhar‘s work for other visualisations of the election results: Congress Sweep |  Overall resultsDistrict-wise results | Party-wise performance on a district map | 

Comments { 4 }

In Pragati: From Open Data to a Culture of Openness

This week, I write in Pragati, the Indian National Interest Review about thinking beyond open data and creating a culture of openness in India:

Building sound public policies requires robust information systems and data. The state has the primary onus on provisioning for the public good that is knowledge – and traditionally the working of the Indian state involved secrecy as a core principle. The state enjoyed a monopoly on generation and access to large parts of knowledge and data relevant to public affairs, with its citizens often kept in the dark. Thanks to almost two decades of campaigns in the country, this paradigm has been overturned significantly.

When it comes to knowledge, information and data – the form of opening up determines how usable it is. Since the passing of the Right to Information Act into law, the focus of many advocacy efforts has shifted to usability and ease of access to government information and data. Release of information suo moto decreases the time cost of accessing information, and when it comes to data it is quite obvious that whether it is shared as a photocopy, PDF or an excel file makes a world of difference. Unfettered open access to data and information generated through taxpayer funds may remain a distant goal, but there are constant signs of progress. The Census of India is an exemplar of openness and friendliness as a data provider, and the Government Data Portal: was launched just last year is being populated with more datasets every month.

Increasing openness of official information is but one leg of improving the ‘public good’ quality of knowledge. Along with openness, the contestability of information has also increased in India over the last two decades. As Pratab Bhanu Mehta mentions, when official pollution numbers are not trusted, you have private organisations like the Centre for Science and Environment to monitor the pollution on their own and provide independent evidence.

Research institutions, not-for-profit organisations and the media are playing a larger role in generating data that is relevant to public affairs. Many private players routinely engage in data collection exercises, and conduct large surveys for research and to answer specific questions. Not only is this being done on a local scale, but nationally as well. Bangalore-based Public Affairs Centre conducted a national analysis of public services in India about a decade ago, and the ASER Centre provides state and district level information on children’s learning levels and education across the country, year after year. These are but two examples out of many.

While citizens of India are demanding more openness from their government and private entities are playing a greater role in contributing to public knowledge, openness as a culture has been far slower off the mark. Barring exceptions, research and not-for-profit organisations are far from open with their data in India today.

Data is collected with great care, cost and effort and is often used with great effect – but rarely more than once or twice. The original researchers often do not have the inclination, incentives or the luxury of going beyond their original mandates in analysis. Just like government data is underutilised if it only remains in an official report, the utility of privately generated data may far exceed this limited use.

Data and information are also network goods. Data sets can complement each other and together they can yield  richer knowledge than they would on their own. A lack of public sharing of data sets and a culture that does not expect that of its knowledge creators prevents this from happening.

None of this is to say that individuals are not generous with what data is at their disposal. While rent-seeking remains a problem, many researchers and organisations are happy to oblige requests for data. However, this remains sub-optimal as personally investigating and enquiring after data results in very high costs for searching – if not monetary, then certainly in time and effort spent. The only way to change this is if more people adopt online, open disclosure of data.

While an open culture around data is desirable, it takes several complementary actions to get there. Access to private data, whether it is generated by a business or by an NGO, cannot be treated as a rights issue. Individual freedom and private property need to be respected while creating enabling incentives and encouraging voluntary efforts to open up information.

As research and data get exposed, the first problem that arises is that faults and errors become evident and can cause people to beat a quick retreat. What needs to be kept in mind is that pioneers who open their data sets before others can reap a large signaling dividend. Modest research with openness could have a higher impact than a superior quality study that remains closed. If sustained, open data researchers can build a reputation that is several notches above those who keep their data closed.

The second problem is that open disclosure results in a loss of control – where unknown anonymous users could use it without giving proper credit and for radically different ends. While this can indeed happen, the public benefit from open disclosure and the credit for it remains higher than such losses.

The third problem that can arise is that as the culture of sharing is weak and data science is still nascent in India, the benefits of opening up data sets may take a lot of time to yield impressive results. What is also needed is a kick starter – perhaps in the form of scholarships for students and researchers to use high quality datasets and publish new results.

Donors and foundations funding research and analysis also have an important role to play here – by committing to openness and providing platforms that enable sharing, they can change the work cultures in organisations that they fund. Official mandates to that effect government-funded research would work the same way. In India there is ample precedent for comfort: agencies such as the World Bank and others have embraced open data, it is for others to keep up.

If India is to transition to a true knowledge economy, open access, availability and contestability of public knowledge is paramount. A narrow campaign to liberate government data will have far less of a lasting impact than a broader attempt at creating a culture of openness and sharing around information and data. It is time we started on the latter.

Comments { 2 }