ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Socio-economic Performance of Constituencies: A Response

socio-economic Performance of constituencies: A Response Laveesh Bhandari related to poverty levels in Murshidabad as published by Mint and the Congress Party. That was, to clarify, not based on any Indicus estimate, and as such I am unable to respond to that debate.

DISCUSSION

Socio-economic Performance of Constituencies: A Response

Laveesh Bhandari

related to poverty levels in Murshidabad as published by Mint and the Congress Party. That was, to clarify, not based on any Indicus estimate, and as such I am unable to respond to that debate.

Another point was the difficulty in estimating constituency-level data. To get

A rejoinder to the article “Electoral Politics and Manipulation of Statistics” (EPW, 9 May 2009), which questioned the quality of data on Lok Sabha constituencies that was put out by some agencies.

Laveesh Bhandari (laveesh@indicus.net) is with Indicus Analytics.

Economic & Political Weekly

EPW
october 3, 2009

T
he article (“Electoral Politics and Manipulation of Statistics” by Himanshu, 9 May 2009) has criti

cised Indicus Analytics’ data on parliamen

tary constituencies as well as a recently

published paper on West Bengal, in addi

tion to some other data and reports.

I will limit myself here to (a) establishing

that there are many methods that make it

possible to estimate socio-economic pa

rameters at fine geographical levels, and

(b) Indicus results on socio-economic progress at finer geographic levels indicate some interesting insights that are also supported by other studies/data. I will refrain myself from other comments to his note.

We were first criticised for using flawed data for a paper on West Bengal that I coauthored with Bibek Debroy (“Transforming West Bengal: Changing the Agenda for an Agenda for Change”, Indicus White paper, mimeo, 2009). The paper as penned by the authors contains their views and not that of their organisations. The point of that paper was, West Bengal was earlier among the topmost states in India, and there has been a relative fall in its position over the last four decades. Almost all the data used in that paper was from the state statistical abstracts and sourced from various government of India and related entities publications. There were a total of 15 graphs and 44 tables (of which 36 in the appendix and eight in the main text). Of these only one graph and one table were Indicus estimates, and two tables from a World Bank study. Indicus data was only used to establish that West Bengal has a smaller middle class than the India average. To repeat, the whole study was almost entirely based on government figures, and mostly from the state statistical abstracts of West Bengal, Tamil Nadu and Maharashtra. But that was not the only criticism.

The article by Himanshu then discusses other unnamed authors and arguments

vol xliv no 40

constituency-level information we need to convert district level information to the constituency level. About x constituencies are the same as the district. But for the rest, one needs to translate district information to constituency-level information. How does one do it? There are many different ways. The Delimitation Commission has shared with the public a mapping of districts, blocks and their population, as well as constituencies. This allows us to take a weighted average of districts (weighted by the population…) to get constituency-level information.

Next consider the “criticism” that our data was showing falling immunisation levels. Yes, immunisation levels have fallen in many states, many of which some regard as “better” administered ones. And this fall in some parts of the country (not including West Bengal) is well known, not just to domain experts or line ministries. Perhaps the author was not aware that all major reproductive and child health surveys have been showing a fall in immunisation in some parts of India, all that Indicus data showed was that this is reflected at the constituency level as well (see National Family Health Survey 3 (NFHS3), and Reproductive and Child Health Survey (RCHS) data for greater insights).

Why might immunisations have fallen? When we look at information at smaller geographical units, we find that the fall in coverage is more pervasive than just in the case of reproductive health indicators. Water levels fall and hand pumps stop working. The Rajiv Gandhi Drinking Water Mission (RgDWM) continues to grapple with this slippage in coverage of safe drinking water. Similarly, electricity wires are stolen and connections fall, households increase but administrations are not able to quickly respond to increased coverage requirements, pucca roads get washed out in torrential rains or

DISCUSSION

floods, or need to be reclassified as kutcha, the examples are many. A village may be classified as having safe drinking water, connected to a pucca road and having electricity one year, but have none of them the next.

Data at disaggregated level are suggesting that the mission-based approach is a flawed path to socio-economic development. It has the advantage of rapid spread but the disadvantage of unsustainability built in. We need to better understand such implementation issues. For, the latest such mission – the Pradhan Mantri Gram Sadak Yojana (PMGSY) – may fall in the same category, with serious consequences for all.

Small Area Statistics

Next, consider the criticism of the methodology. We were criticised for putting out data that is not possible to estimate from conventional techniques and by using publicly available surveys. And that some of our figures are significantly higher than that from the National Sample Survey Organisation (NSSO). Two points: First, if it was possible, using the conventional approaches, to obtain such information then it would have been available in the public domain and our efforts would not have been required. Second, and more important, small area statistics is being used in India; broadly such techniques combine information from many different sources, find patterns that are significant and then conduct the most likely estimations. Parametric and nonparametric estimations, trends, calibration and many other methods using latest available information are used for such estimates. And researchers at Indicus are not the only ones who do this.

The Central Statistical Organisation has advised states to use calibration techniques for estimating GDP at the district level. Expenditure and income estimates using purely household survey data – i ncluding those by the NSSO – rarely account for more than 60% of that reported in the National Accounts Statistics (NAS). (Also see, “What Explains the Sluggish Growth in Consumer Expenditure?”, A K Tripathi, pp 12-15, National Seminar on NSS 61st Round Survey Results, 2007.) Calibration with NAS data is one (but not the only) method of correcting this problem. The National Council for Applied Economic Research, for instance, uses a functional approach where a parti cular distributional form is used for its estimates on city level incomes. Similarly, poverty estimates can be obtained in many ways, not just by merely estimating using household survey data from a district.

One method of correcting the small sample problem at the district level is to estimate for a larger area and then disaggregate it into smaller areas by using other proxies that reflect the economic well-being of those at the bottom of the pyramid, for which information is available. Some researchers prefer to use a combination of occupation and education indicators to e stimate economic wellbeing, and yet others prefer asset ownership for which data are available from other surveys.

Moreover, if we look hard enough large amounts of information is available on d evelopment parameters at the district, block and even village level. The PMGSY and RGDWM are only two examples of government services where coverage data being placed in the public domain at a fine geographical level. Moreover, the NSSO is not the only large-scale survey organisation. The Indian Institute of Population Studies conducted large sample surveys to estimate reproductive health at the district level in its District Level Household and Facility Survey conducted in 2007-08 (with samples of about a thousand households per district). Similarly, the National Family Health Survey has surveyed women and households across the country. The Indian Readership Survey and National Readership Survey collect a large amount of data on Indian households every year, though the raw data is not in the public domain, many of their estimates can be used for value addition.

Indicus takes data from multiple sources, not just the NSSO. Our methods differ across states and districts and across socioeconomic parameters. We estimate using multiple techniques depending upon what we consider to be the most appropriate to correct the problem at hand. We have shared the following set of parameters with various media groups. The bulk of them are socio-economic parameters:

  • (1) Total population, (2) Literacy rate,
  • (3) Female literacy rate, (4) Sex ratio,
  • (5) Work participation rate, (6) Infant
  • CD-ROM 2006

    The digital version of Economic and Political Weekly is now available for 2006 on a single disk.

    This electronic edition contains the complete content of all the issues published in 2006. The CD-ROM 2006 comes equipped with a powerful search as well as utilities to make your browsing experience productive. The contents are indexed and organised as in the print edition, with articles laid out in individual sections in each issue. Users can browse through the sections or use the sophisticated search facility to locate articles and statistics of interest.

    Price for CD-ROM 2006 (in India) Individuals – Rs 285 (Rs 250 plus postage and handling charges of Rs 35) Institutions – Rs 535 (Rs 500 plus postage and handling charges of Rs 35)

    International – US$ 40 (including airmail postage)

    Also available 2003, 2004 and 2005 on three separate CDs, individual CD price as above

    Any queries please email: circulation@epw.in

    To order the CD-ROMs (please specify the year) send a bank draft payable at Mumbai in favour of Economic and Political Weekly. The CDs can also be purchased on-line using a credit card through a secure payment gateway at epw.in.

    Circulation Manager,

    Economic and Political Weekly

    320, 321, A to Z Industrial Estate, Ganpatrao Kadam Marg, Lower Parel, Mumbai 400 013, India

    october 3, 2009 vol xliv no 40

    EPW
    Economic & Political Weekly

    DISCUSSION

    mortality rate, (7) Children (age 12-23 months) fully immunised, (8) Total voting population (+18 years of age), (9) BPL population, (10) Percentage of persons living in urban areas, (11) Under five mortality rate, (12) Households with electricity c onnection, (13) Habitations connected by pucca roads, (14) Households having a ccess to safe drinking water, (15) Violent crimes (% of total crimes), and (16) Crime against women (% of total crimes).

    The above variables were estimated for 2004 and 2008 at the district level. The

    -

    -

    -

    -

    -

    -

    -

    bulk of these have correlations of above 95%, with the latest available data at the state level. Some of the criticism was also towards Indicus income and expenditure figures that were not shared with the media, but placed on our web site and which got scrambled, these were corrected when it came to our team’s notice. Those figures are still on our web site

    There is a reason why Indicus data is respected and established media as well as academia use it. In the past our estimates have shown what few were

    -

    -

    -

    ---

    --

    aware of. We were the first to show how Mizoram was among the more socioeconomically advanced states in India, how agriculture was booming in many districts of Nagaland, how Himachal Pradesh had almost caught up with Kerala in basic e ducation, how Gujarat was booming e conomically despite a horrible law and order record. And many times we were criticised for it. Today much of this is common knowledge. And this is why e stablished organisations use our data.

    -

    -

    --

    -

    -

    Economic & Political Weekly

    EPW
    october 3, 2009 vol xliv no 40

    Dear Reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here

    Back to Top