How are data to be collected and disseminated?

The data needs of a panchayati raj administration are very different from the data needs of policy personnel at more aggregative levels of administration, such as the State, or the country as a whole. Panchayats require data to address local problems, for resource allocation at local levels, and to identify individuals for various purposes. To meet these requirements, we need major changes in how data are collected, maintained, and disseminated. Three such issues are discussed in this blog, drawing on our new book A New Statistical Domain in India: An Enquiry Into Village Panchayat Databases (2016).

The method followed in the book is to examine existing databases in the two selected gram panchayats, and, using innovative methods such as micro-discrepancy analysis, to try to establish how these issues can be addressed.

A shift in the recording principle of data collection

Administrative registers often record data at the point of “occurrence” rather than at the point of “usual residence” of the individual. For example, the land revenue system records data on land where it is located and not where the owner of land resides. Individuals residing in a village may own land in other villages and towns, and land in a particular village may be owned by individuals residing outside the village. Hence, it is difficult to obtain data on land ownership of each resident of the village from official land revenue records, even though such information may be necessary for a panchayat to plan and implement land- and agriculture-related programmes.

Birth, death, and morbidity in India are recorded at the point of occurrence, which is often at hospitals or places far away from the place of usual residence of the mother. While this may pose no problem for a system of economic administration that is concerned with State- or district-level aggregates, it leads to the rather absurd phenomenon that a village panchayat does not know how many children were born to the residents of a village, or how many residents died or were affected by particular diseases.

The panchayat needs data on residents living within its jurisdiction. Such data cannot be collected and aggregated from official registers that record data at the place of occurrence. Hence, there is need to record data at the place of usual residence that can be used by the panchayat for administrative purposes. This leads us to suggest that panchayats require a “people’s list” or a core list of residents.

Panchayats’ need for unit-level data

Unit-level data refers to data for each primary unit of a survey. The primary unit may be an individual, household, enterprise, or involve natural resources such as trees, waterbodies, mines and quarries, depending on the object of the survey. Unit-level databases are ready to aggregate “statistics” and direct bases of “statistics” at an appropriate level of aggregation, such as the village, district or State. In India, as in many other countries, aggregated statistics are disseminated for databases such as the population Census, while household-level data are not disseminated due to privacy laws and concerns. In databases where household-level data are disseminated (such as unit-level data from National Sample Surveys, or National Family Health Surveys), the identity of households and individuals is protected. In some databases such as surveys of below poverty line (BPL) households, unit-level data can be identified and are available in the public domain.

The functions of a panchayat are such as to require unit-level data. Panchayats need to identify problems facing each individual/unit. Who needs a housing subsidy? Who is eligible for a scholarship? Which plots of land are affected by a plant disease and who owns them? What are the types of non-agricultural enterprises that exist in the village and how many people are employed? These are questions that micro-planners of panchayats face, and these questions can only be addressed with micro databases where units can be identified.

Ownership of data

In a centralized statistical system, data are collected by local officials and are sent to higher offices for aggregation and dissemination. The local agencies lose control and rights over the data in the process. For example, once data are sent to higher offices, local agencies may be denied access to unit-data completely (and given only access to aggregate data, such as in the Census of India), or they may have access to unit-level data but cannot make changes or update the data according to their own requirements without the approval of higher authorities (such as the BPL surveys or land revenue records). Hence the important questions that need to be addressed in a decentralized statistical system are: What should be the rights of the panchayats over unit-level data? Who should be authorized to update records and how? How would the changes be verified and incorporated in official statistics?

Gram Sabhas (or Gram Sansads) in India run on the principle of direct democracy, that is, through direct participation of citizens in decision-making. The larger public interest inside a village may justify the disclosure of certain types of information not otherwise be disclosed; indeed, the Right to Information Act, 2005, secures certain rights of access to administrative records at the gram panchayats if the larger public interest justifies the disclosure of such information (Article 8).

The Expert Committee for Basic Statistics for Local Level Development recommended that “the Gram Panchayat should consolidate, maintain and own village leveldata.” Easy as the proposition may sound, it cannot be implemented without far-reaching changes to the present statistical system. We argue that a new system can evolve from the current administrative framework, as there exists a gamut of registers and administrative records that has the potential to evolve into statistical databases “owned and maintained by panchayats” and used by all.