Opinion
The government must review its proposals on access to official micro-data
Summary
- India’s official data-sets are a public asset. But the statistics ministry’s proposed rules on request approvals would restrict rather than ease the access of researchers to important numbers.
With the objective of facilitating easy access to official statistics for all users, while implementing a basic requirement for official data (i.e. protecting confidentiality), so that the purpose of data-driven governance and inclusive development is better served, the Union ministry of statistics and programme implementation (MoSPI) recently prepared a draft of revised Guidelines for Statistical Data Dissemination and placed it in the public domain for comments.
As the theme indicates, this version proposes to revise MoSPI’s 2019 guidelines “to provide seamless access of statistical data to all stakeholders including government, semi-government, private data users, or international agencies, and promote data-driven decision-making."
Also Read: TCA Anant: The budget represents a missed opportunity to strengthen Indian statistics
Since 2019, MoSPI has been sharing aggregated/analysed data from its bank of data classified as shareable (termed ‘open access’). Its micro-data, which includes National Sample Survey (NSS) unit-level data and other numbers, is anonymized and shared through a simple user registration process. Both these types of data are provided free of charge.
Presently, while urban frame survey (UFS) maps are priced, user registration is all it takes to access up to a quarter of the total UFS blocks of a town, with authorization needed beyond this limit. Data categories that require registration are termed ‘restricted access data.’
In the case of the NSS, anonymized unit-level data is placed in the public domain simultaneously with the government’s release of data/reports. This facilitates immediate deeper data analysis. It needs no emphasis that most of the in-depth research done on India’s official statistics is based on such unit-level data.
While the ministry’s professed objectives underlying the revisions are commendable, the proposed changes and implementation mechanism seem to point in a different direction altogether.
For instance, although no modification is proposed for open-access data, major modifications have been suggested in the process of granting access to restricted data. In place of just registration, access is proposed to be granted through “registration and authorization." Also, the process laid down for obtaining this authorization would involve prior approval from MoSPI. Effectively, this entails examining the purpose and intent behind such data requests, presumably on a case-to-case basis.
Also Read: India’s central bank needs reliable and frequent employment data as a policy input
That MoSPI authorization would be a fundamental requirement for data access is unequivocal, since the relevant proposal stipulates that the ministry may request additional information regarding the proposed use of requested data, or other aspects of the project, and that all questions about the data-access application must be resolved before its formal approval. Also, the decision to grant data access will rest with the secretary of MoSPI, who will examine the genuineness and validity of data-usage requests.
All these proposed provisos and qualifiers, such as authorization, approval, necessary aspects, genuineness and validity-of-usage are vague and subjective, to say the least. Whether what is deemed necessary for research should be left to the discretion of MoSPI, or whether researchers should be trusted to judge their own data needs, is a moot issue.
Also, what attributes determine the genuineness of user requests and who can act as an objective arbiter of their intent? Research being an inherently exploratory exercise, can the validity of a research proposal be predetermined? Some deductions made from the data may be tenable but unpalatable; would access to such data be eased or barred? What if the final inferences vary, wholly or partly, from those visualized by a study proposal at its conception stage? Or is the unstated intent behind these preconditions to deprive general users of access to public assets, as official statistics are currently deemed?
Also Read: Data users need clarity on numbers from the statistics ministry
After the 2019 guidelines, with the passage of time and technological advancements, the user community was anticipating increased and widened sharing of official micro-data. However, the proposed draft guidelines demand compliance with nebulous conditions, the fulfilment of which has no objective criteria that can be followed for official approval. This is a rather retrograde step.
Moreover, there is no provision for challenging the denial of data requests; the decision taken by MoSPI is final. The draft is also silent on the circumstances that compelled MoSPI to resort to such stringent conditions for allowing access to a public asset.
In such an environment, in the event of access denials or undue delays, users may begin to wonder if data access is granted only to entities that are expected to support a particular narrative. Needless to say, such perceptions among stakeholders would ultimately work to deprive the government of the benefits of external inputs in the formulation policies. It may even hamper policy course correction.
To conclude, the proposed changes in the guidelines are regressive. If implemented, it is likely that the credibility of official data will suffer a severe blow. MoSPI should relook the guidelines and work to expand data access rather than restrict it. Only then can it assure the country of a positive role in progress towards policy formulation driven by the hard evidence of data.
These are the authors’ personal views.
The authors are, respectively, former additional director general and director general, ministry of statistics & programme implementation.
