It has been a regular complaint that India’s statistical infrastructure has not kept up with time and many may argue that it has in fact deteriorated. This deterioration, according to them, has not been a recent phenomenon. Nonetheless, the fact that the National Statistical Office (NSO) has been subject to several controversies regarding their surveys over the last few decades should be worrying. Is something broken with our data collection process—or is it simply that our methodologies have not evolved with time?
These are issues that are regularly discussed, and they have important implications for measurement of the impact of various policies. Take for instance the 2017–18 Consumption Expenditure Survey, which was withheld due to data quality issues. Initially, many were quick to hail the survey and its findings—this was despite warnings regarding data quality that were flagged by this author along with Dr Surjit S Bhalla. Three years have passed since, and now virtually every serious researcher accepts and acknowledges that something went wrong.
Unfortunately, we do not know what went wrong but we do hope that the NSO would investigate and address any shortcomings before the next round of the survey. The National Council of Applied Economic Research and The University of Maryland conduct a similar household survey that can be a good reference point for the NSO to evolve the methods and bring them up to date with the globally accepted best practices. To address such shortcomings will be fitting for an institution that was the pioneer in conducting household surveys in the 20th century.
The issues are not just regarding measurement and methods but also about the lag with which such datasets are made available. The unit-level data for the Periodic Labour Force Survey is a good example of this as one expected the data to be made available within a few months. This is now feasible thanks to the use of modern equipment such as tablets that can be deployed for data collection. Good data that is available at regular intervals helps with a better understanding of the present state of the economy—which is crucial for fiscal and monetary authorities to respond by way of their policy tools.
There is also an issue regarding the dissemination of these datasets that happens in a format that is unnecessarily complicated. To be able to use any of these unit-level datasets, one needs to spend about an hour just to extract it in a format so that it can be used for analysis. Some of the other government agencies such as the Indian Council of Social Science Research (ICSSR) maintain their database where they provide these data in an easier-to-use format—but again, this happens with a lag. There is no reason why the NSO cannot make these datasets available across formats used by social science researchers. Doing so will only improve the productivity of independent researchers and add to the overall research output. Hopefully, the NSO will also adopt the dissemination approach of the ICSSR and some other agencies in the near future.
It is obvious that these challenges have meant that governments too have had to devise a new way of monitoring their programmes and tracking progress. Most state governments and
the Union government have therefore moved forward with building dynamic dashboards to track outcomes of key government programmes.
The way these dashboards work is simple. The nodal ministry defines the key performance indicators (KPI) for these programmes and then tracks them on a dynamic basis. As an example, for Swachh Bharat, the KPI would be the number of toilets constructed by the government. This information can be collected on a daily basis and with geo-tagging and verification methods, the data can be as reliable as it can get. Similarly, for the Awas Yojana, it will be houses constructed and for the Jal Jeevan Mission, it will be the number of households provided with a tapped water connection.
All these dashboards ensure transparency in governance and provide for a more real-time feedback in the form of monitoring the programme implementation. Many of these dashboards developed by the Centre and various state governments are ushering in a silent data and governance revolution. These dashboards also fill the data gap between two household surveys conducted by various agencies.
India’s statistical architecture is therefore unique as it comprises the NSO, one of the pioneers of household surveys and large administrative data, which is now collected for these dashboards. For a truly modern statistical architecture, it now has to work on a framework for using both kinds of data for governance. More importantly, there is a need to collect the information from all these dashboards and provide a time-series dataset available at a single place—preferably with the unit-level dataset. Doing so will help researchers and analysts better understand India’s development trajectory over the last decade and more importantly help highlight areas where more work needs to be done over the coming decades.
While the NSO works on addressing some of the measurement issues, harnessing the large administrative datasets containing verified geospatial information can help plug in gaps in our understanding, while simultaneously motivating the need for newer forms of household surveys. India’s existing statistical architecture has done a great service for several decades as the country became a $3 trillion economy. However, the same architecture cannot be relied upon to understand and inform policy to aid its transition to a $5 trillion and hopefully a $10 trillion economy.
Karan Bhasin
New York-based economist