India needs stronger statistical system for better policymaking

It was the careful support of policymakers and the political class that contributed to the Indian statistical system commanding respect globally. It is no exaggeration to say that it was in many ways top-notch, and even a pioneer in implementing sample surveys and other statistical system instruments.

But the last 75 years have seen India’s economy and society change. Our statistical system also needed to change to adapt. And it has. Ever increasing data is required to provide a better snapshot of the economy.

Unfortunately, while the last few decades have seen a proliferation of data from multiple sources, its quality has declined. In particular, the statistical system has been politicized and has largely lost its independence. Interventions have taken the form of denial, delegitimization and diversion. While all governments have been complicit, the credibility of the statistical system is now under threat.

Criticism of official statistics by governments and political parties was usual, and there was never an attempt to deny us access to data. This happened for the first time with the 2017-18 Consumer Expenditure Survey (CES) that was junked without any public debate. Questions have been raised even earlier, notably with the 1999-00 CES. But the then government had released the data and allowed a debate on it. Similarly, the fact that for the first time after independence, India is unlikely to have the decennial population Census amounts to another attempt at denying access to vital data. Both data-sets are crucial policy inputs. For example, the CES is essential for updating the national accounts and inflation indices. In their absence, we are using national accounts and inflation data that is more than a decade old and does not reflect prevailing consumption patterns. It is not just survey data, but even administrative data that is no longer readily available as it used to be. Data on distribution of income tax by income classes available since independence was discontinued after 1999-00. It was briefly revived in 2012, but was again discontinued after 2016. Data on caste as part of the SECC 2011 is yet to be released despite demands from the general public and academic community.

Second, even when data has been released, there were attempts to denigrate and delegitimize it. The Periodic Labour Force Survey (PLFS) is an example. It faced unnecessary and irrelevant criticism when it was leaked by a journalist. Employment data also suffered the fate of the CES and Census when the labour bureau’s annual employment surveys, started after the 2008 financial crisis, were discontinued after 2015.

Finally, for every official statistic, there is plenty of alternative data that is thrown up to raise doubts on the credibility of official statistics. Provident fund and pensions data was used to discredit the official employment statistics and create an alternative narrative.

The absence of credible official statistics has obviously given rise to alternative statistics from the private sector, which is unlikely to be a substitute for credible, robust and publicly available data. While these private statistics fill a void, they come with their own set of issues and are unlikely to be viewed in policy circles with the same credibility.

Data-sets on employment, poverty, inequality, education and even basic population numbers are essential inputs for several government programmes. Public availability of such data would let researchers generate debates on various important matters of governance. In fact, credible data is no less essential to a democracy than a free press. It allows society to question the government, analyse social and economic problems, and raise fundamental questions. Denying people access to vital data is not just undemocratic, but also blinds policymaking to changes in the economy and society that may need attention. The need of the hour is to strengthen the country’s statistical system to reflect these changes and provide better-quality data. An essential prerequisite for this is that the system is given the independence needed to generate data without fear or favour.

Himanshu is associate professor at Jawaharlal Nehru University and visiting fellow at the Centre de Sciences Humaines, New Delhi