2 min read.Updated: 04 Apr 2021, 10:53 PM ISTLivemint
The launch of two mega surveys, one to see how migrants are faring and another to gauge our job scenario, should nudge us to ask how useful our data is and how to fix our statistics
A lack of quality data has afflicted Indian policymaking for long. This deficiency was felt most acutely after last year’s overnight lockdown trapped countless migrants in our cities without adequate means of sustenance. Millions braved exhaustion, starvation and mishaps—a few were run over by a train—as they resorted to walking down highways and railway tracks back to their villages hundreds of kilometres away. It was a massive tragedy that shook our conscience. The government’s delayed response was partly pinned on an inability to identify who all were in need of help. We had no database on migrant workers. Santosh Gangwar, India’s labour minister, admitted in Parliament last September that no data was available on how many migrant workers had died and how many had joined the exodus from urban India. Numbers collated from states later gave us an estimate of 10 million returnees, although some pegged that figure vastly higher. A similar statistical scatter attends official snapshots of employment. In an attempt to plug India’s glaring data gaps, our labour ministry announced two large-scale surveys last week. If designed well, these would aid policy in no small measure.
The first of those studies will survey some 300,000 households to get a picture of our migrant population, from their slice of our urban populace to their conditions of life and sources of livelihood. The other, a quarterly one, aims to gather job creation data from 150,000 companies. These are large sample sizes, and while we badly need reliable numbers on daily-wage earners and the job scenario in the country to aid policy formulation and welfare interventions, what matters most is not scale, but accuracy. We must get this right at the very outset. Details of this initiative are not yet available, but the surveys’ designers must ask themselves some questions before hitting the field. One, if a parallel exercise is conducted independently, will it yield the same results? Two, will the sample truly be random, faceless and representative? To the extent possible, will everyone in the target group be equally likely to be included in it? And will all questionnaires be neutral enough to keep out biases brought in by the queries themselves? Beyond this, definitions will also need clarity, as many of our statistical cracks are down to muddles of what qualifies as what. Disguised unemployment, for example, makes a ‘job’ difficult to define.
Standards of living have long been tracked by the National Sample Survey Office (NSSO), but one of its reports ran into controversy in 2019 after it found a decline in household consumption. That episode had sprung a toss-up: Either the survey was flawed, or the Centre’s relationship with reality. A few years earlier, allegations had arisen of data on the Indian economy being dressed up, when the Centre revised how we work out our national income. It weakened confidence in our official data. Yet, nothing is at out-damned-spot levels in our statistical set-up, surely, indelibly beyond correction. Also, some snapshots of India are way too slow to be of policy use. The NSSO’s survey of employment, for instance, is done much too infrequently, leaving analysts to rely on a tracker run by the Centre for Monitoring Indian Economy. The launch of two new mega surveys should nudge our government to review all its statistics, lay all its methods out for open scrutiny, and fix all that needs fixing.