Reading between the lines: The PLFS survey

The much-awaited release of the Periodic Labour Force Survey (PLFS) is a sheer delight for researchers. However, a close look at the data raises several questions. This is important if India were to seriously find a solution to take advantage of its demographic dividend rather than nit picking at the data.

First, the questionnaire needs to be redrafted. We are in a new economy where jobs and earnings are incongruous. For example, the PLFS survey (block 5.1) directly asks the question of the status of employment of the household as per his/her profession. Given that employment in India means a permanent salary every month, if the surveyor simply asks “are you employed”, the immediate answer will be an emphatic “no”. But if one had asked the same person, if his/her income is “zero”, again the answer will be an emphatic “no”. 

Second, how does the PLFS survey account for the clear shift in unemployment age? As per the PLFS survey, the unemployment in the age-group 15-29 years is as much as 17.8 per cent. However, in the 15-59 age group it significantly declines to 6 per cent (overall at 6.1 per cent). The critiques will argue this is a case of serious youth unemployment, but it’s a reflection of changing employment pattern, with the percentage of men/women in the education system being very high until the age of 23-24. Earlier, it used to be only up to 17 years. As per PLFS, these people are not counted in labour force because they are still in colleges. This could thus push up the unemployment rate in the 15-29 age bracket as a pure statistical artefact (as unemployment rate is explained as a percentage of labour force). Interestingly, as per the MHRD data, the total number students enrolled for graduation and diploma (under and post-graduation) was as much as 36 million in FY18, of which 10 million is from the north alone. How does one account for this shifting employment pattern? 

We believe such shift in the employment pattern will result in very different unemployment rates for higher age brackets. Unfortunately, the report does not report unemployment rates for the 30-plus age group. Nevertheless, based on the unemployment estimates and the age group wise population shares provided, it is possible to infer the unemployment rates for the 30-plus age group. Our estimates show that the estimated unemployment rates for the 30-plus group are much lower than that of the 15-29 age group. In fact, there are instances where the unemployment rate turns out to be negative. Such negative values are possibly reflecting incorrect weights, for instance, the 15-29 age group is considered working age, but the entire population in this age group may not be in the labour force, thereby substantiating our case. If this is true, the weights employed in generating state and national level estimates may be flawed, leading to an overestimation of unemployment in the 15-29 age group. 

The possible imprecision in weights could be a result of the change in the criteria in the PLFS survey for the selection of households in the second stage for both rural and urban areas, based on the number of members in the household having general education up to secondary level (10th standard). This brings a huge bias. By doing so, we are assuming the first criteria of having a job (formal or informal) is having secondary education. This might be true for formal, but there is no such criteria for informal jobs. 

Such criteria are also not representative of the population if we look at Census 2011.  The percentage of people above secondary education is 35.2 per cent in urban and is only 15.3 per cent in rural. Given that household-level secondary education was not readily available from Census 2011, the PLFS survey might have used some proxies. Also, overall literacy rate is at 63.07 per cent, but as per PLFS, only 25 per cent is sampled from this population. Thus, by doing second stage stratification dependent on secondary education, we are making the sample highly skewed (under sampling). 

There are other issues also. For example, the data in the report states that the level of self-employment in urban areas has marginally declined between FY12 and FY18. This result is difficult to reconcile with massive government efforts towards self-employment, either through MUDRA or other government schemes and through formal/vocational training (the percentage of persons receive such training is constant between FY12 and FY18). How do we then interpret the Ministry of Skill Development data that indicates than an average of 80 lakh persons were skilled between FY16 and FY18?

The report also reveals that the extent of formalisation in the economy has declined between FY12 and FY18, and that is surprising. Thus, the results show an increasing trend of workers without any formal job contract and an increasing percentage of regular wage/salary earners not eligible for paid leave going up. 

We would like to highlight that today the formal jobs are 2.5 crore in government, Union, state, parastatal, 7.0 crore from the EPFO and 1.5 crore from ESI. Thus 11 crore is the total of people on payroll, who get a salary every month and are formal employees. Even if we assume a working population of 50 crore, this implies formalisation in the economy is around 22 per cent. This is not bad and it makes India the third largest payroll employee country in the world after China and the US.

The PLFS survey is a veritable storehouse of rich data and the NSSO must be complimented for such a stupendous effort. But such data should not be peddled to create a misleading narrative as this deflects us from the actual problem confronting India today. In fact, the protagonists should read the classic papers by Diamond, Mortensen and Pissarides (2010 Nobel winners) on why unemployment remains high even when jobs are available (so relevant in the context of India’s shift in unemployment age). But this does not absolve us from creating more quality jobs that are, sadly, scarce today.
 

Soumya Kanti Ghosh is group chief economic advisor, State Bank of India; Pulak Ghosh is professor, IIM Bangalore. Views are personal