Steep decline in quality of labour data

The contemporary period is unusual in three senses. Firstly, we do not have contemporary and reliable official base on certain aspects of the labour market and industrial relations such as employment, strikes and lockouts. Secondly, doubts have been raised on the veracity of official data on inflation and the new and so-called recalibration of national income data with retrospective effect. And, thirdly, private sector data sets (CMIE, for example) and proxy data sets (payroll data) have emerged.

All have served fodder for “political controversies” and the approach of the general elections has generated much heat over them. But in the desperate “search for numbers” on employment the shortcomings of the existing statistical base have raised serious concerns.

In fact it is not only realistic but even ominous that the government in its press release (dated September 19, 2018) on the introduction of revised unemployment insurance scheme Atal Bimit Yyakti Kalyan Yojna for jobless workers has expressed that “change in employment pattern and the current scenario of employment in India … has transformed from a long term employment to fixed short term engagement in the form of contract and temping.”

So it is clear that the character of jobs has admittedly changed significantly enough to prompt an admission from the government — something which the trade unions and academics have been pointing out.

The government mechanically, and somewhat imperiously, announces that “Labour Statistics play an essential role for the formulation and evaluation of policies, which helps to increase understanding of problems, explain actions and mobilise interests related to employment in the country.”

But the relevant question here is whether the official employment data base is equipped to capture these significant changes? The answer is a clear “No”! Several official agencies collect data on employment/unemployment with differing definitions and classifications of workers and with different frequencies (decadal to annual/quarterly).

The Annual Survey of Industries (ASI) which covers the organised factory sector (comprising primarily the manufacturing sector) collects data annually on workers/employees among others from the establishments covered under the Collection of Statistics (COS) Act, 2008. The employment data relates to directly employed workers (male and female) and those employed through contractors (under the Contract Labour (Regulation & Abolition) Act, 1970, contract workers), supervisory and managerial staff and other employees.

Information deficit

Academic research has shown that directly employed workers comprise permanent workers (whose share has been declining) and non-permanent workers including non-statutory apprentices/trainees and fixed-term employees (FTE). The dataset is silent on nature of employment (based on hours, part or full time), workers employed under government schemes like ‘earn and learn’ and National Employability Enhancement Mission, etc., which according to trade unions offer de jure flexibility as they perform regular work.

However, owing to lack of clarity several scholars of even international repute equate directly employed workers as ‘regular/permanent workers’ and consider the contract workers as the only flexible category. Further, field research and anecdotal/press stories have shown that the extent of contract workers reported in the official data is after all an “underestimate”.

It should be clear even to the lay readers that not only do the analysts grossly underestimate “flexible workers” but also that we do not have even an idea of “transformations” that the government spoke of taking place in the labour market. Can this incomplete and grossly distorting data set serve the basis for the government to frame its labour and employment policies? By the way, data on gender composition of contract workers could be a welcome addition.

The data collected from the households, arguably the most comprehensive dataset, under the quinquennial National Sample Survey (NSS) by NSS Organisation (NSSO) — the latest data relates to 2011-12 — on employment classify workers as self-employed, regular and casual, which again presents problems for comprehension and interpretation.

The fine distinction between “regular” and “casual” is the presence (casual) or absence (regular) of daily or periodic renewal of work contract which distinction may neither be correctly and consistently conveyed by field officials (who recently were hired on contract!) nor be appreciated by the respondents; hence it is difficult to determine whether and to what extent the “regular” category of workers comprise “standard” or “non-standard/flexible” employment.

The Labour Bureau’s Quarterly Employment Surveys (based on establishments) follows the same definition of NSSO and hence suffers from the aforementioned limitations.

There are other data sources like Employment Market Information (EMI) which cover exclusively the organised sector and the latest year for which data under EMI available is 2012 (Economic Survey 2017-18) and disaggregated data sets are difficult to access.

The scarcity of data relating to the unorganised sector for which NSSO conducts non-periodic surveys and they are dated, need hardly be mentioned. The statistics on industrial relations (like industrial disputes or trade unions or labour turnover) provided by the Labour Bureau are pathetically incomplete and utterly unusable, and perhaps it is facing the prospect of natural demise?

We don’t have a separate and good data set on the service sector (which comes under the State-level Shops and Establishments Acts) where reportedly many jobs are getting created. Why couldn’t the Labour Bureau compile and provide sector-wise data as it has well-entrenched state labour departments (SLDs)? The latter, by the way, are terrible defaulters in submitting data and dilution of the administrative system in the name of reforms has made them worse. The question whether the labour administrative reforms are capable of generating reliable and periodic data has not been addressed by the government.

In short, labour market governance is at peril as there is a serious information deficit and it is imperative that the government takes measures to design afresh a statistical system reflective of dynamics of changes in the labour market.

The writer is Professor, HRM Area, XLRI, Xavier School of Management, Jamshedpur.