Data Delayed is Data Denied – for policymaking
The ongoing debate on employment – or unemployment in India reminds one of many debates on the same issue as in 2018. That both these debates occurred one year prior to the general elections is not surprising.
In May 2018, I commented on the challenges in assessing the strength of India’s labor market due to the lack of a reliable high-frequency large sample survey from the NSS. You can check that article out by clicking here.
For an economy as large such as India, NSS should have invested a lot more into data collection efforts. The lack of a reliable series had made many analysts look at other sources to get a sense of the directional change in employment without being able to determine reliable estimates of the impact of India’s growth on employment generation. Luckily, the advent of administrative datasets provided some relief at least for the formal labor markets.
In 2023, we now have not just administrative datasets but also a large representative sample survey in the form of the Periodic Labor Force Survey (PLFS). The PLFS is a high-frequency survey, however, it is published with considerable lags making its use for policymakers and economists difficult. There are also other private surveys that are less reliable – yet, they are more regular, and therefore popular press relies on them extensively. Many academics have been using these surveys even as several people have pointed out issues with respect to their sampling weights and inconsistencies in their surveys. Needless to say, there is work to be done by statistical agencies on improving India’s efforts not just at data collection but also at timely data dissemination.
Let us take a look at how labor market outcomes are measured in the United States – often regarded as the frontier in data collection and dissemination. Labor market statistics in the US are diverse and comprehensive making it the best available model to study and improve India’s labor market statistics.
Labor Market Data Collection Efforts in the US
Some key variables of interest would be the total number of employed and unemployed, the number of fresh vacancies, labor force participation rates, and wages. There are two sides to a labor market, firms, and households. Therefore, one needs two sets of different survey instruments that look at the labor market both from the firms (labor demand) and the households (labor supply) side.
For the US, the Current Establishment Survey is an establishment survey as the name suggests and then for the households, there is the Current Population Survey. In addition, there is data on the number of people who received unemployment insurance benefits and the Job Openings and Labor Turnover Survey (JOLTS) data. Most of this data is available on a monthly frequency. In addition, there are two other useful survey instruments, the Quarterly Census of Employment and Wages (QCEW), and the Quarterly Workforce Indicators (QWI) and then there is also the American Community Survey (which has 1-year and 5-year estimates).
If anything, the discussion above illustrates two important points. The first pertains to the extensive coverage of labor market data by various authorities in the US. Second, it highlights the importance of gathering such information that it is done at various levels and at different frequencies such as monthly or quarterly.
Data Collection Efforts in India
Many suspect that India’s labor market data collection efforts are not as comprehensive as they should but a closer look reveals a lot of information is available. For example, just as the US has an unemployment insurance claims measure available, India too has the MGNREGA Man Days demanded as a statistic that is available. Similarly, we have the payroll data from EPFO which captures monthly changes in formal sector employment (from the establishment side just as CES).
For the household sector, we have the Periodic Labor Force Survey which captures the dynamics in labor force participation and employment status in short time intervals of three months only for the urban areas in Current Weekly Status (CWS) and the second one for both rural and urban areas at an annual frequency which covers both Usual Status and the Current Weekly Status.
There is scope for improvement here as such information must be collected more often for both rural and urban areas – but more importantly, the data from these surveys should be made available in a timely format. That is, even with existing frequency, there are long delays in providing an update on the status of movement in labor market dynamics. This renders the use of the survey for informing policymakers such as the RBI’s monetary policy committee, the Union Government, or even state governments difficult on a regular basis. As a consequence, the policy response to weak labor demand would lag, often extenuating the negative effects of labor market weakness on the Indian economy.
Beyond the PLFS, we have the Annual Survey of Industries and the Annual Survey of Service Sector Enterprises which are establishment surveys. Whether India needs an additional comprehensive annual establishment survey by major sectors that covers both services and industries is something that policymakers must evaluate as the size of India’s economy expands. A regular survey on job openings and labor turnover would fit well within India’s existing statistical exercise of collecting data pertaining to the labor market.
Challenges to accessing data
Even with extensive data collection efforts, India’s statistical agencies falter in two major places.
First is the timely release of the data. A formal statistical calendar – and one which is binding on MOSPI to adhere to would go a long way in re-establishing its credibility. MOSPI must be legally obligated to explain to the parliament in the event of any delays in the release of data or conducting of a survey from the established statistical calendar.
The second pertains to the format in which such data is made available. Much of it is available in files that require researchers to write programs to import the data and clean it for further analysis. Technology now allows for APIs and typically such data is used only in select formats so it should not be difficult to make available all these files in these formats. Indian Council for Social Sciences Research already provides this for some of the old releases – MOSPI can now establish a standard practice to make the releases available in usable formats.
There is a case to add survey rounds and improve the frequency of data collection efforts but the above two points are far more important to make the most of the resources spent on answering the simple question on the number of jobs added and measuring the unemployment rate in every state or by different industries. Eventually, there is a need to add several more statistical products to better understand India’s development journey but here’s hoping that we make the most of the existing spent to collect information for informing policy.