The Indian Express
Since 2014, the poorest communities are earning less
Wage data collected by the Centre for Labour Research and Action reveal a steady decline in real wages of brick-kiln workers in the last 10 years. Brick-kiln work is a fallback occupation for some of India’s poorest groups.
Written by Jean Dreze
Updated: May 25, 2023
In an earlier article (IE, April 13, ‘Wages of distress’), I drew attention to recent evidence of a virtual stagnation of real wages, based on Labour Bureau data. To illustrate, the real wages of male agricultural labourers, non-agricultural labourers and construction workers grew at less — much less — than 1 per cent per year between 2014-15 and 2021-22. Since the point created some interest, an update may be useful.
Just-released Labour Bureau data reinforce the point: In 2022-23, the growth rate of real wages was just 0.2 per cent for the first occupation group, and negative for the other two. If we extend the time-series to 2022-23, the trend growth rates from 2014-15 onwards are as follows: 0.8 per cent per year for agricultural labour, 0.2 per cent for non-agricultural labour and slightly negative for construction workers (men only).
These estimates are based on semi-log regression of real wages (money wages deflated by the Consumer Price Index for Agricultural Labourers, CPIAL) on time. Others have replicated them without difficulty. Even the finance ministry graciously requested my permission to use these figures for a training session on macroeconomic diagnostics. For the doubting Thomases, a spreadsheet is available on request.
If you don’t like regressions, you can take an average of year-on-year growth for this period, the results are almost exactly the same (see table). And if you don’t like maths at all, you can just eyeball the graph of real wages, updated here — that is the most telling clue.
My analysis focused mainly on annual all-India wage figures. These figures are unweighted averages of monthly wages in about 600 centres, distributed across states in rough proportion to their population. Naturally, there are not many annual all-India figures (one per year per occupation group), but they encapsulate a lot of data. It is possible to unpack the dataset and look at more disaggregated figures, as I did down to the state level. That would uncover areas and periods of substantial growth in real wages, but it would not alter the overall picture of sluggish growth in the last nine years.
Little time needs to be wasted on Surjit Bhalla’s rejoinder (IE, April 25, ‘Wages are rising’). Bhalla has presented a summary table that contradicts his main point, namely that real wages were rising fast between 2014-15 and 2018-19. An attempt to replicate the table suggests that some of the entries in the 2014-18 and 2019-21 rows are interchanged. Incidentally, this goof-up created a glaring internal inconsistency in the table. It is surprising that Bhalla did not notice it.
In any case, the method Bhalla uses to estimate wage growth is flawed. Briefly, he calculates state-specific, month-specific year-on-year growth rates, and then aggregates them using unweighted averages over months and population-weighted averages over states. This might be called the “unpack-repack method”.
The inaccuracy of this method can be conveyed with a simple example. Let’s say there are just two months in the year, Primo and Secundo, identical in every respect except the wage rate. The daily wage rate is Rs 80 in Primo and Rs 120 in Secundo in Year 1, and vice-versa in Year 2. What is the growth rate of wages between Year 1 and Year 2? Zero, obviously. But the unpack-repack method arrives at a different answer: 8.3 per cent per year!
As this example illustrates, the correct approach is to put wages on an annual basis before calculating annual growth rates. That is what I had done. Based on this sound method, the growth rates for Bhalla’s reference period (2014-15 to 2018-19) are as follows: 2.2 per cent, 1.4 per cent and 1.3 per cent respectively for male agricultural labourers, non-agricultural labourers and construction workers. We might accept this as evidence of a slight “hump” shape in post-2014 trends, with a semblance of growth up to 2018-19. So what? The fact remains that real wages are much the same today as in 2014-15.
The evidence of near-stagnation in real wages goes much beyond the three occupation groups discussed in my earlier article (based on an RBI summary of Labour Bureau data). As Bhalla rightly points out, the original Labour Bureau series has data for many informal-sector occupations: 25 for men and 16 for women. This dataset is being continuously analysed by Arindam Das, Joint Director of the Foundation for Agrarian Studies. In a forthcoming study, Das examines trends in real wages for all these occupations between 2014-15 and 2021-22. For 21 out of 25 male occupations and 9 out of 16 female occupations, the trend growth rate is below 1 per cent per year. It is above 2 per cent per year for just two occupations (picking and horticulture, women only), that too based on patchy data. The picture is likely to look worse when the series is extended to 2022-23.
In short, Labour Bureau data clearly point to near-stagnation of informal-sector real wages in recent years. A comparison with similar data from agricultural wages in India
and the Periodic Labour Force Surveys would be useful. Meanwhile, there is corroborating evidence from other sources. For instance, wage data collected by the Centre for Labour Research and Action reveal a steady decline in real wages of brick-kiln workers in the last 10 years. This is all the more alarming as brick-kiln work is a fallback occupation for some of India’s poorest communities. Nothing like this has happened for a very long time.
The author is Visiting Professor at the Department of Economics, Ranchi University
