A recent piece in The Hindu Businesline had questioned the sudden boom in wages of the male population in rural India. Dhananjay Sinha, the author, based his findings on data compiled by the Labour Bureau (LB), which showed a 17% year-on-year jump in March 2026 – far above the trend of 5-6% pace seen in prior years. In July 2025, the average daily wage showed a year to year increase of 12.7% across all occupations. He highlighted that in the 10 months to June 2025, rural wages grew 6.23% on a year to year basis but by March 2026, they ballooned to 18.12%. Our study shows that over the years, the Labour Bureau has not followed well established practice of data sampling due to which the all-India rural wages may not have correctly reflected the situation even before the current revision from July 2026.Data collection processBefore analysing the unusual increase in daily wages, it may be instructive to understand the process of data collection and compilation. As per the LB, data is collected for rural India for 25 occupations – 12 agricultural and 13 non-agricultural. For this, 787 villages have been identified by stratifying the states based on agro-climatic conditions and population density.Then allocation of villages, in multiples of three, is made in proportion to the rural population within each state. Primary informants for wage data are village officials like panchayat secretary, patwari, block officials and other prominent village persons.It appears that till June 2000, average wage rates at the all-India level were derived by taking the simple average of state figures. Since July 2000, the averages have been obtained by dividing the sum total of wages of all the states by the number of villages from which data was collected. The LB portal also asserts that till June 2025, average daily wage rate was compiled based on wage rate information collected from 600 sample villages across 20 States/UTs. Consequent upon the revision of the base year of CPI-AL/RL to 2019 and the associated revision in sample size and sample villages, the coverage has been expanded to 787 sample villages across 34 States/UTs. This change became effective from July 2025. Accordingly, users have been advised by LB to exercise due caution while making comparisons of wage rate estimates across the periods before and after July 2025.DiscussionFrom the statistical perspective, the sampling design adopted is not suitable for the purpose of computing average wages. If the intention was to compile average wages of the rural labour, their numbers in the States ought to have governed the sample allocation and not their total populations. For instance, it is seen that a small UT like Daman & Diu, has been allotted 16 villages for collection of data. It had only 84 inhabited villages and merely 0.03% of India’s rural population in 2011. On the other hand, Uttar Pradesh, with nearly 98,000 inhabited villages had 18.7% of the country’s rural population, but data is being collected only from 71 villages from UP (Table below). Another instance is that data is being collected only from eight villages in Delhi vis-a-vis Daman’s 16.Table: Illustrative comparison of structure of rural population and village allocationState/UTPopulation structure (% of India’s population)Village allocated20192011NumberPercentUttar Pradesh19.5318.67719.02Bihar11.9411.11384.83West Bengal7.107.44364.57Rajasthan6.486.19334.19Mizoram0.060.0681.02Goa0.050.0781.02Sikkim0.040.0581.02Delhi0.020.0581.02Daman & Diu0.030.03162.03Source: Computed from relevant RGI and LB data. Minor rounding-off mismatches expected.Also, out of 34 States/UTs, allocation of villages is in multiple of three only in 12. Further, the aggregation methodology given by the LB simply states that average wage is equal to sum of total wages across all states divided by total number of quotations. We find it quite surprising that the Labour Bureau may not have used any weightage to states’ overall number of rural labour or even population. Thus, compilation of average wage rates by LB does not seem to be in tune with statistical logic. We feel that aggregates (from which rates may be later derived) from any sampling procedure require incorporation of appropriate weights at the stage of estimate generation to account for underlying demographic variations. The LB’s extant computation process altogether ignores this fundamental statistical requirement at the state as well as all-India level since neither the number of villages allocated to states nor the distribution of occupations across villages, even intra-state is self-weighting.As a result of this computation process, wages from the smaller states/UTs like Daman, Delhi, Goa, etc., have been allotted samples 20-77 times their population sizes. In these small states/UTs, the wages may be mostly urban-centric and much higher than large labour surplus states like Bihar, Jharkhand, Odisha or Uttar Pradesh. At all-India level this absence of weightage may have had the effect of showing a higher rate of rural wages. This may be appreciated from a simple example of an electrician whose wage in July 2025 was Rs. 800 in Delhi and Rs. 563 in Uttar Pradesh. ConclusionIt is thus clear that the process adopted by LB to compute average wage rates fails to adhere to statistical premises of the term. The labour bureau will do well to follow a robust and well established statistical methodology of sampling and allocating appropriate weightage.Sanjay Kumar retired as Additional Director General of the Ministry of Statistics & Programme Implementation (MOSPI); N.K. Sharma retired as Director General of the MOSPI, and Siraj Hussain is former Union Agriculture Secretary. Views are personal.