CPI weights do not reflect Covid-19 pandemic realities, shows analysis

May well be understating it in some periods in the past two years

In April last year, the Ministry of Statistics and Programme Implementation could not release data on the consumer price index. Given the country was under lockdown from end-March to mid-April 2020, the ministry could not gather responses on price changes in the basket of commodities. Later, it released CPI for these months based on imputed values. While the April period was just an aberration and the ministry has since been publishing CPI data every month, the index has not been adjusted to reflect the changes in consumption patterns during the pandemic. In contrast, in the US, the government changed the basket of goods to reflect pandemic realities.

While India needs to pivot to a new consumption basket after the pandemic, it also needs a revised measure of how weights are assigned to different categories. The last time India updated these weights was in 2014, when the new CPI series reflected changes in weights based on the 2011-12 consumer expenditure survey. The 2017-18 consumer expenditure survey, which would have formed the basis of a revised goods basket, was junked by the government in 2019 due to inefficiencies.

Many of the categories that existed in the 2011-12 consumer expenditure have lost relevance since. Kerosene consumption, for instance, has reduced as many consumers have shifted to LPG connections. Similarly, landlines have become a thing of the past. A Business Standard analysis finds that if all such categories are excluded, then the weight of the commodities in CPI reduces by nearly 3.34 percentage points. By this measure, the increase in consumer prices in October was 0.26 percentage points lower than the actual figure for October. A similar analysis for other months shows that for most of the period, the consumer prices have been overestimated compared to last year.

During the pandemic, the effect would have been much worse as fuel consumption had reduced and spending on many items such as cinema halls had come to a halt. Besides, households cut back on frivolous spending as a result of the widespread uncertainty.

A BCG analysis shows that households were spending more on personal care apps and platforms last year in India. Similarly, in the case of food items, research indicates a reduction in 7-10 per cent from total expenditure, whereas there was a 15-19 per cent reduction in the non-food category.

If only the transportation categories were excluded, the CPI weight would reduce by another 6-7 percentage points.

The analysis does not consider spending on items like fuel, which would further distort the calculations.

Nor does it account for changes in the weight of commodities. The partial results of the consumer expenditure survey, which did find their way to certain publications, indicate that the share of education and rent in total consumption expenditure had risen, whereas the share of cereals had fallen. Private final consumption expenditure shows a similar pattern.

If one accounts for rent and education, then an analysis shows that house rent on average has increased 3.9 per cent, and private tuition and coaching expenses have gone up 5.5 per cent. In contrast, the increases in CPI have averaged 5.6 per cent.

While India certainly needs data from a new consumption survey, the methodology needs a change as well. At present, consumption surveys are a quinquennial exercise. The government needs to adopt international best practices. In the US, for instance, the consumption survey is an ongoing process, as the government keeps altering the sample size and includes new households. India needs a similar approach so that the revision to the statistics is more timely and in tune with the ground realities.