RBI study fills four-month gap in state inflation data using survey inputs | Banking

Reserve Bank of India (RBI) researchers have estimated state-wise urban inflation for four months during the Covid-19 pandemic for which official data were unavailable, using household inflation perceptions, three-month-ahead expectations and past inflation data. The estimates were refined using historical forecast errors to account for the pandemic’s unprecedented impact on prices.

The refined estimates, when aggregated, tracked the official all-India urban CPI inflation series closely, with differences ranging from around 14 to 22 basis points across the four months.

The study, by Purnima Shaw and R K Sinha, uses granular data from the RBI’s Inflation Expectations Survey of Households (IESH) to estimate consumer price index (CPI) inflation for April and May 2020 and the corresponding months in 2021.

Restrictions on movement and the closure of shops during the pandemic disrupted price collection, preventing the compilation of state-wise CPI indices for April and May 2020. While the Ministry of Statistics and Programme Implementation (MoSPI) released imputed all-India inflation figures for these months, state-wise data remained unavailable. This also meant that year-on-year inflation for April and May 2021 could not be calculated using the missing year-ago indices.

The findings highlight the potential of household surveys as an alternative source of information when conventional price collection is disrupted. State-level inflation data are useful for analysing regional price differences arising from factors such as transportation costs, state taxes and supply-chain conditions.

However, the researchers flagged limitations in the methodology. Household inflation perceptions and expectations may not correspond directly to the CPI consumption basket, while survey centres may not fully represent inflation trends across entire states. The initial estimates for April 2020 and April 2021 also struggled to capture the direction of inflation accurately, requiring further refinement.

The researchers said the reconstructed figures could help bridge gaps in state-level inflation time series and support future economic analysis, including studies relevant to India’s inflation-targeting framework. They also suggested comparing the estimates with model-based forecasts and exploring alternative statistical methods to improve their accuracy.