Home / Opinion / Columns /  Need a more representative survey to track inflation expectations

The measurement of inflation is a serious business. Ask Sait Erdal Dincer. For much of his term, the head of the Turkish Statistical Institute was under fire from the opposition for under-reporting inflation. Two months ago, he was fired by Turkey’s President Recep Tayyip Erdogan because the inflation figures were too high. According to Turkish news reports, Erdogan felt that those price-rise figures presented an unflattering picture of the economy.

Thankfully, India’s inflation data hasn’t been politicized yet. But this does not mean that the country’s price-surveillance systems are robust. Three key data issues impact the diagnosis and cure of inflation in India. The first relates to the composition of the consumer price index (CPI). The second issue is the absence of a producer price index (PPI). The third relates to the inflation expectations survey that is conducted by the Reserve Bank of India (RBI).

The first of the three issues has received the most attention. Some economists have expressed fears that our retail inflation gauge may be out of line with reality, since CPI weights are based on the 2011-12 National Sample Survey (NSS) consumer expenditure survey. In particular, they worry that the heavy weightage of food items is distorting the index. They also point to some obsolete items in the ministry’s inflation-tracking basket, such as VCR cassettes, which are hardly used anymore.

These fears are a bit overblown. The old items account for a small share of the inflation basket. The weightage of food may be skewing our inflation gauge, but not in the direction critics are pointing. The 2017-18 NSS consumer expenditure survey showed that the share of food in overall consumption went up 1.6 percentage points to 47.5% between 2011-12 and 2017-18. The survey results were junked by the government because they showed a dip in rural consumption. Had it been used to revise the CPI basket of items tracked, the weightage of food would have gone up, not down. As people cut back on discretionary expenditure during the pandemic, the share of food in overall consumption only went up further. So the ‘heavy’ weight of food is not distorting the CPI index; if anything, its weight should be a bit heavier.

The lack of a PPI is a much more serious issue. Such an index tracks the price pressures faced by firms when they buy inputs. It can provide lead signals on retail inflation. The lack of a PPI index handicaps RBI’s monetary policy committee (MPC), since it has to rely on a flawed proxy—the wholesale price index (WPI). The use of the WPI, in lieu of a PPI, in gross domestic product (GDP) calculations is even more problematic, for it distorts the GDP deflator. In other countries, economists and policymakers can use the deflator as a broad measure of inflation. In India, the GDP deflator is completely unreliable.

The most serious issue, as far as monetary policy is concerned, relates to RBI’s inflation expectations data. One of the key tasks of any central bank is to anchor inflationary expectations. This can be done only when it is measured correctly.

Central banks typically gauge inflation expectations in two ways. One is to calculate the implicit inflation expectation from the yields of inflation-linked bonds. The other is through household surveys, where people are asked explicitly about their inflation expectations. In the absence of inflation-linked bonds in the country, RBI relies solely on its inflation expectations survey. Roughly 6,000 respondents spread across 18 cities are interviewed every two months.

This sample is too narrow to draw any meaningful inference about inflation expectations in a large diverse country such as India. The survey coverage and design have improved over the years but these changes aren’t enough. In 2012, the number of cities covered went up from 12 to 18. In 2014, the survey instrument was changed. In late 2018, the survey was redesigned to use a random sample for the first time; it was based on ad-hoc quota sampling earlier. These changes have also meant that the survey data are not comparable across time. This means that the data is of limited use in forecasting, even though several analysts—including RBI’s own researchers—run time-series regressions on this dataset without accounting for the series breaks.

The big problem with the RBI survey is its lack of rural coverage. Its design is based on household surveys conducted by Western central banks in predominantly urban economies. But India is predominantly rural. Leaving out the biggest chunk of a country’s population does not make sense. It also deprives the central bank of valuable price information that food producers in the countryside tend to have. This is especially important in an economy where food accounts for nearly half of the inflation index. RBI’s own working group on surveys set up in 2008 by the then RBI governor D. Subbarao had recommended that the survey be extended to rural areas.

RBI has a rich data legacy. The All-India Debt and Investment Survey began life as a rural credit survey of RBI in 1951-52. The next few rounds were also conducted by RBI before NSS staff took charge of the survey in the 1970s. RBI set up its data warehouse at a time when the statistics ministry was not even considering one.

RBI’s inflation expectations survey does not live up to that legacy. It needs to be overhauled. Only a large-scale nationally representative survey can provide an accurate measure of inflation expectations in India.

Pramit Bhattacharya is a Chennai-based journalist. His Twitter handle is pramit_b


Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.
Recommended For You
Edit Profile
Get alerts on WhatsApp
Set Preferences My ReadsFeedbackRedeem a Gift CardLogout