Flexible inflation targeting: Fix the underlying first - it is broken

British economist Friedrich August von Hayek once said: “I do not think it is an exaggeration to say history is largely a history of inflation.” So, any study on inflation and that too by one of the deputy governors of the Reserve Bank of India (RBI) and a Monetary Policy Committee (MPC) member is bound to make people sit up and take notice. And especially at a time when there is an impending review the monetary policy framework and the target band. The RBI paper avers that trend inflation has fallen and the monetary policy target inflation level is consistent with the trend. It suggests that the inflation target band of 4% +/-2% must be maintained in the medium term, setting off debates in the policy and media circles.

It is often surmised that low and stable inflation with well-anchored expectations is important for attaining higher sustainable growth. However, the entire discourse on the appropriateness of the inflation targeting framework misses the most important point — the signalling power of the underlying benchmark Inflation index, the quality and its composition. Importantly, this also implies that any discussion around the appropriateness of the inflation target has limited utility if the benchmark inflation Index is not quite appropriate. After all, any policy framework and its working are only as good as the underlying data on which they are based.

There are three major issues with the Consumer Price Index (CPI) data. First, the representativeness of the underlying consumption basket is questionable. The weights for the CPI basket were set a decade ago, with the base year being 2011-12. Per capita incomes have more than doubled since then and household consumption patterns have changed significantly. This change is likely to have been away from food to non-food items implying that the weightage of food, which is the most volatile sub-group in CPI, would have come down. This is exactly what had happened between 1999-00 and 2011-12; as per the Consumption Expenditure Surveys, the share of household spending on food items dropped by 10.8pp and 9.6pp in rural and urban India, respectively, during this period.

Second, the weights of items within the sub-groups also have likely changed. For instance, within the food sub-group the consumption pattern is likely to have moved away from cereals (which is about 25 per cent of food by weight) towards pulses, fruits and processed food, given the rising per capita incomes. Moreover, several items such as VCRs, DVD players, cameras, radios, tape recorders, 2-in-1s, CDs, DVDs, audio/video cassettes, etc have also become redundant and would need to be dropped and new items are likely to be added.

Finally, the data sources for price collection also need to be expanded to include digital channels given the shift towards digital forms of shopping and away from brick-and-mortar stores. Currently, prices are collected from 1,181 village markets across all districts and 1,114 markets across 310 towns in the country — to be fair, these numbers are quite large and seem representative. However, there should be some thoughts on the sampling side at the first level — especially due to the pandemic and how people’s preference for shopping possibly is changing and these behavioural changes may be sticky. Capturing online price points would be important to augment the CPI data.

Also, the base year for other CPI indices, ie the CPI-IW etc, should be made congruent with the national Headline CPI to enable more informed policy decision making from all institutions.

The moot point here is that the current CPI may not be accurately representing the actual inflation levels seen in the economy. This is a particularly dangerous proposition for monetary policy guided by a flexible inflation targeting framework wherein even a 25bp or a 50bp misrepresentation of the inflation trajectory, even if temporary, could lead to a substantial difference in historical as well as future policy outcomes.

The computation of consumer prices must move away from the current system that is quite static, with weights and items in the consumption basket being relooked at once in a decade, to a more dynamic system. This can be done in multiple ways. For instance, in the UK, the basket of goods and services comprising the CPI and their associated expenditure weights are updated on annual basis. This helps avoid many potential biases such as development of new goods and services, redundancy or even substitution towards cheaper goods. This would also ensure that the price indices reflect changes in consumer spending patterns. The alternative is a measure of Personal Consumption Expenditures, like in US, which uses expenditure data instead of choosing a fixed basket to allow for changing weights, substitution effect and would even take care of the problem of the inability of consuming some goods and services during a brief period as was seen during the pandemic. Interestingly, even in the US, where the Federal Open Market Committee (FOMC) uses the Personal Consumption Expenditures (PCE), the CPI is re-based and the weights are adjusted once every two years.

The RBI must push the government to put the periodical rebasing of the CPI into the inflation targeting agreement. This would be key to measure underlying inflation accurately, ensure policy credibility and avoiding unintentional policy mistakes. Thus, while the RBI paper quips that “if it ain’t broke, don’t fix it”, we believe the underlying CPI index definitely needs fixing first and that too on an urgent basis. 

The author is Chief economist, M&M Group

(With inputs from Rahul Agrawal, economist, M&M)