Column: Wait a minute, Dr Rajan

Updated: Sep 25 2013, 03:55am hrs
The availability of year-on-year headline CPI inflation data from January 2012 has attracted the attention of market participants, analysts and policymakers. For the first time, a composite retail price index, with rural-urban disaggregate sub-components, is published at a fortnights lag. The new series is intended to overcome deficiencies in Indias price statistics, viz lack of a comprehensive retail price indicator; existing indices were sector-specific (e.g. CPI, Industrial workers or CPI(IW), CPI-Agricultural Labourers, or CPI(AL), with different index baskets. Such issues obstructed finer monetary policy decisions, leaving the wholesale price index (WPI) as the headline inflation indicator.

Specifically, the increasing divergence in WPI-CPI inflation rates has prompted the RBI to attach more significance to the latter in monetary policy formulation. While the central bank continues to anchor inflationary expectations to WPI inflation, its policy rate has increasingly tracked core CPI . Analysts believe that the new CPI is currently in a transition phase and its only a matter of time before the RBI formally switches to it as inflation anchor.

Notwithstanding the inherent appeal in targeting retail price inflation for monetary policy purposes, should the new index be taken at face value, presuming the data series are robust Statisticians would surely caution against any hurry with advice to wait until the series is long enough and stabilised. But given the appeal, few would like to wait.

Therefore, an alternative approach to test out the new index is to examine its rural-urban components and see if initial price trends conform to common priors about price formation in both segments. The twenty data points available so far, although not long enough for reliable statistical scrutiny, provide early indications on evolving inflation trends. These raise a few questions, and some disquiet, as to how robust the index might be.

The first of the accompanying charts shows an initial gap of 1.8 points between rural and urban retail inflation, which disappears with the two series converging by July 2012. As a crude counterfactual, the other chart plots CPI (IW) and CPI (AL), despite the limitations for comparison with the new CPI index. The old CPI rural-urban inflation rates exhibit a measure of divergenceinflation for industrial workers stayed above inflation for agricultural labourers broadly until September 2012, the trend reversing thereafter.

So, why should rural and urban inflation rates converge as per the new CPI Thats puzzling, given that the sub-components of each sector carry different weights. Isnt it extraordinary that inflationary trends in both the sectors have remained almost similar in level as well as direction, considering that rural housing segment carries a zero weight vis--vis 22.53 for the urban counterpart, and that weightage for food in rural segment is almost double that of urban segment

An obvious question is if this convergence is a mere coincidence, e.g. there could be significant divergences at further disaggregate-levels, which would then imply that an aggregate level convergence is a short-run phenomenon. While such an outcome could be a statistical possibility originating from construction of the indices, especially over short periods, this does not shed any light on a theoretical explanation of similar aggregate demand pressures across two diverse economic sub-sectors.

We also examine the four major sub-components of the respective rural-urban indices of the new CPI. These are food, beverages, etc; fuel and light; clothing, bedding and footwear; and miscellaneous items that mostly consists of services like education, medical care, transport and recreation. What intrigues is the fair degree of convergence in inflation trends here too; fuel, where prices are mostly administered, is the only exception.

At an individual commodity group and services level, some inflation differential is observed, as the recently published RBI Annual Report notes. For example, in respective services sub-groups (under miscellaneous services component), we find some convergence in rural and urban prices in recreation and amusement, transport and communication, and personal care. Amazingly however, the differences between rural and urban inflation rates in medical care and education, two key sub-components, indicate little difference! This raises suspicion about the reliability of the source data.

Assuming the new CPI is initially robust with no uneven fluctuations, what do these trends signify Do they suggest that rural demand is sustained That rural demand is as high as its urban counterpart And if that is so, what factors are closing the inflation gap

In theory, convergence in price levels and inflation differentials across regions (or for that matter, countries) can be driven by cyclical and/or structural dynamics. Typical drivers include rising incomes, productivity growth, market integration, competition-induced effects, price deregulation, etc. Which of these explain rural-urban inflation convergence

Market integration, often a key force for price-inflation convergence in goods markets, is hard to apply in the case of services; nearly half the new-CPI basket consists of various services. Even so, a market-integration hypothesis to explain inflation convergence in rural-urban services, leads to an inference that rural demand is rising at a similar pace as in urban areas, or that there are supply-side issues. But if there are supply-side constraints, then it is strange coincidence that inflation rates are similar in both sectors.

Perhaps the new CPI is reflecting hitherto unknown, maybe deeper, fundamental facts. But if the disappearance of rural-urban inflation differentials is recent, there is a need to probe deeper, for example, structural changes induced by strong growth from 2004, or any other.

It is also possible too much significance is being assigned to a data series that has only twenty observations. Perhaps the new index hasnt stabilised yet and it would be best to wait until it passes statistical scrutiny for robustness and reliability. Prudence demands that theoretical explanations be explored thereafter.

Renu Kohli

The author is a Delhi-based macroeconomist