Should RBI prefer CPI over WPI

Updated: Nov 21 2013, 11:27am hrs
RBI has, for long, articulated the inflation goal in terms of the WPI but has always considered the CPI along with other variables under its multiple indicators approach. The current debate is whether the inflation target ought to be enumerated in terms of wholesale or consumer price inflation. There are sound arguments favouring a retail price inflation target, the most significant being it represents actual living costs faced by consumers. A country-specific context must inform theoretical principles though. Here, statistical issues as well as the level of inflation divergence between the WPI and the new CPI suggest caution in shifting to the CPI as inflation anchor.

The composite CPI is less than three years old; its reliability, statistical robustness is yet to be established; and its relationship with key macroeconomic variables central to the monetary policy decision-making process is not yet known or understood. For example, rigid persistence of core-CPI inflation in 8.2-8.6% range contrasts perplexingly with contractionary indications on the real side as well the downtrending WPI inflation. The magnitude and lag of wholesale-to-retail prices pass-through is critical here and answers to which can hardly be found from 22 retail inflation data points.

By construction too, three-fifths of the index is weighted with food and fuel prices that not only tend to be very volatile, but whose role in conduct of monetary policy is debatable. Another 9.77% weight is for house rents that are largely imputed, raising measurement concerns. Eliminating these along with food and fuel prices where supply shocks dominate significantly narrows the price information embedded in the core-CPI inflation. It is also geographically sensitive, whereas monetary policy decisions are aggregate. So, an early adoption of CPI could raise the probability of policy errors induced by erratic data points in an evolving, short time series.

The second issue relates to the wide WPI-CPI divergence, close to 4-5 percentage points. RBI hasnt really stated its official position on this or how it proposes to handle this issue even while informally inching ever-closer to the CPI. In the context of high food-price inflation persistence and the central banks belief that this gets generalised if left unaddressed, identification of the spillover process is very important as explained above. A formal shift could entail an upward shift in the yield curve and an across-the-board increase in the general level of interest rates in the economy. This could potentially hurt growth.

Then there is the argument that households inflationary expectations, an input into monetary policy decisions, more track CPI rather than WPI inflation. This is true, for food and fuel costs are what consumers face at a daily, retail level. These items have a relatively higher proportionate share in the average Indian consumption basket vis--vis the advanced economies, for example. It is also true that statistical agencies cannot deviate much in the construction of the CPI in this light. Here, policymakers need to exercise their discretionary judgement as to whether a country-specific adaptation in adopting the CPI as inflation anchor is called for. Given the dominant role of food and fuel prices in forming inflationary expectationsthat these are usually backward-looking as well as adaptiveand the thin, often contradictory empirical evidence of their translation into realised inflation, the issue deserves serious thought.

In this light, the WPI-CPI debate must be more nuanced and cautious than a simple binary choice. Its also important to not lose sight of its genesis too, one of which is RBIs failure to bring inflation under control over 2009-11. Something new was needed to cover the loss of credibility and the new CPI happened to come along at the right time. First, there was talk of a new normal for inflation, implying a higher target. If I am not mistaken, it was the IMF that first suggested shifting to the CPI for better inflation control; several analysts then took this up. That monetary policy was too loose in retrospect is something that must not be ignored in debating the relative merits of WPI and CPI.

Renu Kohli

The author is a New Delhi-based macroeconomist

The coexistence of negative output gap on the back of deteriorating growth momentum over the last two years along with elevated and sticky inflation has been perplexing for Indian policymakers. In this context, RBIs discomfort with inflation along with its coherent pronouncements for anchoring inflation expectations and ensuring positive real returns for savers has revived the debate on what should be the appropriate inflation measure for an economy like India that renders itself suitable for policy-making. Like most countries, India has multiple inflation indices. At the retail, wholesale and economy wide level, CPI, WPI, and GDP deflators are used to capture trends in price levels respectively. But unlike most countries where CPI inflation is regarded as a benchmark, monetary policy in India has largely centred around WPI inflation (the closest global proxy for WPI would be PPI, measuring prices at the producer level).

To be fair, larger focus on WPI had its merits. The index was available on a timely basis, both commodity and geographical coverage was better and there was sufficient availability of history for policy research and analysis. The correlation between WPI and CPI (based on Industrial Workers series) was moderate at 43% with average CPI remaining lower than the average WPI inflation by 61 bps during 2000-08.

But with change in the global and domestic economic structure since the 2008 crisis, the moderate level of similarity between WPI and CPI inflation gave way to greater disparity as CPI inflation not only became structurally elevated but also outpaced WPI inflation. While average WPI inflation moved up from 5.6% during 2000-08 to 7.1% since 2009, average CPI inflation accelerated sharply from 5.0% in 2000-08 to 10.3% since 2009. The deterioration in CPI inflation is not only absolute but also relative. A sample of top 100 economies (based on GDP size) reveals that Indias CPI inflation is expected to be the 8th highest in 2013. This is a marked deterioration from 2000-08 when India on an average used to be ranked above 40 with respect to CPI inflation.

Keeping in mind the long-term adverse impact on savings from structurally high CPI inflation, recent commentary from RBI suggests a growing preference for including CPI inflation for policy-making.

A gradual move towards according equal importance (if not greater) to CPI inflation is justified as the WPI suffers from significantly large impact from price movement in tradables while also ignoring the services side of the economy. Furthermore, it has been observed that WPI inflation is relatively more volatile and is subject to frequent and large revisions vis--vis CPI inflation that is less volatile and prone to fewer revisions.

While the concept of using core inflation (ideally derived by filtering out inflation in volatile items like food & fuel from the headline inflation) for monetary policy has started to find global acceptance, the evidence for emerging markets is not yet conclusive. A case in point for India is the post-crisis stimulus led recovery in domestic demand in FY10, when CPI inflation averaged 12.3% while core inflation (as captured by the non-food manufacturing prices in WPI) was flat at 0.2%. CPI inflation was led by over 20% rise in housing inflation and moderate rise in services inflation as demand for services remained supported by over 10% growth in the sector. Since significant part of growth at this period was supported by government stimulus, improvement in demand conditions was aptly manifested in the trajectory of CPI inflation. On the contrary, WPI inflation was unduly influenced by the glut in global commodities post the Lehman shock.

For an economy like India, the role of inflation expectations which are driven largely by non-core inflation as expenditure on food and fuel inflation constitutes a significant part (59.2%) of household income cannot be ignored as they have important ramifications for saving-investment behaviour. As such, headline inflation (WPI and CPI) should continue to guide monetary policy in the near term. Going forward, with provision of adequate time series with greater granularity, CPI inflation can start assuming a dominant role in shaping monetary policy framework.

Shubhada Rao

The author is chief economist, Yes Bank