It has been more than a year since RBI began to target CPI inflation as its nominal anchor, replacing WPI inflation. On September 20, 2013, RBI Governor Raghuram Rajan signaled the change by raising the policy or repo rate by 25 bps, beginning the alignment with the much higher CPI inflation, though a formal announcement came four months later. In its January 28, 2014, policy review, RBI accepted the “glide path” recommended by the Patel committee report on a new monetary policy framework.
Expectedly, there were oppositions to RBI’s move from several quarters to which we too contributed. Our concerns were manifold, but the main focus was upon a potential, secular rise in cost of funds and its consequent impact upon output persisting into the long run. In a recent FE column, we highlighted how poor credit off-take as a leading indicator was pointing to low manufacturing activity amidst extraordinary market optimism. The subdued manufacturing growth of recent months has brought our concerns to the fore. Today’s column looks back at all the issues that we raised to examine their worth.
The issues raised against adoption of the new CPI as nominal anchor were threefold: i) statistical; ii) representativeness; and iii) timing. The statistical issues flagged if the new CPI series launched in January 2011 was long enough to instil confidence about its use as a policy variable; a related issue was an observed stickiness in CPI-core (excluding food and oil). Apprehensions about representativeness referred to the disproportionately higher weights on food and related items in the index. Finally, the issue of timing questioned if RBI should not wait for a shift to CPI until the WPI and CPI inflation rates converged for a smoother transition.
How have these concerns played out with the new monetary policy framework in place, albeit informally?
Analysts opposing the shift to new-CPI mostly focused upon its representativeness as food, etc, have a nearly 50% weight in the basket. With persistent, high food inflation for past several years, the fear was that the CPI index carried an inherent, upward bias relative to the previous anchor, the WPI, thus pushing up the policy rate. Therefore, some suggested that if the shift had to happen, CPI-core would be a better choice as the monetary policy anchor instead of headline CPI. This, they felt, would also spare policymakers from the high degree of volatility associated with food prices in our economy.
Although fully endorsing this view, our apprehension focused more upon the high and sticky CPI-core inflation. In several articles in this newspaper, we reiterated this concern by graphically showing how RBI’s policy rates mostly tracked WPI-core in the previous regime (2008-2011) when WPI-headline inflation was the official anchor, inching closer to CPI-core from January 2012; and even under the new regime with CPI-headline as the official anchor from late 2013. Our argument was that the upward bias in the policy rate would arise mostly from the observed statistical divergence between the new CPI-core and WPI-core inflation rates. While the food price inflation differential between the two inflation indices could narrow substantially in the medium-term if the government ably responded to supply constraints and reformed agriculture markets, we remained sceptical that the gap in CPI-core and WPI-core would be bridged.
Our concerns arose from the general trend observed in a developing economy characterised by a typical Balassa-Samuelson framework: CPI-core, which largely represents services inflation (i.e. non-tradable inflation), remains relatively higher than the WPI-core, which wholly represents manufacturing inflation (tradable inflation). This divergence between the two kinds of inflation rates is generally attributed to productivity differentials, often exaggerated by faster economic growth and excessive government spending. Therefore, as long as these three factors remain in play, respective inflation differentials could persist and impart an upward bias to the policy rate vis-à-vis the older monetary policy regime.
A natural question therefore, was the size of this divergence. And whether this was significantly above historical trend? According to then available data, July-August 2013, the core-CPI and core-WPI divergence was as high as six percentage points! Unfortunately, historical data for core-CPI was unavailable. Even the Patel committee report, which published a back-casted series of the new CPI using the existing CPI-industrial workers, did not provide data for CPI-core. Our past research, based on the implicit GDP deflator, indicated an average divergence of around one percentage point in 1991-2003, i.e., before India’s high growth phase began. While one would expect the divergence to widen further in the following years as growth accelerated, manufacturing gained in productivity and fiscal balances worsened, we were surprised by a retail-wholesale, or nontradable-tradable inflation gap of 5-6 percentage points.
This led us to question data quality in one of our articles. The stickiness of CPI-core inflation raised doubts about teething problems in early days of data collection; at the very least, we remained skeptical about the data on housing rentals. These apprehensions were misunderstood by some analysts that our fear was CPI-core would remain stuck at around 8% for long because of suspect data, regardless of the monetary policy stance. To the contrary, our argument was quite different: if indeed the data were robust, we would be stuck with a high divergence between tradable and non-tradable inflation; the latter, in turn, would get linked to a relatively much higher policy rate in the medium- to long-term, irrespective of the monetary policy regime. This is what led us to suggest RBI should wait until both the inflation series converged.
In the one year of CPI-headline inflation as anchor, the divergence in CPI-core and WPI-core declined to about 3 percentage points because WPI-core rose from the currency depreciation pass-through. As CPI-core witnessed sharp correction in August-September 2014, so did WPI-core. It remains to be seen if the divergence either stabilises at this level or again goes up once the pass-though effect is fully internalised.
Our stress on this divergence is linked to important developments in the credit market, where the transmission channel has remained delinked from the monetary policy stance throughout the period of informal, inflation-targeting regime. And inflation has moderated in this one year without any transmission of monetary policy signals. In the absence of loan demand, banks have refused to raise their base rates despite three rounds of a cumulative 75bps interest rate increases by RBI over September 2013-January 2014. Vigorous supporters of the new monetary policy regime, who anticipated the central bank would raise the policy rate even further in line with the Taylor rule, have largely ignored the sharply decelerating credit growth. The central bank’s argument for sustaining a high policy rate to stimulate financial savings has not quite yielded the expected dividend. To the contrary, banks have started reducing their deposit rates and effective lending rates to selective sectors to push up credit.
As RBI stands firm, holding on to its policy rate at 8% in its battle against inflationary expectations, two important questions emerge. When will RBI change course? And when it does, by how much? The answer to the first question would depend on how the central bank interprets its own inflationary expectations surveys, for there is too much credibility at stake. While significant numbers of professional economists seem to have well read Governor Rajan’s mind, disagreements appear to have emerged in the expert committee advising the central bank. But our focus was on the answer to the second question: At what policy rate would the monetary policy transmission channel restart? We believe both the level of CPI-core as well its divergence from the WPI-core could be an important pointer!
The author is a New Delhi-based macroeconomist.