The CPI data has just been released, and contrary to expectations, the annual rate of change (year-on-year) in the index was a very respectable 7.8 per cent. This is the fourth lowest monthly CPI number since February 2008, when a weighted average of rural and urban inflation rates registered 6 per cent. The lowest CPI inflation since then was in June of this year — 7.2 per cent. That is 78 months of data since February 2008, and two of the four lowest inflation months have occurred in the last three months.
The RBI has been fighting the good fight against inflation and many would be tempted to conclude that monetary policy is working towards reducing inflation and inflationary expectations. There have been three clear policy initiatives outlined by the RBI — the need to “anchor” inflationary expectations, the need to keep policy rates “high” and the need to implement “inflation targeting”. Perhaps all three measures have worked to reduce inflation — or, perhaps as documented here, none of these three “initiatives” has had any role in lowering inflation.
Tight monetary policy has as much to do with Indian inflation as it has to do with India’s magnificent loss to England in the recently concluded Test series. In a series of articles (academic and journalistic) written over the last decade, I have attempted to document that the traditional measures of monetary and fiscal policy — some version of money supply or credit growth or fiscal deficits — do not explain, at all, the path of inflation in India. Econometric models based on these variables, no matter how manipulated, explain zilch about Indian inflation.
Two variables are key to explaining Indian inflation. First, and obviously, international inflation plays an important role. However, the model that Indian inflation is some function of international inflation has broken down over the last decade. So what does explain Indian inflation, not only for the last 10 years but for the last 40? The policy of minimum support prices (MSP) for food determined by politicians at the Centre. MSP inflation explains a large fraction of the variance in CPI inflation around a 5 per cent mean.
The Indian inflation model is a one-variable, one-trick pony — CPI inflation is explained by MSP inflation of the previous year. The model simply says that the predicted inflation in any year is 5.1 per cent plus a third of the lagged MSP inflation. That is it. No repo rates, no real interest rates, no money supply growth, no fiscal deficits, nothing else. The model is estimated from 1978 (the policy of MSP came into being in 1975) till 2004. This end-year gives one a decade of out-of-sample prediction, as reported in the table.
A lot has happened in India, and the world, since 2004 — the last year for which inflation and MSP data has been used to “determine” the model. The average error of the model 2004-14 is only 0.25 per cent, with the largest over-prediction in 2009 (actual 10.9 per cent, predicted 13.2 per cent) and the largest under-prediction in 2010 (actual 11.9 per cent, predicted 8.2 per cent). The fact that the large errors follow each other is a positive aspect of the model. Excluding these two large errors, the average prediction error is a low 0.1 per cent, with a range of -1.6 to 2 per cent. Such accuracy is not a function of luck or randomness.
Apart from accuracy, the model has strong implications for the RBI’s repo policy aimed at reducing inflation. Since May 2013, the repo rate has increased from 7.25 per cent to 8 per cent, while GDP growth has plumbed to multi-decade lows. But inflation has declined by 200 basis points — from around 9.5 to around 7.5 per cent. Before (the monetarist) one gets carried away by saying “we told you so”, note that the repo rate increased from 6.25 per cent in end-2010 to 8.5 per cent end-2011; this not only had zero effect on inflation but was actually correlated with increasing inflation!
If not deficient demand and not tight money, then why is inflation down by over 200 basis points in 2014? It’s because of the MSP, stupid. In 2012, the MSP was increased by 16.2 per cent, so predicted (and actual) inflation in 2013 was close to 10.5 per cent. In 2013, the MSP increase was 6 per cent and both predicted and actual inflation were close to 7 per cent. Bulls eye (for details, see table).
RBI Governor Raghuram Rajan has a CPI inflation target for end-December 2014 at 8 per cent and end-December 2015 at 6 per cent. The MSP model says that both targets will easily be met, with the average inflation for 2015 predicted to be 5.8 per cent. Given that inflation rates are falling from lofty heights, it is likely that December 2015 inflation will be closer to 5 rather than 6 per cent.
I don’t know if more documentation or confirmation is necessary to verify that monetary policy has played a zero role in either its rise to 10 per cent-plus for six years or its predicted fall to around 5 per cent by end-2015. A friendly challenge is offered to the RBI and monetarists (hawks and doves) to either prove where my analysis is wrong or even suspect — or to offer a monetarist, inflation-targeting, anchoring inflation expectations, real interest rates model to explain any of the CPI inflation over the last decade. A fair duel deal?
And yes, an additional request to the RBI. Please junk your inflation expectations survey before it causes more damage to your enviable reputation. Making repo policy on the basis of the inflationary expectations survey is exactly the same as making policy according to a random noise generator. The latest June 2014 survey has the following gems, and I quote from the RBI report: “current perceived median inflation rate is 13.3 per cent, whereas median inflation expectations are 14.0 per cent for the three-month ahead period and 15.0 per cent for one-year ahead period”; and “The proportion of respondents expecting double-digit inflation in next three-month period have declined marginally to about 72.0 per cent”. Since January 2011 (43 months), the maximum y-o-y inflation was 12 per cent, and the average was 9.8 per cent. Over the last 24 months, average inflation has been 9.5 per cent and maximum inflation 11.2 per cent. Pure junk was never defined better than the results of the RBI expectations survey, the very same that the RBI uses for information on “anchoring” inflationary expectations.
The simple point is that inflation in India is a structural problem, and one whose structure is not determined by excess demand or deficient supply, or by mis-measured inflationary expectations, but by what the political arm of the government decides to do about winning elections and consequent vote-priming of rich farmers. We have just seen the disastrous consequences of such a policy for the political side of the equation. Food is 50 per cent of the CPI, so aiming monetary policy at food inflation is the same as raising interest rates to force Saudi Arabia to cut the price of oil.
The writer is chairman of Oxus Investments, an emerging market advisory firm, and a senior advisor to Zyfin, a leading financial information company