The numbers that matter: Mint 6th January 2011

Inflation, growth and production figures are influential in policymaking. Yet they contain structural inconsistencies

Agricultural growth will be robust, Wholesale Price Index (WPI) inflation is edging down, industrial growth will decelerate in the months to come, and exports are growing rapidly and may cross the $200 billion mark. These are facts cloaked with some idiosyncratic statistical proclivities.
The term that pre-empts our ears is “base-year effect”, which is a valid explanation of most of what has been stated above. This has at times made growth numbers meaningless, conveying trends that may be divorced from reality. When vegetables are relentlessly selling at Rs 40/kg from Rs 20/kg a few months back, it is hard to accept a low inflation number. More significantly, it obfuscates the actual factors at work.
Financial year 2011 (FY11) has been a contrast to FY10. There was a drought in FY10, while industry topped with double-digit growth on a low base of 2.7%. Inflation was in the double-digit zone by the end of March. Further, exports had declined in FY10 as global demand had slumped. Now, with a gentle reversal of most of these indicators, the picture looks rosier, though disguised.
Take, for example, agricultural production. We are satisfied with the good monsoon and the kharif harvest, and the first advance estimates of farm output show foodgrain and oilseed production at 115 million tonnes (mt) and 17.3 mt, respectively. While these numbers show significant growth over FY10, they are still lower than the 118 mt and 17.8 mt, respectively, in FY09. This means the current growth has at best brought us back to the pre-drought level. The same holds for exports, which grew by more than 25% till November; but at $140 billion, was higher by just 3% over FY09.
At a different level, WPI inflation appears to be coming down, and the Reserve Bank of India’s target of 6% looks achievable by March. However, this is more because of the high base-year effect of FY10, when WPI climbed sharply during the November-March period. But month-on-month index numbers are still increasing even in the harvest period, while the inflation rate is coming down. The Consumer Price Index (CPI) inflation number, on the other hand, has shown an increase of 9.7% in October over 13.5% in 2009. How does one evaluate inflation then?
Second, the situation has been further obfuscated by adopting 2004-05 as the base year for WPI and gross domestic product (GDP). The WPI number has come down by 1 percentage point right away on doing this. The CPI, on the other hand, goes with a base of 2001 and a different set of goods, further queering the pitch for comparison. There is talk of the Index of Industrial Production (IIP) base also being aligned to the others. This makes sense, but perspectives could change and the current high numbers could become lower with the new base year.
Third, there is a logistics issue relating to revisions in numbers. The WPI is susceptible to changes, as is the IIP. In the last six months of 2010, the variation between the announced and revised IIP growth rates was 0.6-1.3%. Given that the numbers at times were of a low order, this could make a difference in interpretation. Even recent WPI numbers are susceptible to change, with an upward revision for September.
There are basically two policy implications here. The first is that data collection has to be made more robust. This is a challenge, given that 40-45% of the economy is unorganized. Farm prices are not homogeneous, while industry associations may not be reporting in time, making data collation difficult. Ideally, we need to elongate the data period, and can consider monthly numbers for even primary and fuel products.
Second, the focus of monetary policy has to change from “inflation rate” to “price level”, as the base-year effect cannot be eschewed. We have rarely had five successive years of high or steady industrial growth or inflation. This has led to a debate globally over whether the monetary authority should target inflation or the price levels. For example, if inflation is a target for RBI, should it be satisfied that the inflation rate is, say, 6% after 10.3% in March? The normal tendency during droughts is for prices to increase by 20-40%, and then drop by 20-30% to settle at a higher level than the pre-drought level. In such a case, has inflation really come down? As a corollary, should the monetary strings be loosened, assuming there is no latent inflation?
One way out is to actually compare the present with a moving average of the past two or three years. This will bring proper focus to policy, or else the targeted variables could be at variance from the ground situation. Also, the base-year changes must be synchronized and data released at the same time for easier comprehension. Above all, given the logistical challenge in an economy where a significant part is not organized, we should concentrate more on quality of data rather than frequency.