## India’s Core inflation calculation — confusions and clarifications

This is a followup to the post on calculating India’s core inflation. Actually, I should have  expanded the post then but kept forgetting it.

In the previous post, I explained how one can estimate core inflation from the headline WPI inflation numbers. The idea was to exclude Manufactured Food Index from Manufacturing Products Index. This then becomes Non-Food Manufacturing Products Index. This is nothing but core WPI index (as RBI calls it) and based on the index we can calculate core inflation.

Actually there is another method as well and this is a source of confusion. The data based on the two methods should be same, but this is not the case because of rounding effect. Sometimes the differences are quite large leading to added confusion on which one to report.

Where the first method helps calculate core via exclusion, the second one is via summation method. This is how it goes:

• Manufacturing Products Index is a sum of twelve sub-sectors:
 Weights MANUFACTURED PRODUCTS 65.0 1 Food Products 10.0 2 Beverages Tobacco & Tobacco Products 1.8 3 Textiles 7.3 4 Wood & Wood Products 0.6 5 Paper & Paper Products 2.0 6 Leather & Leather Products 0.8 7 Rubber & Plastic Products 3.0 8 Chemicals & Chemical Products 12.0 9 Non-Mettallic Mineral Products 2.6 10 Basic Metals Alloys & Metals Products 10.7 11 Machinery & Machine Tools 8.9 12 Transport Equipment & Parts 5.2
• Ideally both should give the same answers but because of rounding, we see differencesIn the first, we subtract (Mfd Products Index*weight) – (Food Products* weight). Based on this we get core index for May-12 at 78.0850. May-11 reading is 74.4797.
• In the second, we add (weight*index) of the remaining eleven sectors (from Beverages to Transport).  In the below table, I have also put May-12 index of different sub-sectors. So if you add all (weight*index) except of Food Products, you get core index for May-12 as well. Based on this we get core index for May-12 at 78.0766. May-11 reading based on same is 74.4588.
 Weights May-12 Index Weight * Index 2 Beverages, Tobacco & Tobacco Products 1.8 170 3.0 3 Textiles 7.3 129.2 9.5 4 Wood & Wood Products 0.6 166.8 1.0 5 Paper & Paper Products 2.0 134.1 2.7 6 Leather & Leather Products 0.8 132.6 1.1 7 Rubber & Plastic Products 3.0 135.4 4.0 8 Chemicals & Chemical Products 12.0 140.7 16.9 9 Non-Mettallic Mineral Products 2.6 160.2 4.1 10 Basic Metals Alloys & Metals Products 10.7 165.8 17.8 11 Machinery & Machine Tools 8.9 126.6 11.3 12 Transport Equipment & Parts 5.2 127.1 6.6 78.0766
• Summing the two methods we get:
 May-11 May-12 Inflation Exclusion Method 74.4797 78.0850 4.8407 Summation method 74.4588 78.0766 4.8588
• So based on exclusion we report core inflation at 4.84% and based on summation we get 4.86%.
• In Apr-12, we had 4.85% based on first method and 4.77% on second. So, analysts looking at first say core inflaiton has declined from 4.85% in Apr to 4.84% in May. In second, they say it has risen from 4.77% to 4.86%.
• The % differences are small but the interpretation is very different. Those looking at first indicate softening of inflation. The second camp says core hardening and rate cuts not needed.

The question is which one to follow? I tracked RBI’s reporting of core inflation data (given in DBIE) and this is how we have inflation for three series:

 Summation Exclusion RBI Apr-11 6.99 6.97 6.97 May-11 7.26 7.34 7.34 Jun-11 7.78 7.73 7.73 Jul-11 7.69 7.65 7.65 Aug-11 7.78 7.77 7.77 Sep-11 7.93 7.99 7.99 Oct-11 8.07 8.13 8.13 Nov-11 8.42 8.35 8.36 Dec-11 7.99 7.91 7.91 Jan-12 6.98 6.96 6.95 Feb-12 5.84 5.87 5.86 Mar-12 5.02 4.96 4.95 Apr-12 4.77 4.85 4.86 May-12 4.86 4.84 4.84

Though there are differences in all three data series, we see RBI’s reporting closer to the exclusion series. And in some months RBI data is similar to Exclusion method. Ideally, RBI data should have been same for one of the two but this is also not the case somehow.

Given whatever we have, exclusion approach looks like the method to follow.

So, this is how core inflation data calculation looks as of now.I am not sure whether the post leads to any clarifications. It must surely be leading to more confusions..