Thursday, April 28, 2011

Kicking the wheels of the new CPI

by Ashish Kumar.

Inflation measurement is a critical component of macroeconomic policy. In a recent paper, Patnaik et. al. have argued that while the CPI-IW has many problems, these difficulties are not first order, and that the CPI-IW can yield a reasonable measure of inflation today.

On 18 February 2011, CSO released a new CPI with base year 2010 (Jan-Dec =100). This new CPI has five important new features:

  1. It is disaggregated at the rural and urban levels. The new overall all India CPI is a weighted average of the two. This is in contrast with the earlier CPIs which represented subsets of the population (industrial workers, agricultural labourers, rural labourers, etc.).
  2. The new series has better geographical as well as commodity coverage. The basket of consumer goods has risen from 25 to 250.
  3. The weights have been derived from the 61st round of the NSS consumer expenditure survey (2004-05).
  4. Data for the urban CPI will be collected from 310 towns (compared to 78 in the current CPI-IW, for all India). The rural CPI will use data from 1181 villages. Field officers of the NSSO and the Department of Post will be the price collection agents for urban and rural centers respectively.
  5. Since the two series are not comparable, year-on-year inflation numbers based on the new CPI will be available only from February 2012.
Sub Group New CPI
Rural Urban All India CPI IW
Food, beverages and tobacco 59.31 37.15 49.71 50.20
Fuel and Light 10.42 8.40 9.49 6.25
Clothing, bedding and footwear 5.36 3.91 4.73 13.28
Housing 0.00 22.53 9.77 5.33
Miscellaneous 24.91 28.00 26.31 24.94

The share of food in the new CPI series has seen a small dip in comparison to the CPI-IW while the share of services has risen. The share of housing has also seen a sharp rise. In CPI-IW, the price of housing services was imputed from the house rent allowance given to civil servants. For the new CPI series, housing prices will be collected through surveys of a sample of rented dwellings in 310 towns.

The weights in the new CPI are taken from a household survey by NSSO. This is, however, already quite dated given that it was conducted in 2004-05. It is hence interesting to compare these weights with those seen in the CMIE Consumer Pyramids dataset, which goes upto the quarter ending Dec 2010. This is a panel dataset where 140,000 households are measured every quarter.

The household basket as shown by CMIE gives a weight to rent based on households that report rent. The CPI uses an imputed rent. An imputed rent calculation for the CMIE data is not feasible based on the information presently given out by CMIE. In order to render the two comparable, we purge both consumption baskets of rent.

Sub Group New CPI, rural CMIE: Oct-Dec 2010, rural
Date 2004-05 2010-11
Food, beverages and tobacco 59.31 59.57
Fuel and Light 10.42 12.12
Clothing, bedding and footwear 5.36 3.75
Miscellaneous 24.91 24.54

In rural India, the weights of food and miscellaneous in the new CPI match that seen in the CMIE consumer pyramids even though the CMIE dataset is much more timely. In comparison to the Consumer Pyramids weights, fuel and light is under-weighted while the clothing category is over-weighted in the new CPI. The fact that these differences are small gives us increased confidence in the NSSO and in the new CPI.

Sub Group New CPI, urban CMIE: Oct-Dec 2010, urban
survey in 2004-05 2010-11
Food, beverages and tobacco 47.96 45.95
Fuel and Light 10.84 16.45
Clothing, bedding and footwear 5.04 3.78
Miscellaneous 36.14 33.82

A similar comparison in urban India shows noticeable differences in all categories. The weights for the food group is lower in the CMIE data. Both the clothing and the miscellaneous categories exhibit similar patterns. The fuel group has a significantly higher weight in Consumer Pyramids. Over time, the role of fuel has risen.

Sub Group New CPI, all India CMIE: Oct-Dec 2010, all India
survey in 2004-05 2010-11
Food, beverages and tobacco 55.09 53.48
Fuel and Light 10.52 14.06
Clothing, bedding and footwear 5.24 3.77
Miscellaneous 29.16 28.69

All India weights reveal similar patterns as urban India weights. This is not surprising because all India figures are weighted averages of rural and urban weights.

Assuming NSSO did a good job of measurement in 2004-05, this suggests that over a short period of time, the expenditure pattern of Indian households has been changing at a fast pace.

Despite the issue of weights, the new CPI series is a welcome step. Improvements in inflation measurement will be an important component of the Indian process of refashioning monetary policy to deliver low and stable inflation,

No comments:

Post a Comment

Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticising me is perfectly okay; uncivilised language is not. I delete any comment which is spam, has personal attacks against anyone, or uses foul language. I delete any comment which does not contribute to the intellectual discussion about the blog article in question.

Please note: LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.