Inflation measurement issues in India

G V Nadhanael and Sitikantha Pattanaik of RBI have written an insightful paper on the issue. It is an excellent primer on inflation. If there is someone who wants to understand inflation issues in India and has to read one paper, this one is just right. Though this study came a while ago, I am just pointing in case some people have not read it.

Ambiguity in inflation assessment resulting from deficiencies in data on prices could pose significant challenges for the conduct of monetary policy. In India, in view of the large divergence between CPI and WPI inflation trends in the past, wide dispersion in inflation across commodity groups within WPI, and significant volatility in headline WPI inflation under the influence of supply shocks, the statistical limitations of prices data have received increasing attention in the policy debates. This paper presents the key issues in the current context, while also explaining how policy analyses relevant for the conduct of monetary policy could yield ambiguous results if inflation data used in such analyses have serious limitations.

A comparison of the WPI inflation against the GDP deflator suggests that both track each other almost perfectly, which may lead one to draw the wrong inference that the usual arguments against WPI in terms of non-inclusion of services and non-revision of the index to capture the structural changes in the economy are not very relevant. This paper highlights the scope for possible misleading inferences due to data deficiencies and also suggests the areas where improvements in data collection and dissemination may serve the monetary policy needs of India better in future.

The analysis about GDP Services deflator and WPI Inflation is very interesting. I am actually wondering how the two move together. Just a plain statistical coincidence!  Then the authors present analysis comparing GDP Agri deflator with WPI Primary Articles and GDP Manufacturing deflator with WPI Manufactured Products deflator. Thankfully we some co-movement here. Though I would not be surprised if the authors were not able to find any co-movement. Such is the beauty of statistics . And this isn’t true only for India though may apply more to India.

Though now we have a more updated WPI inflation based on 2004-05 series, much of the issues still remain. There is an interesting discussion on biases that remain in making the index more accurate.

Accuracy of the Price Index

How accurately the price index captures the movements in general price level is the next key issue in the measurement of price indices. Edey (1994), Diewert (1998) and Hill (2004) identified the major sources of bias that could enter into measurement of any price indices, with special refere*/ce to consumer price index. These are:

Product substitution bias

Most price indices are based on a fixed weight (Laspeyers Index with base period weights or Pasche Index with current period weights) and do not take into account the impact of product substitution within the production/ consumption basket. Consumers may substitute some products whose relative prices have increased over time with some other goods whose relative prices have fallen, in order to maintain the same level of utility. Therefore, fixed weight based price indices tend to either overestimate or underestimate the true cost of living.

Quality bias

Quality change arises when a firm produces/provides a new improved version of a product/service. Frequently there is no overlap between the two versions. That is, the old version is discontinued as soon as the new version appears. Nevertheless, statistical agencies try to splice the two price series together, often without making any adjustment for improved quality which leads to a bias. To avoid this, hedonic estimation methods are used by many countries to account for changes in quality. Boskin (1996) report estimated that in the US the inflation is upwardly biased by 60 basis points on account of the quality changes in products.

New goods bias

New goods introduce an upward bias into the measurement of inflation for two reasons. First, when a new good appears in the market, it would take some time for the good to be represented in the basket of commodities in the price index. Second, introduction of a new good in itself represents a price fall. Hicks (1940) argued that we should view the price in the period before a new good is introduced as the minimum price at which demand is zero. This price is referred to as the reservation price. Therefore, when a new good first appears, we can interpret its price as falling from the reservation price to its initial selling price. This reservation price can be estimated econometrically.

Outlet substitution bias

If there are substantial price differences between different outlets, consumers tend to substitute those outlets with higher price by lower price outlets. If the collection of price data is based on fixed number of outlets, which is not changed frequently, the consumer price index thus constructed may overestimate/ underestimate inflation. For estimating the outlet substitution bias the difference in quality across different outlets need to be controlled.

The other major issue with regard to the measurement of price indices is how to account for seasonality of different products. Certain products enter the market only in a few seasons/months and are absent for the rest of the year. This could make the comparison of prices across different months difficult. Annex I illustrates how the measurement issues are addressed in different countries.

In the end authors say:

This paper highlights that the Indian inflation path has been significantly conditioned by two major supply shocks, i.e., oil and food. Even the exclusion of these two items, representing the most conventional measure of core inflation, does not impart greater stability to the inflation path. Also, such exclusion makes the core measure much less representative, since the common man is primarily affected by food and fuel inflation. In this context, addressing the statistical issues relating to weights and coverage consistent with the changing economic structure assumes importance.

This paper also shows that year-onyear inflation has been much less volatile than sequential month over month (seasonally adjusted) inflation, suggesting the relevance of the former for conduct of monetary policy. The distribution of inflation across commodities at any point of time shows large dispersion, and the assessment of generalised inflation, given the large dispersion, could also complicate decision making relating to use of monetary policy actions to contain overall inflation. While different prices data covering specific segments of the population/regions of the country are necessary to assess their economic conditions in relation to developments taking place at the aggregate level in the economy, for the conduct of monetary policy, however, a single representative measure of inflation that could be available with limited time lag may have to be aimed at over time.

Current initiatives like updating the base in WPI and collection of new data on CPI-Urban and CPI-Rural as well suggestions for constructing services price series are steps that could be expected to facilitate convergence to a single measure of inflation over time that would be more suitable for use in the conduct of monetary policy.

Very useful..

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.


%d bloggers like this: