Opinion Indexing inflation

In recent years,consumer price inflation in India has slowly crept up. There are four measures of CPI,which have broadly moved together. The CPI for industrial workers (CPI-IW) is the one with the most recent weights. The year-on-year change of the CPI measured by this index has exceeded 5 per cent in every month of the last five years. This contrasts with most other emerging economies which have,in general,witnessed low inflation,especially after the global financial crisis of 2008.

At present,monetary policy in India is not explicitly charged with delivering low and stable inflation. Even then,monetary policy needs to care about inflation,given that in the long run inflation is always a monetary phenomenon. In this context,a major problem identified by the Reserve Bank of India (RBI) is the measurement of inflation in India.

In a recent paper on inflation measurement in India,my colleagues and I examined detailed price data,expenditure patterns of households and the composition of different price indexes available in India.

Internationally,“producer price indexes” are widely used. These reflect “factory-gate prices”,valued from the producer’s perspective. In contrast,the Indian phrase “wholesale price” may record prices paid at various stages of the distribution chain: starting from prices of raw materials for intermediate and final consumption,or prices of intermediate goods,to prices of finished goods up to the retail stage. Furthermore,the price data used in the WPI may sometimes contain discounts and rebates,taxes and subsidies on products,as well as trade and transport margins. In addition,the weights in the WPI do not reflect the consumption of the average household.

Increasing trade integration coupled with the domestic liberalisation of administered prices has turned a growing fraction of the WPI basket into tradable goods,whose prices are determined in international markets. The evidence shows that the WPI tends to move with producer prices of other countries as a consequence of the substantial share of tradables in the WPI.

The domestic WPI is thus strongly influenced by the fluctuations of global prices of tradables and the fluctuations of the rupee. Domestic monetary policy has no impact on global tradable prices. In addition,now that India has moved towards a flexible exchange rate policy,domestic monetary policy does not involve an administrative control of the exchange rate.

There is a sharp contrast between wholesale price inflation,where it is seen that a range of countries has similar tradables inflation,and consumer prices which show a marked divergence of consumer price inflation across the same countries.

For these reasons,the WPI,while continuing to be a valuable source of data,should be de-emphasised in the discussion of inflation outcomes. The central bank should focus on the unique features of each domestic economy,rather than on the common factor of global tradables inflation.

To gauge the extent of the information delays in the CPI basket,we compare it with the CMIE Consumer Pyramids,a data set drawn from a panel data set where over 120,000 households are surveyed each quarter,for which we have a detailed breakdown. The weight of food in the Consumer Pyramids data set is similar to that seen in the CPI-IW. Within food also,we find that the distribution of expenditure is not too dissimilar across the two sets of weights. This increases our confidence in the CPI-IW.

The national CPI,which is going to be soon released,will measure both urban and rural consumer prices. It will be based on the latest weights in the most recent 2004 round of the National Sample Survey of household expenditure. We may expect that these changes will further increase confidence in the CPI.

Two other issues in data quality of the CPI are food prices and the prices of services. When food prices in the CPI are compared to those from other sources like the ministry of agriculture or data collected by NCDEX,it is found that the data in the CPI broadly corresponds with the information from other sources. This inspires further confidence in the CPI.

Second,while services account for half of India’s GDP and a large share of household consumption expenditure,there is no price index for output prices in this sector,neither at the consumer nor at the producer level. The only price series available for some services are those that have always been routinely collected in the CPI surveys (in particular for the CPI-IW).

Confidence in the information content of the CPI-IW suggests that the acceleration in year-on-year inflation beyond 5 per cent from early 2006 onwards should be seen as a serious problem. The problem of high and volatile inflation should not be downplayed on the grounds that it is based on low quality information.

The Central Statistical Office’s plans of releasing a new CPI series for India this month lends fresh salience to this question. This new CPI is likely to become the best

option for a headline inflation indicator,through significant improvement upon existing price indexes in terms of representation,quality of price collection and weighting. The release of this new CPI is a natural opportunity for RBI to de-emphasise other inflation measures and

focus on the new CPI.

The writer is a professor at the National Institute of Public Finance and Policy,Delhi,express@expressindia.com