In India, inflation has always been one of the most closely monitored macroeconomic indicators. In recent times, double-digit inflation has made headlines and the rising food prices have been a cause of concern for policy makers. The inflationary trends highlight the need to have an appropriate price index. Till recently there were two indices to measure inflation, the wholesale price index (WPI) and the consumer price index (CPI). WPI is more widely used because it has a wider coverage of goods and is available on a weekly basis for primary articles. We have two distinct CPI numbers — for industrial workers (CPI-IW) and for agricultural and rural labourers (CPI-AL/RL). A new series of WPI was launched in September, 2010 with the base year 2004-05.
Recently the Central Statistical Organisation (CSO) also released a new series of consumer price indices for all-India and states and Union Territories, separately for rural, urban and combined. The new series includes some services under “housing”, namely education, medical care, recreation & amusement, transport & communication and personal care & effects. The detailed methodology, however, has not been made public. The release of CSO’s series begs some questions. First, given the multiplicity of price indices, which price index should be used as deflator? Second, with CSO now releasing CPI, what is the significance and relevance of CPI released by the Labour Bureau? Also, we need to note that CPI usually has two different series for urban and rural consumers. Thus, it is plausible to argue that the difference in the consumption basket of the urban and rural population shall be more pronounced in case of services than in case of agricultural and manufacturing items. Also because “services” are expected to have a higher income elasticity of demand, it is likely that the kind and quality of services available in urban and rural areas are very different.
The other two indices mentioned above comprise only the tangible goods and ignore the intangible services. A price index without services would not reflect the true cost of living because services contribute to more than 50 per cent of the GDP. For example, an increase in fuel prices gets captured in WPI and CPI, however, the consequent increase in transport fares does not get captured and hence the two indices paint an incomplete picture. A way out can be a “goods and services price index” (GSPI)that incorporates goods and services as a composite price index. GSPI should be a true cost of living index and it should reflect the expenditure required to maintain a certain standard of living. Compilation of GSPI will certainly be difficult because of several reasons. First, collection of price data for services is problematic. Services are highly customised and hence it’s extremely difficult to get a representative price for any service. For example, in a city like Delhi the consultation fee of a doctor varies from Rs 50 to Rs 1,000 for a visit. There is also a problem of administered or regulated prices with respect to certain services like railways and postal services. In case of railways, prices usually change only with the annual Railway Budget. There are sectors in which “price” per se is difficult to identify. For example, in the banking sector, financial intermediation is the primary role of the banks. But the question arises, what is the price of “financial intermediation”? Theoretically, interest rate is the cost of the capital and hence we should subtract this from earning on loans to arrive at price for advances. The concept paper on “Banking Service Price Index” prepared by the Office of the Economic Adviser (OEA), Ministry of Commerce & Industry has sought to address this issue by the proposed “reference rate”. The concept paper estimates this price as loan price = interest rate received on loans - reference rate and deposit price = reference rate - interest rate paid on deposits. The issue of what should be the reference rate is yet to be resolved.
Now to come to technical issues. There is always a possibility of a “new outlet bias” if we use Laspeyre’s index (as is the case of WPI and CPI) due to rotation of outlets or units into and out of the index sample. At the time of rotation, any difference in the price between items in the old outlet and the new outlet is implicitly assumed to reflect a difference in quality. Further, the “qualitative change” that we experience in services is difficult to capture. Also, since the product basket in case of services would change rapidly, it would necessitate releasing of new or revised series (with a new base) more frequently. As for weighting diagram, traditionally it is devised by assigning weights to various products in accordance with their shares in the total output or revenue. But in case of services we may not be able to capture the disaggregation at every level, for instance, in case of healthcare it would be appropriate to look at the consumption expenditure data of NSSO while deciding on the weights. Weighting bias in case of services is also likely because these fall largely in the unorganised sector.
The Paasche index is another index that is not so widely used. It uses quantities in the current year for devising the weighting diagram. It has limitations of understating inflation. The effect of substitution would be that a greater importance would be placed on the goods that are relatively cheaper in the current year. The Paasche index also requires information on current quantities, which is not feasible. The Fisher’s index, a geometric average of Laspeyers’ and Paasche Indices, is theoretically an “ideal” index because it has no substitution bias and it has the time reversibility property. It is also consistent with the revealed preference theory. Fisher’s Index also comes closest to Robert Pollak’s social cost of living index, which is the ratio of the total minimum cost or expenditure required to enable each of the households present in the two periods to attain their reference utility levels in the two time periods*.
A step forward in this direction has already been taken in the form of the development of business service price index (BSPI) scheme. The idea was mooted when the Rangarajan Commission set up by the Ministry of Statistics & Programme Implementation in its report in 2001 had recommended that a separate service sector index should be developed initially as a component to WPI and later to be merged with WPI once it attains some stability. The present scheme of BSPI of OEA covers ten services and sectors, namely banking, trade, business services, postal, telecommunication, air transport, port services, insurance, rail transport and road transport for development of experimental service price index. Surprisingly, the most dynamic services – information technology and information technology-enabled service – are not part of this basket. OEA has already placed concept papers for railway service price index and banking service price index in public domain. One must view both these indices with caution as these prices are largely administered prices. Nevertheless, it is a step in the right direction. Finally, when OEA merges WPI and BSPI, it may come up with an index close to suggested, yet elusive to GSPI.
*W Erwin Diewert — Index Number Issues in the Consumer Price Index, Journal of Economic Perspectives, Winter 1998
The author is an officer of the Indian Economic Service. The views expressed are personal. The author is grateful to Dr Kaushik Basu, chief economic adviser, Ministry of Finance for his comments on an earlier draft