Measuring inflation accurately is vital to policymakers. The SBP designs monetary policy in accordance with the expected inflation rates along with other determinants.

When banks lend money, lending rates are adjusted with inflation rates and T-bills' rates move with inflation rates. The Federal Bureau of Statistics (FBS) calculates and releases four different price indices to measure inflation.

Most widely used and quoted index is the Consumer Price Index (CPI). CPI covers retail prices of 375 items in 35 major centres and shows the cost of living in urban area.

The Wholesale Price Index (WPI) is used to calculate price changes of a basket of 425 items at wholesale level prevailing in the cities of origin of the commodities. The Sensitive Price Indicator (SPI) measures changes in 53 kitchen items consumed by low-income households whose monthly income is up to Rs3000.

In the West, CPI is used not only to index wages but index pensions and social security benefits also. CPI data helps governments to adjust official fees and charges, payments in commercial contracts, and for assessing and formulating trade and exchange rate policies.

While the CPI is considered to be the best indictor of inflation, it is not free from criticism in Pakistan because of the weaknesses in measuring inflation. These are identified as follows:

In all, 375 items have been included in the CPI basket of goods and services with weights assigned to individual commodities based on the Family Budget Survey conducted in 2000.

Food, Beverages and Tobacco constitute major portion at 44 per cent of the basket, followed by House Rent at 21.69 per cent and Fuel and lighting at 8.10 per cent.

The weights are held constant normally 10 years and during this time the buying habits and quantity bought by consumers are not properly reflected in the CPI. Also relative price changes are not reflected in the index. This may overstate or understate the index.

The index fails to acknowledge the quality difference between two identical or standardized products. For example, why price of Sony TV is higher than LG or any other brand? Also various outlets in a city offer discounts that are not recorded by FBS field staff.

The CPI uses a fixed market basket, but many new products have been brought to the market during 1990-2000. For instance, computers back in 1990 were not in high supply and prices averaged around Rs40,000 for a piece.

The price of a computer in 2004 has fallen on average to Rs12,000 but the CPI fails to indicate why prices have fallen. This leads us to suggest that the price decline, common in the product cycle, particularly for durable goods, never shows up in the CPI.

The CPI shows movements in inflation only with a considerable time lag and portrays an artificially smooth path for the running inflation rate as International Labour Organization (ILO), reports.

The CPI takes into account 375 items and that too focuses on buying habits of urban population. Items like TV, VCR, computers, washing machines, carpets, D-Freezers, ACs, household servant, tissue papers, toilet paper rolls are included in the CPI basket of goods and services.

In practice, a very high percentage of population living in small towns and semi-urban areas is unable to afford many of these items. Inclusion of such commodities in the basket may not truly reflect the impact of inflation rate on the population living in those areas which are considered semi-urban or small towns.

The basket of goods and services used to compile CPI doesn't include the price of land, bonds, shares, or other marketable securities purchased by households. In theory, if markets are perfect relative prices should reflect relative costs of production and relative utilities.

For example, one grocery store in Islamabad market sells a product at quite different prices than the other. This aspect is not reflected correctly in the CPI where the average price of various markets for a commodity is taken.

The CPI does not record the price discrimination in some cases--different prices being charged from different customers. The coverage of CPI is limited to selected markets only, hence prices may vary within the city or between two cities.

The CPI is supposed to be a price index that measures changes in the prices of consumption goods and services acquired by households. Only cities (65 per cent of urban population) are covered in computing CPI and majority of rural areas like Gilgit and Skurdu have been ignored.

A separate CPI is needed to be computed and released every month under the title "CPI-Rural (R)". This will help policymakers to assess the impact of a change in the CPI for rural dwellers correctly where more than 60 per cent of poverty-ridden population live and work.

In Pakistan, smuggled items are openly sold in markets without paying any tax or duty to the government. The CPI doesn't distinguish between the illegal and legal goods. In many cases the basket might understate or overstate the inflation rate.

Statistical theory suggests that the sample size and sample selection methods for both outlets and goods and services for which price movements over time are to be observed should ensure that the prices collected are representative and sufficient to meet the requirements for the accuracy of data.

FBS selects cities following the stratified random sampling method. Little is known as to what method the FBS has adopted for selection of outlets/markets to collect price data.

If it is a purposive selection (biased sampling) as happens to be the case, then it may be cost effective but results may not truly reflect the price trends. Random sampling may become too costly and too difficult but results obtained may reflect the true picture of price trends.

The selection of sample has been much debated in the literature and the selection decision is dependent upon the resources available to the agency/department.

The United States Bureau of Labour Statistics makes extensive use of random selection procedures to select both outlets and products within outlets. The study suggests use of random sampling by the FBS in Pakistan for collection of price data.

Timing of price collection plays an important role in making the statistics reliable. In case of perishable goods, prices need to be collected on the same day of the week and not just before closing time, when stocks may be low, or sold cheaply to minimize wastage.

Sale prices, discounts cut prices, and special offers should be included when applicable to all the customers. Use of computer technology must be exploited to the fullest extent possible.

If budget permits the FBS may use potable computers and transmit their results electronically to the central office that could also help cover remote areas. Even use of mobile phone (MSN) may be another option in this regard. The FBS may also wish to get its data validated by a competent third party. This will ensure transparency in the system.

There is a strong feeling among the public at times that the official data is fudged. There must be public confidence in the data reliability. The FBS releases CPI data every month.

However, effective efforts appear to be lacking to educate the public properly about the methodology adopted to compile CPI. Even FBS website http://www.statpak.gov.pk/depts/index.html does not share the methodology used to compute CPI. The method used to compile CPI data must be documented, transparent, and open to public scrutiny.

Table I indicates that in the developed world, the CPI is measured on monthly basis except Australia where data is released on quarterly basis. The base year and weights are revised after every 5-year in most of the cases.

Number of items varies from country to country depending upon the tastes and preferences of households and standards of living. Interestingly, in the UK a private sector company collects domestic prices. Pakistan's performance has not been that bad when we compare its CPI methodology with other countries.

Table II shows that Pakistan has more updated weights and base year as compared to India. Not much difference is noted in number of items. Number of centres in India (59) is not that great as compared to Pakistan (35), given the geography and size of population of both countries. However, India has four CPIs whereas Pakistan uses only one CPI.

Like other countries, inflation in Pakistan is also measured on the basis of the CPI. Inflation has been a crucial macroeconomic problem for most countries of the world and Pakistan is no exception. There are, however, certain problems associated with measuring inflation through the CPI.

First, the CPI is based only on a small fraction of the many goods and services that are available in an economy. In case of Pakistan, 375 items are included in the basket as compared 650 in Canada, 741 in Singapore, 596 in Japan, and 650 in the UK.

If a person's buying habits differ substantially from the market basket on which the index is based, that person may experience a very different change in his cost of living than what the CPI shows.

The weights of basket of CPI are derived from the Family Budget Survey, which, in most cases, are revised after every 5-year period. In Pakistan, however, weights are revised after 10-year period.

Coverage of cities and small towns excludes that northern areas and rural areas and only 65 per cent of urban population is surveyed. Sample size be enlarged enough to give due coverage to rural population and leftover areas.

The study recommends inclusion of northern and rural areas in the sample size of price data collection centres. Buying habits, tastes and preferences and standards of living in rural areas differ substantially from the urban class.

It's the rural population, which is facing poverty in most instances, and therefore any changes in prices affect rural population more than the urban population. The FBS may also wish to consider releasing a separate CPI called the CPI-Rural which is more reflective of buying habits of rural population.

(Article written by a senior faculty member and the deputy director SZABIST, Islamabad Campus.)