Consumer Price Index Numbers for Industrial Workers-Frequently Asked Questions

Consumer Price Index (CPI) is designed to measure the changes over time in the level of retail prices of a fixed set of goods and services consumed by an average family of a defined population group in a given area with reference to base year. It does not measure costliness of a place as the same involves quantity as well as price but in Consumer Price Index only price changes are reflected and not quantity changes. However, this aspect could be taken care of to some extent, by revising the base year of the index at short intervals, so that relative importance of items in the family budget of the given population based on latest consumption pattern is reflected in the index.

Consumer Price Index is different from Cost of Living Index as it does not take into account the level of living i.e. level relating to health, education, working conditions, service conditions, housing conditions, indebtedness, savings etc., of the population.

The Consumer Price Index differs from the Wholesale Price Index in the sense that WPI is based on value of domestic production, imports and wholesale prices whereas Consumer Price Index is based on average expenditure reported on the items consumed by given population and retail prices.

In India, Consumer Price Indices are compiled for five different segments of population namely industrial workers, agricultural labourers, rural labourers, urban population and rural population. The first three indices are being compiled by Labour Bureau, Shimla and the last one by Central Statistical Organisation based at New Delhi. The Consumer Price Index Numbers for Industrial Workers (CPI-IW) is being compiled for the industrial workers employed in any one of the seven sectors viz.

• Factories
• Mines
• Plantations
• Railways
• Public motor transport undertakings
• Electricity generation and distribution establishments
• Ports and docks.

A working class family has been defined as one wherein one of the family members work as a manual worker in one of the 7 sectors as listed above and which derive one-half or more of its income through manual work.

The Consumer Price Index Numbers for Industrial Workers (CPI-IW) for 50 centres and All-India on base 1960=100 was started w.e.f. August, 1968 index on the basis of Weighting Diagram drawn by conducting the Family Living Survey (FLS) in 1958-59. The old series (1960=100) was replaced by a new series with base 1982=100 with effect from October, 1988. This was further updated and a new series with base 2001=100 has subsequently replaced the series on base 1982=100 with the release of index of January, 2006 and still continuing.

Keeping in view the main use of Consumer Price Index i.e., regulation of dearness allowance of the working class at local as well as national level, it was decided to compile the index at centre level as well as at all-India level. Therefore, at the time of selection of centres not only the importance of a particular centre was considered from the point of view of industrial employment but its representativeness in the all-India industrial employment was also given due importance. The present series (2001=100) of Consumer Price Index Numbers covers 78 important industrial centres as against 70 centres covered in the 1982 series of Consumer Price Index Numbers. These centres were firstly allocated among the factory, mining and plantation sectors in proportion to the total employment in each sector in the country. Secondly, the number of centres allocated to each sector was distributed amongst States on the basis of industrial employment in a state subject to a maximum allotment of 5 centres in a state in a sector. Lastly, the actual centres were selected on the basis of centrewise industrial employment in consultation with the respective State Governments. It may, however, be mentioned that after the centre was selected, workers belonging to all the sectors were covered for this survey as against one predominant sector in a centre during 1960 series. Keeping in view the resources available, it has not been possible for Labour Bureau to cover larger number of centres and some old centres had to be dropped to select new developing centres. This issue had been considered by Technical Advisory Committee on Statistics of Prices and Cost of Living (TAC on SPCL) also and they decided that other centres if found useful, may be covered by the respective State Governments.

Firstly, it is essential to find out different items being consumed by the population and the relative importance of those items in their budgets. For this purpose a detailed family budget enquiry is conducted amongst the industrial population following well established sampling techniques at regular intervals. The Labour Bureau has so far conducted four such enquiries during 1958-59, 1971-72, 1981-82 and 1999-2000. These enquiries involved sample size determination, selection of workers’ families, canvassing of schedules, scrutiny of schedules, tabulation of results, preparation of weighting diagrams etc. On the basis of such enquiries, a list of items normally being consumed by the working class is prepared for regular collection of prices as well as index compilation. The second requirement of Consumer Price Index is base year prices for each of the items listed for price collection. This set of prices is collected after organising the price collection work in all the centres through selection of markets, listing retail shops in the markets, fixing specifications and units, fixing price collection day and time for one full year which when averaged over months provides base year prices. Thereafter, prices for listed items are collected regularly every month for the compilation of index numbers. From the results of working class family income and expenditure survey, proportion of expenditure on each of the items consumed is worked out and is utilised for assigning weights to different items. Thus, an index is a weighted average of the relative price change of the list of items included in the index.

The centre-wise sample size was determined so as to provide sufficiently precise estimate of the consumption pattern of the population for a given centre. For this purpose, the centres were divided into two categories namely, centres which were common with the 1982 series and those which were new centres. For 69 centres common with 1982 series, subsample- wise weighting diagrams were prepared and using centre level price data for 36 months period, i.e., from January, 1993 to July, 1995 sub-sample-wise consumer price

indices were compiled. For each set of price data, sub-sample indices were used to work out the co-efficient of variation (C.V) and the average of these 36 C.V. estimates was calculated. The sample size was then fixed for each of the 69 centres.For the new centres, sample size was determined after taking into consideration the working class population at these centres and C.V. observed at a nearby or similar centre. The minimum sample size for a centre was fixed at 216 for Schedule ‘A’ (Income & Expenditure) and 84 for Schedule ‘B’ (House Rent) for operation convenience. The sample size for other centres was adjusted to multiples of 216 and 84 respectively to achieve equitable distribution of work amongst field investigators. The sample size varied at different centres from 216 to 3024.

Preliminary enquiries were conducted by National Sample Survey Organisation (NSSO) during the year 1999 with a view to define the boundaries of each centre, examine the feasibility of adopting tenement or payroll method of sampling and collect information regarding concentration of working class population of different sectors, availability of addresses of the workers with the establishments, etc. Basically, in each centre the working class families were selected through standard sampling techniques in two stages. In the first stage, a sample of blocks (areas or localities having 150 to 200 houses) or a sample of establishments was selected. In the second stage, a sample of workers’ families was drawn from the selected blocks or establishments. Tenement method of sampling was preferred only in those cases where payrolls/addresses of working class were not maintained properly by the establishments and workers of a particular sector were concentrated in a limited area. Efforts were also made to draw the samples in such a way as to give representation to different characteristics of population having bearing on consumption pattern such as state of origin, expenditure class, single member families, employment, type of ownership etc., wherever possible. For this survey, family was defined as consisting of persons generally related by blood, marriage or adoption, usually living together and/or are being served from the same kitchen and pooling a major part of their income and/or depending on a common pool of income for a major part of their expenditure. Relatives or friends who were absent temporarily were treated as family members. Servants, temporary guests and paying guests who pay a fixed amount for boarding, lodging were not considered as family members in brief, family was taken as the sample unit and not the household.

Mainly two schedules, one family budget schedule or Schedule-A and the other House Rent Survey Schedule or Schedule-B1, were canvassed amongst the working class families selected for the survey. In the Schedule-A, besides identification particulars of the family, the information on demographic particulars of the family such as sex, age, marital status, education standard, economic status, activity status, industry/occupation etc., sources of income of the family such as family members’ enterprise, paid employment, selfemployment, net rent, other incomes, other receipts were collected, on the one side. On the other side, details of the family expenditure incurred not only on consumption items such as cereals, pulses, oils, fats, meat, fish and eggs, milk and related products, condiments and spices, vegetables and fruits, sugar, honey, non alcoholic beverages, other intoxicants, fuel and light, house rent and related charges, repairs, clothing, bedding and footwear, medical care, personal care, education and reading, recreating and amusement, transport and communication, personal effects, washing, furniture, household appliances, expenditure incurred on ceremonies but also on non consumption items such as sub ion, taxes, litigation, remittances sent to dependants, savings and investments loans repaid, etc. were collected. In addition, details relating to indebtedness of the working class families such as purpose of loan, amount of loan, source of loan, rate of interest paid, nature of security given etc. were collected. In the Schedule-B1, besides identification particulars of the family, details of the dwelling in which the family resides such as location, type of building, number of living rooms, provision of kitchen, bath room, latrine, varandah, courtyard, drinking water source, furniture, other appliances, rent particulars, category of house such as rented/owned/ free, particulars of tenancy expenditure incurred on electricity/water/sweeper charges, local taxes, repairs, etc., were also collected. Details collected in the Schedule-A formed the basis of weighting diagrams and the details collected in Schedule-B1 formed the basis of 6 monthly house rent surveys for determining variations in rents over time.

The responsibility of fieldwork was given to trained staff of NSSO. They were supplied detailed instructions on each of the items given in the schedules. In addition, zonal training, conferences were convened by Labour Bureau for the officers of NSSO where points of doubt were also clarified. The NSSO officers conducted the primary supervision of the fieldwork, which was supplemented by Labour Bureau officers. Besides, the detailed scrutiny of the filled-in schedules was also done by Labour Bureau as regards inconsistency in specified codes, total and sub-totals. The results were also examined from the point of view of prevailing prices and results of earlier surveys. All inconsistencies were reconciled with NSSO before finally adopting the data. The expenditure details wherever relating to households (i.e. family & owners staying with the family) were adjusted for the family by multiplying with the factor f÷(f+e) where ‘f’ stands for family members and ‘e’ stands for other members of the household. In this connection, it may be mentioned that expenditure reported on ceremonies was suitably split into two parts i.e., for family members and nonfamily members. Only the portion of expenditure relating to family members was accounted for in the index against relevant items. The survey was spread over 12 months so as to account for seasonal variations in expenditure of working class families. A detailed tabulation plan was then prepared and given to Regional Computer Centre (RCC), Chandigarh along with instructions for each entry in the tables for tabulating the data as per the approved estimation procedure. This involved blowing up of sample results to entire population using sampling fraction. The tables received from RCC were also checked for inconsistencies by Labour Bureau before adopting them for preparation of weighting diagrams.

The list of items on which working class families reported expenditure were divided into two categories namely consumption items and non-consumption items. The nonconsumption items included gifts and charity, ceremonies, fines and penalties, trade union sub ion, income tax, road tax, municipal tax, interest paid, litigation, remittances to dependants, precious ornaments, land and building, live stock, life insurance premium, provident fund contribution, savings, chance game and lottery, revenue stamp, loans advanced, investment in shares and securities, etc. All other items were normally treated as consumption items and were included in the weighting diagram of the concerned centre. However, some items which could not be satisfactorily priced overtime were not directly retained but imputed to related items such as other cereals, mixed pulses, prepared meals, dung cake, readymade garments, library charges, hostel charges, pocket expenses, etc. There were some other items on which expenditure reported was negligible and number of families reporting expenditure were not significant, such items were also imputed to other related items. The total expenditure on consumption items was divided into 6 main groups namely Food; Pan, Supari, Tobacco & Intoxicants; Fuel & Light; Housing; Clothing, Bedding & Footwear; and Miscellaneous. The Food group was further subdivided into 8 sub-groups namely Cereals & Products; Pulses & Products; Oils & Fats, Milk & Products; Meat, Fish & Eggs; Condiments & Spices; Vegetables & Fruits; and Other Food. The Miscellaneous group was also sub-divided into 5 sub-groups namely Medical Care; Education, Recreation & Amusement; Transport & Communication; Personal Care & Effects; and Other Miscellaneous Items. For imputing items to related items, the method used was to add the expenditure to the items having a broad similarity of price behaviour, satisfying similar want, produced under the same manufacturing process, etc. Where it was not possible to impute expenditure to a particular item, it was imputed to a group of items called section, sub-group or group. This was done on the assumption that group of items taken together reflect the price behaviour of item(s) imputed. All items which had at least one percent expenditure reported in the sub-group/group expenditure and/or had significant families reporting expenditure, could be priced over time satisfactorily were directly retained in the index. It has been found that each centre has more that 100 items directly retained for price collection. The percentage expenditure on each item in the sub-group/group represents its weight. Similarly, the percentage expenditure on sub-group/group in the group/total consumption expenditure represents their weight. The entire set of weights so prepared is called weighting diagram of the centre. It may be clarified that with change in coverage and changes in consumption pattern, relative weights of different groups cannot be expected to be same in old and new series.

No difference was made between durable or non-durable items. The items were included in the index if they satisfied the criteria given above. Similarly no distinction was made between essential and non-essential items. Intoxicant items were included in the index, if they were preferred by the working class in a centre and there was no prohibition in force on sale of these items.

For selection of markets, Labour Bureau officials visited all the centres and contacted representatives of employers organisations, workers unions and concerned State Governments officials to identify areas of working class concentration and retail markets catering to them. Thereafter, these markets were visited to verify availability of various items being consumed by the working class on a regular basis. After these studies, such a number of markets were selected which represent purchases of larger segment of the working class of the centre. The markets once selected are not changed in the life span of an index series to maintain comparability of current prices with base prices. In each of the selected markets, two shops were selected for each of the items retained in the price schedule so as to account shop to shop variations of the markets. In addition, two more shops were selected and kept as reserve so that prices could be collected from these shops in case of non-availability of any item from selected shops. Keeping in view the earlier experience about frequent nonavailability of clothing varieties, it was decided to select retail mills shops wherever possible for price collection of clothing items on the presumption that price trend revealed by these shops will represent price trend of the clothing items of the centre. For collecting details relating to ration items, four ration shops were selected from each market provided they were available. While selecting the shops, factors like co-operation of the shopkeeper, regular availability of the stocks, predominantly retail transactions was also kept in view. The shops once selected are changed only in rare cases such as change of business by shopkeeper to avoid spurious variations in prices.

For each of the selected items, normally, one variety was selected for regular price collection on the presumption that it will reflect price trend of all the varieties of the item. Efforts were made to fix variety in a detailed manner so that it could be easily identified by the price collector over time. Similarly, for each item a unit was fixed after consulting shopkeepers and local people who could represent majority transactions of the working class. The list of items retained for price collection had been divided into 3 main categories, on the basis of frequency of price collection namely, weekly, monthly and six monthly. In the weekly price schedule items such as cereals, vegetables, oils, etc., are covered where prices are sensitive and vary frequently. Prices of items like cinema, furniture, utensils, household appliances, transport etc. are collected once in a month as their prices do not change very frequently. Prices of all types of clothing and footwear items covered are also collected once in a month for the same reason. The prices of items like house-rent, school/college fees, school/college books are collected once in six months/year. For all the items prices are collected on a fixed price collection day and time (in case of perishable items) every week/month. The day for this purpose had been fixed after considering the extent of transactions taken place on various days in the week and views of the State Governments.

The State Government employees who are locally posted and normally belong to Department/Bureau of Economics and Statistics or Labour Department collect these prices from the selected markets. For each market, a separate Price Collector has been appointed who does this job on part time basis and receives honorarium from Labour Bureau for this work. They have been provided detailed instructions in this regard. The middle rank officers of the State Governments under whom the Price Collectors work have normally been appointed as Price Supervisors for a group of markets. These Price Supervisors have been instructed to pay visits to the selected market to keep watch on work of Price Collectors and send his/her detailed monthly report directly to the Labour Bureau. Over and above this arrangement, Labour Bureau has posted field staff in its four Regional Offices at Ahmedabad, Kanpur, Kolkata, Chennai and one sub-regional office at Mumbai, who regularly visit the markets, guide the Price Collectors and Price Supervisors in price collection work, verify prices from the concerned shop-keepers and send their detailed reports to the Headquarter at Shimla. At the Headquarter, the prices received are examined for inter-period, inter-shop, inter-market and inter-centre variations. All doubtful points are referred to the field for verification and confirmation. Thus, it would be clear that three levels checking of the prices are done before finally adopting them. Moreover, prices are adjusted for quality as well as quantity variations with reference to the specifications fixed in base year before utilising them in the index. Prices for some essential items are also published in the monthly publication of Labour Bureau, the Indian Labour Journal. In addition, it has been made known to the index users that they can visit Regional Office of the Labour Bureau for verifying the prices collected at any time. But the lists of shops from which prices are collected are not made known to the users so as to avoid non-cooperation of shop-keepers and industrial disputes of workers and employers on the prices collected. For the same reason, representatives of workers and employers are not associated in price collection work. All methodological issues relating to utilisation of prices are invariably decided in consultation with the TAC of Government of India.

There are two situations prevailing in the country at present. In some centres of West Bengal such as Kolkata, Howrah, Statutory rationing is in force i.e. the sale of ration items in the open market has been banned. In such cases, only ration prices are utilised in the index and no account is taken of unauthorised or black market prices. However, when supply of cereals through ration shops falls short of the requirement of an average family, corrective action as approved by TAC is applied which involves distributing the portion of weight of the item-corresponding to short-fall among all other food items pro-rata. In many other centres, informal rationing is in force which implies that ration items are available freely in the open market also, in such cases, so far Labour Bureau had been giving weight to the ration price on the basis of admissibility of supplies by rationing authorities but for the new series, the ration prices are weighted with the actual availability of the ration items in selected ratio shops. This is an improvement over the old practice as it better reflects actual consumption of working class and supplies actually made available by the Government. For working out actual availability, opening balance at the beginning of the month is added to the quantity received during the month for each ration item and divided by number of units registered with the ration shop. This availability provides the weight for ration price and rest of the requirement of an average family provides weight for open market price. These weights are worked out at the shop level and then averaged to get market level weights. For each market, separate weighted price is worked out for each ration item. These prices are then averaged to get the centre-weighted prices for each ration item. No adjustment is made in ration price for quality variations with open market price as it is not possible because of involvement of subsidy element.

The practice under the old series on base 1960=100 was to revise the index for housing group twice in a year i.e., in January and July on the basis of half-yearly house rent survey conducted in the preceding half-year in respect of rented houses only. But housing index was being kept frozen at 100 in mining and plantation centres in view of predominantly free and owned houses there. However, in the subsequent series on base 1982=100 and 2001=100, actual rents of rented houses and comparable rents of owned houses are collected in the course of half-yearly house-rent surveys and zero rent issued for free houses. For compiling the housing index, three separate indices are compiled for free, rented and owned houses which are combined using weights, which are proportions of families residing in three categories of houses. For half-yearly house-rent survey, a sample of dwelling is drawn from the sample covered under the family budget enquiries a nd is normally kept fixed in the life of the series. In no centre less than 84 dwellings have been selected for these surveys. The housing index is compiled by Chain Base Method, in which rent movements are compared with the last six monthly periods and not with the base period as this method better takes care of depreciation aspect of housing. Expenditure incurred on minor repairs of the sample dwellings by the workers is also accounted for working out rent index. New houses are not included in the sample as it is felt that relative movement of rents in new houses is not going to be different from that of old houses. Moreover, coverage of new houses in the sample can result in operational problems.

A particular fruit/vegetable is priced during the availability period only which is determined after studying the prices of two to three years from different markets of the centre. For off-season months, the weight for a particular item is distributed to the corresponding fruits/vegetables section in the months concerned. Therefore, in respect of fruits and vegetables the weighting pattern change every month, depending upon their availability and thus avoiding the need of estimating prices for the off-season months. This is also an improvement over the earlier practice where only such items were priced which were available throughout the year. For some other items like cold drinks, maize atta, umbrella, blanket etc., which are also sometimes not available throughout the year, no monthly pattern has been prepared but the problem is dealt with in the same manner whenever the prices are not available for these items

In case an item is supplied free to entire working class in a centre, zero price is used in the index. However, if a proportion of working class only gets an item free then a weighted price, with proportions of working class getting the item free of cost and paying market price as weights, is utilized in the index. No attempt is made to estimate the prices of free items or impute their weights to other items, as it will not reflect the real situation appropriately. In case an item was free during base period, its price relative is kept frozen at 100 if it is supplied free now.

It has been found by economists that Laspeyre’s Formula when used for an index series for a long period results in upward bias in the index. Further, changes in consumption pattern over time necessitate revision of weighting diagrams at short intervals to the extent possible. But in India, working class family income and expenditure surveys are generally conducted after an interval of 10 years, as it had not been feasible so far to conduct such surveys at the shorter interval. Normally, a year during which the family budget enquiry is conducted or a period not very distant from survey period for which reliable price data are available is adopted as the base year. A period affected by developments of serious nature such as war is not adopted as the base year because it cannot be treated as a normal year economically.

An all-India index is worked out as a weighted average of 78 centre indices. The weight assigned to each centre is worked out as a proportion of the consumption of expenditure of estimated number of families allocated to a centre in the State to sum total of all such expenditure over all centres in the country. For this purpose, each centre is presumed to represent equal share of the working class families in a State. However, if the actual number of families in a centre exceeds the assumed share, its share is taken as the actual number of families. For deriving the number of families in a State, the average daily employment in respect of the sectors covered for the survey is divided by the average number of earners in the State. It may be mentioned here that with the change in number of centres in a state, change in average consumption expenditure and number of families represented by a centre, centre-weights in all-India index are bound to be different in old and new series.

Normally, arithmetic method which involves ratio of the old series indices to the new series indices for the base period is utilised for working out the linking or conversion factor of the two series. However, ratio method which involves average of the ratio of monthly indices of the two series for the base period has also been tried in the past and it gives similar results as the arithmetic method. A period longer than base period or later than base period is normally not preferred because it can affect continuity of two series. However, at times due to administrative and operational difficulties the release of new series is inordinately delayed and in the meantime the old series is continued. In such cases, two series are linked using ratio of averages of indices for the 12 months proceeding the month of introduction of new series. This latter approach was followed for linking 1982 series with 2001 series which was introduced with January, 2006 index. Generally, the two series are linked as such, i.e., without any revision of the old series due to methodological improvements in the new series. It may be added here that two series cannot be expected to show similar trends because of differences in coverage, methodology, changes in consumption patterns over time, etc.