Specimen Data Sheet

The data are presented in each national, zonal, state, and district pages. The data definitions are comparable across all geographical units and are discussed in the following pages. Different parts or sections of the data pages have been assigned letters from ‘A’ through till ‘K’. The relevant data definitions in the following pages are presented next to these letters.

1. Explanations

The district market size is defined as the annual total expenditures of all the households in the district estimated for the year 2003-04 in current rupee value. (The unit is Rs Crores, which equals Rs 10 million.)

The market size of the following broad categories of items are also presented:

• Food Products: Includes high value food items (milk, milk products, ghee, edible oils, egg, fish, meat, and their products, vegetables, fruit, nuts, pan, tobacco, intoxicants and other food items such as sugar, spices, beverages such as tea, coffee, cold drinks, processed food such as biscuits, namkeen, etc.) and basic food items (cereals and cereals product; pulses and pulses product).

• FMCG products: Fast Moving Consumer Goods are considered to include toiletries, soaps, cosmetics, teeth cleaning products, shaving products, detergents, other non-durables such as glassware, bulbs, batteries, plastic goods such as buckets, etc.

• Durables: Includes furniture, utensils, ornaments, kitchen equipment, vehicles, clocks, watches, cassettes, records, TV sets, radio and other household appliances. It also includes repairs and maintenance expenses though these are minor shares of total durables expenditures.

• Miscellaneous Goods and Services: Includes clothing and footwear, fuel and light; cinema/theatre/video show; tuition fees; newspaper, magazines, fiction; regular and other journeys; house rent; school books and educational articles; non-institutional medical expenses; institutional medical expenses such as hospital, nursing home; and other miscellaneous goods and services.

The market size, or total expenditures, on these items are calculated as the total annual expenditures of households within the district. The market sizes have been calculated from National Sample Survey Organization’s (NSSO) surveys (discussed later).

Briefly however three adjustments were done. First data from expenditure and employment round of NSSOs large sample surveys were combined using methodology developed by Dubey and Bhandari (2003). Second, the estimates were adjusted for expected growth till 2004. Third, the estimates were adjusted for under-reporting of survey data as per multipliers derived by Surjeet S. Bhalla (2002, 2003). See section ‘Estimations’ below for details.

This presents the number of households in ‘000s according to the annual total expenditure of the households. Higher expenditures could be due to larger incomes or due to larger households.

These variables have been segregated into five annual household expenditure categories: Rs 0 to 35,000; Rs 35,000 to 70,000; Rs 70,000 to 140,000; Rs 140,000 to 200,000 and greater than Rs 200,000.

The data are from the National Sample Survey Organization that have been updated and adjusted to account for lower reporting as well as for the year calendar year 2003-2004. The same set of adjustments as briefed in Part A above were also conducted for these items – these are discussed in the last part of this section – ‘Estimations’.

This presents the number of individuals in ‘000s according to the Per Capita Monthly Expenditure (PCME) of the household that they belong to. Higher PCMEs tend to imply better economic profile of the individuals. The data also includes the numbers of high school graduates or those having greater education in ‘000s under the same expenditure classification.

These variables have been segregated into five Per Capita Monthly Expenditure (PCME) categories: less than Rs 400 per person per month; Rs 400 to 550 per person per month; Rs 550 to 750 per person per month; Rs 750 to 1,500 per person per month and greater than Rs 1,500 per month.

The data are from the National Sample Survey Organization that have been updated and adjusted to account for lower reporting as well as for the year calendar year 2003-2004. The same set of adjustments as briefed in Part A above were also conducted for these items – these are discussed in the last part of this section.

Markets grow for various reasons and different components of the markets grow in different ways. It is well known that past growth strongly reflects future growth prospects of the market.

In some districts, market growth is pushed by population growth, in others it is due to greater expenditure of the incumbent population. We include five measures – population, employment, literacy, savings and credit. Annualized growth rates of population, employment, and literacy have been calculated from data available for 1991 and 2001 (Census of India). However, the annualized growth rates for savings and credit have been calculated from available RBI data for 1996 and 2001.

• Annual Population Growth (Census of India-1991, 2001) is by far the strongest reflector of market growth. High population growth tends to be in those locations where the economic opportunities are growing rapidly. However, given high poverty rates in India, it should not be taken as the sole measure of market growth.

• Annual Employment Growth (Census of India-1991, 2001) measures the growth in main workers and is another important measure of market growth. High employment growth would also happen where there is rapid growth of economic opportunities. High employment would mean higher earnings and hence, would suggest greater spending capacity.

• Annual Literacy Growth (Census of India-1991, 2001) Higher literacy growth would suggest increasing awareness amongst the population of the district.

• Annual Saving Growth (RBI-1996, 2001) Higher nominal savings growth (not adjusted for inflation). This indicates an increase in the wealth of the population and hence, greater ability towards investment in assets.

• Annual Credit Growth (RBI-1996, 2001) is again a good reflector of market growth. Higher nominal credit growth would signify greater economic activities relative to other districts.

Households by number of persons and rooms in a house: This component presents the number of households with different categories of number of dwelling rooms and household size. The categories of household size are households with 1-2 persons, 3-5 persons, 6 & above persons.

Likewise categories of number of dwelling rooms are 1 room, 2 room, 3 room, 4 rooms and greater than 4 rooms. The estimates are derived from Census 2001 and have been updated to represent 2004 levels using the expected population growth between 2001 and 2004. The number of households with no exclusive rooms can be ascertained by subtracting from the total households.

Penetration of modern technology and equipments reflect the nature of consumers. The data here are related to the percentage of households who have the said asset available to them. Many households for instance do not own their cars but use those owned by the employer of the income earner. The data are from the Census of India 2001. Penetration levels imply the percentage of households who have access to the use of the asset as of 2001. These have not been forecasted to 2004 levels.

This is the percentage of expenditure in the district to the expenditure in the state. The contribution helps identify the relative share of the district in the states market. (In the case of state pages it is the share of the state in the national market.) This is done for the following categories of expenditures: 

• Total market size: Total annual household expenditure 

• High value food: All food items not including cereals and pulses. This would include expenditures on fruits and vegetables, meat, edible oils, etc. Basic foods such as cereals tend to be distributed in a similar manner as the size of the population, but ‘high value foods’ better reveal the relative purchasing power. 

• FMCG or Fast Moving Consumer Goods: As mentioned above. 

• Durables 

• Clothing & footwear 

• Other goods & services

The contribution is graphed for total expenditure (or market size) and a break up for particular item categories is also available. The same set of adjustments as briefed in Part A above were also conducted for these items – these are discussed in the last part of this section – ‘Estimations’.

These sets of estimates can best be used by those interested in identifying a few districts within a region or a state where they would like to focus there marketing/selling efforts.

A general classification of households has been made based on the family type, number of adults and number of children. Numbers of households according to family type include the following:

• Nuclear Households: Households with a husband, a wife, and one or more unmarried child residing together.

• Joint Households: Households with more than one married sibling residing together.

• Extended & Others Households: Households with a married child living with parents are classified as extended. ‘Other households’ include many different kinds of households such as aunts or uncles living with nephew/niece etc as well as other

The term residing together implies that all members of the household share the same kitchen.

In addition the table also includes the distribution of households according to the number of households members, these include:

• Households by number of adults: This component has been classified into three different categories i.e. households having 2 adults, 3 adults, and 4 or more adults. Households that have only one adult are not included in this categorization but can be calculated by taking the difference from total number of households given in part B (explained above).

• Households by number of children: This component is also classified into three different categories, i.e., households with 1 child, 2 children, and more than 2 children. Households that have no children are not included in this categorization but can be calculated by taking the difference from total number of households given in part B (explained above).

The percentage distribution of households has been estimated using data from National Sample Survey Organization’s (NSSO) 55th round employment and expenditure surveys. These two large sample surveys have then been combined to create a sample of 240,000 households with approximately 1.2 million individuals. The estimates were relevant for 1999-2000. These were then combined with Census 2001 as well as growth rates derived from Census 1991 and 2001 to provide a distribution of different types of households in numbers for the year 2004.

This part presents the distribution of expenditure across quarters. It is separately plotted for the various item categories and the market size of the district. Seasonality is an important parameter as expenditure is not uniform across a year. People tend to spend more in festive seasons and after harvesting in rural areas. Seasonality tends to be more marked in categories such as durables. An appreciation of seasonality allows a better time-focus to sales efforts. The following broad categories are covered:

• High value food: All food items not including cereals and pulses. This would include expenditures on fruits and vegetables, meat, edible oils, etc. Basic foods such as cereals tend not to show any seasonal fluctuations, this is less true of ‘high value foods’.

• FMCG or Fast Moving Consumer Goods: Typically seasonal variations are low in these items unless the population of the district is highly affected seasonal income variations.

• Durables: This class of expenditures tends to show some of the highest seasonal fluctuations.

• Clothing & footwear: This class of expenditure tends to more seasonal in rural and semi-urban areas than in urban ones.

• Total market size: Total household expenditure on all items.

The same set of adjustments as briefed in Part A above were also conducted for these items – these are discussed in the last part of this section.

It should however be noted that the categories for which seasonality is provided are very broad and provide an aggregated picture. Moreover, seasonality is also affected most by when particular festivals of importance in the area in that year. Last, seasonality is becoming less of an issue in urban areas but continues to be of importance in rural areas.

Many, if not most, marketing professionals use socio-economic classification (SEC) of urban consumers as an indicator of the propensity of a consumer to purchase different items. The number of urban households under each SEC category (A, B, C, D, and E) is provided for each city-market. Note that this does not imply that households with low SEC profile do not purchase high value consumer goods – they merely have a low propensity to do so.

A ‘high’ SEC rating suggests that the household has a high propensity to purchase high value items. In other words, SEC A households have a higher propensity to purchase automobiles than SEC B, who in turn have a higher propensity to purchase than SEC C and so forth. The table below provides a delineation of the different types of education-occupation characteristics that have been used to categorize households.

The estimates are based on raw data from the NSSOs large sample expenditure and employment rounds that together make a sample of 240,000 households across the country.

Another interesting socio-economic classification is in terms of the number of earning members within a household. The following are presented:

• Single Income Households: This includes the number of households with single income earning member.

• Two or More Income Households: This includes the number of households with two or more earning members. The earning members could be either a couple or siblings or parents and their children.

• Own house: This is the number of households staying in homes that are self-owned. Since rent is not paid in these cases, typically larger amounts are available for other discretionary purchases.

• Separate Kitchen: This is the number of households that have a kitchen in a separate room. This indicates the potential for a large variety of products.

See estimations on demographic data in ‘Estimations’ section for a brief on the method followed for these estimates for the year 2004.

This section brings forth data at the sub-district level. The reported data and estimates are the more important urban areas in the district. The following data are reported for these key (top seven in terms of number of households) urban areas in each district:

• Households: estimated number of households as of 2004.

• % of Urban Total: Households in the given urban sub-district as a percentage of the total urban households in the district as in 2004. This suggests the importance of the particular area in the total urban component of the district.

In addition the penetration levels of key consumption assets are also included for each urban-sub-area. That is the data provide the percentage of households that have access to: Four wheelers, two wheelers, bicycles, and TVs. All estimates are for the year 2001 and have not been estimated for 2004, as an appropriate methodology could not be identified. However it is unlikely that there would be any change in the relative importance of different sub-urban areas in the first two-three years of the 2000s.

This volume brings together data from different sources. The bulk of the raw data used here is from the year 2000-01. This includes the decennial Census (2001), National Sample Survey Organization’s large sample expenditure and employment surveys (1999-2000), Reserve Bank of India (various years), Central Statistical Organization (various years), etc. However, the estimates presented here are estimates for the year 2003-2004. This has involved forecasting the earlier data based on observed growth rates. This in turn has also involved us to take into consideration similar data for earlier years (for instance Census 1991) to estimate the growth rates.

In addition, our usage of survey-based data also involves under-estimation of some expenditure that need to be adjusted. This is because respondents tend to under-report expenditures. Fortunately for us, others have published studies on the extent of under-estimation in survey-based data such as that from NSSO. Hence not only do we adjust for observed and expected growth but also for under-reporting. (See Surjit Bhalla, ‘Imagine There’s No Country: Poverty, Inequality, and Growth in the Era of Globalization’, Penguin, New Delhi, 2002. Also see by the same author, ‘Not as Poor. Nor as Unequal As You Think: Poverty, Inequality, and Growth in India, 1950-2000, mimeo. New Delhi (2003)).

But that is not all. There are many gaps in published and available data, especially when we are attempting to work at the district level. This has required us to estimate some of the data at the district level. This happens to be more true for the smaller districts with lower populations such as those in the northeastern parts of India or areas such as The Dangs in Gujarat, or the more interior parts of Jammu and Kashmir, etc. The quality of the estimates therefore tends to be much better for the districts with larger populations and larger market sizes. Broadly three types of estimations have been done. These are described below.

Demographic data: The demographic data (households and individuals) are from the Census of India as well as the large sample employment and expenditure surveys National Sample Survey Organization held in 1999-2000. Typically demographic structures (percentage distribution across categories) change slowly but the actual numbers of households and population change much more rapidly. We therefore use the Census figures from 2001 as well as the observed growth rate from the Census to shoot up the figures to make them relevant for 2003-04. The structures (percentage distribution across different categories) are however calculated from the NSSO’s data as they are not available from any other source.

To improve the quality of the estimates we have combined the two data large sample surveys of 120,000 households each (expenditure and employment surveys) both held during 1999-2000. This ensures a combined raw sample data of more than 240,000 households and about 1,200,000 individuals. Since many of the queries were the same in both the samples we were able to exploit the similarities and the large sample nature to get district level estimates. Appropriate modifications to the multipliers had to be conducted to ensure relevance of the estimates. These distributions were then combined with the data from the Census of India to get a picture that is comparable with the rest of the data.

Expenditure data: All data in Rupee terms are presented in nominal values for the year 2003-04. They have been first estimated from the NSSO using their large sample surveys held in 1999-2000. Three steps have then been conducted on these raw estimates.

• The data on expenditures from the two large sample surveys were combined using the methodology as outlined by Dubey, Amaresh and Laveesh Bhandari in the work entitled "District Level Poverty in India" published by the Rajiv Gandhi Institute of Contemporary Studies, New Delhi, 2003.

• The degree of under-estimation has been removed as per the methodology prescribed by Bhalla (2002, 2003) referred above.

• The growth in expenditures between 1999-2000 and 2003-04 has been estimated using the nominal growth in personal consumption using National Accounts Statistics for the period 1993-94 to 2000. It has been assumed that this same growth will be observed in the first part of 2000s as in the middle/latter nineties.

It should however be noted that for the extremely high expenditure/income groups we are still not capturing the full extent of their expenditures. Though we would have liked to adjust for this underestimation as well, we could not identify any satisfactory methodology for doing so.

Asset usage/penetration data: The Census of India has recently brought out household tables related to access to assets by households. These data are for the year 2001. All the penetration data described above derives from this data source. The data on penetration has not been updated to that expected for 2004 as we could not identify an acceptable methodology for doing so. However, the estimated numbers of households for 2004 are provided and the reader is welcome to make her own judgment of the same.

Notes on Data

As per the census there are 593 districts in India. The Census reports data for 592 districts; the census was not carried out in Kutchh district of Gujarat. We have reported most data for all the districts. Data is not available in the NSS for the following districts: The Dangs, Lahul & Spiti, Leh (Laddakh), Doda, Punch, and Kargil. Some of the consumer characteristics for these districts have not been estimated.

In certain cases the data available was for undivided district. The values for the missing data have been estimated using 2001 population as a ratio of the population of the parent district from which the major portion was carved out.

The consumption expenditure has been estimated for 2003-04 using GDP growth rate (Gross Domestic Product at Current Prices, 1993 up to 2000) at the state level.

Adjustments have also been made to account for consumption structure. There has been a shift in consumption pattern between 1993 and 2000. This shift in pattern has been accounted for in the estimates.

Urban households data has been estimated for 2004 using urban population growth at the district level.

Demographic estimates for 2004 are based on population growth at the district level.

For certain districts the sample sizes in combined NSS survey data set are very low. For these districts (that have fewer than 350 sampled households), the low sample problem has been circumvented estimating consumer characteristics on the basis of a combination of districts within the state that have similar characteristics (are socioeconomically homogenous).

In case of seasonality as well the regional data have been used where sample sizes are found to be very low.

Due to "rounding" of figures the aggregates may deviate from the given totals.

Note: Key Urban Areas are urban regions of the district. Data for major cities falling within the zones have been provided among the comparative pages.

The Item Categories :