NCAER – A History of Independent Research in Applied Economics

The National Council for Applied Economic Research is an independent, non-profit research institution that is committed to assist government, civil society and the private sector to make informed policy choices. Set up in 1956 at New Delhi, as a quasi-governmental institution, it has been a significant player in the arena of capturing the socio-economic development of the country providing inputs to policy and research through sponsored projects and independent research. It has significant long term partnerships with institutions such as with the London school of Economics, the CEPR Maryland, Brookings, etc, collaborating on creating intellectual assets. Over the years NCAER has built up a loyal base of customers, who depend on it for the unique information that helps build their policy and strategy. The NCAER legacy in Income, Spending and Saving surveys

The NCAER legacy in Income, Spending and Saving surveys

NCAER’s strengths in economic analysis have been rooted in its mastery of large data, survey based information tools and techniques, and its resource base as an analytical powerhouse. The particular legacy in various domains of research, include the MISH – NSHIE surveys in household income and spending. As the sole series of rigorous estimates of household income, it has generated a strong following in both policy and corporate sectors. The legacy is particularly strong: it started in 1962 with the All India Rural Household Survey of Savings, Income and Investment, and continued with the Urban Income and Saving Survey again in 1962. It continued with the ARIS/REDS (1971, 81, 99, 06), the significant contribution to understanding various dimensions of income and its effects through the MIMAP and Human Development Indicators (1993, 2005). The regular series of surveys of household income was thoroughly established with MISH 1985-2001, and culminated in the landmark National Survey of Household Income and Expenditure, or NSHIE, 2004-05. Indian institutions needed hard data on India’s rapidly changing socio-economic landscape, and issues of poverty, consumption, etc.

Both the Policy analysis establishment and corporate India, are regular users of these surveys for their Pan India initiatives. Developmental efforts first require an understanding of the socio-economic landscape that requires initiatives, and thereafter the impact assessment, through “state of the nation” snapshots to capture the holistic effects at household level – how households in India live their lives. This has led the NSHIE 2004-05 analyses to focus its most recent publication on, and title it, the "Living in India" series. It provides detailed consumption data, and links household consumption with income, assets, savings/debt and demographic profile.

In the next round, NSHIE -2011, the agenda has been suitably carried forward with a more comprehensive view of Indian households and their life. It includes preliminary listing survey, main survey (comprehensive Living in India data), followed by panel surveys covering themes in depth.

Agenda – Living in India 2011

Continuing with the earlier trend, NSHIE 2011 has taken up the agenda of focusing on the needs of three key stakeholders whom the NSHIE will serve: business, policy and research entities within and outside India.

The primary focus would be on the need for Consumer insights, trends & tracking, projections, and a larger understanding of the Indian consumer landscape and market structure. For instance the demographic dividend may be rather well known fact internationally, but what are the income and spending patterns of these youth segments? Are they global in their perspective, and are they well equipped to harness and boost the growth in GDP that India has witnessed? What will they consume? Will they spend on education and health, and will they be employed or generate entrepreneurial growth? These are all critical questions that key stakeholders may wish to explore through the current round.

Further, to capture this holistic growth, a “State of the nation” dashboard of factors such as wellbeing, financial and mainstream inclusion, migration, health, education, etc. has been chosen as specific coverage areas of NSHIE 2011.

Finally, some new sectoral themes are being taken up in NSHIE 2011 such as Remote payment (Bill Gates foundation); Gold Investment (World Gold Council), Automobile sector, household cost of acquisition of cash for domestic use, health and insurance, FMCG and TV viewership.

Objectives – Living in India 2011

The broad agenda mentioned previously (viz., consumer insights, state of the nation dashboard, and new themes), was concretized into more specific agenda, and objectives were stated as below. Continuity with past data would be ensured to permit robust projections. In addition, an expanded list of variables will be canvassed, including consumer confidence, buyer behaviour, mindscape of consumers, wellbeing and aspirations.

Income, spending and savings: The research involves the study of Income, Spending, Savings, of households, and demographics. This will be studied along with measures of related variables of debt and assets, housing, remote payments, liabilities and commitments, financial optimism, inflation and income expectations, etc. The survey has over the years fulfilled the need for robust estimates of income that are otherwise unavailable, and has been used for policy and corporate decision making.

The analytical design proposes to describe, classify/cluster the population into various income/socio-economic strata, based on econometric as well grounded research. Further, projections/trends, modeling of the structure of the market, its drivers, are also planned.

There will also be scope of tracking on a more continuous basis than in the past since this time it is planned that the 5 lakh population sample for listing, will become a platform for continuous measurement, including the Main survey and then later, rolling panel data.

The measurement of many critical variables such as income and spending on household routine expenditure, will then become far more robust, and will continue to attain stringent international benchmarks for validity of the measures – a hallmark of the NSHIE and NCAER data in general.

State of the nation: The study of household health and health insurance, education level, place and aspirations as linked to occupations, occupational aspirations and changes, wellbeing, financial optimism, consumer confidence, income and inflation expectations, household amenities and water, MNREGA, etc. is an integral part of the Main survey. This will provide a detailed look at the state of the nation on various parameters such as social background, socio-economic conditions, and related amenities and aspirations of society at large, and of households in given states and regions. The creation of an index in various domains, such as income, consumer confidence, wellbeing, health, financial inclusion, etc., is also planned as a secondary objective, and will provide a handy dashboard to understand, track and project the social and other developmental parameters.

Sectoral themes: Sectoral themes specific to health and health insurance, automobiles, insurance, remote payment, gold purchases, unique identification, media and TV usage, cost of cash at household level and FMCG consumption will be covered in this round.  In all there are twenty three sections of covering different data sets that will be captured in the Main round of the Survey.

In all there are twenty three sections of covering different data sets that will be captured in the Main round of the Survey.

Analytics and outcomes

There would thus be the much deeper, more granular classification of households, and consequent segmentation of the Indian market, a more detailed market structure that will be reported. The level of details used in describing the segment profiles would be much richer. For instance, sections on income and spending, can also be informed by survey data on savings, buying behaviour, asset ownership in numbers and type of assets, housing, amenities, FMCG and non essential consumption, to segment the market. Looking further, the section on the mind of the consumer promises to reveal a lot more about the materialism, individualism and buying tendencies of the Indian consumer, and the data will be used in creating lifestyle based clusters, and lifestyle indicators will be used for further refining the consumer classification and market structures.

The broad socio-economic parameters reveal a lot about life in India, including developmental impact: does education really impact income levels, aspirations, occupational shifts, etc. significantly, within the generation? Does it lead to better health? Can one re-explore the logic of smaller family sizes with better health, higher incomes? To what extent does the "progress" in specific infrastructure terms (electricity, roads, access to amenities and banking, or a combination of the same) really lead to better quality of life and material wellbeing? Are the social classes really bridging the divide on critical indicators? It is planned that this area of analysis will involve collaborative efforts with leading global academics who, along with CMCR experts, will be focusing on specific questions using the data.

Based on the specific agenda above, a set of specific objectives was framed for the Living in India 2011 Survey.

Primary Objectives: Status, Classification and Distribution, and Index

 

Consumer Insights

                  

 

 

 

Secondary objectives: CMS, State of Nation - Models, drivers and scenarios.

More advanced statistical treatment to meet the objectives of estimation of composite constructs such as health, wellbeing and other “state of nation” index, and their projections, through models, identification of drivers, scenario generation through appropriate techniques will be taken in the second stage of analytics. These will also involve other databases such as the NSSO consumption data etc.

A set of objectives relates to the study of upward social mobility of households of various classes, and their changes in lifestyle given changes in socio-economic parameters vis-à-vis income. For instance, identifying the determinants of wellbeing and financial security, may help explore whether education and place of education, quality of education (from village level survey) are linked to financial and socio-economic mobility in rural households. The degree of freedom, modernity and conformance to traditions within family and society, aspirations and exposure (place of study) may shed further light on the patterns of education, financial inclusion, income growth, occupational changes.

Yet another major area of study relates to the buying behavior that will be captured as part of the primary objectives. As a secondary objective, the changes in patterns of buying and material lifestyles amongst various socio-economic classes may reveal the correlates of the socio-cultural changes that can be observed in households and family life across India.

The Research Design – Living in India 2011

Overall approach to the National Household Survey 2011 was to meet the research objectives stated above, using the best mix of baseline-panel-longitudinal and cross-sectional surveys, in a phased manner. The research is designed as a multistage survey, with the Listing Survey canvassing over five lakh households, the Main round covering around one lakh households, and subsequent Panel rounds, all on the same households covered during Listing round.

The detailed objectives were derived from the agenda and primary and secondary objectives, and an intensive effort at identifying the constructs were carried out to meet these objectives. Then specific objectives at each level were stated. The chief wage earner of the household, identified as one with the highest income, would be the main respondent, for all sections except one. The data for the household level would be captured for the reference period April 2010 - March 2011.

It was planned that each phase would draw upon the same households – the listing survey would help identify household characteristics for stratified random sampling requirements of the min survey. Main survey would help capture key household data, and main themes could be explored in depth and tracked for more robust estimation and modelling, though panel surveys on smaller subsets of the same households.

The listing phase included face to face interviews of chief wage earner, using a 61 item schedule capturing household SEC, income, expenditure and other data including wellbeing perceptions. The main highlights of the listing survey include use of PDA (personal digital assistant) to help capture data, and ensured higher efficiency and data quality and control, fast and accurate back-checks, etc. It helped generate the digital image of the chief wage earner for future identification.

The Main round would comprise sample drawn as per the strata decided upon and selected based on stratified random selection of households based on Listing survey. It would cover the variables identified while operationalising the objectives, as above, through face to face interviews with chief wage earner, using a schedule with twenty three sections of items, which are detailed below.

Based on the preliminary analysis, panel studies would be initiated to study specific themes, using international best practises in panel surveys, with rolling sample. This would cover tracking of indices created through prior cross section study, in-depth study of income and social mobility etc., as mentioned earlier.

 

1.1.Data collection Instrument

Listing survey schedule: The survey schedule used for interviews in the listing phase contained ten sections, with a total of 61 items.

 

Main survey schedule: The schedule used in the main survey contained twenty three sections. The first section contained sample identification data. The twenty two sections, with 196 subsections/questions, and over 900 possible items, collected detailed information about households as given below. Specific items would be relevant for a given household, such as income: out of eight subsections covering income from all types of sources as per strata, only one or two would be relevant for each household generally, and of 130 items only a few would require a response from the household. For example, if the household belonged to agriculture self employed strata, then the relevant section would be the income from self employment in agriculture, and sections on professional self employment would not generally apply. Pre-testing helped identify items that required modification, and re-coding of options, among the other things. The items in each section are detailed below.

 

 

The sections above provide an exhaustive coverage of households and their members, their socio-economic and living indicators. Each section will be analyzed to provide a detailed picture of Living in India, and how life differs across various household categories. It will represent the Indian population, as per the sampling design given below. This ensures that survey results obtained will be amenable to projections for the entire Indian population, given appropriate weights and use of relevant measures.

2.      Sample design

The household is the target unit of the study, with states and urban/rural categories as sub-populations or target groups, for whom representative estimates will be sought. The geographical coverage of the survey includes the entire country comprising of major states and UTs. The sampling design is kept broadly similar to the last round, which was drawn after reviewing best international practices and proved robust and effective.

A three-stage stratified sample design has been adopted for the present survey in which a ready-made frame is used for the first two stages and a sampling frame is developed in the last stage. Districts, villages and households form the first, second and third stages of sampling, respectively for selection of the rural sample while cities/towns, urban blocks and households are the three stages of selection for the urban sample. Sampling for rural as well as urban areas is done independently within each state/UTs. Following the basics of sample survey principle that geographical spread is more important than the sample size, both rural and urban sample is selected from a wide cross-section of the country. The rural sample has been selected from a representative number of districts from across the country, while the urban sample covers a range from big metropolitan cities to small towns with population below 5,000. The sample sizes at first, second and third stages in rural and urban areas are determined on the basis of available resources and the derived level of precision for key estimates from the survey, taking into account the experience in conducting the earlier round of NSHIE-2004 survey.

Coverage: The survey will be carried out in 32 states and Union Territories of the country. Within each state an NSS Region -which is a group of districts within a state similar to each other in respect of agro climatic features- formed the strata for both Urban and Rural Sample. All the 32 States are divided into 75 NSS Regions which form the strata or domain of study below the level of state/UT. Each NSS Region is assigned a 3 digit code where the first two digits indicate state/UT and third indicate region number within a state/UT. The rural and urban areas are taken as defined in the census.

Sampling Frame: The list of 2001 census districts and villages constitute the sampling frame for rural and list of 2001 census cities/towns for urban. In the absence of the definitive list of households (sampling frame), specially designed listing proforma will be used to list households in the selected villages and urban blocks to collect information on various auxiliary variables. In the case of large villages/urban blocks, a fraction of households will be listed based on sampling fraction.

Sample size: A total sample of 99,384 households will be covered in this study. Out of which, 45,144 households spread over 2,508 villages in 292 districts and 75 NSS regions covering the 32 states/UTs will be covered in rural areas. In urban areas, 54,240 households spread over 2,712 urban wards in 363 towns and 75 NSS regions will be covered.

Rural Sample: In rural, a sample of 292 districts was allocated to the 75 NSS regions within the 32 covered states/UTs in proportion to the total number of districts in an NSS region. From each of the NSS regions, the allocated number of districts was selected, as the first stage sample units, with probability proportional to size and replacement, where rural population of each district as per Census 2001 was used as size measure. Villages formed the second stage of selection pro­cedure. A total sample of 2,508 villages (second-stage sampling units) was allocated to the selected 292 districts, approximately in proportion to rural population of each selected district. The allocated number of sample villages was chosen with equal probability sampling approach. A total of 500,400 households were listed comprising 250,800 households from entire rural India.

The variables of occupation of chief wage earner (CWE) and land possessed by household, were used for stratifying the listed households into nine strata as follows:

Stratum 1: CWE was self-employed in agriculture and household possessed 0-2 acres land;

Stratum 2: CWE was self-employed in agriculture and land possessed was 2-4 acres;

Stratum 3: CWE was self-employed in agriculture and land possessed was above 4 acres;

Stratum 4: CWE was labour (agricultural/other casual) and having no land;

Stratum 5: CWE was labour (agricultural/other casual) and having land;

Stratum 6: CWE was self-employed in non-agriculture and having no land;

Stratum 7: CWE was self-employed in non-agriculture and having land;

Stratum 8: CWE was regular salary/wages and other sources and having no land

Stratum 9: CWE was regular salary/wages and having land.

 

Following the above sampling design in rural areas, the realised sample of 45,144 households out of preliminary listed sample of 250,800 households was spread over 2,508 villages in 292 districts and 75 NSS regions covering the 32 States/UTs. 

Urban Sample: In the urban, within the 32 covered States/UTs, the 75 NSS regions were again treated as the main strata and a sample of 363 towns (first stage units) were selected covering all sample states. All the cities/towns of varying sizes (less than 5,000 to over 10 million populations) in sample districts were grouped into nine categories based on the population. Those cities/towns having population over 200,000 that formed first five groups (over 190 cities) were selected with a probability one. The remaining four town groups were considered as separate strata on the basis of their population size and from each stratum a sample of towns was selected independently. These town groups may be called sub-strata which are formed within the main strata of NSS regions in each State. Out of 363 sample towns in the listing survey, 189 cities and towns were selected by equal probability sampling approach from remaining four towns groups. A total sample size of 2712 urban wards was allocated among the sample towns more or less in proportion to the number of wards in the respective towns, maintaining an equal number of wards allocated to each selected town in a town group. The number of towns and wards in respective town group is given as follows:

At the time of listing of the sample households, information on household size, household consumption expenditure for last month (MPCE), and occupation of chief wage earner (CWE) were collected for stratifying the listed households into 10 strata. Then random selection of households later on a weighted basis was made from these strata. The strata is as follows:

Stratum 1: CWE was regular salary/wage earnings and sources like remittances, pension, etc. and MPCE of Rs. less than 800;

Stratum 2: CWE same as in stratum 1 but MPCE Rs. 801-2500;

Stratum 3: CWE same as stratum 1 but MPCE above Rs. 2500;

Stratum 4: CWE was employed as casual labour (agricultural or non-agricultural) and MPCE of Rs. less than 800;

Stratum 5: CWE was employed as casual labour (agricultural or non-agricultural) and MPCE above Rs. 801;

Stratum 6: CWE was own account worker (Petty traders, shop owners, businessman with no employee, self-employed professional and non-professional) and MPCE less than Rs. 800;

Stratum 7: CWE was own account worker (Petty traders, shop owners, businessman with no employee, self-employed professional and non-professional) and MPCE Rs. 801-2500;

Stratum 8: CWE was own account worker (Petty traders, shop owners, businessman with no employee, self-employed professional and non-professional) and MPCE above Rs. 2500;

Stratum 9: CWE was Employer (Business/self-employed) and MPCE Rs. less than 2500; · Stratum 10: CWE was Employer (Business and self-employed) and MPCE above Rs. 2500.

Following the above sampling design in urban areas, the realised sample of 54,240 households, out of preliminary listed sample of 2,71,200 households, was spread over 2,712 urban wards in 363 towns and 75 NSS regions covering the 32 States/UTs, used for the Main Survey.

3.      Analytic design and objectives

The analytical design revolved around the primary objectives. For each objective, the relevant sections and variables were identified, and statistical tests of normality etc, were used to analyse the data structure. The major objectives required the use of classificatory statistics, and the use index, factor and cluster analysis, hierarchical analysis, trees, tests of association, etc. would be used to develop the classification of consumers on various basis mentioned in the objectives. Thereafter, the market structure and projection related tools will be used for analysis. Deeper profiling would use multidimensional scaling in addition to other tools. Modelling would use exploratory and CFA; regression, path analysis etc. A detailed mapping of analytical objectives mentioned in the prior section, with analytical tools and outcomes has been prepared. An indicative description of the same is given below, organised around the consumer, state of nation and specific themes mentioned earlier.

 

Consumer insights

 

 

 

For each of the riders and specific themes, there would be analysis based on status of households on that sector, household classification based status and distribution, and modelling to understand the drivers in that sector, along with creation of index, such as for health, as mentioned above. These would be exactly in line with the analytical objectives mentioned in the prior section.

Overall, the objectives are to provide a clear description of the income, spending and savings of the households in India. Further, their comprehensive demographic profile and socio-economic status will be captured. From there, modeling the various phenomenon of interest to key stakeholders would be carried out. Indices will help capture quick measures of the status of the various domains of interest, and track their changes over time and across regions. As discussed above, the specific interest areas covered in the study may help in the formulation of policymaking and decision-making, in both corporate and governance / developmental fields.

For corporate strategy, marketing and operations, market structure and insight support is required for identifying and mapping the contours of new product/market opportunities, “hard” numerical estimation of demand and operational parameters such as for assembling, distribution and logistics. It is also required for determining the strategy for market operations, such segmentation, targeting customers and partners, operational clustering and geo-logistics, etc. However, deeper insights into the socio-cultural and behavioral dimensions and groupings, often with psychographic measures, are key for business and particularly marketing: segmentation, aspects of positioning, branding, creating and delivering offers that are competitive, such as pricing, access, operational capabilities, product development, sales and distribution, communication. Many a times they also inform the corporate of the requirements for operating in specific segments which are distinct, at the micro level, sub-cultural and district levels.

For policy, it is essential that there is a proper understanding of the number and scale of, say, target beneficiaries of a developmental program, or socio-economic class or industry for whom a regulatory framework is being developed. But it is also equally important that there is an understanding of how, for example, target beneficiaries would become aware of, would access, would adopt, and would eventually benefit from a developmental program or a new policy. So the information provided will try to include both aspects – of scale and distribution, well as collateral information of how demographic, infrastructure, community, socio-cultural norms, media usage, may affect the phenomenon under study.

For both policy as well as corporate, the whole set of behavioral variables, related to a certain status of ownership or class (income, demographic or social), would prove invaluable information. The determinants of such patterns, such as awareness and media usage, access to new media such as telecom and internet, situational and psychographic factors such as beliefs, value systems, self orientations, aspirations etc would be much needed for effectiveness of the programs/strategy. Further, for facilitating the market forces, it is essential to help understand the whole value chain that a developmental program or a product launch will impact, including the parties affected, their status and role. The efficiency of operations will be an outcome of the quality and level of detail of such information. That is the ultimate goal that NSHIE 2011 is attempting to achieve.

Form of output

It is planned that the major output as earlier, in the form of reports such as the proprietary Living in India: How India Earns, Spends and Saves, as well as other major reports on socio-economic trends and analysis will continue. Sector specific reports will also be published in greater numbers now, given the larger coverage and involvement of collaborating domain experts.

Databases and software based products will also be brought out, such as fact finders, query based search engines and higher analytical software utilizing the rich NHSIE data in the form of a data warehouse utilizing unit level data. Index formulation and panel based indices, as well as business briefs and specific reports will also be part of the outputs on a regular basis.

Given an open platform concept, publications in leading journals and other publications will also happen as a matter of course with the engaged team members and collaborators bringing out specific theme papers and articles in topics of relevance.

Policy advocacy as well as advisory basis the NSHIE is also a feasible output that shall be consciously adopted, and may be tied with conferences and workshops as well as presentations to stakeholders and target audiences relevant to policy.

Further, the business stakeholders would probably require in-depth analysis of a specific topic, and that also shall be an important area for dissemination of analysis. This will be in the form of customized reports. Thus a wide range of output areas and forms are planned with the NSHIE data platform as a basis.