An Essay on National Income Statistics of India

Makoto Kojima

1. National Income Statistics: Background

Scholars attempting to estimate national income statistics for pre-independence India have confronted innumerable difficulties in finding reliable data. Whatever estimates they make, they are forced to rely on macrolevel data and/or to make numerous assumptions. Only for the period after India became independent in 1947 is it possible to find reliable official statistics. However, Pakistan's secession and the consequences difficulties experienced by the Indian territory have meant that no one has as yet made a concerted effort to link Indian national income statistics for the periods before 1946 and after 1947.

In 1949, soon after the establishment of independent India, the National Income Committee (NIC) was formed to compile statistics and estimate national income. The committee was headed by P.C. Maharanobis and included D.R. Gadgil and V.K.N.V. Rao. Assisting the NIC was the National Income Unit (NIU), directed to prepare estimates of national income every year. The NIU was under the jurisdiction of the Ministry of Finance, later changed to the Central Statistical Organisation.

Estimates of Indian national income have been based upon four sets of statistical series. First, The Estimates of National Income was published in 1956, and had a 1948-49 base year. Second, The Estimates of National Product was published in 1957 as a revision of the conventional series. The revised edition had a 1960-61 base year and gradually came to include personal consumption expenditure, savings, capital formation, factor incomes, consolidated accounts, and public sector accounts. Renamed The National Accounts Statistics (NAS) in 1975, it included estimated values for the period 1950-51 through 1972-73. Thirdly, another revised edition was published in 1978, which shifted the base year from 1960-61 to 1970-71. In 1980 estimated values starting from 1950-51 were published. Fourthly, a new series was published in 1988, and it had a 1980-81 base year. In 1989, estimated values from 1950-51 through 1979-80 were published.

2. Special Features of the 1980-81 Series

The main method adopted to compile estimates of domestic production by industrial origin is to estimate gross value added in terms of the difference between input and output values. Among the sectors examined by this production approach are agriculture/forestry/fishing, manufacturing (excluding unregistered manufacturing), and construction. The income approach is adopted for other sectors, such as electrical power, transportation, commerce, and administration. Under the income approach, factor payments for organized sectors (registered sectors plus the public sector) are estimated on the basis of budget documents and annual reports of enterprises. Applied to unorganized sectors, the income approach means multiplying value added per worker by labor force (number of workers). In addition, the commodity flow method has been adopted to estimate private final consumption expenditure and capital formation.

The new series features revisions regarding data on and methods of estimating forestry (production of firewood); textiles (categorized in India as an unregistered sector and also regarded as a decentralized sector) and kutcha construction (which is in contrast to pucca construction, is labor intensive and based on crude materials), as well as different areas of domestic industrial production. The most important revisions were made with regard to consumption, savings, and capital formation.

First, because the depreciation entered in enterprise account books has usually not been shown as a cost of replacement of fixed capital, and there has been no regulation regarding fixed capital in the government departments, the consumption of fixed capital has been underestimated. As a result, there has been criticism that net savings as well as capital consumption have been underestimated. In line with recommendations of The United Nations System of National Accounts, 1968 (SNA), the perpetual inventory method was adopted for estimating the consumption of fixed capital.

Second, private final consumption expenditure was subdivided into 38 sub-groups from the previous 24 sub-groups, while the market surplus ratio was adopted to estimate final consumption expenditure for agricultural commodities. In addition, new data was used to make estimates for household consumption items such as fuel and power; spices, salt, medical care/health services, and educational services; transportation services; and consumer durable goods.

Third, until recently, the Reserve Bank of India (RBI) and the Central Statistical Organisation (CSO) both estimated savings but used different methods and data to do so. They have now coordinated their practices so that the CSO compiles estimates for the public sector and the household sector (physical assets and life funds, provident and pension funds), while the RBI is responsible for the private corporate sector and the household sector (other financial savings).

Fourth, with regard to areas of inventory change in capital formation, until recently, grain stocks were estimated according to stocks held by merchants. Now stocks are estimated by subtracting the amount consumed from the net availability of grains available for public consumption. Another change is that the RBI now directly estimates inventory changes for the joint-stock companies of the private corporate sector.

3. Compiling Estimates and Gathering Raw Data for the NAS

(1) Agriculture

A gross value added for agriculture is estimated by subtracting intermediate consumption from the value of output and adding in the gross value-added of government irrigation facilities. To estimate a gross added value for government irrigation facilities an income approach was adopted which combined elements such as employee costs/allowances and consumption of fixed capital.

In order to measure producer income in the agricultural sector as accurately as possible, the average wholesale price of crops sent to market is used. The price is estimated for the period immediately following the harvest.

Raw data for agricultural statistics is provided by the Department of Agriculture (DEAg). Land use statistics are based on Agricultural Statistics in India, which boasts a history of publication dating back to 1884. The final forecast estimates for both area and production for principal crops (45 items, known as forecast crops) are based on Area and Production of Principal Crops in India (containing state-level data), which appears three to five months following harvests. Data on minor crops (ten items, known as non-cast crops including plantation crops) are also provided by the DEAg.

The formula for measuring costs of cultivating major crops was established by the DEAg in 1970-71. The formula takes into account the amount seed rate by crops per hectare, oil use per tractor pump, consumption of electricity, fodder fed to animals, cost of fertilizer, chemicals, and repairs and maintenance expenses of farm machinery.

(2) Manufacturing

We begin by examining the registered manufacturing sector. The term registered manufacturing sector, which are regulated under the Indian Factories Act (IFA), 1948 and the Bidi and Cigar Workers (Conditions of Employment) Act, 1966 applies to all factories employing 10 or more workers and using electrical or mechanical power, or 20 or more workers and not using electrical or mechanical power. However, factories in the "service" sector and all electricity undertakings registered with the Central Electricity Authority are exempted. Thus, manufacturing covers India*s industry divisions 2 & 3 and major group 97(two-digit level), according to the National Industrial Classification (NIC-1987).

A figure for gross value-added for manufacturing is obtained by subtracting the value of input at purchaser's prices from the value of output at ex-factory prices. The output of manufacturing industries includes, in addition to ex-factory values of all products, receipts for services rendered to others and the value of fixed assets produced for use in factories.

The raw data is taken from the Annual Survey of Industries (ASI), which is published by the National Sample Survey Organization (NSSO). The ASI includes an annual survey of all the census sector with factories employing 50 or more workers and using electrical or mechanical power, or 100 or more workers and not using electrical or mechanical power. Of non-census sectors not covered by the above criteria (those employing 10-49 workers and using electrical or mechanical power, or 20-99 workers and not using electrical or mechanical power), one third are sample surveyed every year.

The ASI data is highly reliable but several years are necessary before the results become clear. For this reason, for the years for which ASI results are not available, the NAS utilizes the relevant indices of industrial production statistics and wholesale prices to estimate output and gross value-added. Another factor in the ASI data is that it does not reflect figures for gross value-added for factories which do not respond to the survey. However, by using the employment ratio of the non-responding factories to the census sector, upwardly adjusted estimates of gross value-added are made in the ASI.

Next, let us examine the unregistered manufacturing sector. This designation applies to all manufacturing, processing, maintenance and service units employing less than 10 workers and using electrical or mechanical power, or less than 20 workers and not using electrical or mechanical power. Like other unorganized (non-agricultural) sectors, gross value-added of unregistered\ manufacturing is estimated by multiplying labor force by the gross value-added per worker. Once estimated values for the base year are calculated, an indicator representing the physical volume of activities is used to compile estimates.

Raw data is available from the Economic Census for unregistered sectors and the National Sample Survey (NSS). The Economic Censuses, conducted in 1977 and 1980, covered all business establishments with less than ten employees (those with six to ten employees were categorized as Directory Establishments, and those with one to five employees were classified as Non-Directory Establishments). The Economic Census, conducted in 1990, applied to every business in the country by including not only the above but even so-called Own-Account Enterprises, which have no employees. However, the NSS is conducted only once every few years, which presents the problem of how to obtain estimated values for in-between periods.

(3) Final Private Consumption Expenditure

Final private consumption expenditure covers the final consumption expenditures of households and non-profit institutions serving households. It is estimated through the commodity flow method. Intermediate expenditure consumption for each industry and all final consumption (including imports and exports) other than household and non-profit institutions are taken from the total amount of goods and services at market prices.

The subject expenditures are classified into 8 categories: (1) food; (2) clothing and footwear; (3) gross rent, fuel and power; (4) furniture, furnishings, appliances, and services; (5) medical care and health services; (6) transportation and communication; (7) recreation, education, and cultural activities; (8) and miscellaneous goods and services.

In the case of food items, product consumed by producers is estimated from producer's prices, and the marketed portion is estimated from retail prices. In the case of manufactured items, excise duty and trade and transport margin are excluded from production costs. Import duties are added to the value of imports.

In the case of agricultural commodities, it is necessary to obtain data on output, seed, feed, and wastage to grasp commodity balances available for private consumption. Basic data on inputs and outputs are obtained in the same way as GDP estimates.

(4) Domestic Saving

Domestic saving consists of (1) the public sector; (2) the private corporate sector; and (3) the household sector. The lack of data causes many problems in estimating domestic saving, so up to now, only estimates at current prices have been compiled.

Public sector saving is composed of government administration and departmental enterprises, and non-departmental enterprises. Gross saving from government administration and public enterprises is defined as the surplus of current receipts over expenditures. For non-departmental enterprises, general saving is estimated from analysis of their annual income and expenditures.

The private corporate sector is composed of all non-government financial/non-financial companies, and cooperative institutions. Estimates for this sector utilize basic data from RBI surveys.

The household sector accounts for the biggest portion by far of domestic saving. This is because this sector includes, in addition to households themselves, the non-corporate sector covering all unregistered firms and even some portion of registered firms. It is in fact the non-corporate sector which comprises the major part of the household sector. Because it is not possible to use raw data to estimate figures for the household sector, calculations are based on the residual method.

Domestic saving is grouped in two general classifications, financial assets and physical assets. Financial saving includes currency (cash), net bank deposits, shares/debentures and bonds, net claims on the government, life insurance funds, and provident/pension funds. Included among physical assets are construction and investment in fixed assets such as construction and machinery & equipment as well as changes in stocks.

(5) Domestic Capital Formation

Gross capital formation is the total of gross additions to fixed assets and changes in stocks. Fixed assets consist of construction and machinery *o equipment (including transportation machinery and livestock). Construction for military use and defense equipment are not included in capital formation, but capital outlays for production in military-oriented enterprises are included. Additions to non-reproducible tangible assets such as land and mineral resources are not included in gross capital formation, but improvement of mining sites, plantations, and forests do account for part of capital formation.

Capital formation in construction is estimated by subtracting expenditures on current maintenance and repairs from the total value of output of new construction during the year. Pucca construction is calculated on the basis of the commodity flow method, while kutcha construction uses the expenditure approach.

Capital formation for machinery and equipment is calculated through the commodity flow method. The various items of machinery and equipment, domestically produced, imported and exported are classified as: (1) capital goods; (2) parts of capital goods ; (3) partly capital goods; and (4) parts of partly capital goods. The first category is 100% capital formation, the second is 50%, and in the third and fourth, the ratio is determined according to the item.

Gross capital formation is also estimated according to the industry of use through the expenditures approach method. Almost 70% of the organized sector is estimated on the basis of direct annual data, but there are many problems in compiling estimates for the unorganized sector. Estimates of registered manufacturing utilize ASI factory sector data, whose reliability is high. However, as gross capital formation of unregistered manufacturing is concerned, the estimates for 1980-81 were first prepared using the sample data of 1968-69 and 1971-72, which were simply applied to later years.

(6) Capital Stock

The pioneering work for estimating capital stock is "An Estimate of the Reproducible Tangible Wealth of India," Review of Income and Wealth, Series 8, 1959, by M. Mukherjee and N.S.R. Shastry, which estimated net fixed capital formation at the end of 1949-50. Following the Report of the Working Group on Savings, under the chairmanship of K.N.

Raj, 1982, CSO introduced the perpetual inventory method to carry forward the above bench-mark estimates of net fixed capital formation for 1949-50, using the official estimates of net fixed capital formation. In 1988, figures for capital stock for the end of March 1981 were announced along with the introduction of a new series.

4. The Tasks Ahead

The essay has presented certain particular features and problems involved in compiling national income statistics for India. We will conclude by mentioning two major points related to the major problems which remain.

First, as has long been the case, nearly 40% of the data used to compile GDP statistics does not draw on direct current data for the current year. For 1985-86, the use of direct current data for the current year for mining, registered manufacturing industries, electrical power, railways, communications, and banking and insurance is nearly 100%, but for the economy as a whole it is only 62.2% (CSO, National Accounts Statistics: Sources and Methods, 1989, Table 1.1). A substantial portion of GDP estimates continue to be based on data other than current direct data for the current year. The lack of objective basis for these estimates may well mean that political and administrative factors bias their compilation.

Second, even when current direct data is available, value added tends to be under-reported. This problem is especially acute in taxed sectors of the economy where it is very likely that production costs are inflated and sales are under-reported. For this reason, it is almost certain that India's GDP is underestimated to some extent.

Makoto Kojima
Chiba University of Commerce
Faculty of Commerce and Economics