ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Revision of National Accounts Statistics

The Central Statistical Organisation's revision of the National Accounts Statistics series, by shifting the base year from 1993-94 to 1999-2000 is a welcome step. What is missing in this so-called comprehensive revision is any reference to the serious decline in the quality and reliability of the statistics that form the building blocks for the NAS, as revealed by the National Statistical Commission. There have been attempts though to improve the existing database sources and all changes, except for two, namely, the introduction of a new category of "valuables" in the estimation of gross capital formation and the alignment of industry-wise estimate of GCF with estimation by asset or institution, are to be welcomed.

����������������

Revision of National Accounts Statistics

A Welcome Step, Good in Parts

The Central Statistical Organisation’s revision of the National Accounts Statistics series, by shifting the base year from 1993-94 to 1999-2000 is a welcome step. What is missing in this so-called comprehensive revision is any reference to the serious decline in the quality and reliability of the statistics that form the building blocks for the NAS, as revealed by the National Statistical Commission. There have been attempts though to improve the existing database sources and all changes, except for two, namely, the introduction of a new category of “valuables” in the estimation of gross capital formation and the alignment of industry-wise estimate of GCF with estimation by asset or institution, are to be welcomed.

S L SHETTY

IIIII
IntroductionIntroductionIntroductionIntroductionIntroduction

T
he Central Statistical Organisation’s revision of the National Accounts Statistics (NAS) series by shifting the base year forward to 1999-2000 from the earlier 1993-94 base, is indeed a welcome step. Such a revision of the base year for the NAS series within a six-year interval deserves to be commended as it contrasts with all other index number series for wholesale and consumer prices,1 agricultural and industrial output and export-import trade, which have remained unchanged for more than a decade or even two. At the time of the 1993-94 revision itself, the CSO had agreed that it was possible from then on to bring forward the base year for NAS purposes every five to six years with a view to capturing the rapidly changing structural features of the economy. This is because of the availability of nationwide quinquennium survey results of the National Sample Survey Organisation (NSSO) on employment and unemployment and household consumption expenditures which, along with other surveys on demography, land and livestock, manufacturing, services sectors, and household debt and investment, have provided a rich source material, though secondary in character, for the estimation of different components of national accounts.

While the CSO-NSSO’s efforts in the above respects are thus commendable, the absence of uniformity in setting some common base-year periods for the index number series covering different economic statistics certainly speaks poorly of coordination at the government level, despite the growing need for more accurate and timely statistics for planning and policy formulation purposes. This lapse no doubt appears minor when compared with persistent delays in implementing the recommendations of the National Statistical Commission (NSC) [NSC 2001; chairman: C Rangarajan], which have advanced suggestions for improvement in every component of the Indian statistical system.

IIIIIIIIII
Basic Weaknesses in Primary Data RemainBasic Weaknesses in Primary Data RemainBasic Weaknesses in Primary Data RemainBasic Weaknesses in Primary Data RemainBasic Weaknesses in Primary Data Remain
to be Addressedto be Addressedto be Addressedto be Addressedto be Addressed

The entire set of national income and related aggregates and accounts are derived statistics, based as they are on the basic data available from a multitude of primary sources. While the CSO’s picturesque presentation of varied accounts under the NAS appears very sophisticated, guided as it is by the framework set out in the UN System of National Accounts (UN SNA), there is wide concern that the quality of official statistics generated for different sectors of the economy, constituting the building blocks for the Indian NAS, leaves much to be desired as they have deteriorated over a fairly long period of years. This is as much true of the organised segment of the manufacturing, corporate and non-banking financial sectors as it is of the unorganised segments in manufacturing as well as the vast and growing services sector.2 In this respect, the CSO’s claim in the latest brochure on the revised NAS [CSO 2006] that it has undertaken an exercise variously described as “a complete review” or “a comprehensive review of both the databases and the methodology employed in the estimation of various macroeconomic aggregates,” appears untenable. What the CSO has apparently attempted is to undertake a series of improvements by and large with the existing basic data available from primary sources, particularly in respect of agricultural and industrial sectors.

What is missing in the so-called comprehensive revision exercise is any reference to the serious decline in the quality and reliability of statistics generated in respect of these important sectors as revealed by the NSC. The commission had clearly opined that their recommendations with regard to improving the national

Economic and Political Weekly June 10, 2006 statistical system would have an important bearing on the issue of national income estimation, that “the recommendations in respect of agriculture, industry, trade and services sector would also contribute towards improving the quality, reliability and timeliness of direct estimates (of national income)” [NSC 2001:358, Vol II]. It added that these recommendations “in respect of official statistics relating to different sectors of the economy be implemented speedily so as to improve the quality of data going into the compilation of National Accounts Statistics from primary source agencies” [NSC 2001:392]. Therefore, at this stage of statistical development, the derived macroeconomic aggregates dependent on sectoral data should have the nature and extent of improvements achieved based on NSC recommendations as the starting point. On the other hand, there is sufficient evidence that shows the recommendations of the Rangarajan Commission have hardly been implemented in any of the ministries/departments of the government of India which produce primary real sector data [Shetty 2006]. When this is the position, the CSO cannot claim to have undertaken a complete review of the database, though technically it cannot be blamed for the weaknesses of sectoral data; the NAS constructed by it can only be as good as the data provided by the source agencies.

In agriculture, for instance, though statistics of crop production

– both area and yield – had been based on scientifically designed methodologies, “the quality of land use and crop output data has suffered seriously for a variety of reasons” [NSC 2001:10, Vol I]. It is estimated that area enumeration in respect of about 40 per cent of the land under major agricultural crops is unscientific. Yield measurements in crops, based on crop-cutting experiments, are in bad shape “on account of the poor performance of field operations” [Rangarajan 2002; see also DoS 1999]. As for industrial statistics, “the data generated by the ASI system based upon this deficient ASI frame do not therefore depict the true situation of organised industrial sector” [Rangarajan 2001]. Also, the functioning of the source agencies providing the primary data of industrial production to the CSO is afflicted with a number of serious deficiencies [Rangarajan 2001; see also Shetty 2006]. The private corporate sector data used for estimating, particularly, saving and capital formation are known to be faulty for want of a sound census frame and due to the absence of current data. This is equally true of the data for non-banking financial companies (NBFCs) and unorganised financial enterprises. When such are the obvious blemishes in the database for different sectors and when all of these are yet to be addressed, the CSO cannot adopt an ostrich-like approach. It is better that it offers a more sober and restrained claim about its revision of the database for NAS purposes and clearly indicates the sectors for which the NSC recommendations have been implemented and the sectors for which the suggested improvements are yet to take place.

IIIIIIIIIIIIIII
Concerted Attempts at Improving ExistingConcerted Attempts at Improving ExistingConcerted Attempts at Improving ExistingConcerted Attempts at Improving ExistingConcerted Attempts at Improving Existing
Database SourcesDatabase SourcesDatabase SourcesDatabase SourcesDatabase Sources

The above is not intended to belittle the importance of extensive exercises undertaken by the CSO to increase the coverage of items; change the procedures and methodology of NAS compilation; implement the 1993 SNA recommendations, make studied adjustments to the normally survey-based estimates of the workforce as well as average value added per worker in respect of the unorganised segments of manufacturing (non-SSI) and services sectors, which have been otherwise most intractable; organise type and other studies from states and independent agencies to update various rates and ratios used; adopt detailed economic and purpose classifications of government finance accounts; and assume consultative processes in effecting changes in methodology and use of alternative data sources. No doubt, these changes have been of a significant nature, and except for two – the introduction of a new category of “valuables” in the estimation of gross capital formation (GCF) and alignment of the industry-wise estimation of GCF with the estimation by asset or institution (more on it later) – all changes deserve to be welcomed.

Discrepancies between ‘Accounts’ ContinueDiscrepancies between ‘Accounts’ ContinueDiscrepancies between ‘Accounts’ ContinueDiscrepancies between ‘Accounts’ ContinueDiscrepancies between ‘Accounts’ Continue
To Be a WorryTo Be a WorryTo Be a WorryTo Be a WorryTo Be a Worry

The above changes, while they are wide-ranging in nature and quite meticulously done, cannot, it appears, contribute to any significant improvement in the NAS system unless the NSC recommendations are put in place. That is, as stated above, “to improve the quality of primary data collected by the reporting source agencies” [Rangarajan 2002]. As indicated in the next section, changes in the levels of GDP due to the revision in the national accounts series have been minimal and there have not been any noticeable alterations in the overall growth of real GDP either.

But, a more potent measure of judging if the database revisions and procedural changes effected by the CSO in the new 1999-2000

Table 1: Comparison of ‘Discrepancies’ and ‘Errors and Omissions’ in SNA Accounts in New and Old SeriesTable 1: Comparison of ‘Discrepancies’ and ‘Errors and Omissions’ in SNA Accounts in New and Old SeriesTable 1: Comparison of ‘Discrepancies’ and ‘Errors and Omissions’ in SNA Accounts in New and Old SeriesTable 1: Comparison of ‘Discrepancies’ and ‘Errors and Omissions’ in SNA Accounts in New and Old SeriesTable 1: Comparison of ‘Discrepancies’ and ‘Errors and Omissions’ in SNA Accounts in New and Old Series

(in rupees crore)

1 Account 1 (GDP and Expenditure) New Old (1999-2000) (1993-94) Series Series 2 3 Account 3 (National Disposable Income and Its Appropriation) New Old (1999-2000) (1993-94) Series Series 4 5 New (1999-2000) Series (Without Valuables) 6 Account 5 (Capital Finance) New (1999-2000) Series (With Valuables) 7 Old (1993-94) Series 8
1999-2000 2000-01 2001-02 2002-03 2003-04 2004-05 -30,076 6,837 -542 27,357 44,147 53,017 -2,192 18,243 19,239 54,365 86,029 – -5,467 38,838 48,962 35,491 23,239 45,171 -12,915 22,207 28,481 11,663 449 – -2,925 -2,564 -30,731 1,494 26,502 50,310 12,594 12,160 -16,544 15,451 50,986 90,456 32,407 25,471 9,531 52,330 91,174 –
Source: CSO (2006).
2318 Economic and Political Weekly June 10, 2006

series have in fact contributed to an improvement in the quality of NAS, lies in the extent to which the series of “statistical discrepancies” or “errors and omissions”, have been pruned. These are generally noticed as existing between various accounts when an attempt is made to integrate them under the “consolidated accounts of the nation”. When the 1993-94 base revision took place, then revised series had tended to narrow the sizes of discrepancies in some accounts, while no such improvement occurred in some other accounts [EPWRF 2004]. Thus, discrepancies between GDP estimates and estimates of key expenditures on GDP (Account 1 under UN-SNA) and those between national disposable income and its appropriation (Account 3) had reduced from the 1980-81 to the 1993-94 series. On the other hand, the “errors and omissions” under the capital finance account (Account 5), juxtaposing sources of finance for capital accumulation and estimation of capital accumulation itself, had remained stubbornly high. There is also another set of discrepancies largely specific to the Indian NAS, which relate to the differences between the estimation of capital formation by type of assets and that by industry of use; these discrepancies too had widened after the 1993-94 revision.

The CSO has just released the details of some components under the system of “consolidated accounts of the nation” as per the 1999-2000 series compared with the 1993-94 series. In the existing 1993-94 series, the aforesaid discrepancies in almost all accounts had begun to increase in recent years and more significantly, tended to widely fluctuate from year-to-year [CSO 2005] casting misgivings on the overall quality of estimates, probably because of the differing and deteriorating quality of data sources. In the latest revision with 1999-2000 as the base, there is no evidence of any systematic reduction in discrepancies between accounts except for Account 1 (GDP and Expenditure) (Table 1). The information available in respect of “errors and omissions” in the capital finance account (Account 5) concerning capital accumulation and its sources, presents an interesting story. No doubt, these “errors and omissions” in the new series have come down from a range of Rs 9,531 crore to Rs 91,174 crore for the period 1999-2000 to 2003-04 as per the 1993-94 series, to a range of (-) Rs 30,731 crore to Rs 26,502 crore for the same period.

But, apart from behaving in an unstable manner, these estimates of discrepancies under the new series have a major flaw in that they are partly the result of the inclusion of “valuables” in gross capital formation (GCF), without a counterpart incorporation in domestic saving to finance such investment. As explained in a subsequent section, if the equivalent of “valuables” is added to the estimation of domestic savings, the consequential errors and omissions almost double. In respect of Account 3 concerning national disposable income and its appropriation, “statistical discrepancies” have substantially increased between the old and the new series; of course, in this case, “valuables” have been included neither under personal consumption nor under domestic saving. What is more, the new series has sought to duck the discrepancies between GCF estimates obtained from the commodity flow approach and those obtained by industry of use. By assuming that gross fixed capital formation (GFCF) estimates from the commodity flow approach for the entire economy are firmer estimates, the CSO has adjusted the industry-wise GFCF estimates compiled by the expenditure method in respect of private corporate and household sectors proportionately with the estimates compiled by institutions through the commodity flow method. Thus, the discrepancies have been obliterated by a definitional fiat. Therefore, the improvement in the database and the methodology of estimation is not reflected in any noticeable narrowing of different types of discrepancies in national accounts.

IVIVIVIVIV
Significant Change in Sectoral GDP SharesSignificant Change in Sectoral GDP SharesSignificant Change in Sectoral GDP SharesSignificant Change in Sectoral GDP SharesSignificant Change in Sectoral GDP Shares

As the CSO itself has emphasised, at the aggregate level, the changes in the levels of GDP following the introduction of the new series appear minimal and range from 0 to 1.7 per cent. More significantly, there are hardly any changes in the overall growth rates of real GDP between the new and old series. Except for one year 2002-03, the growth rates appear identical for four years between 2000-2001 and 2003-04. There appears to be a major change in 2004-05, which is because of the database differences between the earlier advance estimates for the year based on indicators and the subsequent quick estimates based on more detailed data at the sectoral level.

But there are some noticeable differences at the sectoral level, in which there is also a distinct trend noticed. That is, as shown in Table 2, estimates for GDP originating in agriculture and allied activities have been consistently lower in the new series as compared with the old and this negative difference has rapidly grown during 1999-2000 to 2004-05 (ranging from (-)1.7 per cent to (-)7 per cent), while those for the industrial sector have shown a negligible difference. However, the revision has produced a uniformly positive impact on all components of the services sectors. Consequently, the relative share of agricultural and allied activities in aggregate GDP has dipped much more sharply, by about 1 to 1.5 percentage points, in the new series as compared with that in the old series and that of industry has remained at the old level. Thus, the losses suffered by agriculture

Table 2: Gross Domestic Product at Factor Cost by Industry atTable 2: Gross Domestic Product at Factor Cost by Industry atTable 2: Gross Domestic Product at Factor Cost by Industry atTable 2: Gross Domestic Product at Factor Cost by Industry atTable 2: Gross Domestic Product at Factor Cost by Industry at
Current Prices:Current Prices:Current Prices:Current Prices:Current Prices:
Difference between New (1999-2000)Difference between New (1999-2000)Difference between New (1999-2000)Difference between New (1999-2000)Difference between New (1999-2000)
and Old (1993-94) Seriesand Old (1993-94) Seriesand Old (1993-94) Seriesand Old (1993-94) Seriesand Old (1993-94) Series

(Percentage change)

Industry 2004 2003 2002 2001 2000 1999
05 04 03 02 01 2000
(QE) (PE)
Agriculture, forestry
and fishing -7.0 -7.1 -7.7 -4.5 -2.4 -1.7
Agriculture na -7.5 -8.3 -4.6 -2.1 -1.7
Industry 1.0 0.9 -0.3 -0.5 0.3 0.0
Mining and quarrying -6.0 4.3 1.1 1.0 0.8 0.7
Manufacturing -0.5 -0.6 -1.6 -1.5 -0.4 -1.0
Registered na 1.9 1.3 1.3 1.5 1.3
Unregistered na -5.6 -7.3 -7.1 -3.9 -5.0
Electricity, gas and water
supply 4.6 4.2 8.0 6.6 7.4 5.8
Construction 6.8 2.3 -0.6 -0.8 -0.9 -0.1
Services 3.3 4.5 3.7 4.0 4.0 4.5
Trade, hotels and
restaurants na 5.9 4.2 3.6 3.3 3.3
Transport, storage and
communication na 3.6 3.0 3.6 3.9 6.0
Finance, insurance,
real estate and business 4.6 8.6 6.7 8.1 6.3 5.5
Community, social and
personal services 1.2 -0.3 0.6 1.0 2.8 4.2
Total gross domestic product
at factor cost 0.5 0.9 0.0 0.8 1.4 1.7

Note: QE: quick estimates, PE: provisional estimates, na: not available.

Economic and Political Weekly June 10, 2006

Figure: Percentage Difference between New Series and OldFigure: Percentage Difference between New Series and OldFigure: Percentage Difference between New Series and OldFigure: Percentage Difference between New Series and OldFigure: Percentage Difference between New Series and Old
Series of GDP: Agriculture, Industry and ServicesSeries of GDP: Agriculture, Industry and ServicesSeries of GDP: Agriculture, Industry and ServicesSeries of GDP: Agriculture, Industry and ServicesSeries of GDP: Agriculture, Industry and Services

6.00

4.00

2.00

0.00 –2.00 –4.00 –6.00 –8.00

–10.00

1999-2000 2000-01 2001-02 2002-03 2003-04 2004-05

Agriculture Industry Services

Percentage Difference

in its share of GDP for the above years have been by and large gained by the services sectors.

While all service sector areas have gained in the latest revision of GDP, sizeably large increases are noticed in a few service sector areas; they are: hotels and restaurants, the railways, storage, communication, and “real estate, ownership of dwellings, business and legal services”. All of these have contributed to an augmentation of the services sector share in GDP. It is said that in almost all these segments, adoption of the latest data has contributed to the upward revision. For “hotels and restaurants”, value added per worker (VAPW) from the Enterprise Survey conducted during the 57th round (2001-02) of NSSO and workforce estimates from the 1999-2000 employment and unemployment survey (55th round) along with the Population Census 2001 have been used; these replaced the 50th round (1993-94) data employed along with the Population Census 1991 for the old series. “Communication services” have experienced considerably improved coverage under private communication, including separate estimates for cellular mobile, courier services and public call offices (PCO). Apart from increases in GVA estimates from real estate and ownership of dwellings due to the adoption of latest data, a new item, “renting of machinery and equipment without operator” has been covered for the first time in the new series, using the VAPW and workforce in the activity from the results of the NSSO 57th and 55th rounds, respectively.

It appears from the above results for various segments of the services sectors, that the CSO/NSSO have achieved considerable advance in updating the database for the services sectors. In the absence of direct estimates, the use of a variety of economic censuses, enterprise and other quinquennium or periodic surveys, combined with the indicators approach, has enabled the CSO to adopt probably the second-best method for estimating GDP of the services sectors. There may be further scope for improving the survey results so as to minimise sampling and non-sampling errors. There are also certain flaws in the sample survey procedures, as explained in a subsequent section. The surveys themselves are done once in five years by the NSSO and they also exclude trading activity, which is the largest segment amongst the services sectors. The quinquennium benchmark estimates are projected through extrapolation with the help of indicators for specific service sectors. Consequently, successive survey results tend to depart from the earlier survey results. Despite all this, it could be said that attempts to improve the database for the services sectors are in the right direction.

Contrariwise, it is the gap in the quality of real sector data that stands out. For instance, a closer examination of the changes effected in the estimation procedure suggests that three crucial factors have contributed to the losses in the relative share of agriculture in total GDP. First, in the revised series, the price data furnished by states for non-procured quantities of paddy and wheat have been used, as distinct from the minimum support prices (MSP) that were used earlier when the former were lower than the latter. Second, in the case of fruits and vegetables, a sub-committee has recommended that the CSO use the National Horticultural Board (NHB) data and not the data collected under the centrally-sponsored plan scheme, ‘Crop Estimation Survey on Fruits, Vegetables and Minor Crops’, because it had not yet been evaluated in terms of the quality of data generated. Both of these have meant some reductions in agricultural output and GDP originating in it. These approaches, particularly the one on farm prices, are contrary to what has been recommended by the NSC. The NSC (2001:385) report states thus: “Sometimes, it is observed that these prices reported by the states are under-stated, as devolution of central funds, to some extent, depends on the estimated per capita state income. It is necessary that such price data is provided by the Directorate of Economics and Statistics, Ministry of Agriculture (DESMOA), which would then be free from bias or alternatively the devolution of funds may be de-linked from the per capita income of the states.”

These brief observations are in addition to the general point that the primary source data for agriculture are yet to be improved. A third factor that has been responsible for the reduction in agricultural GDP has emanated from the input cost in the form of market charges, which have doubled from 1.29 per cent output of agricultural crops to 2.358 per cent based on a fresh market study undertaken by the ministry of agriculture, which appears to be an improvement. While on input costs, it may be noted that with the upward revision of the yield rate for fodder crops, their value of output has substantially gone up but it has not affected the estimate of agricultural GDP as the entire output is treated as input, namely, feed of livestock, in the agriculture sector.

As for industry, the primary data sources are truly in a shambles, as evident in the fact that the CSO itself has not found it worthwhile to use the ASI data for national accounts purposes beyond 1999-2000 [CSO 2006:16]. In the absence of ASI data, the index of industrial production (IIP) is used for extrapolation of the benchmark data of ASI for more number of years than in the past. But, studies have also shown how there are serious gaps in the coverage of industrial units for the IIP.

It is because of these drawbacks in the real sector database that there is a lurking misgiving that the higher, and growing, relative share of the services sectors in GDP in recent years could also be attributable to these weaknesses. While it is difficult to foresee how the output of agricultural and industrial sectors will fare when improved data are put in place for them, there is enough corroborative evidence in the NSC report as well as in independent research studies to show that, at any rate, the output of the organised manufacturing, which forms the basis for the CSO’s estimation of GDP for the sector, is underestimated in official statistics.

Differing NSSO Survey ResultsDiffering NSSO Survey ResultsDiffering NSSO Survey ResultsDiffering NSSO Survey ResultsDiffering NSSO Survey Results

Before concluding this section, it must be pointed out that a preponderant part of the upward or downward revisions in the sectoral estimates of GDP as well as PFCE are hardly based on improved coverage of production or consumption items. They are rather based on fresh NSSO survey results, which have generally tended to be different from the earlier survey results, or the inter-survey growth rates used earlier have been disproved by the fresh survey results. These bring home the basic fact that the NSSO survey results or the methodology of using them for estimational purposes, has not stabilised as yet. It cannot, for if the NSC report is to be believed, the existing approach adopted for data collection through follow-up Enterprise Surveys on different segments of the services sector has a basic flaw in that it seeks to collect data for all types of enterprises, irrespective of their size. It is said that such integrated surveys do not ensure adequate representation to big-size units, which tends to distort the estimated results. The NSC has, therefore, recommended a survey of non-manufacturing industries for bigger units, say with at least 10 workers (other than those in the public sector) and a separate sample survey for the residual category of smaller units similar to the existing follow-up Enterprise Surveys. To be able to do this on a regular basis, it is necessary to have a sample frame and for this the NSC has recommended the building up, though in phases, of a nationwide business register for the defined large units [NSC 2001:194-95, 635-38, Vol II].

VVVVV
Coverage and Procedural Changes in SavingCoverage and Procedural Changes in SavingCoverage and Procedural Changes in SavingCoverage and Procedural Changes in SavingCoverage and Procedural Changes in Saving
and Capital Formation Estimatesand Capital Formation Estimatesand Capital Formation Estimatesand Capital Formation Estimatesand Capital Formation Estimates

More than the coverage of production and consumption data, it is the procedural changes and that in the coverage of estimating saving and capital formation which have produced far-reaching alterations in the latest revision in NAS. The estimates have catapulted the Indian economy from the image of a low saving and low investment economy to that of a high saving and high investment economy, that is, from a saving-investment level of 23-24 per cent of GDP to that of 29-30 per cent of GDP. This sudden upgradation of the saving-investment scenario has been achieved in the new series of NAS through a number of estimational devices, which require closer scrutiny. Before doing so, let us look at the artefacts of changes in saving and investment rates.

First, upward revisions had been introduced of late in domestic saving rates in the old series itself, which were closely examined in literature [Shetty 2005 and Chaudhuri 2005]. Now, the domestic saving rates have been further revised upwards in the new series. As shown in Table 3, gross domestic saving as a percentage of GDP at current market prices stands pulled up by about one percentage point (from a range of 24 to 28 per cent in the old series to 25 to 29 per cent in the revised series) in each of the years from 1999-2000 to 2003-04. Also, within the revised series, the gross saving rate has risen somewhat faster from 24.9 per cent to 28.9 per cent during the above five-year period and it has further risen to 29.1 per cent in 2004-05.

Table 3: Estimations of Gross Domestic SavingTable 3: Estimations of Gross Domestic SavingTable 3: Estimations of Gross Domestic SavingTable 3: Estimations of Gross Domestic SavingTable 3: Estimations of Gross Domestic Saving
and Capital Formation at Current Pricesand Capital Formation at Current Pricesand Capital Formation at Current Pricesand Capital Formation at Current Pricesand Capital Formation at Current Prices

(in rupees crore)

Year New Old As Percentage of 1999-2000 1993-94 Respective GDP Series Series New Series Old Series

AAAAA
Gross Domestic SavingGross Domestic SavingGross Domestic SavingGross Domestic SavingGross Domestic Saving
1999-00 487,301 468,681 24.9 24.2 2000-01 496,272 490,049 23.5 23.5 2001-02 537,966 532,274 23.6 23.4 2002-03 648,994 642,298 26.5 26.1 2003-04 797,512 776,420 28.9 28.1 2004-05 907,416 – 29.1 –

BBBBB
Gross Domestic Capital Formation*Gross Domestic Capital Formation*Gross Domestic Capital Formation*Gross Domestic Capital Formation*Gross Domestic Capital Formation*
1999-00 509,289 490,669 26.0 25.3 2000-01 509,026 498,179 24.2 23.8 2001-02 523,737 513,543 23.0 22.6 2002-03 620,508 610,288 25.3 24.8 2003-04 752,132 726,868 27.2 26.3 2004-05 939,555 – 30.1 –

Notes: * Including valuables in the new series and also adjusted for errors and omissions. Old series are not available beyond 2003-04.

Source: CSO (2006).

Table 4: Estimates of GDS by Institutional Sectors,Table 4: Estimates of GDS by Institutional Sectors,Table 4: Estimates of GDS by Institutional Sectors,Table 4: Estimates of GDS by Institutional Sectors,Table 4: Estimates of GDS by Institutional Sectors,
1999-2000 to 2004-051999-2000 to 2004-051999-2000 to 2004-051999-2000 to 2004-051999-2000 to 2004-05

(in rupees crore)

Sector 2004-2003-2002-2001-2000-199905 04 03 02 01 2000

AAAAA
Comparison of Old and Revised EstimatesComparison of Old and Revised EstimatesComparison of Old and Revised EstimatesComparison of Old and Revised EstimatesComparison of Old and Revised Estimates

1 Public sector Old (1993-94) series -9,429 -26,652 -61,912 -48,361 -20,049 New series 69,390 28,026 -16,181 -46,377 -37,062 -16,659 Difference (new-old) 37,455 10,471 15,535 11,299 3,390

2 Private corporate sector Old (1993-94) series 114,157 94,269 81,076 86,142 84,329 New series 150,947 120,852 99,767 81,669 87,017 87,234 Difference (new-old) 6,695 5,498 593 875 2,905

3 Household sector

3.1 Financial saving Old (1993-94) series 314,261 254,439 253,964 216,774 205,743 New series 320,777 316,444 253,256 247,476 215,219 206,602 Difference (new-old) 2,183 -1183 -6,488 -1,555 859

3.2 Saving in physical assets Old (1993-94) series 357,431 320,242 259,146 235,494 198,658 New Series 366,302 332,190 312,152 255,198 231,098 210,124 Difference (new-old) -25,241 -8090 -3,948 -4,396 11,466

4 Total economy Old (1993-94) series 776,420 642,298 532,274 490,049 468,681 New series 907,416 797,512 648,994 537,966 496,272 487,301 Difference (new-old) 21,092 6,696 5,692 6,223 18,620

BBBBB
GDS as Percentage of GDP at Current Market PricesGDS as Percentage of GDP at Current Market PricesGDS as Percentage of GDP at Current Market PricesGDS as Percentage of GDP at Current Market PricesGDS as Percentage of GDP at Current Market Prices

1 Public sector
Old (1993-94) series -0.34 -1.08 -2.73 -2.31 -1.04
New series 2.22 1.02 -0.66 -2.03 -1.76 -0.85
Difference (new-old) 1.36 0.42 0.69 0.56 0.18
2 Private corporate sector
Old (1993-94) series 4.14 3.83 3.57 4.12 4.35
New series 4.84 4.38 4.07 3.58 4.13 4.45
Difference (new-old) 0.24 0.25 0.01 0.01 0.10
3 Household sector
3.1 Financial saving
Old (1993-94) series 11.39 10.33 11.18 10.37 10.62
New series 10.28 11.46 10.34 10.85 10.21 10.55
Difference (new-old) 0.08 0.01 -0.33 -0.16 -0.08
3.2 Saving in physical assets
Old (1993-94) series 12.95 13.00 11.41 11.27 10.26
New series 11.74 12.03 12.74 11.19 10.96 10.73
Difference (new-old) -0.92 -0.26 -0.22 -0.31 0.47

4 Total Economy Old (1993-94) series 28.13 26.07 23.43 23.45 24.20 New series 29.07 28.89 26.49 23.58 23.55 24.88 Difference (new-old) 0.76 0.42 0.15 0.09 0.68

Source: CSO (2006).

Economic and Political Weekly June 10, 2006

Second, as shown in Table 4, the upward revision in total savings numbers for all the five years from 1999-2000 to 200304 is due to sizeable increases in the saving estimates of the public sector, which, as explained below, have been brought about by the newer economic and purpose-wise classification of government expenditure. Office expenses have been classified for the first time as between current and capital expenditures and losses of departmental enterprises have been treated as subsidies. Besides, undistributed profits accrued to the unit holders have been excluded from the saving of UTI and included under household saving. But, apart from these estimational changes, the improvement in public saving, particularly in years 2003-04 and thereafter, has been buttressed by a steady decline in revenue deficits of the central and state governments. There has been some upward revision in the savings of the private corporate sector too, but both the financial and physical components of household sector savings have faced downward revisions, with the physical assets components experiencing generally steeper revisions.

Finally, within the public sector, the upward revisions in savings estimates have occurred under departmental and nondepartmental commercial enterprises but the bulk of change has occurred under departmental enterprises (Table 5) due to a major change effected in the treatment of operating losses as imputed subsidies instead of treating them as a negative operating surplus. For administrative departments, there has been a significant improvement in the year 2003-04, and probably thereafter, following the decline in revenue deficits, as shown in Table 6.

As for capital formation estimates, as shown in Table 3 earlier, there are no dramatic changes between the old and the revised series at the aggregate level. As in the case of domestic saving rates, the gross capital formation (GCF) rates are higher by about 1 percentage point in each of the five years, 1999-2000 to 2003-04. However, there are significant differences in capital formation estimation procedures, which have led to differences in the composition of GCF between the old and new series. The foremost one concerns the category of “valuables”, included for the first time as part of GCF in the revised series. This category covers expenditures on such valuable goods as gold, silver, precious stones and jewellery. As may be seen in Table 7, this new category alone accounts for the differences in GCF estimates at the aggregate level between the old and new series (more on it later).

The second major difference concerns the GCF estimation by institutional categories. As depicted in Table 8, GCF estimates attributable to the public sector and the private corporate sector have been considerably revised upwards, the former by about 9 to 17 per cent and the latter by about 12 to 52 per cent during the past five-year period. On the other hand, GCF estimates for the household sector have been revised downwards in almost all the years except for the base year 1999-2000 when there was a 6 per cent upward revision for this sector too. No doubt, the downward revisions in respect of the household sector have been meagre, ranging from (-)2 per cent to (-)7 per cent (Table 10). The upward revisions in the GCF estimates of public and private corporate sectors have to perforce lead to a downward revision in GCF estimates for the household sector because the latter are derived as a “residual”. Even so, downward revisions in the estimates of household estimates have been meagre because there have been upward revisions in the estimates of aggregate GCF too. As a result of the revisions made above for the institutional categories, private corporate sector investment as percentage of GDP has shot up from a range of 4 to 6 per cent to 5.50 per cent to 8.50 per cent. The public sector investment rate too has gone up by about 1 to 2 percentage points, while the household sector rate has experienced some fractional fall (Table 9).

As is widely known, the composition of GCF by institutional categories is done only at the unadjusted level. Hence, it does not cover the “errors and omissions”, which are used for adjusting the aggregate GCF estimated through the commodity flow method, against the yardstick of investible funds available through domestic saving plus capital inflow from abroad. This brings us to the third aspect of change in GCF estimates, which is that in the new series the size of such “errors and omissions” has been drastically brought down from the old series. For

Table 5: Gross Savings of the Public Sector, 1999-2000 toTable 5: Gross Savings of the Public Sector, 1999-2000 toTable 5: Gross Savings of the Public Sector, 1999-2000 toTable 5: Gross Savings of the Public Sector, 1999-2000 toTable 5: Gross Savings of the Public Sector, 1999-2000 to
2004-05: A Comparison of Old and Revised Estimates2004-05: A Comparison of Old and Revised Estimates2004-05: A Comparison of Old and Revised Estimates2004-05: A Comparison of Old and Revised Estimates2004-05: A Comparison of Old and Revised Estimates

(in rupees crore)

2004-2003-2002-2001-2000-199905 04 03 02 012000

1 Public sector Old (1993-94) series -9,429 -26,652 -61,912 -48,361 -20,049 New (1999-00) series 69,390 28,026 -16,181 -46,377 -37,062 -16,659 Difference (new-old) 37,455 10,471 15,535 11,299 3,390

  • (i) Administrative dept and quasi government bodies Old (1993-94) series -118,744-126,341 -137,772 -113,125 -96,151 New (1999-00) series -84,552 -101,206-128,194 -137,845 -115,491 -98,409 Difference (new-old) 17,538 -1,853 -73 -2,366 -2,258
  • (ii) Departmental enterprises Old (1993-94) series 2,417 3,104 578 4,905 11,099 New (1999-00) series 16,468 14,494 13,119 12,588 17,281 23,277 Difference (new-old) 12,077 10,015 12,010 12,376 12,178
  • (iii) Non-departmental enterprises

    Old (1993-94) series 106,898 96,585 75,282 59,859 65,003 New (1999-00) series 137,474 114,738 98,894 78,880 61,148 58,473 Difference (new-old) 7,840 2,309 3,598 1,289 -6,530

    Source: CSO (2005, 2006).

    Table 6: Trend in Revenue Deficit,Table 6: Trend in Revenue Deficit,Table 6: Trend in Revenue Deficit,Table 6: Trend in Revenue Deficit,Table 6: Trend in Revenue Deficit,
    Combined forCombined forCombined forCombined forCombined for
    Centre and StatesCentre and StatesCentre and StatesCentre and StatesCentre and States

    Year Amount (Rs Crore) Per Cent to GDP

    1999-2000 121,393 6.2 2000-01 138,803 6.6 2001-02 159,350 7.0 2002-03 162,990 6.7 2003-04 159,500 5.8 2004-05(RE) 128,355 4.1 2005-06(BE) 119,806 3.4

    Source:Handbook of Statistics on the Indian Economy 2004-05, RBI, 2005.

    Table 7: Role of ‘Valuables’ in New SeriesTable 7: Role of ‘Valuables’ in New SeriesTable 7: Role of ‘Valuables’ in New SeriesTable 7: Role of ‘Valuables’ in New SeriesTable 7: Role of ‘Valuables’ in New Series

    (Percentage of GDP at current market prices)

    Year GCF Estimates ‘Valuables’

    New Series Old Series Difference Included in the (2)–(3) New Series

    (1) (2) (3) (4) (5)

    1999-2000 26.0 25.3 0.7 0.8 2000-01 24.2 23.8 0.4 0.7 2001-02 23.0 22.6 0.4 0.6 2002-03 25.3 24.8 0.5 0.6 2003-04 27.2 26.3 0.9 0.9 2004-05 30.1 – – 1.3

    Source: CSO (2006).

    1999-2000, it has come down from 1.67 per cent to (-) 0.15 per cent, and for 2003-04, from 3.30 per cent to 0.96 per cent of GDP (Table 9). As commented upon at a later stage, a part of this has been due to the inclusion of “valuables” as a new item in the estimates of GCF (Table 10). No doubt, even after accounting for the “valuables”, there is some reduction in “errors and omissions” due probably to the improvements in estimation procedures.

    A third and final major revision in the GCF estimates emanates from a heroic methodological change effected in the new NAS with the intention of eliminating discrepancies between gross fixed capital formation (GCFC) estimates made by the commodity flow method and the industry-wise GFCF estimates made by the expenditure method. What the CSO has done is to assume that GFCF estimates obtained from the commodity flow method for the entire economy as firm estimates and adjust the industry-wise GFCF estimates for private corporate and household sectors, pro rata with the estimates compiled by institutional categories. This methodological change has brought about large

    Table 10: ‘Errors and Omissions’ and ‘Valuables’Table 10: ‘Errors and Omissions’ and ‘Valuables’Table 10: ‘Errors and Omissions’ and ‘Valuables’Table 10: ‘Errors and Omissions’ and ‘Valuables’Table 10: ‘Errors and Omissions’ and ‘Valuables’
    in GCF Estimationin GCF Estimationin GCF Estimationin GCF Estimationin GCF Estimation

    (Percentage to GDP at current market prices)

    Year Errors and Omissions ‘Valuables’

    New Series Old Series Difference Included in the (New-Old) New Series

    1999-2000 -0.2 1.7 -1.9 0.8 2000-2001 -0.1 1.2 -1.3 0.7 2001-2002 -1.4 0.4 -1.8 0.6 2002-2003 0.1 2.1 -2.0 0.6 2003-2004 1.0 3.3 -2.3 0.9 2004-2005 1.6 – – 1.3

    Note: For details, see Table 8; see also the text in Section V.

    Table 8: Capital Formation by Type of Institution at Current Prices: Difference between New and Old SeriesTable 8: Capital Formation by Type of Institution at Current Prices: Difference between New and Old SeriesTable 8: Capital Formation by Type of Institution at Current Prices: Difference between New and Old SeriesTable 8: Capital Formation by Type of Institution at Current Prices: Difference between New and Old SeriesTable 8: Capital Formation by Type of Institution at Current Prices: Difference between New and Old Series

    (in rupees crore)

    New (1999-2000) Series Old (1993-94) Series 2004-05 2003-04 2002-03 2001-02 2000-01 1999-00 2003-04 2002-03 2001-02 2000-01 1999-00 (QE) (PE) (QE)

    1 Gross capital formation 889,245 725,630 619,014 554,468 511,590 512,214 635,694 557,958 504,012 472,708 458,262

    1.1 Public sector 225,319 180,228 151,246 157,580 145,775 146,483 154,086 131,966 140,095 131,505 134,484

    1.2 Private corporate sector 257,478 188,728 141,659 127,503 119,993 140,088 124,177 105,750 104,771 105,709 125,120

    1.3 Household sector 366,302 332,190 312,152 255,198 231,098 210,124 357,431 320,242 259,146 235,494 198,658

    1.4 Valuables 40,146 24,484 13,957 14,187 14,724 15,519 2 Finances for gross capital formation 939,555 752,132 620,508 523,737 509,026 509,289 726,868 610,288 513,542 498,179 490,669

    2.1 Gross domestic savings 907,416 797,512 648,994 537,966 496,272 487,301 776,420 642,298 532,274 490,049 468,681

    2.2 Net capital inflow 32,139 -45,380 -28,486 -14,229 12,754 21,988 -49,552 -32,010 -18,732 8,130 21,988 3 Errors and omissions (2-1)’ 50,310 26,502 1,494 -30,731 -2,564 -2,925 91,174 52,330 9,531 25,471 32,407 4 Gross capital formation (1+3) 939,555 752,132 620,508 523,737 509,026 509,289 726,868 610,288 513,543 498,179 490,669

    (adjusted for errors and omissions)

    Absolute Difference between New and Old Series Percentage Difference between New and Old Series 2004-05 2003-04 2002-03 2001-02 2000-01 1999-00 2003-04 2002-03 2001-02 2000-01 1999-00 (QE) (PE) (QE)

    1 Gross capital formation 89,936 61,056 50,456 38,882 53,952 14.1 10.9 10.0 8.2 11.8

    1.1 Public sector 26,142 19,280 17,485 14,270 11,999 17.0 14.6 12.5 10.9 8.9

    1.2 Private corporate sector 64,551 35,909 22,732 14,284 14,968 52.0 34.0 21.7 13.5 12.0

    1.3 Household sector -25,241 -8,090 -3,948 -4,396 11,466 -7.1 -2.5 -1.5 -1.9 5.8

    1.4 Valuables 24,484 13,957 14,187 14,724 15,519 100.0 100.0 100.0 100.0 100.0 2 Finances for gross capital formation 25,264 10,220 10,195 10,847 18,620 3.5 1.7 2.0 2.2 3.8

    2.1 Gross domestic savings 21,092 6,696 5,692 6,223 18,620 2.7 1.0 1.1 1.3 4.0

    2.2 Net capital inflow 4,172 3,524 4,503 4,624 0 -8.4 -11.0 -24.0 56.9 0.0 3 Errors and omissions (2-1) -64,672 -50,836 -40,262 -28,035 -35,332 -70.9 -97.1 -422.4 -110.1 -109.0 4 Gross capital formation (1+3) 25,264 10,220 10,194 10,847 18,620 3.5 1.7 2.0 2.2 3.8

    (adjusted for errors and omissions)

    Source: CSO (2005, 2006).

    Table 9: Capital Formation by Type of Institution at Current PricesTable 9: Capital Formation by Type of Institution at Current PricesTable 9: Capital Formation by Type of Institution at Current PricesTable 9: Capital Formation by Type of Institution at Current PricesTable 9: Capital Formation by Type of Institution at Current Prices

    (Percentage to GDP at current market prices)

    New (1999-2000) Series 1993-94 Series 2004-05 2003-04 2002-03 2001-02 2000-01 1999-00 2003-04 2002-03 2001-02 2000-01 1999-00 (QE) (PE) (QE)

    1 Gross capital formation 28.49 26.29 25.27 24.30 24.27 26.15 23.03 22.65 22.18 22.62 23.66

    1.1 Public sector 7.22 6.53 6.17 6.91 6.92 7.48 5.58 5.36 6.17 6.29 6.94

    1.2 Private corporate sector 8.25 6.84 5.78 5.59 5.69 7.15 4.50 4.29 4.61 5.06 6.46

    1.3 Household sector 11.74 12.03 12.74 11.19 10.96 10.73 12.95 13.00 11.41 11.27 10.26

    1.4 Valuables 1.29 0.89 0.57 0.62 0.70 0.79 0.00 0.00 0.00 0.00 0.00 2 Finances for gross capital formation 30.10 27.25 25.33 22.96 24.15 26.00 26.34 24.77 22.60 23.84 25.33

    2.1 Gross domestic savings 29.07 28.89 26.49 23.58 23.55 24.88 28.13 26.07 23.43 23.45 24.20

    2.2 Net capital inflow 1.03 -1.64 -1.16 -0.62 0.61 1.12 -1.80 -1.30 -0.82 0.39 1.14 3 Errors and omissions (2-1) 1.61 0.96 0.06 -1.35 -0.12 -0.15 3.30 2.12 0.42 1.22 1.67 4 Gross capital formation (1+3) 30.10 27.25 25.33 22.96 24.15 26.00 26.34 24.77 22.60 23.84 25.33

    (adjusted for errors and omissions) 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00

    Source: CSO (2005, 2006).

    Economic and Political Weekly June 10, 2006 increases in the GCF estimates for all sectors of the economy. The largest and persistent change in GCF ranging from 42 per cent to 236 per cent has occurred under “construction” – an area for which the new series has made as explained below, a significant upward revision even before the proportionate adjustment. While the agricultural sector has been revised upwardly by 42 per cent to 72 per cent in investment, many areas of the services sector have been much more; similarly, for unregistered manufacturing. The logic of these relatively sharper upward revisions in respect of informal sectors lies in the fact that the entire NAS revision exercise seems to have updated the database for such sectors and this would have tended to revise upward the GCF estimates for these sectors even before the pro rata adjustment.

    VIVIVIVIVI
    Estimational Issues in SavingEstimational Issues in SavingEstimational Issues in SavingEstimational Issues in SavingEstimational Issues in Saving
    and Capital Formationand Capital Formationand Capital Formationand Capital Formationand Capital Formation

    As explained earlier, following the revision in the series, there have been sizeable increases in the savings of the public sector, some increase in those of the private corporate sector, and some downward revisions in those of the household sector. These revisions have been based on some changes in coverage and procedures effected in the estimational processes, which have been of a significant nature and except for some, appear realistic and should be welcomed. But, these revisions have hardly addressed the long-standing issues of improving the database for the sectoral estimations of saving and capital formation.

    Saving EstimationSaving EstimationSaving EstimationSaving EstimationSaving Estimation

    Admittedly, there are serious gaps in the database for the estimation of household savings (i) in deposits of non-banking financial companies (NBFCs), (ii) in informal sector financial assets, (iii) in currency holdings with the public (93 per cent of which is treated as part of household saving since 1985-86), and

    (iv) in shares of companies for want of regular surveys on ownership of company shares. In the case of the corporate sector, until the ministry of company affairs’ MCA 21 programme, which envisages online filing of accounts by companies, becomes a success and until those accounts are cleaned and processed, estimates of savings for the corporate sector will continue to depend on a weak database.

    No doubt, the procedural changes effected in classifying government expenditures as capital expenditures that impact saving estimation are substantial. They are: (i) head office expenditures of government departments, in which it is revealed that substantial expenditure is incurred on machines and equipments, furnitures, photocopiers, computer hardware, software and other accessories that were all earlier treated as consumption expenditure; (ii) inclusion of data in respect of some government schemes like Indira Awas Yojana (IAY), Prime Minister’s Grameen Sadak Yojana (PMGSY), in GFCF; and

    (iii) some parts of the outlays on schemes like Sarva Shiksha Abhiyan (SSA), district primary education programmes and mid-day meal schemes, which are now included in GFCF. Besides, the losses of departmental enterprises in power, manufacturing, transport, etc, were being treated as a negative operating surplus earlier. As recommended by the SNA 1993, these have now been treated as imputed subsidies and hence savings and other components of GDP have gone up in respect of the departmental enterprises, as shown earlier.

    It is not that saving estimations of household and corporate sectors have remained untouched; as brought out above, their numbers have undergone noticeable changes. In the case of the private corporate sector, saving estimates stand revised as a result of two factors: (i) revised earnings of foreign companies, a component of FDI in India, has been excluded from the saving of private corporate sectors and included as part of property and entrepreneurial income from the rest of the world; and (ii) upward revision in the estimates of gross saving of cooperative societies an quasi-corporate bodies on account of the availability of fresh data. Likewise, the estimates household sector savings have been revised due to four factors: (i) incomes accrued to households but not paid on account of investment in UTI units are now included in the estimates of household saving in the form of shares and debentures; earlier they were treated as part of public sector savings; (ii) incorporation of fresh data in respect of household investment in private insurance companies has been done since 2002-03; (iii) households’ net deposits have been revised downwards due to revision in the deposits of NBFCs and advances by cooperative societies to households; and (iv) finally, the household saving in capital assets has been substantially revised upwards due to the general upward revision of capital formation estimates.

    Capital Formation EstimatesCapital Formation EstimatesCapital Formation EstimatesCapital Formation EstimatesCapital Formation Estimates

    As shown earlier, capital formation estimation has undergone changes in all categories; by type of assets, by institutions and by industry of use. As per assets, substantial change has occurred in the estimation of new construction. Under the accounted (pucca) construction, inclusion of import of wood products, and use of ASI 1999-2000 results on basic materials, etc, have resulted in a sizeable change in estimation. The coverage of unaccounted (kutcha) construction has been expanded to include civilian construction in installing wind energy systems and seven more additional plantation crops in cultivated assets. As for machinery and equipment, the main changes introduced relate to: (i) preparation of a revised capital goods item basket using the detailed results of the Annual Survey of Industries, 1999-2000 based on NIC 1998 classification; and

    (ii) the results of the NSSO 56th round of unregistered manufacturing for 2001-02.

    Under institutional sectors, estimates for the public sector have been revised upwards mainly due to the reclassification of expenditures of government in the new series, as explained above under the subsection on saving estimates. Similarly, private corporate sector investment has gone up due to the inclusion of expenditures on software, wind energy systems, additional coverage of cultivated assets and capital expenditures made by new industries before commencing production. Finally, the rise in household sector capital formation has been mainly due to the overall increase in gross fixed capital formation, estimated through the commodity flow approach. This was in turn due to the increase in the output of construction, revision in the itembaskets of machinery and equipment as per NIC 1998 and the use of the latest available data on registered and unregistered manufacturing sectors.

    Two Key MisgivingsTwo Key MisgivingsTwo Key MisgivingsTwo Key MisgivingsTwo Key Misgivings

    All of the above changes obviously constitute measures of improvement in the estimation procedures, which deserve to be commended. But, there are two crucial changes in methodology and coverage effected, which raise misgivings about their tenability and hence call for their reconsideration.

    The first one concerns the pro rata adjustment made to the industry-wise estimates with the estimates made by the commodity-flow method. Amongst the three sets of assumptions involved in this adjustment, two, namely: (i) estimates made by the commodity flow method are more reliable than those made by the expenditure method industry-wise, and (ii) adjustments, if any, are required to be made in the estimates for private corporate and household sectors, appear uncontestable. What is contestable is the third assumption that the differences between the two methods are distributed industry-wise in proportion to their GCF estimates prior to the adjustment. There does not appear to be any basis for this, and as the sums involved are very large, the results may appear sensitive to the errors in respect of even a few industries. That apart, such pro rata adjustment of the industry-wise estimates has tended to hide the discrepancies found in the two sets of estimates and hence, deprived researchers and even official agencies of keeping a close watch on the differences in the two estimation procedures as the years roll by.

    The second most conspicuous, and apparently inadmissible method adopted, is the coverage given to a new category of “valuables” in the estimation of capital formation. This is attributed to the recommendations of 1993 SNA and has further enhanced the GCF rate by 1.3 per cent of GDP (or by Rs 40,146 crore) for 2004-05. The method adopted is a patently incorrect interpretation of the UN SNA recommendations. UN SNA 1993 had said that “Valuables are expensive durable goods that do not deteriorate over time, are not used up in consumption or production, and are acquired primarily as stores of value”. It had also said that acquisitions and disposition of “valuables” are to be recorded in the capital account, but not as a component of capital formation. The UN SNA systematically excludes “acquisitions less disposals of valuables” from the composition of gross fixed assets formation and locates it after “changes in inventories”. There are a number of other items in the capital account which do not form part of gross capital formation. It has been explicitly recognised that such net acquisitions of “valuables” do not contribute to production nor can there be any imputed earning or rentals for such assets.

    If the CSO’s method of including valuables as part of GCF is accepted, there ought to be a counterpart saving to finance such investment. Household saving estimation in India is not directly done as the difference between household income and consumption, rather it is done indirectly through the estimates of saving in the form of financial and physical assets formation without taking account of “valuables”. Hence, net acquisitions of valuables get completely ignored in the CSO’s saving estimation procedure, whether now or in the past. This contradicts CSO’s acceptance now of the UN SNA thesis that valuables ought not to be treated as part of consumption. To be consistent with the CSO’s treatment of valuables in GCF, an equivalent amount has to be added to the estimation of domestic saving. If this is

    Table 12: GCF Estimates sansTable 12: GCF Estimates sansTable 12: GCF Estimates sansTable 12: GCF Estimates sansTable 12: GCF Estimates sans
    ‘‘‘‘‘
    Valuables’ and ErrorsValuables’ and ErrorsValuables’ and ErrorsValuables’ and ErrorsValuables’ and Errors
    and Omissionsand Omissionsand Omissionsand Omissionsand Omissions

    (In rupees crore)

    Year Final GCF as Valuables Errors and GCF Estimated by Estimated Omissions Commodity by CSO Flow Method (2-3-4)

    (1) (2) (3) (4) (5)

    1999-00 509,289 26.0 15,519 -2,925 496,695 25.4 2000-01 509,026 24.2 14,724 -2,564 496,866 23.6 2001-02 523,737 23.0 14,187 -30,731 540,281 23.7 2002-03 620,508 25.3 13,957 1,494 605,057 24.7 2003-04 752,132 27.2 24,484 26,502 701,146 25.4 2004-05 939,555 30.1 40,146 50,310 849,099 27.2

    Note: Figures in italics are percentages to GDP at current market prices. Source: CSO (2006); col (5) is derived by us; see also Tables 10 and 11.

    (In rupees crore)

    Table 11: Impact of ‘Valuables’ on ‘Errors and Omissions’Table 11: Impact of ‘Valuables’ on ‘Errors and Omissions’Table 11: Impact of ‘Valuables’ on ‘Errors and Omissions’Table 11: Impact of ‘Valuables’ on ‘Errors and Omissions’Table 11: Impact of ‘Valuables’ on ‘Errors and Omissions’

    2004-05 2003-04 2002-03 2001-02 2000-01 1999-00
    1 Gross domestic capital formation (GDCF) 889,245 725,630 619,014 554,468 511,590 512,214
    including valuables (28.5) (26.3) (25.3) (24.3) (24.3) (26.1)
    1.1 Valuables 40,146 24,484 13,957 14,187 14,724 15,519
    1.2 Gross domestic capital formation(GDCF) 849,099 701,146 605,057 540,281 496,866 496,695
    excluding valuables (commodity flow method) (27.2) (25.4) (24.7) (23.7) (23.6) (25.4)
    2 Finances for GDCF 939,555 752,132 620,508 523,737 509,026 509,289
    2.1 Gross domestic savings 907,416 797,512 648,994 537,966 496,272 487,301
    2.2 Net capital inflow 32,139 -45,380 -28,486 -14,229 12,754 21,988
    3 Errors and omissions (2-1) 50,310 26,502 1,494 -30,731 -2,564 -2,925
    (1.6) (1.0) (0.1) (-1.3) (-0.1) (-1.5)
    4 Finances for GDCF (adjusted) 979,701 776,616 634,465 537,924 523,750 524,808
    4.1 Gross domestic savings 907,416 797,512 648,994 537,966 496,272 487,301
    4.2 Valuables 40,146 24,484 13,957 14,187 14,724 15,519
    4.3 Net capital inflow 32,139 -45,380 -28,486 -14,229 12,754 21,988
    5 Adjusted errors and omissions (4-1) 90,456 50,986 15,451 -16,544 12,160 12,594
    (2.9) (1.8) (0.6) (-0.7) (0.6) (0.6)
    Memo item
    (4.1+4.2) as per cent of GDP (30.4) (29.8) (27.1) (24.2) (24.2) (25.7)
    Note: Figures in brackets are percentages to GDP at current market prices.
    Source: CSO (2005, 2006).
    Economic and Political Weekly June 10, 2006 2325

    done as shown in Table 11, the domestic saving rate goes up by 1.3 percentage points (or by Rs 40,146 crore) to 30.4 per cent and the consequential “errors and omissions” between the finances for GCF and the unadjusted GCF almost doubles from Rs 50,310 crore (1.6 per cent of GDP) to Rs 90,456 crore

    (2.9 per cent). Overall, it thus appears improper to include “acquisitions of valuables” as part of GCF both conceptually and from the view point of its operational significance (see also EPW editorial of February 11, 2006).

    A Long-standing Methodological ErrorA Long-standing Methodological ErrorA Long-standing Methodological ErrorA Long-standing Methodological ErrorA Long-standing Methodological Error

    Finally, an issue that has remained unattended to by the CSO, despite the recommendations of the Raj working group [RBI 1982] as well as the Chelliah working group [DoS 1996], relates to the inadvisability of adjusting the GCF estimates arrived at by the commodity-flow method with its difference from gross saving, including capital inflow from abroad. The Raj working group had opined that it was improper to conceive the measure of saving plus foreign capital inflow as the controlling total and then adjust the independently-estimated gross capital formation numbers by the difference and treat this difference as “errors and omissions” [EPWRF 2004]. It had proposed that the difference be retained as “statistical discrepancy” without making any adjustment to any of the independent estimates. The fear expressed by the Raj working group that as the adjustment is made only at the aggregate level and not at the sectoral level, “the present method has created a serious analytical problem in the sense that no consistent sectoral shares in gross or net capital formation could be worked out” (p 128), remains unaddressed, despite the claim to the contrary (see Table 9). On examining the status of Raj working group recommendations, the Chelliah expert group’s status report [DoS 1996] said that the recommendation had been implemented and the term “errors and omissions” substituted by the term “difference”. But, in reality, no such change has been effected and the CSO’s National Accounts Statistics year after year continues to present, as in the past, the figures of “errors and omissions” representing the difference between total savings (domestic and foreign) and the estimated figures of gross capital formation by the commodity flow method (see item 3 in Table 11). Notwithstanding such a strong case for avoiding it, the CSO has retained the erroneous method of deriving“errors and omissions” and presenting an adjusted capital formation series as the final investment estimates even in the revised series.

    Final GCF RateFinal GCF RateFinal GCF RateFinal GCF RateFinal GCF Rate
    sans Valuables and ‘Errorssans Valuables and ‘Errorssans Valuables and ‘Errorssans Valuables and ‘Errorssans Valuables and ‘Errors
    and Omissions’and Omissions’and Omissions’and Omissions’and Omissions’

    If the adjustment for “errors and omissions” are thus not made and if the value of valuables are excluded from the GCF estimates on the grounds that they do not constitute any productive capital asset creation, the figure of GCF rate for 2004-05, which has been placed by the CSO at a level as high as 30.1 per cent, should be appropriately placed at 27.2 per cent (Table 12). Thus, the image of the economy gets sobered down to a statistically more realistic investment level of 25-27 per cent, instead of 27-30 per cent as per the data for the past two years.

    A legitimate question that may be posed is whether it is realistic to place gross capital formation rate at a level (27.2 per cent) that is lower than the domestic saving rate (29.1 per cent), particularly for a year, as in 2004-05, when there is a current account deficit in the external sector. The obvious answer is in the affirmative because there can be leakages of domestic saving into unproductive assets. Expenditures on valuables are one such distinct example. There are possibilities of overestimation of household financial savings [Rakshit 1982, 1983; EPWRF 1995, 1996]. Some others may be subsumed in the mass of data that cannot be captured overtly.

    m

    Email: epwrf@vsnl.com

    NotesNotesNotesNotesNotes

    [Tabulations for this note, prepared by R Krishna Swamy, are gratefully acknowledged; so also the efforts of K Srinivasan in inputting the note].

    1 The Labour Bureau (government of India) has just revised, after 25 years, the base year for the Consumer Price Index for Industrial Workers from 1980 to 2001. It is also reported that efforts are being made to update the other price indices – WPI (1993-94=100), CPI-UME (1984-85=100), and CPI-AL (1986-87=100) – with new base levels.

    2 Apart from the series of research papers pointing to this deterioration as quoted in Shetty (2006), the Report of the National Statistical Commission cited above has been an eye-opener in this respect.

    ReferencesReferencesReferencesReferencesReferences

    Chaudhuri, Saumitra (2005): ‘A Note on Investment and Savings’, Money and Finance, ICRA Bulletin, January-June. CSO (2005): National Accounts Statistics, Central Statistical Organisation, May.

    – (2006): New Series of National Accounts Statistics (Base Year 1999-2000), Central Statistical Organisation, February.

    DoS (1996): ‘Saving and Capital Formation in India: 1950-51 to 1994-95’, Report of the Expert Group on Saving and Capital Formation, chairman Raja J Chelliah, December.

  • (1999): ‘Comments on Press Reports Relating to Advance Estimates of National Income’, Department of Statistics, Government of India, February 12.
  • EPWRF (1995): ‘Economic Reform and Rate of Saving’, Economic and Political Weekly, May 6-13.
  • (1996): ‘Economic Reform and Rate of Saving’, Special Number, Economic and Political Weekly, September.
  • (2004): National Accounts Statistics of India 1950-51 to 2002-03: Linked Series with 1993-94 as the Base Year, Fifth Edition, EPW Research Foundation, December 20.
  • NSC (2001): Report of the National Statistical Commission, chairman C Rangarajan, Volumes I and II, August. Rakshit, Mihir (1982): ‘Income, Saving and Capital Formation in India’, Economic and Political Weekly, Annual Number, April.

  • (1983): ‘On Assessment and Interpretation of Saving-Investment Estimates in India’, Economic and Political Weekly, Annual Number, May.
  • Rangarajan, C (2001): ‘National Statistical Commission: An Overview of the Recommendations’, Economic and Political Weekly, October 20.
  • (2002): ‘Reform of the Indian Statistical System’, inaugural address, January 14, see Data, Models and Policies, Proceedings of the 38th Annual Conference of the Indian Econometric Society.
  • RBI (1982): Capital Formation and Saving in India 1950-51 to 1979-80: Report of the Working Group on Savings, chairman K N Raj, February, Reserve Bank of India.

    Shetty, S L (2005): ‘Saving and Investment Estimates: Time to Take a Fresh Look’, Economic and Political Weekly, February 12.

    – (2006): Paper presented at seminar on ‘Macroeconomic Policy, Rural Institutions and Agricultural Development in India’(A National Conference in Honour of A Vaidyanathan), held at the Institute for Social and Economic Change (ISEC), Bangalore, April 9-10.

    Dear reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here

    Comments

    (-) Hide

    EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

    Back to Top