The domestic savings rate or the ratio of gross savings to GDP is estimated by the CSO to have touched a record level of 29.1 per cent in 2004-05. This implies an increase of 5.5 percentage points since 2001-02, before which the rate had remained stagnant and even declined since the mid-1990s. The recent revision of the weighting system and revision of the base year from 1993-94 to 1999-00 has made a difference to the savings rate estimates for individual years for which estimates for both series are available. But the trend remains broadly the same. The initial rise in the savings rate was captured by the old series. And provisional estimates for 2004-05 from the new series suggest that the trend has continued. Overall, the rise in the savings rate has coincided with an increase in the rate of growth of GDP over the last three years, suggesting that the economy is transiting to a sustainable, higher growth trajectory.

The question naturally arises as to the factors that are responsible for the rise in savings. In particular, given the necessarily complex way in which savings and investment are estimated in an economy like India with a large household sector, including unincorporated enterprises, the question arises as to whether the sudden increase is more statistical than actual. In fact, since the current savings surge coincides with a massive increase in the inflow of portfolio FII investment in India’s stock and debt markets, it had been argued (see S.L. Shetty in Economic and Political Weekly, February 12, 2005, for example) that some of this capital may have been lodged with the banks and got erroneously recorded as constituting household savings invested in financial assets.

A priori this is indeed a possibility. Savings are estimated by dividing the universe of savers into the public sector, the private corporate sector and the household sector (including unincorporated enterprises). Public sector saving is computed from budgetary data, and captures the excess of government expenditure at the central and state levels over revenue; and private corporate sector savings in the form of retained earnings are obtained from company balance sheet data.

Household savings are estimated in two parts. First, household financial savings, estimated using data on the assets and liabilities of the financial sector, adjusted for its outstanding positions with the public and private corporate sector. Second, household savings in physical assets which is the excess of aggregate capital formation estimated by the product flow method (or the availability of items like machinery, equipment and construction material that enter into capital formation) over the estimated capital formation in the public and private corporate sectors based on budgetary data and company balance sheets.

It should be clear from the above that household saving is derived as a residual category in the case of both household financial saving and household savings in physical assets (or household capital formation). It should be clear that if funds transferred by FIIs for investment purposes are parked in bank deposits or other financial instruments before being invested in corporate equity and bonds, there is a possibility that they could be attributed to the household sector. Similarly if aggregate capital formation derived through the product flow method is overestimated, then assuming that estimates of capital formation in the public and private sectors are correct, there would be an overestimation of ''physical savings'' in the household sector.

These issues are of relevance because the gross savings rate in the household sector had indeed risen between 2000-01 and 2003-04 (provisional) (Chart 2). However, there are two features of movements in the sectoral savings rate that need to be noted. First, while the increase in the economy-wide gross savings rate between these two years amounted to 5.4 percentage points, that in the case of the household sector totaled 2.3 percentage points. That is, movements in household sector savings account for much less than half of the increase in the aggregate savings rate between these two years. Second, if we consider the quick estimates for 2004-05, we find that the savings rate in the household sector fell by 1.5 percentage points, whereas the aggregate savings rate rose by a further 0.2 of a percentage point. Overall, between 2001-02 and 2004-05 while the aggregate savings rate rose by 5.5 percentage points the household savings rate was stagnant.

Another point to note is that the composition of the household savings rate (Chart 3) has not changed significantly in favour of financial instruments. In fact the share of financial savings has been more or less stagnant, falling in individual years like 2002-03 and 2004-05. This weakens the argument that the rise of in the aggregate savings rate could have been the result of a mis-categorisation of float FII funds as household savings.

In fact, a close examination of sectoral savings trends (Chart 2) suggests that the sector responsible for the rise in the savings rate is the public sector, which has seen a sharp transformation of its dissaving into saving in the years since 2001-02 and 2004-05. During this period, the public sector saving rate rose from a negative 2.0 per cent to a positive 2.2 per cent, contributing a remarkable 4.2 percentage points increase to the savings rate. Together with a small contribution from the private corporate sector, this been primarily responsible for the statistically recorded savings surge. This comes through from Chart 4 which shows the major role played by reduced public sector dissaving or increased public saving to the increase in the aggregate savings rate between 2002-03 and 2004-05.

The public sector itself consists of three sub-sectors: government administration, departmental enterprises and non-departmental enterprises. An interesting question to ask is which of these contributed to reduced dissaving or the increase in saving. There is a common perception that some of the non-departmental public enterprises, in the wake of liberalization of public sector pricing practices, have been able to accumulate large surpluses that are accumulated in the form of reserves. While this is indeed true, it does not seem to be the case that this substantially explains the improvement in public saving. Rather, as Chart 5 shows, while the savings of non-departmental enterprises have contributed to improved savings, especially in 2004-05, it is the reduced dissaving in government administration that substantially accounts for the estimated improvement in the savings rate. This sub-sector accounted for approximately 60, 90 and 80 per cent respectively of the contribution of the public sector to increases in the savings rate in 2002-03, 2003-04 and 2004-05.

The final question to ask then relates to the factors that contributed to the improvement in public savings. As Chart 6 shows both stagnation and decline in government expenditures and a significant improvement in government revenues seem to have delivered the reduction in net dissaving of the administrative departments of government.

This is indeed a puzzle inasmuch as the liberalization years have seen a sharp reduction in customs tariffs and that while the deficit at the central level has come down relative to GDP, no sharp reduction has been recorded. There could thus be three factors that could have played a role here. First, a growing curtailment of public employment, which is by no means positive given the massive deficit in public services, especially in rural areas.. Second, improved customs duty collections despite reduced tariffs, because of an increase in the quantum of imports and the sharp increases in oil prices. Third, the contribution of special dividends from to-be-privatised entities and receipts from disinvestment and privatization to government ''revenues''.

However, as we noted, the adjustment in the deficit of government administration must have in significant measure occurred at the state level. And states could not have benefited much from customs tariffs and privatization. In all likelihood then, the cutback in the size of government administration must have occurred in large measure at the state level, which cannot but have adverse implications for development and the provision of public services.

Needless to say, all this is predicated on the numbers being reliable. Two features of the savings and investment estimates released in recent times give cause for concern on this front. First, the rather sharp revisions made to the estimates as we move from the quick to the provisional, revised and final estimates. Second, the huge statistical discrepancy between savings and investment recorded in the ''errors and omissions'' category. To the extent that the capital formation estimate falls short of the sum total of public, private corporate, household and foreign savings, the discrepancy is recorded under ''errors and omissions''. This practice has been controversial for some time and was to be given up and substituted with an honest admission of a ''statistical discrepancy'' between savings and investment estimates. But neither has this been done nor has the discrepancy come down significantly. In the circumstance we may be debating an issue which itself is only a statistical mirage.