would remain in doubt.

O

The “tax expenditure statement” (TES) or the statement of “revenue foregone”, as indicated by the finance minister in his budget speech, is yet another attempt on his part to add meaning to the budget. Last year, the finance minister, for the first time, introduced the “gender budget”. This year, in addition to the TES presented in Annex 12 of the Receipts Budget, he has also promised to present shortly an “outcome budget” that is expected to inform the Parliament and the public of the actual outcomes of the expenditure incurred by the government. All these efforts fit well into the overall scheme of things envisaged under the FRBM regime.

It is well known that the budgets reflect implicit tax preferences towards different groups of taxpayers. Such tax preferences take different forms. In addition to outright exemptions provided under the statute, they could be in the form of special rates of tax, deductions in the taxable income, rebates, tax deferment or tax credits. Whatever be the form, these are all hidden subsidies of one kind or another. As long as they are shown implicitly as revenue foregone on the receipts side, their cost-benefit implication for the economy remains non-transparent. If their real cost to the economy could be quantified and stated as lost revenue, they could be shown as an item of expenditure and their impact stated more explicitly. Theoretically, therefore, the TES is indeed a positive step forward in the overall effort being made by the government towards more and more transparency.

The TES has another additional advantage. Tax exemptions per se involve a good deal of administrative processing that includes certification, verification and final acceptance of a given exemption. Wherever there is an administrative process, there is a subjective element that creates scope for harassment and corruption. Good governance in tax administration therefore requires total elimination of all exemptions. To this extent, the TES should hopefully facilitate a wider discussion on this important aspect in Parliament.

The TES as presented in this year’s budget broadly covers corporate and personal income taxes, direct taxes relevant to the cooperatives and excise and customs duties. As far as corporate taxes are concerned, there are five important items left out in the TES calculations. These are the tax deductions allowed for donations to charitable institutions, donations to institutions engaged in scientific, social science or statistical research, donations to political parties, expenditure on schemes for promoting social or economic welfare of the public and the concessional tax regime for the shipping industry. Sections 80G, 35(1), 80GGB and other similar provisions of the Income Tax Act are relevant to these exemptions.

In the case of excise duty, the TES calculations exclude the revenue foregone as a result of ad hoc exemptions issued under Section 5A(2) of the Central Excise Act, 1944. Similarly, in the case of customs duty, the TES estimates do not cover the revenue loss on account of exemptions granted under Section 25(2) of the Customs Act, 1962. Despite the exclusions referred to above, one could readily agree that in its present form, the coverage under the TES in its very first year is fairly comprehensive.

The usefulness of the TES would really depend on how up to date its estimates are. The kind of estimation that is attempted in the TES requires a fairly comprehensive and accurate database. Unfortunately, despite the efforts that have been made over the years to computerise the information systems in the different wings of tax administration, there are still many large gaps in the information that is available, both in terms of the extent of coverage and the time lag in compiling the information at one place. The TES itself is a set of estimates for the financial year 2004-05. Thus, it is a somewhat outdated statement. Even this is not based entirely on the tax information that is relevant to 2004-05. For example, the “revenue foregone” estimates for corporate tax were based on the information available in respect of 1,689 companies (out of a list of 9,000 companies initially identified) for

Economic and Political Weekly April 8, 2006

the financial year 2003-04. The revenue foregone, as estimated on this basis, was extrapolated to cover the entire population of companies in 2004-05.

Similarly, the corresponding estimates for cooperatives were based on taxrelated information available in respect of scheduled urban cooperative banks (UCBs) for the financial years 1999-2000 and 2000-01. Estimates on the basis of this information were extrapolated for the financial year 2004-05. Even though there were other types of cooperative banks, namely, state and central cooperative banks, the same basis had been adopted to arrive at the revenue foregone in their case also.

In the case of individuals paying income tax, one would require the typical taxpayer profile for 2004-05, for estimating the national tax expenditure for that year. Since such an up to date tax profile could not be readily constructed, the profile worked out on the basis of data for the four previous assessment years, namely, 1998-99 to 2001-02 had been taken as the starting point and adjusted suitably, by factoring in the tax changes that had come into effect later, in arriving at the taxpayer profile for 2004-05.

In the case of excise duty, the information available was somewhat more realistic. During 2004-05, the excise department’s own database covered excise duty returns that captured the details for 93 per cent of the total gross revenue collections for that year. The duty-foregone in respect of this group of taxpayers had been extrapolated to cover all those that paid the duty.

Finally, in the case of customs duty, the revenue-foregone estimates were based on the department’s own electronic data system known as ICES. The ICES information was available only to the extent of the bills of entry filed at 20 Electronic Data Interchange (EDI) locations. In other words, such a database could not capture the information available at these same 20 locations, to the extent that the bills of entry were not filed with the EDIs. Also, the information available at other locations that were manually operated was not readily available for the purpose of the present study.

The structure of taxation changes from year to year. Therefore, it would not be realistic to arrive at the revenue-foregone estimate for one year by linearly extrapolating the estimate for another year. Moreover, unless a statistically representative and significant sample is chosen, estimates such as these could turn out to be misleading.

The revenue foregone estimates crucially depend on the manner in which the “benchmark” taxation levels are established. The difference between the benchmark and the actual tax rate, when applied to the gross income for a given group of taxpayers represents the national revenue that is foregone in the case of that group.

Before discussing the benchmarks chosen in this year’s TES, it should be pointed out that the absolute levels of taxation vary from year to year, depending on the political preferences of the incumbent governments, as also the pressures they face to raise resources for their schemes. In other words, the structure of taxation in any given year is one that has already captured the tax preferences of the incumbent government. Therefore, the so-called “benchmark” tax for any given group of taxpayers in a particular year is not a static one with reference to which one could readily compare the effective tax rate, in estimating the revenue foregone. As a result, it would not be strictly meaningful to compare the estimates of revenue foregone arrived at on this basis for a given year, with corresponding estimates for other years.

The current year’s TES refers to this issue and glosses over it by saying “this issue has been addressed by attempting to estimate the revenue loss only in respect of major items of tax preferences over which there is no ambiguity”. This is certainly not a satisfactory way of addressing the problem of defining the benchmark tax rate!

There is another area of concern as far as the methodology is concerned. In taxation, the relative levels are as important as their absolute levels, to the extent they affect revenue buoyancy. Also, taxes have both short-term and long-term effects. Often, the long-term effects could prove to be more important. This is particularly relevant in the case of tax preferences, as it is the longterm impact of tax preferences that the government is concerned about. The implications of both these have been ignored in current year’s TES and the credibility of the statement suffers to that extent.

A reference has already been made to the need to base the estimates on up to date tax-related information, rather than outdated databases. There are a few minor methodological issues that need clarification. For example, in the case of corporation tax, the benchmark level is assumed at 36.59 per cent, even though the statutory rate had been indicated at 35.875 per cent.

Economic and Political Weekly April 8, 2006 for corporation tax, 24 per cent for personal income tax, 99 per cent for customs duty and 31 per cent for excise duty.

It would also be of interest to compare these revenue loss estimates with actual arrears of tax revenue as they stood at the end of 2004-05 (Annex 10 of the Receipts Budget). Taking first the total picture into account, the arrears at that point of time were Rs 1,11,108 crore, compared to an estimated revenue loss of Rs 1,58,661 crore for 2004-05. According to the information disclosed, about 59 per cent of the arrears were under disputes and perhaps subject to some kind of litigation. Even then, the revenue from the remaining 41 per cent that were not under any dispute could be realised straightaway! In one of his earlier budget speeches, the finance minister had assured Parliament that he would assign priority to collection of tax arrears, but the progress made in that direction seems to be somewhat tardy.

Since the government is genuinely interested in ensuring transparency of the budget, one would have expected the finance minister to have disclosed more details about the tax arrears in Annex 7. For example, the age-wise break up of the arrears under each head could have been placed before the Parliament, as these figures are readily available with the concerned departments. Similarly, the finance minister could furnish to Parliament the list of the top 100 defaulters of corporation tax, in the case of which the undisputed arrears stood at a staggering level of Rs 9,938 crore!

As a first-time initiative, the TES deserves commendation. However, a great deal of effort is needed to ensure that this statement adds value to the discussion on the budget within the Parliament and outside. The TES estimates can have some practical value, provided the concerned departments of revenue modernise their respective databases at the earliest. Unless a more rational definition is found for the “benchmark” tax rates, the credibility and the usefulness of TES would remain dubious. The accuracy of the revenue loss estimates would also remain questionable unless the estimation takes into account the interactive aspect of taxes on different items, as also the long-term effect of tax on the buoyancy of revenue.

EPW

Economic and Political Weekly April 8, 2006