While most pundits, including those who are not journalists, are predicting a slew of tax reforms in the coming Budget, a point not explicitly stated is that unless these "reforms" happen, the country's tax collections are going to be way below what is needed to meet our expenditure needs.

That is, the existing tax structures are simply not delivering the goods, certainly not enough of them. A logical corollary is if the system isn't delivering, it needs a dramatic overhaul.

One good example of how old ideas can't just be reworked a bit is what's happening on excise duties.

While many cite the slowdown in excise duty collections even at a time when manufacturing is booming as a sign of poor tax compliance, as the Kelkar report on the implementation of the Fiscal Responsibility and Budget Management Act points out, excise revenues from the non-oil industrial sector have been falling steadily for many years, from 16.4 per cent of industrial GDP in 1980-81 to 7.7 per cent in 2003-04.

Indeed, the report says the buoyancy of excise revenues (less those on the oil sector) is a mere 0.5 -- that is, if industrial output rises by 10 per cent, excise collections will rise by around half this, something quite akin to the situation today presumably.

Since excise collections as a proportion of total taxes are still around 35 per cent, as compared to a 24 per cent share of industry in GDP, it is likely such taxes will continue to fall in the years to come.

In which case, pinning hopes on being able to squeeze too much more out of this sector may not be too great an idea.

Indeed, it is worth paying attention to other fundamental changes that have taken place in the tax structure over the years.

Next year (the year for which the Budget will be presented on February 28), income tax collections are likely to exceed those from customs duties for the first time in the country's history.

In the year after that, Kelkar projects corporate taxes will become the single-largest tax, beating traditional winner excise duties.

In Kelkar's business-as-usual scenario, corporate taxes will account for 3.1 per cent of GDP in 2006-07 as compared to 3.0 per cent for excise -- in the reforms scenario these are projected to be 3.4 per cent and 3.3 per cent, respectively.

While the use of the Tax Information Network is an important part of the exercise to be able to get more people into the net, and to ensure they file their incomes correctly, there is an even more basic issue here, first pointed out by Surjit Bhalla and later sanctified by the Kelkar task force.

It appears the people evading (not avoiding, which is legal) taxes are not so much the fat cats as it is ordinary middle class folk. While over 43 per cent of those earning above Rs 10 lakh (Rs 1 million) pay their taxes this compares with a mere 12�13 per cent of those earning Rs 200,00-500,000 a year.

Indeed, it is for this reason that Bhalla argues for a flat tax of 15 per cent on all incomes above Rs 100,000 -- since this is well below the rate applicable to such incomes today, compliance will clearly increase and therefore ensure there is no loss of revenue.

But what of equity, is the obvious question, and not just from the left parties.

Well, for one, as public finance expert M Govinda Rao argues, the current global view is that equity can no longer be assured through tax policies but is best done through generating jobs, and it is the rich who have the surplus funds to create employment.

But more than that, a look at some numbers should make it obvious why there's no point only chasing the rich. According to Bhalla, there are 230,000 individuals in the country who have an income of over Rs 10 lakh (Rs 1 million) per annum.

Since they should have paid a tax of Rs 18,900 crore (Rs 189 billion) in 2003-04 versus the Rs 8,100 crore (Rs 81 billion) they paid, there is Rs 10,800 crore (Rs 108 billion) to be made by getting all these people into the net.

There are, however, 12.8 million people who have an income of between Rs 200,000 and Rs 500,000 but just 1.6 million pay their taxes. Thus, of a potential Rs 70,000 crore (Rs 700 billion) of tax hardly Rs 9,000 crore (Rs 90 billion) is collected.

Since the spending levels of the middle classes aren't high enough to get captured in the TIN, lowering tax rates is the only solution, and having one rate also reduces the incentive to show lower incomes in order to pay lower tax rates.

Another area where flat taxes make sense is customs duties. While graded duty structures like 5 per cent on raw materials and 10 per cent on finished products look nice and fair, in actual fact they greatly distort markets.

Assume the import duty on crude is nil and that on petrol/diesel is 10 per cent. Given the conversion norms, take a global refinery, which buys crude at Rs 100 and converts this into petrol/diesel worth Rs 120.

Now, if an Indian refinery is to import the crude, this will also cost it Rs 100 since there is no import duty. Once again, it will convert this into petrol/diesel at Rs 120, and create a value addition of Rs 20.

Since imported products will cost 1.1 times Rs 120, or Rs 132, domestic producers can sell petrol/diesel at up to Rs 132 -- that's a protection of Rs 12 for a value addition of only Rs 20, or an effective protection of 60 per cent!

But surely this is good because it is helping build up capital-intensive refineries in the country. True, but also think of the benefits the entire economy would get (even if the refinery didn't come up) in terms of reduced transport costs if it was able to buy petrol/diesel at Rs 120 instead of for Rs 132.