Sunil Jain: Numbers don't add up

RATIONAL EXPECTATIONS

While one section of the government battles the political fight over whether SEZs should be allowed to take over large tracts of land, at prices often a fraction of the true value, another is caught up in the maths of it "" how do the tax losses compare with the benefits from SEZs. Most thought the matter was pretty much over when, despite the finance ministry's calculations of tax losses running into Rs 175,000 crore over a five-year period, or around 5 per cent of annual tax revenues, the Prime Minister decided to press ahead. But obviously the battle is not over since, for some reason, the finance ministry commissioned think-tank ICRIER to do a study on it. ICRIER's study says the benefits far outweigh the tax losses "" if nothing else, this should settle the debate on whether consultants tailor their studies to suit their clients!

The ICRIER study makes a set of very elaborate assumptions. So, Rs 100,000 crore is to be invested each year (the study does the numbers for three years); 50 per cent in Year 1, 30 per cent in Year 2 and the rest in Year 3; of this 50 per cent in Year 1, 60 per cent will be used to create infrastructure in Year 1; of the 30 per cent to be spent in Year 2, 30 per cent will be on infrastructure; 44 per cent of the productive investment (the part that's spent on plant and machinery, after removing the spending on creating infrastructure and large townships) will be for IT exports in Year 1 ... you get the drift, the assumptions are humungous.

At the end of it, what do you get? The study says the tax loss ranges from Rs 19,429 crore to Rs 24,261 crore over three years. Based on current tax collections, that's a tax loss of around 1.3 per cent for investments that are around 4-5 per cent of total investments in the country. If the tax loss figures are correct (they certainly don't match those put out by the finance ministry), the bargain may not be a bad one provided (that's in capitals) none of this investment would have come into the country had it not been for the SEZs. The ICRIER study tries to make this argument and talks of how SEZs attract FDI, but it's not too convincing "" most SEZ applications so far are from local firms and the huge real estate segment of the SEZs makes it apparent exports are a smaller part of the story.

Having argued the investment in SEZs is incremental in nature, ICRIER justifies the tax loss by stating that the country earns an extra Rs 76,482 crore to Rs 109,875 crore by the extra exports over three years, or a figure that's around 15 per cent of the current level of exports. That's a huge reason in favour of SEZs.

But this is where some pretty serious problems crop up. The report assumes three Incremental Capital Output Ratios (ICOR) for its calculations. The first one is 0.58, a figure got from an article in The Economic Times in 2006 which had a survey on 150 companies; another ICOR used is 3.5, which is the overall ICOR for the economy; and given that exporting units are generally believed to be more efficient than others, the study uses a lower ICOR of 2.5 as a third option.

Take a look at the 10th Five Year Plan document and you realise most of these numbers are incorrect. Sure, the economy-wide ICOR is 3.58, but that's because agriculture's ICOR is 1.99, trade's is 0.91, construction's is 0.99 and financial services' is 1.56. But the SEZ won't have agriculture, trade and other such services "" it'll have manufacturing (for which the ICOR is 7.77) and IT (the ICOR for communication, the backbone of IT, is 8.33). If the SEZ has some serious infrastructure, as reports indicate the larger ones will have, the SEZ's ICOR will go up even more "" the ICOR for railways is 14.66. Plug in these numbers and the extra exports, and the justification for SEZs come down dramatically.

For purposes of calculating the tax loss, ICRIER also assumes the profit margins on non-IT products are 6 per cent (more than half the SEZ production is assumed to be non-IT in ICRIER's calculations). Taking even the most favourable ICOR of 0.58, this means if Rs 100 is invested, the output will be Rs 172, on which the profits will be Rs 10.34. A 10.34 per cent return on investment is not something that will satisfy either the lenders or the equity investors in the project. Clearly, the profit levels assumed are too low. But if you increase the profit levels, the tax collections would also go up "" and since these are not collected if you're in an SEZ, the tax losses rise significantly.

Apart from the shaky assumptions, the other issue is of reconciling the ICRIER estimates with those of the finance ministry, whose tax losses are much higher, and on much lower investment figures than those in the ICRIER study. Presumably that's what Finance Minister P Chidambaram will ask both ICRIER and his officials to explain when he gives ICRIER a date for making a presentation on its study.

sunil.jain@bsmail.in