In my previous article (goo.gl/meM9kY), I had argued that the imposition of a long-term capital gains tax (LTCG) was not a good idea because the tax revenue obtained from its imposition was not likely to be large. In this article, I document, according to various methods, that the LTCG formulation is not only flawed in design, but also likely to yield very little tax revenue. At the same time, it manifestly increases the discretionary power of the taxman and this has its own set of governance problems.

In the run-up to the budget, I had pointed out (as had others) that the government had set up a committee headed by Arbind Modi to provide a new direct tax policy; hence, few expected that finance minister Arun Jaitley would tinker with the direct tax schedule. This was not to be, because he did tinker. That it was manifestly not worth the tinkering is documented below.
But first–the legitimate problem of stock-price manipulation and the evasion of taxes. There is a well-known promoter-operator network which manipulates the prices of penny stocks, drives the price upwards, and the operator-promoters then cash out at a higher price (a year or so later), and pay no tax. There is, thus, an element of double corruption involved. There is very little data available on the exact magnitude of this manipulation/tax avoidance, but it is something that should be handled by the tax authorities in a better manner than “punishing” 99.9% of investors because of the corruption of very few.
As I have repeatedly emphasised, the best tax policy for the taxman is one that maximises tax revenue, and not one that maximises his morality, or his employment, or his discretionary powers. With that in mind, let us explore what the imposition of the LTCG tax will do for tax revenue.
In his budget speech, FM Arun Jaitley made the following statement: “The total amount of exempted capital gains from listed shares and units is around `3,67,000 crores as per returns filed for A.Y.17-18.” Subsequently, this was repeated by finance secretary Hasmukh Adhia. This is an eye-popping tax revenue gazing figure; at a tax-rate of 10%, this is an additional `36.7 thousand crore of tax-revenue. But, is this estimate of tax-revenue likely to be correct for an average year in the future? Very unlikely in our view, and we attempt to document this conclusion according to several different assumptions.
First: The stock market gained an average amount of only 3.9% in FY17 (or assessment year 17-18), the year for which the taxman claims to have lots of exempted capital gain income. The ministry of finance (MoF) has provided detailed income data for all taxpayers for the four years FY12 through FY15 (see the website goo.gl/zq9JEg). Between FY11 and FY15, the average value of Sensex increased from 18,607 to 26,531, a gain of 43%, or a compounded annual gain of 8.8% a year, i.e., more than twice the gain recorded in FY17.
According to the MoF website, cumulative and all LTCG income (that accruing from stocks, property, gold, etc) for four years was `279 thousand crore. In FY17, with just a 2.8% increase in the stock market, it is very unlikely that an income gain of `367 thousand crore was possible. (see table).
Second: The accumulated STCG (short-term capital gains) income, for these four years, was just `125 thousand crore, yielding an annual tax revenue of `4,700 crore (at a 15% tax rate). But what percentage of stock market gains are short term (and hence what proportion is long term?). A difficult question to answer, but one extreme case is that short-term trades are only 25% of total trades. This implies that LTCG income is three times the STCG income, i.e., for the four years, around `375 thousand crore. But, all long-term capital gains (stocks, property, etc) were just `279 thousand crore or (279/125) 2.2 times STCG income—so three times is a huge over-estimate.
A reasonable liberal estimate (based on the recent FY12 –FY15 experience) is that LTCG income for stocks alone is 1.5 times STCG income. Hence, for the four years for which we have tax data, the cumulative four-year long term gains from stock market is 1.5*125 or `188 thousand crore, or a maximum of `47 thousand crore a year. And remember, this is with the Sensex gaining an average 8.8% a year, and not the paltry 2.8% increase experienced in FY17. This number is just 13% of the `367 thousand crore claimed by finance ministry officials.
Now let us take the high market growth year of FY15—reported STCG income for this year was `73 thousand crore. It is likely that in years of high stock market increases, more long-term sellers enter the market, in order to avail of a zero tax benefit, i.e., they sell, so they can pay no tax on “high” income gains. This would mean that exempted long-term stock market gains in a good stock market year would be no more than `73 thousand crore—less than a fifth of that assumed by the MoF.
There is an alternative long-term calculation. The Kelkar Direct Taxes task force of 2002 (with Arbind Modi as a senior member and myself as a pro-bono researcher) estimated that the long-term tax revenue being foregone in FY03 via the elimination of a 20% tax rate on LTCG was `1,000 crore. This implies that the estimated LTCG income in FY03 would have been `5,000 crore. The stock market increased by 8.5 times between FY03 (Sensex equal to 3,203) and FY17 (Sensex equal to 27,340). Hence, LTCG income (assuming buying and selling behaviour has stayed the same on average over the last 15 years) in FY17 would be estimated to be approximately 8.5*5000 or approximately `43 thousand crore.
No matter what the method, or the assumption, LTCG income in FY17 is unlikely to have exceeded `50 thousand crore, and most likely a lower number. This is less than one-sixth of the MoF estimate. Hence, imposition of a 10% LTCG tax rate is unlikely to yield more than `5,000 crore in tax revenue.
One final note about the tax revenues gained from both long-term and short-term capital gains. The MoF website also contains an estimate of a loss set-off for each of the four years. This set-off includes losses in business, because such losses are allowed to be set-off against LTCG. If the new LTCG tax is implemented, there will be an additional category under set-off—long-term capital losses. In any case, it is revealing that after accounting for set-offs, the MoF has gained, on average, just `3,000 crore a year!
The 2002 Kelkar report had advocated the abolishment of LTCG tax, the abolishment of dividend tax, the retention of short-term tax to 10%, and the introduction of the securities transaction tax (STT). Average annual STT revenue has been around `7,000 crore, and is forecast to be `11,000 crore in FY19. Regardless of market direction, STT makes revenue for the taxman. Isn’t it much simpler, and more tax revenues, to increase STT by 25%. This will yield `3,000 crore more per year.
What is clear is that there is no historical data which suggests that exempted LTCG amount to anything close to the `367 thousand crore figure claimed by the MoF for FY17. It would seem to be prudent for the ministry to withdraw this tax proposal and state that they will act once the Arbind Modi report becomes available. Until then, the MoF should find a method to identify the penny stock manipulators. That will be a win-win for all.
PS: There is one consistent explanation for the possibility that `367 thousand crore were actually declared in income-tax returns despite FY17 not showing much returns for the average investor. Could this be the effect of demonetisation, i.e., the stock market being used to launder black cash? Even if this is the case (like the promoter-operator nexus) it requires a very different policy response than the imposition of a tax on future long-term capital gains.