Finding India's Nemo - a low tax-to-GDP ratio

In "Earn poor, spend rich", I had documented how India's tax/GDP ratio was not low in an international context.

Not only was it not low, but it was almost exactly as could be predicted, given India's income level. The response to the article has been -- it can't be so. And why not? Because several experts have claimed that India's tax/GDP ratio is low, and so I must be doing something wrong.

The more charitable critics have correctly stated that the calculation of tax/GDP ratio is plagued with three major problems. First, that the definition of "tax" needs to be explicit.

For example, does it include or exclude state and local taxes; does it include or exclude social/security taxes. Second, what is the source of these data -- the World Bank, or the International Monetary Fund, or the Organisation for Economic Cooperation and Development? The three organisations' tally of taxes should not differ because they all use the same source, that is, ministry of finance data originating in each country.

But they do differ -- mostly because of differing distributions and "patches" made by the different organisations. The third problem is the selection of countries. Clearly, sample selection can affect the results and bias it in the direction "desired" by the researcher or policy advocate.

Because of these formidable data problems, the critics contend, we cannot arrive at any accurate judgement about whether the tax/GDP ratio is indeed lower than "expected" or not. So we can all go along with our initial priors and make policy recommendations accordingly.

This confident assertion about the tax/GDP ratio in India being too low is in many ways the IMF's answer to the World Bank assertion on poverty in India. The latter concludes, irrespective of all evidence to the contrary, that dollar-a-day poverty in India is upwards of 35 per cent of the population when most data, and even the most pessimistic "adjustments" place poverty in India less than 27 per cent (using the same poverty line, of course).

But the influence of the World Bank is such that even Indian governments believe that poverty in India has not been much reduced. It is another story that this is a marriage of convenience because the present United Progressive Alliance government desperately (ideologically?) wants to believe that it got voted in because the poor lost out under the National Democratic Alliance regime.

It is completely another story that the 1990s' poverty decline data refers to 1993-94 to 1999-2000, that is, years for which the Congress and its allies were responsible about what happened to the economy and the poor!

Analogously, perhaps because of the radical chicness within the organisation (the spillover effect of the virus emanating from across 19th street in Washington!), the ostensibly objective IMF also believes that Indian tax/GDP ratio is too low.

So India taxing too "little" has the IMF and the Left parties in India both agreeing. Therefore, the conclusion must be right -- and my results, which showed that this was not the case are interesting and provocative but ultimately not believable. As Indira Rajaraman, noted fiscal expert opined -- they must be the basis of a hunch!

What about the hunches of experts reaching the opposite conclusion? Here is a sampling. India's parliament: "Expressing 'grave concern' over the low tax:GDP ratio in India compared to that in many developing countries, the Standing Committee on Finance said the distortion was mainly due to the fact that large number of prospective tax-payers are yet to be trapped and brought under the tax net".

A leading advocate of more expenditures and more taxes, National Advisory Council member, Jean Dreze: "India's tax-GDP is very low in international perspective: about 15 per cent (Centre plus states) compared with, say, 37 per cent in OECD countries". Two IMF authors, Yougesh Khatri and Kalpana Kochhar: "Tax and total general government revenue relative to GDP in India are low by international standards…India's revenue to GDP ratio is low, even compared with Asian countries".

Kelkar Task Force study on taxation: "India has one of the lowest levels of the tax/GDP ratio in the world. This low tax/GDP ratio has been a central feature of India's fiscal problem".

I should add that there are innumerable esteemed economists, policy makers, bureaucrats and so on, who also believe that the Indian tax/GDP ratio is too low. Indeed, it is fair to say that except for a few (significantly less than five), all believe in the universal wisdom documented above.

But how accurate is this expert wisdom? The table presents the data on the tax/GDP ratio for nine different classifications and two different definitions of taxes: government revenue from all taxes including social security contributions, and government revenues excluding SS funds.

If tax ratio without SS is examined, then the Indian ratio is lower than what it should be ("should" is the predicted value obtained from a regression of tax/GDP ratio and per capita income level, 19993 PPP$, World Bank data) for only three of the first eight classifications. If SS is included, the tax/GDP ratio is lower in six of eight. In none of the 16 cases, is the Indian tax/GDP ratio significantly different than "predicted".

The ninth classification is a special one. These voodoo or jadu countries are chosen to yield the result that India taxes too little of its GDP.  Its construction was a major research undertaking but it essentially consisted of choosing all countries poorer than India and having a higher tax rate, and choosing counties richer than India with not only a higher tax/GDP ratio, but a significantly higher such ratio (note that a higher ratio by itself is not sufficient because some increase is expected with development).

We were able to identify 16 such sure-shot countries; unfortunately for our too-low-tax experts, most of these countries are difficult to place on the world map. Some of these countries are Moldova, Namibia, Latvia, Belarus, Burundi, Swaziland and so on.

Nine of the 16 countries have population below 10 million; several of them are there because of commodity taxation, a luxury India clearly does not have. So these voodoo countries are the best jadu evidence we have to assert that India's tax/GDP ratio is too low.

While not too low, India's tax/GDP ratio should rise by 2 to 3 percentage points in the next few years (largely through increases in direct taxes, personal and corporate).

Very few of the protagonists in this debate, and not including the IMF study cited earlier, believe that excess expenditure is the major problem causing the large fiscal deficits. If only the few would become many. But it won't happen, because that is not politically correct in these trying human face expenditure times.