A subsidy is anything that alters relative prices. Since it is a subsidy, the price charged is below what a market-determined price should be. Who pays for these subsidies? The government does and this doesn’t mean only the Union government. State governments also have substantial subsidies and so do local governments, although the extent of latter is difficult to quantify. Governments collect tax revenue, though the government also has some non-tax channels for collecting revenue. Estimates that float around on the tax/GDP ratio are typically for the Union government. If one also includes state-level taxes (local body taxes are very difficult to estimate), the tax/GDP ratio goes up to around 17%. These taxes are collected so that the government (at various levels) can deliver law and order, security (defence), physical infrastructure (roads, water, electricity), social infrastructure (schools, health centres) and other services.

A basic principle of economics is that of opportunity costs, though this is a principle that is often forgotten. If resources are spent on something, they won’t be available on something else. Therefore, if the government spends its tax revenue on subsidies, money will not be available for performing these other desirable activities. That’s the first problem with a subsidy. To compound the problem for citizens, the effect of a change in subsidies is immediately observable; the opportunity cost of lost resources isn’t that easily visible.

Conceptually, all subsidies are flawed. They distort resource allocation by affecting relative prices. If water is priced cheap, there will be a tendency to overuse it, whether for irrigation or for something else. It is far superior to have income transfers. Why should one decide to provide cheap food to a poor person, with ‘food’ being defined as foodgrains? His/her requirement is perhaps salt, or something else. An unconditional income transfer permits the poor person to spend on whatever is desired. It is a complete red herring to argue that subsidies are meant to push ‘public goods’.

Going back to the days of Paul Samuelson, economists have an agreed upon the definition of a public good. Such a public good has two characteristics. First, my consumption of the good doesn’t reduce your consumption of the good. Second, there is no way to exclude anyone from consuming that good. Fresh air and national security might be examples of public goods. But if these are ignored, there is nothing that we subsidize that is a public good; they are all private goods. The expression ‘public good’ is bandied around unnecessarily. There is yet another problem with subsidies; they are discretionary and complicate public policy and tax policy. A case can be made for subsidizing everything, especially since we forget about opportunity costs. Once the thin end of discretion comes in, there is no putting stop to it.

 

Let us be clear, subsidies aren’t only to consumers. They can be to producers too, including producers who produce inputs. When we are making a case against subsidies, that’s also a case against producer subsidies. Subsidies to consumers, which is what the discussion is most often about, is only part of the problem. For instance, the budget papers now have a revenue foregone statement, an account of how much tax revenue is lost because of tax exemptions. Depending on the year, that’s between 5 and 5.5% of GDP. Conceptually, this is a subsidy too.

Not all subsidies come out of budgets, Union or state. If they come directly out of the budget, they are explicit subsidies and easier to compute. However, there are hidden or implicit subsidies too. There is a NIPFP paper, commissioned by the Finance Ministry, which goes back to 2004-05. I am not aware of it having been updated. But since I am going to give a percentage figure, it may not have changed much. According to this, all subsidies, explicit and implicit, amounted to 14% of GDP. 17% collected as tax revenue. 14% spent on subsidies. Where is the money for the government to do what it should be doing?

Subsidies are ostensibly meant to serve a redistributive purpose, transfer resources from the relatively rich to the relatively poor. There are any number of studies, especially for PDS, to document the strong pro-urban bias of the current subsidy regime. Subsidies are pro-urban and pro-rich. They benefit the relatively rich in the urban segments, who masquerade as ‘aam aadmi’ or ‘aam aurat’ in TV interviews when questions are raised about eliminating subsidies. NSS data show that out of the 600,000 villages in India, around 125,000 have no physical or social infrastructure worth the name, not to speak of the absence of any law and order machinery. Why are we surprised? We should know what the government’s money is being spent on. Contrary to what was expected, from the redistributive angle, subsidies have been regressive, not progressive.

 

There is another red herring about merit goods, an argument I don’t buy. Conceptually a merit good, though not a public good, is a good (or a service) that has strong positive externalities. Examples are given of elementary education, public health, roads and bridges. The idea is that merit goods must be subsidized. To reiterate, none of these is a public good. They are all private goods. At best, they are collective, rather than individual, private goods. Therefore, one shouldn’t keep the prices of these low. One should instead ensure that poor people are not deprived access through income transfers.

I am unhappy with the discourse because it tends to only focus on explicit consumption subsidies in the Union government’s budget – food, petroleum products and fertilizers. The desire to reduce these is driven by fiscal considerations. While fiscal considerations are important, there is a broader agenda of questioning the policy on subsidies. Subsidy targeting also runs into a problem of identifying BPL (below the poverty line). As far as I can see, there are three problems with this and we go round and round in circles because of this. First, poverty is an individual household characteristic. It has nothing to do with collective identities like ethnicity, religion or caste. If one equates it with collective categories, there is a double kind of mistake. There are errors of commission, because one includes those who aren’t BPL, but happen to be in the collective category. There are errors of omission, because one excludes those who are BPL, but don’t happen to be in the collective category.

 

Second, identifying BPL on the basis of NSS data is a non-starter. NSS large samples occur at infrequent intervals, an average of five years. There are thus time lags. In addition, NSS is a survey. It’s not a census. That’s no way to identify those who are BPL. The Union government wants to use NSS, and the consequent BPL figures churned out by the Planning Commission (though this body is in limbo now), because it wants to fix the share of subsidy expenditure that will be borne by the Union government, as opposed to the states bearing the costs. The only way to identify BPL properly is through a decentralized process, at the level of the gram sabha, not through socio-economic surveys. If that BPL list is publicly put up, errors of commission and omission can both be rectified. I see no reason why this shouldn’t work for rural India, though urban India does constitute a problem.

Third, despite the urban caveat, it should be easy to identify BPL (those who deserve subsidies) and APL (above the poverty line, those who clearly don’t deserve subsidies). The entire debate has been hijacked by those who are not clearly BPL and not clearly APL either. We are going round and round in circles because of them.

In the present discourse, we are talking primarily about petroleum products, especially LPG. As far as I can make out, the idea is direct benefit transfers, linked to bank accounts, linked to Aadhaar. However, Aadhaar is a tool, it doesn’t solve the problem of BPL identification. That’s a socio-economic cum political problem. There are recognized problems with Aadhaar. What it nevertheless does do is to eliminate multiplicity, because biometry rules out the chance of two individuals, with two different names and the same biometry, registering at the same address. Moving beyond that, there are still problems with financial inclusion, the Pradhan Mantri Jan-Dhan Yojana notwithstanding.

Consider, for instance, RBI’s mandate of financial inclusion. That extends to habitations that have 2000 households, and may go lower, perhaps to habitations that have 1000 households. The problem is the following. We don’t have data from the 2011 Census. But if the 2001 Census is any indication, there are probably 125,000 villages that have less than 1000 households. Those remain marginalized and bypassed and must be progressively integrated.

Stated differently, what’s being pushed by the Union government is only a beginning and must be taken to its logical step of dismantling the edifice of subsidies. That has to be spliced with decentralization of public expenditure, down to the level of local bodies. That’s the layer of government where most goods and services are actually delivered. Citizens have begun to demand better delivery. For the supply-side to adjust, decentralization is necessary. This is also part of the subsidy overhaul agenda.