Search interesting materials

Saturday, April 16, 2016

Subsidies are the last refuge of a failed policy maker

by Ajay Shah.

The government wants to subsidise your use of energy efficient household appliances. RBI wants to subsidise merchants who accept cards. Many people want to increase Internet access by using subsidies, e.g. the money disbursed through the `USOF'. Are all these subsidies appropriate?

Market failures versus your pet peeve

My pet peeve about the world is that there isn't enough classical music. Does this justify State intervention? The careful answer of an economist is: No, the scope of interference must be limited to `market failures'. Market failures are situations where the free market gets the resource allocation wrong. These are also generally situations where the free market gets the price wrong.

A useful four-part classification of market failures helps us look for them. The four categories are: market power (e.g. a monopoly that jacks up the price and cuts the quantity produced), asymmetric information (e.g. you buy medicines at a shop but you don't know whether they're adulterated), public goods (e.g. law and order, where the private sector will always under-produce), and externalities (e.g. your smoking gives me cancer).

The absence of an opportunity to trade widget $x$ at time $t$ is not, in general, a market failure. It may be a perfectly rational outcome for the retailing industry to not sell woolens in the desert. It may be a perfectly rational financial market where limit orders of a certain kind are absent. We may want more prosperity, more development, through which these markets would be more active, but the absence of trading in widget $x$ at time $t$ is not, in general, a market failure. In India, often times, we find that abnormally low trading is caused by government failures where the government is banning or interfering with various kinds of economic activity.

Similarly, when we see person $p$, and we dearly wish that he could buy butter, but he is too poor to buy butter, this is not a market failure. It's poverty.

When we see something that is going wrong in the world, the first hygiene check should be: Is this just my value judgement, or is this actually a market failure? Just because I like classical music, this does not mean that the slow death of classical music is a market failure.

Addressing market failures

When faced with a market failure, we may try to come up with a State intervention which would address this market failure. This is a dark art which involves the following considerations:

  1. The smallest use of force is the best. The best intervention that gets the job done is the one which uses the least coercive power of the State.  Spending money is grounded in taxation, which is a high use of the coercive power of the State. Each rupee of money spent by the government probably imposes a cost of Rs.3 upon society. We should be cautious before going there. In general, the engineering question -- the search for tools which address a market failure using the least force -- requires scientific knowledge in the form of a good understanding of the market failure. As an example, a careful analysis of the problems of energy efficiency leads to solutions very different from subsidising energy-efficient equipment.
  2. Taking implementation constraints seriously. Some interventions are infeasible in the light of the constraints of State capacity. Sometimes, the incentives of politicians and officials are impossible to correct, and while we can conceive of a nice tool for addressing a market failure, this may prove to be administratively infeasible. When implementation constraints are unsurmountable, we should just walk away leaving the market failure untouched. The lower the State capacity in a country, the more we should push towards laissez faire. Addressing a large class of market failures through State intervention is the luxury of people who possess State capacity. There is libertarianism of necessity and there is libertarianism of choice; we in India have to often engage in the former.
  3. Subsidies are the tool of choice when faced with one kind of market failure: externalities of the positive kind. Consider education. When person $p$ gets more educated, he captures certain gains, but there are also gains to society at large which are not captured by him. These positive externalities are not valued by person $p$, who would tend to under-invest in education. This under-investment is a market failure: the free market outcome is the wrong resource allocation. We can correct this market failure by having a subsidy. While this logic is true some of the time, not all positive externalities can be addressed using subsidies. Here's an example : of market failure in the undersupply of criticism.
  4. How big should the subsidy be? This requires two pieces of measurement. We should measure the positive externalities. The magnitude of the subsidy should be set to the point where the marginal gains to society from the last unit of subsidy is equal to the marginal social cost of funds. At present, in India, we have little empirical economics ability on measuring either. When we don't measure these things, caution requires a bias in the downward direction: push towards low or zero subsidies.

Addressing the `digital divide' using subsidies? 

Consider recent debates about the `digital divide'. Facebook and others proposed net non-neutrality as a way to increase their profit rates, and claimed that this would address access problems to the Internet. Proponents of net neutrality said that if access was a concern, this could be achieved by issuing bandwidth vouchers through which the government pays for (say) the first Rs.50 of data consumption per month by each citizen of India. Yes, this can be done. But should it be done?

Suppose most consumers use data comm to watch cat videos. In this case, why should the violence of the State (taxation) be brought to bear on increasing the consumption of cat videos? Before we propose a subsidy that would increase consumption of data communications, we should have measures of the spillovers, the positive externalities.

Ordinarily, a subsidy is funded by general tax revenues. In this case, people often rush to proposing the use of the `Universal Services Obligation Fund' (USOF). This is a less efficient means of raising public resources as it is a tax on one sector. The marginal social cost of USOF money is higher than the marginal social cost of funds. Intuitively, if the marginal social cost of funds is 3, I suspect the marginal social cost of USOF funds is 6. A subsidy funded through USOF would thus have to be significantly lower than a subsidy funded through general tax revenues.

Aadhaar and the Indian debate on subsidies

Nandan Nilekani and his team were very controlled in their arguments. They said: We have no opinion on whether the purchase of LPG should be subsidised, but assuming you want to do this, here's a way to better implement this subsidy. The re-engineering of many subsidy programs using Aadhaar has occupied public attention and absorbed management capacity in government.

But this does not change the basic question: Why should we have the LPG subsidy in the first place? Now that the Aadhaar system is built, we should go back to asking deeper  questions. Do we really want to have a fertiliser subsidy?

The rampant use of subsidies

We can't get our policies on ATMs right, so let's just subsidise ATM placement in northeast states. This sort of policy reflex is found all over the Indian policy landscape.

We failed to get our policy framework on payments right. Now we see a point of pain: acceptance infrastructure is lacking. The solution: subsidies in the form of an `Acceptance Development Fund'.

We failed to get prices right and interest rates right. Hence, households have incentives to buy energy inefficient equipment. The solution: subsidies.

We failed to make the Bond-Currency-Derivatives Nexus work. The solution: create a specialised class of financial firms named `primary dealers' and subsidise credit to them.

We failed to make infrastructure financing and the Bond-Currency-Derivatives Nexus work. Infrastructure projects are unable to sell corporate bonds. The solution: subsidies in the form of tax exemption for these bonds.

A conjecture about the rampant use of subsidies

Why are subsidies rampantly used in India? Why are policy makers so quick to reach into their toolbox of diverse possible interventions and pick on the subsidy?

Perhaps there are many market failures out there. Perhaps there are pervasive failures of policy analysis, misbehaviour by politicians and officials, and a shortage of State capacity. Hence, many market failures are not addressed. Solving these problems at the root cause is hard. The easy way out is a subsidy.

From a political point of view, sound policy work on addressing market failures involves using the coercive power of the State in order to force certain persons to behave differently. This is always unpleasant and makes some people unhappy. Spending money, on the other hand, is always popular. The unhappiness is concentrated against the Ministry of Finance which has to use State violence in collecting taxes. Once taxes are collected, every department of government, and every government agency, is too happy spending money. This is the soft option, compared with actually addressing market failures.

Actually fixing policy institutions is hard work. The leadership of most policy institutions lacks the ability to achieve State capacity. Subsidies are a tempting alternative for an `action-oriented government'.  Samuel Johnson said `patriotism is the last refuge of a scoundrel'. In similar fashion, subsidies are the last refuge of a failed policy maker.


If you believed that income inequality was a problem, the subsidy which addresses that is a transfer of Rs.1000 per person, every month, to the bottom decile of society. There is no need to interfere with the working of the market economy, for the purpose of reducing inequality.


Many times in India, subsidies are being used to express sheer value judgments; they are just the faddish thinking of one bunch of hausfraus running policy versus another. At other times, a market failure is indeed present. But instead of more subtle interventions and the minimum use of force -- based on a sound scientific understanding of the anatomy of the market failure -- we tend to rush to the excessive use of force that is a subsidy. Every subsidy is grounded in the monopoly of violence of the State that is required for tax collection. We should be far more circumspect before doling out subsidies. Subsidies are the last refuge of a failed policy maker.


  1. I would suggest collecting (say 10) of these basic (but profound) articles on policy and making an ebook and making it available for free, or some nominal charge on kindle, etc. Maybe give it away for free with Chetan Bhagat novels to educate the young. How useful will it be if this level of understanding becomes consensus in our country?!

  2. Ajay, your essays have become clear, concise and more direct, a welcome development. Wish there are more practitioners writing as directly as you do. Having said that, please be careful in pushing guaranteed basic income approach that has become the fad of the global progressives. Incentives matter.

  3. Ajay, while the economic notion of state intervention only in the case of market failure is reasonable, the term market failure needs careful scrutiny. What is your view, when subsidies are used (tax exemption or other) to promote the industrial development of a region by maximum participation of private sector, at the same time avoiding the direct involvement of state undertakings?

    Secondly, on use of Aadhar, do you see the rationalisation of subsidy recipients as a strong enough tool to correct the excesses of subsidies in India?



Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticism is perfectly okay; uncivilised language is not. We delete any comment which is spam, has personal attacks against anyone, or uses foul language. We delete any comment which does not contribute to the intellectual discussion about the blog article in question.

LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.