The perils of pay-as-you-go

When countries embark on building pension systems, one of the first "solutions" which appears handy is like this: the government pays a pension to the old, which is financed by taxing the young. This is called a "pay as you go" system. Such a system calculates the required outgo to the elderly, and then finds out the required tax rate.

In numerous countries, such systems came about at a time when the number of the elderly, and the tax burden on the young, was small. However, when the demographic transition takes place (as it does in every country), the number of the elderly starts rising sharply. People live longer, and fewer children are born. Both these combine to generate extreme fiscal stress.

When paygo systems run into trouble, they are very painful to reform. The elderly do not want to see any reduction in the benefits that they had been promised all their life; the benefits which had always figured in their old age planning. The elderly often mobilise themselves into political coalitions to block any reduction to their benefits.

But the arithmetic is simple and brutal. Suppose the average wage is Rs.100 and the pension is Rs.50. Suppose there are two workers for each elderly person. Then the payout to the elderly person (Rs.50), has to go with a 25% tax rate purely to finance the pension outgo. All other government expenditures have to be met from additional taxation.

Going beyond the demographic transition, paygo systems suffer from other profound problems. In the paygo system, the contributors are forcibly paying taxes. They only have an incentive to pay as little tax as possible. The recipients are only receiving a dole; they only have an incentive to obtain as much as possible. Neither the young nor the elderly have any incentive to check the government when it is irresponsible. This gives the government tremendous political leeway in making pension policies. It is common to see a slow and steady creep upwards of benefits, since raising benefits is always popular.

All this is in gross violation of elementary good sense in finance. If a monthly pension has to be paid to a person until he dies, then this is a large liability. Using life tables, it is possible to precisely attach a number to it (it is the price of an annuity).

For example, in India, if a person is to get Rs.1,000 a month from age 60 till death, this is worth around Rs.200,000 today. If the government has an obligation to pay Rs.1000/month till death to one person, it ought to have an asset of Rs.200,000 in hand.

From a financial perspective, paygo systems are simply bizarre. A financially sound approach would require that every liability should be matched by a corresponding asset. Hence, if there is a promise to pay pension, then it should be matched by assets which reflect a sound valuation of the net present value of the pension.

In a paygo system, even if an attempt to build up assets is made, it is fraught with political risk. From a workers point of view, once the government has made a promise about a future pension, he has no incentive to act in the sphere of politics when it comes to the management of the assets. Whether the government invests these assets wisely or the assets are squandered; whether the assets are sufficient or not -- the worker has no incentive to worry about these questions because he is promised a fixed pension. Further, the worker has every incentive to insist that his contribution should be as small as possible, and weak governments are likely to succumb to the temptation of promising generous benefits while charging small contributions.

Hence, the modern wisdom in pension reforms is to avoid defined benefit programs and to avoid paygo programs. Both these alternatives have their own merits in some respects. However, the political risks are large, and pension assets are sizeable enough that mistakes of this nature could destabilise the economy.

Turning to India, we do have these problem with two parts of the pension system:

  1. Civil servants pension The pension that the government of India pays to its employees (including railways, defence, etc) is a pure paygo. The benefits obtained by each retiree are defined and inflation indexed. The money needed by the government in order to meet these obligations is just raised through tax revenues (or deficit financing). The fifth pay commission has done projections which suggest that by 2005 AD, the pension outgo of the government will exceed its wage bill.
  2. EPS 1995 is a defined benefit program. This suffers from all the political risks described above. The asset management of EPS 1995 leaves a lot to be desired, but individual participants have no incentive to undertake political actions to change this since better fund management will not change their benefits. There is a tremendous political risk that in the future, government will be tempted to be over-generous in payouts.

Both these problems have been discussed in public debates and there is a wide consensus about the need for surgery. India's problems here are not unique: most countries do have a problem with programs like EPS 1995 and most countries do have a civil servants pension program which needs to be converted into being fully funded. Where we are unique is in the much bigger scale of paygo programs. Most countries suffer from paygo pensions; we also suffer from paygo banking!

The post office savings programs, PPF, etc. are paygo savings schemes! When an individual deposits money into the program, it is consumed by the government right away. When the individual withdraws money, the government funds this out of taxes.

In this situation, the "interest rate" paid by the government on these programs is just a fiction. Indeed, government can set these "interest rates" to just about any number that can be imagined. It is possible (though highly unwise) for government to pay 12% interest rate on these products, even though this translates into a 9% real rate of return.

The way forward. The way out of this mess is not hard to describe. There should be no paygo programs. Every liability should be backed by a visible asset. The liabilities that exist, today, should be converted into assets by a one--time bond issuance. For example, PPF should be converted into a professionally run fund, and the assets of the fund should be initialised with government bonds which are issued in place of the future liabilities.

The fiscal fragility of the government of India is much worse than is apparent, once these liabilities are added into the stock of government debt. Yet, we will be on the right path if these underhand liabilities are made explicitly visible. Further, if these paygo programs are converted into benefits that are backed by assets, then participants will have an incentive to worry about the rate of return that their assets are getting and undertake political actions to improve the governance of these assets.

Back up to Ajay Shah's media page