OPS, NPS and the AP alternative

The outcomes (in terms of pension as a percentage of final salary) of the New Pension Scheme (NPS, or the National Pension System) were estimated in the first part of this article (May 30). It was noted that the Old Pension Scheme (OPS) will definitely have a larger present value of pension payouts compared to the NPS. What is the extra government contribution that would be needed to fund such a higher pension?

Suppose pension payouts equal to OPS are funded with a contributory scheme (‘funded OPS’) having higher government contribution and no employee input. If, at retirement, the Mean ETV (Expected Terminal Value) of the amount of corpus (i.e. sum of contributions plus investment return) is equal to the Mean EPV (Expected Present Value) of OPS pension payments, then the government would have enough money to provide a pension that is (on average) equal to that under OPS.

Using the assumptions in Part I of this article, the EPV of pension under the OPS was modelled. To find the new required government contribution rate, the ratio of EPV of OPS to the ETV of NPS Corpus was calculated. Using the average of these ratios over 1,00,000 simulations, the required total contribution to match OPS is given in Column (II) of Table 2.

The required contribution rate is around 50 per cent for Rajasthan and 40 per cent for Chhattisgarh (difference due to varying life expectancy). As there is no employee contribution in the OPS, this total contribution has to be paid by the government. It would have to pay 5.003/3.945 times the original NPS Government Contribution, respectively. Under the NPS, these State governments were paying only 10 per cent of the salary; now, they would have to pay 50/40 per cent.

We can now compare OPS and NPS, as now OPS is being represented by the funded OPS; we have two baskets of apples to compare. The initial cost of the shift is estimated in Table 3 using the following data:

So, as an answer to all the suppositions about the cost to shift being significant or insignificant, it would cost Rajasthan around ₹49,000 crore. And it would cost Chhattisgarh around ₹12,000 crore. This is the one-time amount required to top up the NPS corpus if it is to be converted to a funded OPS. This is just the cost till now of transferring all existing staff under NPS to OPS.

In future there will be extra costs for the existing staff. And, every year, new people will also be added to OPS. These people will not contribute to their pension cost, and the entire cost of their pensions is on the government. So, apart from the figures of ‘Initial Fiscal Cost’ in Table 3, there will be an annual accrual of costs for existing NPS-holders and for new staff. Since the contribution rate is rising from 10 per cent to 40-50 per cent, the government will incur a cost 4-5 times the current annual cost.

Budgetary impact

What would the budgetary impact be? Using NPS Corpus data, government contribution to NPS was estimated at ₹2,732 crore (Rajasthan) and ₹1,060 crore (Chhattisgarh) for 2021-22. In shifting to an unfunded OPS (i.e. pay as you go), they would stop making these contributions.

In cash terms Rajasthan is likely to save more than ₹2,732 crore. in 2022-23, a reduction of over 10 per cent of the pension budget; Chhattisgarh is likely to save at least ₹1,060 crore, a reduction of 14 per cent.

However, for a funded OPS, the total pension cost for Rajasthan for 2021-22 would increase by ₹10,936 crore, and for Chhattisgarh by ₹3,122 crore; it would be even more for 2022-23. A prudent approach would be to account for such a liability each year. But with the proposed OPS being an unfunded scheme, this cost would not be provisioned annually, and would be loaded onto future governments.

Hence, though the liability for post-2004 staff is actually increasing with the shift, their cost would disappear from the Budget (which may be the main reason governments are attracted to revert to OPS). Costs of the shift would become known only when the staff hired from 2004 start to retire. This will hurt governments of the 2030’s, while current governments benefit from higher fiscal space.

One also cannot argue that over the years inflation will reduce the costs of the shift, as the OPS is also inflation indexed (through Dearness Relief).

The figures above are approximations based on modelling of publicly available data and reasonable assumptions. However, they establish that the fiscal costs are remarkably high.

AP model

The main argument of staff against the NPS is that its benefit is undefined and that the pension depends on unpredictable markets. Is NPS the only alternative to OPS that is fiscally prudent? An alternative proposed by the government of Andhra Pradesh is a guaranteed level annuity as pension. Being a defined benefit scheme, the government takes over the investment risk. However, staff and government contributions both continue. This has two benefits. First, only a part of the cost of pensions is on the government; second, the cost is funded in each year when it accrues and is not transferred to future generations and governments. This scheme would be sustainable if the proportion of the guaranteed level pension is chosen wisely.

If implemented properly, this could potentially make pensions stable for staff and fiscally prudent for governments, avoiding the opulence of OPS and the niggardly NPS.

Overall, the figures indicate that OPS is a ticking time bomb that was defused insufficiently with the NPS; it’s starting to tick again.

The writer is an economist and student of the Institute & Faculty of Actuaries, London