## Saturday, May 16, 2015

### Voluntary participation in the National Pension System: What does the evidence show?

by Renuka Sane.

Long-term saving is challenging in most parts of the world. Individuals are impatient, and old age is too far away. Rising life expectancy and potential poverty in old age have led countries to set up state funded pension programs or mandate contributions through the employer. Both these are difficult to implement in India. For example, the EPFO covers only about 8-10 percent of the workforce. This makes the voluntary build-up of savings important. Informal sector workers often do not have access to formal finance, and are unable to save large sums of money in one transaction. Poor people may also find it difficult to forgo current consumption and get invested in illiquid pension assets. There is a case for the State to facilitate a formal savings mechanism, and encourage pension accumulation through co-contribution.

These ideas started gaining ground after the `National Pension System (NPS)' (which used to be called the New Pension System) had been in operation for a few years. The NPS is grounded in the philosophy of self-help and thrift. It is mandatory for central government employees since 2004, and accessible to all citizens of India. The NPS-Lite model was introduced for the informal sector, followed by the launch of the NPS-Swavalamban (NPS-S) scheme in 2010. Under the Swavalamban scheme, if a subscriber in the informal sector contributes a minimum of Rs.1000 in a financial year into her NPS account, she receives a co-contribution of Rs.1000 from the government. The scheme has been operational for four years now, and the co-contribution was promised to last until March 2017.

In the recent Budget, the Finance Minister announced another informal sector pension scheme, the Atal Pension Yojana (APY) which promises a fixed pension of at least Rs.1,000 at age 60 if subscribers contribute pre-defined amounts over their working life. While the APY has several design flaws, it seems likely that it will replace the Swavalamban scheme before 2017, at least for those between 18-40 years of age. NPS-Lite, i.e. the NPS without the co-contribution, is likely to remain in place.

It is important to take stock of what has been the response of the informal sector to NPS-Lite/Swavalamban before taking policy measures on the same. Are people enrolling in the scheme? What kinds of contributions are they able to make? Do we have the policy and processes in place for when customers retire?

### How are enrollments and accumulations faring?

There is often skepticism about the ability of poor people to save. However, research has demonstrated that when provided with formal channels, poor people do save, sometimes at high cost. For example, Mukherjee (2014) finds that the willingness of people to save in a co-contribution pension scheme is high.

Aggregate official data also show that subscriptions to the scheme have been rising. According to the 2013-14 Annual Report (Table 1.6, page 22) of the PFRDA, there were a total of 2.8 million customers of NPS-Lite/Swavalamban. They made up 43 percent of the total NPS subscribers, and were the largest category of subscribers - more than government employees who make up a total of 30 percent of the subscriber base. The AUM under NPS-Lite/Swavalamban was Rs.8.4 billion, about 2 percent of the total NPS AUM. The percentage growth of AUM at almost 94 percent, between 2012-13 and 2013-14 was the highest for NPS-Lite/Swavalamban.

Sane and Thomas (2015) analyse participation and contributions of customers over the first three years of the scheme in more detail, using data from one financial services provider, the Kshetriya Grameen Financial Services (KGFS). They find that voluntary participation in an individual account DC pension system is feasible. In fact, it is the relatively poor in the sample that are more likely to open Swavalamban accounts. The evidence on persistence, is however, not as optimistic: only about 50 percent of the participants had managed to contribute more than Rs.1000 at least in one financial year in their NPS-S account. However, non-contribution in one year did not mean dormant accounts - several customers came back the next year. In terms of total contributions, members stop short of contributing more than Rs.1000 in a financial year. Part of this seems to be driven by the scheme becoming centered around the threshold for the Rs.1000 co-contribution. Members could actually contribute larger amounts, but often do not, because the scheme is sold as a Rs.1000 per year contribution scheme.

### The problems in the draw-down phase

A pension scheme is ultimately judged by its ability to provide for an adequate consumption in retirement. Accumulations are only one part of the story. Since the accumulated wealth has to provide for a meaningful consumption over the lifetime of the individual, how this wealth is drawn-down becomes important. All the NPS models, including NPS-Lite/Swavalamban require that 40 percent of the account balances be used to purchase an annuity, while the remainder may be drawn-down as a lumpsum. There is currently no option of a programmed withdrawal, where part of the retirement fund is used for a draw-down (as income withdrawal) while leaving the rest of it invested.

Annuities can be expensive for the poor as they have a lower life expectancy than the rich. If they die early, they effectively end up subsidising the rich. We, therefore, need to think more carefully about the choice between annuitisation and programmed withdrawal. Different countries have approached the question of annuitisation differently, and are largely influenced by existence of a state funded pension which offers protection from poverty in retirement. The Chilean approach, for example, has been to restrict lump-sum distributions, and mandate the use of fixed inflation-indexed annuities or lifetime phased withdrawals. The Australians, are more flexible in allowing lump sums. Most recently, the UK has done away with its rule of mandating the purchase of an annuity by the age of 75, and allows for programmed withdrawals. The US has very little mandatory annuitisation.

Life insurance companies are often reluctant to enter into annuity markets because of the lack of availability of good mortality tables as well as instruments for hedging longevity and inflation risk. Lack of good mortality data is especially true in the case of low-income customers. The nominal annuity may also not be able to buy a minimum consumption basket if inflation rises over the lifetime of the retiree. If the administrative costs charged by insurance companies are high, then the value of the annuity will fall further.

Benefit policies of NPS-Lite/Swavalamban thus require a re-think. Enabling the development of mortality tables, market for inflation indexed bonds, changing the procurement of annuity service providers so as to minimise the costs of the annuity, designing default options for those who cannot choose the optimal combination of annuity and lumpsum are some of the policy initiatives that the PFRDA needs to undertake.
Similar questions are pertinent for the NPS as well. However, government employees who were enrolled in the NPS starting 2004 still have some time before they retire. The NPS-Lite/Swavalamban members who enrolled in their late 40s will get to the retirement threshold sooner, and bad design of the draw-down policy or delays in providing benefits can potentially destroy the foundation that has been built for improving informal sector participation.

### Conclusion

There are several take-aways from the experience of the NPS-Lite/Swavalamban schemes:

1. The number of people contributing Rs.1000 is gradually increasing.
2. Non-contribution in one year does not mean subsequent non-contribution.
3. There seems to be a hump in contributions at Rs.1000, most likely driven by the threshold design of the co-contribution.
4. Benefit design policies and processes require a re-think.

The NPS-Lite/Swavalamban is gradually taking root, people are beginning to understand the scheme, and intermediaries are learning how to distribute it. Familiarity with the scheme and intermediaries is likely to build the trust that is important in fostering long-term illiquid pension contributions. The infrastructure required for channeling contributions to the fund managers seems to be largely in place.

Analysis shows that the APY by itself is not enough to meet consumption needs in retirement. Thus, even if the APY replaces Swavalamban, intermediaries should consider continuing to distribute the NPS-Lite, as a combination of APY and NPS-Lite may allow customers to enjoy higher returns than the APY alone. A minimum amount of contribution could be made to the APY towards the guaranteed pension, while the remaining can be invested in the NPS-Lite for potentially higher returns. The PFRDA needs to incentivise the sale of both the APY and the NPS-Lite, dislodge the mental threshold of Rs.1000 to encourage contributions of larger amounts, and do a rethink of the draw-down phase.

### References

Mukherjee (2014), Micropensions: Helping the Poor Save for Old Age, Paper presented at the 5th Emerging Markets Finance Conference.

Sane, R. and S. Thomas (2015), In search of inclusion: informal sector participation in a voluntary, defined contribution pension systemJournal of Development Studies (forthcoming).

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