Retirement study: High IRR?

I refer to the report “Prudent housing choices ensure enough CPF savings for retirement: survey” (Channel NewsAsia, Nov 14).

It states that “And men earning median incomes at the 50th percentile should be able to replace 70 per cent of their wages on retirement at 65”.

According to the study:-

“In contrast to previous studies which assumed constant wage growth, we used data collected by the Ministry of Manpower (MOM) for the Labour Force Survey (LFS) over 2001 to 2011 to simulate real wage growth paths for individuals”.

“It is a hump-shaped distribution of earnings by age where wage growth is faster when the worker is young and tapers off into the negative as he gets older”.

From the graph in the study, it appears that real earnings start to decline from around age 38, for males at the 50th percentile for earnings.

So, my understanding is that at age 55, which is the age used to compute the Income Replacement Ratio (IRR), the real earnings would be about the same as that at around age 33.

So, does this mean that we are assuming that one would be earning at age 55, the real earnings equivalent of what one earned at around 33?

In other words, the amount that is used to determine the IRR may not be what most layman may think that as they grow older – they would earn more.

IRR – Based on net earnings?

“To calculate net IRR, the denominator is the assumption of 85% employment density factored in. The after-tax pre-retirement earnings are computed by deducting personal income taxes and employee CPF contributions from the gross pre-retirement earning of workers at age 55”.

Does this mean that the earnings at 55 are further discounted for “85% employment density” (periods of unemployment or economic inactivity), personal income taxes and employee CPF contributions?

In other words, the amount that is further discounted to determine the IRR may not be what most layman may think that it is based on their gross pay at age 55.

IRR – Include Imputed Rental?

“Adjusted Net IRR with Imputed Rent

Including imputed rent in the calculation lifts the net replacement rate by an additional 4 to 9 percentage points for men and additional 7 to 11 percentage points for women. When comparing Singapore’s IRR with those of countries with low home ownership such as Germany, where rental is the norm, it would be more appropriate to use IRR with imputed rent”.

My understanding of this is that as it is assumed that most Singaporeans own their homes, and do not pay rental, an adjustment is made to increase the IRR to include imputed rent.

In other words, it may not be what most layman may think that the IRR is the actual cash that they can use, but rather an inflated figure assuming that they would otherwise have to pay rental.

In summary, the IRR is based on age 55 earnings that are discounted for declining real earnings as one gets older, periods of assumed unemployment and economic inactivity, imputed rental, etc.

Are there any other countries in the world that analyse retirement adequacy on this basis?

CPF Life assumptions?

“Retirement Adequacy when up to the Maximum of MS is Annuitized

In the base case model, we do not put a cap on the maximum amount that can be annuitized. At age 55, members can use the accumulated CPF savings to buy CPF LIFE products up to the maximum of MS. CPF savings in excess of the MS are left in their respective CPF accounts and accrue interest at the CPF interest rates. Subsequently, at age 65, this money together with further CPF contributions from work after age 55 is used to buy an additional life annuity that is similar to CPF LIFE. We assume the existence of a private annuity market which will give payouts comparable to CPF LIFE ”

Currently, “a private annuity market which will give payouts comparable to CPF LIFE “, does not exist. So, the IRR computed may be a bit over-optimistic.

The study assumes a floor rate of 4% for the CPF Special, Medisave and Retirement Accounts, whereas when the current 4% guarantee ends on 31 December, 2013, the floor rate is 2.5% if the current average 10-year government bond yield plus 1% remains at it current low level.

Moreover, the CPF Life annuity payouts are based on the 4% instead of the 2.5% floor rate. I would like to suggest that another projection be done based on the 2.5% floor rate.

Other studies – Low IRR?

This study is also not inflation-indexed for the retirement annuity, like the “OECD (2012) reports the gross IRR for Singapore to be 13% for a working career of 40 years and 9.3% for a shorter career of 30 years (See OECD, p. 36)”.

Of course the study explains in great detail why the OECD and other studies differ in methodology and assumptions which I shall not dwell into, as they are very complicated.

I understand that all studies done so far on Singapore assumes a constant wage increase (such as 2%) methodology, against the subject study’s declining real earnings as one gets older assumption.

In my opinion, a constant wage assumption may be more appropriate as arguably retirement adequacy should be about being able to meet expenses which generally increase with inflation, regardless of the Singapore context of declining real wages with age.

The reason given for using age 55 earnings is “we modeled earnings using empirical wage growth for different age bands as described in Section 2.1, where earnings generally decline in later life. Earning just prior to retirement at age 65 is therefore too low a proxy for pre-retirement earning”.

In my view, if like the other studies whereby the earnings at 65 (the retirement age) is used instead of 55, the “discount” factor may be higher due to even lower earnings at 65, such that the resulting IRR may be so high that nobody in the world may believe it.

An ideal life story?

The study is fundamentally based on an ideal scenario – buying a BTO flat with HDB grant at age 30, without the realism that at its peak, more than 40,000 HDB households were unable to pay, and we have no regular statistics as to how many people are in arrears for more than three months on their HDB and HDB bank loans, HDB and bank foreclosures, etc.

Also, with the current trend particularly of PMETs losing jobs, difficulty in finding a similar reasonably paying one, etc, how many Singaporeans can expect a life that is smooth sailing all the way?

In my view, an independent academic study commissioned by the Government on such an important subject as the retirement adequacy of the nation, can be more robust by giving more attention and mention to the statistics, problems, issues and concerns of the present. The study seems to be overly focused on projections of the future.

For example, I feel that the study may not have adequately addressed the fact that currently, for every active CPF member, there is almost one other inactive CPF member who may have very little savings for retirement.

Now vs future projection?

You may also like to read an analysis on the current retirement adequacy here  
(not projection of IRR in the future).

Leong Sze Hian

About the Author

Leong Sze Hian has served as the president of 4 professional bodies, honorary consul of 2 countries, an alumnus of Harvard University, authored 4 books, quoted over 1500 times in the media , has been a radio talkshow host, a newspaper daily columnist, Wharton Fellow, SEACeM Fellow, columnist for theonlinecitizen and Malaysiakini, executive producer of Ilo Ilo (40 international awards), Hotel Mumbai (associate producer), invited to speak more than 200 times in about 40 countries, CIFA advisory board member, founding advisor to the Financial Planning Associations of 2 countries. He has 3 Masters, 2 Bachelors degrees and 13 professional  qualifications.