A comparative analysis of real CPF returns and other provident funds
“If that is the case…does the Government short-change Singaporeans by giving CPF members 3.5 per cent of the interest rate while the GIC makes 9 per cent (and) pockets the balance of 5.5 per cent?” -Worker’s Party Secretary-General Low Thia Khiang, parliamentary session on 18th Sept 2007.
“Sir, the poor rate of returns on CPF balances is one of the main reasons why the existing CPF scheme, without reform, will be inadequate to meet Singaporeans’ retirement needs. As a long-term retirement savings plan, the CPF’s rate of return is crucial. Unfortunately, the rate of return enjoyed by CPF members has been poor.” – Then-NMP Siew Kum Hong, parliamentary session 18th Sept 2007.
“You can’t completely guarantee this system that you are providing now. This is because there isn’t any risk-free asset that can be certain to achieve the 2.5 per annum minimum return.” – Then-Manpower Minister Ng Eng Hen in reply to Low and Siew.
How does the Singapore’s CPF rate of return measure up against those of similar pension funds of other countries? This was a question which I sought to answer back in 2009. Back then I promised readers (mostly consisting of myself) that I would update the post with real returns of other provident funds instead of mere nominal figures. I barely found time (and motivation, it’s a lot of work to source for and dig up data) to do so only until somewhat recently. While nominal CPF returns may appear comparatively low, Singapore’s low inflation rate (disregarding issues of whether inflation is adequately measured in Singapore) leaves open the possibility that CPF provides a better real rate of return as compared that of other countries.
Notes: This post may be considered an update of the older 2009 one, but much of its contents will be plagiarised; there’s no need to read the earlier 2009 post if you’re reading this one.
Singapore’s CPF earns the dubious honour for having the highest mandatory contribution rate (thus robbing its depositors the most of their take-home pay) and providing the lowest return (annual interest accrued where relevant). For this article, I examined four listed countries’ provident funds in Wikipedia, namely Singapore, India, Hong Kong and Malaysia. There may well be more provident funds, but for some reason Wikipedia only has four listed. These four schemes would be examined in detail below.
Malaysia’s provident fund is named the Employees’ Provident Fund (EPF). It’s a mandatory retirement plan whose contribution rate stands at 11% from the employee’s monthly payroll and 12% from the employer for a total of 23%.
As noted by Wikipedia, EPF is legally obligated to provide only a 2.5% return rate annually, but historically it has been far more generous than that. The CPI data for calculating inflation rate is provided by the Malaysia time series on consumer price index here. With this data, we may derive the real rate of interest paid on EPF deposits as follows.
|Inflation rate||EPF Interest rate||
Real interest rate
One might notice a gradual reduction in the dividend rate (M’sian counterpart of interest rate in Singapore’s CPF) throughout the years, but even at its lowest it gives a return of 4.5%, a rate which is higher than CPF’s for the SMRA account of 4% and 2.5% for the OA. For the years of 2009 and 2010 the EPF interest rate was hiked back up to above 5%.
Unlike Singapore’s CPF where a Minimum Sum of money (currently S$117,000 and increasing and separately S$32,000 for the Medisave Minimum Sum) has to be retained in addition to a compulsory annuity scheme upon reaching 62 years of age, Malaysia’s EPF allows its depositors to withdraw up to 30% of their money upon reaching 50, and everything else by the age 55.
India’s EPFO has a defined 12% contribution rate (compare this to CPF’s 35.5% contribution rate) from both employees and employers in total. For the year of 2008, the annual interest rate on an individual’s EPFO account was 8.5%. Here are the numbers from earlier years in the decade.
|Year||Nominal EPF interest rate||Inflation rate for||Real EPF interest rate for|
Notes: Real interest = nominal rate – inflation rate
The CPI data for India provided by India’s central bank, the Reserve Bank of India, is calculated separately for industrial workers (IW), white-collar employees (UNME) and agricultural labourers (AL). I was not able to find an overall historical CPI time series for all Indian employees and decided to calculate all three groups separately. Real EPF rates would differ for employees depending on their industry. Similarly historical EPF interest rates had to be sourced from various news sources as their main website did not appear to have archived all the past EPF rates on one readily accessible page.
For those interested, here’s an opinionated analysis of India’s EPF interest rates dated from Jul 2004.
Like Malaysia’s EPF, India’s EPFO depositors are able to withdraw the full amount in their accounts when they hit 55 years of age. That’s right, no sneaky Minimum Sum or compulsory annuity. Employees are also allowed to withdraw all their money in the following cases:
- A member who has not attained the age of 55 year at the time of termination of service.
- A member is retired on account of permanent and total disablement due to bodily or mental infirmity.
- On migration from India for permanent settlement abroad or for taking employment abroad.
- In the case of mass or individual retrenchment.
In the case of “termination of service” stated above, this refers to early retirement ie. termination of all employment services. India’s EPFO also allows an early withdrawal of up to 90% of their account when they reach 54 years old.
Now for Hong Kong’s Mandatory Provident Fund (MPF). HK’s mandatory pension plan resembles the 401(k) pension plan of the United States more than any of the other three. This will be explained in greater detail.
The contribution rate is 10%, with 5% each from both employer and employee, unless one earns less than HK$5000 a month, in which case the employee does not have to contribute, so total contribution stands at 5%. Those earning above HK$20,000 a month have their contributions from both employers and employees are capped at HK$1000 each. Employees may choose to invest more of their savings than stipulated should they wish to do so.
Those employed in the catering and construction industries have a more complicated defined contribution rate. Refer to here for details.
As explained above, HK’s pension plan consists of MPF schemes which are chosen by the employer. Once the employer has selected a suitable MPF scheme for the employee, the employee is free to choose which investment funds available in that scheme he would prefer to invest in. All schemes and funds have to be approved by the Hong Kong government.
How does Hong Kong’s MPF work? A 2002 paper by libertarian think-tank CATO explains as follows:
The MPF system is basically a forced retirement saving program. Investment funds, called MPF schemes, are set up by private companies to invest the savings. All MPF schemes must be established under trust arrangement and governed by Hong Kong law. The trust arrangement means that scheme assets will be held separately from the assets of the trustees or the investment managers. This safeguards the interest of the scheme members from unnecessary financial risks.
Investment managers are appointed by the trustees of MPF schemes to make long-term investment of scheme assets. Each scheme typically offers a couple of investment funds to its members. One of the funds must be a Capital Preservation Fund, which is basically a money market fund. There are strict guidelines on the types of assets in the investment funds.
Employers are mandated to choose at least one scheme for their employees.6 Employers therefore indirectly choose the investment managers. Employees can choose their own investment portfolio out of the funds provided by the scheme chosen by their employers.
Such a pension program carries the same risks as the 401(k) plan especially since Hong Kongers, unlike Americans, do not have guaranteed minimum returns or any defined benefits plan such as Social Security. Although at the same time this allows them to enjoy a higher rate of market-based returns. Hence there are no cited interest rates on MPF accounts. Like 401(k) accounts, Hong Kong’s MPF accounts were badly hit by the 2008-2009 global financial downturn. However, note that as compared to CPF, Hong Kong’s MPF has the lowest mandated contribution rate of 10%, so the employee has limited exposure to financial risks.
As for the withdrawal of benefits, Hong Kong only allows employees to withdraw all their savings when they hit the age of 65 barring special considerations such the following:
- early retirement at the age 60; or
- permanent departure from Hong Kong; or
- total incapacity; or
- death (note that the MPF will be regarded as part of the member’s estate and can be claimed by the personal representative of estate); or
- small balance account of less than $5,000, no contributions made to a scheme for 12 months, and declared not to become employed or self-employed within the foreseeable future.
Again unlike Singapore’s CPF, there is no required Minimum Sum or compulsory annuities. Because HK’s MPF function so differently from CPF and Malaysia’s and India’s EPF, I am not able to provide tabulated data which may be compared with the other three countries which have explicit declared interest rates for their pension accounts. In lieu of that here’s what renowned pension expert and economist Professor Mukul Asher wrote in a 2001 paper comparing Singapore and Hong Kong’s pension scheme:
Compared to Singapore’s 50-year-old CPF, which is centrally managed, non-transparent and non-accountable, Hong Kong’s MPF scheme, begun in 2000, incorporates nearly all the best international pension fund management and governance practices. Moreover, while members of the MPF earn market rate of returns and enjoy considerable choice, CPF members earn interest that bears no resemblance to returns from investing their assets, and its members enjoy very limited choice. However, both the MPF and CPF members do not enjoy protection against longevity and inflation risks, and do not address the needs of the lifetime poor.
Asia Sentinel has a good comparison between Singapore’s CPF and Hong Kong’s MPF here (dated from 2007). Among other things, it notes that CPF’s measly rate of 2.5% is barely above (if at all) the annual inflation rate:
The interest supplement is tacit acknowledgement of how far the forced savers have been subsidizing the borrowers – the Singapore government and ultimately US and other consumers who are being financed by Singapore’s savings excess. The current normal interest rate on CPF balances is 2.5 percent — barely above the rate of inflation. Indeed, for years, the interest rate has been about nil in real terms. It is noteworthy that while the giant state investment corporation Temasek boasts double-digit returns on investments, Singapore’s forced savers have been receiving a quarter of that amount.
The first consequence of this is that savings have not in practice earned anything, so balances are now far from adequate to sustain a reasonable standard of life for low-income retirees despite the fact that contributions to the CPF are 36 percent of income and were once as high as 40 percent.
Indeed, from what has been written above it’s a mystery as to why CPF provides such a low return rate of 2.5% especially given that its mandatory contribution rate of 35.5% is the highest of the four provident funds. To make things worse, the interest rate on the CPF SMRA accounts will no longer be a fixed 4% in the future but will be allowed to vary without a floor. More specifically the interest rate on that account would be pegged to the average 12-month annual 10-year SGS yield + 1%. Currently that rate stands at 3.37% for 2010. Here are the historical yields for the 10-year SGS yield. It’s not terribly high, to say the least. Note that in order to achieve 4% as it currently stands, the yield has to exceed at least 3%.
Finally here’s the data for Singapore. I assume most readers are familiar with how CPF works and there’s no need to provide an explanation. Accordingly I go straight to the tabulated data as follows. CPI data found here, while historical CPF interest rates may be obtained here. The CPF SA/RA variable interest rate is a purely hypothetical interest rate which would have been paid on members’ CPF SA account had the government all along adopted a 12-month average of annual 10-year-SGS yield + 1% as benchmark interest on that account. The government intends to adopt such a measure after Dec 2011. The data for that was sourced from here.
Note: The SA and RA accounts, while in actual fact are two separate accounts, are treated interchangeably here because the same rate of interest is credited to both.
|Year||Annual Inflation rate||CPF OA interest rate||CPF SA Interest rate||CPF OA real interest rate||CPF RA real interest rate||CPF variable SA Interest rate||CPF SA variable real Interest rate|
The following table sums up the preceding discussion on how the above-mentioned provident funds differ in vital characteristics. As stated above it’s apparent at least in nominal terms that Singapore’s CPF mandates the highest contribution rate amongst the provident funds while providing the lowest returns.
|Contribution Rate||Nominal interest rate||Partial Withdrawal||Full Withdrawal||Minimum Sum/ Mandatory annuity?|
|Malaysia||12%||11%||23%||5.8% (as of 2010)||50||30%||55||No|
|India||1.11%||12%||13%||9.5% (as of 2010)||54||90%||55||No|
|Hong Kong||5%||5%||10%||N/A (returns depend on performance of investment funds)||N/A||Not allowed||65||No|
|Singapore||15.5%||20%||35.5%||2.5% (OA) and 4% (SMRA)
Variable rate for SA/RA after Dec 2011
|55||Everything except Min. Sum + Medisave Min. Sum||N/A (not possible to withdraw everything due to CPF Life)||
Comparison of real returns from 1999-2009
In order to provide an adequate perspective on how real CPF returns measure up against that of the other countries, I’ve crunched all the above tabulated data and graphed them for better visualisation (click to enlarge) of real returns by the provident funds over the period of 10 years as follows.
Notes: IW- Industrial worker, UNME – Urban non-manual employee, AL – agricultural labourer
This graph requires some explanation. I would have preferred to use data available right up till 2010, but apparently India’s statistical office tracks changes in the CPI and decides on EPF contribution rates from a financial work-year from April to April, which explains why the tabulated data for India was presented as it was above. The data for 2010 April to 2011 April is still not available hence the omission. Hence for India I have taken the rate stated for year X to X+1 be for year X in order to do a comparative analysis. Secondly, the graph is consistent with what you might expect for real returns for all the pension funds in the year 2008, where a sharp spike in global inflation effectively eroded all real increments.
The pink line Singapore CPF Variable RA refers to the real rate CPF would have paid on all its depositors’ Retirement/Special Account as if it were 10-yr-SGS-yield + 1% all along. Currently Special and Retirement Accounts enjoy a fixed rate of 4% which has been so since 1999. Take note that this is a strictly hypothetical rate and not an actual one.
Some things are apparent from this graph. Firstly, there is absolutely no doubt just by glancing at the above that the 2.5% interest rate paid on CPF Ordinary Account is downright pathetic in comparison to the rest. Yes CPF OA’s 2.5% real interest rate loses out to both impoverished India’s EPF rate as well as Malaysia. Singapore consistently lost out to what India and Malaysia were able to provide for their citizens. This is especially worrisome given that the bulk of CPF contributions goes into the CPF Ordinary Account. How are 3rd-world countries able to deliver so much more generous real returns over a ten-year period where 1st world Singapore utterly fails to?
But lest I be accused of omitting the other CPF account, I’ve also calculated and included the real CPF returns for the Special Account (RA) which offers a more generous nominal 4% interest rate. Apparently it performs much better compared to the measly CPF OA which pays only 2.5%. But even then it still underperforms Malaysia’s EPF in real terms when average compounded real returns are compared. Of all the three countries, India consistently provided the highest real rate of return for its despositors. This can be more easily seen below:
|Country/Pension||Real compounded annual
return rates (CAGR) 1999-2009
|Singapore CPF OA||1.34%|
|Singapore CPF RA||2.99%|
|Singapore CPF Variable RA||3.21%|
|India EPF IW||4.01%|
|India EPF UNME||3.84%|
|India EPF AL||4.58%|
Notes: IW-Industrial workers, UNME-Urban non-manual employees, AL – Agricultural labourers
The best scenario CPF is able to offer is 3.21% over the 10-year period. This compounded real rate of return is still slightly below what Malaysia is able to pay its EPF depositors (3.26%) over the same period. To make matters worse, 3.21% refers to a strictly hypothetical rate CPF would have paid on its depositors’ Retirement Account if it had implemented its variable 10-year-SGS-yield + 1% right from the start. The actual rate paid on the CPF RA over the ten-year period is found to be a smaller 2.99%.
It can be seen from the graph that in the later half of the decade, even the hypothetical CPF RA rate tracks the actual real RA interest rate paid with little improvement. Does this mean that all those oft-repeated claims that floating the interest rate on the RA would pay depositors a rate closer to market based returns is nothing but a big exaggeration? This was in fact proposed by some experts which argued that CPF returns should be pegged to a floating long term bond yield. This hypothetical scenario shows that doing so might not be much different from existing reality, as evidenced in recent years.
Needless to say, both Malaysia and India handily beat Singapore on real pension fund returns. A sobering conclusion to reach, but nonetheless. Perhaps Singaporeans can take some consolation from the fact that India’s EPF is likely to deliver lower real rates of return in future due to soaring inflation in India at present. Then again, given that India’s Labour ministry has recently shown itself to be capable of standing up and fighting for a higher rate of return on EPF balances for its citizens, perhaps not. Time will tell.
Expert opinion on CPF’s adequacy
This post would not be complete if it didn’t include what pension studies experts have said or commented on the (in)adequacy of CPF returns. Hence this section is dedicated to what pension experts and academics have had to say about CPF.
Take for example this 2006 World Bank paper:
CPF balances are invested in government bonds, which carry lower risk but also bring relatively low returns. Since the Singapore government routinely runs a surplus, it does not need the proceeds of these bond issues for funding purposes. These funds are instead invested in external assets by the Government Investment Corporation. So while CPF members are sheltered from investment risks, the downside is that members are deprived of the opportunity to reap higher returns. On the other hand, although the actual investments in the GIC portfolio are not disclosed, given the long term nature of CPF assets, the government can maximize return by managing long term risk. It is therefore questionable whether there is a justification for not sharing these high returns from the forced savings of its citizens (Asher 2004).
A 2007 paper by NTU economics professor David Reisman:
The GIC has been building up a stock of higher return, higher risk investments as a reserve against bad times. CPF savers have been doing national service by lending to the state at a low rate of interest. Assuming that the need for a buffer has to some extent been met, it might be possible to pay CPF savers a proportion of the gains. At least the CPF rate could be rebased on the equity rate, the bond rate, or some other long-term rate. The interest-rate charged by the three main local banks is a short-term rate. A long-term rate would pay CPF members a higher return. A higher return might keep them away from excessive house-purchase which could harm their ﬁnancial health.
The bubbles and the falls have shown that even houses are not “as safe as houses”. The real return on housing in Singapore averaged between 4 per cent (for ﬂats and condominiums) to 9.5 per cent (for resale HDB properties) in the 15 years between 1984 and 1999 (McCarthy et al., 2002, p. 221).While respectable, the CPF Board, which offers a risk free investment, does not have very far to go to offer a competitive rate of return.
Don’t Singaporeans ever feel patriotic that they’re doing a service to their country through their CPF contributions and tolerance of low returns?
CPF has to choose between serving members or state, between more market-based rates of return or be hamstrung by non-economic and political economy considerations. The residualist approach to the social safety net and welfare protection may have to change to a more proactive stance, as the globalisation backlash worsens. Beyond state welfare and social assistance in monetary terms, skills training and other ways of helping workers adjust competitively in the globalised new economy require mindset changes
On governance, CPF–ﬁscal process may be accountable, but not transparent. The co-mingling of funds for GIC of Singapore investment and non-accountability by individual funds is an open question. While it is inappropriate compare investment performance of a social security scheme with shorter-term private ﬁnancial investment, based solely on the annual net rate of return, CPF rates of return of 2.5 to 4 per cent may be re-examined. As capital markets become more developed with possible diversiﬁcation of investment to external markets, learning from US 401(k) and the MPF schemes ﬁrst requires a political decision on what role the CPF should play, serving state or members. The CPF –ﬁscal process is founded on an upright and honest government which intervenes for the collective good, deﬁned with some measure of paternalism. Pristine and trustworthy leadership cannot be taken for granted. Self-determination and choice call for more democratised checks-and-balance. Implicit taxation of returns rather than poor investment in the political context contributes to ‘asset rich, cash poor’ inadequacy.
Actuarial studies show that, on average, a person will require a monthly retirement income equal to at least two-thirds of his last-drawn monthly salary to maintain his standard of living.
The monthly annuity payments (ANN) and replacement rates (RR) corresponding to the monthly wage levels across the 3 age groups are shown in table 3. The 50-54 age group has the lowest average replacement rates of 30 percent (for a withdrawal rate of 45 percent) and 27 percent (for a withdrawal rate of 60 percent). The average rates for the 45-49 age group are 36 and 31 percent, respectively. Even in the best-case scenario, the average replacement rate is only 40 percent for the 40-44 age group with a withdrawal rate of 45 percent. The rate falls to 33 percent with a withdrawal rate of 60 percent. These rates are far from adequate when compared with the benchmark replacement rate. Thus, CPF saving by itself may not be adequate for retirement, even for those who have longer contribution histories.
2004 paper by NUS LKY School of Public Policy economics professor Mukul Asher:
To the extent the government holding companies earn higher than what is paid to the CPF members, implicit tax on CPF wealth occurs. IMF has estimated that the Singapore Government Investment Corporation (SGIC) earned about 10.0 percent per annum during the 1990s, substantially higher than the average nominal return of 3.4 percent credited by the CPF. The implicit tax for 2000 is (10.0-3.4=6.6) times $90.3 billion, or $5.96 billion, equivalent to 42 percent of contributions or 3.75 percent of GDP. The implicit tax is recurrent, and it is regressive as low-income individuals hold proportionally greater wealth in the form of CPF balances. At the minimum, the implicit tax is the difference between what is earned on insurance funds and the returns to members on their balance as shown in Figure 1.
In a 2000 paper, an earlier study by Prof Mukul Asher concluded that Malaysia’s EPF real returns rate had outperformed CPF in earlier years not covered in this post:
Between 1961 and 1999, there were only three years (1973, 1974, and 1981) when the real rate of return [of Malaysia’s EPF] was negative (Table 5). Compound Annual rate for the 1961-99 period was 3.37 percent; while for the 1987-99 period, the corresponding rate was 4.40 percent (Table 5). Thus, in spite of the 1985 recession and the 1997 economic crisis, Malaysia has been able to provide a satisfactory real rate of return to its members.
The real rate of return on CPF balances (nominal rate minus GDP deflator) averaged only 1.54 percent during the 1983-98 period; and only 0.07 percent for the 1997-98 period, the period when the floating rate was introduced (Table 6). The above rates are quite low, and therefore they negate the potential advantage of mandatory saving in financing retirement.
In recent years CPF’s inadequacy has becoming increasingly noticed even by the foreign press which often lavish generous (and un-deserving praise) on a retirement scheme whose only attraction is it’s highly unlikely to ever become insolvent. In the wake of the failed 2005 Republican attempt to reform and dismantle Social Security, journalists have studied and rejected the CPF model as a viable alternative to Social Security.
To mitigate this glaring shortfall in providing adequate real returns to its depositors (and without raising CPF interest rates), the Singapore government routinely runs pro-family filial piety campaigns (most recently this in 2010) in order to instil within Singaporeans a sense of responsibility (and guilt) while evading their responsibility to care for the elderly, many of whom cannot rely on their CPF alone to survive.
This problem is made much worse by the fact that the current generation of elderly workers were actively encouraged (and discriminated against in some cases if they choose not to comply) to bear fewer children back in the 1960s-1980s when the anti-natalist and quasi-eugenics campaign by the ruling party was in full swing. This apparent pushing of responsibility to the current generation of working adults to care for the elders is made most apparent in the introduction of the 1994 Maintainance of Parents Act which provides a legal means for parents to sue their children if they do not provide adequate financial upkeep for their parents if they are unable to subsist on their meagre post-retirement annuity and income.
In the years to come, it is foreseen that the rise of China and India in today’s globalised world would mean a faster pace of development which often brings with it higher inflation. If you thought that the first decade of the millenia was bad enough, subsequent ones are likely to be more inflationary. Leading liberal economist Krugman explains this point best in a post when he says that the sharp divergence of economic fortunes between advanced Western economies and developing economies mean most of the world inflation is disproportionately concentrated in developing countries which are reluctant to allow their currencies to rise for fear of jeopardizing economic growth. While Singapore clearly no say in the conduct of monetary policy of much larger foreign economies, its government has every means and tools at their legislative disposal to help its retired and retiring folk tide through a likely period of high inflation in the years to come. The only question is whether such efforts will go beyond half-hearted one-time electioneering budget handouts.