CPF Reforms: Why Government's Critics Mostly Miss the Point

The major piece of CPF reform was supposed to be about the Compulsory Annuity Scheme (my views here and here). Nevertheless, the relationship between Government investment vehicles (namely GIC and Temasek) and CPF has received much of the attention instead. I was absolutely tickled with Chua Mui Hoong's quip, "The more you know, the more you realise you don't know."

Indeed, we do not have complete knowledge on how the government channels savings into the various investment vehicles. But the fundamental dissatisfaction with the CPF system is simple - the low returns relative to what the government reports on its investments. The bugbear is therefore the difference between the high investment returns of the government's investment vehicles (reported as 9.5 per cent for GIC and 18 per cent for Temasek) and low CPF interest rates.

The Famous Bank Analogy

When pressed on the difference, the Manpower Minister deflected the questions with the bank analogy - that depositors enjoy a fixed rate and cannot go to the bank to ask for higher returns. Not surprisingly, many bloggers disagree with this line of argument and continue to argue for CPF rates to be in some manner pegged to GIC or Temasek returns. Since some of these calls come from politicians, I shall refrain from making any explicit links here. My point is totally apolitical.

Why Returns are Not Low

My concern is strictly that of an economist. I maintain my stated view that CPF returns are not low on several accounts. First, it is risk-free, a position also reiterated by the government. Second, young Singaporeans also borrow at this preferential rate (0.1 percentage points above) to fund their housing purchase. Borrowers do benefit from this same cheap rate. Third, there are also other investment possibilities through CPFIS if a member is willing to tolerate higher risks and takes responsibilities for his own financial decisions. Fourth, property prices have seen a long run secular appreciation way in excess of 2.5 per cent. Those who borrowed at the cheap preferential rates would have made good excess returns, in addition to not having to pay rent.

Is a GIC Peg the Solution?

Let's just assume that there is one government investment vehicle, GIC. Let me then move on to address the pegging of CPF to GIC returns: Is this a good idea? Again, I am only concerned about the optimality of the institutional arrangement between CPF and GIC. I am therefore not addressing the socio-economic or even political consequence of the current relationship between GIC and CPF.

Let's take at face value that CPF board purchases government bonds that pay a 4 per cent return, risk-free, year on year. The government then, directly or through various channels, invests the proceeds which over the long run get 9 per cent returns. Is there an institutional arrangement that can somehow transfer these returns to the millions of CPF account holders?

A Principal-Agent Problem

The difficulty here lies in the principal-agent problem. Consider a simple example. The boss (principal) pays the worker a fixed wage for the worker (agent) to put in effort to maximise profits for the company, but the boss does not observe the amount of effort which the worker puts in. The optimal response of the worker is always to shirk since the boss cannot tell. Because of the asymmetry of information, this contractual arrangement results in suboptimal outcomes. The worker will never put in effort for the company.

The only way to overcome this is to make the worker the residual claimant or the profit owner. That is, the boss collects a fixed rent and leaves the rest of the profits to the worker. Since the worker now gains the benefits, he puts in effort and this arrangement results in optimal outcomes. The economic principle is simple: whoever controls the amount of unobservable effort must be made the final claimant of the rewards that come from this effort.

GIC and CPF: An Optimal Arrangement?

The institutional arrangement between CPF and GIC is therefore an optimal one. Here, the CPF Board, which is the custodian of people's savings, can be thought of as the principal. GIC, which invests the money, is the agent.

The optimal arrangement is therefore one that CPF receives a lower fixed amount while the GIC receives the higher excess. This point is further reinforced by the fact that the principal in this case is risk-averse. It therefore makes most economic sense for risk-averse citizens to get a fixed rate while allowing the government to bear the risk. To extend the argument a little, can CPF board demand a high fixed return, say 7 per cent, and leave GIC with the rest? Think about a real life principal-agent problem. If you pay the CEO with stock options that have very high strike price, the CEO simply takes on huge risks in order to have any chance of seeing the money, again to the detriment of the company. The principal-agent problem strikes again!

To peg CPF returns to GIC is therefore bad on several counts. First, it transfers investment risks to risk-averse citizens. Second, it potentially makes GIC behave sub-optimally. Worse, since investment assets are 'fungible' and returns not fully observable, it creates incentives for false-reporting. The agent will find ways to realise losses, postpone gains, and cook the accounts to reduce the payout to the principal.

In the end, the point is a simple one. The only way to make government watch over the country's money as best as it can is to let it behave as if it owns the money! Any other institutional arrangement will not achieve this optimality. Rather than increasing returns to CPF account holders, a GIC peg may become an institutional nightmare that creates more problems than it solves. The argument will run on.