Thursday, June 27, 2019

The Model Thinker #24 - Mechanism Design.

Image result for mechanism

This chapter focuses on designing political and economic institutions. I will shoehorn this chapter into a discussion about the relative strengths and weaknesses of our CPF system.

There are four aspects to such institutions.

Information consists of what participants know and what should be revealed to them. All Singaporeans know what interest rates they should expect to receive for their CPF accounts. Singaporeans are probably less clear about what kind of payouts they expect to receive from CPF Life when they hit 65.

Incentives cover the costs and benefits of taking some actions. Contributing your money via the Retirement Sum Topping-Up scheme leads to tax deductions that can benefit folks in the higher income brackets more than those in the lower brackets.

Aggregation show how individual actions translate into collective outcomes. The collective outcome of CPF is that Singapore insures itself from the bad financial habits of other Singaporeans and have a tool to keep inflation low compared to other OECD countries.

Computational costs measure the cognitive demand placed on participants. In this aspect, the complexity of CPF, such as the need to select from a menu of different CPF Life schemes,  makes it one of its biggest flaws.

No perfect mechanism exists, but when we design a mechanism, here are some aspects of the system that makes it good.

The lowest threshold is Pareto efficiency. There simply should not be a case where an alternative exists that everyone should prefer. Cynics will say that Singaporeans do not need CPF Life. But for folks who are very long lived, CPF Life will hedge against longevity risk. So CPF Life meets the requirements of Pareto efficiency. It may be much harder to argue that HK extradition laws are Pareto Efficient.

The next aspect is dominant strategy implementable. The best strategy of a participant should not depend on the actions of others. CPF Life is dominant strategy implementable. Unless your lifespan depends on how others select their CPF Life schemes, you can just choose the program that is best given what you know about how long you are likely to live. For me, I will elect for BRS plus the Basic CPF Life scheme since I am diabetic and not likely to face longevity risk. Our financial markets are not dominant strategy implementable. We can buy dividend stocks so long as yields do not get compressed too much by old uncles who need passive income to buy 4D.

The third aspect is voluntary participation. In this case, unlike many other countries CPF is involuntary. CPF is obviously not perfect.

The fourth aspect if budget balance. If a mechanism involves a transfer of resources, we cannot put in additional money and resources. I am inclined to argue that CPF does not meet this requirement because we have to pay for CPF IT systems, infrastructure and staff. However, if the money eventually gets invested well, the dividends, after paying off citizens for their 2.5%/4% may be able to sustain the CPF infrastructure, then it can be argued that this aspect is met.

Lastly, a mechanism is good if there is incentive compatibility. Participant's best choices should reflect their true preferences. There is simply no reason to game the CPF system by choosing a scheme that does not reflect your true preferences. Systems that do not have incentive compatibility include trading systems that enable high frequency trading.

So by reading this chapter, we get the impression that the two weaknesses of CPF is that it is involuntary and exacts a heavy cognitive load on Singaporeans. We should be simplifying it to make it easier to understand, maybe even merge SA and MA. A voluntary pension fund scheme where members could elect to take on some market risks on a global scale may also make it less involuntary.

No comments:

Post a Comment