Wednesday, 13 March 2013

Tweaking the COE scheme.

13 Mar 2013

Most people agree that there is something wrong with the COE scheme. If you ask them for specifics and dig deep, they are universally agreed on the problem of the COE scheme: they can't get one cheap.

Asked for a solution, and you will be amazed by their creative suggestions as to how the COE scheme should be tweaked in order to give them the best chance (or a sure chance) of getting a COE, and if they are particularly inventive, cheaply too.

So you can be sure if someone suggests that families with more than 3 children be granted a special exemption from COE or priority in getting a COE, you can be sure he (or she) has more than 3 children. Of if families with elderly or disabled should have special treatment, they have elderly or disabled in their families.

However you want to tweak the COE scheme, it should address 3 issues:

1) the COE scheme allocates COE based on ability to pay, not on the basis of need. How should it be tweaked to separate the wealthy car-collector who can afford any COE premium, no matter how high, from the person who just needs dedicated transport (whether he be disabled, or with a disabled or elderly family member, or many small children). One less than perfect option is to a) categorise COE by OMV, and b) reduce the COE quota for the high OMV category, while increasing the quota for the Low OMV category.

2) Singapore Drivers are using the vehicle more than most other drivers in larger countries (with more urban sprawl) because of the "sunk cost fallacy". Because of the high cost of COE, car owners feel the need to justify the high "investment" they have made (sunk costs), by using the vehicle as much as possible. This has been borne out by at least one study. Because of the high "sunk costs", pay-to-use measures like ERP is less effective, or even ineffective. What is $1 more when COE is $80k? To reduce the Sunk Cost Fallacy (a study found that high COE leads to more driving, and correspondingly, lower COE correlates to less driving), COE should be lowered.

3) The final issue is related to the second issue and that is the COE should be tweaked to re-emphasise a usage-based charging. One suggestion is to sell permits for blocks of distance (say 25,000 km), rather than a COE based on time (10 years). Here is another suggestion that addresses both the sunk cost effect, and tying the COE to distance driven:

The point is that COE was supposed to lower costs of car ownership, while we transited to a usage-based "tax" or charges (as the COE scheme was introduced vehicle tax were adjusted down from 140% to 100%).

There is another issue and that is, how will the current system transit into the new system. If we suddenly move to a "COE for a block of distance" instead of the current COE for 10 years, how does it affect current COE and vehicle owners? What is the Cat A and Cat B equivalent under the new "Distance COE" system?

I don't know. It may be an irrelevant problem since it is a suggestion or recommendation that has not been accepted. 


1) Categorising COE by OMV would not completely separate those who need a car and those who "collect" cars, but it would help somewhat, especially if the quota for the low OMV Category is sufficiently generous. I would shed no tears for the wealthy who may bid $200k for a COE to match their million dollar Ferraris. 

2) Reducing the Sunk Cost Effect/Fallacy, and changing the COE to a charge-by-use system. The problem with the purchase of a car is that the price of the car is also a significant "sunk cost". The proposal for a permit for a block distance ("distance permit") is one way, but it would not remove the sunk cost effect completely. (I explain below)

Another approach is to pay the COE as you use it ("Pay as you Use"). In this approach, people bid for COE as usual, but on successful bid, the "Winners" of the COE bid pays only a percentage of the bid, say 10%. So if you bid $100,000, you only pay $10,000 up front. This seems like a great bargain, right?

BUT, this allows you to drive the vehicle up to 500 km per month. For the next 500 km, you will need to pay 1% of the COE bid. So if you exceed the 500 km, you need to pay $1000 for the next 500 km. And after 1000 km, you need to pay 1% for every 100km.

[April 2016 comment: 500 km per month works out to about 17 km per day. If you stay in Toa Payoh and work at Plaza Singapura or it's vicinity, your daily return drive would take 15 km. If you live in Yishun and work in the CBD, you would cover 20 km both ways. If you do not use the car on weekends or use it sparingly, you could still stay within the 500 km allowance.]

So under the Pay as you Use, if you consistently drive only 500 km or less per month, your one-time COE payment of 10% is ALL you will ever pay. So in effect, you got a 90% discount. And the distance you travel is 6000km per year or 60,000 km in 10 years. Currently, Singapore Drivers drive on average 190,000 km per year. So your 10% COE allows you 30% of the distance. This is a good deal for the infrequent driver.

However, what if you drive between 501 and 1000 km every month? Then in addition to the 10% "Deposit", you would also have to pay 1% per month or 12% per year, or 120%  over 10 years, plus the initial 10% or 130%. The Original COE bid of $100k has become $130k.

The average Singapore Driver drives 19,000 km a year, or about 1,500 km per month. If you drive this distance, your monthly COE would be 6% (1% for 501 to 1000th, then 1% per 100 km for the next 500 km). In a year, you would pay 72%, and in 10 years, 730%. The 100k would become $730k.

[April 2016 comment: COE is currently about $45,000 - $50,000. Say $50,000. 10% would be $5,000. 730% of $50,000 over 10 years would mean $365,000 in COE. However, as buyers would only need to pay 10%, it is likely that the "ceiling" on COE would easily rise to $100,000 or more. ]

Which is the better approach?

The Distance Permit approach has a fatal flaw. Once you use up the block of distance you "bought" - 25,000 km - you cannot drive anymore. The car sits in the parking lot and it is an unusable "sunk cost". To "rescue" that sunk costs, drivers will buy a Distance Permit, and once they have a distance permit they will drive happily until they are down to the last 1000 km. Or even 500 km.

[The ST Sr Transport Correspondent suggest a block of 120,000 km. Given that the average distance driven by Singapore drivers is 19,000 km a year, the 120,000 km will last a little over 6 years. And this block given all at once will still lead to the sunk cost effect.]

The Pay-As-You-Use approach is somewhat better because it rations the distance on a month by month basis. So if you used up the 500 km this month, you have a choice, leave the car alone for a few days, or be prepared to pay a premium to "overdrive". If you choose not to "overdrive", the car is not a unusable sunk costs. In a few days (in the new month), you can drive it again, and recover your sunk costs.

You may initially choose to pay the "overdrive" premium because of the sunk cost effect, but subsequently, the costs will become significant and even unbearable. This is what we want. For drivers/car owners to actively ration and limit their driving. So if they drive furiously or extravagantly in the early part of the month, they would run out of mileage and have to stop driving in the last week or so. Subsequently, they may space out and be more judicious about their driving. 

Between the 2, the Distance Permit system is easier to implement. However, I suspect it would not work very well. Certainly, the distance permit cannot be put on a quota. Basically, every car owner MUST be able to buy a distance permit whenever and however often he wants. It's like ERP or a Road Tax. You can charge it, but you can't withhold it. 

The Pay-As-You-Use  is a tweak of the existing COE scheme and would be compatible with a revised COE category approach (by OMV). Motorist will bid for the COE as usual under the new categories (by OMV). When the results is announced, all those who got a COE will just pay the 10% (or whatever percentage). They then have to comply with the 500 km per month limit, or be aware of the COE premium that would be charged for the next 500 km and subsequent 100 km.

The problem with a monthly limit is recording the distance travelled. It would be more complex than the Distance Permit approach. But this may be resolved with ERP 2.0.

Another problem that applies to both approach is how to discount travel in Malaysia. An obvious answer is to record the odometer when the vehicle leaves and when it returns.

For the Pay as you use, cars that are in Malaysia for a long time or had been in the workshop for a while may not use up the 500 km. Drivers may ask for the distance to be "rolled over" to the next month.

There are other issues and loopholes to be addressed, but the problems are not insurmountable, and would in effect support the intent of minimising driving.

No comments: