COEs: Nice, But No Cigar
Car ownership in Singapore
Singapore has one of the lowest car ownership rates, among advanced economies, worldwide. In 2022, there were only 149 cars per thousand inhabitants here, a rate less than half that of the UAE or Taiwan, a quarter that of the United Kingdom and Switzerland, and six times less than New Zealand and the United States.
The upshot of this is that our streets are among the least congested in the world. Compared to global cities such as London, Chicago, or Paris—where drivers spent more than a 130 hours in their cars sitting in traffic last year—local commuters only lost 26 hours to jams, and correspondingly less in terms of economic costs from lost time and petrol.
This is admirable for a global city, and one that we should not take for granted. This is evident for anyone that has had to endure rush hour in the major metropolitans of our neighbours, whether it be Bangkok, Kuala Lumpur, Jakarta, or Manila. And a major reason for this favourable anomaly is, undoubtedly, the system of electronic road pricing (ERP) and the vehicle quota system (VQS), of which certificates of entitlement (COEs) are the much-reviled offshoot.
COE prices hit a new high in October, following repeated new records set in prior months. The open category now changes hands at $152,000, more than five times the recommended retail price of a brand-new Honda Civic or Toyota Corolla sold elsewhere. This is one of the major drivers of ever-rising pressure on our nation’s cost of living, and COEs were implicated in the most recent increase in headline inflation.
Some may argue that cars in Singapore are a luxury, and that a car-light society is both better for the environment, and more than made up for by an excellent public transportation system. Surely the price of COEs are only the concern of an elite few?
But for middle-income families with younger children—as so many of my Sengkang residents are—a car has almost a necessity for getting the kids to and from childcare and school, cocurricular activities, enrichment classes, and weekend shopping trips. Similarly, those who are elderly or less mobile may well feel that a car in the family is less of a luxury, and more of a need.
And car ownership is more widespread than we may think; two in five households own one. And even for those who do not own a vehicle, the price of cars and COEs bleeds into our costs every time we hop onto a taxi or book a private hire car. The high prices of COEs for cars had even spilled over into that of motorcycles (at least until recently), where the relatively low costs of ownership used to be a refuge for those of more modest means wishing to have their own mode of transport.
A brief history of the COE system
The VQS and COEs were introduced in 1990. The objective was to restrain ownership, so that road traffic remains smooth.
Inherent in the policy is a decision to exchange certainty in congestion management, for uncertainty in prices faced by aspiring car owners, along with the possibility of inequity in access and usage. Put another way, the government consciously chose a quota system, knowing that it could lead to rising costs. If so, then it also follows that policymakers should step in to manage prices when they get too out of whack.
Since its introduction, the system has been repeatedly tweaked. These included minor adjustments—tenders became monthly instead of quarterly, for example—but others have been more consequential, such as a decision to go from closed to open bidding, to make COEs nontransferable, or the collapsing of private-car categories into the present two. But the government chose to lock in vehicle quotas—with some adjustment for growth—at the levels associated with new car registrations established in the years following the initial rollout of the VQS. This has led to significant variations in the total quota in different years; the difference in the highest and lowest supply years has been, on average, as large as 4,300 every month.
In 2018, the rate of growth of cars was frozen. This means that the stock of available COEs is now indefinitely fixed, and hence, the allocation of COEs is effectively zero-sum. If you are an aspiring car owner, but are priced out of the market, then too bad for you. If a well-off family can afford to buy multiple cars for each of their children, it will come at the expense of others who are less able to do so. This could be a delivery person who may need a vehicle for their work, a household with a sick or disabled member that needs a car for routine hospital appointments, or a family of five who would benefit from one ride to shuttle their kids to and from school. But the government has emphasized that—while it aims to keep COEs affordable—they were decidedly not a tool for enhancing progressivity.
This capsule history lesson is instructive, because it helps us understand not only the motivations behind why a COE system was rolled out, but also, why this history has unwittingly introduced shackles that inhibits the successful and equitable functioning of the system today.
The economics of COEs and the VQS
Perhaps the best way to think about COEs and the VQS is that it is a two-step system for controlling the vehicle population. The VQS first sets the total number of cars that may ply our roads. After this is established, bidding via the COE efficiently allocates the available quotas to potential car owners.
COE bidding requires that each bidder make offers, but only pay the price of the lowest successful offer. It was inspired by auction theory, and in particular, Nobel Prize-winning insights into how best to ensure that each bidder bids as much as their own private valuations will carry them. At the same time, no bidder has any incentive to deviate from the bids they put forward. The outcome will, in theory, not only raise the same revenue as an auction awarded to the top bidder, but also allocate these COEs efficiently.
Herein lies the wrinkle between theory and practice: while the COE system was modeled after this so-called second-price, sealed-bid auction, real-world considerations have meant a departure from this theoretical ideal. Bidding is now open, rather than sealed. There are multiple certificates to be issued, not just one. And perhaps most importantly, to make things easier for car purchasers, we now allow dealers to bid on the buyers’ behalf, for multiple COEs.
Recent efforts are insufficient
Sir, the government has made some efforts to address the recent spike in COE prices. They have tweaked supply, mainly by bringing forward confirmed 5-year COE deregistrations, and have increased bid deposits for motorcycles, while shortening the temporary validity period. But these have had limited effect, which is unsurprising, since the moves are one-off. Measures that appeared to contain prices in the motorcycle COE market are unlikely to work for car COEs, since dealer strategies for selling cars differ. Perhaps most importantly, the freed COEs still constitute only a drop in the bucket of the overall quota shortfall relative to high-quota years. Even with the cut-and-paste measures announced in November, as shared by SMS Chee Hong Tat yesterday, my calculations suggest that these adjustments inject about 10 percent more COEs into supply, on average within each category, per month, whereas a more decisive cut and paste is required to genuinely move the needle.
To be clear, the government is well aware of how relative quota differences are unequal across the years, and has been aware of this for a long time. In response to Workers’ Party MP Low Thia Khiang’s query, in 1998, about whether high COE prices were due to such quota imbalances, then-Minister of Communications Mah Bow Tan simply explained that they were fixed at the beginning of each quota year, conveniently sidestepping the issue of whether a more balanced distribution of quotas was called for.
Members of the public also recognize this. A letter in the Straits Times on Oct 14 called for managing COE demand and supply, not alternative, newfangled systems. Another online commenter to CAN claimed that they are being “penalized unnecessarily by a good system gone rogue.” Transportation experts also expressed their displeasure that “the COE system has not gone through a ‘more fundamental review’ to deal with… changes [since 33 years ago].”
And unlike Vegas, what happens in the car COE market doesn’t stay in the car COE market. Rising COE premiums affect the prices of used cars, of course. But they also affect the cost of motorcycles, which are frequently used not only for transport, but to make a living. High COEs also add fuel to other expenses, already on fire, ranging from private hire fees to motor repair.
Yet in response to a Parliamentary Questions (PQs) filed by the honourable member Yip Hong Weng in 2022, the Ministry of Transport stated that they would not be exercising any controls on private-hire car (PHC) market. The response to another PQ filed by my honourable friend, Louis Chua, likewise shunned the idea of restricting PHCs from the COE bidding process. Indeed, the Ministry has gone as far as to say that “LTA has no plans to review the COE rules.”
Potential solutions
One approach is to further segment the COE market in some way. This suggestion is not radical. Since the scheme was first introduced, there has been changes to the number of categories, before being continually refined to the present five. Today, there are routine calls to cater to separate, specialized categories of COEs. One way would be to drop the open category altogether, or further tweak the criteria for the other categories. One could also create a category specifically for vehicles driven primarily for commercial purposes, or—as I had suggested in my speech on the Electric Vehicles (EVs) Charging Bill late last year—for EVs.
However, one potential implication of introducing additional categories is that—depending on the distribution of demand among potential buyers—there could well be greater upward pressure on COE prices. Overall, the evidence suggests that categorization is often gamed, and as a result, would likely do little to improve equity considerations. It does not appear to offer a substantive way forward, in terms of arresting price increases, even if it may be used to fulfil other objectives.
The increased pervasiveness of private-hire cars—their number has grown almost fivefold over the past decade—has also been blamed for rising COE prices. One aggressive solution is to ban PHCs from the COE market entirely, managing their quotas separately. The Ministry of Transport has suggested that there is no evidence for this, and has even gone as far as to suggest that doing so could even drive up costs for consumers.
Even so, it makes sense to remove PHCs from regular Cat A and B bidding. Instead of banning such vehicles from the bidding process altogether, bidders for high-usage PHCs should do so in the Open category, but be allowed to pay the respective Cat A or B prices. This is currently the practice for taxicabs, and given how PHCs essentially replicate many functions of the taxi—insofar as road usage is concerned—standardizing the COE treatment for vehicles effectively used for commercial purposes still makes sense.
The practical question of how PHCs may be properly classified naturally arises. I do not think that—with ride-share data easily obtainable—that a data-savvy government such as this cannot successfully make a determination between “low-use PHCs”—that is, cars that are only occasionally operated as PHCs, not exceeding a certain threshold of rides a month—and “high-use PHC,” which are primarily operated as a rideshare fleet car. PHC drivers will be required to comply with their usage group, much like how weekend cars were restricted (they will be monitored by asking ride-hailing app companies to furnish these data). Only high-use PHCs would be required to bid in the Open category.
Another strategy is to alter the bid structure. One such suggestion—which has previously been rejected by the government—is to pay for exactly what one bids, instead of the lowest successful bid. Alternatively, bids could be set in terms of percentages of the open-market value of the vehicle—a so-called ad valorem system—which has been championed by some academics. Relatedly, categorizations could be altered, and pegged not to cubic capacity or engine horsepower, but market value. Proposals of this nature are reasonable, and have the added advantage of making the system more progressive. In my view, such proposals will make only a small dent on COE prices.
What would make the most impact is to smooth the vehicle quota supply, by transferring excess quotas from high-supply years to low-supply ones, such that quotas are broadly equal across future years. Importantly, this involves more than just the marginal transfers from expected deregistrations that are currently done today.
As mentioned earlier, the feast-and-famine nature of COE market has long been recognized as a problem. What has made this approach even more problematic is that, in low-supply years, high COE prices discourage deregistrations (since car buyers need to purchase new COEs at high prevailing prices), whereas in high-supply years, those who secured COEs at high prices have incentives to sell their COEs, pocket the prorated reimbursement value (while eating some depreciation), and reenter the COE market. Both practices exacerbate the cycle of supply imbalances.
Hence, the existing practice of reimbursing early-deregistered COEs at their the book value should end. We may even wish to reconsider transferability, since there is very little evidence that the practice increases speculation. But importantly, transferability should be allowed not to other car owners, but to other cars in the same category, for a given COE holder. This preserves flexibility for car owners to sell early, should their circumstances change, without introducing perverse incentives to game the system.
Thus far, I had focused mainly on supply-side proposals. What about demand? Hopefully, it should be clear that the enormous variation in COE prices has far more to do with (known and alterable) supply factors, rather than demand, which the data suggests is secondary. But perhaps more importantly, demand effects should be expected to moderate even further in the future, since increases in per capita incomes—which alter structural demand growth—should also be expected to moderate, given our status as a high-income country.
Still, two demand-side proposals merit consideration. First, dealers should no longer be allowed to bid. There is substantial evidence that speculators bid up COE prices. Because dealers have a greater willingness to pass on costs—or occasionally absorb loss—they could contribute to upward price pressure, as some have suggested. Others have argued that dealers have an incentive to keep COE bid low, since they sell cars and COEs in a bundle, and can pocket the difference if COE prices are low.
Regardless, the evidence suggests that dealer intermediation has likely resulted in greater concentration in the distributor industry, and such market power is seldom good for competitive prices. After all, price discovery has already improved substantially with open bidding, thereby diminishing some of the benefits of dealer hand-holding. In my view, the presence of distributors—despite the benefits of convenience of having a one-stop shop for purchasing a car—interferes enough with auction prices as to render their exclusion the lesser evil. Better for such distributors to offer advisory and guidance, for a fee, than be in a market-making position of collecting and making multiple bids.
Second, require second car purchases to be in the open category. The inequity of second (or third) car purchases has bothered many, and one proposal has been to apply some form of additional buyer duty on such purchases, under the premise that these purchases—however small—could still fuel demand and push up COE prices, as my honourable friend Gerald Giam has suggested. While the Ministry has declined to pursue such restrictions, one alternative—which admittedly would require some policing—would be to require households seeking a second or third car to do so in a distinct category, thereby limiting their price pressure on Cat A and B.
On motorcycle COEs
Sir, I have focused on solutions in the COE market for cars. Even so, I will reiterate what my honourable friend Faisal Manap had previously shared about our position on motorcycle COEs. The issues in this category are somewhat distinct, not least because many riders are lower-income, and require access for not just private transportation but to make a living. Progressivity should therefore be an even greater concern in this market. To this end, one is left to wonder why motorcycle growth is constrained in the same manner as cars, given their much smaller vehicular footprint.
To this end, we could introduce subcategories in Cat D by engine capacity, and to consider replacing the bidding system for Class 2B with a balloting one. Mr Faisal also stressed the need to allow motorcycle buyers to bid for COEs under their own name rather than a dealership, which is consistent with the point I made earlier about excluding dealers from the car COE market.
Conclusion
The recent cyclical high in COE prices has come at an inopportune time, and has contributed to already rising costs of living. The existing VQS-COE system is broken, and while the government has made tweaks to the system, these do not address the true underlying problem that has led to boom-bust cycles in COE prices. Let’s not ignore the most important driver of supply—the imbalanced monthly quota—and some of the most prominent sources of demand, from PHCs and wealthier multiple-car buyers. Let’s better manage the COE system, an important driver of rising costs of living, consistent with the goals of the motion, which I support.