انعطاف پذیری تخصیص و بهره وری قیمت در نظام سهمیه بندی خودرو در سنگاپور

Permit or license plate quotas are highly effective albeit controversial policy tools for managing growth in the vehicle population and thereby, adverse traffic congestion. A judicious distribution of the scarce permits that targets the dual objectives of price efficiency and social diversity in vehicle ownership can however mitigate the controversy. This paper scrutinizes the attainment levels of these two objectives within Singapore’s multi-categorical Vehicle Quota System in two time periods (1991–1998 and 2002–2011) which differ in the number of permit allocation categories, the auction format used (sealed versus open bids) and in the frequency of distribution (monthly versus semimonthly). The lessons derived are contrasted to the other two jurisdictions which have also implemented quotas on their vehicle registrations namely, Shanghai and Beijing.

South-east Asian city-state Singapore and Chinese metropolises, Shanghai and Beijing, currently use the frontline tool of permit quotas to manage vehicle ownership and thereby, the adverse effects of traffic congestion within their jurisdictions. Beijing, which joined this select club in January 2011, was pertinently ranked among the three most stressed cities in the IBM’s Global Commuter Pain Index constructed from surveys carried out in 2010 and 2011 (IBM, 2010). 1.1. Singapore Singapore implemented the world’s first Vehicle Quota System (VQS) in May 1990 after systematic raises of vehicle and gasoline taxes since the 1970s were deemed insufficient towards curbing vehicle ownership aspirations – see for example, Phang (1993). Another complementary innovation implemented in 1998 was Electronic Road Pricing (ERP) which replaced the manual cordon-style Area Licensing Scheme (ALS), introduced in 1975 to charge for entry in the Central Business District, and the Road Pricing Scheme started in 1995 for expressways utilization (Chin (2002) and Olszewski and Xie (2005)). The ALS is recognized as the first ever successful implementation of congestion-pricing worldwide. The VQS is deployed to the public via regular uniform-price auctions of quotas of Certificates of Entitlement (COE). These are permits required in the registration process as they entitle vehicles to run on the local roads for up to 10 years. COEs are offered in distinct categories catering to cars and taxis, commercial vehicles and motorcycles as well as in an unrestricted or “open” category whose objective is to avail flexibility in the mix of vehicles. Car categories are demarcated by engine capacity to promote social equity and progressivity in car ownership. The original four car categories (1000 cc or less, 1001–1600 cc, 1601–2000 cc and above 2000 cc) have been merged into two since May 1999 with the cut-off set at 1601 cc. A COE obtained in the “open” category is transferable and it can be used to register any vehicle. COEs in the other categories are non-transferable except for those within the commercial vehicles category. The evolution of the COE scheme is traced in several articles, e.g. Chin and Smith, 1997 and Barter, 2005, etc. Recently, Chu (2011) contrasted the COE premiums in the “open” category following the auction format change from sealed to open bids in July 2001 and inferred an average price reduction of about 16% ceteris paribus as well as a more pronounced manifestation of demand and supply forces. 1.2. Shanghai Beijing and Shanghai have opted for simpler methods for allocating their quotas of vehicle permits (or license plates, as they are more commonly known there). Shanghai has used an auction system since 1986 (Song and Zhou, 2010). The latest reform pertaining to private cars took place in January 2008 with a shift in format from sealed to open bidding (Press Conference of the Shanghai Municipal Government, 2008). The 2-rounds auction, accessible over both Internet and telephone channels, takes place monthly on a designated Saturday. The first round of the auction typically takes place between 10 and 11 am. Its objectives are to register the initial bid from each participant and thereby, to determine a minimum bidding price for the second round. During the latter’s 30 min duration, each participant is allowed up to two bid revisions which are constrained to be within ±RMB3001 of the evolving price. If the current clearing price is say RMB30,000, a revised bid can then only be between RMB29,700 and RMB30,300. Successful bidders pay the final amount they bid. Interestingly, the winner’s curse effect is mitigated as the range constraint on the revised bids positions the auction closer to the uniform price. In June 2012, the auction attracted 24,774 bidders for the 9500 license plates on offer. The minimum and mean successful bids were respectively RMB55,800 and RMB58,2272 (Bloomberg News, 16 June 2012). In a rare published analysis of the Shanghai auction data, Song and Zhou (2010) infer that the auction format change from sealed to open bids in January 2008 has led to an average price reduction of RMB3020 and higher price efficiency, as inferred through higher prediction errors. As a result of the ownership restraint, there were only about 1.03 million private cars registered in Shanghai at the end of 2010 (Shanghai Daily, 5 April 2011). This is a remarkable feat as Shanghai is widely regarded as China’s iconic business hub and its population size of 23 million inhabitants is comparable to the capital city Beijing. The latter is discussed next. 1.3. Beijing Beijing issued some 700,000 license plates during 2010 and ended the year with a park of about 4.8 million vehicles. Compared to its vehicle population of some 1 million in 1997, this implies an average yearly growth of about 13% since then. To counter this unsustainable growth rate, Beijing introduced in January 2011 a monthly quota of 20,000 new license plates catering to individuals (17,600), business (400) and government bodies (2000). Allocation is effected via an egalitarian ballot or lottery system (see e.g. The New York Times, 2 January 2011). A possible rationale for the adoption of balloting in the Chinese capital city is the questionable legality of the auction system in Shanghai as it ran counter to national policies pertaining to road usage and the development of an automotive industry (Song and Zhou, 2010). The very first computerized ballot draw carried out on live TV attracted 187,420 eligible candidates for the 17,600 private vehicle license plates in January 2011. By March 2012, the number of participants had swollen to 990,806 (amounting to a chance of success of about 1 in 50) due to the automatic re-eligibility of the unlucky ones from previous draws. This number however dropped by about 120,000 in April 2012 when long-time unlucky applicants were asked to renew their applications (China Daily USA, 10 April 2012). The competition for a license plate remains way above that observed in the auctions held in Singapore and Shanghai. It highlights the inefficiency of a lucky dip allocation system which tends to attract an inflated demand coming from those who either have no immediate need for a vehicle or who are just helping out friends or relatives. 1.4. Research agenda: allocation flexibility and price efficiency Economists define efficiency as the allocation of resources to their highest valued use. A pragmatic way to achieve this goal is through auctions where scarce resources like license plates are allocated to those who value them the most (e.g. Vickrey, 1961 and Klemperer, 1999, etc.). In comparison, lotteries (or queuing) engender “… inefficiency … in the fact that those who draw the goods … may not be the ones who value the goods the highest …” (Boyce, 1994). This appears to be the case with the allocation of license plates via the lottery system in Beijing, as reported in the previous sub-section. Singapore and Shanghai demarcate from Beijing in their adoption of the auction mechanism to allocate license plates. Whereas Shanghai auctions off license plates in a single category, Singapore uses several categories as well as an unrestricted “open” category for flexibility purposes. Singapore’s multi-category strategy aims to promote ownership diversity compared to Shanghai where potentially the richest could dominate the single category license plate auctions and thereby aspire to vehicle ownership. The research agenda in this paper centers on the allocation flexibility and the price efficiency attained across the diverse categories in Singapore. Allocation flexibility is ascertained by analyzing the engine size profile of the car population over time. Price efficiency is gauged via the relationships among the COE premiums from different categories. Of specific interest are the “open” category COEs: in which categories are they converted and how are their premiums inter-related? For example, suppose the “open” category COEs are used mostly to register cars with higher engine sizes. Do the historical data demonstrate significant price correlations between the COE prices in these categories and in the “open” category? Such a consistency in COE premiums is desirable from the perspective of price efficiency. A closely related contribution is Koh and Lee (1994) who address social equity and economic efficiency within their derivation of optimal bidding strategies for transferable and nontransferable COEs. They support their theoretical predictions with data from the first 29 auctions held between May 1990 and October 1992. The data analysis carried out in this present paper extends to a much longer 21-year period starting 1992 and ending 2011. This paper pursues its journey as follows. The next section provides a chronology of critical changes in Singapore’s COE scheme prior to addressing the issue of diversity in car ownership. Section 3 compares the annual number of COEs allocated and the engine size profiles to infer in which categories the “open” category COEs were converted. Following unit root and Granger causality tests, regression models are then developed to assess how COE premiums in the “open” category are linked to those in the other categories. Analyses are carried out separately for the 1991–1998 and 2002–2011 time periods which have different number of car categories, bidding format (sealed versus open bids) and auction frequency (monthly versus semimonthly). Final remarks are posted in Section 4.

The analyses carried out in this paper have demonstrated that Singapore’s multi-categorical COE scheme has attained over its 21 years existence a high degree of price efficiency across its different categories. Car ownership diversity was also shown to correlate with rising affluence. These have been achieved thanks to a policy planning process that made several good decision calls when implementing the VQS. This includes the allocation of the Certificates of Entitlement (COEs) in specific vehicle categories as well as an “open” category to cater for flexibility in the vehicle mix, the shift in the uniform price auction format from sealed to open bids which has improved transparency and mitigated price volatility (Chu, 2011) and last but not least, the decision to allow vehicle distributors to bid for COEs on behalf of their clients. This is a key difference between Singapore and Shanghai or Beijing where applications for vehicle permits are mostly made by individual buyers. Vehicle distributors account for almost all the bids for COEs as they offer a convenient service bundle which comprises an attractive credit financing of both the vehicle purchase and the COE premium7 as well as the administrative chore of placing a bid on behalf of the vehicle buyer and if needed, update the bid amount – see for example Koh (2003) for details. While accumulating orders ahead of each auction, they are in a better position to gauge overall demand for COEs and therefore, they are more informed relative to individual bidders under both sealed and open bids auction formats. This is a likely explanation for the remarkable price efficiencies noted in this paper across different COE categories, especially amidst the opacity of sealed bids between 1991 and 1998. It is quite unlikely that Singapore will ever resort to the auction of COEs in a single car category, as in Shanghai. This is demonstrated by two official announcements. In January 2011, the Transport Ministry rejected a request by the Motor Traders Association for the elimination of the “open” CatE. It highlighted the flexibility that category played in meeting vehicle demand changes. This is supported by Fig. 1, Fig. 6 and Fig. 7 in this paper which demonstrate how the “open” COE conversions have aligned with increasing affluence. With the trend towards green vehicles with smaller engines, the “open” Cat7 COEs can likewise play a useful role to meet such changing demand in the future. Also in 1999, the possibility of a single car category with the COE premium scaled according to either the engine capacity or import value of the car (also suggested in Koh and Lee (1994)) was rejected by the VQS Review Committee as it opined (p15, items 23–25) that it would be hard to calibrate a scaling mechanism and furthermore that would portray the VQS as a taxation measure rather than an allocator of scarce COEs. The rationing of license plates in Singapore, as well as in Shanghai and Beijing, are worth following by policy makers as more cities (e.g. Mexico City ranked at the top of IBM’s Commuter Pain surveys in both 2010 and 2011) become entangled in road congestion. A recent publication by Muthukrishnan (2010) demonstrates interestingly that a shift from fixed to usage charges for vehicles may not be welfare-improving. When the construct was tested on Singapore data, the same results were obtained when fixed charges (i.e. COEs) were reduced and variable charges (e.g. ERP) increased. This finding provides support to the effectiveness of license plate auctions which extract fixed charges from vehicle buyers.