Chinese megacities ration new car sales by capping the number of license plates they issue that permit driving within city limits. Concerns regarding the fairness of this policy have led city governments to use lotteries to allocate either all or a share of the license plates. Lotteries create gains from trade that have stimulated black markets for license plates in such cities as Beijing and Hangzhou. Though black markets are usually considered undesirable, they can correct for distortionary regulations. A recent literature evaluates the trade-offs between equity and efficiency that are inherent to a city′s choice of mechanism to allocate its license plates. In spite of the substantial estimated volume of trade in Beijing’s black market for license plates, most of the gains from trading license plates are lost to transaction costs.
Due to concerns about escalating congestion and pollution, Chinese megacities have adopted schemes to ration license plates, as a means to control the number of cars permitted to drive within the city limits. The choice of allocation mechanism for the license plate quota creates a tension between economic efficiency and fairness. Shanghai’s auction mechanism, in place since 2002, allocates the license plates to those who value them the most. However, the escalation of Shanghai auction prices, which are currently on par with the price of a new car, has led to concerns regarding fairness. Survey evidence shows that citizens disapprove of auctions (Li, 2018).Subsequently, Beijing adopted a lottery mechanism in 2011 to maintain a perception of equity (Huang & Wen, 2019; Li, 2018). Later, the cities of Guangzhou, Tianjin, Hangzhou, and Shenzhen chose hybrid mechanisms that allocate part of the plates by auction, and part by lottery. See Table 1 for an overview.
Table 1: Table based on Table 1 in Huang and Wen (2019). All prices in RMB.
These hybrid allocation mechanisms strike a balance between efficiency and equity in places where auctions are politically undesirable. However, lotteries create incentives among lottery winners and rationed car buyers to trade plates. Though license plates are not transferable, Daljord, Pouliot, Hu, and Xiao (2019) documents numerous news reports of the emergence in Beijing of a black market for license plates soon after the introduction of the lottery. Trade takes place on dedicated online sites and through intermediaries, such as car dealerships. Rentals seem to be the most common kind of transaction, but formal license plate ownership can reportedly be acquired through corrupt officials (e.g., New York Times, 2016)(see Note 1). News outlets also report black markets in other cities (see Note 2). Examples abound of rent-seeking behavior online, such as advertisements for license plates sales or rentals and for companies that broker rentals. Figure 1 shows an example from the website of the company atzuche.com, which acts as an intermediary in the black market. In spite of stringent enforcement of the requirement for a license plate to drive within the city limits, there is anecdotal evidence suggesting that there is more lenient enforcement regarding the transfer of license plates.
atzuche.com. “Is your idle license plate late expiring soon? A license
plate needs to be attached to an automobile within one year from
acquisition. Otherwise the license plate will expire. If you don’t use
the plate yourself, you can purchase a car model specified by AT and
then rent it to AT. While you manage to retain the plate, you can also
get a stable rental income.” Retrieved December 7, 2019.
Given the ease with which governments could foreclose black markets by permitting resale, the costs of prohibitive enforcement seem insufficient to explain the lax enforcement of black-market trade. Black markets may however serve an economic purpose: they can counter inefficiencies by reallocating license plates to the car buyers who value the license plates the most. The combination of a lottery and a black market serves effectively as a hybrid allocation mechanism that can correct some of the most severe misallocations while maintaining a sense of fairness, similar to the auction/lottery hybrids in cities like Tianjin and Guangzhou. By a Coase theorem argument, the black market can realize a fully efficient allocation in the absence of transaction costs: transactions take place until all gains from trade are exhausted. The fact that black markets are illegal, however, creates transaction costs, which include non-pecuniary costs, such as search costs and potential legal liabilities. Daljord et al. (2019) estimates that 7 to 28 percent of the potential gains from trade are lost to transaction costs in the Beijing black market.Due to their illegal nature, black market transactions are not directly observed, but can be inferred from from data on connected markets. Figure 2 shows the average prices of new cars in Beijing and the nearby cities of Tianjin and Shijiazhuang. The prices show a conspicuous, post-lottery upward shift in Beijing of about 20 percent, with no similar shift in Tianjin and Shijiazhuang. Following the rationing, the prices in Beijing have caught up with and, as of 2014, even exceed the prices in Shanghai. If the lottery randomly selected car buyers, if there was no change in the pricing of cars, and if preferences were stable, then one would expect the average price to follow the trends. Instead, the upward shift in prices in Beijing is consistent with a black market that reallocates the license plates to wealthier car buyers, who buy more expensive cars (Li, 2018; Xiao, Zhou, & Hu, 2017). Tianjin experienced a similar jump in prices to the one in Beijing, albeit at a later time. As Figure 2, shows, Tianjin’s jump in prices coincides with the introduction of a hybrid auction/lottery rationing mechanism in January 2014, which allocated 44 percent of the license plates by auction and by design selects a wealthier subset of car buyers. The comparable circumstances and magnitude of the jump in prices suggest that Beijing’s black market reallocated a material share of license plates to wealthier car buyers.
Other policies may have affected car sales. The Beijing subway system has expanded rapidly over two decades, from 39 stations in 2002 to more than 300 in 2014. In particular, a number of new lines opened in December 2010, when the rationing went into effect, and which expanded the network from 228 to 336 km (Xie, 2016). Public transportation appeals disproportionately to lower-income households, which may subsequently select out of the car market, leading to a shift in the composition of demand that is unrelated to the rationing. Since the subway expansion in 2010 coincided with the rationing, it is hard to directly test this alternative account in the data. There were however three other major subway expansions: in December 2012 (70 km), May 2013 (57 km), and December 2014 (62 km). Figure 2 shows no similar jumps in the Beijing prices during any of these expansions, which suggests the effect of the expansion in December 2010 was also modest.The price shift could also reflect a supply-side price response to the market contraction. However, a national policy of resale price maintenance precluded Chinese dealerships from adjusting their prices in response to the market contraction (Li, 2018). A comparison of same-model prices before and after the lottery confirms no change.
A simple and intuitive means to estimate black-market trade determines the smallest amount mass that must shift from the pre-lottery distribution to give the post-lottery distribution. This estimator detects trade when a license plate buyer purchases a different car than the license plate seller would have purchased, had the latter not sold her license. However, a transaction goes undetected if it is between a license plate buyer who purchases the same car as the license plate seller would have purchased, had she not sold the license plate. The estimator therefore generates a lower bound on the black market trade.A specific example is instructive. Suppose cars sell either at a price of 1 or a price of 2. Pre-lottery, 80 thousand cars sell: 60 thousand at the price of 1, and 20 thousand at price 2. The population sales distribution is . Post-lottery, the quota is set at 40 thousand cars. In that instance, 10 thousand sell at the price of 1, and 30 thousand at the price of 2, such that the population sales distribution is . In this case, at least 20 thousand license plates must be traded to rationalize the shift in the sales distribution.
Note 1: The former head of the Beijing Traffic Management Bureau News was sentenced to life in prison in 2015 for selling license plates.Note 2: See e.g. from Shenzhen, Hangzhou, and Guangzhou.
(Oystein Daljord, Chicago Booth School of Business, University of Chicago; Mandy Hu, CUHK, Department of Marketing; Guillaume Pouliot, Harris School of Public Policy, University of Chicago; Junji Xiao, Department of Economics, UTS Business School, University of Technology Sydney.)
Xie, L. (2016). Automobile usage and urban rail transit expansion: evidence from a natural experiment in Beijing, China. Environment and Development Economics 21 (5), 557–580.