WELFARE ANALYSIS OF THE VEHICLE QUOTA SYSTEM IN CHINA

• Author: Xiaolan Zhou,Wei‐Min Hu,Junji Xiao
• DOI: http://doi.org/10.1111/iere.12229
• Published date: 01 May 2017
• Date: 01 May 2017

INTERNATIONAL ECONOMIC REVIEW

Vol. 58, No. 2, May 2017

WELFARE ANALYSIS OF THE VEHICLE QUOTA SYSTEM IN CHINA∗

BYJUNJI XIAO,XIAOLAN ZHOU,AND WEI-MIN HU1

The Chinese University of Hong Kong, Hong Kong; East China Normal University, China,and

Shanghai University of Finance and Economics, China; National Chengchi University,Taiwan

This article presents a welfare analysis of the vehicle quota system of Shanghai, China. The empirical findings

suggest that the quota system leads to both welfare loss as a result of reduction in vehicle transactions and welfare

gain because of less externality of auto consumption. The net effect depends on the shadow price of the marginal

externality, the assumption of vehicle lifetime, and market conditions such as consumers’ intrinsic preference

for vehicles. Compared to a progressive tax system, the quota system is less effective in vehicle control but more

efficient in improving social welfare.

1. INTRODUCTION

The Chinese auto market has developed rapidly since the early 1990s and accelerated further

after China’s entry into the World Trade Organization in 2001. However, ancillary facilities,

such as parking lots2and road capacity were not ready to support such rapid development,

which has led to serious traffic congestion in cities such as Beijing and Shanghai. In addition,

the booming auto consumption has created some other serious problems such as air pollution

and energy shortages. To control the vehicle population and thereby resolve these problems,

the Chinese central government has applied tax policies such as a fuel tax and a consumption

tax, which have proven to be effective in controlling emissions (Xiao and Ju, 2014). Some local

governments have implemented more stringent policies. For example, Beijing applied an odd–

even license plate rule,3whereas Shanghai, the largest city by population in China, imposed a

vehicle quota system (VQS) and allocated the quota through auction.

The effectiveness of the VQS seems evident, but its effect has never actually been quantified.

In 2010, the number of vehicles in Shanghai was only one-third of that in Beijing, compared with

∗Manuscript received August 2013; revised November 2015.

1Xiaolan Zhou gratefully acknowledges the financial support of the National Natural Science Foundation of China

(grant #71603159), the Innovation Program of the Shanghai Municipal Education Commission (grant 14ZSO77), and

the support from NVIDIA Corporation with the donation of the Titan Black used for this research. We thank Shanjun

Li, Gautam Gowrisankaran, and Thomas Ross for their comments and suggestions. We are grateful to the editor and

three referees for their helpful comments. All remaining errors are ours. Please address correspondence to: Xiaolan

Zhou, Faculty of Economics and Management, East China Normal University, 3663 N. Zhongshan Road, Putuo District,

Shanghai 200062, China. Phone: (86)2162232025. Fax: (86)2154344955. E-mail: xiaolan.z.zhou@gmail.com.

2For example, Beijing had parking space for 1.3 million vehicles by the end of 2010, but the number of vehicles

owned by residents was about 5 million. (“Beijing’s Parking Problem,” China Daily, February 4, 2011, p. 5).

3The regulation was introduced on July 20, 2008, to ease congestion and reduce pollution during the Olympics and

Paralympics. It limited the city’s 3.3 million private cars to alternate days on the road from 7 am to 8 pm, according to

whether they had even- or odd-numbered license plates. Since this policy successfully improved Beijing’s air quality

and increased road space availability, the Beijing Traffic Management Bureau issued an end-number policy right after

the expiry of this regulation in September, 2008. Following this end-number policy, the automobiles in Beijing city

(inside the 5th Ring Road) are prohibited on public roads for one day per week based on the last digit of their license

plates. Davis (2008) finds that a similar vehicle control policy in Mexico was eventually ineffective in vehicle control

and improving air quality. His empirical evidence indicates that the restrictions led to an increase in the total number

of vehicles in circulation and changed the fleet composition toward high-emissions vehicles since consumers would

purchase more than one vehicle in response to the policy. This is also one of the reasons Beijing adopted the VQS in

January 2011.

617

C

(2017) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social

and Economic Research Association

618 XIAO,ZHOU,AND HU

TABLE 1

TIMETABLE FOR THE PASSENGER-VEHICLE CONTROL POLICIES IN MAJOR CITIES OF CHINA

Cities Time Policies

Shanghai January 2001 Auction

Beijing January 2011 Lottery

Guiyang July 2011 Lottery

Guangzhou July 2012 Auction and lottery

Tianjin January 2014 Auction and lottery

Hangzhou May 2014 Auction and lottery

Shenzhen December 2014 Auction and lottery

two-thirds in 1994 when the VQS was introduced. However, this reduction could be attributed

to other confounding factors, such as difference in aggregate demand, instead of the VQS.4

Previous studies on a similar VQS in Singapore also suggest that it is effective in vehicle control.

However, these studies cannot disentangle the effect of VQS from the changes in other factors

such as the economic situation (Seik, 1998) or conditions affecting purchasing decisions, which

may include the improvement of public transportation or increasing traffic congestion. So, the

impact of the VQS on vehicle control may not be that obvious.

It is imperative to identify and analyze the effect of the VQS since more and more Chinese

cities are adopting this policy.5If the quota system becomes a national policy, its effect, either

positive or negative, would be tremendous.

This article presents a welfare analysis of Shanghai’s VQS. Since the target of this policy is to

improve consumers’ welfare by reducing traffic congestion and air pollution, a welfare analysis

is valid for measuring the policy’s effectiveness. The welfare analysis consists of two parts:

(1) measuring the direct welfare effect related to vehicle transactions, including the consumer

welfare, firm profits, and government revenue; and (2) quantifying the externality of vehicle

consumption, which refers to the costs imposed on others that the vehicle consumers do not take

into account, including the cost of damage from carbon emission, air pollution, traffic accidents,

and congestion. The ambiguity of the welfare effect arises since the VQS could lead to both

welfare loss due to fewer vehicle transactions and welfare gain due to fewer externalities. This

article will quantify the net effect. Our empirical findings suggest that the welfare effect of VQS

eventually depends on the assumption of the vehicle lifetime, the market conditions such as

consumers’ intrinsic preference for vehicles, and marginal externality. Given Shanghai’s market

conditions in 2010 and the marginal externality estimated by Parry et al. (2014), with a vehicle

lifetime of 10 years the overall welfare effect of the VQS is a loss since savings in externality

are less than the direct welfare loss due to fewer transactions. However, the manufacturers

could strategically respond to the VQS by increasing vehicle prices, which would lead to more

reduction in vehicle sales and so make savings in externality exceed the direct welfare loss. In

contrast, with a vehicle lifetime of 15 years the VQS has an overall welfare gain since savings

in externality are more than the direct welfare loss. However, we may have overestimated

the VQS welfare gain if the parameters of the marginal externality are lower in China than

the estimates in the previous literature. Likewise, we may have underestimated the gain if the

parameters are higher. Because of these dual findings, we provide two sets of estimated shadow

prices of externality: The first equalizes the externalities to the changes in consumer surplus

whereas the second makes the social welfare in various scenarios unchanged from the null

4See Subsection 2.3 for detailed discussion.

5Shanghai, Beijing, Guiyang (capital city of Guizhou Province), Guangzhou (capital city of Guangdong Province),

Tianjin, Hangzhou (capital city of Zhejiang Province), and Shenzhen (major city of Guangdong Province with sub-

provincial administrative status) have adopted this policy to control the number of vehicles. These quota policies are

different from the Shanghai VQS in their quota allocation rules: Beijing and Guiyang use lottery, whereas Guangzhou,

Hangzhou, Shenzhen, and Tianjin use a combination of lottery and auction. Table 1 lists the time line and allocation

method for these policies.

WELFARE ANALYSIS OF VQS IN CHINA 619

scenario. The first set of shadow prices emphasizes the superior role of consumer welfare over

firm profits and government revenue in our welfare analysis, whereas the second set of shadow

prices assigns equal weight to all the components of welfare analysis. These shadow prices can

be used for contrast in various conditions to evaluate the policy’s effectiveness: If the realized

marginal externality were higher than the shadow prices in Shanghai in 2010, the VQS would

have increased social welfare, and if shadow prices were higher, the VQS would have decreased

social welfare.

Most previous studies about policy interventions in the auto industry have investigated the

policy influence on automobile externalities, such as air pollution (Dahl, 1979; Fullerton and

Gan, 2005; Bento et al., 2009; Feng et al., 2013; and Xiao and Ju, 2014, study various taxes,

whereas Crandall, 1992; Sterner et al., 1992; Koopman, 1995; Agras, 1999; and West, 2004, ana-

lyze the compulsory regulations such as the corporate average fuel economy), traffic congestion,

and accidents (Newbery, 1988, 1990; Parry, 2004) or combinations of these issues (Parry and

Small, 2005). This article conducts a more comprehensive welfare analysis of the VQS, includ-

ing the welfare loss due to fewer vehicle transactions, because such a welfare loss is always

the target of criticism by VQS opponents. The cost–benefit analysis we adopt is able to give a

complete policy evaluation.

There is a bulk of research on the welfare analysis of vehicle transaction. For example, Petrin

(2002) quantifies the benefits of introducing the minivan into the market, using counterfactual

analysis after estimating consumers’ preference over vehicles. Following this methodology, we

estimate the demand for vehicles by applying the revealed preference method proposed by

Berry et al. (1995, hereafter BLP) to the car registration data of the Chinese market; we then

use the estimates and the observations of Shanghai to compute the welfare loss due to the VQS

by examining the changes in consumer surplus, firm profits, and government revenues, among

scenarios with or without the VQS. The comparative statics between scenarios suggest that the

high license fee in Shanghai has decreased new car sales by at least 60%, so the license quota is

effective in controlling the vehicle population. But this reduction in transactions also resulted in

a transaction-related welfare loss of RMB 12.57 billion. Taking into account the manufacturers’

strategic response to the policy intervention, this welfare loss could be even larger.

This article uses the lifetime reduction in the externality of new vehicle sales—the ow

information of the vehicle fleet—to evaluate the welfare gain from the VQS, whereas, as Parry

et al. (2007) point out, most studies use the average cost or marginal cost of all active vehicles—

the stock information of the vehicle fleet—to evaluate the externality (e.g., Vickrey, 1963).

Our methodology is particularly efficient in evaluating the vehicle control policies designed

to decrease traffic flow by curbing new vehicle sales. We calculate the externality using the

exogenous parameter of marginal externality, the lifetime vehicle miles traveled, vehicle fuel

efficiency, and new vehicle sales. We adopt the exogenous parameter of marginal externality

per liter of gas consumption in China estimated by Parry et al. (2014). Our empirical findings

suggest that the VQS dramatically decreases the 10-year or 15-year lifetime externality from

extra auto consumption by RMB 11.25 billion or 16.88 billion compared with the scenario if

VQS were unavailable in 2010.

Using cost–benefit analysis, we also compare the VQS to an alternative policy—Hong Kong’s

first registration tax. The VQS imposes the same cost on all vehicles across various price ranges,

so it works in the same way as a regressive tax. The license premium forms a relatively smaller

percentage of the total price of a luxury car than of a smaller car and hence shifts the composition

of sales in favor of relatively more expensive items, which are usually less fuel efficient.6In

contrast, the Hong Kong tax system is more effective in vehicle control; moreover, it uses a

progressive scheme in price, which shifts the demand distribution to low-end vehicles that usually

consume less fuel and hence further reduces the emissions. The tax system also is favorable

to low-income consumers since the low-end vehicles are now at lower costs and therefore are

6Luxury vehicles usually have larger weights and displacement for safety or better performance such as acceleration.

These features, however, are negatively correlated with fuel efficiency (Klier and Linn, 2012).