**1. Introduction**

The total number of vehicles in China has increased dramatically over the past decades and reached 231.2 million in 2018 (excluding 9.06 million three-wheeled vehicles and low-speed trucks) [1]. The rapid growth in vehicle numbers has caused serious environmental and energy security problems. In 2015, nearly 700 million tons of carbon dioxide (CO2) was produced from China's road traffic, and these emissions are increasing [2]. China's dependency on foreign countries for oil reached 68.4% in 2017 [3]. Automotive gasoline and diesel accounted for 80% of all gasoline and diesel consumption in 2017. To reduce the fuel consumption from passenger cars, China has issued a series of fuel consumption rate (FCR) standards and regulations for passenger vehicles. China released its first FCR limit standard for passenger vehicles in 2004 [4]. The corporate average FCR standard for passenger vehicles was released in 2011 [5] and updated in 2014 [6]. China also set the FCR targets for new passenger vehicles to 5.0 L/100 km in 2020 [7] and 4.0 L/100 km in 2025.

Globally, the transport sector contributes about one-fourth of total fossil fuel greenhouse gas (GHG) emissions, about three-quarters of that amount come from road transport [8]. To curb the GHG emissions from road transport, ten countries and regions including China, U.S., EU, and Japan, etc., have established mandatory or voluntary standards for light-duty vehicles [9,10].

There are two pathways to achieve the increasingly stringent passenger vehicle FCR target for carmakers. One is to produce more new energy vehicles (NEVs), as the electricity consumption of NEVs is calculated as zero, and they have multipliers in the Phase IV standard. In the Phase IV standard [6], the equivalent energy consumption of battery electric vehicles (BEV), the electric-drive part of plug-in hybrid vehicles (PHEV) and fuel cell vehicles (FCV) are calculated as zero. The multiplier of NEVs is set at five from 2016 to 2017, decreasing to three from 2018 to 2019, and two in 2020. The other pathway for compliance is to improve the FCR of conventional vehicles by deploying fuel-efficient technologies or adjusting portfolios by producing more smaller and lighter vehicles. As there are still many barriers to the promotion of NEVs, such as high retail prices, short electric ranges and a shortage of charging infrastructure [11], improving the FCR of conventional vehicles has become one of the necessary paths for carmakers to comply with their FCR targets.

With the tightening passenger vehicle FCR targets and regulations, China achieved 1.7% FCR improvement annually for the period 2009–2017. The market penetration of fuel-efficient technologies is increasing rapidly. The adoption rate of gasoline direct injection (GDI) and turbocharging for new passenger vehicles in China reached 39.39% and 45.11% in 2017 [3], from 0.5% and 3.4% in 2009 [12], respectively.

Previous research found that the official tested FCR is highly correlated with curb weight, power, acceleration time, and other characteristics. By using the features of the U.S. car from 1975 to 2009, the research found that a 1% increase in weight results in a 0.69% increase in FCR, and a 1% reduction in 0–97 km/h acceleration time results in a 0.44% increase in FCR when holding all else equal [13]. Similar results were also found by using the technical specifications and fuel consumption information of automobiles for sale in Europe from 1975 to 2015 [14]. After reviewing relevant studies, the fuel-mass coefficients (the ratio of FCR change (%) and weight change (%)) was found to be in the range of 0.315−0.71 [15]. By evaluating a wide range of vehicle case studies of gasoline turbocharged cars, which represent the 2015 European market, the fuel reduction value coefficient (the ratio of FCR achieved through mass reduction to vehicle mass reduction) was found within the range of 0.159−0.237 L/100km\*100kg for the mass reduction only and 0.252−0.477L/100km\*100kg for the secondary effect [16]. The combination of life cycle assessment with the traditional design procedure was also proposed to assess the environmental performances of automotive component light weighting [17]. Fuel economy standards and regulations aim to improve the FCR of new passenger vehicles. In addition to FCR, consumers also pay attention to vehicle performance, etc., which are highly correlated with FCR. Thus, to analyze the feasibility of achieving the targets of 5.0 L/100 km in 2020 and 4.0 L/100 km in 2025 for new passenger vehicles in China, it is necessary to quantify the tradeoff between FCR and other attribute parameters of passenger vehicles in China.

Many studies have been conducted on the trend of passenger vehicle FCRs and the tradeoff between official tested FCR and other vehicle attributes. A new index called the Performance-Size-Fuel economy Index (PSFI) was proposed by An, which is defined as the product of the vehicle performance, size, and fuel economy [18]. The PSFI was used to assess the technical efficiency improvement rates of cars and trucks from 1977 to 2005 in the U.S. The PSFI showed good correlations and appeared quite linear for both cars and trucks by using the 1977 to 2005 data from the Fuel Economy Trends Report of the Environmental Protection Agency (EPA). The PSFI provides a way to measure the technological progress of vehicles, but it simply sets the coefficient values of the three variables to one, which requires further study and explanation. Compared with recent research results [13,14], the impact of curb weight on FCR could be exaggerated in the definition of PSFI. To better estimate the technological progress and the tradeoff between FCR and other attributes, Knittel built a regression model on fuel economy, weight, engine power, and torque, and also introduced the production possibilities frontier (PPF) to capture the technological progress. The result showed that if the power, torque and curb weight of the light-duty vehicle in the U.S. stayed at the same level as 1980, the fuel economy from 1980 to 2006 could have improved by nearly 50% for both passenger cars and light trucks [19]. Klier and Linn expanded Knittel's analysis by matching engine data to vehicle production data. The changes in the rate were examined, and the results showed that recent changes in the U.S. and European fuel economy standards had increased the speed of technology adoption [20]. MacKenzie and Heywood extended

Knittel's econometric approach by adopting both vehicle system attributes and consumer amenities as independent variables. They found that between 1975 and 2009, per-mile fuel consumption could have been reduced by approximately 70%, or an average of 3.4% per year, if not for reductions in acceleration time and the introduction of new attributes and functionality to vehicles [13]. To the best of our knowledge, this topic for China has not been carefully studied before due to a lack of sufficient data.

In this paper, we mainly aimed to address the following points:


The rest of the paper is organized as follows: The data used in this paper is discussed in Section 2. The methodology is presented in Section 3. The results are detailed in Section 4. The conclusion is presented in Section 5.
