European countries have been expressing deep environmental concerns for some time and now play a leading role worldwide in the fight against pollution. To achieve this purpose, the European Union (EU) has been implementing environmental policies to counteract the degradation of the ozone layer and to bring the green house effect to an end. The EU has established directives for its member states in order to restrain and diminish the emission of greenhouse gases (GHG), namely carbon dioxide (CO₂), chlorofluorocarbons, methane, nitric acid and ozone. Since CO₂ is the major GHG released into the atmosphere (98% in 2007 for the EU15), it is essential to reduce its emissions in order to work against global warming and climate change. Substantial CO₂ emissions originate in the transport sector (25% in 2007 for the EU15, excluding the international traffic departing from the EU) and almost all of this comes from road transporttation (93% in 2007 for the EU15). This large contribution makes this sector one of the largest polluters with respect to oil fuel combustion.

The road transport sector includes both motorcycles and automobiles. The latter consist of: 1) passenger cars (PC) (84.4% of the number of automobiles sold in 2007 for the EU15); 2) commercial vehicles (15.2%); and 3) buses and coaches (0.4%). Since PCs constitute the majority of automobiles on European roads, they play a crucial role in road transport CO₂ emissions. As a consequence, the EU decided to make voluntary agreements with the automobile manufacturers’ associations, the ACEA [1] JAMA [2] and KAMA [3], in order to promote the decrease of the average CO₂ emissions per km, by each new PC.

The literature regarding CO₂ emissions from PCs brings to the fore a vast normative perspective, but it suffers from scarce empirical support. This paper contributes to the empirical evidence, focusing on the drivers of road transport CO₂ emissions. Overall, the nature of drivers can be socio-economic, demographic, energetic, manufacturer or market. In particular, we work on the questions: 1) is dieselization actually reducing CO₂ emissions released by PCs? And 2) how does GDP per capita influence CO₂ emissions? The responses may define important policy measures to facilitate a reduction in road transport CO₂ emissions. For this purpose, we use a panel dataset for thirteen years (1995-2007) from the EU15 (except Greece). These countries belong to Europe, which has been in the front line of the reduction of road transport CO₂ emissions, and they are selected to fulfill the criteria of the longest time span with available data for drivers we control. In accordance with the common policies guidance from the EU, the econometric methods take into account the contemporaneous correlation.

We extend the literature on road transport CO₂ emissions by: 1) showing the relevant role of the drivers of new diesel PC registrations per 1000 inhabitants, and the average power of new diesel PCs registered; 2) shedding light on the debate of the pros and cons of dieselization; 3) discussing the importance of car sharing and the use of public transport in the reduction of CO₂ emissions; and 4) applying panel econometric techniques that cope well in the presence of common political guidance.

The paper is organized as follows: the second section consists of a literature review, the third presents the data and methodology used in this work. Section 4 provides the results obtained, the fifth section discusses those outcomes and section 6 concludes.

In a modern society, CO₂ emissions are generated by numerous sectors. Energy industries, manufacturing, construction, transport and other sectors, like comercial/institutional, residences, agriculture/forestry/fishery, all contribute to environmental damage. According to the source of CO₂ emissions, different literature is applied and several methodologies can be found. The literature on road transport CO₂ emissions, particularly from PCs, evolves according to two main perspectives: 1) the normative; and 2) the empirical. The normative focuses on the analysis of CO₂ emissions, considering both characteristics and fleet composition of PCs [4]. The empirical perspective includes several techniques, namely the decomposition analysis of CO₂ emissions [5], and the panel data approach [6]. The influence of the various vehicle characteristics on the changes in CO₂ emissions from PCs was analysed by [5], in Greece and Denmark between 1990 and 2005. In their turn, Ryan et al. [6] focused on the relationship among variables like fuel price, vehicle taxes, income and population density.

As noted by Stead [7], PCs using different fuel types release different amounts of CO₂. Indeed, the average diesel PC releases smaller quantities of CO₂ per km in comparison to the average gasoline car [8]. A diesel engine consumes 20% to 30% less fuel per km than a gasoline engine equivalent [9]. Nevertheless, while consuming 20% less, it only releases 9% fewer grams (g) of CO₂ per km than gasoline engines [10]. Apart from the fuel economy of diesel PCs, as Pock [11] pointed out, diesel cars have also been upgraded, namely in comfort and driveability, and their retail price is lower in relation to gasoline cars in most European countries. This has all contributed to the trend known in Europe as dieselizetion, which consisted of a sustained diesel market growth. On the one hand, authors such as Fontaras and Samaras [12] and Cuenot [13], connect dieselization to a reduction in CO₂ emissions, as a consequence of the increased fuel efficiency of diesel engines. On the other hand, recent literature minimizes the impact of this trend in reducing CO₂ emissions, because of the higher distance travelled by diesel PCs [14]. This phenomenon of longer trips taken by diesel PCs deserves further analysis.

Diesel PCs release inferior average CO₂ emissions per km than gasoline cars, when travelling the same distance. Nevertheless, as Schipper [14] points out, these type of vehicles in Europe travel 40 to 100% more than their gasoline counterparts, namely since most taxi drivers, salesmen and businessmen use them. For example, in 2005, in France diesel PCs were driven 64% further than gasoline ones and, in Germany, 80% more [15]. Also in Denmark, in 2007, Papagiannaki and Diakoulaki [5] mentioned that diesel cars travelled twice as far as gasoline PCs.

As stated earlier, the increasing demand for diesel is due to its lower retail price compared to gasoline in most European countries. This asymmetry is a consequence of the lower taxation applied to diesel, which results partly from the professional transport sector lobby, as noted by Pock [11]. Moreover, this author points out that, in the short run, higher fuel prices decrease vehicle use, while in the long run, they cause a reorganization of the PC fleet to more efficient gasoline cars and diesel ones. In the former case, this is true since diesel price and diesel PC ownership expenses are reasonably low. Therefore, in the long run, given the correlation between fuel consumption and road transport CO₂ emissions [6], as the former decreases, so CO₂ emissions diminish. Such fuel consumption reduction is directly caused, on the one hand, by fewer kilometers driven in the long run [16,17] and, on the other hand, by lower speeds on roads. In fact, fuel consumption diminishes as more drivers circulate at optimum speeds [18]. All these consequences of high fuel prices arise from its impact on families and individuals’ income.

When there are higher incomes, two opposite behaviors can arise. According to Storchmann [19], in the short run, individuals tend to drive more, increasing road transport CO₂ emissions. In contrast, over time buyers have greater opportunity to acquire powerful vehicles, but also better equipped with regard to fuel efficiency and technology [18]. Hamilton and Turton [20], when studying GHG emissions in OECD countries from 1982 to 1997, and Hatzigeorgiou et al. [21], when analysing CO₂ emissions in Greece between 1990 and 2002, pointed out GDP as the greatest contributor to CO₂ emissions. Nonetheless, Tapio et al. [22] noted that in the EU15 countries, from 1960 to 2000, GDP growth decoupled from energy use and, therefore, from CO₂ emissions. Another socio-economic factor affecting CO₂ emissions is population. Although it makes a positive contribution to road transport CO₂ emissions, its effect is not very noteworthy due to the small variations in population figures over time [5]. Nonetheless, it is worthwhile mentioning that increasing population density reduces the number of gasoline PC [6], favouring the use of diesel cars.

Another contributor to road transport CO₂ emissions is PC power, which is highly correlated with PC weight. Zervas [4] reported a rise in the average maximum power of both gasoline and diesel cars, from 1995 to 2003, as a result of the improved combustion efficiency. The increase in PC weight and power were in part a result of dieselization [10]. Diesel PCs have experienced a greater growth in power than gasoline ones. Since diesel PCs had to find more torque to increase their power/weight ratio in comparison to gasoline cars, they became more powerful. As a consequence, fuel consumption and CO₂ emissions also increased, counteracting the advance of technological standards in fuel efficiency. Regardless of the technical aspects, the last word about the average power of PCs, as Bonilla [18] points out, belongs to consumers, whose preferences when buying a new PC depend on their income.

The greater demand for diesel in Europe produces, however, a negative outcome in the whole CO₂ emissions, because it generates inefficiency in the entire fuel supply chain. Indeed, the adjustment of European refineries to the production of diesel causes an increase in CO₂ emissions due to higher energy loss. Exportation of gasoline and importation of diesel associated with the lower and higher demand, respectively, of the European PC fleet increases CO₂ emissions due to international transportation [23].

In the EU, most decisions aimed at reducing CO₂ emissions have a common guidance. To the best of our knowledge, the scarce empirical literature on road transport CO₂ emissions has not yet taken into account the possible existence of contemporaneous correlation between the EU countries as a result of the similar policies measures taken in all member states. To that extent, apart from the variables mostly suggested by literature (GDP per capita, population density, and gasoline price), we control for the effect of new diesel PC registrations and average power of new diesel PCs registered on road transport CO₂ emissions. The next section describes the data, method and estimation process.

In order to select the appropriate methodology that will give us a full understanding of the object on which we are focused, we must have a thorough understanding of the available data. In this section we present the data, their sources and main characteristics, as well as pursuing a discussion about the methodological choices.