Rail and public transports are usually deemed essential features for a well-functioning mobility system. How could we get around in our cities without buses, trams and metros? Wouldn’t there be huge traffic jams? Isn’t rail “a key contributor to the EU’s 2050 climate neutrality goal”? 
And how could we have public transport without public investments and subsidies?
Today, governments spend a lot of money to fund public transportation, including railway systems. According to a recent analysis (Schäfer and Götz, 2018), between 2001 and 2015 public contributions to railway companies in France, Great Britain, Germany, Italy, Spain and Sweden amounted to some 659 billion euro. This means that public expenditure for railways in the EU exceeds seventy billion Euros per year. The figure for local public transport in Italy, France and Germany is about thirty billion per year. Thus, we can estimate that in the whole EU transfers are of the same magnitude of those for rail.
At the same time, the tax revenues from motor vehicles exceed 400 billion euro per year or three to four times public expenditures for road construction and maintenance.
Cars and coaches currently count for 85% of passenger transport, while trucks cover 95% of freight. Thus, the EU countries have failed to move transportation from roads to other modes, rail in particular. Yet, air pollution and accidents have decreased dramatically. Today’s vehicle emissions are less than a tenth of what they were 40 years ago and so is the mortality rate (number of people killed per kilometer driven). Noise has also decreased, and fuel consumption and CO2 emissions per kilometer driven have been roughly halved while remaining unchanged overall. In other words, subsidising public transportation has had and has very limited benefits on the environment.
What about congestion? As mentioned above, public transportation plays a limited role on a national scale. However, it plays a significant role in the main urban areas. Since subsidies typically cover about two-thirds of costs, their elimination would certainly provoke a shift to cars and an increase in congestion. Is that unavoidable? Let us consider that most of the road network and all the urban streets are owned by the state or local authorities. With some exceptions, congestion is due to the fact that there is no price signal that ensures an efficient allocation of the scarce space. The solution, therefore, is charging access, rather than public transportation. London had an excellent network of buses and subways, but average speed of cars was just 14 km/h. It rose only after the introduction of the congestion charge to access the city centre.
Subsidised public transportation is often deemed desirable since it helps the poor. Yet, most trips made by low-income people are currently by car, not by rail or public transport.
A recent analysis by ISFORT shows that in poor Italian urban areas (where average income per capita is below € 15,000) the share of trips by bus/tramway is just 6%, while it is 16% in the wealthier (and more densely populated) areas. In fact, the key variable that explains the share of public transportation is not income, but geography.
Would it be possible to run market-based collective transportation services in the urban areas? The answer is generally positive. For example, in the UK (outside London) public transport has been deregulated in the 1980s and 90% of the vehicle miles of local buses are supplied without direct subsidies. The local authorities only ensure that people with a disability or older than 60 years get free tickets. After deregulation, the cost per vehicle kilometre on local buses outside London dropped by 50%.
To compare, in 2019 the unit cost was 228 pence within the UK metropolitan areas, about half of the average cost in Continental Europe. This means that in the UK, a bus with on average ten people on board breaks even with a revenue per passenger-km of less than 0.3 €. A typical urban trip, five kilometres long, would then cost 1.5€.
In the medium-term, robotaxis could provide affordable mobility in the less densely populated area. According to Elon Musk, Tesla could charge $0.40 per mile (or 0.25 €) for this service.
What about railways? Would they disappear if they were no longer subsidised? No, they would not.
In the UK, just a few years after the privatization of the railways in the ‘90s, the network was re-nationalized and received several billion pounds of public funds per year. Services – i.e. cars and locomotives — remained in private hands; subsidies were progressively decreased while the number of passengers grew more than in any other European country (except Sweden) in the same period. Entirely (network and services) private railways for freight do exist today in the United States. In the same vein, revenues from passengers repaid capital and operational expenditures for high-speed lines between Paris and Lyon and between Tokyo and Osaka and, probably, for the Rome-Florence connection.
Changes in cost efficiency should also be taken into account. As the cost of infrastructure costs and part of those for the services today are paid by the taxpayers, there’s no strong incentive to minimize them. A private company, on the contrary, would do everything possible to do it.
The space sector offers an interesting example of a situation in which private investors brought about a drastic increase in efficiency. Elon Musk’s SpaceX reduced the cost of launching payloads into space by 80-85%, compared to NASA’s traditional spaceflight programs before the era of reusable rockets.
To sum up, today about 90% of the trips are paid for by users and taxpayers, with amounts that exceed by far public expenditures for the road networks. In brief, the efforts to shift demand from cars (and trucks) to train and public transport failed. Moreover, the distributive effects are somewhat paradoxical: some low-income people are advantaged, but most of them keep using car transportation and are heavily taxed. A system without coercive transfers and where everyone pays for its direct cost and for the externalities generated would indeed be feasible and less expensive.
