Statistics Explained

Railway safety statistics in the EU


Data from January 2022.

Planned article update: January 2023

Highlights

In 2020, there were 1 331 significant railway accidents in the EU, with a total of 687 persons killed and 468 seriously injured.

There were 898 fewer significant railway accidents in 2020 than in 2010, a reduction of 40%.

In 2020, more than half of fatalities from railway accidents in the EU involved unauthorised persons on the tracks (60%) and almost one third occurred at level crossings (31%).

Railway accidents, EU, 2010-2020
Source: Eurostat (tran_sf_railac)

In 2020, 1 331 significant railway accidents were reported in the EU. A total of 687 persons were killed in these accidents, while another 468 persons were seriously injured. At EU level, the number of fatalities in railway accidents decreased gradually over the last decade, from 1 245 in 2010 to 687 in 2020, a fall by 45 %. However, it should be noted that from 2019 to 2020, the decreases in railway accidents, fatalities and seriously injured persons coincided with a sharp drop in passenger transport by rail caused by the Covid-19 pandemic. The large increase in remote working and home schooling, combined with recommendations to avoid unnecessary travel during the pandemic, contributed to a rail passenger transport almost halving in the EU – see the article on railway passenger transport statistics for more details.

Suicides occurring on railways are reported separately. With 2 204 reported cases in 2020, suicides outnumber the victims accounted for by railway accidents.

Eurostat publishes data collected by the European Union Agency for Railways (ERA) in Eurobase. This information is also published by ERA. It should be noted that there are no railways in the EU Member States Cyprus and Malta, nor in the EEA/EFTA country Iceland.

Full article

Continued fall in the number of railway accidents

12 % fewer railway accidents in the EU in 2020 compared with 2019.

The number of significant railway accidents in the EU fell almost continuously between 2010 and 2020, with the only exceptions a sharp increase in 2014 and a slight increase in 2017. In 2020, the number of significant accidents decreased by 184 compared with 2019, to a total of 1 331 accidents (-12 %). Railway safety has generally improved in the EU, with 898 fewer accidents in 2020 compared with 2010, a reduction of 40 %. Following a sharp increase in 2014 (+7 %), the decrease compared with the previous year was particularly marked in 2015 (-13 %). The decrease in accidents from 2019 to 2020 did concern all accident categories except collisions and the ‘other’ category. The number of fires in rolling stock fell by -41 % to 10 accidents in 2020, level crossing accidents by -19 % to 350, accidents to persons by rolling stock in motion (excl. suicides) by -14 % to 685 and derailments -5 % to 69. In contrast, the number of collisions increased by 7 % to 110 accidents, and ‘other significant railway accidents’ by +11 % to 107.

Accident figures for EU countries are comparable from 2010 onwards, following the implementation of common definitions across all Member States. Prior to 2010, Belgium, Poland and Slovakia generally reported all railway accidents instead of only significant accidents. As a result, there was a lower number of accidents in several categories from 2010 onwards compared with previous years.

Looking at the detailed 2020 figures on significant railway accidents (Figure 1), the largest category at EU level was accidents to persons caused by rolling stock in motion, with the 685 registered accidents representing 51 % of the total. Typically, these accidents involve persons on railway tracks (unauthorised persons or trespassers) that are hit by a running train. Accidents at level crossings, including pedestrians, is the other main category, with a total number of 350 accidents in 2020 (26 % of the total). Together, these two categories represented 77 % of the total number of railway accidents in the EU.

Figure 1: Rail accidents by type of accident, 2020
Source: Eurostat (tran_sf_railac)

Germany registered 294 accidents, the highest number of railway accidents among the Member States, followed by Poland with 179. Together those two countries recorded more than one third (36 %) of all significant railway accidents in the EU in 2020. At some distance, France (104 accidents), Romania (102) and Hungary (98) followed, each with between 8 % and 7 % of the total. By contrast, Luxembourg reported one significant railway accident in 2020 (an accident to persons by rolling stock in motion) and Estonia and Slovenia five accidents each. In 2020 there were no significant railway accidents registered in Ireland, nor in the Channel Tunnel.

Downward trend in fatalities persists

The number of persons killed in railway accidents in the EU was 45 % lower in 2020 than in 2010.

Figure 2 shows the number of persons killed in railway accidents in the EU for each reference year from 2010 to 2020. The total number of fatalities gradually declined from 1 245 persons killed in railway accidents in 2010 to 687 in 2020. With the exception of slight increases in 2013 (+3 fatalities) and 2016 (+12), the number of persons who lost their lives decreased year-on-year throughout the period. The strongest decrease was recorded from 2019 to 2020, with 115 fewer persons killed in such accidents (-14 %). Over the period from 2010 to 2020, the number of persons killed in railway accidents was almost halved (-45 %).

Figure 2: Persons killed in railway accidents, EU, 2010-2020
(number)
Source: Eurostat (tran_sf_railvi)

Figure 3 presents the number of rail deaths per million inhabitants in 2020. The EU average of 1.5 masks wide differences between Member States. Two Member States registered more than 4 deaths per million inhabitants: Slovakia (4.9) and Romania (4.1). Nine Member States registered less than one fatality per million inhabitants (Slovenia, Finland, Italy, France, the Netherlands, Sweden, Spain, Ireland and Greece), while Luxembourg recorded no fatalities in 2020.

Figure 3: Persons killed in railway accidents, 2020
(per million inhabitants)
Source: Eurostat (tran_sf_railvi), (demo_pjan)

The overall trend in the data shows that travelling by railway is safe, with few fatalities among rail passengers. Based on the most recent data, ERA has estimated that the fatality risk for passengers travelling on trains is 0.033 fatalities per billion passenger-kilometres at EU level, or correspondingly one fatality each 30 billion passenger-kilometres. [1] However, it should be noted that there are substantial differences between Member States.

Most fatalities were persons unauthorised on tracks and railway premises

The main type of accidents was accidents caused by unauthorised persons on the tracks

Focusing on 2020, fatalities in the category “Unauthorised persons” (Table 1) remained the largest category of victims, with 411 cases (60 % of the total number of persons killed in railway accidents). The second largest category was “Level crossing users” with 213 deaths (31 %). Only a fraction of the registered fatalities were railway passengers. In most of the years for which harmonised data are available (i.e. from 2010 onwards), railway passenger fatalities represented only a marginal share (1 % - 5 %) of the persons killed. The exception was 2013, when the 97 passengers killed in railway accidents represented almost 9 % of the total. This was solely attributable to a railway accident in Santiago de Compostela in July 2013. The 79 fatalities from this accident represented all of Spain’s fatalities in the category “Railway passengers” and 81 % of all railway passengers killed in accidents in the EU in 2013. In 2020, the share of railway passengers in the total number of fatalities remained at 1.5 % (10 railway passengers killed), the same share as the previous year.

Table 1: Persons killed in railway accidents by category of user, 2020
(number)
Source: Eurostat (tran_sf_railvi)

Table 2 outlines the fatalities from railway accidents according to the type of accident. In 2020, close to two thirds (66 %) of these fatalities in the EU were caused by “Accidents to persons by rolling stock in motion”, typically involving persons that are unauthorised on the railway tracks and are hit by a running train. Together with level-crossing accidents, which caused 31 % of fatalities, these accidents were responsible for over 97 % of all deaths occurring in railway accidents in the EU.

Table 2: Persons killed in railway accidents by type of accident, 2020
(number)
Source: Eurostat (tran_sf_railvi)

The number of persons killed in ‘accidents to persons by rolling stock in motion’ was particularly high in Germany (109 fatalities) and Poland (98 fatalities) in 2020.Together, the five Member States Germany, Poland, Romania, Italy and France registered almost three quarters (73%) of the persons killed in accidents to persons by rolling stock in motion.

Fewer persons seriously injured

Significant reduction in persons seriously injured in railway accidents since 2010, especially for railway passengers

Over the period 2010-2020, the number of persons seriously injured in railway accidents at EU level fell for all types of persons (Figure 4). This trend is noticeable despite occasional year-to-year fluctuations. Whereas the average annual decrease in the total number of seriously injured persons averaged 11 % per year from 2010 to 2015, an increase of 21 % in the number of injured persons was observed in 2016. The high number of injured railway passengers registered in 2010 was due to a severe train collision at Buizingen (Belgium) with 171 persons seriously injured. In 2016, the number of injured passengers registered a significant increase compared with the previous year, with a number of larger accidents involving passenger trains recorded across the EU. The number of seriously injured persons fell again in 2017, but remained at the same level in 2018. However, the number of injured persons from railway accidents fell significantly both from 2018 to 2019 (-18 % to 612 injured persons) and from 2019 to 2020 (-24 % to 468 injured persons).

Figure 4: Persons injured in railway accidents by category of person, EU, 2010-2020
(number)
Source: Eurostat (tran_sf_railvi)

Focusing on the year 2020, Table 3 illustrates the correlation between the various accident categories and the number of persons injured in the different types of accidents. More than half of the seriously injured persons were registered in accidents involving rolling stock in motion (251 injured persons, or 54 % of the total), followed by level crossing accidents (179 persons, or 38 %).

Far fewer persons were injured in other types of accidents. Train collisions seriously injured 29 persons in 2020, of which 12 persons were injured in accidents in Czechia and 8 in Germany. In 18 of the 25 Member States with railways, no persons were seriously injured in train collisions in 2020. Throughout the EU, no person was injured in train derailments or fires in rolling stock in 2020. Germany (54 seriously injured persons), Poland (23), Romania (22), Italy (21) and France (20) recorded the highest numbers of persons injured in accidents to persons by rolling stock in motion. It should be noted that the comparison of seriously injured persons between countries may be slightly biased due to differences in reporting regimes.

Table 3: Persons injured in railway accidents by type of accident, 2020
(number)
Source: Eurostat (tran_sf_railvi)

Suicides cost far more lives than accidents

Suicides on the railways outnumber the number of persons killed and seriously injured in accidents by far

Suicides occurring on railways are reported separately from persons killed or injured in railway accidents. For the EU as a whole, the number of such suicides remained at between 2 200 and 2 800 per year in the period 2010-2020. The highest number was recorded in 2012 with 2 734 suicides on railway premises. In the following years, the numbers fluctuated, changing by between -7 % and +6 % year-on-year. The numbers of suicides on railway premises have fallen consistently since 2017.

In 2020, 2 204 suicides were reported, 109 fewer (-5 %) than in the previous year. The number of suicides on railway premises is significant in most of the Member States. With 678 recorded suicides in 2020, Germany alone accounted for almost one third (31 %) of the EU total. There were also significant numbers of such suicides in Czechia (204 recorded cases), France (203) and the Netherlands (198).

Table 4: Suicides on railway premises, 2010-2020
(number)
Source: Eurostat (tran_sf_railsu)

Source data for tables and graphs

Data sources

The sources used for the statistics in this publication are data reported to the European Union Agency for Railways (ERA). Railway safety data have been collected by ERA since 2006 through the Common Safety Indicators (CSIs). These were introduced by Annex I to the Railway Safety Directive (Directive 2004/49/EC). Member States have a legal obligation to submit their CSI data to the ERA. ERA publishes an overview of safety-related CSIs as soon as data have been consolidated. The CSIs data are reported via and available through the ERAIL system. The full set of CSI data is made available in the annually published Railway Safety Performance Report. Accident figures are reliable from 2010 onwards, following the strict application of standard definitions. In the past, Belgium, Poland and Slovakia typically reported all railway accidents instead of significant accidents only. This meant a lower count in several categories of accidents since 2010.

Eurostat has signed an agreement with ERA to disseminate these railway safety data through Eurobase. The data disseminated by Eurostat constitute a subset of the data available at ERA.

Railway accident data were also collected through Annex H to Regulation (EC) No 91/2003 on rail transport statistics. This Regulation has been recast and replaced by Regulation (EC) No 2018/643. The data collection through Annex H was phased out and replaced with the data collected by ERA. However, Montenegro, North Macedonia and Turkey continue to provide data according to Annex H to Regulation (EC) No 91/2003, as they do not have an agreement with ERA.

The railway accident data collected by ERA data are located in the “Multimodal data (tran)” section under “Transport safety (tran_sf)” in Eurobase. Historic data based on Annex H can be found in the section “Railway transport - Historical data (2004-2015) (rail_ac_h)”.

Some differences may occasionally exist between these data, as ERA handles its own compilation procedures and quality checks. Also, whereas data reported to ERA are provided by the national safety authorities, data reported to Eurostat under the framework of Regulation (EC) No 91/2003 were reported by the national statistical institutes. The NSIs might have depended on data from the same national safety authorities, but not necessarily.


Composition of EU aggregates:

EU: European Union composed of 27 Member States: Belgium, Bulgaria, Czechia, Denmark, Germany, Estonia, Ireland, Greece, Spain, France, Croatia, Italy, Cyprus, Latvia, Lithuania, Luxembourg, Hungary, Malta, the Netherlands, Austria, Poland, Portugal, Romania, Slovenia, Slovakia, Finland and Sweden.

It should be noted that the EU Member States Cyprus and Malta have no railways.

Regarding the data for EFTA countries, Iceland has no railways, while Liechtenstein's railways are included in the Austrian data as they are operated by the ÖBB.

The tables include the Channel Tunnel as a separate entity, as data referring to railway accidents in the Channel Tunnel cannot be assigned to either France or the United Kingdom. EU aggregates do not include Channel Tunnel figures.

Data for Montenegro, North Macedonia and Turkey are extracted from Annex H to Regulation (EC) No 91/2003.

Some data for the most recent reference year may remain provisional for some time. This is linked to ongoing investigations and hence decisions whether to include or exclude certain accidents and or their categorisation.

Context

National rail networks have different technical specifications for infrastructure – gauge widths, electrification standards and safety and signalling systems – which make it more difficult and costly to run a train from one country to another. EU policies exist to overcome such differences. Creating an integrated European railway area thus requires better technical compatibility – 'interoperability' – of infrastructure, rolling stock, signalling and other subsystems of the rail system. Procedures for authorising the use of rolling stock across the EU's rail network also need to be simplified.

The European Union Agency for Railways (ERA) helps promoting interoperability and develop uniform technical standards, a process in which cooperation between EU countries and rail stakeholders is essential.

ERA, based in Lille/Valenciennes (France), is helping to build an integrated European railway area by improving rail safety and interoperability. Set up in 2006, it develops shared technical specifications and approaches to safety, working closely with stakeholders from the rail sector and national authorities, the EU institutions and other interested parties. Featuring a dedicated safety unit, the Agency also monitors and reports on rail safety in the EU.

Direct access to

Other articles
Tables
Database
Dedicated section
Publications
Methodology
Visualisations





Database

Multimodal data (tran)
Transport safety (tran_sf)
Rail transport safety (tran_sf_rail)
Rail accidents by type of accident (ERA data) (tran_sf_railac)
Rail accidents victims by type of accident (ERA data) (tran_sf_railvi)
Rail accidents involving the transport of dangerous goods (ERA data) (tran_sf_raildg)
Suicides involving railways (ERA data) (tran_sf_railsu)

Data collected through Annex H of Regulation (EC) No 91/2003:

Railway transport (rail)
Railway transport – Accidents – Historical data (2004-2015) (rail_ac_h)
Annual number of victims by type of accident (2004-2015) (rail_ac_catvict)
Annual number of accidents by type of accident (2004-2015) (rail_ac_catnmbr)
Annual number of accidents involving the transport of dangerous goods (2004-2015) (rail_ac_dnggood)


Dedicated section

Methodology


Notes

  1. For further details, see ERA’s Annual overview for Safety - 2021.