Shaping the Balkan corridor: development and changes in the migration route 2015-16

In 2015-16 Europe witnessed the biggest refugee crisis since the Second World War. Over a million people moved via different migration routes. The key route was the Balkan corridor running from Turkey, via Greece, to Central, Western and Northern Europe. The aim of this paper is to describe the development and changes in the route and provide an analysis of transit via Balkan countries, looking at factors that influenced the shape of the corridor. This refugee crisis was challenging for European countries and the whole European Union (EU). This corridor was unique, being de facto formalized semi-legal territory, which the EU had never faced before. An official reaction to the crisis was necessary due to the substantial number of people who were on the move, seeking to cross the Balkan countries in the fastest manner possible. Therefore, the response was focused on arranging transport and providing only shortterm accommodation. This paper uses the PESTLE framework to examine the key political, economic, social, technological, legal and environmental aspects that influenced the Balkan corridor, as well as changes in the route over time and responses to them. The role of the media in shaping the corridor is also acknowledged, thus resulting in a PESTLE-M framework. Findings from the research are important, as it is likely that the EU will face a similar crisis in the near future. Therefore, there is a need to prepare and develop a plan in case such a situation


INTRODUCTION
In 2015, the total number of arrivals to Europe was 1,046,599. In 2016, the majority (1,034,269) entered through Greece, while 47,136 migrated through Bulgaria. The International Organization for Migration (IOM) identified this situation as the biggest European refugee crisis since the Second World War (IOM, 2016). This paper examines the movement of migrants in 2015-16 through the Balkans, looking at factors that shaped the corridor. The term "Balkan corridor" was initially used in the media and applied to the route through the Balkan states. This de facto humanitarian corridor was used by people who entered countries under certain conditions and later continued their journey without having to present any identity documents . The Balkan corridor started in Turkey -an entry point to Europe -and extended through Greece, Macedonia, Serbia, Hungary, Croatia, Slovenia and Austria, with different configurations over time (Arsenijević et al., 2017). Most of the countries on the way served as transit countries; migrants aimed at used at different levels of analysis, from a single company to a market to a whole country or region. During our analysis, a new dimension emerged: the role of the media was observed to be an important factor in shaping the corridor. The media cannot be classified under individual PESTLE dimensions, as they cut across all of them. The media use Technology, are subject to Legal regulations, can be non-profit or for-profit (Economic factor), may function as a bottomup form of communication such as Social media, and finally they might not be objective, but support certain Political views. In this paper then, the media are treated as a separate, new factor that emerged from the data. This is in line with other research that has modified the PESTLE framework to add or remove certain groups of factors (PEST, STEP, STEEP, etc.). The weakness of the PESTLE analysis is that it allows for generalization, without capturing differences at the lower (in this case city or organizational) level.

BALKAN CORRIDOR
Until the end of August 2015, the Balkan corridor went largely unnoticed (Kasparek, 2016). However, in August 2015, Austrian authorities discovered the bodies of 71 migrants in a refrigeration truck near the border with Hungary (Spindler, 2015). Afterwards, the Austrian police reinforced border controls, which led to blocking the passage to Germany. The "March of Hope" (Kallius et al., 2016), when thousands of migrants left Budapest Keleti train station marching to the Hungarian-Austrian border, changed the "invisible practice" of Balkan migration into a "highly visible phenomenon" (Kasparek, 2016: 4), widely reported in the media and later used for domestic political purposes. Over time, different key events shaped the Balkan corridor (Table 1). Start of the "Balkan corridor" migration. Deaths on railway tracks in Macedonia. August 2015 Open borders in Germany and Austria. Discovery of 71 bodies in a truck near Austrian-Hungarian border. September 2015 Death of Aylan Kurdi during sea crossing to Greece.
Hungary closed border with Serbia (corridor re-routed). October 2015 Hungary closed border with Croatia (corridor re-routed). November 2015 Border-crossing allowed for people from Syria, Iraq and Afghanistan only.
Macedonia refused to let Afghans pass through. March 2016 Transit countries closed their borders. Agreement between EU and Turkey to stop migration.
During September and October 2015, EU civil protection mechanisms were launched for the countries along the corridor (ACAPS, 2016). The process of closing the Balkan corridor began on 18 November 2015, when only people from Syria, Afghanistan and Iraq were allowed to cross the Macedonian, Slovenian, Croatian and Austrian borders (Beznec et al., 2016). In February 2016, Afghans were also banned. The Austrian decision to introduce daily and annual caps on the number of asylum-seekers affected other countries, and "the collapse of the coalition [between Germany and Austria] in early 2016 left Berlin alone to manage the crisis" (Weber, 2017: 7). The whole corridor was closed on 8 March 2016(ME, 2016a. Closure started in Austria and "provoked a chain reaction in other countries of the formalized corridor" (Beznec et al., 2016: 22). As a result, at the end of 2016, more than 75,500 migrants were stuck in transit countries (Batha, 2016), unable to move on.
A growing number of fatalities in the Central Mediterranean corridor (via the sea) contributed to an increase in the usage of the Balkan corridor. More unaccompanied minors, as well as families with children, started using it (Hintjens, 2019). The Balkan corridor began in Turkey, however, many people had already travelled long distances to arrive there. From Turkey, migrants went by sea to the Greek islands, where they were registered and then transferred to the Greek mainland. From there, they continued by land to Macedonia, Serbia, Hungary and to the free-movement Schengen zone (Kasparek, 2016;Arsenijević et al., 2017). Slovenia and Croatia became part of the corridor (Figure 1) when Hungary closed its Serbian and later Croatian borders (Sardelić, 2017). The following sections provide country descriptions in the geographical sequence of how migrants entered and crossed these countries. Each country profile contains general information related to the crisis, followed by information about transportation, shelter and movement, and then the situation after the closure of the Balkan corridor. Key events and dates are listed chronologically. Austria is not included, as it was both a transit country and, like Germany, a destination country (Bučar et al., 2017).

Greece
Hundreds of migrants were arriving daily from Turkey to the Aegean islands. Greece adopted the "hotspot" approach, presented in May 2015 at the European Agenda on Migration (EC, 2015b). The aim was to help the frontline Member States (bordering with non-EU countries), i.e. Italy and Greece, to identify, register and fingerprint incoming people. Five hotspots were created in Greece on the islands of Chios, Kos, Leros, Lesbos and Samos (FRA, 2018). From May 2015, the number of arrivals increased sharply, and the capacity of the hotspots was exceeded. The Greek authorities were not able to provide basic services, and people were kept waiting for registration. Importantly, the conditions differed between islands (ACAPS, 2016). Almost 60% of the migrants arrived at Lesbos with daily reception capacity amounting to 1480 people (MSF, 2015), while Greece reached its peak with 211,663 arrivals in October 2015 (Anderson et al., 2016). At Lesbos, buses and vans transported migrants from beaches to the reception sites (UNHCR, 2015a). In October 2015, Syrian families had to register in the Kara Tepe camp on Lesbos where they usually spent the night in tents or in dedicated "Ikea shelters". The next day, they moved to the port of Mytilene, and were then transported to Athens. Single Syrian men and other nationalities had to register at Moria camp -the first hotspot -where they spent around three days with little or no assistance or accommodation (MSF, 2015). On Kos island, between 200 and 500 people arrived daily in the same period, notwithstanding the fact that there was no reception centre.
In Greece, migrants had to process their papers with national authorities. This registration enabled them to travel to the mainland and exempted them from being deported or arrested as their papers were valid for 30 days (ACAPS, 2016). Migrants were able continue to the Athens ports of Piraeus or Kavala (Sekeris andVasilakis, 2016), or to Thessaloniki (ACAPS, 2016). Migrants travelled by ferry (tickets were required) or were transported for free by NGOs. Then, they mostly continued from Athens through the Idomeni border-crossing to Macedonia on commercial buses (Anastasiadou et al., 2018). From 18 November 2015, Macedonia allowed Syrian, Afghan and Iraqi nationals to enter only through the Idomeni crossing. After weeks of protests and clashes with the police, on 9 December Greek authorities forcibly transported people waiting on the Greek side to the temporary camp in Athens (ACAPS, 2016).
After the closure of the Macedonian-Greek border, migrants were stuck in Idomeni (Beznec et al., 2016), and long-term assistance was required. Thousands of refugees and migrants (as of 2019) are still stuck in Greece, waiting for resettlement, repatriation, deportation or an asylum decision (Farhat, 2018).

Macedonia
Macedonia holds candidate status for EU membership. In the 1990s it opened its borders to displaced people during the conflicts following the dissolution of the Socialist Federative Republic of Yugoslavia (Beznec et al., 2016). In 2015, Macedonia tried to facilitate the quick transport of migrants, establishing centres for a few hours' stay (ACAPS, 2016), then allowing transit.
Until spring 2015, transit migration had scarcely been noticed; then, an increase in numbers was apparent. People were walking along railway tracks towards Serbia, usually at night. The last destination before crossing the Serbian border were two villages, Lojane or Vaksince, both with informal camps, "run by smugglers" according to UNHCR (Beznec et al., 2016: 16). Changes were introduced following media attention due to fatal railway accidents -14 people were killed in April 2015 (Šabić and Borić, 2016). Firstly, Macedonian citizens and NGOs initiated assistance. Secondly, the policy on asylum-seekers was changed. In compliance with the Law on Asylum and Temporary Protection, Macedonia introduced a travel permit (Beznec et al., 2016) so that asylum-seekers could register their intention to apply for asylum in Macedonia at the border. A 72-hour travel permit would be issued and migrants could travel to any police station within that time period and register the asylum claim (UNHCR, 2015b). Frontex (the European Border and Coast Guard Agency) reacted to this legislation by stating that it "accelerated migrants' movement" across Macedonia and "increased the pressure on the common border with Serbia" (Frontex, 2015: 18). The permit allowed migrants legal transit through Macedonia and to "use public or private housing and transport as well as free medical aid at any local or state health facility" (Beznec et al., 2016: 18). However, on 20 August 2015, Macedonia declared a state of emergency (Frontex, 2015) and temporarily closed its borders (Beznec et al., 2016). This closure resulted in a growing number of stranded migrants and in clashes with the Macedonian police (Šabić and Borić, 2016), which brought the attention of media (Beznec et al., 2016).
The border was re-opened on August 22. Afterwards, the state dispatched trains from the Macedonian-Greek border to the Serbian border (Kasparek, 2016) and coordinated the movement. A new camp was built in Gevgelija (near the Greek-Macedonian border) to issue 72-hour travel permits (Beznec et al., 2016). UNHCR began monitoring Gevgelija as of July 2015. Approximately 775,823 people transited until the end of February 2016(UNHCR, 2016a. To cross Macedonia, migrants had to walk from Idomeni to the Vinojug centre in Gevgelija, and wait for transportation to Tabanovce, a camp located only few meters from the Macedonian-Serbian border. Transport was by train, bus or taxi (Beznec et al., 2016). Special trains for migrants ended at the Tabanovce camp without any stops, and migrants were charged 25EUR for the trip, as compared to the regular price of 10EUR (Beznec et al., 2016). At Tabanovce, people usually had to wait a couple of hours and then walked two kilometres to cross the border into Serbia at Miratovac (Šabić and Borić, 2016). In October 2015, arrivals reached up to 10,000 people daily (C&DI, 2016). From November 2015, only people from Syria, Afghanistan and Iraq were allowed entry, while other nationalities were stuck in Greece (HRW, 2015). Dozens of people, the majority of whom originated from Morocco and Iran, were held in Gazi Baba Reception Centre for Foreigners in Skopje. The corridor was officially closed in March 2016 when Macedonia suddenly closed its border, which left approximately 1600 migrants stuck in Macedonia (Beznec et al., 2016). Importantly, according to UNHCR (2015b), Macedonia does not qualify as a safe third country.
After closing the corridor, Macedonia abolished travel permits and declared all neighbouring countries as safe. This opened up the possibility for immediate deportation of all newly arriving people (Beznec et al., 2016). Migrants caught by the police were refused the right to claim asylum (Weber, 2017).

Serbia
Serbia was also affected by the displacement caused by conflicts after the dissolution of the Socialist Federative Republic of Yugoslavia in the 1990s. During 2015-16, Serbia became a crucial gateway and transit point for people arriving to the EU via Hungary, Croatia and Romania (Arsenijević et al., 2017). The Serbian Commissariat for Refugees appeared as "the most prominent national organization (…) taking on the responsibility for the admission of refugees" (Beznec et al., 2016: 30). According to Serbian Asylum Law, the police should issue a document to anyone asking for asylum, thereby confirming the person's intention to seek asylum in Serbia, and allowing them to be admitted for 72 hours in one of the asylum centres run by the Commissariat (Beznec et al., 2016).
The number of people arriving in Serbia was growing with the better weather during spring. In March 2015, the parks were full of tents, and around 1000 people slept near train and bus stations in Savski Venac in Belgrade (BCHR, 2016), a transit hub on their journey to Central Europe. Despite the creation of informal camps, the Serbian authorities did not try to remove migrants or smugglers operating there (Beznec et al., 2016).
The new corridor became visible with the growing number of arrivals in Preševo (a city near the Serbian-Macedonian border). After crossing Macedonia, migrants walked for a couple of kilometres through the fields and crossed the border with Serbia. Following the intervention of a local imam, a road was built to make walking easier (Mitrovic, 2019). UNHCR provided bus transport from the border to Preševo, where Serbian authorities established a One Stop centre. Migrants could obtain 72-hour travel permits there. When Hungary closed the border with Serbia, Serbian authorities sent buses to the border with Croatia (Jovanovic, 2015). From Preševo, migrants could go to Belgrade or Šid (a city near the Serbian-Croatian border) by bus or train. Only registered migrants were transported and they had to pay their fare (Šabić and Borić, 2016). Until November 2015, migrants crossed the Serbian-Croatian border at Berkasovo/Bapska on foot (ACAPS, 2016). In November 2015, a new transit camp was opened in Adaševci, and free train transport to Croatia was provided from Šid to the reception centre in Slavonski Brod. Some migrants also came to Serbia from Bulgaria. They usually registered in the centre in Dimitrovgrad, obtained travel permits and continued to Belgrade by bus (ACAPS, 2016), where they spent some time in parks and entered the corridor in Šid (Beznec et al., 2016). The second entry point from Bulgaria was Zaječar (Šabić and Borić, 2016).
Closing the corridor led to an increasing number of migrants stuck in Serbia. Some were deported from Slovenia to Croatia or Serbia (PI, 2016). However, according to UNHCR, Serbia was not classified as a safe country (Šabić and Borić, 2016). During spring 2016, Belgrade was full of migrants again. Transit camps in Šid, Principovac and Adaševci became increasingly crowded. However, closures of borders between countries on the Balkan corridor reduced the number of registered migrants in Serbia. At the same time, there was an increase in reported violent events (Arsenijević et al., 2017): human rights organizations estimate 100-140 pushbacks to Bulgaria per day, while the number of illegal push-backs to Macedonia is not clear (Weber, 2017).

Hungary
Over 350,000 migrants transited through Hungary in summer 2015 (Čapo, 2016), although the country had little experience with large numbers of migrants. The first response of the authorities was to detain all migrants who had illegally crossed the border. At the same time, the construction of a 175km fence along the Serbian border was announced and completed in September 2015. Afterwards, migrants shifted from the Hungarian-Serbian to the Hungarian-Croatian border, which was also later sealed (Svensson et al., 2017).
The Hungarian Office of Immigration and Nationality established reception centres. After crossing, migrants received free tickets at the transit camp in Röszke to other camps in Bicske (near Budapest), Debrecen, and Vámosszabadi (near the border with Slovakia and Austria). Migrants had to take trains through Budapest to reach these destinations. However, the majority continued to Western Europe, which was their target destination, instead of traveling to "overcrowded camps, where the prospect of quality legal aid [was] minimal and reception conditions [were] poor" (Kallius et al., 2016: 2).
The Budapest Keleti train station became the unofficial transit point. However, transport by train and bus to Western Europe was prohibited for migrants by the government (Kallius et al., Kasparek, 2016). Between 31 August and 4 September 2015, thousands of people were stuck at Keleti train station, forcing the Hungarian government to decide whether they should allow migrants to use trains or suspend international rail traffic (Kasparek, 2016). In early September, migrants staged a protest (Kallius et al., 2016). On 4 September 2015, migrants started a march toward a border crossing with Austria. On the way, several people collapsed, and this forced the Hungarian government to provide transport by bus to the Austrian border (Kallius et al., 2016). The tensions at Keleti train station peaked and the government decided to open the border to Austria and provided buses from Keleti (Kallius et al., 2016). Hungarian and Austrian citizens also offered transport to migrants from Budapest.
During summer and autumn 2015, Hungary was the only country that tried to protect the national and the EU border and constructed a barbed-wire fence. After the border closed, clashes between migrants and Hungarian police took place (Beznec et al., 2016) and were covered by the global media, showing police use water cannons and tear gas against migrants (Zuvela and Than, 2015). After fencing its border, Hungary created transit zones close to the border-crossing points at Horgos-Röszke and Kelebia-Tompa. Those were the only places where migrants could legally cross the border. Hungary also declared Serbia a safe third country, which meant that authorities were able to send people back to Serbian territory. Due to limited access to these zones, two illegal camps formed in front of the crossings in April 2016. Hungarian police were pushing thousands of people per month back to Serbia and there is evidence of violence during these push-backs (Weber, 2017). According to Human Rights Watch (2016), migrants were beaten, attacked with dogs, kicked, handcuffed and forced to go through small openings in the razor fence, leading to injuries, when they tried to cross the Hungarian border illegally.

Croatia
The crisis began immediately after Hungary closed its border with Serbia (16 September 2015). The Balkan corridor was re-routed and there was a massive influx of people which was impossible for the government to manage (Šabić and Borić, 2016). In Tovarnik, the State Administration for Protection and Rescue, the police, and the Croatian Red Cross were prepared for reception, expecting up to 500 people daily. However, 5650 migrants crossed the Croatian border on the first day. Faced with this influx, the migration policy was suspended, and Croatia focused on helping the incoming people (Čapo, 2016), thereby facilitating the migration movement (Kasparek, 2016).
Authorities organized trains for transit (Čapo, 2016;Beznec et al., 2016). Refugees were separated from the local population. Migrants crossed the Croatian border on foot at Bapska and Tovarnik, and were then transported to the reception points by bus or train (Šabić and Borić, 2016). After several hours at reception centres, they were transported by bus to the destinations prearranged by Croatia, Hungary or Slovenia (Čapo, 2016). In November 2015, Croatia and Serbia agreed to provide free train transport from Šid (Serbia) to Slavonski Brod (Croatia) (ACAPS, 2016). From Slavonski Brod, migrants were moved to Sibinj, where free trains or buses took them to Dobova or Mursko Središće (Slovenia). Initially migrants were accommodated in centres for asylum seekers in Kutina and Zagreb "on an ad hoc basis" (Šabić and Borić, 2016: 12). However, the centres ran out of capacity within a few hours. Incoming people filled Tovarnik city while waiting for the train. Therefore, a temporary centre in Opatovac was opened, followed by the winter reception transit centre in Slavonski Brod.
According to Kasparek (2016), some migrants were transported from Croatia to Hungary. There, on a field behind the border, they changed trains and were able to continue to Austria. On 16 October, Hungary built a fence along the border with Croatia. Afterwards, the corridor was re-routed to Slovenia (Šabić and Borić, 2016). Until 17 October, Hungary allowed migrants to cross its border with Croatia through Botovo by train or through Petrovo Selo by bus (Čapo, 2016). Since Slovenia was not prepared to accept a large number of migrants, Croatia tried to slow down the transit through the centre in Opatovac. This led to a growing number of people waiting at the border with Serbia (Čapo, 2016). New restrictive measures implemented in February 2016 resulted in a growing number of asylum applications (Šabić and Borić, 2016).
Like other transit countries, Croatia adopted a policy of push-backs. According to UNHCR, approximately 150 people were pushed back to Serbia per month. These push-backs were also violent; refugees had their money and smartphones taken from them, or were stripped of their clothes and shoes (Weber, 2017).

Slovenia
The massive inflow of people occurred in Slovenia after the Croatian-Hungarian border closed and peaked during spring and summer 2015 . Slovenia started preparing a contingency plan in case of a massive influx. However, the plan was not applicable, however, as the government had not foreseen that most incoming people would not apply for asylum in Slovenia or that Croatia would not accept them back . Initially Slovenia implemented the same approach as Croatia, closing borders to everyone without documents (Šalamon, 2016b). Faced with failure, Slovenia followed the Croatian approach (Kasparek, 2016), facilitating transit.
On 17 September, approximately 200 migrants arrived at Dobova from Croatia by train. They did not apply for asylum, so the Slovenian police treated them as irregular migrants, trying to send them back, but they were not accepted (Ladić and Vučko, 2016). Then, railway passenger transport between Croatia and Slovenia was suspended (Šabić and Borić, 2016), and migrants were not able to reach Slovenia by train. Those who chose different ways to reach the border were not allowed to enter Slovenia (Ladić and Vučko, 2016). People got stuck at the border without any shelter, food, sanitation, water or medical help (Videmšek, cited in Ladić and Vučko, 2016). As a result of the blockade, migrants entered Slovenia illegally at the green border; those captured were detained. The number of arrivals decreased after five days, when Hungary reopened its border. However, on 17 October the border with Hungary was closed again, so migrants were only able to continue the journey via Slovenia .
During the first weeks, migrants travelled by train to the border with Slovenia and then crossed on foot to Rigonce village (ACAPS, 2016). Crossing the border was illegal, but migrants were not returned to Croatia. On 28 October 2015, transport from Opatovac or later Slavonski Brod (Croatia) to Dobova (Slovenia) began after an agreement was reached between the countries (ACAPS, 2016). However, migrants had to wait for several hours without access to medical care, food or water (PI, 2015). The registration of arrivals began in November 2015 around the Dobova border-crossing point, where assistance was provided by the Administration for Civil Protection and Disaster Relief (Šabić and Borić, 2016). From Dobova, migrants continued by train to the Sentilj/Spielfeld or Jesenice/Villach crossings with Austria. During the first weeks, people were transported by bus to reception centres in Brežice, Gruškovje, Dobova and Livarna by the police. When the capacity of the Dobova centre was exceeded, migrants were transported to transit facilities at Gornja Radgona, Šentilj or Gruškovje for registration. Migrants did not have any freedom of movement in Slovenia. Their journey was fully controlled by the authorities, who organized free transport and took care of accommodation and registration centres (Šalamon, 2016b(Šalamon, , ACAPS, 2016. According to statistics from the Slovenian police, 396,240 migrants crossed the Slovenian territory between September 2015 and January 2016 (Šalamon and Bajt, 2016). On 11 November 2015, Slovenia began fence construction along the Croatian border (Šabić and Borić, 2016). In the following weeks, Austria did likewise.

DISCUSSION
Transit countries adopted one of two policies: either "turning a blind eye" or actively supporting the transit of migrants. The term "policy" is used here not in the legal sense, as not all actions were regulated by law, but by national practice, in some cases supported by law. The first policy was "turning a blind eye" on people who wanted to cross the territory (Macedonia and Serbia). Following this policy, both countries provided 72-hour travel permits, provided that migrants expressed their intention to apply for asylum. The second policy was to transport people (Croatia and Slovenia). As a frontline European country (and frontline EU member state), Greece was placed in a difficult position. Being a seaside country, it was impossible to close the sea border on the Greek side. Before re-routing the corridor in September/October 2015, Hungary was the second "gateway" to the EU and the Schengen zone. Hungary was able to close its land borders through the construction of fences.
The PESTLE model (Figure 2) is used to analyse factors that affected the shape of the Balkan corridor and changes to it. The following paragraphs focus on each group of macroenvironmental factors.

Political
The countries along the Balkan corridor are characterized by emigration more than immigration. Therefore, the situation in 2015-16 was challenging (Svensson et al., 2017). The influence of national governmental policies and of the EU on the Balkan corridor is apparent. Bulgaria, for instance, was not part of the main corridor, and one of the reasons for this was that the country "treats illegal refugees in a more rigorous and restrictive way" (Kyuchukov, 2016: 3). The decision of Hungarian authorities to build fences and close its border led to rerouting. Transit countries faced the decision to try to stop the movement and close their borders or to allow migrants to transit. They were also able to restore border controls, i.e. at the German-Austrian, Austrian-Hungarian and Hungarian-Slovenian borders (October 2015). Another example is the decision of the Slovenian government to allow only Syrians, Afghans and Iraqis, which was followed by Croatia, Serbia and Macedonia. As a result, different nationalities were stuck in Greece. In Weber's opinion (2017: 17), the whole shift from the policy of transporting people via national territories to the policy of closing borders was "purely political, without any legal underpinning". According to Sekeris and Vasilakis (2016), the presence of migrants led to increased xenophobia in Greece: after the election in September 2015, support for the antiimmigration party Golden Dawn increased. As Macedonia was undergoing its own profound political crisis, the migrant crisis was ignored during a political debate (Šabić, 2017). Serbia had the most positive rhetoric towards migrants (Šabić and Borić, 2016) and wanted to demonstrate its abilities to be a reliable and responsible European state with potential for EU membership (Šabić, 2017). The Croatian government approached the crisis from both humanitarian and security angles. However, with upcoming elections, the government had to show its ability to control the crisis as it started to dominate the political agenda. The Croatian discourse was also affected by being a member of the EU but not of the Schengen Area. Therefore, its responsibility was reduced by the legal framework (Šabić, 2017). In Slovenia, the question of migration was on the daily agenda before the first people even crossed its border. Politicians focused mostly on security (Šabić and Borić, 2016). In Hungary, the government framed the question of migration as a topic of terrorism or extra costs, pointing out that EU officials paid attention to Italy and Greece while ignoring other countries, namely Hungary, which was also affected by the crisis (Juhász et al., 2015). Hungary was criticized by the EU, the Council of Europe and UNHCR afterward. The EU did not escape changes in mood and rhetoric either. In February 2016, the term "migrants" replaced "refugees", and there was a noticeable shift in European documents toward pro-security that "legitimizes any deterrence practices regarding the movement [of migrants], and undermines [their] rights" (Tsitselikis, 2018:162). Part of the political discussion was a question of national and regional security. A pro-security rhetoric was common to most of the governments of Balkan transit countries. This securitization led to a certain dehumanization of migrants, depicting them as a "threat to national security and social cohesion" (Mušić and Agović, 2017: 16). However, the rhetoric used by Balkan politicians was not new. Already in 2000, Huysmans mentioned that migrants were often presented as a security threat to the well-being of host communities. Šabić (2017: 61) also speaks about the fear that migration will become "a trigger for conflict in the Balkans", showing another angle of this securitization. Some of the concerns were legitimate; according to Šabić (2017), security concerns arose from the crisis itself, including new terrorist attacks in Europe, radicalization and criminal networks of smugglers trafficking drugs, weapons and humans. Huysmans (2020) added that fear strengthens the securitization of migration in general. Pozniak (2019: 69) remarks that migration is "affected by the dynamic between the humanitarian domain and the domain of security, and that these two domains are part of the same wider system". Therefore, the securitization aspect cannot be considered without the humanitarian.

Economic
The economic situation of transit countries affected their ability and will to react. Greece was facing a financial crisis (Skleparis, 2017); it therefore struggled to respond to the crisis due to the "negative results of economic austerity on the state's infrastructure, the overall financial disability of the Greek economy, the high level of unemployment and the high percentage of households living below the poverty line" (Agelopoulos, 2015: 9).
It seems that a greater economic burden was placed on the countries that fully controlled the movement of migrants, providing accommodation and transportation. At the same time, the position of EU member countries seems to be more comfortable because the EU covered various expenses of its member states (Pavlic, 2015). However, the EU provided financial support to non-member states as well (EC, 2016). Macedonia, for example, requested assistance and hoped that using the army would force Brussels to provide aid (Šabić and Borić, 2016).
Macedonia also generated some income through the state-owned railway company Makedonski Železnici as migrants had to pay for train transport. In September 2015, the media informed about "hiking fares" (Marusic, 2015). During summer 2015 the fare was 5-10EUR, whereas in autumn of the same year it amounted to 25EUR. Macedonian taxi drivers were making "extraordinary profits by charging large sums" for transporting migrants to the Macedonian-Serbian border (Šabić, 2017: 55). The grey economy thrived due to the presence of migrants and illegal trade practices. Water, food and bicycles were sold at higher prices to migrants than to local people (Beznec et al., 2016). Migrants were also paying smugglers for their services. There was an economic impact on trade when the international rail traffic was suspended, influencing the movement of goods and people.

Social
In some cases, there was a direct impact on or engagement of society (donations or volunteering) along the Balkan corridor. However, society has another impact in Europe, as people have the right to vote and express their view for (or against) anti-refugee and antimigrant parties, consequently affecting EU and national migration policies. This was highly visible in Hungary where the whole public discourse in 2015 was dominated by the existence of the Balkan corridor. UNHCR reacted to this atmosphere in May 2015 concerning rising xenophobic tendencies (Juhász et al., 2015).
When speaking about society, it has to be mentioned that local civilians were among the first responders in this crisis. In Greece, citizens were handing out water, food and clothes to incoming people (Kingsley, 2015). In Serbia, citizens were distributing food, clothes and hygienic items and also providing information and free Wi-Fi (Beznec et al., 2015). Hungarian citizens transported migrants from Budapest by private car (Kallius et al., 2016) and they also provided in-kind assistance. The Deputy Mayor of Szeged commented on this involvement, stating that citizens "took over duties of the state" (Svensson et al., 2017: 10).
National and international NGOs provided support. Aside from direct assistance, NGOs and wider society also influenced political decisions. This scenario occurred in Hungary and Slovenia where NGOs were speaking against the decisions of governments (Juhász et al., 2015;Šalamon and Bajt, 2016). Similar tendencies also emerged in Macedonia where NGOs helped to change national policy in 2015 (Juhász et al., 2015).
The states which had faced waves of refugees and undergone such experiences in the 1990s took a more pro-refugee and solidarity stance. Serbian discourse, aimed at retaining the basic humanitarian approach, was also shaped by the experience with refugees during the wars in the 1990s. A similar attitude could be seen in Croatia. Šabić (2017: 53) even speaks about "recalling personal tragedies from the wars". These findings are in line with Weber (2017). However, there were evident changes in the approach in 2015-2016 compared to the situation in the 1990s. Sardelić (2017) demonstrates this shift with the case of Slovenia. Despite hosting 70,000 war refugees from Bosnia and Herzegovina in the 1990s (Dragoš, 2016in Sardelić, 2017, Slovenia anticipated only 1000 asylum seekers in 2015-2016, according to its contingency plan. There was another impact of the 90s war. The Croatian Minister of Internal Affairs stated that the transit of migrants was fully facilitated as a protective measure due to landmines in Croatia (Čapo, 2015). The attitude of society was influenced by earlier experience with migration, but also economic and political factors; all were interconnected.

Technological
Mobile technologies influenced the choice of transportation or a decision to use smugglers' networks (Liempt and Zijlstra, 2017). The website AppsForRefugees.com published a list of apps downloaded directly from Google Play or the Apple App Store. Travel apps included InfoAid, w2eu.info, RefugeeMap.com (covering all Balkan transit countries), Love Europe app for Refugees, first-contact.org, Speakfree or OsmAND (AppsForRefugees, n.d.). Apps allowed communication between migrants or between migrants and volunteers.
However, apps were not the only tool for communication. Migrants used social media alongside geolocation applications, WhatsApp or Viber (Zijlstra and Liempt, 2017). They were thus able to communicate, gain information about the police or border controls. There were Facebook pages with smuggler offers (Brunwasser, 2015), including a social media group called "Smuggle Yourself to Europe Without a Trafficker". Having GPS and mobile tools, people were able to move via transit countries without smugglers. Anonymity and accessibility of social networks led to a boom in the amount of "fake news" about the Balkan and European migration crisis (Meusburger, 2018). Politicians and governments used social media, i.e. Twitter as a channel for communication (Triandafyllidou, 2018) to inform about their approach to the event.
Using the Internet and mobile technologies became a new observable phenomenon. According to AppsForRefugees.com (n.d.), "smartphones are the most important tools for refugees". Another, more traditional, phenomenon was building fences within the EU area to act as barriers to people intending to cross borders. Bulgaria erected a fence in 2013 (Arsenijević et al., 2017) on the border with Turkey. However, it refused to build a fence along the border with Greece (Kyuchukov, 2016). Some fences were put up along borders with non-EU countries. However, Hungary, Austria, Slovenia and Macedonia also built fences within the EU. The Hungarian government was the first to adopt this approach, and then Macedonia and Slovenia followed. Nonetheless, fences were only successful in stopping the migration flow at one point. In the end, they just caused the corridor to be rerouted.

Legal
Transit countries on the Balkan corridor have their own national legislative frameworks. In response to the crisis, changes were implemented. Greece modified its asylum law to make the EU-Turkey agreement operable, creating an "exceptional asylum regime on the Greek islands and other border areas" (Skleparis, 2017: 04). Macedonia amended its Law on Asylum and Temporary Protection, and Croatia introduced amendments that allowed the armed forces to cooperate with the police in protecting the border (Šabić and Borić, 2016). The Slovenian government approved amendments to its Defence Act to control and limit migration (Government of the Republic of Slovenia, 2015). Hungary prepared a draft amendment to its national refugee law; however, it breached UN regulations. During the following months, refugee laws were tightened, and after further changes, it became almost impossible to gain refugee status in Hungary. In September 2015, the Hungarian Helsinki Committee presented its evaluation and observed that refugee rights had ceased to exist there (Juhász et al., 2015). Another legal change made by national governments was the extension of the list of safe third countries. At the European level, the Dublin Regulation system was de facto interrupted (Triandafyllidou, 2018), and transit countries accepted it. European countries failed to act collectively (Hintjens, 2019) as existing EU rules were not designed for mass influx (Šalamon, 2016b). People were able to move only inside the corridor. Outside, they were "subject to the regime of asylum, detention, and deportation" (Kasparek, 2016: 6). Therefore, the corridor became a space where laws were suspended, and the legal space was not formally constituted. Inside the corridor, people were under a different legal regime than national or international law. As mentioned by Šalamon (2016b: 46), "entry of people in the corridor may be considered unauthorized, however, in practice it is undoubtedly authorized". The corridor therefore became the most practical option to reach Western Europe when compared to legal channels. Upon the closure of the corridor, legal exit routes from Greece were limited (Tsitselikis, 2018).

Environmental
Environment and geography -climate, terrain, seasonality and weather -influenced the shape of the corridor and had an impact on the number of people crossing it, as did the number and character of border obstacles (rivers, mountains, sea access, islands etc.). Geography is linked to the infrastructure network. An example is train transport from Greece via Gevgelija (on the Greek/Macedonian border) to Serbian border crossings, as travel via Macedonia is much more convenient than via Albania, where mass transport infrastructure is non-existent (C&DI, 2016).
The Balkan corridor in 2015-16 followed some Pan-European transport corridors on main European road/train routes, such as corridor X (Austria-Greece) (Žunić, 2017), which runs through Idomeni, Tabanovce or Tovarnik. One branch of PECX connects Budapest and Beograd, running through Subotica, Röszke and Szeged (Miltiadou et al., 2012). The principal route runs via Slavonski Brod (Ilić and Orešić, 2004) and a railway variant to Dobova. The road variant of PECX has two branches: one to Graz, with Sentilj as a crossing point, and the second branch to Salzburg, with a crossing point at Villach (CEMT, 2003). Corridor IV (Germany-Turkey) was partially used. The Hungarian M1 highway is part of it and was a key road during the "March of Hope" in 2015.
The weather affected the situation along the Balkan corridor in several ways. When the weather was favourable, the number of people travelling the corridor increased (Beznec et al., 2016). As winter was coming, Balkan countries "increasingly coordinated migrant policy and action, both logistically and politically" (Šabić and Borić, 2016: 4). Macedonia adapted its reception places to the needs of winter (EC, 2015a), whereas Croatia built a new winter reception centre in November (Šabić and Borić, 2016). The influence of weather was even more crucial when there was a water obstacle, such as a sea or a river. This was a significant phenomenon mostly before the corridor was "formalized" and when migrants were crossing borders illegally. However, even after formalization migrants had to overcome the Aegean Sea. Despite the border closure, migrants continued to cross the river on the Greek-Macedonian border (Moving Europe, 2016b), which resulted in drownings (Euronews, 2016). These incidents were widely covered by the media.

Media
According to Rastogi and Trivedi (2016), the media should be included in social factors. However, in this paper we took a different approach, as the media played a major role, intersecting with all of the PESTLE categories. The media are not contained within the scope of the Social dimension, but cross the boundaries between PESTLE categories, resulting in a PESTLE-M framework. This is especially visible when looking at social media, where content is edited by individuals or ad-hoc networks using new technology available. At the same time, the media represent different political views (Political factor), and publish information for profit (Economic factor). As mentioned by Bobić and Šantić (2019), political discourse together with public opinion and media coverage depicted forced migrations as a security crisis. Therefore, the media helped shape the corridor, influencing society and politicians (national and EU), informed NGOs, civil society and migrants in the corridor. The manner in which migrants were represented in the media influenced different social and political groups. The first turning point was the discovery of the body of a toddler, Aylan Kurdi, in September 2015. This led to greater empathy of society, as well as "large-scale charity giving, petition-signing and demonstrations" (El-Enany, 2016: 14).
The media in Macedonia reflected a view held by society and government: the primary focus was on security. A similar approach could be seen in Slovenia. In Greece, however, the media focused more on helping migrants, and newspapers emphasized the geopolitical reasons behind the crisis (Georgiou and Zaborowski, 2017). The Serbian media published news about inhumane conditions at the borders and underlined the humanitarian character of the crisis (Šabić and Borić, 2016). The Croatian media also responded positively to the crisis. They compared the experience of incoming people to that of Croatians in the 1990s (Šabić and Borić, 2016) and covered their treatment by the Croatian police or the response from humanitarian organizations (Šabić, 2017). In Hungary, the government dominated the discourse of the crisis in 2015-16. The governmental media used the term "migrant" more than "refugee", in contrast with non-governmental sources, and at the peak of the crisis, the term "illegal" was frequently used (Sik and Simonovits, 2019). The media were a source of information for migrants, allowing fast distribution of information and reaction to changes in the corridor.

CONCLUSION
The aim of this paper is to describe the development and changes in the transit corridor and then to provide an analysis of factors that shaped the movement of people via Balkan countries in 2015-16. The corridor was not static, but rather highly dynamic, adjusting as the macroenvironment changed. All the Political, Economic, Social, Technological, Legal and Environmental factors included in the PESTLE framework influenced the corridor and the flow of people, to which we added the Media factor, resulting in a PESTLE-M framework. To understand the changes in the corridor we need to look at all the macro-environmental factors as they are interconnected. For example, when the political decision to erect fences was announced, the direction of the flow changed. Dynamic changes were possible due to fast information flows. The media, particularly social media, communicated all the political decisions and obstacles, and the shape of the corridor adjusted almost immediately. This clearly indicates that the approach to migration cannot be determined solely at the national level. Local change merely changes the path of the flow from one country to another, so there should be coordination at the EU level and beyond.
To effectively address the question of people in transit in the future, the past has to be analysed. In 2015, the formal mechanisms of the EU border regime failed, and the Dublin regulations were in fact suspended. Local preparedness plans could not handle the movement of such a considerable number or people. Countries on the Balkan corridor were not willing or able to process asylum applications (Šalamon, 2016a), and people were transiting through Balkan countries too fast to be able to receive aid or support other than basic food and non-food items (ACAPS, 2016). Dealing with mass migration was extremely difficult in terms of logistics (Šalamon, 2016b), and was beyond the capacity of a single country.
The nature of humanitarian operations involving transit migration is different from more static refugee or migration crises. The humanitarian response in 2015 and 2016 showed that European countries and the EU itself do not have any plan for crises like this. Long-term planning was missing, and the level of organization was poor. If Europe wants to be prepared for the future, all these aspects have to be solved together with the improvement of cooperation and coordination. Responsibility levels and better control mechanisms have to be considered as well. European countries and the EU also underestimated preparedness and prevention by not monitoring what was happening in Turkey, Lebanon, Jordan or even in the Balkan countries. Development aid and assistance directed to the countries of origin of incoming people might avert or reduce the magnitude of the crisis. The EU and all countries should also focus on protection and not only of people in transit. The level of protection was not sufficient in 2015-2016, nor is it today. There is also a question of border protection, security and human/refugee rights, especially when the nature of this crisis switched from a humanitarian to a security one, which should be addressed in future research.
Future research should also examine which services were offered to people transiting via the Balkan corridor and who provided them, with the aim of optimizing the "logistics for people on the move". This has emerged as a new phenomenon, as humanitarian organizations are familiar with the "static" situation of people in camps, but not with that of people travelling long distances daily. Additional work could be focused on individual countries, their policies, logistical support, border protection, the role of the media, including social media, NGOs and citizens, and research on countries such as Turkey and Austria. Individual cases (regarding organizations, camps, and transit points) could be analysed in depth. There is also the question of the costs and to what extent resources used in the EU could be redirected to humanitarian assistance in countries near conflict areas, protecting people, providing shelter, food, medical services and jobs.