DISCOVER NPL IN EU POLITICS

The European Union's (EU) management of political crises and international negotiations between 2007 and 2021 revealed institutional patterns that directly influenced its capacity for collective action. Two emblematic cases – the 2015-2016 migration crisis and the negotiations with Mercosur – illustrate how fragmented decision-making structures and internal conflicts shaped the outcomes.

Refugee Crisis (2015-2016): Fragmentation and Provisional Solutions Germany’s unilateral suspension of the Dublin III Regulation in 2015, under migratory pressure, exposed the lack of effective solidarity mechanisms between Member States.

Despite 2.5 million asylum applications in the two-year period, the EU failed to reform the Common European Asylum System (CEAS), opting instead for: Temporary agreements with third countries (e.g. Turkey) Transfer of responsibilities to border states Reliance on unsustainable "palliative solutions" The analysis of the case shows that: Shared competences created conflicts between decision-making levels (EU vs. national governments) Asymmetric pressures (border countries vs. others) made lasting consensus unfeasible Lack of institutional "bonding" (systemic solidarity) undermined coordinated responses EU-Mercosur Negotiations (2007-2021): Sectoral Conflicts and Geopolitics

Trade negotiations highlighted: Structural opposition from the French agricultural sector, which blocked tariff concessions for 14 years Strategic use of environmental issues as a non-tariff barrier (e.g. criticism of deforestation in 2019) Subordination of negotiations to parallel multilateral processes (WTO, COP21) Main institutional obstacles: Fact r Impact Agricultural lobby Sectoral protectionism (e.g. beef) Shared competence in trade Difficulty in aligning the positions of the 27 Normative agenda Conflict between economic and environmental interests Analysis Method via Natural Language Processing (NLP) To map decision-making patterns, a 4-step framework is proposed: Data Collection Textual corpus: minutes of European Council meetings, political speeches, legislative documents (2007-2021)

python import requests from bs4 import BeautifulSoup collection of documents from EUR-Lex def fetch_eu_docs(keyword): url = "https://eur-lex.europa.eu/search.html" params = {'text': keyword, 'scope': 'EURLEX'} response = requests.get(url, params=params) soup = BeautifulSoup(response.text, 'html.parser') return [doc['href'] for soup.select('.SearchResult')] Topic Analysis LDA (Latent Dirichlet Allocation) modeling to identify recurring thematic clusters

Sentiment analysis in debates on migration and trade Semantic Network Construction of graphs relating political actors, institutions and key concepts python import spacy nlp = spacy.load("pt_core_news_lg") doc = nlp(text) # Extraction of named entities entities = [(ent.text, for ent in doc.ents] Scenario Simulation Predictive models using transformers (e.g. BERT) to assess the impact of different rhetorical frames

This approach allows: Identifying institutional veto patterns (e.g. France in agriculture) Mapping networks of influence between interest groups and legislators Quantifying the evolution of positions during political crises The cases analyzed demonstrate that the EU's institutional architecture, marked by shared competences and mechanisms fragile multilevel governance, tends to favor minimal solutions when under external pressure. The proposed method offers tools to anticipate blocking points and optimize negotiation processes.