arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2507.02439econ.GNq-fin.EC
收藏

引入一个新的英国脱欧相关不确定性指数:其演变与经济后果

Introducing a New Brexit-Related Uncertainty Index: Its Evolution and Economic Consequences

Ismet Gocer, Julia Darby, Serdar Ongan

AI总结:

本文构建英国脱欧相关不确定性指数BRUI,结合文本挖掘、NLP与大语言模型分析经济学人智库报告,并用VAR检验其对宏观经济变量的影响,为英国政策制定提供稳健工具。

摘要:

重要的改变游戏规则的经济事件与转型会引发不确定性,这些不确定性可能影响投资决策、资本流动、国际贸易和宏观经济变量。其中一项重大转型是英国脱欧,它指的是英国退出欧盟的多年过程。本研究开发并使用了一个新的英国脱欧相关不确定性指数(BRUI)。在构建该指数时,我们应用了文本挖掘、上下文窗口、自然语言处理(NLP)以及来自深度学习技术的大语言模型(LLMs),对经济学人智库2012年5月至2025年1月的月度国家报告进行分析。此外,我们采用标准向量自回归(VAR)分析来考察各宏观经济变量对BRUI冲击的模型隐含响应。在开发BRUI的同时,我们还构建了一个互补的新冠疫情相关不确定性指数(CRUI),以区分源于这些不同事件的不确定性。实证结果以及BRUI与其他早期开发的不确定性指数的比较证明了该新指数的稳健性。这一新指数可以帮助英国政策制定者衡量和理解英国脱欧相关不确定性的影响,从而支持更有效的政策制定。

英文摘要:

Important game-changer economic events and transformations cause uncertainties that may affect investment decisions, capital flows, international trade, and macroeconomic variables. One such major transformation is Brexit, which refers to the multiyear process through which the UK withdrew from the EU. This study develops and uses a new Brexit-Related Uncertainty Index (BRUI). In creating this index, we apply Text Mining, Context Window, Natural Language Processing (NLP), and Large Language Models (LLMs) from Deep Learning techniques to analyse the monthly country reports of the Economist Intelligence Unit from May 2012 to January 2025. Additionally, we employ a standard vector autoregression (VAR) analysis to examine the model-implied responses of various macroeconomic variables to BRUI shocks. While developing the BRUI, we also create a complementary COVID-19 Related Uncertainty Index (CRUI) to distinguish the uncertainties stemming from these distinct events. Empirical findings and comparisons of BRUI with other earlier-developed uncertainty indexes demonstrate the robustness of the new index. This new index can assist British policymakers in measuring and understanding the impacts of Brexit-related uncertainties, enabling more effective policy formulation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑