Highly skilled professionals' forced migration from Ukraine was triggered by the conflict in Ukraine in 2014 and amplified by the Russian invasion in 2022. Here, we utilize LinkedIn estimates and official refugee data from the World Bank and the United Nations Refugee Agency, to understand which are the main pull factors that drive the decision-making process of the host country. We identify an ongoing and escalating exodus of educated individuals, largely drawn to Poland and Germany, and underscore the crucial role of pre-existing networks in shaping these migration flows. Key findings include a strong correlation between LinkedIn's estimates of highly educated Ukrainian displaced people and official UN refugee statistics, pointing to the significance of prior relationships with Ukraine in determining migration destinations. We train a series of multilinear regression models and the SHAP method revealing that the existence of a support network is the most critical factor in choosing a destination country, while distance is less important. Our main findings show that the migration patterns of Ukraine's highly skilled workforce, and their impact on both the origin and host countries, are largely influenced by preexisting networks and communities. This insight can inform strategies to tackle the economic challenges posed by this loss of talent and maximize the benefits of such migration for both Ukraine and the receiving nations.

翻译：2014年乌克兰冲突引发了该国高技能专业人士的被迫迁移，而2022年俄罗斯的入侵则加剧了这一现象。本文利用领英估算数据以及世界银行和联合国难民署的官方难民数据，探讨哪些主要拉动因素主导了接收国的决策过程。我们识别出持续且加剧的高学历人才外流，这些人主要流向波兰和德国，并强调现有网络在塑造这些迁移流动中的关键作用。主要发现包括：领英对乌克兰高学历流离失所者的估算数据与联合国官方难民统计数据之间存在强相关性，这凸显了与乌克兰的既有关系对决定迁移目的地的重要性。我们训练了一系列多元线性回归模型并应用SHAP方法，结果显示，支持网络的存在是选择目的地国家的最关键因素，而距离则相对次要。本研究的核心发现表明，乌克兰高技能劳动力的迁移模式及其对来源国和接收国的影响，在很大程度上受既有网络和社群的影响。这一见解可为应对人才流失带来的经济挑战、并最大化此类迁移对乌克兰及接收国利益的策略提供参考。