So far, the effect of multinational corporations on political stability in a host country is ambiguous. While their exploitation of resources and labour force in the primary sector is theorized to exacerbate grievances and greed among the population, recent research suggests a conflict-reducing effect of foreign companies in the service sector. Prior empirical analyses either examine the role of foreign firms at the country-year level or focus on indicators of political instability separately. Yet, as both economic activities and incidents of political instability tend to be regionally clustered, an analysis at the macro-level may not adequately capture the relationship. In our paper, we contribute to this debate by linking geographically referenced data on affiliates of multinational corporations from the World Bank’s Enterprise Survey with geo-referenced event data on various types of social and armed conflict. Constructing a novel geo-coded data set for developing countries in Africa and Latin America from 2000 to 2016, our paper is the first to assess spatial dynamics of sector-specific economic activities on protest, riots, and violent conflict. In order to explore causal mechanisms, we control for firm-level characteristics and other regional information such as ethnic power relations. The analysis also accounts for reverse causality as prior research argues that political instability adversely affects the investment decision of foreign investors. Our findings provide more refined insights on whether international economic cooperations promote or deteriorate political stability abroad.