Lost in the multiverse: Methodological uncertainty in studying global equity returns
We examine the role of methodological uncertainty in studies of international variation in country equity returns, analyzing 15 predictors across 69,120 unique research designs. By varying ten key methodological decisions-spanning data sources, sample preparation, and portfolio implementation-we reveal substantial differences in results. Many established patterns like momentum and valuation effects prove surprisingly fragile. Research designs emphasizing small, segmented markets imrove observed performance, while those focusing on more liquid and accessible peers diminish returns. Using a bootstrap-based test, we identify only a few robust factors, such as market size, issuance, and political risk. Our findings highlight the need for methodological transparency in future research.