One without the Other? Prediction and Policy in International Studies

Abstract

Salient events such as the COVID-19 pandemic and the Ukrainian crisis and Russian invasion lead to interest and debate on how research can inform policy responses. Research can provide important evidence on the potential consequences of different actions with regard to specific objectives. However, research conclusions also remain uncertain and typically depend on many assumptions. I argue that since policy prescriptions entail claims about future consequences, they must be seen as predictions. Although prediction is difficult, especially about the future, we can have better and more informed discussions about policy consequences if we pay attention to what we have learned about predictions and predicting better. Moreover, beyond whether predictions are ultimately correct or not, it is useful to consider what we can learn from them. In some cases, it is more helpful to understand how specific inputs influence predictions than to focus only on approaches that maximize overall fit. Predictions may not be intended to influence policy, but research is more likely to be useful to non-academics if presented in a clear and accessible manner. Finally, since the outcomes we try to predict depend on policy choices, we can potentially improve predictions by thinking about how decisions are made.