Although in theory the long-run effect of uncertainty on investment is ambiguous, available econometric evidence widely supports a negative association between aggregate investment and political instability. A shortcoming of this body of evidence is that it has failed to investigate the existence and direction of causality between these two variables. This paper fills this gap by testing for such causal and negative long-run relationship between political instability and investment. We find there is a causal relation going from instability to investment, but it is positive and particularly strong in low-income countries. This finding is robust to various sensitivity checks.