Race as a Cause

This paper explores the use of race as a cause, as in claims that people's race sometimes causes racial discrimination or that race causes poverty. Such claims are common both in everyday discourse and, to some extent, in the social science literature. One objection to the use of race as a cause derives from an interventionist account of causation: race does not seem to the sort of property that can be manipulated or changed through interventions. I suggest that when a property has the feature of unmanipulability, this is often an indication that causal claims involving the property are unclear or in need in disambiguation. Such claims often can be replaced by causal claims involving other variables that better capture what is intended. Two sorts of claims in which race is proposed as a cause are discussed: the claim we need to treat race as a cause to make sense of racial discrimination and claims that race can be regarded as a cause of outcomes such as income and educational attainment. I argue that in both cases alternative formulations are preferable. In social scientific cases, claims that race is a cause of some outcome often commit us to unclear counterfactuals that we don't know how to assess. Moreover, such claims violate other plausible conditions that causal claims ought to satisfy-- they are associated with modularity violations, ambiguous interventions and relationships that are not "autonomous" or "structural" in the econometric sense. I argue that such claims are better replaced with investigations that focus on specific racialized economic and social mechanisms that produce racial disadvantage. Although some of my criticisms of claims about causation by race echo those of scholars who focus on "constitutive" rather than causal relations in understanding racial inequality, I nonetheless defend the use of causal analysis as central to addressing such inequalities.