If believing P will result in epistemically good outcomes, does this generate an epistemic reason to believe P, or just a pragmatic reason? Conceiving of such reasons as epistemic reasons seems to lead to absurdity, e.g. by allowing that someone can rationally hold beliefs that conflict with her assessment of her evidence's probative force. We explain how this and other intuitively unwelcome results can be avoided. We also suggest a positive case for conceiving of such reasons as epistemic reasons, namely, that they exhibit a form of interpersonal normative parity that's typical of epistemic reasons but not pragmatic reasons. We then link this discussion to religious belief, suggesting that there are sometimes indirect epistemic reasons for religious belief, and that certain characterizations of religious belief are instructive in thinking about how to take account of indirect epistemic reasons.