A growing literature characterizes climate change damages by relating temperature shocks to GDP. But theory does not clearly prescribe estimable forms of this relationship, yielding discretion to researchers and generating potentially considerable model uncertainty.