Testing Efficient Risk Sharing with Heterogeneous Risk Preferences: Comment