He is against “one-size-fits-all” policy rules and favors experimentation, but he is skeptical of the experimental claims by the “randomistas.”