Yes, as a movement it doesn’t, but the competing hypotheses of @agauger and myself are both testable by the same experiment:
Catalytic antibodies are structurally constrained and must function in the context of the immunoglobulin (Ig) fold.
If their prevalence is representative of the prevalence of function in sequence space, if we experimentally release them from that constraint by adding random sequence on each end, at the end of several selection cycles we should observe that enzymatic activities are significantly greater than those of the parent catalytic antibodies that were selected as antibodies.
If their prevalence is not representative of the prevalence of function in sequence space (for virtually any reason, including those offered by @Agauger and @bjmiller), if we experimentally release them from that constraint by adding random sequence on each end, at the end of several selection cycles we should observe that enzymatic activities are the same or lower than those of the parent catalytic antibodies that were selected as antibodies.