On the (ir)relevance of direct supply-side effects of monetary policy