The permanent effects of innovation on financial depth:: Theory and US historical evidence from unobservable components models