Large banks have imposed huge costs on society and the call for their break up has been justified on this grounds.
Some recent work has identified scale economies still exist in these large institutions.
An excellent keynote by Bob De Young highlights two recent papers:
which both find significant scale economies in large US banks. Hughes and Mester 2011 go much further in capturing the risk preference of management, something that was missing in previous studies. Wheelock and Wilson 2011 use non-parametric methods to more precisely measure the tails in the production function.
Of course one large caveat of these studies is the lack of a quantified social cost.
While I am no advocate for big banks I do think a more accurate cost benefit analysis is necessary.
A colleague, Professor John Turner, quite rightly points out that with political economy we may end up with banks that are either two small or too big.