The current financial crisis has raised queries about the adequacy of the present regulatory regime. Whilst the immediate priority may be to plug the obvious holes in the system, there are some long-term generic problems with almost any system of financial regulation. This paper explores one such concern, i.e. the boundary problem. This arises because effective regulation, one that actually bites, is likely to penalise those within the regulated sector, relative to those just outside, causing substitution flows towards the unregulated. This boundary problem impacts on many proposals, such as ‘narrow banking’ and my own, with Avinash Persaud, for state and time-varying capital adequacy requirements. The question of how and where to set the boundary is considered. Such boundaries will always be criticised as leading to disintermediation, competitive inequality (no level-playing-field), inefficiency and higher spreads and borrowing rates; and such criticisms will be valid up to a point. The paper ends by discussing how best to respond.