I think that those are two different approaches. E.g., in this howto
http://www.howtoforge.com/howto_bind_chroot_debian we don't need all the libraries etc. in the chroot jail because we tell Bind's init script to run Bind chrooted (by putting
Code:
OPTIONS="-u bind -t /var/lib/named"
into /etc/default/bind9). I think it's a lot easier than putting all the libraries etc. into the chroot jail...