One of my favorite biologists is Gunter Wagner. He makes the claim in Genome Biology and Evolution that evolvability and modularity are highly associated. While not proof of a requirement, I think that Wagner is on the right track.
In fact, this sort of research can actually bridge the gap between Intelligent Design and Evolutionary biology.
The main critique ID has for evolutionary biology is that the haphazard mutation/selection paradigm does not create organisms. That doesn’t necessarily mean that they didn’t evolve in some way, but it does rule out the haphazard mechanisms.
As I pointed out in 2008, there is a difference between “parameterized” evolution and “open-ended” evolution. Parameterized evolution requires information about the most likely productive ways to evolve. It means that evolution happens in discrete units, not just in haphazard molecular alterations. A similar concept was developed in 2006 by Lynn Caporale called The Implicit Genome.
In any case, what all of these ideas and suggestions have in common is that information precedes evolution rather than arising out of it. If modularity enables evolvability, what happens before the modularity? Where did the modularity originally come from? This is the core of ID – that information and its similar entities are requirements of evolution, not products of it.
NOTE – I don’t claim that evolution can’t produce any information, only that the information that it does produce requires large amounts pre-existing information to already be there (see also here).
EDIT – I forgot to mention some interesting articles I found when looking through the links to this one on Google Scholar: