I studied this problem while I was at the University of Alaska Fairbanks,
but unfortunately left all the materials there.

My ultimate conclusions were 1. an average recovery across institutions
is not conceptually meaningful because the recovery rate depends on they
type of institution: one heavily into medical sciences will have a high rate
because NIH generally pays IDC, whereas an institution more heavily
dependent on the Dept of Ed. or Ag. will appear to do less well.
2. It is not possible to make any methodological comparison between
institutions because the data simply is not available in comparable
form. I had hoped to compare land grant universities, or institutions of
comparable size, using data that COGR had compiled and data compiled by the
InterMountain University Research Administrators Group. Howver there were too
many uncontrolled variables. Some included the medical or vet school or ag.
center; others did not, for example. 3. Institutions vary as to the extent
to which they will waive IDC. 4. Certain assistance programs are valuable to
the institution, whether the grant carries full IDC or not - they help pay for
necessary things.

I think the accountants have to defer to the chancellor, president or
provost who approves the policy relating to IDC waivers and acceptance of
awards that restrict IDC recovery. Such policies are based on instiutional
circumstances and strategies.One must realize that rejecting an award that
incompletely funds IDC may have impacts that are not primarily financial.
In addition, because many institutions have discretionary funds, it is not
(from an acconting viewpoint) unreasonable to fund part of the (indirect) cost
of projects that are in the institution's interest to perform.

I hope these comments are helpful. Charlie