When you apply for a grant, you add an "overhead" amount, which is specified as a percentage of the grant. A typical number is 50%: if you want $X for your lab, you apply for $X * 1.5, with the reasoning that the extra 50% is for the university to provide you with facilities, tuition wavers for your students, computer support, etc. So far, so good.
What many people don't know is that of that 50% that the university collects, about 20% (actual figure can vary a lot) comes to the Dean's slush fund (so, about 10% of the grant itself using above figures). The Dean can use this money to give young faculty starter research grants; but more often, the Dean can use this money to dole out discretionary bonuses (which can be very large) to the top research faculty. So if the grantee gets a $19MM grant, you can bet that he's making 100s of 1000s of dollars in bonuses from the Dean.
Anyways, the system is quite messy if you take a closer look...
I've never heard of this kind of kickback scheme, in >10 years at several large research universities in the US.
Here's a recent article about universities resisting efforts to curb overheads: http://www.bostonglobe.com/news/nation/2013/03/17/harvard-mi...
He most likely did this to get more resources for research work. E.g. it was motivated by a desire for fame and status.
I mean, you pay the University $XX for "misc facilities and support". That goes into the school's fund. Later, the Dean gives you a "completely unrelated and coincidental" Xmas bonus of $X. Also out of the school's fund. It's not grant money at that point. It's from a general discretionary fund that just happens to get deposits for services provided to grant based science.
So... money laundering basically I guess?