This is an interesting idea. As far as I understand, many grantors such as NSF place strict limit on each categories (PI salary, student stipend, material, trav

This is an interesting idea. As far as I understand, many grantors such as NSF place strict limit on each categories (PI salary, student stipend, material, travel), and generally you cannot move extra $ from one category to another. I think your proposed method does look like stream-lining the accounting, though I do not have a strong opinion. From the grantors' perspective, they might prefer the first as the second reduces transparency on where $ goes?