I've got some interesting feedback / questions for clarity from the last post in the comments and off line. Thought I'd post some more thoughts here, on whether 5 years is too long... The reason I like 5 is that I think it's hard to know (sometimes) whether something is going well or not until year 3, and I'm nervous about constantly writing big grants.
and yes, 5/3 cycle funding would be a replacement for "operating" multi-year project grants (2-4 years) are a different good idea.
Maybe the whole package is something that looks like:
Short turn around micro grants / recommenders: There are little ($500-$2,000) things that need to get done quickly.
Project Grants: Pretty much like what we have.
Multiyear project grants: Covering 2-4 years and focused on a single "project" but including some operating expenses (covering activity outside of the single project like admin and ancillary projects)
Operating Cycle Funding: Described in the post. Allows for flexibility and stability and isn't as focused on a single project, but on a breadth of activity.
Also I'm in favour of radically changing peer jury system - changing especially how disciplines and scales are thought of. Right now most of my frustrations are about the decisions being made by the juries, not the system. But the changes in the system could help. Again, there's more detail on options in Shannon Litzenberger’s Metcalf paper, Choreographing our Future.