Well, charge pump current is a factor in the open and closed loop performance, but the VCO gain, loop division ratio, and loop filter components all come into play here. There really isn't a single equation that relates the charge pump current to lock time, phase noise, etc. Since all of these parameters are multiplied together, you can adjust things to compensate for other thing, and you can adjust the loop filter components to compensate for changes in the other loop parameters.
For a charge pump, I'd have to say that
1-up/down mismatch
2-leakage current
3-Dead Zone
4-Current output noise
are perhaps more important than the exact value of the charge pump current. It sounds like you need to spend some time looking at the linear, closed loop performance to get a better understanding of the trade off's.
If you have access to MathCAD, you might want to download Lance Lascari's MathCAD PLL worksheets and get a better understanding of the trade off's.
http://tools.rfdude.com/ is the link.
If you don't have MathCAD, I'm in the process of writing a PLL tool in Excel, that's located at
www.keystoneradio.com/?n=Main.PLLDesign. It's not quite done yet, but it may be useful.
Dave