Still, Ulm in there (the classical "Order/Productivity" combo) does pretty well. The initial difference in income with Pan was around 400 gold, and the final difference is of order 10000 gold over 30 turns (30x400=12000). Now, Pan got some nice boosts from random events, but still, Ulm presumably got a better start.
Now all we need is for someone bored to run this kind of test enough times that we'd have an idea of variance