I would find it most interesting to give out estimates, like Erik did initially.
Right, but estimates from a government agency are always in dollar (pound, euro, whatever) figures. They don't express it as a percent of previous research spending. I'm saying estimates should be in Production Points--that is not to say that you might go overbudget and spend more, or make a breakthrough and spend less, but that it should be expressed in a simple fashion. If you say "We estimate cost to be 5PP, barring unforseen circumstances" and the player invests 5PP, and it ultimately winds up costing him 6PP, it went overbudget. If he is cautious, he can invest more, or he can go with the esimate.
It's much more logical (and realistic) than saying "We estimate the cost to be 1/3 of the previous year's research budget," which requires they player to do the math himself and adds absolutely nothing to the mechanic. It's the same thing, but with unnecessarily more work.
Unless I get a really good suggestion otherwise, I will keep reactivation at 1:3.
The old DasNES economic system actually does this easily. For example, you have economic levels, like the following:
Depression (-2)-Bankrupt (-1)-Recession (-1)-Very Poor (0)-Poor (0)-Not Bad (+1)-Normal (+1)-Good Enough (+2)-Growing (+2)-Rich (+3)-Very Rich (+3)-Richer (+4)-Richest (+4)-Economic Powerhouse (+5)-Monopoly (+6)
You cannot spend any of that money the turn you upgrade. This is effectively the same operation as being forced to spend your current level to get to the next level. If you are at Good Enough(+2), you spend 2EP to get to Growing(+2). At the transition points (eg: Growing(+2) to Rich(+3)) you spend 2, and gain 1. Hence 2:1. But as you move up the chain (say, (+5) to (+6)) you are spending 5 to gain 1. Hence 5:1 It introduces a system of diminishing returns very simply.
This idea could be staggered out and made to fit a numeric system like that used in this NES as well. For example:
Between 0 to 40PP is 1PP to upgrade for 1PP gained.
Between 41 to 60PP is 2PP to upgrade for 1PP gained.
Between 61 to 80PP is 3PP to upgrade for 1PP gained.
So on and so forth. It's not complex and solves the problem rather simply. You could make the tiers an exponential decay of value to make it difficult to get a planet beyond a certain point (reflecting the difficulty of maximizing planetary output).
Also, depending on the cost of military vessels (presuming a Frigate costs, say, 7PP) I would think a ballpark of 1:3 through 1:5 would be good for colonies, otherwise you have the opposite problem of there being no reason to colonize when you can upgrade your homeworld (even if you use a flat 1:3 upgrade). Vertical vs. Horizontal Expansion should be a player preference, I think, not a decided advantage either way (or each should have its own advantages to compensate).