The merits of having a separated budget are obvious. The AIs should have such a system to allow them to grow dynamically, if nothing else.
Less so... is why do these things cost the player money at all? IRL, the player would raise cash through long-term loans, sell stock or otherwise defer direct investment as much as possible. It's reasonable that the player has to invest, say, $100K into a new Colony, but after that, if it's growing and stable, it should be able to grow on its own.
So instead of having a direct cash-for-stuff-then-wait-and-be-bored™ mechanic, there should be a guns-vs-butter system, where the player can direct how much of the Colony cash flow is used for defense vs. growth. 100% into growth means that the player has to go and physically defend everything constantly, but gets huge growth numbers, defense 100% with endgame-size colonies should mean no more bothering with Path / Pirates / Raids; only things that are real threats are determined attempts to sat-bomb Colonies back to the Stone Age or [[Really Scary Endgame Stuff]].
From there:
Colonies should start small. They default to Size 3 now; why not Size 1, where they represent a small core of engineers, biologists, etc.?
Direct income or loss, monthly, to the player represents what's left over, if there are any profits at all. This shouldn't be a gravy train until Colonies are pretty big (Size 5 or more) but it should be a steady, if non-spectacular, source of income past Size 3. Until then, they're spec projects with huge monthly losses on the books, but it shouldn't cost the player other than contributing initial capital. If the player wishes to skip this phase, they can import more people and pay more up front- say, a million credits to start at Size 3. Obviously, Size 1 Colonies are growing quickly, but they're so weak that even a successful Raid probably means their populations are gone.
Running maximal Defense pushes all but the strongest, most-stable economies deeply into the red, and should cost the player real money past, say 70%, representing them digging deep into their personal resources to keep their worlds' economies from crashing during a war; during early lategame (i.e., early colonies) a 30/70 mix of Defense / Growth results in a steady, unspectacular level of profit and smallish defensive fleets that aren't totally worthless vs. small Pirate fleets.
Economies that have a bad mix- Defense vs. Growth is pushing them into the red- should steadily lose Stability and take some time to restore to order and profitability, unless the player can provide the cash.
This opens up Factions being able to change naturally without much intervention- they can spend their internal budgets in occasional decision-making that doesn't need to be "smart", but fixes obvious problems over time. It also provides a very natural way for Stability to cause Factions to declare and choose to exit states of war; as they fight, they'll raise Defense, eroding Stability, and quit when "exhausted". This allows them to have reasonably-realistic periods of warfare and real game mechanics underlying their long periods of peaceful distrust that may even reverse over time, with the exceptions of the Pact and Pirates. It also explains why they aren't annihilating each other- the costs of war are so great that it's not sensible to get into conflicts for long periods.
Pirates and the Pact are both explained by such a system. They're always spending more on Defense than most Factions would, even at peace- 50/50 or worse. They're destabilized and have smaller Faction budgets as a result. Bases established reflect this, and should probably be counted as a "building", dragging down Growth in the short term, but expanding it once built (the loot fuels economic growth, basically), so that there's a real limit on their spam, and they would have cycles where they can build big, nasty bases, but they can't just do it forever without pushing Growth upwards and giving up some spending on Defense.