I noticed this on my last game. I was Montezuma on Archipelago. I tend to play Archipelagos over-heavy on Navy and commerce with the idea of a UN or Domination victory in mind.
Anyhow, I founded Hinduism and Judaism in my capital, Tenochtitlan, before founding my 2nd & 3rd cities.
After bee-lining to Optics and spreading my religions and converting everybody but Alexander (who was a devout Christian), I noticed my capital had a ridiculous amount of raw gold production even before Banking.
Do you REX your first 6 cities? I always run into major -

issues with expansion because of maintenance and distance.
I often wonder if my city placement is just off or something. Most of the time, I REX 3 cities and stomp my nearest neighbor after slingshotting CoL.
I think this is where I'm going wrong. Not only do I tend to neglect my specialists, but I tend to use my GPs for slingshotting in the early game. I guess I focused so much recently on
starting good that I lost focus of the long term repercussions.
Yeah ... I'm totally doing something wrong. No wonder I'm finding the jump to Monarch so unnerving.
You guys got any end-game saves you'd be willing to post for me to reverse engineer? I'm going way wrong somewhere along the lines of city sites, tile improvement, specialist assignments and wonder-placement. Hopefully I can glean even more useful insight by spying on your cities

.
Thanx to everyone so far for your help ... I can't wait to see if paying more attention to this stuff has helped (it sure seems to have helped my confidence, at the very least.)
Looking over the responses to this thread, it's a sign of a well-designed game when people have multiple ways of winning and they all work. I don't even care about SE v CE, I start off SE for production and whipping and start hybridizing later. Like I said, my style is intensely bureaucratic right now, though I've experimented with other playing styles and light-bulbing instead, and lately I have not been founding ANY religions but just going for early conquest and subsequent rexing.
I don't rex six cities early due to maintenance problems; I usually only have 3 or maybe 4 cities when it's time to pick up the axe and crush my neighbor, keeping his best cities and razing the rest, backfilling later. Re: my five-city problem: the problem was that I was stuck on the smaller continent all alone for my last game and wasn't used to having to actually, you know, FOUND that many cities instead of CAPTURING them, so I lost track and came up with only five cities instead of six at a time when my capital was ready to Oxford-ize. By the time my sixth city cranked a university out, my capital was in the middle of Statue of Liberty, so I had to wait forever for that to finish before I could Oxford-ize, since others already had democracy and I was worried that if I paused to Oxford-ize, I'd lose SoL. I still won by bribing the giants of the map to fight each other but had to space-race it and won in the 1960s, but that's not my preferred victory condition.
Also, island maps are somewhat unusual, so you might want to experiment and figure out what works best on those maps. There is nothing inherently wrong with using a specialist to light-bulb, but you better know what you're sacrificing that GP for.
A super-scientist generates 9 research base with representation, for instance, and library, uni, observatory, oxford, bureaucracy = +225%, so that super-scientist would be producing 29.25 beakers per turn under those conditions, not to mention generating a baseline hammer (and culture if you have Sistine Chapel). Over the course of, say, 100 turns, that's 2925 beakers difference NOT INCLUDING the side benefits of that hammer--faster wonder production and also, hammers convert to beakers at 100% rate if you tell the capital to produce research. My capital tends to be a production superpower in midgame thanks to having dozens and dozens of specialists/superspecialists + Bureaucracy, so frequently I can build even wonders very quickly (with industrious and Organized Religion especially, and I usually have stone, copper, and iron on hand.. also marble sometimes) and let the capital produce research for the rest of the time.
So it's actually best to light-bulb critical stuff early OR to light-bulb nearer the end of the game when the game won't last long enough to make settlling worthwhile.
And at higher difficulty levels, you almost HAVE to light bulb a little bit at first just to catch up to the AI's cheating.
Also, Angkor Wat makes priests into all-purpose specialists that generate 2 hammers, 5 gold, and 6 beakers baseline, making that a favorite Wonder of mine, as it lets you sit at 90% or even 100% research pretty easily if you have any decent infrastructure at all. (However, be prepared to increase engineer and scientist count to match, unless you want all of your great people to be prophets.)
P.S. I agree with the above about not founding a holy shrine early; the payoff sucks if only 2 cities have your holy city's religion; I'd rather either pop the prophet or (you guessed it) settle him.