It likely won’t, because it will be ineffective at the proposed rate such that few people will choose to use it (noob trap). People who are advocating for this change will want it to be a viable proposition relative to the other options. If the mechanic was ever added, but was too weak, players would complain until it was buffed to a sufficient level.
Let’s look at the proposal:
Taking a somewhat generous view of this with a number skewed somewhat towards the optimal end of the range (10 merchants equal to 1 relic), that gives an investment of 1000 Food, 200 gold, and 10 pop space for the equivalent gold generation of 1 relic (30/min). Putting aside the opportunity cost of 20 villagers’ worth of food, 200 gold, and 10 pop space of units that can only do 1 thing, let’s get a reasonable baseline by looking at the existing gold-generating mechanics in the game:
Relics: Assuming you go for self-trade in lieu of competing hard for relics, and your opponent gets 4-5 over you, you have to train 40-50 merchants just to break even with the gold generation provided by his relics (or 20-25 even under nearly completely optimized conditions of 5 merchants corresponding to 1 relic). Lacking this population space for the rest of your eco or military will leave you at a major disadvantage. And more importantly:
Market Trade: Assuming bottomed out market prices, a civ without guilds gets 14 gold/100 food/wood. A civ with guilds gets 17. A single Imperial lumberjack with upgrades takes about 2:46 to gather 100 wood. Let’s round up to 3 min to account for bumping and supoptimal lumbercamp refreshment. That still provides 200 wood in about 6 villager minutes, or 6 villagers on wood can generate about 200 wood in 1 minute. A civ with guilds can then trade this for 34 gold, even at bottomed out market prices (and obviously much before before then). These 6 villagers are generating more gold than the 10 merchants referred to before, at a lower creation cost, and with the benefit of much greater flexibility (and with lesser raidability). Even a civ without guilds gets 28 gold from these 6 villagers, again, almost as much as 10 merchants under the previous assumptions. For both groups of civs, this is better gold generation than 7 merchants (under your stated situation of 7 merchants being the best that could reasonably be expected, and providing the gold gen of 1 relic).
This also doesn’t factor in civs that get more favorable trades per unit time (Saracens due to market bonus, Slavs/Celts due to faster res gathering), or civs with other relevant gold bonuses (Aztecs/Hindustanis getting more from relics, Burgundians/Vietnamese getting added gold from their UTs, or the Portuguese with the Feitoria, which provides 20 merchants’ worth of gold on top of food, wood, stone, lower cost, and much decreased raidability.)
So yes, under this ideation, self-trade wouldn’t be “broken,” but it’s also not strong enough to be useful except (barely) at the very extreme end of the suggested range, with perhaps 3-4 merchants generating the gold of 1 relic. Which just isn’t good design, and will so rarely be worth using that it’s not worth implementing. Maybe we recognize the strength of the much simpler, more intuitive (and more powerful than is being recognized) form of “self-trade” that has existed all along via buying gold at the market.
I think the existing gold-generation mechanics are not being respected nearly enough, and in order to make “self-trade” viable, you’d have to bump those numbers up quite a bit (which then opens the pandora’s box of how to make it balanced). At the end of the day, this idea just seems like an overly convoluted way of reinventing the wheel (pun intended) for an already robust and balanced (infinite, but limited rate) gold generation system. At the risk of being pretty unfair, as you might say, it strikes me as being similar to the idea that, because chocolate cake is enjoyable, one should eat chocolate cake for breakfast every day. Mod or separate game mode? Be my guest; apart from that it seems like a very poor fit for the game (in 1v1s) as it currently exists.