I will now try to explain why I think that we need both, an "opinion" mechanic (species-species and species-empire, or something like that) and a separate "stability/happiness" mechanic.
The fundamental issue here is that we basically want two things, which
seem to be very similar, if not the same (and probably hence the idea we could merge that into one thing):
We want to keep track of how much the populations of our colonies
like our empire (aka loyalty/allegiance), and we want to keep track how
happy/stable our colonies (or their populations) are.
The purpose is to have a mechanic that allows for population discontentedness, unrest up to rebellion/defection from the owning empire; that provides means of peaceful integration of independent worlds, or more general, that allows populations to react to the actions/decisions of the player, etc.
While those two are of course closely related and (in reality) affect each other strongly, these are two
different, distinct concepts (an important fact that cannot be stressed enough!). And as long as we want to model
both of them in our game, we consequently need two separate mechanics/stats. You cannot model both with just one mechanic just as much as you cannot have the concepts of both research and production with only one mechanic. That's basically like wanting to eat the cake and keep it.
If you want to have only one mechanic, you have to give up one or the other concept. If you want to integrate both concepts, you need two mechanics.
Currently population is just a "mindless" resource, which is required for a couple of things (base for other resources like industry and research, crews for ships), and which comes in several different "variations" (species). These variations provide different boni/mali, and the more of the different variations you are able to claim, the better.
Incorporating additional species into your empire is more or less always a good thing and a no-brainer. The incorporation process itself is uncomplicated.
There are only some placeholder exceptions to that rule, like the current xenophobic and rudimentary happiness mechanics. Both are practically negligible.
The unanimous consensus is that this isn't what we want. We want our populations to be not-mindless entities players need to deal with and react to:
1) While multi-species empires should of course be viable, having more species in your empire should come with certain drawbacks/costs. Like multi-ethnic nations/empires in reality, a multi-ethnic empire should be more difficult to hold together than an empire with only a few or even one species.
2) Incorporating a new species into your empire should be a task that requires effort/costs.
3) The population of your empire should react to your decisions and actions, and those reactions should depend on the species. Populations of different species should react differently to the decisions/actions of the player.
4) Players should have different, clearly distinct, viable ways to deal with the population of their empire and their reactions to the players decisions and actions. More precisely, we want to offer them "peaceful"/"diplomatic" approaches versus "aggressive"/"oppressive" ones. A player should e.g. have the choice to handle discontentedness/unrest on a colony (which, in game mechanic terms, means, raising stability/happiness) either by trying to appease/calm the population, or by suppressing it (send in the military). Another example, a player should be able to acquire a native world by other means than military conquest.
5) The way a player treats one colony should not only affect that colony alone. (If I happily slaughter those nasty Abbadoni on one of my worlds, the Abbadoni on my other worlds shouldn't celebrate me like a saviour.)
6) This one might be a bit controversial, as apparently not everyone likes it, but I think it's very important because it addresses something that has bothered me to no end in other 4X space games: ship crew loyalty. In FO, ship crew species is tracked (because species can provide boni to ship stats). Loyalty of ship crews should be a thing. Because that way, you can't just snatch this colony of a super-pilot species, build all your new warships there and use them to slaughter the rest of their former empire (which maybe contains a lot of other colonies with the same species), without suffering severe repercussions (all you brand new ships defecting to that empire and turning on you).
7) Populations should have at least some memory of past decisions/actions of the player, otherwise you could wipe out populations and only have to handle the repercussions while e.g. your concentration camps are active. Once the job is done, scrap them, and a turn later everyone is peachy fine. Not.
8.) One final factor, which is very important regarding this discussion: In FO, species!=empire is one of the fundamental design decisions.
How on earth are you going to model all this with just a planetary stat? This is going to be difficult enough with the proposed opinion and stability mechanics, completely and utterly impossible to do with only one of those.
Point 6 alone is only doable with species-empire relations.
Staying true to point 8 the same, unless you want either all your population react the same regardless of species (scrap 3, which goes against the spirit of 8.) or have no memory of past actions (scrap 7). You can't have pops of different species react differently AND have a memory of past player actions without a per-species bookkeeping of said memory (which is what species-empire relations is about).
Point 4 is simply impossible without tracking stability and loyalty separately. The moment we want to offer the player these different, distinct options to deal with their populations, you need both mechanics to model all possible variations: a player can ensure stability of their colonies by making the population like their empire, or by suppressing unrest. In both cases you get higher stability, but the former only works if you get the pop to like you, the latter will work regardless of how much the pop likes you (and make them dislike you).
Point 1 requires species-species relations if we want to model that in a meaningful way (and not just a plain and boring more species -> more tension). There should be species that get along with each other better than other species. A xenophilic species should be easier to integrate into your empire than a xenophobic one. Holding together an empire of different species of peace-loving tree-huggers should be easier than an empire of different species of warmongering xenophobes.
Point 2 requires point 7, unless we want to do it in a very simple and boring way (like an influence project that costs a certain amount of IP and takes a certain number of turns and then bang, the species is yours).
Point 7 in conjunction with most other points require some kind of per-species bookkeeping in general.
I could probably go on and on. Point is, if we only want happiness/stability, or only species-empire relations etc., most of the above just can't be done. Only a very reduced, simplified version of some of the points maybe. Which means we would sacrifice a lot of interesting possibilities and choices we could otherwise offer players.
IMO these possibilities and choices are (by far!) sufficiently interesting and fun to justify a more complex setup with species-empire relations, species-species relations and stability/happiness mechanics.