But making all ships capable of doing any job equally well removes meaningful choice.
Some ships should be specialised, as the passenger ships were, while others should remain multirole, jack of all trades, master of none.
I believe this would make more ships useful. Otherwise you can just pick whichever ship you like the look of and can afford and just stick with that one.
Quite the opposite, having choice requires things to be balanced as otherwise a single option will dominate. Asking a player to choose between 1 million credits per hour or 0.5 million credits per hour isn't a choice at all - 1 million is clearly better than 0.5 million. However, as you have pointed out, a choice between 1 million credits and 1 million credits isn't really a choice either. Continuing this analogy, what is needed is a choice between 1 million credits and something that is nominally worth 1 million credits that may be situationally better than simply taking the cash (or might be worse, depending on situation).
You also seem to be conflating balance and similarity, when the two are quite different. Granted, many developers choose to balance games by making things the same as it is a very easy way to balance things, but equally it is very much possible to create a game with a great variety that still remains balanced. A good example is StarCraft, where each of the races are quite different in execution and yet are all fully capable of performing well under all but the most contrived set ups. These differences are possible because there are many different aspects and methods to a particular activity, and this also applies very much to Elite as under most situations a ship's overall performance would depend on a large variety of different ship statistics and variables.
For example, traders very much like having a large cargo capacity, but they also depend on jump range, supercruise maneuverability, realspace maneuverability for docking and even heat generated from the FSD for those that frequent planetary bases (I've not included combat specific stats because NPCs interdict like an unconscious drunkard in zero-g trying to walk). This means that, even without going into cosmetic or lore reasons, it is quite possible to have a large variety of traders that all offer similar performance under normal conditions (there will always be extreme conditions that benefit certain ships while negating the strengths of others, but they should be minimised at the BGS level rather than attempting to balance ships around them) yet are actually quite different to play as. It wouldn't be simply picking a ship you like the look of, but picking one based upon your playstyle (both what you are good at as flying as well as what you enjoy flying) as well as after careful consideration of the tasks specifics.
For passenger ships, making them the best PAX ships but useless at everything else is the very antithesis of choice. If the Dolphin, Orca and Beluga were decisively the best PAX ships in their price range, then that would effectively remove every other option in the game for passenger running. Rather than a careful consideration of a large variety of different ship's relative strengths and weaknesses, it simply becomes PAX -> Saud Kruger. Rather than making a choice yourself, the choice is made for you the moment you decided to try moving some passengers. Similarly, if the SK ships were to be nerfed into oblivion in every other regard to "balance" their passenger superiority, then they are effectively removed as a choice for every other activity, players would no longer even have to consider them. If this happened across the board, then players would pretty much only ever have a single ship as a viable option.
Ideally, when someone states a budget and asks what ship they should buy for a particular activity, they shouldn't receive a simple answer. There shouldn't be a "best" option that everyone can agree on, but instead it should spark a discussion regarding the relative merits of a variety of different ships coupled with requests for more information regarding the task at hand and finding out what the prospective pilot is experienced at flying already.
The only way that specialised ships could allow for choice is if there were a
lot of them. As in, several times the current number of ships in the game at minimum, enough so that for each given activity there is a range of different ships at each price point. That same choice can be achieved with less than half the number of slightly varied ships (not even full-spectrum multirole, but even just dual-role ships will help), as individual ships can pull double-duty or even triple-duty as they can be considered for multiple tasks. Considering the rate of implementation of new ships in the game, it would be impossible to have actual choice all throughout the variety of different activities and at all price points.
Obviously, the more multirole a ship is the more it should cost relative to its specialised peers. However, that extra cost shouldn't serve to render them invalid as a choice compared to their more specialised brethren, as that would render them completely useless. In this regard, I actually see the multirole tax as being a touch too high at the moment (particularly with the recent buffs to the T9), as at the moment it's pretty much that no trader would ever consider buying a multirole ship for trading except for the AspX (which is incredibly good for its cost, it could literally have its hull cost doubled and it would still remain a solid investment) and the Python (which has the special position of being the decisively best medium freighter).