I think the very first thing we have to understand is what Base Tax really represents. Paradox has been pretty obscure in this matter, so it is pretty difficult to imagine what they wanted to do.
Once this has been said, I would just try to focus on how I see it.
I usually see people try to equate wealth with BT. I do think that this a mistake. We have to consider that there are three ways of earning money: BT, production and trade. In this case, I would say that BT has to be raw taxes over population.
I have been working for quite a lot of months in a project to update BT values for the M&T mod. I have been able to wrote down a large list of data to modify both BT and MP values, so I think I can share a little of my experience.
First of all, I have just worked with one data: the XVII Century. I have chosen this data for two simple reasons, it is a middle ground and there is a good amount of data available. Why do I need a middle ground? Well, because not only colecting data for all the timeline would be crazy (if not impossible), but we have not dynamic BT values. We would have to change it by event and it would be a real mess.
I have considered BT as urban taxes. This is a good aproach because I consider that taking taxes in urban areas, where people are more "controled" is easier than in sparsely populated and/or rural areas. Wealth of a low populated region is easily done with trade and production values. Provinces where a good amount of cities existed have those values added to each other to make the final calculus.
So I have been making calculus to convert urban values to Base Tax, and provincial values to Manpower. I do not know if my method is the best, but I'm pretty proud of it because at the very least it gives consistency since all regions follow the same rules.
If we look at the vanilla map, we'll see that regions out of Europe that were pretty rich are far poorer than those in Europe. Does it make sense? I think it doesn't. So now, after studying that data, I have come to a new map where a lot of regions out of Europe are fairly rich. Besides, with the new authonomy modifiers we can play and introduce new "corruption or tax efficiency" modifiers to rectify when needed.
Using my scale, I have to say that Greece is far poorer than what it should be. At least, it is not consistent neither among their provinces, nor among other similar provinces. This can be applied to Iberia, Italy and a lot of regions in Eastern Europe.
For example, in 1600 Greece have these values (just some examples):
Athens: 33.000 inhabitants.
Ioannina: 12.000 inhabitants.
Iraklion: 14.000 inhabitants
Kerkyra: 13.533 inhabitants.
Mystras: 30.000 inhabitants (in the case of M&T, in the same province, there are Kalamai [3.000], Nafplion [6.000], Tripolis [10.000]; for a total of 49.000 inhabitants)
Serres: 22.000 inhabitants (+8.000 in Dráma)
Thessaloniki: 50.000 inhabitants.
And so on. There are at least another 7 10.000+ inhabitant cities and another 5 4.000+ inhabitants cities. We have to consider that anything over 4.000 inhabitants was fairly big for the era.
The biggest city in the world (Beijing) had 706.000 inhabitants. Is this considered in the map? I don't think so.
I hope this post serves its purpose and throws a little of light over the issue (at least, how I see it).