Most of this data has been discussed elsewhere on this forum and others. The main problem is that it didn't test to see if people were gifted or not, but rather conducted a survey on a "gifted" program. These were the people who were interested in being a part of a gifted program. I'm sorry, I cannot imagine even the most gifted and intelligent ESFP teenagers being "interested" in a gifted program on average, though obviously some will be.
The monetary statistics come from an online site whose tests survey the respondents. Problem number 1 is that it asked for "household" income, not personal income. Problem number 2 is that there is no controlling for age/experience. I can imagine plenty of younger, single people just starting out in life, and of these, the ones investigating typology are going to be the INxx types. And I can imagine plenty of older, married, established people having to discover their MBTI type as part of a "team building" exercise at their place of work. So yeah, you're going to get a lot of INxx types looking like they're financially struggling (they're predominantly kids in college or just out of college) while the ESxx types who wouldn't look into typology unless explicitly told to finally do so as part of work later in life when they have plenty of money. It should be possible to publish results that correct for age, but I doubt that's going to happen. Not that it would reverse the trends you see here, but it would make them much less drastic.
This doesn't mean that the conclusions drawn from these flawed sets of data are incorrect, per se. They're just not as strongly supported by the data as people might think. Keep in mind, these same two sets of data are brought up over and over again in threads like these. No one is actually doing any research on the topic. While the statistics do use many data points, we conceptually have only two data points: these two flawed studies. They're suggestive, but not conclusive.