They curate broadly. The median Venezuelan+Spanish user votes for 27 distinct authors; Cuban+Spanish: 123. The median American user: 9. Self-vote rates are lowest in the Spanish-speaking groups (0.2% median)
So, they curate more, but it's with a small select group of their own friends? Could it be to make sure those rewards are getting to their community so they have it to spend, specifically? That looks like a calculated effort to extract to me.
They produce substantial content. The median Venezuelan+Spanish post is 5,199 characters — nearly 3× the USA median. Only 1% average under 500 characters, versus 20% for USA.
They tend to double their blogs word counts through posting in at least 2 languages, so that's not a valid figure. They are still 50% ahead, but that doesn't decide quality.
They do power down more (37% vs. 16% for USA). This is the one finding that partially supports the critique. But the difference is in amounts, not commitment: Cuban+Spanish users power up at the highest rate of any group (43%), and the average Venezuelan stakes 11 HIVE versus 205 for an American. That's a wealth gap, not a behavior gap.
Just looking at power downs is only part of the story. Especially after the haircut rule triggers, we receive earnings in liquid Hive and HBD when the haircut isn't in effect that doesn't need to be powered up in the first place. This means many of your figures are quite low from the reality. The truth is those numbers could be double the figure you cite.
I'd like to see total earnings from the Communities and the percentage of that powered up, down and moved off platform in all the forms. Then we start to get an accurate real picture. In an objective way as well. Added rationales for or against aren't objective, it's creating a narrative.
Also, since they are relying on a specific community to curate them, are the blogs/input value actually quality? This of course is subjective, but the community as a whole has rules and these users are following them, which is good, but selectively curating isn't really broadly beneficial.
RE: Extracting Value From Hive? What the Data Actually Shows