> But isn't China's growth still mostly in the non-poor parts of it,
No.
https://en.wikipedia.org/wiki/List_of_Chinese_administrative...
Scroll to: By share of national GDP per capita
It'll show you: List of Chinese provinces by share (%) of national GDP per capita 2000 - 2017
You'd expect if all provinces grew equally (e.g. doubled in size), that their relative shares would not change between 2000 and 2017. (A) If poorer provinces grew faster than richer provinces, they'd grow in relative share. (B) The same holds for if rich provinces grew faster than poorer provinces, then rich provinces would grow in share. (C)
So sort on the year 2000. You see Shanghai at the top with 374% of the nation's average GDP per capita. At the bottom is Guizhou with 35% of the nation's average. Shanghai goes down to 214% while Guizhou goes down to 64%. Both grew in absolute terms, but the poorer province grew much faster.
A bit more data. It tends to hold across the board with few exceptions.
top 3
Shanghai 374 -> 214
Beijing 304 -> 218
Tianjin 219 -> 201
bottom 3
Tibet 58 -> 66
Gansu 52 -> 48
Guizhou 35 -> 64
> isn't that 12k figure pulled down by the exchange rate
No, the measure is: GDP per capita adjusted for price changes over time (inflation) and price differences between countries. It should take this into account.
> the massive rural population
I'm not sure how it does... First I'd want to reiterate that, uplifting rural people from poverty, both by developing these areas and by creating jobs (causing population shifts away from rural areas) has been a huge part of the economic development in China. If you'd dismiss this growth, you'd have never reached $12k in the first place.
Second, dismissing the rural population is exactly one of the biggest problems you could imagine for China. One of their biggest fears is that they leave hundreds of millions of people behind, upset, in poverty. The only reason they can keep an orderly dictatorial system is because they deliver on the promise to the vast majority of the population, many of whom poor: lots of jobs, annual wage growth, hope and optimism. Low growth prospects for a massive rural population would be disastrous. As such, even if GDP per Capita was $100k in some small areas, if that $12k (or lower) figure is not rapidly developing for China's massive rural population, that'd be seen as a huge policy issue right there. The argument that the $12k figure may be pulled down by rural poor should underline how worrisome relatively low-growth is.