Thread by @Nrg8000

A new report from me and @jleibold. We found that the precipitous and unprecedented decrease in Xinjiang's birth-rate since 2017 has disproportionately come from Uyghur areas. Drastic decreases in indigenous-majority areas, roughly stable in Han areas.
aspi.org.au/report/family-…

In late April 2017 the Party Secretary of Xinjiang convened a meeting and instructed local officials to "reduce fertility... at a moderate level" focusing specifically on the four Uyghur-majority Southern Prefectures, ensuring the officials "contain illegal births".
2/

Since then, the birth-rate in Xinjiang has dropped from about 125% the national birth-rate (a ratio that had been relatively stable since 1990) to less than 80%.
3/

By 2019, the birth rate in Xinjiang had fallen by almost half (48.74%) since 2017. See this tweet by @jnzst.

https://twitter.com/jnzst/status/1364530097701576706?s=19

4/

What we found, when we drilled in deeper and at a finer geographic scale, was that almost all of this decline was coming from counties in Xinjiang with a large indigenous proportion of the population.
5/

Compared to the 48.7% decline overall in Xinjiang, among the Indigenous-majority counties that had published 2019 or 2020 birth-rate figures (these are increasingly hidden), the birth-rate fell 58.5% compared to a pre-crackdown baseline.
Counties 90%+ indigenous fell 66.3%.
6/

The birth rate in Hotan county (خوتەن ناھىيىسى) is 99.4% indigenous fell from 25.4 per thousand in 2012 to 7.4 in 2019, a decline of over 70%.

Kashgar city (قەشقەر) fell from 41.8 in 2017 to 6.5 in 2019 (archive.vn/R3yq8), an 85% decline.
7/

These two graphs show the extremely clear correlation between the birth-rate decline since 2017 and the indigenous proportion of counties and Prefectures. The first graph shows complete figures from 2018 and the second is partial figures from 2019.
/8

What we found was that in 2018, among Uyghur-majority counties (and other indigenous nationalities), there were 162,700 fewer children born than would've been expected if the birth-rate stayed stable.
In Han majority counties, 1000 more babies were born (so roughly stable).
/9

In an interview with state media, the deputy mayor of Bay county said it was important to "punish and punish resolutely" those who violated increasingly strict birth-control policies. "Never tolerate, never soften".
/10

These policies and extremely coercive enforcement tactics (extreme fines, extrajudicial detention ect) are efforts by party officials to "alter and optimise the ethnic structure" of Xinjiang,
archive.is/jH8ab
/11

In Chapchal County, over a four month period in 2017, officials collected nearly $1mil USD in fines from women in the county who violated family planning policies (from around 630 women).
archive.is/iPIRT
/12

Violations of family planning policies and failure to pay associated fines was the most frequently cited reason for people being detained in the leaked Karakash List.
See @adrianzenz' figure here jpolrisk.com/karakax/
/13

These policies, the eugenic-inspired rationale behind them and government statistics demonstrate that these declines in birth-rates are disproportionately affecting indigenous communities and appear to be systematically discriminatory and aiming to reduce births among Uyghurs
/14

The determination of genocide requires an assessment on the mental element of the crime (intent) which we can't offer any assesment about, but this research solidly suggests these policies may constitute acts of genocide under clause D, article II of the genocide convention.
/15

Which - to be clear - only goes any of the way in demonstrating the physical element of the crime - independent assessments are necessary to establish the mental element of genocide to whatever evidentiary standard governments and policy makers deem appropriate.
/16

Internationally, however, this drop is remarkable and unprecidented, even among countries during the midst of some of the 20th century's worst genocides. During the worst of the Khmer Rouge genocide, the birth-rate among women aged 20-28 declined 22% in 2 years. Vs 49% in XJ.
/17

Over the course of two years, the birth-rate in Xinjiang has declined from being similar to neighbouring countries (such as Kazakhstan or Japan) to being almost as low as Japan's birth-rate, a country facing a "national crisis" with their birth-rate.
/18

The policies towards childbirth for Uyghurs (et al) in XJ sharply contrast with the policy direction of China more generally, which now reguarly talks about the necessity of encouraging and incentivising births across the country (if you're not Uyghur)
pbc.gov.cn/redianzhuanti/…
/19

Several Chinese provinces likely provide more relevant benchmarks to Xinjiang than the national average, but even among Ningxia, Qinghai & Yunnan (similar demographics) and Sichuan, Jiangxi & Shanxi (similar economics), Xinjiang CLEARLY stands out.
/20

The increasing sparcity of Chinese statistics has proved a massive barrier to research in these areas. Before 2017 (pictured is 2014), official data separated the overall birth-rate and "minority" birth-rate by county. Unsurprisingly, this stopped following the crackdown.
/21

The latest statistical yearbook (2020 for 2019 data) simply stopped releasing any breakdown of the region's birth-rate. This is why we have had to rely on finding data on individual county websites, where about 50 published 2019/2020 rates.
See

https://twitter.com/adrianzenz/status/1382397174684680197

/22

Surprisingly, Chinese state propaganda actually agrees with a lot of our findings, but frames them differently. In a widely promoted article addressing Zenz' previous research on the topic, sociologist Li Xiaoxia confirms these declines.
chinadaily.com.cn/a/202101/07/WS…
/23

Li Xiaoxia also claims that previously high-birthrates (when XJ's birth-rate was similar to neighbouring countries) were due to “religious extremists” who “bewitched or even coerced people” into marriage and child birth. Also neatly proving our point.
/24

But I'm certain that the Chinese government will continue to 'refute' these findings (that were based entirely off of official statistics) using misleading statistics, other forms of misinformation and ad-hominem attacks.
/25

But the reality of these statistics is hard to explain away without mention the highly discriminatory birthing restrictions for indigenous communities in Xinjiang. Which is likely why they've simply stopped publishing these statistics...
/26

... (@BadChinaTake had a good deleted thread on this - pressure him to repost it)

I consider these findings highly relevant to determinations of genocide in Xinjiang. I believe that this research provides the most compelling evidence so far that these birth-control policies amount to acts of genocide (and qualify for the physical element of genocide).
/27

It is then up to individual governments and others to make a determination about the mental element of genocide - something that OSINT cannot begin to prove to an ICJ standard - based on the balance of probabilities and the information we do have available
/28.

Now, get your hands on the data!
Here is the full report in pdf format - s3-ap-southeast-2.amazonaws.com/ad-aspi/2021-0…
Here is the full datasheet we have curated - docs.google.com/spreadsheets/d…
Here are the relevant sections of Xinjiang's Statistical Yearbooks - mega.nz/file/jZ5QgDDT#…
/29

Thanks to @GroseTimothy, @adrianzenz & @PLMattis (among others) in helping us review drafts of this text, and to @DaniellesCave , @FergusHanson @KelseyMunro, @He_Shumei & @michael_ASPI for additional help.
And massively to Tila Hoja and @DariImpio for research assistance.
/30

Additionally, @dtbyler, @GroseTimothy, @PLMattis, @xu_xiuzhong and @adrianzenz are all magnets for primary source documents and were a massive help in supplementing what we were able to find and fleshing out parts of the report!
/31

Finally, check out the cover art, which is altered from a wonderful photograph of a young girl in Kashgar by @mimokhair. She is obscured by a family-planning target document from Uchturpan county and her face has been replaced with a mostly-AI-generated composite face.
/32