From Brad DeLong, economic historian at U.C. / Berkeley, who places the internal migration within the context of the past 60 years (from his substack blog):

For 40 years now, I have been a short-run China bull and a long run China bear. It has seemed clear to me, at every point in time. It is, of course, very clear and obvious that at each stage in the past I have been wrong.
China has, since 1980, at every point in time… been able to successfully rejigger its development model at each stage to continue economic progress. I think Barry Naughton’s periodization of post-Mao China is the most useful way of thinking about the stages of this process:
1. Agricultural de-communization
2. Labor-intensive manufactures in the Chinese countryside via [Town and Village Enterprises]
3. China’s cityscapes via retail decontrol
4. The drastic and painful 1996 to 2002 shrinkage of [state owned enterprises] work force by more than 40 percent— the flip side was that alternate businesses and ownership forms had reached sufficient scale to absorb the workers, land, and structures
5. Greatest of all: 200 million migrants flooded into the urban economy
6. The decision in 1998 to privatize urban housing
7. Finally, the decision to enter the WTO—the China-export shock

From here: The census in 2000 found that there were more than 120 million migrant workers in Chinese cities. More-recent estimates go as high as 200 million. This massive internal migration has appeared especially dramatic from a Chinese perspective because mobility was severely restricted in Maoist times, making it almost impossible for rural people to leave their villages.

The pull factors for rural migrants were to be found in the growing prosperity of the coastal zone, where China’s extraordinarily rapid economic growth has been concentrated. Some migrants went to established cities, where they worked as petty traders or in the service sector, as well as in manufacturing and the construction industry. Others were drawn to the new urban areas of the coastal zone, such as Shenzhen, where booming export industries created an ever-increasing demand for assembly-line labor.

Also from here: In 2009, there were 145 million rural-urban migrants in China, accounting for about 11 percent of the total population. Among them, an estimated 85 million to 100 million were born after 1980 — a period when three distinct government policies converged to shape the circumstances for increased rural-to-urban migration within China.

After its introduction in 1979, the controversial One Child Policy, which promoted late marriage and delayed child bearing and limited the number of children born in rural families to 1.5 (two for a first-born girl, otherwise one), was firmly implemented and shifted the vast rural China household structure — and thus, agricultural workforce — dramatically to fewer children.

Then in the mid-1980s, the Hukou System — a residence registration system devised in the 1950s to record and control internal migration and which ultimately hindered rural-to-urban movements — began to loosen in response to the demands of both the market and rural residents wishing to seek greater economic opportunity in cities.

International migration has more than tripled in size since 1960, rising from 77 million to almost 281 million by 2020. The costs of doing so declined:

First, the costs of movement. Air flights have reduced the cost, delays and uncertainties of travel. As more persons in developing countries gain more income, air travel is more affordable.

Second, the costs of settlement. As immigrant communities in destination countries increase in size, the ability of would-be and arriving migrants to to find housing, get jobs and fit in grows. This applies perhaps most to unauthorized migration – how to avoid deportation. But it also applies to legal immigration when channels such as refugee migration broaden. The Trump administration wants to stop the catch-and-release practices when people cross the Mexican border and then meet up with earlier migrants. This most recently has enabled unaccompanied minors to cross the southern border.

Third, the costs of keeping in touch with the host country. The most obvious improvement is in phone calls. It also includes air travel. Many immigrants today return for temporary visits. And consider remittances. Immigrant households in developed can probably better afford to share incomes in ways that can be used meaningfully in the country of origin. Methods of sending remittances have improved over the decades, and still are. African taxi drivers in Washington DC tell me they routinely back for stretches — their nuclear families live there.

Trump’s policy was essentially the increase the cost of migrating to the U.S. Just one example: it became much harder for spouses of H-1B workers to legally work.

Much of this is spelled out in Paul Collier, Exodus: How Migration is Changing our World.