History-economics-psychology (particularly of groups)

The big factor that seems to be in effect right now is amortizing R&D costs over a long launcher lifetime - e.g. soyuz-proton-long march-ariane

once it works, continue with it - use of proven design lowers the launch failure rate as well

also your tooling costs are paid, and your production training costs are paid, no need to spend them again.

However, at some time the demand for increased capabilities and new technologies make an upgrade desirable - e.g. ariane V, soyuz 2, angara, long march 5

In Japan and the US other factors are at work. In Japan, its the need for assured access for comsats to support the electronics industry, in the US its tha GDFMN.

Defense needs are a separate driver.

If you really want to learn how it works, read Mark Wade's Encyclopedia Astronatutica for about 3 or 4 months - just take any countries program, and follow it through time.

Bottom line, the trade off is money for launchers.