Publicly available startup data includes firms that exist just as online profiles. So, maybe these firms will do their product some other time or they will disappear. It's better to exclude such startups from stats and look at who survives.
Funding is a good filter here. Getting seed funding means a startup at least has a team and idea. But over the years, the fraction of series A deals decreases:
If a smaller fraction of startups gets next-stage funding, it means that fewer startups survive after getting seed money. This survival rate indicates how well startups get prepared for doing business. The fewer firms lost on the way, the lower risks investors bear.
The major startup nations from CrunchBase:
China and Israel do well here. The US makes other countries look like dwarfs on charts, so it has a separate graph:
About 80% of startups live their first to fifth funding stage. Having more stages isn't that common. By the later stages, a startup either becomes a company with more conventional funding (revenue, bank loans, bonds, public equity), or gets acquired by another company, or disappears.
Replication files: https://github.com/antontarasenko/blog-replication-files/tree/master/2014-09/08_cb_funding_stage