So, student subsidies are the odd one out here, because they're not money you're given to survive, they're given in exchange for a specific activity. They do, however, count as transfer income, and the source I have for 60% didn't break it down.
Yes, unemployment benefits can provide a backstop that enables entrepreneurship, but Denmark has more generous benefits than most other countries, yet the levels of entrepreneurship don't follow.
Any other direct subsidy towards entrepreneurship (although often misguided) aren't transfer income in the sense being suggested by the GP.
The people who can just sit around and come up with exciting new research or who can become successful entrepreneurs are generally also the kind of people who can get a job if they want or need one.
I guess it'd be interesting to read a study quantifying it, but I'd be surprised if it were really true that not much research comes out of people on transfer incomes, especially if you include the people who used transfer incomes to bootstrap the early stages of a startup.