Is there value in participating in a startup accelerator program?

There’s reason to believe that the answer is yes. To date, startups that have participated in one of the 200 such programs have raised a combined $20 billion in funding since 2005, according to data published on Friday by Jed Christiansen, a product manager at startup accelerator Techstars and who also runs an online database about accelerators.

Over the past decade, startup accelerators, which select small groups of promising startups and provide them with resources like mentorship, business advice, and often even funding, have increased both in popularity and number.

Get Data Sheet, Fortune’s technology newsletter.

Of course, a few exceptional outliers are responsible for the huge cash haul. Airbnb, for example, which graduated from the prestigious Y Combinator program in 2009, has raised almost $4 billion in debt and equity funding to date. Video gaming network Twitch and self-driving car startup Cruise Automation, both of which also participated in Y Combinator, have been acquired for almost $1 billion each.

Overall, accelerator alums have been responsible for more than $5 billion in exits, venture capital jargon for acquisitions or going public. Those that are still operating privately are valued at a combined $80 billion.

Pitchbook, an investment database, also found that one-third of startups that raised a Series A round in 2015 previously graduated from a startup accelerator. This suggests that participating in one increases the chance that a startup will “succeed” (if we define success as growing and continuing to raise funds), since these startups account for around 10% of all startups founded in the last few years, according to Christiansen.

But that still raises the question of whether these startups are succeeding because of the coaching and knowledge they gain through the accelerator programs, helping them build businesses attractive to investors—or if the accelerators’ brands are acting as a stamp of approval to investors.

I’d bet the accelerators would argue it’s the former, of course.