This is a basic statistical misunderstanding because a huge part of that winrate is correlation.
Same reason Rabadon's deathcap has a high winrate. Not because it's broken, but because you buy it when you're already snowballing. In other words, it gets correlated with winning teams.
So it is with drakes. The team that was better/played better/was winning anyway would pick up the drakes along the way, so of course drakes would have a high winrate.
Here's an example. It's probable that firstblood has a surprisingly high correlation with winrate. But that doesn't imply first blood's reward is too strong. It implies that the better player, who is going to win lane, is likely to be the one to have gotten first blood anyway.