The variance overlaps from age 20-59 so you can't statistically differentiate anything conclusive about the difference between those groups. An alternative conclusion is that retired people (65+) don't drive as much and are able to avoid rush hour / high risk times to drive.