First of all, he's unaware of how that number is generated. 2.2% is the percentage who primarily bike commute. It ignores fair weather or occasional bike commuters. I don't qualify because I telecommute more often than bike. But I suspect there are a lot of fair weather cyclists (which is fine, IMO).
Second, he doesn't consider that this is an average. On some days that number is much higher. And on those days bike commuting significantly decreases congestion and pollution.
Third, he fails to note that transportation is about more than commuting. There may be more trips by bicycle than commutes by bicycle. Update: In fact, there are. According to the LAB, bike trips are three times larger a percentage of total trips than bike commutes are of total commuting trips. Which makes sense. People are more pressed for time and more interested in looking clean and coiffed when going to work.
Fourth, he uses 1990 census data. Why not go back to 1890. Driving doesn't rank that high there, does it?
Fifth, in DC the drive-to-bike ratio is only 30:1, which is really not bad. I'm not sure what the ratio of dollars spent for drivers to cyclists is, but my money is on a number larger than 30.
Sixth, what Dave Jamieson says.
Seventh, Rome<>one day