Matthew Yglesias suggests that pervasive traffic-law breaking by both bicycles and automobiles is partly attributable to an excess of rules including (but not necessarily limited to) overuse of traffic-control devices. Ryan Avent stakes out a better position in suggesting that traffic control properly should be aimed at restricting the behavior of the potentially most-dangerous vehicles on the road, i.e. automobiles.
In this light, the primary fault is in subjecting bikes and autos to the same traffic controls. Particularly in urban and suburban areas, much automobile control is largely for the purpose of “traffic calming,” e.g. impeding traffic via quasi-random stop signs, speed humps, lane narrowing, and the like, so that drivers are discouraged from barreling down 25 mph residential streets at 40. That’s rarely an issue for bike traffic. In fact, if a car slowly rolls through a stop sign installed for calming purposes, as often happens, the driver really isn’t grossly violating the spirit of the sign; the snooty spandex-clad cyclist who just rolls on through even less so — though in either case, this argument should not be expected to win the day with a traffic cop. Nevertheless, being prepared to yield the right-of-way is more important in many cases than coming to a complete stop.
Less-restrictive controls for bikes are an option. Indeed, along the heavily-trafficked Southwest Path in Madison, the city replaced stop signs with yield signs at some campus-vicinity intersections where the path carries more traffic than the cross streets. On shared streets, setting separate controls for bikes and cars is not beyond comprehension; for example, it’s already present where streets are marked as dead-ending for cars but not for bikes and peds.
More broadly, being assigned the rights and responsibilities of automobiles is a mixed blessing for cyclists. Merely having a right to the road is fine for a (small) segment of the cycling population, but not having to live with auto-centric traffic engineering is better.