I’ve had occasion in the past to mention Larry Summers’ “ketchup economics” paper. Niklas Blanchard reports that there is in fact an empirical investigation of the ketchup market, and it’s not efficient:
If you would like to see this divergence from principle on the micro level, you can read about a study of London’s ketchup market done by James Monetier in his book, Behavioral Finance (pp 29-31). He found deviations of up to 43% from the theoretically predicted price. Where are the arbitrageurs taking advantage of this free lunch (and maybe even eating a tastier lunch)? Of course, in the real world, arbitrage opportunities take time to discover, come and go, and can be hampered by various barriers.
The issue is that ketchup is cheap, so across a wide range of ketchup pricing schemes it doesn’t make much sense to invest time and energy in ketchup arbitrage.