Is spectrum a rival good? I recently stated that since spectrum was a rival good, strong property rights would ensure efficient allocation. Staton pointed out people who disagreed, and then wondered how it should be allocated if it was non-rival. If spectrum is non-rival, it should be completely unregulated, but that doesn't matter since, contrary to dissenters, it is rival after all.

David Reed claims that there is actually no such thing as interference, since "radio signals just add to each other, non-destructively." This is actually false since if a signal came across its perfect reflection, they would (additively) cancel each other out, and the information would be (destructively) lost. This is how those noise-canceller airplane thingies work.

What Reed really means is that processing power in the receiver can deal with dirtier signal, and so use spectrum more efficiently. This statement is not the same as "there is no such thing as interference" and is actually true. There is a trade off between processing expense and spectrum consumption, and only a market can correctly allocate between the two.

This is why strong spectrum property rights are still the answer. If you bought 50Mhz of spectrum, and receiver technology improves so now you only need half that much, you can simply sell off the unneeded remainder. But for this to work you must own the spectrum in perpetuity and be able to use it for anything.

All the considerable waste in spectrum exists because the spectrum users don't own it in perpetuity or it has encumbrances on its use, which means it can't be freely sold. Strong property rights would solve both problems. Also note that if the signal attenuates quickly, as is the case in 802.11b, and there's no interference problem, then spectrum actually is a non-rival good and it should remain completely unregulated.