We've just seen a classic example of how the threat of entry can impose competitive constraint on an incumbent - Air wars: Air New Zealand slashes fares ahead of Jetstar arrival.
In the airline market, it's clear that actual entry on new routes is not far away, but in the extreme case, incumbents might be held in check merely by the possibility of entry, without entry ever actually happening - the theoretical world of the 'perfectly contestable' market, where incumbents perpetually have to keep looking over their shoulder to check that 'hit and run' entrants aren't on their way.
I have to confess at this point that I get little red dots in front of my eyes when people diss the general idea of contestability. Of course it's true that the 'perfectly contestable' market is a straw man, and no regulator or competition authority in their right minds would rely on it when (say) thinking about approving a merger. But equally the general idea of contestability - which for me looks very like the reverse side of the barriers to entry coin - makes sense. My go-to resource, Viscusi/Vernon/Harrington's Economics of Regulation and Antitrust, says (p164) that "the theory of contestable markets is quite controversial", and that's certainly right, but it also says that "if nothing else, contestability has been instrumental in causing antitrust analyses to reduce their emphasis on concentration and take proper account of potential competition". Right on.
Doing so, however, is no easy matter for a competition authority, especially when it comes to the L-for-likely leg of the LET test: "The LET test is satisfied when entry or expansion in response to a price increase or other
exercise of market power is Likely, and sufficient in Extent and Timely enough to constrain the
merged firm" (from para 3.96 of the Commerce Commission's Merger and Acquisitions Guidelines). It doesn't want to be a soft touch - waving through every merger because it thinks someone or other will turn up sooner or later and compete effectively with the merged entity. As the guidelines say (para 3.98), "The mere possibility of entry or expansion is insufficient". Equally though it wouldn't want to go to the other extreme, either. There will be occasions (like this airlines one) when it can clearly see who's coming, and when, and how much. But there will also be occasions when there isn't a competition problem, even when the authority won't be able to tell exactly in whose colours the planes will be painted.
Isn't it strange, by the way, that every man and his dog can see the immediate and positive connection between competition and good outcomes for consumers when it comes to air travel, yet many of those same people would die in a ditch to stop the same competitive pressures bringing us better outcomes in health or education or infrastructure?