I had forgotten of its effect on the reward pool. True.
The psychology is a really interesting bit of it. The airdrop wouldn't immediately change the market cap, but it would almost certainly result in the price going down as people sold off their new "free" stuff. Wouldn't be too quick if it were airdropped as SP, I suppose. But its eventual effect would almost certainly be to make the market cap drop.
The burn would drop the market cap, but then possibly make the price rise. But I'd bet it wouldn't really make the price rise much, because the utility of each STEEM wouldn't actually change at all (because @steemit isn't using their stake for voting). So it could actually make the price of STEEM drop, because now each STEEM would be controlling the same fraction of a now-smaller reward pool.
Interesting.
Have a look at @misterdelegation. It's an account controlled by Steemit Inc that has delegated out 8 million SP, with several more million to come. This, plus freedom's 7.5 million SP, and other delegation services means nearly 20 million SP that has hit the reward pool. The end result is net influence is 1/3rd what it was a couple of months ago, per Steem Power. The price is inversely correlated to this change, in fact. Of course, not causation, at all - that'll be the booming market and greater adoption of Steem in general. Indeed, the whims of crypto whales outweighs everything else. So if a burn or airdrop is favourable to them, the price will rise regardless. As for which method of the above (and others) they like best, I don't know.