With the stacks you described, I'm pretty sure it is not +EV to take a coinflip in this situation. You can keep pushing on the blinds and putting the pressure on them until the bubble bursts.
As set out above, it's +$EV, even though it's slightly -chipEV.
That assumes you're definitely coin flipping and villain is definitely calling though (ie: villain's range is 22-TT and he'll never fold to a shove).
If you make villain's calling range 22-AA / AJ+ and allow for the fact that he'll fold some of the time, which is a much more realistic scenario, the figures will change.
We're a bit behind against their calling range now, but we gain some equity back because they're folding some of the time. It's way too early in the morning on this side of the world to be doing this, but I'll give the figures a bash.
I haven't got PokerStove on this computer, but let's give ourselves 40:60 against the villain's calling range, and say they're going to have that range and call us 80% of the time:
40% of 80% is 32%, so 32% of the time we get called and win third-place money
60% of 80% is 48%, so 48% of the time we get called and knocked out
20% of the time villain folds, and we win the 1050 in the pot
Add those together, and 52% of the time we either win the pot or cash, and 48% of the time we get knocked out.
If they only call 65% of the time, however...
40% of 65% is 26% of the time we get called and cash
60% of 65% is 39% of the time we get called and knocked out
35% of the time villain folds, and we win the 1050 pot
So in that case, 61% of the time we either cash or win the pot, and 39% we get knocked out.
On those figures (anyone who's got the actual equity figures for AJ vs 22-AA / AJ+ feel free to update the above), even if villain calls four times out of five, we're still ahead. And if they fold just a third of the time, we're further ahead.