Ah, Mannich, I knew it had a name but I'm losing my touch.
Decarboxylation leads to the loss of only one carboxylate, so after that you have intermediate form known as an alpha-carboxylate dianion. This is a pretty basic compound (usually it is prepared by reacting a carboxylic acid with LDA or even stronger bases). A rapid equilibration will happen, but I think that the small amount present will react readily with the Mannich reagent, shifting the equilibrium. After that, another acid/base shift will lose the diethylamine and convert the whole thing to an exomethylene.
At least, that's my idea. I might be wrong.