From my understanding of DFT that makes sense but it seems like in that case you are using the wrong tool. Approximate theories are good when the approximations are good. If your phenomena is strongly correlated, throwing out those correlations is bad.

I have no idea if the use of DFT here is appropriate or not.