In addition to Dimitri's comment and user35980's answer above, the following comes to mind: so much USD has been printed by the FED already and now with Democratic majority likely, more fiscal stimulus and even more FED balance sheet expansion is being priced in over the past few days.
Because the market is awash with USD liquidity, everyone who sits on USD is trying to somehow put their money to work. One way of doing that is swapping USD for MXN via FX swaps and Xccy basis swaps: that way the institution that lends out USD will be receiving USD Libor for the duration of the swap whilst paying MXN TIIE rate plus the basis (if the basis is positive) or minus the basis (if the basis is negative) on the MXN that it borrows.
(The question that then arises is: what will the institution do with the MXN on which it then sits instead of USD, whilst having to pay the MXN TIIE rate plus or minus basis on it (depending on whether the basis is positive or negative)? If the institution has full legal entities in both USA as well as Mexico, it can raise the USD domestically in the US money-market at sub-Libor cost and lend it out at Libor (thereby making a bit of money on the USD leg alone). It has also access to the Cenral Bank of Mexico deposit facility, which pays around the MXN TIIE rate: so if it can raise MXN at sub-TIIE rate and then deposit the MXN at near TIIE rate, it will make money on the MXN leg too.
Even if the institution doesn't have access to the Central Bank of Mexico deposit facility, it can invest the MXN into Mexican bonds or lend the MXN in interbank market and still probably make profit, as long as the USD-MXN basis is sufficiently negative).
In any case, if the above hypothesis is true and many market participants are trying to do a similar thing, it should make the USD-MXN basis less negative (or more positive, if it's already positive) (I haven't been following the USD-MXN basis but I suppose when you say "the curve has been decreasing" you mean the basis is becoming less negative? If it's in fact becoming more negative, then it signals flow in the opposite direction: more demand for dollar vs. MXN, in which case my hypothesis is wrong).