But I believe difficulty lags price by a month or so.
There was a pretty clear pattern (difficulty rises lagged behind price increases by three to eight weeks) when GPUs were the primary technology used. Currently the rise thanks to ASICs will rise regardless of the direction and rate of change of the exchange rate.
But for sure, there's no reason a higher difficulty yields a higher exchange rate unless for some reason there was not enough hashing occurring where the exchange rate was depressed due to concern that it wasn't high enough to provide sufficient protection. Dan Kaminsky thinks we might be there again simply because a well-funded single supplier of ASICs can easily reach 51%. Others think he is blowing smoke as who has five or ten million dollars to blow trying to prove his point. Who knows.