My gut says "statistical fluctuation".
Look at the maps in Figure 1. The largest cold spot happens to lie just to the right of center, and the largest hot spot happens to lie just to the right of that. The hot spot roughly lines up with one of the two hot spots in the quadrupole plot (Figure 14a) and the cold spot roughly lines up with one of the two cold spots in the quadrupole plot.
Furthermore, that hot spot and cold spot also roughly line up with hot and cold spots in the octopole plot. Since the alignment of the quadrupole and octopole moments is dominated by those two features in the CMB map, perhaps it's not surprising that they roughly line up.
What are the odds of a random fluctuation of that size, in the absence of any preferred direction in space? I haven't the foggiest idea. I suspect it's not negligible, but I could be wrong.
I'll go ask Max.
From the paper:
What does all this mean? Although we have presented these low multipole results merely in an exploratory spirit, and more thorough modeling of the foreground contribution to l=2 and l=3 is certainly warranted, it is difficult not to be intrigued by the similarities of Figure 13 with what is expected in some non-standard models, for instance ones involving a flat "small Universe" with a compact topology [...] and one of the three dimensions being relatively small (of order the Horizon size or smaller).
The Times article certainly did miss the point of the paper. The quote from the paper could hardly have been more circumspect.