After a bit of Googling it appears that anodization greatly increases the emissivity of the aluminium surface, depending on the thickness of the oxide layer of course.
Indeed, your link points to between 4 times and 17 times better emissivity after anodization.
How and why it does this is still a mystery to me, at least. As you rightly point out, it would have been more effective to anodize the heatsinks black.
You have Stefan-Boltzmann Law of Radiation
(it's in the middle of that long page) : Energy radiated per second: H = e * A * omega * T^4
You want to increase H, keep T low (it's the temperature of the heatsink), omega is a physical constant. So you are left with A (the radiating surface) and e which is the emissivity
In our practical situation, you can't expand A too much size-wise, because it has to fit in the PSU. But you can increase the surface locally by giving it a rough finish. If the surface is rough, it increases the total radiating surface (/\/\/\ rather than ____ )
As to the emissivity, if the surface is closer to a perfect black body
, it increases the heat radiated. With temperature found in a PSU, perfect applies to the ability to dissipate IR. I'm not sure if that's still regular black (which is perfect in the visible spectrum). Emissivity equals absorptivity if you believe Kirchhoff's law of thermal radiation, and we all know from day-to-day practical life, that under the visible light of the sun, black aluminum would be hotter than pink aluminim. So it's maybe a safe assumption to expand that to the IR spectrum and say that black would have been better than pink to radiate heat.
Hence the conclusion : anodized was good, but black finish would have been better than pink finish.
The question is to know if the pink is a painting or is anodized aluminium (I don't know the color it's supposed to look like). Is the pink easily removable, or does it seems embedded in the aluminum ?
Maybe they couldn't be bothered to paint it black, or the strange factor was supposed to make people believe it's better that way.
Said the marketroïd : "Black is so boring you know, you can't hype a 170$ PSU with black. Put pink in there, stupid engineer !" /rant
In your pdf, article 3.1 page 5 is especially important. They explain how a diffuse surface would be good for improving radiation. A rough finish does that a bit, but anodizing achieves that even better.
They conclude the paper with
"Further research should be conducted to find the emissivity of aluminum surfaces that have received other treatments, as for instance painted aluminum profiles." Maybe someone can find that paper, because experiment is always better than some theoritical rambling.