Do accounting changes make economic changes?

This seems like old non-news to me, but here is an article on Bloomberg regarding AIG's credit default swaps and the ongoing risks of such things.  

An excerpt:

The $192.6 billion figure for the swaps is comprised mostly of $99.4 billion tied to corporate loans and $90.2 billion linked to prime residential mortgages, the insurer said in a May 7 filing. The combined total was reduced from $234.4 billion on Dec. 31.

Most of the home loans tied to the European swaps are first-lien mortgages for owner-occupied properties, the insurer said in March. The other transactions include secured and unsecured corporate loans.

The fair value of the derivative liability was $393 million as of March 31, compared with $379 million on Dec. 31, according to AIG filings.

But I think the new news part of it has to do do with the reclassifying of the reporting of these securities as "risks".  I could be wrong as the article is a bit shy on nitty-gritty-detail.  Missmc is confused because the last she heard on the topic was that the diminished value of these securities did not have to be accounted for as losses by the banks if they were held to maturity.  And that might be why Citigroup is still alive.  Murky murky murky complex changes.  Do accounting changes make economic changes?

When the loss is lost (but you can't say that), it is risk.  It does all make sense, but, only theoretically.