Opposing Goals as a Method of Regulation

The rating agencies that failed so spectacularly during the recent housing boom have a serious conflict of interest problem: if they don't rate bonds highly, the bond owners will take their business elsewhere. Since they're funded by the companies whose products they rate, this is a serious problem. There are many proposed ways to address this, such as changing who funds them or basing rewards on prediction accuracy, but I was also thinking that requiring them to offer another service that creates a bias in the opposite direction could also work. If a rating agency were required to offer insurance on a product based on their own rating, then they'd have a powerful financial stake in its accuracy. Of course they'd need to have enough capital requirements to ensure their insurance was viable, and the two competing areas of business need to be roughly matched in value. This is fundamentally different than having third party insurance entities, because the rating agency has complete knowledge of the internal mechanisms used to arrive at their ratings. I feel a fundamental overall of the entire system is more appropriate, but still found this idea rather intriguing.

I couldn't think of a good picture to accompany this. Any suggestions?