[lordnacho]

If not an automatic sunset, I would propose a different rule: for every law, we stipulate what statistical measures we thought would be improved by the law over some time period. Like "we think if kids are all fed at school, pass rates will increase 5%". Then check after 5 years to see if that happened, and we can decide what to do.

At the moment every law has an imprecise expected outcome and we never go back to think about whether the intervention worked.

I'm not saying use the stats mechanically to cancel the law if it fails, just that there ought to be a post-fact consideration of whether something worked.

[candiddevmike]

This is great. Have some kind of "definition of success" with laws that will be used in the future to gauge their effectiveness. If the law isn't having the desired outcome, it gets repealed.

Truthfully, outcome based legislature would be quite a bit more readable too, as the outcome bit will spell out exactly the intention of the law.