Children have surprisingly perceptive risk intuition. They know when parents are serious about a threat and when their threats are empty. And they respond to the real risks of parental punishment, not the inflated risks based on parental rhetoric... There have to be real consequences.
Hah, just had an argument with my mom over this the other day - I'm 23 (living separately), my sister is 12 (living with mom), and her and my mom were visiting. I had just emptied a piggy bank and my sister (12 yo) was counting the coins. First my mom said "oi, those are dirty, you shouldn't touch them", and was promptly ignored. She kept escalating her rhetoric, finally describing the unimaginable diseases my sister could pick up. Still ignored, until she raised her voice. At that point, I stepped in, and the argument turned into whether or not I should question my mother's authority in front of my sis. Whether or not my sister learned it, I still think that questioning authority calmly and firmly is a great lesson. :)
The concept of a "risk thermostat" is interesting -- people seem inclined to maintain a more or less constant level of risk. I've noticed an effect that interacts with that: my risk inclination seems to have some initial static friction that changes to a lower level of kinetic friction once I'm moving. Example: some years ago I was feeling a little hesitant to try snowboarding because it seemed "risky" (I knew I'd fall a lot). The first time I tried it, I was having so much fun that I pushed myself well into exhaustion and took some crazy tumbles that didn't stop me from getting up and riding more. I suppose my initial response might just have been an underestimation of the "reward" part, rather than an incorrect estimation of the "risk" part.
Aside: When I used to ride a motorbike, I found the combination exhilarating of the more severe consequences of an accident, with greater ability to be aware of it and greater ability to deal with it. Unlike a car, a motorbike has a full field of vision, and you can also see over cars because you are higher. It's also more agile: you can turn, brake and accelerate more quickly, it being lighter; and of course you can pass through smaller gaps to get out of trouble.
It's incredible fun.
But I limited my riding to a year, because you can't avoid some accidents, and the consequences are severe: a friend worked in prosthetics, and half of the recipients are motorcycle riders.
A friend of mine was in hospital for a quite serious motorcycle accident (as in "lost half a leg" type serious), and was in a ward with 3 other motorcycle accident patients. The four of them had a lot of fun, whenever any visitors started questioning their intentions about continuing to ride, by pointing out the the _next_ ward had four guys all of whom had fallen off ladders.
Yeah, motorcycling is dangerous - you're around 10 times more likely to be hospitalised or killed per kilometer travelled according to local (Australian) stats - but _life_ is dangerous, and not having fun isn't living...
One additional thing is that the punishments have to be realistic. If your policy is to fire someone who is lax in security, your managers are likely to cover up infractions so that they don't have to fire their employees.
Children have surprisingly perceptive risk intuition. They know when parents are serious about a threat and when their threats are empty. And they respond to the real risks of parental punishment, not the inflated risks based on parental rhetoric... There have to be real consequences.