The problem is that, in many cases, risk hasn't declined at all — it's just that we experienced a sequence of lucky draws where nothing bad happened. So when we respond to our perception by topping up our risk taking, we’re actually exposing ourselves to more risk than we realize. When a bad event happens, we freak out and then go to the opposite extreme. Then after our fear subsides, we realize that the opposite extreme isn’t really appropriate either, so we start the cycle all over again.
The adaptive nature of human cognitions also means investors are subject to misperception from time to time. After a period when we don't experience any negative events, our perception is that the risk has declined. So naturally, we respond by taking on additional risk because that’s what we were comfortable with before the perceived decline.
Think of seat belts. I wear mine all the time even though I haven't had an accident in a long time. A typical response to a great safety record might be, "Gee, I'm a pretty skillful driver; I don't need seat belts anymore.” But we don’t think that way because we've learned that we can't predict when we're going to have an accident, Even though we may be very skillful, what about other drivers on the road who aren’t? This is an example of how humans have adapted to our risk-perception bias.
Most of us haven't done the same with financial markets. Many of us feel like, "Gee, the stock market just went up by another 5%. I've been making money for the last three or four years. The market is just not that risky. I can afford to take on more risk than what I'm doing now.” And that's the wrong lesson.