One explanation for bubbles is the greater fool theory. People figure they can sell the stock to some greater fool, who will pay more for it.
Another thing that can cause bubbles: People don't always have great information about what's going on. So they just follow the crowd.
And once bubbles start, it can be hard to put the brakes on. If you think home prices are going up, you can buy a house, or two houses. But it's hard for the average investor to bet the other way — that home prices are going to fall.
So for all these reasons, we see bubbles and crashes, in the lab and in the real world. The people in Williams' experiments sometimes buy stock for more than it could ever pay out. This gives us a definition of a bubble: When the price of something rises way above its fundamental value.
That definition, though leaves one in somewhat of a quandary when it's applied to gold. Since there's no way to calculate gold's fundamental value, it's conservative in the accountant's sense to assume that gold has none because it pays no interest and dividends. Thus, a conservative accountant would have to conclude that gold has been in a bubble for at least four thousand years.
Clearly, this line of reasoning is a reductio ad absurdum. The definition of "value" is clearly too narrow, as can be seen when it's applied to any valuable that doesn't generate a return or stream of services. Yes, this includes money: it too pays no interest or dividends. The same criterion used by narrow-minded stock jocks to claim that gold has "no intrinsic value" can be used to claim that any money, fiat or no, is worthless. Exactly the same line of reasoning: just substitute any currency for gold and run through to the answer.