By their very nature, networks - of customers, Web sites, weather patterns, evolutionary systems, or economic competitors - are subject to random and unpredictable movements.
Networks of interrelated things, whether we are talking about the Web or a collection of stock and bond prices, often grow in a kind of "cascading" pattern. For instance, because one person buys a stock, others think it must be valuable, too - so they buy it. Influential customers become more influential at a faster rate than those who aren't already influential. The most visited Web sites gain visitors faster than others, and wealthy people become wealthier at a faster rate than others. The reason this kind of cascading occurs is easy to understand, but the effects of cascading like this can be quite random. Cascading means that two systems could be set up with nearly identical initial conditions, but even very slight, almost undetectable differences between them will inevitably escalate into major differences after just a bit of time, creating what has come to be known as the "butterfly effect."
All reality is, in fact, time-dependent. Everyone's life today is a culmination of all the sometimes random events that have shaped it. Think about how you met your spouse, for instance, or why you chose your college major, or how you landed your first job, or how you just missed an important flight, or how you got rear-ended by someone. Each of these events is memorable to you and will have had an important role in shaping your life, but many of them could easily have turned out differently, in which case you might have a completely different life today, with a completely different perspective.
One consequence of network structure is that periodically a whole system will go through a very significant change for no obvious reason. For instance, while current thinking is that the dinosaurs and many other species were apparently eradicated by a catastrophic meteor strike, it is in the nature of evolutionary networks that periodic massive species extinctions will occur with a certain predictable probability, with or without massive externally generated disasters. This is simply the way complex systems behave.
As technology brings increasing "connectedness" to our world, the volume and speed of the various feedback loops that link human beings and all economic actors are increasing. Our system is speeding up, in other words. It is cycling into different states at a faster pace, so the butterfly effects of minor changes and perturbations will be felt even faster and more significantly in the future.
While we can all try to pin the blame for the current financial crisis on various factors - deregulation or over-regulation, greedy business executives or cynical politicians - the fact is that any economic system as complex and interconnected as ours will almost certainly go through occasionally comprehensive and wrenching changes.
This doesn't mean we shouldn't try to repair and upgrade our economic system once in a while, in order to try to avoid the catastrophic consequences of these ups and downs. But it does mean that no matter how good our efforts are, they will be temporary at best, and the next cycle hit us with just as much surprise as the current one has.