In an 'efficient' market all future risk is included in the analysis of current value meaning that if it was an 'efficient' market this would actually not be a problem. However it is just one of many* ways that markets are not in reality 'efficient' in the economics sense as the assumptions required to prove them just do not match up to the real world.
The EMH claims that all future time discounted risk known to market participants is included in the analysis of current value.
If you want to show the markets are not efficient, go achieve excess risk-adjusted returns. The sole claim of the EMH is that you can't do that. The EMH doesn't claim that market participants will not take risks, nor does it claim that investors/customers will not apply time discounting to those risks.
Well, yes. The economists know that.
You might have noticed a little recent global credit crunch which was not predicted by most economists but was predicted and modeled by the author of the book I linked who is an economist (so I don't lump them all together) but the overall level of the state of economics is so poor as a discipline at understanding the overall economy it should be embarrassing to them.
I am genuinely interested if someone has a critical analysis of the Debunking Economics book that points out how and where it wrong but when I last looked I couldn't find any serious attacks online (minor nitpicks only) but I think it is largely being ignored by those who disagree so I haven't found their counter arguments.
If you take this as your starting point for critique (the weathermen didn't predict Weather Event X!!!), you will always win the argument because you're beating up a strawman. No economist has ever seriously claimed specific predictive power.
All an economist can give you is generalised statements of causality, most of which will be unobservable. Steve Keen was not the first to point this out. Quite a few economists from various schools have picked flaws with general equilibria models of macroeconomic phenomena (ie, using calculus to describe people, markets and countries).
I don't expect a model to tell me that the markets will crash tomorrow but I would have expected widely use models to indicate that we were in dangerous period in 2005-2007 and that the upwards path was impossible to sustain over a 10 year period.
Predicting a crash immediately before it happens is not so difficult. Lots of economists were clanging the alarm bells all through the mid-00s. Predicting exactly when and exactly what the trigger would be? Basically impossible. Economists don't do that (it's left to advisors, pundits and newsletter salesmen).
So there were a few but they generally weren't in the mainstream of economics (from Krugman to the Chicago School) which generally did a very bad job. Roubini did call it but none of the others on the list I linked to were people I had heard of before 2008 (but I haven't formally studied economics).
Many of the common models taught and used never indicate crashes, this sort of thing should just be thrown out. Most of them also don't really include the financial industry (including debt).