It's mostly ideology with mathematical excuses.
At best, it's non predictive modeling.
Too easy to dismiss that much work. And don't blame economists for the crisis. That's politics. There are a lot more lawyers in politics than economists...
We don't discredit the whole of psychological science for the failures of a load of social pop-psychologists. Perhaps, perhaps we can discredit the whole of nutritional science. And even there I'd say we have learned some but that the financial incentives are so biased against proper learning that it's hard to blame the field as a whole.
Probably the best piece of economics I've seen is Coase's theorem. While we base our laws on this idea, it's still a shaggy dog story, and basing it on dialectical materialism, the divine right of kings or Schmitt's friend/enemy principle would work just as well. Citation: places which never heard of Coase's theorem work/worked just as well.
That’s too facile. I’d say, insofar as it is a science, it’s not about the real world, and insofar as it’s about the real world, it's not a science.
But even that doesn’t give it enough credit.
And then on the question. Perfect information is false and there are degrees in the definition of how efficient markets are, with the most common one approximating 'somewhat' on the not too short timescale. Mileage varies between commodities, stocks, your staple teaching example of the market for used cars (lemons) going towards the more micro models with network effects etc.
Institutional economics will teach you that no market functions without surrounding institutional arrangements and that there is some relationship between the nature of the product and the optimal 'institutional burden' (my words). Mechanism design is the subject where they try to design optimally structured markets.
Economics is awesome and anyone teaching only perfect information even to undergrads or business students should be fired. Ceterum censeo