I've just recently started following Paul Krugman's blog and I find his style of writing to be very clear. In this blog post he talks about a longer article he published in the September 2nd issue of the NY Times magazine. I just finished reading the longer article and it too is a very clearly written piece.
In the blog post he talks about natural (non human) systems that can be compared to economics. And references a piece on the Discovery web site that talks about his article, the Discovery piece is titled "Mistaking Beauty for Truth".
Paul Krugman's comment is that rather than use the two concepts General Relativity and Newtonian dynamics as suggested in the Discovery article a better concept would be Meteorology.
Meteorology can be thought of as a non linear dynamical system that is chaotic at it's core with things call attractors. A chaotic system with attractors is one that resist prediction but still stays 'inside-the-lines' and does not go wildly astray.
I think the day to day operation of the Economic System could well be modeled by such a thing.
However the economic system is susceptible to 'opportunistic blooms' which do naturally arise in all sufficiently complex systems. The main characteristic of an opportunistic bloom, and its inevitable collapse is a growing (blooming) bubble and then a bursting bubble.
So my point is you can model economic markets and your model must contain the concept of a bubble (or bloom). This argues for something folks call 'smart-regulation' of the financial system where you do not publish some set of fixed rules and set some group of fixed policies to enforce those rules but instead have flexible rules and enforcement that are themselves policed by being transparent and constantly audited.