He mainly argues that George Soros understands, and explains in his book, that markets can behave irrationally. They do not always operate according to strict theories or rationality imposed upon them by economists who use computer models to make predictions.
Financial markets perform two essential roles in the economy: (1) they take money from those with no immediate use for it, such as people saving for retirement and the hereditary rich, and put it into the hands of firms and entrepreneurial individuals with productive investment ideas but a shortage of cash to finance them; (2) they allow individuals and institutions to reapportion risk to those more willing to bear it. If Wall Street didn't exist, another method of allocating savings and risks would have to be found. One alternative is diktat, but the history of the Soviet Union and other Communist countries amply demonstrated the difficulties involved in centralizing economic decisions.
The great advantage of a market system is that it draws on information from throughout the economy and translates it into public signals—prices—that investors and firms can react to. Earlier this year, investors woke up to the fact that Detroit had ignored the threat of dwindling oil stocks and had bet its future on gas-guzzling SUVs: the stock prices of American car companies plummeted, making it much more expensive for them to sell equity in their corporations. Toyota and Honda, which had invested heavily in smaller, more fuel-efficient vehicles, have seen their stocks hold up much better, enabling them to raise funds cheaply. Nobody planned it, but in this instance the market rewarded foresight and innovation.
For financial markets to allocate resources to their most productive uses on an ongoing basis, the price signals they send must be the right ones day after day after day. Is this a realistic goal? A typical investor following the Dow's gyrations on CNBC or Yahoo Finance might be tempted to say no, but then the typical investor doesn't have the benefit of an economics Ph.D. from the University of Chicago.
Soros's theory of "reflexivity" is especially interesting and Cassidy even made up a helpful little diagram, but this isn't, unfortunately, in the online version of the article.