The subtitle should be "why free market types are like religious fundamentalists", because he loves using the cute phrase 'market fundamentalism' the way MSNBC uses the phrase 'tea baggers'.
A nice thing about being a Nobel prize winner in economics is when you write books with the same policy recommendations, but use inconsistent arguments in each book, you still get the front table at Barnes and Noble. FreeFall makes the standard talking points you hear on AirAmerica, MoveOn.org, or Noam Chomsky:
1) every economy needs more regulation and higher taxes, especially on the rich
2) He, along with everyone who agrees with him that markets are irrational, predicted the 2008 financial crisis
3) the Community Reinvestment Act and similar government programs had absolutely nothing to do with the crisis.
4) the bubble was fueled by Greenspan's easy money from 2002-2007.
Why do we need more taxes and regulation? Well, according to Stiglitz, the rich are generally immune to incentives, probably parasitic, and generally monopolistic. Like Russian Kulaks, Jews in pre-War Germany, Indians in Uganda, the rich are impediments to growth & justice. Yet he also notes a large amount of financial innovation is to skirt taxes and regulations. I agree that many finacial innovations are focused on taxes and regulations, but that's only because these exogenous rules tend to create mutual gains from trade, and so are bad only if you think prices are wrong (ie, market participants are predictably wrong, or fail to capture externalities).
His proof that the CRA and other government housing initiative had nothing at all to do with the housing bubble, is proven thusly: AIG failed, and they didn't issue mortgages, just bought bonds and derivatives. QED. Further, subprime mortgages failed at rates similar to CRA related subprime mortgages. Yet, current law does not allow a bank to charge different prices based on race. To increase the amount of loans in historically underserved demographics--poor people--you have to lower the bar for all loans.
The old and seemingly irrational rules of thumb of having certain levels of wealth, credit score, validation of income, and downpayment, were diminished because of the fact that banks historically had low losses on mortgages, because there was little evidence of large year-over-year aggregate declines in real housing prices. Thus, for example, we went from requiring 20% down in the 1990's, to zero percent down in the bubble (now upped to 3.5% by the US's FHA). Indeed, Stiglitz himself wrote a white paper arguing that Fannie Mae's 2% capital requirement was more than adequate in 2002 (he estimated an expected loss on $1 Trillion by Fannie of only $2 Million--pre hindsight). Indeed, Fannie Mae was one of his examples of beneficial government policy in his 2002 book The Roaring Nineties. Fannie and Freddie have already cost the government $127B, and it's not done. That's 90 Nick Leesons and counting, but the nice thing about being an intellectual is you aren't accountable for for how policies that were aided and abbetted by your arguments actually worked, because if it fails, it wasn't implemented correctly (eg, Socialism didn't fail, rather, Soviet-style socialism failed).
His assertion that he called the subprime crisis is pretty weak. Look in vain for any strong statement by Stiglitz that underwriting for home lending was too easy prior to 2007, and you won't find it. He did reference an argument he made in 1992 that mortgage asset backed securities may be problematic, but this was a hedged statement, and not important enough to reassert over the subsequent 15 years.
In Stiglitz's Globalization and Its Discontents, written in 2002 just after the internet bubble, the big policy blunders where high interest rates, free markets, the 'fear of default' by lenders, and a lack of concern for the poor. He specifically mentions that Greenspan was excessively concerned with inflation during the 1990's, constraining the Clinton's ability to create economic growth. So, the easy money, lack of concern with default, and insufficient initiatives targeted to the poor, all arguments that he now asserts caused the 2008 crash, would seem to be right out of his playbook. Consistency is for non-experts, I guess.
In his 2006 book, Making Globalization Work, Stiglitz praises Japan, Korea, and China, for their high-minded government policies and trade restrictions as an impetus for growth. Now, modern economies all have government of at least 25% of the economy, and many regulations on businesses. Any success could be a result of some governmental interference, which are many. But to be convincing, one would have to do a true cross-country comparison, and like typical theorist his idea of data is anecdote. The Asian Tigers, West Germany's Adenauer, Chile, recent ascent of Ireland are all ignored. As is the fact that China and India moved considerably towards free markets at the same time their growth rate rose, or that Japan and Korea are less regulated than most nations.
Stiglitz's two most prominent papers are papers he coauthored: "Credit Rationing in Markets with Imperfect Information", The American Economic Review (1981), and his “On the Impossibility of Informationally Efficient. Markets”, American Economic Review (1980). He has always been a theorist, not an empiricist.
Grossman and Stiglitz's "On the Impossibility of Informationally Efficient Markets" flows naturally from Grossman's work on getting information into prices, and Stiglitz's obsession with market imperfection (which came from his fawning work on Samuelson's collected works, which tended to emphasize market imperfections). I have never seen an empirical application of this paper. It's often used as a profound proof that markets aren't efficient, for those who think the relevant standard is perfection. If all information is totally transparent, symmetric, and logical, nobody trades securities with anyone, like what Stokey and Milgrom proved in their No-Trade Theorem (1982, Information, Trade and Common Knowledge, Journal of Economic Theory).
What really bothers me about the paper is its pretentiousness, as if they proved something profound. They proved something obvious. Sure the proof is hard (you solve a differential equation and there's a chi-squared distribution), but the results aren't surprising to anyone. A really neat theory should be important (in this case: maybe), succinct proof (no), and slightly surprising(not!). G&S have all these results in the paper that are really obvious, like the greater the noise, the less informative the price system will be, or the lower the utility of the uninformed, etc. Stigler and Hayek basically argued the same idea, that all the theoretical papers that assume 'perfect competition' and perfect foresight basically assume no need for the market: the data sufficient for a central planner is assumed available. Yet, economies need entrepreneurs and businessmen seizing profits precisely because relevant information is parochial and flawed. Decentralized decision-making takes advantage of decentralized information, and profits are the incentive. So, his demonstration the market perfection is never, technically, true, isn't a critique of the market, it's why many free marketers believe what they do.
In their seminal paper on credit rationing, Stiglitz and Weiss (1981) posed the question: “Why is credit rationed?”. In a perfect capital market with all information available to everyone banks give risk adjusted credits to borrowers of different types. Banks choose in a perfect competition the interest rate such that they achieve zero profits in equilibrium. Stiglitz and Weiss by contrast consider an imperfect credit market in which banks cannot observe the types of the borrowers. With two borrower types, that means that a bank does not know whether a safe or a risky borrower is applying for credit. They assume that all borrower types have the same expected return, but riskier projects offer a higher return in case of success (because of the standard debt contract that has limited upside and 100% loss downside), at the cost of a lower probability of success compared to safe projects.
At any interest rate, the expected profit is thus higher for the risky borrowers than for the safe borrowers. Therefore, the risky borrowers are willing to pay a higher interest rate and still make non-negative profit. So it would be effective for a bank to charge lower interest rates from the safe borrowers and higher interest rates from the riskier investors, but because banks lack the knowledge of the risk types, they set a common interest rate for both borrower classes, which yields zero profit. An equilibrium with credit rationing is necessary for banks give credit to both risk types at a common interest rate. Safe borrowers effectively cross subsidize risky borrowers. This is an inefficient outcome because it rations safe borrowers.
Notice this is inefficient because lenders make safe borrowers subsidize risky borrowers, because of ‘imperfect information’. Yet, banks usually do not have zero information on borrowers as Stiglitz and Weiss assume, but rather, imperfect information. Risky borrowers, via FICO scores and down payment (ie, equity investment in the collateral), pay more on average. In contrast, the major government initiatives, such as Health Care programs, charge everyone the same, which is why people don't like these programs when they aren't cross-subsidized by the non-participants (eg, Medicaid). Thus, to the extent charging everyone the same is inefficient, it is much more prevalent in government initiatives than private ones. Yet, because he proved the market is imperfect relative to perfect information, Stiglitz then merely notes this proves markets are inefficient, and thus government is better, a highly dubious inference given the way government works even when lots of information is available (ie, without a pnl objective, or better information, or less irrationality or selfishness by individuals working for the government).
In sum, Stiglitz generates the same platitudes you hear from typical far left-wing types (eg, the market is inefficient therefore dominated by government, the market leads to a race to the bottom). His arguments supporting his policy preferences are inconsistent. His Nobel prize winnung research did not, does not, forcefully prove we should always be increasing government as he assumes (he notes government gave us the internet, biotech, and ‘research and developement’ --i.e., Tang).