Book Review: Too Big to Save: How To Fix The U.S. Financial System

Knight Ridder/Tribune

Robert Pozen's Too Big to Save: How To Fix The U.S. Financial System (Wiley/Nov 2009) is a story of what happened, an education in 'modern finance,' a criticism of the bank bailouts, and a number of policy recommendations. Despite being one of seemingly innumerable books on the 'Great Recession of 2008,' Pozen's book will likely provide important insights for readers.

Two of those insights are: 1) Prepayment penalties cause problems for mortgagors wanting to refinance. (Honest, I hadn't realized that!) 2) Fannie Mae and Freddie Mac held their own mortgage-backed-securities (MBS), along with the associated prepayment and default risks, so they could also function as hedge funds for their investors.

"Special Purpose Entity" (SPE) is a term often used, but only briefly explained elsewhere. Too Big to Fail provides the needed background. Wall Street banks avoided normal capital and disclosure requirements by creating small shell companies (SPEs) that would purchase pools of mortgages from brokers, regional banks, and thrifts, repackage them as MBS -- usually sliced into varying tranches of risk, and then sell them to institutional investors around the globe. The SPEs would also hire mortgage servicers to collect payments and direct those funds to appropriate tranche holders. SPE assets and liabilities are off-balance sheet to the holding banks -- the assumption is that most of the risks are assumed by other investors in the SPE. Regulations required doing so only if an entity held over 5% of voting shares. Most SPEs were 'orphans' and did not appear on anyone's balance sheet. However, in 2007-08 billions of SPE assets returned to bank balance sheets because of credit guarantees or 'liquidity puts' made by the issuing banks. (A liquidity put occurs when the bank is obligated as a last resort to buy the SPE's short-term commercial paper used to finance the assets.) Another advantage of using SPEs is that banks could book profits upon selling loans to an SPE, and not wait for the sale to a final customer. Finally, SPEs are not required to file annual reports if the number of investors and sponsors number less than 300.

Other authors believe Denmark's mortgage laws offer a safer and less-speculative model for the U.S. -- Pozen does not believe this change would be accepted, but also provides a good summary of their operation. Danish mortgages may only be issued by mortgage credit institutions (MCIs), limited to that function. Each MCI retains most of the risk of losses on their mortgages. All mortgages use a standard format, and cannot exceed 80% of the home's value. Documentation and appraisals are verified by an independent Mortgage Institute. Mortgagors retire their mortgages early by purchasing the same amount of bonds at market price -- thus, no prepayment risk to the original mortgage-holder.

Credit default swaps (CDS) played a key role in the 2008 financial collapse. CDS buyers paid a premium for effectively insuring against a debt default, and received a lump sum payment if the debt instrument is defaulted. CDS were not regulated as insurance, with associated reserve and capital requirements, because buyers did not need have an insurable interest. Most buyers were hedge funds -- John Paulson made a personal profit of $6+ billion doing so with his fund. AIG was a major CDS issuer, and was eventually bailed out with $85+ billion in taxpayer funds. That action is now being investigated as a 'sweetheart deal' that paid Goldman, Sachs and other 100% of their AIG assets while other entities were accepting far less. CDS are also being investigated for possibly allowing buyers to manipulate the market downward for their own benefit. The methodology supposedly involves starting by buying CDS for the debts of a company that uses short-term financing on a permanent basis (most financial companies), shorting the stock and thereby forcing the price of the CDS upward to panic lenders into cut off lending to the firm. Ergo, a default that pays off both the short-sellers and CDS holders.

Bond rating agencies earned fees at least twice that from ordinary mono-corporate and municipal bonds ($30,000 -- $300,000) when they instead rated MBS. Rating firms also sometimes received fees to give advice on how to structure complex deals -- this conflict of interest is now banned by the SEC, similar to Sarbanes-Oxley vs. accounting firms. Competing for MBS rating business reportedly led to overly generous ratings -- Pozen suggests raters be approved by the SEC. Bond raters' prior experience being mostly limited to single-issuers played a major role in their ignoring the highly systematic (correlation) risks of mortgage credit-default obligations (CDOs), instead assuming the creation of desirable risk diversification. (CDOs are bonds backed by pools of mortgages that generate tiered cash flows from the mortgages, based on their risk levels. Every tranche carries a debt rating ranging from the highest credit rating of AAA (0.02% annualized default rate) on down to junk bond status. (AAA-rated securities were particularly attractive because they only require banks to carry half as much capital in reserve.)

Derivatives are also at the forefront of blame for 2008 downturn, but never explained in books I have read that were targeted for general readership. Here, Pozen makes his greatest contribution, though indirectly by citing an excellent 2008 Harvard Business School paper on the topic ("The Economics of Structured Finance," by Joshua Coval et al) that details how CDOs and synthetic CDOs-squared seemed to 'create yield out of thin air.' (Pozen's explanation was unsatisfactory.) The objective is to create some tranches that are safer than the average of the underlying asset pool; the problem is that high risk correlations between mortgages made CDOs far riskier than advertised. In addition, they carried an unrecognized magnified sensitivity to errors in estimating the original mortgage risks (common because of limited experience with subprime mortgages), making tranches rated AAA much more likely to default than was believed. Yet, by 2007 there were an estimated 37,000 structured CDOs with top ratings, vs. less than 1% for single corporation offerings.

Coval et al begin by explaining that tranches are prioritized in how they absorb losses from the underlying portfolio, and then show how two tranches against a simplified two mortgage portfolio, each with a 10% uncorrelated default risk, results in the senior tranche (defaults only if both bonds default) having only a 1% change of default, vs. 19% for the First (junior) Tranche. Using larger numbers of assets in the pool allow a greater proportion of the tranches to acquire a risk rating superior to the average. Using three bonds to illustrate, they report the senior (3rd) tranche has a default rate of 0.1%, the middle of 2.8%, and the junior rates a 27.1% expected default rate. This 'magic' can also be extended by reapplying the process to junior tranches created directly from underlying mortgages.

Combining junior tranches from two two-asset pools creates a senior tranche with default probability of 3.6% and junior tranche risk of 15.4% , and is known as a CDO-squared. However, perfect positive correlation increases the senior tranche default rate in all instances back to that of the underlying average -- 10%. Finally, they also note that many CDOs and CDOs-squared were constructed from subprime 'non-conforming' mortgages (violating borrower credit quality standards), and underestimating the underlying risk. At least two major rating agencies later realized and admitted that their models were inaccurate -- eg. assuming continued increases in housing values, underestimating risk. Clearly, risk premiums were under-priced prior to the crash and the housing boom benefited. The 'good news,' say Coval et al is that "investors are now reluctant to invest in securities they do not understand," and the likelihood of the preceding processes recurring is forever diminished.

Back to Pozen -- we've had three major finance crises in a ten-year period: 1) The 1997-98 Asian financial crisis and LTCM bailout. 2) The Dot.com bubble of 2000-02. 3) The subprime debacle of 2007-08. Over-leveraging is an underlying problem. It's easy to say 'get out before the bubble bursts,' but Pozen also points out that respected Yale economist Robert Shiller warned U.S. housing prices were grossly inflated in 2003, yet the gains kept coming for three more years. Pozen suggests resisting protectionism (no rationale offered), higher capital requirements for banks, and not allowing big financial institutions to get any bigger. Finally, in case you thought the bailouts given by Treasury were overly generous, Pozen agrees with you, and documents that conclusion by comparing terms given for Buffett's $5 billion and Treasury's $10 billion. Both were negotiated in the same time frame, yet, despite Treasury's $10 billion commitment being double that of Buffett's, guess who got the better deal from Goldman!

Bottom-Line: Too Big to Save provides good background on the technical aspects involved in the current crisis (even better if the referenced Harvard article is also read)!

Robert Pozen is Chairman of MFS Investment Management®, which manages over $150 billion in assets for individual and institutional investors. He currently is a senior lecturer at the Harvard Business School and was chairman of the SEC advisory committee on improving financial reporting, 2007 through 2008. In 2001 and 2002, Pozen served on President Bush's Commission to Strengthen Social Security. In 2003, he served as Secretary of Economic Affairs for Massachusetts Governor Mitt Romney. Pozen was also formerly vice chairman of Fidelity Investments and president of Fidelity Management & Research Company.

Book Review: Too Big To Fail By Andrew Ross Sorkin

Loyd Eskildson is retired from a life of computer programming, teaching economics and finance, education and health care administration, and cross-country truck driving. He's now a reviewer for Basil & Spice.

----

More book reviews

More book news

Copyright 2012 by Knight Ridder/Tribune