January 11, 2009

Honesty: The First Step to Progress

I despise intellectual dishonesty as counterproductive and potentially harmful. Here is a case in point. I was recently asked, as an officer of a small financial institution, to write an article for a publication regarding the state of the financial markets. Since I'm a former investment banker who used to structure mortgage related investments and also provided consulting services to the mortgage industry, I had a lot to say on the subject. Once the article was written, I presented it to my public relations people, where it was killed. The reason? No one could argue with the accuracy of the content. Neither could anyone deny that it was an insightful summary of recent developments. The article was killed because it made what were considered disparaging remarks about regulatory and rating agencies. Since having an investment-grade rating is important to a financial organization, no one wanted to risk the ire of the companies that provide them.

I can understand the sentiment, but there's a serious problem with an inability to speak the truth. In short, if we can't be honest about the problems we face, how can we solve them?

At any rate, here's the article.


The subprime market’s collapse didn’t surprise most experienced lending officers. They understood that extending credit to borrowers who had little means to repay and insufficient collateral resulted in foreclosures waiting to happen. The real surprise was that such high-risk financings became commonplace at all. Given the subprime market’s history of excess, a knowledgeable observer might ask: How did it survive for as long as it did? By answering the question, we gain clarity to the underlying causes behind today’s credit crunch.

Easy Money Years

For the decade leading up to 2004, the subprime market grew from $35 billion in originations to over $510 billion, a compounded annual increase of nearly 28%. Much of the growth was an indirect result of the Federal Reserve’s reaction to the tragedy of 9-11, which had fueled the unease of financial markets already suffering from the bursting of the Dot.com bubble. To keep the US economy from dipping further into recession, the Federal Reserve embraced a monetary policy that resulted in a steady decline of the Fed Funds rate from 6.5% in January of 2001 to 1.0% in July of 2003.

In retrospect, the effect of that monetary policy was more impactful than anyone had anticipated. Due to its influence in reducing market interest rates:

  • Housing suddenly became more affordable as mortgage rates and resulting average monthly payments fell.
  • Real estate values rose as housing demand increased.
  • A combination of increased property values and low interest rates encouraged homeowners to borrow against home equity. Cashout refinancings added $600 billion to the US economy in 2004—more than twice the value of that year’s tax cuts.
  • Much of the added liquidity was used to purchase second homes and investment properties. The percentage of second home purchases, for example, increased from 7% in 2000 to 14% in 2004 and this further fueled the increase in property values.
  • While all this was going on, the government (particularly HUD) inadvertantly pushed lenders into speculative lending practices through aggressive home ownership goals
These developments had a profound influence on how Wall Street assessed risk in subprime lending. In an environment of declining interest rates and rising property values, defaulting homeowners had more options to avoid foreclosure, including the refinance or sale of mortgaged properties. Even when defaults did occur, loss severity was mitigated by improving collateral values. Over time rating agencies, investment banks and some of the best managed hedge funds in the world accepted these conditions to be the norm and, by doing so, made critical errors in judgment.

Garbage In, Garbage Out

To understand how rising property markets affected Wall Street’s thinking, consider how we make decisions. When contemplating where to go for lunch, for instance, we consider the options already known to us—places where we’ve eaten before and locations our friends and acquaintances may have recommended. Decisions, in other words, are generally rooted in past experience and knowledge. Given this natural dependence on prior events to guide our future choices, what did the marketplace know about subprime lending?
Not enough. And here’s why:
  • Until 1994, subprime was an inconspicuous niche in the broader mortgage market, amounting to less than 5% of originations. Although a decade later the product would represent one out of five new mortgage loans originated, the history of subprime lending as a mainstream financial option was relatively short—not much more than ten years.
  • The period during which subprime grew from a niche product to a borrowing option with broad appeal was atypical of most business cycles. Interest rates were in a nearly continuous decline and, aside from a few regional dislocations, real estate markets were buoyant.
  • Subprime lending changed significantly during the period as guidelines became less restrictive. Virtually every metric used to measure loan risk—including metrics meant to gauge collateral value and a borrower’s ability to pay—were gradually relaxed.
All of this led to a classic “garbage in, garbage out” problem. As rating agencies, Wall Street firms and investment houses analyzed the risks in subprime loan pools, they relied on past performance data that reflected: 1) a short and unusually sunny history and 2) loan guidelines that had once been far more restrictive. While they did assume default rates as high as 30%, their financial models tended to predict that strong collateral values would mitigate against severe losses. That, of course, proved to be false. In keeping with our earlier restaurant analogy, Wall Street made lunch reservations at Le Benardin, but ended up at IHOP.

A House of Cards

In the third quarter of 2004, when the Fed reversed its four-year old easy money policy, it set in motion a series of events that uncovered weaknesses in the subprime market and reversed the earlier rise in property values. The effect was to reduce liquidity, leading to fewer borrowing options, which in turn spawned rising foreclosure rates and further reductions in liquidity as lenders abandoned the market.

The subsequent decline in property values was like an ebbing tide that exposed the decay of subprime excess in its retreat. As losses mounted, lenders quickly learned that much of what they’d presumed to be credit problems were instances of fraud instead. These misrepresentations had gone undetected—or ignored—while the home price tide was rising.
Consider the following:
  • Basepoint Analytics estimates that up to 75% of Early Payment Defaults (defaults that occur within the first few months after loan origination) can be tied to material misrepresentations in the mortgage application. This result, in part, is due to the subprime market’s overuse of State Income Loans (SILs) that don’t require proof of income entered on the loan application. It is estimated that 90% of SILs overstate income by 5% and that 60% of SIL applicants overstate income by 50%, or more.
  • The FBI estimates that one in ten mortgage loans contain some type of fraud, 80% of which include over-inflated appraisals. As the Mortgage Brokers Association for Responsible Lending has said, “Pressure is applied to appraisers who are forced into coming up with predetermined values for homes in order for deals to go through.”
Unfortunately, investors have suffered heavily for their lax underwriting and inaccurate financial models, but are all their losses related to defaults? The answer is yes, but not in the way many people think. Consider the example of Countrywide, which in October last year announced in its 10Q that its delinquencies were up 1.81%. Was that a problem severe enough to fell the mortgage giant? Probably not, but it was only one measure of the lender’s predicament—an aspect that paled in comparison to another result mentioned in the same 10Q. Countrywide’s mortgage volume for the quarter had dropped 44.3%! In short, the company had lost its subprime investors and could no longer originate the volume of loans it once had. Suddenly it had too much costly infrastructure chasing too few deals.

In this way, much of the writedowns attributed to mortgage lending are due not to direct default losses but their secondary effects. These include:
  • Operational writedowns associated with the downsizing (or closing) of once flourishing businesses. Many recently bankrupt lenders had made a practice of selling all their loans into the secondary market and retained only limited exposure to credit risk. Their financial problems were due to revenue shortfalls rather than borrower defaults.
  • Contingent liabilities stemming from the possibility that fraudulent loans might be subject to buy-back provisions. This problem receives little mention in the press, but it’s a significant concern to large originators. For the time being, a lack of clarity regarding the true ownership of loans is a saving grace. It’s not apparent who has the right to demand repayment!
  • Mark-to-market losses arising from investments in mortgage backed securities and whole loans. The overwhelming majority of mortgage investments are still paying according to schedule, but there is little demand for the assets and their market prices have fallen, in some cases 30% to 70%.
Where Are We Now?

Obviously the market is in a shambles, but just how bad is it? In some respects it’s worse than anyone is letting on and here’s why.

The majority of subprime loans eventually find their way into Collateralized Mortgage Obligations (CDOs). These are structured transactions that employ a kind of financial alchemy to turn pools of mortgages into securities that can be bought by institutional investors. The securities come in various risk classes or tranches. The highest classes—generally rated AAA and perceived to be low risk—are purchased by insurance companies, banks, mutual funds and similar institutions. By comparison, the other classes are more prone to losses and receive lower (or no) credit ratings.

The reason why some tranches are more risky than others is that they are placed in a payment hierarchy referred to as subordination. In a typical deal, the lowest subordinated tranche (often called the “hurdle” or equity class) is the first to be hurt by defaults in the loan pool. It will generally take the first losses—which reduce its pro forma yield—until it is completely wiped out. At that point, the next higher class begins to absorb defaults and so on according to the increasing level of seniority. Obviously, the higher in the hierarchy a tranche appears, the less likely that it will ever suffer a loss.

For the subprime market to flourish, it must have investors ready and willing to purchase the range of securities that are created through CDO issuance. However, the market for these securities is virtually non-existent today. In particular, the lowest class securities—once purchased by specialized hedge funds and REITs—attract few, if any, buyers currently. Many of the hedge funds, such as those once run by Bear Stearns, are no longer with us. This is a noteworthy development, since the investors in equity class securities once crafted the loan programs and guidelines that added hundreds of billions of dollars in liquidity to the mortgage marketplace. Until such investors are resurrected or replaced, we will continue to experience today’s liquidity squeeze. However, when a better day comes—as surely it will—let’s hope the investment community will possess improved data and financial models that can translate into more appropriate and sustainable loan programs.


troutbirder said...

I believe my father, who was a banker, how deceased surely would have turned over in his grave had he read your description of the causes of the present disaster.

Alan Bahr said...

I'm sure he would. The people on the ground--the loan officers and underwriters who were doing the work--new the loan programs they were given were troublesome.