System Crash on Wall Street

By Kurt Cagle
October 6, 2008


The credit markets are seizing up, Congress-critters are trying to make the case for spending billions in a "rescue" package, the stock market gyrations are giving people whiplash, banks are popping like sulfur-filled bubbles and companies are suddenly having to make some hard decision about payroll at a time even when they have more than enough work. The end of the world as we know it seems to have come about all at once, and even as people are scrambling to protect themselves, not a few people are wondering just how everything went bad so quickly.

Most of the time, finance works very much like technology - the end user of financial services generally do not think that much about how high finance works, and prefer in fact not to know any more except for information which directly impinges on them. They get their checks in the bank and don't have to worry about how it gets there, they take out loans for mortgages, and as long as the monthly terms are right they don't really care how much they spend in the long run ... or how the mortgage itself gets to the users.

A useful analogy in this regard is that such transactions are abstractions in the same way that serving up a web page is generally an abstraction - from the standpoint of the end user, when they point their browser to a URL, what they get back is an HTML document, and that user can generally assume that there's a real document on the other end. From the web developer standpoint, of course, there's much more involved - components or scripts generate HTML, databases are accessed to provide specific content or make decisions, state and environment variables have to be written or updated for every page, and so on. In other words, to create the illusion of a document sent to the client, the web developer generally needs to do much more heavy lifting than is apparent from the other side.

Finance works much the same way. The modern banking system is based upon a number of illusions. One of the most obvious is the idea that a bank is the place where money is kept - you put your money in the bank and this money is put into a vault, and when you want your money back, they retrieve money from the vault. If you want to take out a loan, you can also get money from this vault. It's a convincing illusion because of course of old this is precisely how banks operated. Today, its complete fiction.

Any bank branch office does have a certain amount of cash on hand in a vault, that much is true, but that cash is primarily for day to day transactions, and is strictly limited (typically no more than $20,000 is actually kept on hand for most branch offices, and in many cases it's even lower than that). Any balance beyond that is usually trucked to a centralized office, and redistributed through a banks branches as necessary. Yet this is just for immediate cash.

Typically, the operating capital of a bank exists as less liquid vehicles. One of the most typical such vehicles are US treasury bills, which are usually sold in denominations starting at $10,000. Such T-Bills are of course redeemable, but banks prefer not to move outside of this for their immediate operating capital unless the T-Bills have matured because such are interest bearing vehicles (at admittedly a fairly low rate of interest).

Financial Tools for New Times

The second edition of the O'Reilly Money:Tech Conference will be an even deeper dive into the space where Wall Street meets Web 2.0, using technology as a lens to provide a unique view of the most pressing issues facing the industry now, ranging from securitization and trading velocity to risk measurement and the evolution of research. Happening February 4-6, 2009 in New York City.

Banks generally make very little off of checking and savings accounts directly due to interest (though they rake in a fair amount in fees). Part of this is because the Federal Deposit Insurance Corporation (FDIC) requires that they maintain a certain amount of core capitalization on accounts in order to cover the $100,000 per account guarantee - a person setting up an account in a commercial bank is insured such that they will get at least $100,000 of any account if what they have in the account exceeds that amount.

Again, however, this does not in fact assume that the FDIC actually maintains that amount; indeed, if you think about it, you can understand how absurd that is on the face of it (there are hundreds of millions of bank accounts in the United States alone, which would potentially require that the FDIC be capitalized at more than $10 trillion dollars).

The reality is that each bank is required to maintain a specific reserve against default, and if they are unable to meet that reserve, then the first move of the FDIC is generally to borrow money from other member banks reserves first, at a low but non-zero interest rate, and to pay out directly only in those cases where it can't raise the money. This is usually sufficient for most day-to-day crises, and it means that the actual amount that the FDIC has on hand is in the neighborhood of $80-$100billion or so.

Most of the money that banks make come not from interest on deposits, but on loans. A bank is able to loan out a certain multiple on its reserves - with the assumption that this percentage should be able to fully cover the default of a borrower. If this fraction is 10%, say, then if a bank is capitalized to $10 million, then it can make up to $90 million in loans. In normal times, the likelihood that more that 10% of investors will go into default is comparatively low.

Most smaller commercial banks tend to be very conservative (with fractions usually up around 15-18%) because the risk of a default event is higher, and the damage from such an event could be considerably more devastating to a small bank. The larger the bank, however, the less the likelihood that a default event would prove devastating, and the easier it is for the bank to take care of such a loss (and consequently the smaller the fraction of the reserve to total outstanding debt).

Of course, this also means that the loans themselves are larger, and it also means that larger commercial banks usually find it profitable to buy up smaller banks. The total number of independent banks in the US (and anywhere else where a weak regulatory regime held sway) has dropped by a factor of 10 in the last thirty years, with many purchased during recessions by larger banks when the smaller banks were forced to liquidate their reserves. This has meant that the results of a bank failure have generally proved more catastrophic to a larger number of investors over that same time.

Certain types of loans tie up capital for longer periods of time than others do. A loan for a car, for instance, will only tie up capital for perhaps five years, whereas a mortgage will tie up capital for between 15 and 30 years in most cases. Of course, the house mortgage also has a much higher total rate of return - if you pay $2000 a month for a 30 year 6% fixed mortgage, you could afford a $350,000 house ... and are paying the bank another $350,000 in interest for that house, not counting fees and property taxes. (All of these are making the assumptions of $0 for a down payment, which was unlikely early in the 1990s but all too common later.)

Returns like this tended to be the catalyst for the creation of mortgage banks (and for a number of commercial banks to get into mortgage lending). A debt vehicle like this essentially returns the equivalent of the purchase price of the house to the holder of that debt, and as such it has considerable value in and of itself. There is a certain risk that the homeowner will default, of course, but (at least as was sold to many potential investors) that risk is small - less than 0.05% in most cases.

A long standing rule of thumb said that your total house payment should never exceed 25% of your pre-tax income. This means that, using the above numbers, to be able to afford a $350,000 house at those rates, your total income needed to be approximately $100,000 a year. In 1990, a $100,000 a year income in the US was typical of senior management with small to mid-sized companies, and the median household income was about $40,382 a year in 2000 inflation adjusted dollars ($30,000 before adjustment). In 2000, the media household income was $42,000, meaning that the median income only rose about 4%. Median income in 2008 is essentially unchanged in adjusted dollars.

In 1990, the median cost of a house was $100,000 in 2000 dollars. In 2000, it was $119,000. In 2007, it was $203,000 ($280,000 unadjusted). In other words, the cost of a house has gone up 100% in the last twenty years, most of which was in the last eight.

There were a lot of factors for that. Alan Greenspan and the Federal Reserve reduced the prime lending rate to 1% after the attacks of 9/11, then kept it around that rate for several years. This meant that mortgage lenders could offer far more house for the money, but rather than selling the same houses for less, they encouraged people to purchase larger (and consequently more expensive) houses for the same amount, giving the best rates for adjustable rate mortgages. They also encouraged homeowners to borrow against the equity already paid down in a second or even third mortgage at refinanced rates. Finally, the Bush administration liberalized a number of the rules on mortgage accounting in order to foster "The Ownership Society" as a campaign pledge in 2004.

Lax regulatory practices in the late 1990s were exacerbated after the start of the Bush administration, including reducing the fees to mortgage companies to sell mortgages to investors, and making it possible for commercial banks to both buy and sell mortgages and to become involved in investment banking (which had considerably lower requirements for maintaining reserves). This practice had been forbidden in 1933 after the crash of 1929 in the Glass-Steagall act, which not only set up this wall but was also responsible for the FDIC in the first place, but was repealed in 1999 as part of the Gramm-Leach-Bliley (GLC) act.

One immediate effect of this was that houses became perceived as investment (and speculation) vehicles, and the cost of both new and used housing skyrocketed. As the costs went up, so too did the number of builders starting new housing complexes as did the rising revenues coming from property taxes. Additionally, because banks also made fees of the transactions, they may have also influenced property inspectors who would tend to raise appraisal rates for all houses in a given neighborhood where selling was marked.

Yet as the quality of the borrower diminished (both because that same borrower was looking at housing taking up a higher percentage of their income in housing and because other factors - food, fuel costs, education, health care, etc. - were also cutting into the same income) the risks of default on these mortgages increased as well, and mortgage companies began trying to sell more of these mortgages into the investment markets rather than have these mortgages potentially explode in their faces.

Investment banks differ from commercial banks in a few key areas. One of the first is that such banks are (or were) considered to be speculative corporations rather than fiducial ones - their goal was increasing yield, and hence the return on investments for their clients. Investment banks do not need to retain as much capital, because their investments are not guaranteed - an investor going in knew that the possibility existed that he could lose some or all of his investment. This meant that reserves for such banks were often in the single digits (at the time it went bankrupt, Merrill Lynch had a reserve fraction of 2 ½ percent).

With the repeal of Glass-Steagall, it meant that commercial banks could create a branch of their business that handled investments, and it also meant that assets of increasingly high risk could be moved "off the books" for accounting purposes, into their speculative arms. Banks and investment firms do have to provide financial statements for both the government and for their investors. The Securities and Exchange Commission (the SEC) also established three "levels" for their assets - Level 1 assets were those for which an existing price already existed in the market if the assets had to be sold (market-to-model), Level 2 assets were those for which the precise value could only be determined by creating a model of similar vehicles (what was called marked-to-model accounting), and Level 3 assets which consisted of assets whose values were unknowable on the market.


Marked-to-market accounting started out in the 19th century as a practice among futures traders, though it became entrenched in the late 1980s. In essence, a future trader, when taking a position (a bet on where the price will be), puts down a deposit with the exchange called a margin, intended to protect the exchange from losses. At the end of the trading day, the contract for that particular position is reset to the value for that day. If the trader bets correctly, his account is increased by the contract difference. If he bets wrong, he pays the difference. If he can't pay, he forfeits the margin.

Marking to market opened up a number of abuses in the market itself in the 1980s and 1990s, but ironically, it was the creation of marketing to model in the 2000s in financial company accounting that has become the biggest problem. New financial vehicles that have not yet found an effective market (perhaps because the vehicle hasn't matured yet) can't really be marked to an existing market value, because there may be enough difference between these vehicles and what's already out there to make the comparisons become apples vs. oranges.

As such, the banks had to create models indicating what they felt the securities were likely worth if they were to achieve a market value. This mark-to-model accounting (or Level 2 Accounting) was of course only as good as the model, and since the people holding the securities were the ones making the models, the temptation to value them highly was pretty much irresistible. Within the industry such accounting became known as mark-to-fantasy, as the values given to these assets were as likely as vampires and unicorns wandering down Wall Street.

Backtracking for a moment, as the number of mortgages being dumped into the investment houses increased, these same houses began to realize that they would end up with a significant amount of under-performing loans if they sold only the ones that were safe. As a consequence, they came together and agreed upon a system whereby mortgages would be divided up into one of five tranches (slices), based upon the "quality" of the mortgage. AAA tranches had the highest quality (least likely default percentage) and as such had the lowest returns of investment upon them, mezzanine (AA to BB) tranches had higher default potential and higher rates, and equity (sub A or sub-prime) transches had the highest risk and highest returns. These tranches were then sold as Collateralized Debt Obligations (CDOs) to investors; the ones that were risk averse would stick with AAA CDOs, the ones that were seeking the highest rates of return would play in the sub-A equity CDO space.

Enter the hedge funds. If an investment bank sold themselves to those seeking better returns for their investments, hedge funds catered to the well-heeled elite. Many hedge funds required initial net worth for investors of at least $1 million dollars or higher, and all of them were usually tightly limited in membership. Hedge funds can thus be best be thought of as exclusive investor clubs, and they are even less well regulated that investment banks.

Hedge funds played in the derivatives market. A derivative is a financial instrument whose value depends upon other financial instruments, and are divided into futures, options, forwards and swaps. Each of these are bets of some sort that the market will increase or decrease to some value (as the futures trader in the previous example did with margin). Margin traders could enter into the market with a comparatively small position and make a bet that had a relatively low probability but a high potential return. This principle, called leveraging, could very easily be used to turn a small stake into a much larger stake (in much the same way that a bank can leverage to create money for loans) - in essence, taking advantage of the randomness of the market to get in and out of a position before it becomes time to mark an asset price to market.

Derivatives can also be used, however, to create insurance. One form of insurance is in an exotic financial vehicle called a credit default swap, or CDS. In a CDS, one party of a transaction takes a position that a credit vehicle (such as a mortgage CDO) won't default, while the other side (the counter-party) takes the position that it won't. The first party pays either an up-front payment or a periodic premium for this insurance so long as the credit remains good. For nearly twenty years, this was a winning formula for both sides - the likelihood that a mortage would default was low enough that the counter-party could effectively keep a fairly low cash position while still raking in the premiums from the first party.

As hedge fund managers began to become more concerned about the quality of the CDOs, however, they began to apply ever larger default swaps to these investments. Similar things were happening with other types of credit vehicles, including automobile loans and credit card debt, and the ability of mortgage companies and even banks to sell the CDOs and other credit vehicles began dropping off in the summer of 2007, and by late fall, mortgage banks were in a situation where they had a large amount of unsold mortgages that were eating into their books. By October 2007 these mortgage banks were failing at a rate of ten a week, either to be bought up by other banks or closed by the FDIC.

In November 2007, the Financial Account Standards Board established a set of rules in 2007 called FASB 157 "Fair Value Measurements" in order to provide some kind of guidance on how both mark-to-market and mark-to-model valuations should be determined. Formally, FASB 157 defines values as "The price that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date". It emphasizes that value should be market based rather than model based, and set up a set of hierarchical divisions for determining how such asset prices should be determined. These rules were set up in order to enhance confidence in the markets, but one of the most significant effects of this was to underline just how fragile the market really was.

CDOs and similar vehicles were definitely beginning to reek as foreclosures accelerated into 2008, especially as initial "teaser-rate" mortgages of 1%-2% reset to rates in excess of 6% and mortgage rates begun to break its link that had held for several years of following the prime lending rate. Adjustable rate mortgages and "balloon" mortgages where the rates ratcheted up the more mature the mortgage became put even further strain on already overstretched households, and even as the economy seemed to be improving somewhat in March, the term "jingle-mail" was entering into the American lexicon, describing people who just walked away from a house and its mortgage, sending the keys back to the lender.

The largest holders of these toxic assets were the investment banks, and the new rules introduced by FASB and the SEC forced these companies to move these assets out of the Level 2 accounting (Mark to Model) and into Level 1 accounting (Mark to Market) as part of their quarterly filings. Yet because there was so many of these unmarkable assets, there was no market for them, and the banks were forced to take huge losses that significantly cut into their reserves. What's more, the credit default swaps that were designed to provide insurance against much lower defaults were beginning to kick in, and the counter-parties to these swaps, such as insurance giant AIG, were now having to pay out on these swaps at an alarming rate.

By April 2008, Bear Stearns, the fifth largest investment bank on Wall Street, collapsed when it couldn't find buyers for its assets. Through the next six months, investment banks that had been doing business on Wall Street for decades (in a couple cases more than a hundred years) were either acquired by commercial banks or went out of business until only two - Goldman Sachs and Morgan Stanley. Both of these petitioned (and were granted) the option of becoming commercial holding companies, granting them a certain degree of FDIC protection in exchange for effectively walking away from their investment business. By September 2008, investment banking had died in the United States.

Yet the largest holders of mortgages were the two quasi-governmental lending banks - Fannie Mae (Federal National Mortgage Association) and Freddie Mac (Federal Home Loan Mortgage Corporation). As mortgage lenders of last result, these two organizations essentially ended up with most of the poor quality mortgages that were being released from the CDOs as they were broken up, and they too had used credit default swaps in order to provide the financial backing for the mortgages.

In late August, 2008, both of these companies were forced to account for their poor quality mortgage assets, and announced that they were bankrupt. This began a series of triages by the Federal Reserve, the White House and Congress, as institutions were falling faster than they could be propped up, with the US adding obligation after obligation to an already huge debt load. However, what was perhaps worse than this (in the short term) was the fact that such a bankruptcy was considered a Default Event. Up until this point, most CDS obligations were able to be held in abeyance, but with this default, it forced a huge number of credit default swaps in the market, that wiped out the reinsurer market (such as AIG, which had to be nationalized), and led to faults in one of the most stable financial markets - Money Market Funds.

Money Market Funds are large and extraordinarily conservative funds that in theory should never be undercapitalized. Because of the stability of these funds, large banks typically borrow from these funds in order to provide capital to secondary banks at a rate known as the London Interchange Bank Offered Rate, or LIBOR. While originating in England, the LIBOR rate is now used in 60 nations for over 200 large bank members. LIBOR is considerably below commercial market rates, but it has a strong influence on those rates.

On September 16th, as the CDS market unraveled further, three money market funds "broke the buck", in which investors saw $1 of investment return only 96 cents. Investors in these markets saw that the CDS market was collapsing, and withdrew $173 billion dollars in reserves (essentially $3 trillion dollars of obligations) in the space of a few days The Treasury Department stepped in on Sept. 19th by announcing a money market insurance program to cover investors in these accounts and outflows slowed.


However, the damage had been done. LIBOR rates climber dramatically (and are still rising). One of the most significant measures of economic instability is the The TED or (Euro-Dollar rate spread), which originally was defined as the difference between three month treasury-bill (T-Bill) rates and three month Eurodollar contracts, but is now defined as the difference between the T-Bill rate and the LIBOR rate.

Until 2007, the TED spread typically traded in a range between 10 and 50 basis points (where a basis point is 0.01%, so that a TED of 50 is half a percent point difference in rates). Through 2007, the TED spread had grown to 150-200 bps. On September 17, the day that the S&P fell more than 8%, the TED spread had grown to more than 300 points, and by October 4, it had reached 387 bps, the highest its been since its inception thirty years before.

In general, the higher the TED, the more that liquidity is being drawn out of the banking sector. Liquidity in turn can be thought of as the degree to which loans can be made. If you go in and ask for a loan, if you're considered sufficiently credit worthy, the bank will secure the money from another bank at the LIBOR rate plus an overage rate. As LIBOR rates rise, this gets reflected in the interest rate that you pay for the loan (which are also rising dramatically) but as the TED rate rises, this also indicates the likelihood that the bank can in fact purchase the funds for the loan in the first place. Thus, at a spread of nearly 400 bps, the TED rate is indicating that there is effectively only about 13% of the money available to loan out as there was two years ago, and when you can get it, you're looking at very high interest rates.

At this stage, the banks can't purchase those funds (what is often called Commercial Paper in the industry) - there is simply no money to lend right now. Ironically, at least for the moment small loans (which can be handled by the banks themselves) are still possible, but larger loans are frozen because they can't get the funds ... and many banks are in turn reducing their own exposure on loans as they become increasingly concerned about their ability to receive funds in the future.

On October 3, Congress passed (and the president signed) the the Economic Stabilization Act of 2008 (HR1424) which effectively authorized $700 billion dollars to be used for the purchase of CDOs from the banks - for the primary purpose of attempting to set a market value on these instruments. Until that happens, the CDO market will continue to spin out of control, setting off further credit default swaps time bombs.

Additionally, it reinterprets the FASB 157 accounting rules to take into account the fact that fair value cannot be interpreted when the market is reacting in crisis - a controversial rule that will take some pressure off the banks in the short term but that has the potential to let additional problems fester under the surface in the longer term. The act also places limits on executive compensation, provides tax breaks to individuals and organizations in high risk categories, and formalizes the insurance funds to money market and similar critical lenders to protect and woo back investors into these critical funds.

Meanwhile, the $10 trillion dollars currently leveraged in hedge funds are now coming unleveraged with a vengeance. First, to put this in perspective - The current annual gross domestic product of the US is approximated $14 trillion dollars -  about $40,000 per person in the US. As these hedge funds and derivatives collapse, it could take out anywhere from $10,000 to $30,000 per person, depending upon where the market finally finds a floor. This means effectively that roughly half of the money within this country is now disappearing over the space of a couple of months.

What happens from here? That is not an easy question to answer, and it depends to a great extent on the speed at which the program becomes implemented and how readily the banks will be to take a potentially significant markdown of these assets. In the immediate term, the effects are likely to be minimal until after the presidential election, as the economic policies of each of of the candidates will likely have a significant effect upon the overall trust in the market place.

Already, the credit crunch is having a serious impact. Both the States of Massachusetts and California were denied routine lines of credit, leading to budget shortfalls at a time when revenues from both business taxes and home property taxes are dropping precipitously. Businesses such as AT&T have also been denied credit, and are now having to make largely unexpected cuts in operating budgets. Many businesses use commercial lines of credit in order to smooth out irregular revenue streams and handle payroll; as these come up for renewal, it's likely many of them will be denied, forcing companies to go to a cash accounting system that will likely have the upshot of forcing mass lay-offs.

There are a number of indications that this is already happening. Hewlett-Packard announced recently that they will be laying off more than 25,000 employees over the next few months, making it one of the first to be severely affected by the crunch and eBay is planning on laying off 10%, 1000 or so, employees in the next few months. Large IT companies in general are fairly well positioned to weather this storm because many of them had developed moderate to large cash reserves, and the market has been relatively benign to the IT sector in particular even as it's been hammering housing and the FIRE (Finance, Insurance, Real Estate) sectors.

However, there are signs that outside of pure IT companies, the effects of this recession on IT departments within companies may not be good. Coca-Cola, for instance, recently announced that they are radically trimming their IT departments world-wide, and many companies that had been exploring new initiatives in IT are now scaling back or cutting these efforts off completely, resulting in fewer new jobs overall.

On the other hand, smaller technology firms may be facing serious problems either in closing VC rounds or getting already committed VC funds because many such venture capital companies (and their investors) have lost significant amounts of money in the stock, bond or hedge fund arenas. While the San Francisco IT and VC communities definitely are managing to weather the storm, the drying up of the bond market (even at the junk bond level) has also significantly crimped the ability of such VCs to arrange financing, and a "circle the wagons" mindset will likely mean that most new startups will be launching in the harshest economic conditions of the last eighty years.

An additional effect of the credit crunch has been the cratering of mergers and acquisitions (M&As), beyond the forced "shotgun marriages" being arranged by the Treasury Department where dying banks are bought up by the merely badly ailing. Most mergers are handled using junk-bonds - lower grade bonds that can be purchased cheaply and leveraged to cover the cost of a buyout. With even junk-bonds now in extremely short supply, a number of companies who were through much of the merger process are now walking away from the deals, fearful that the continued credit crunch will keep both companies involved in the merger in suspended limbo and thus vulnerable to other economic shocks.

What we're seeing today is the equivalent of the markets having a cardiac arrest. The problem with heart attacks is not that the heart stops beating altogether, but that it spasms, damaging the heart muscle, often triggering damage elsewhere if corruption is too entrenched, then starving the body for oxygen. The global economy is so interconnected at this point that the damage in turn is a feedback loop, causing a rapid, painful deflation in just about all asset classes.

Inflation occurs as more and more credit gets pumped into an economy without there being underlying assets covering that credit; its a desired state of banks, because the more that money moves through the system, the faster the "velocity" of that money, and the more that they can profit on assessing interest rates and fees as that money moves through their institutions.

Systemic deflation is a phenomenon that Americans in particular have not encountered in the last 75 years. The velocity of money slows, and credit dries up. Prices of nearly all asset classes - from houses to cars to oil to stocks and bonds - all drop, but so too do wages, as companies shed jobs and new jobs at the same level pay less.

However, all is not necessarily gloom and doom. Inflationary periods generally tend to be ones where wealth migrates upwards - the middle class shrinks, capital shrinks, institutions become more exclusionary and disparity between the richest and poorest climb dramatically. In a deflationary regime, on the other hand, the economy tends to become more level, though it may take up to a decade for that process to happen. Savings go up - savings that aren't tied into risky investments. Trust in banks drops dramatically and government re-regulates such banks. Inflation becomes non-existent (a small amount of inflation is necessary when populations are growing, but demographically, the population is probably close to peaking and will actually start to fall within the next decade).

Long term, what's happening now is systemic, part of a much broader cycle of economics. It may very well be that this deflationary prelude is a necessary prerequisite for a shift to a more sustainable economy moving forward, one in which financial institutions (and assumptions) that were valid during the period of industrialization (and less so during post-industrialization) no longer work in an era of near-instantaneous communication and 24/7 global markets. This is definitely fodder for a future article.

Kurt Cagle is Online Editor for O'Reilly Media. He blogs on web and economic issues for O'Reilly's Broadcast, and also has an individual blog at Metaphorical Web.

You might also be interested in:

News Topics

Recommended for You

Got a Question?