Monday, December 14, 2009

The Crash

If you have been patient enough to follow the story this far, I hope that our simple car-crash schema is now fully developed and we have the four basic components we need in order to have our spectacular "accident":

1. The Engine. A fractional-reserve banking system that is inherently dangerous, but extremely efficient at providing liquidity and credit expansion. The main parties who stand to lose if and when the system fails---depositors---no longer care much about risk levels because of generous depositor insurance (which, thanks to clever ploys like CDARS, are effectively unlimited in terms of loss coverage), and bank owners have been compromised by concepts like the "Greenspan put".

2. The Fuel. The Federal Reserve pumped so much money into the banking system starting in 2002 that the Fed Funds Rate had turned negative in real terms, making saving a loser's game and borrowing the breakfast of champions. Serious departures from moderate monetary policy regimes, such as that described by the Taylor Rule, combined with the bank multiplier effect of the fractional-reserve system, led to a vast amount of credit being available.

3. The Steering Wheel. The vast amount of credit was directed towards the housing market. The US Government, through FHA, Fannie, Freddie, and the Community Reinvestment Act, has made affordable housing a priority for many years. Fannie and Freddie, who had incestuous relationships with key players in the House and Senate (mostly Democrats, but some Republicans were also involved), were particularly instrumental in creating an insatiable appetite for subprime mortgages, mortgages that the two GSEs purchased from the banks (freeing them to generate more), then securitized and initially sold into the market. At a certain point, Fannie and Freddie began acting like giant hedge funds and using their AAA credit rating (courtesy of government backstops) to borrow cheaply and repurchase mortgage-backed securities for their own balance sheets. Perhaps $1.5 trillion in toxic subprime material was being carried in this way, until Fannie and Freddie suffered catastrophic losses, failed, and were placed under government control.

4. The Gas Pedal. Sophisticated financial engineering contrivances such as CDOs and CDS allowed banks to remove mortgages from their own balance sheets and also created a tiered structure for new derivative securities that gave purchasers the ability to daintily select from a menu of offerings, with choices dependent on product credit quality and risk tolerance. The methods used by market operators and ratings agencies to assess default probabilities suffered from statistically optimistic assumptions regarding market behavior that have now been falsified. The result of this apparatus was a money machine that performed very well for years, and then generated huge losses in 2007 and 2008. Of the five major investment banks in the US, one failed outright, two were sold off (by heavy-handed government intervention and as-yet-undisclosed threats/promises)to other banks, and two converted to bank holding companies in time to avoid being taken under.

Basic Things That Went Wrong

In the 1980s, the economists Bob Van Order and Chet Foster were hired by Freddie Mac to further develop their earlier, option-based mortgage default model. The Foster-Van Order model essentially predicts that mortgage defaults will rise as the owners' equity in their homes becomes negative. The key inputs to the model involve A) the endogenous drivers of owner equity (the down-payment as a percentage of home price being the biggest for new mortgages) and B) the exogenous drivers of owner equity (i.e., home price appreciation).

The basic logic is this: if a home is currently worth $350,000 and the owners have $70,000 in equity (either by a 20% down payment or a combination of down payment and accumulated equity from having made payments over time), then the owners would rather sell the home for $350,000 if they suddenly were placed under financial distress than allow it to be foreclosed upon by the bank. A sale will still yield $70,000 for them. If, on the other hand, the house is now worth $350,000 but the outstanding mortgage is for $400,000, the owners are said to have negative equity in the home and selling it will not be sufficient to even pay off the mortgage, let alone realize any gains, hence the much higher probability of default.

Applications of the Foster-Van Order model thus involve specifying a probability distribution for the future path of home prices. To stress-test the model for harsh conditions, one would specify periods of house declines: Arnold Kling, formerly at Freddie Mac, has reported that one such stress test used during this time period involved four consecutive years of 10% declines in home prices (a very harsh test indeed).

Another key input in the model was the size of the down-payment, and here is why: if the assumption is that negative equity status creates an incentive for default, then a larger down-payment gives a natural buffer against home price depreciation. A very low down-payment---say, 0%---means that, for a new mortgage, equity is ENTIRELY dependent on home price appreciation for many years. There is no buffer---any downward shock to housing prices immediately puts these owners underwater and thus far more likely to default, and the strength of this relationship increases at an accelerating, non-linear pace as home prices continue to fall.

Note how different the Foster-Van Order approach is from the Gaussian copula approach, which used correlations between derivatives (credit default swaps) instead of explicitly conducting downside sensitivity tests for home prices themselves.

Unfortunately for everyone involved, there is an inescapable tension between the noble goal of increasing housing affordability and the need to protect securitization issuers/purchasers like Fannie Mae and Freddie Mac from the dire results predicted by the Foster-Van Order model should the unholy combination of falling house prices and very low down-payments be achieved. As we now know, the faction favoring housing affordability won out in the early 2000s and a grand social experiment took place after that.

It is possible that housing affordability, as a public policy goal, should in fact win out, despite the recent catastrophe, and that larger down-payments would disenfranchise a huge percentage of the population and create terrible social consequences down the road. This is impossible to state with certainty, of course, but perhaps we need to accept the risk of periodic market crashes as part of doing business, a basic operating cost that comes with attempts to link free market profit incentives for innovation and risk-taking with the moral hazard of socialized losses when things inevitably do go wrong.

My personal feeling is that the orientation towards lower down-payments also fits within the Keynesian war against savers (since higher down-payments will require that people save a far greater percentage of their incomes, this would represent a problem for the Keynesian utopia. Edward Bellamy, an early American socialist, perhaps best described the socialist disdain for personal savings in 1888: "No man has any care for the morrow, either for himself or his children,for the nation guarantees the nurture, education, and comfortable maintenance of every citizen from the cradle to the grave"). Any truly new policy regime that strives for internal coherence is going to involve an overhaul of basic assumptions regarding the proper roles of fiscal and monetary policy in shaping or distorting economic outcomes.

The Minsky Moment

The economist Hyman Minsky proposed that financial markets are inherently prone to bubbles and crashes, as periods of low volatility encourage excessive, debt-fueled risk-taking (his proposed solution thus was heavy on government intervention policies). He believed that booms and busts were the result of three different waves of financing methods that were employed by investors to take positions in financial assets: hedge, speculation, and Ponzi. In the hedge phase, cash receipts from investments easily covered interest and principal on debt. In the speculative phase, cash receipts from investments covered interest on debt, and there was a belief that future increases in cash receipts would allow for principal to be retired. In the final, Ponzi phase, cash receipts did not even cover interest payments on existing debt, so more debt had to be taken out to service the original.

Minsky believed that a normal yield curve would encourage a move towards speculative financing, since there would be profits to be made from borrowing at lower, short-term debt rates and loaning at the longer-term, higher rate. The danger was interest rate risk---needing to rollover short-term rates meant that the speculative financing position could be exposed should short rates exceed what was being earned on the long-term position. A Ponzi phase was simply unsustainable because the leverage would ultimately result in even a small shock bringing disaster. The catalytic event in which the Ponzi is finally exposed and the turmoil begins has been termed the "Minsky Moment".

I think there is something to be gained from Minsky's model, although I do not agree that the more dramatic swings are necessarily inherent features of the financial markets so much as they are an unfortunate consequence of a series of preventable problems, many of them stemming from distortions and moral hazards created by government intervention. I personally prefer the Austrian theory of the business cycle because it incorporates the role of monetary policy mistakes in creating a debt-friendly environment.

Regulations and Reforms

-Among the many regulatory initiatives that are being put forward is a return of the Glass-Steagall Act that formerly split commercial and investment banking functions. However, it was the revocation of Glass-Steagall that made the emergency, Treasury-pushed shotgun marriages of Bear Sterns and Merrill Lynch to JP Morgan Chase and Bank of America possible, and also allowed Goldman Sachs and Morgan Stanley to survive the crisis by converting to bank holding companies so that they could be recapitalized.

-Fannie Mae and Freddie Mac will almost certainly need to return to their old conforming-loans safety model, and to employ Foster-Van Order type risk modeling.

-Occasionally I hear or read the S&L crisis of the 1970s and early 80s being brought up as an example of another de-regulation-induced banking system disaster, but this is revisionist history. The Savings & Loans were limited by Regulation Q to paying depositors interest rates of 3% (eventually Congress raised this to 4%), in an environment of double-digit inflation. Money market funds were created to offer higher returns, and of course depositors fled the S&Ls. Congress first raised the allowable interest rate that the S&Ls could pay depositors, and then gave up and revoked Q when it was clear that the S&Ls were facing runs. This finally allowed the S&Ls to pay competitive rates and retain depositors. Unfortunately, these firms had also made long-term loans out at (low) fixed-rates, so they now found themselves underwater (i.e., paying a higher rate to short-term depositors than they received from mortgage payments). Advanced, creative stages of risk-taking were incentivized because of predictable principal-agent problems (managers knew that the firms were dead men walking, anyway, so they started taking wild bets) and some very bad loans were one result.

In other words, the S&Ls were really killed by inflation and the time/interest rate coordination problems that come with it, by monetary policy errors that started in the 1960s and culminated in Nixon taking the United States off of the gold standard so that the printing presses could be put to work.

Incidentally, mortgage securitization was developed by Fannie and Freddie because the S&Ls were prohibited from acting across state lines. Securitization allowed for mortgage originators in, say, California to access interested investor money in Illinois. If we wish to point the finger at CDOs and the like, we should also consider where mortgage-backed securities got their start.

-It is possible that FDIC insurance is too generous (currently at $250,000, but we already discussed how this is easily circumvented) and has removed depositor risk monitoring discipline from the banking system, but I do not believe that a reduction in FDIC coverage is politically feasible.

-Risk modeling approaches for CDOs and synthetic CDOs will be changed away from the Gaussian copula, probably by a third iteration of the Basel Accords. This would be very reasonable.

-CDS may be regulated as insurance contracts or placed on exchanges. My initial reaction to this is that it would be a sensible regulatory change, but I am wary of adverse unintended consequences, I am not a credit derivatives professional, and I have not studied the pros and cons to any degree.

In future posts we will discuss the investment climate that this crisis, and the government's response to it, have created and some future macroeconomic scenarios that may unfold as a result.

For Further Reading: Hyman Minsky's "Stabilizing an Unstable Economy" was a hard one for me because Minsky has a difficult style at times, but I think the book can be read with profit if the reader is sufficiently motivated. "The Austrian Theory of The Trade Cycle", a compilation of essays by many of the great economists of that (beloved) school, is an alternative that many readers will enjoy.

A 50-minute youtube lecture on the Austrian theory is available and I have posted a window below:

1 comment:

  1. Seb, I've read this series MANY times now as I struggle through my auto-didactic program in remedial finance and econ. I really think it's the most succinct analysis of the crash that is actually understandable without being overly simplistic; congratulations on a great job. You also do an admirable job of exposing your potential biases.

    It seems like there is quite a lot of the old Red Queen effect going on here, especially in the disconnect between the people actually selling street-level mortgages and the higher level finance and government types. It doesn't seem like the latter, no matter how idealistic, would have condoned selling mortgages to people who would inevitably default, but once one group starts doing it, everybody else jumps in to keep up.

    In your conclusions in this post, though, it's not clear how you see free-market solutions being able to auto-regulate a market that is almost definitionally opaque to shareholders. Presumably, if I'm a shareholder in a company that's up to no good, I should be able to punish it by selling or shorting its shares. If there's little of the shadow-bank system exposed to normal shareholders, however, how can the market effectively punish bad behavior?