Bernanke Departs the Fed In Order To State the Obvious
It is not enough to simply infer or explicitly state that the Federal Reserve and its rigid econometrics are ill-suited for the job it has created for itself. I think it goes far deeper beyond the realm of economics, past politics and into the bedrock of American reality. On the one hand, it is very much dangerous to keep reliving the past as there is no doubt the future will contain its own elements and rabid combinations that will be impossible to foresee in very granular terms.
History, on the other hand, is circular, not linear. That fact only seems more relevant given the particular economic and financial history of just the past few decades, where the "cycle" of repeated depression is most conspicuously condensed; a matter unexplained by that impulse to leave the past in the past. The particular emphasis on 2008 serves as the most obvious choke point in between blissful ignorance and the true hope of putting the economic and financial systems back on the path to true prosperity. The illusions must end.
This goes well beyond the very evident and observable incapacity of policymakers to do what they say or believe of themselves. The history of 2008, in strict policy terms, is a comedy of errors linked by an unimaginable chasm between what the orthodoxy understood and expected, and what actually transpired. Just this week, former Federal Reserve Chairman Ben Bernanke admitted, stating the obvious now that the 2008 transcripts are public, that the FOMC was "overconfident."
If that were the extent of the illusion, the solutions would be easy and installable - learn to be more humble and alert, less arrogance, and adjust accordingly. None of that accounts for what has transpired since, including the consistent and persistent undershoot of even official measures of economic activity. Such divergence encompasses both the modeled expectations of the FOMC and mainstream economists, and the public's perceptions of the ongoing disaster in labor. The incongruence between even weak official "growth" and the labor markets suggest very much that there is far more at work, and at stake, than the Fed "learning its lesson."
Mr. Bernanke made another statement in addition to that admission of overconfidence. He said, "I came from the academic background and I was used to making hypothetical examples and ... I learned I can't do that because the markets do not understand hypotheticals." He doesn't specify the exact hypotheticals to which he was referring, or if he did it wasn't picked up by the media recounting his words, but we can make a reasonable guess as to his meaning. And that, I believe, fits very neatly into that extraordinary facet of the 2008 history, and thus its primary relevance to 2014.
There are, of course, more than a few aspects of that crisis and panic that stand out, and will continue to do so for perhaps generations. To me, there were extremes that I think, and hope, will never recur. One of the first examples of that came at the very same moment FAS 157 went live in November 2007. Here was an accounting rule created out of the Enron fraud and fiasco, of all things, being put upon the financial system just as it was beginning to seriously buckle under financial strain. At the time, most investors and observers were blissfully unaware of the dollar strain in global funding markets, but the banks were not.
And the banks knew exactly what was driving those strains, whereas the FOMC, as the 2007 transcripts leave no doubt, had only passing familiarity with the surface of what was a truly deep and multi-dimensional. FAS 157 required financial firms to mark-to-market, but more than that, do so on the basis of "observable inputs" where possible. You may recall the introduction of Levels I-III in bank SEC filing statements, an outgrowth of this new "commitment" to financial "transparency." The timing, with that in mind, was both fortuitous and frightening all at once.
As I said, in November 2007, as banks were updating their accounting systems to incorporate new mark-to-market standards, Morgan Stanley disclosed an unexpected $3.7 billion loss on a subprime mortgage "trade." That, as you might expect, wasn't really the interesting part given the common knowledge of such "toxic" assets. Rather, as then-CFO Colm Kelleher would add, the loss was derived from a short position on subprime mortgages. In other words, the firm had bet, with its own money in a proprietary trade, against subprime at the very time the entire mortgage system was falling asunder - and they still lost an immense sum.
How could that be? As Kelleher would dryly explain, "We began with a short position in the subprime asset class, which went right through to the first quarter; as the structure of this book had big negative convexity and the markets continued to decline, our risk exposure swung from short to flat to long." The two words that should jump out of that statement are "negative convexity." Some investors, particularly in mortgage security trading, are intimately acquainted with negative convexity, but to the vast majority it is beyond mathematical esoterica.
Negative convexity manifests itself in several ways, but it is very peculiar in the inner workings of structured finance - securitizations. Tranching a pool of assets accomplishes a hierarchy of loss positioning which, theoretically, enhances the risk characteristics of parts related to the whole. If you own a senior piece, for example, you are comforted by the "thickness" of the tranches above you who will take losses before you do. That was the "genius" of securitizations in turning risky assets, like subprime mortgages, and making them "safe" enough to be rated AAA. It wasn't that ratings agencies didn't realize there were subprime mortgages contained in those structures, it was that ratings agencies surmised, as did the entire mortgage "market", that the thickness of each structure's junior tranches were sufficient to cover any anticipated episode of expected loss situations.
There was no shortage of mathematical modeling, elegant and sophisticated stuff, backing up those assertions. And the "market" itself largely accepted them on their face. However, there still was the issue of pricing, the missing ingredient in turning this specialized financial element into a mass-"market" product. Since it was impossible to price individual mortgages, Wall Street had to devise a statistical process of what really amounted to representations.
In pricing credit, perhaps the primary statistical factor is correlation. That is especially true of illiquid instruments, as inferred correlation takes on the heavy-lifting. That makes intuitive sense, given that if correlations are modeled to be high and some common factor is believed to be probable, then that factor's potential to create defaults will likely spread across wide proportions. In other words, in incidents of high correlation, one default means many more to come. If correlation ever achieved 100%, that would effectively mean one single default will signal that every loan will eventually.
There was no obvious means to measure correlation in mortgages, particularly since, again, individual loans do not trade. But in 2000, David X. Li published a paper where he mined correlation from credit spreads of credit default swaps. Without getting too far into the math, he essentially surmised, mathematically, that if two assets had similar spread curves they were correlated to some degree. Using the Gaussian copula, he modeled those characteristics as correlation. It was very similar in nature to the way various option programs infer volatility from related market inputs.
The introduction of the Gaussian copula was the spark that ignited the securitization frenzy. Each Wall Street bank used it in some manner, adding their own black box models to it, and began selling structured finance far and wide. The pricing inputs were largely derived from the most liquid segment of the MBS markets, the ABX indices. Again, the theory here was that "markets" were pricing these factors on their own and that these ABX prices incorporated elements like correlation that these mathematical formulas were simply extracting. The ABX and correlation inferences allowed the creation of synthetic structures, and more exotic intonations like CDO^2.
By the time 2007 rolled around and Morgan Stanley had put on its "short" trade, correlations were inferred to be rising as ABX prices began falling. Inside tranches, pricing exhibits what is known as a correlation "smile." In other words, correlation is not consistent throughout the structure. That is owed to the fact that the Gaussian copula and its descendants did not perfectly extract correlations, and that the only way to make the math work using these "market" inputs was to show different correlation values across tranches - the smile (or skew) was simply where priced correlations were higher in the ends than the middle.
What that meant for traders on the ends was a greater velocity of price changes for each inferred change in correlation. There is some intuitiveness to that that transcends the math - you would expect correlation to be much more important in the first-loss piece, for example, because of what I wrote above. High correlation means that if default events are expected to be low, then there won't be many defaults across your sector or loan pool - everyone will default or no one will. That holds for the senior pieces as well, because if correlation is inferred to be higher, the risk of wiping out the junior tranches increases.
But that pricing/correlation mechanism doesn't change in linear fashion, thus negative convexity, as there is often a non-linear and disjointed progression that can turn geometric. In Morgan Stanley's trade, the short position was likely paired with a long hedge position in super seniors of greater size (the estimates were for $10 billion). Where negative convexity plays a role is in its impact across tranches. While prices of the less-senior tranches were declining due to fears about default and cash flow deterioration, the price of the super senior would exhibit very low volatility, and even rise as an expected counterbalance - until implied correlations hit that magical point where the correlation smile produces larger price volatility in the ends relative to the middle. Where this trade seemed like a perfect expression of a hedged bearish outlook (make money on the middle tranche losing value while the super senior adds a hedge protection), it was instead an example of not just bad trade mechanics, but an entirely fragile paradigm.
The problem of this illiquidity transforming into pricing inputs was the Gaussian copula. As estimated prices began to decline once "markets" realized their modeled assumptions were too narrowly attuned, that led to a positive feedback loop. Lower prices increased the demand for hedging, both for cash flow and regulatory leverage needs. But hedging illiquid mortgage structures meant, more often than not, shorting some version of ABX or similar, or buying default insurance on them (credit default swaps). Hedging in these ways meant not only increased downward exertion on ABX prices, but doing so in much the same manner and degree - to which the Gaussian copulas inferred as rising correlation. As further inference of correlation rose in response, prices in structured tranches fell further, especially the behemoth super senior pieces, thus increasing the need to hedge still more; and so on.
As that cesspool of negative convexity and ABX deducing met the new FAS 157 accounting standard, now requiring banks to mark-to-market based on these correlation prices (and tranches were, in fact, quoted pricing in terms of correlation), it began the process of large bank losses that featured no cash component - it was strictly an accounting loss based on observable inputs. That was why, on the whole, bank managers were adamant the mortgage problem (at least as it related to banks, there was certainly a systemic credit problem) was being overstated in late 2007, and remained defiant that these huge losses were temporary and not fully representative of their books.
As 2007 turned to 2008, the accounting losses fed into other channels of rising dysfunction. Funding markets, already strained, became fully unworkable, as fragmentation appeared in eurodollars and elsewhere, making the wholesale system, really, fully irreparable. That also meant that "insurance providers" in default swaps, including AIG and the monolines, were unable to fund, and thus write, additional protection where the "market" was now unable to absorb the massive increase in hedging demand. And so the feedback loop, now effectively a noose, grew tighter and tighter. The more illiquid default insurance and swaps grew, the more volatile and sinister "observable inputs" pushed loss projections at the banks, visibly confirming the ultimate mistrust of funding participants that the FOMC was working so hard to bridge.
That even extended directly into the stock market. As credit default insurance grew more sparse and illiquid itself, banks found some hedging shelter by shorting the stocks of other banks or the entire banking sector. The "evil" speculator short sellers so vilified in that age were, in large part, those very same institutions forced into such desperation. And the more ubiquitous downward pressure grew due to illiquid inputs, the more dysfunctional funding markets became, the fewer avenues to obtain hedging, and it all piled and rippled into the prices of nearly every financial asset.
And into that vicious cycle is introduced the doubt about "markets." At the lowest point, even as early as the weeks before Bear's demise, there began questions, and really conceit, about what these financial prices were referencing. Because of the heavy strains of illiquidity flowing through these processes, it was almost easy to dismiss prices of "toxic" assets as a function of that illiquidity rather than any meaningful cash loss or actual default risk. You can hear that sentiment echoed throughout the discussions in 2008, inside the FOMC and elsewhere, which only fed the distrust of "markets" that I contend is both growing and inherent in this ever-centralizing system.
In some very important ways, I think this relates to Bernanke's assertion of the markets not understanding "hypotheticals", as cloaked within that sentiment is this ultimate distrust of "markets" in their pure form. And you can understand, to some degree given these episodes and mathematical fancies of 2008, why policymakers might believe as much. What were markets actually pricing in 2008; was it simply illiquidity or was it true fundamental value? Knowing the inside of these systems, you cannot help but conclude that it was the former rather than the latter as conventional wisdom holds. These were "toxic" assets only in popular sentiment, at least extending to what banks actually held - the senior pieces.
As hard as that might be for some to recognize and accept, it is confirmed by everything that has occurred since 2008 (in that narrow financial regard). AIG's assets never lost a nickel, and the Fed, through Maiden Lane(s), "made money." To my knowledge, no super senior tranche actually experienced a cash-basis loss, and any of the "lucky" hedge funds that were conveniently positioned for banks to offload their balance sheets were rewarded in multiples of their investment.
In some sense, then, you might conclude, as I believe Bernanke and his compatriots have, that 2008 represented the absence of "markets", thus justifying all the heavy lifting that took place then (even though it was ultimately futile), and has taken place since. In other words, the Fed was, and has been, trying to restore markets that were broken in panic.
Despite everything I have laid out here, I have to come to the exact opposite conclusion. We may think of prices and markets as representations of "true" or fundamental value, and that is unarguable in many circumstances, but I think the panic of 2008 shows that there are other dimensions to markets that contradict these orthodox interpretations.
Around the time Bear Stearns was failing there appeared anecdotes and stories of correlation trading of MBS tranches defying logic and mathematical reality. I'm talking about correlations, both inferred via observable inputs and quoted, that were approaching and even above 100%. A correlation of 120%, for example, makes no logical sense; it is a mathematical impossibility. In all honestly, a correlation of even 60% makes little sense, as, thinking broadly, you would never expect such widespread default short of Armageddon. In a similar manner, the credit default swap pricing on national debt during those days, including the US, jumped even though actual default was also a likelihood of almost exclusively the worst of the worst cases.
What those mathematical expressions were telling us, breaking out as "market" dysfunction, was essentially that the previous "market" paradigm was ever so badly calibrated to the real world and all its messy possibilities. In the context of 100+% correlations, you can see, I hope, how that is just such an expression of models that were unrealistic in both their creation and evolution. What might cause such widespread miscalibration, and thus such misallocation? I don't think you have to look very far for an answer, as all these mathematical shortcomings can be reduced to the common denominator of mispriced risk. That itself is an element of nothing other than monetary policy.
Illiquidity of that age was just an expression of all of these models breaking down in exactly, or nearly so, the same fashion. It was a systemic recalibration of not just mathematical understanding, but of rebuking the previous paradigm that countenanced such dangerous imprecision. In other words, pricing may not have been relating fundamental or cash flow-basis "value", but those prices were more "market" than anything seen in the years (decades?) prior; or since. There is not enough explanatory power in saying the "market broke" in 2008, as that would not sufficiently describe everything that occurred when you observe all the profit motives of all involved. Morgan Stanley's experience shows that, as they themselves were trying to arbitrage that as a market breakdown, instead burned by their own inability to see beyond what was really an artificial paradigm.
It cannot be that the entire system became abhorrent in the same fashion at the same time, and that is the dysfunction - the malformation can only have predated the breakdown, not have been contained within it. How could everybody, from the most sophisticated on down, been so wrong at the very same time? Real markets are not so narrow without artificial constraints. We are not talking about soybean futures here, we are talking tens of trillions, hundreds if you include derivatives, spread across the entire face of Earth. No such true market would be so engrossingly fragile.
That raises the issue now of what are we seeing in asset prices again. Have we seen another dichotomy grow in much the same fashion, where a paradigm of narrowness has been instilled into much the same structures (with different names; where subprime MBS stood years ago now fills with "leveraged loans")? We know without fail that these artificial markets manifest themselves as profit motives not necessarily concerned with fundamental values, but rather leveraged positions of flow and distorted behavior. I have argued that we have seen the contours of that sentiment expressed again in the past ten months or so. In short, there is a perpetual battle between the artificial market, bred under cover of policy and institutional distrust, and the organic market as it adheres only in discrete episodes (some more evident than others).
If the mistrust of markets is both genuine and valid, that would back the statist position of perpetual intervention "for our own good." If, however, I am correct that true markets were only revealed particularly so destructively in 2008, and that intervention is sowed solely to regain that artificiality, then efficiency itself demands the true marketplace. The results of inefficiency would be a perpetual sequencing, or what you might call bubble repetition and the condensed cycle of depression.