|About ADBINews & EventsSpecial ProgramsPartnerships|
|HomePublicationsCatalogThe Unfolding Turmoil of 2007–2008: Lessons and ResponsesOrigins of the Crisis: What's New, What's the Same, and What We Don't Understand|
Origins of the Crisis: What's New, What's the Same, and What We Don't Understand
Most commentaries about the crisis have focused on the unusual structures of the financial system and the role of excessively complex financial innovations. It is true that an important factor in the run-up to the turmoil was the reckless use of recent financial innovations, especially in markets for credit risk transfer. However, as Borio (2008) argued, these are but idiosyncratic elements that represent “more fundamental common causes.” Indeed the current turmoil has displayed a number of important features that it shares with previous crises. At the same time, it is important to acknowledge that there are critical characteristics of the current crisis that we do not fully understand.
A. The chain of events
To understand what triggered the crisis and how it spread, it helps to follow the chain of events that preceded the crisis' onset. Two types of interest rate spreads are especially helpful in tracking events as they unfolded: the spreads of three-month London Interbank Offered Rate (LIBOR) over the overnight index swap (OIS) and spreads on credit default swaps (CDS). The former is a good indicator of liquidity in interbank markets and the latter of credit risk premia. As shown in Figure 1 [ PDF 72.8KB | 1 page ], LIBOR-OIS spreads in the United States (US), the euro area, and the United Kingdom (UK) rose sharply in August and September 2007, rose again in November and December 2007, and yet again in March and April 2008, each case indicating a lack of liquidity. For their part, average credit default swap (CDS) spreads in the same three regions followed almost the same pattern, except that the widening of these spreads became more pronounced in each successive episode, especially in February and March 2008. There was also a jump in CDS spreads in July 2008 that was not echoed in LIBOR-OIS markets. What led these spreads to behave in this way?
Although US subprime mortgage default rates, and the spreads on associated securities, had been rising since late 2006, the first significant event in the broader financial-market turmoil seems to have been the emergence of rumors during the third week of June 2007 about heavy losses in two hedge funds managed by Bear Stearns.1 The losses were later confirmed, and they turned out to be related to positions in US subprime-backed structured securities. By July, major credit rating agencies had either downgraded or placed on review a large number of collateralized debt obligations (CDOs) that relied on mortgages as collateral. In August, the troubles spread to asset-backed commercial paper (ABCP) issued by entities that had invested in CDOs of mortgage-backed securities, and interbank markets around the world began to experience shortages of liquidity. On 9 August, the market received the shocking news that BNP Paribas, France's biggest bank, halted withdrawals from three of its investment funds because it could not “fairly” value their holdings. All these events culminated in September with a run on Northern Rock, a UK mortgage lender, when its liquidity problems became known. During this phase, the LIBOR-OIS spread rose to close to 100 basis points in the US interbank market and even higher in the UK market.
The second major event occurred in October 2007 as participants in the interbank and credit markets were again caught by surprise when large monoline bond insurers revealed losses related to credit enhancements they had provided to structured securities; not many observers knew that the insurers had even been involved in this business. The losses were large enough to threaten the AAA/Aaa ratings that the monolines needed in order to operate. In December, mounting concerns by various lenders, waves of margin calls in the repurchase markets (hereafter known as repo markets) and the anticipation of increased liquidity demands over the year-end led to widening CDS spreads and a second liquidity crisis in the money markets, prompting five major central banks to announce that they would take concerted action to make more funding liquidity available. The LIBOR-OIS spread in the US and euro area interbank markets reached unprecedented levels, while in the UK these spreads rescaled the peaks they had reached in September.
The third and perhaps most alarming period of the turmoil is also the most difficult to explain. In late February and early March 2008, a new wave of deleveraging suddenly engulfed the fixed-income markets. There seems to have been no significant event that could have precipitated this episode, although mounting concerns about monoline insurers, the continued worsening in the US economic outlook, and associated valuation and liquidity problems in high-yield corporate debt and in both prime and non-prime US housing-related paper certainly were part of the background.2 What is evident is that concerns about counterparty risk became extraordinarily intense. Dealers in mortgage-backed securities and in over-the-counter (OTC) derivatives started asking for more collateral from their counterparties. In repo markets, lenders sharply increased their margin calls and refused to accept as collateral anything but US Treasury securities or German government bonds (Bundesanleihen, or bunds). Since bond dealers finance themselves in the repo markets, they abruptly withdrew from making markets in the broader fixed-income markets. Liquidity in US and European fixed-income markets seemed to vanish overnight.
In early March 2008, the news headlines began to report what has come to be known as an adverse liquidity spiral, in which Bear Stearns, a US investment bank, was rumored to be caught. On 16 March, after several days of customer outflows, shrinking capital, and fevered activity on Wall Street and at the US Federal Reserve, the troubled investment bank was taken over by a US commercial bank, JP MorganChase, with the help of an arrangement by which the Federal Reserve would lend up to US$30 billion (later reduced to US$29 billion) in order to finance Bear Stearns' portfolio of troubled securities.
Markets appeared to stabilize in the aftermath of the Bear Stearns episode, but remained subject to episodes of turbulence. Equity and debt markets were buffeted by concerns about the prospects for various segments of the US financial system, as losses spread to government-sponsored enterprises (GSEs) and small- and medium-sized banks. As of the summer of 2008, the impact of the financial system's troubles on the real economy remains a major source of uncertainty worldwide.
B. What's new
The blame for the turmoil has been linked to a wide variety of financial innovations. In particular, many observers have argued that, while these financial innovations have been fundamentally beneficial for the financial system, a reckless use of them has led to the crisis. Borio (2008), for example, identified the two most salient innovations as structured credit and the originate-to-distribute business model. As will be discussed in more detail below, we would suggest that, while these have been sources of substantial losses and uncertain valuations over the past twelve months, a little-noticed innovation in the repo market—the triparty repo3—also contributed significantly to the crisis.
The innovation of structured credit includes both CDOs and CDSs. CDOs use the device of subordination to transform instruments with high credit risk into instruments that receive high credit ratings. The introduction of CDS contracts in turn allows CDOs to be created more easily by serving as the underlying instruments for what are called “synthetic CDOs.” The pricing of these instruments relies critically on assumptions about default correlations, which have been intractably difficult to model or measure.4
These instruments were developed in the 1990s and proved to be fundamentally important in improving the pricing and distribution of credit risks. But in the environment of the past few years, their use became quite widespread and complex variations on the instruments proliferated rapidly, aided by advances in modeling techniques. These variations included CDOs based on asset-backed securities (ABS), as well as ABCP issued by structured investment vehicles (SIVs) that held highly-rated credit instruments, including CDO, tranches. The sheer scale and variety of the use of these innovations outstripped the capacity of even the most sophisticated dealers and investors to understand and manage the risks associated with them.
The proliferation of CDOs owed much to the originate-to-distribute business model pursued by many of the world's largest commercial and investment banks. The model itself is not new, having been used in the syndicated loan market for years. Nonetheless, it achieved new prominence in banks' business strategies with the securitization of mortgages. Under this model, a mortgage lender would routinely package its loans into mortgage-backed securities, which it would sell to investors, thereby providing funding for the loans. In the period leading up to the turmoil, the originate-to-distribute model contributed to the rapid growth of the US mortgage market (as Frankel  has emphasized), but also evidently weakened the incentives of mortgage originators to properly screen loans. Once the first subprime mortgage defaults materialized, standard covenants of the securitizations forced the originators to take back newly minted loans. However, these types of securities had become so widely dispersed that a generalized crisis of confidence ensued. Banks were stuck with “warehouses” of unpackaged loans that could not be sold and rapidly lost value as markets fell. They also discovered that the “super-senior” CDO tranches (the tranches with the highest priority claims within the CDO structure), which the banks typically retained in order to facilitate the sale of other parts of the structure, were far riskier and less liquid than they had expected.
While the problems in CDS and CDO markets have been much commented upon during the crisis, the role of the tri-party repo has attracted less attention. This innovation has become the standard way of transacting in repo markets. In a tri-party repo, the third party is a clearing bank, which knows both the lender and borrower of a repo transaction and takes custody of the collateral. The arrangement has the advantages of avoiding delivery of collateral, facilitating substitution of collateral, and reducing clearing costs. In recent years, the tri-party repo has allowed the repo market to accept a wide range of collateral, including mortgage-backed securities, CDOs, and almost any asset that the clearing banks could hold in custody. Hence, it has allowed investors in various securities to more easily secure financing in the repo market by simply putting up their positions as collateral. However, an important factor contributing to the loss of liquidity in fixed-income markets in February and March 2008 was the sudden refusal of lenders in the repo market to accept as collateral the same wide range of assets as before. This made it difficult to impossible for holders to value the instruments and led to a sharp worsening of the liquidity profile of institutions—in the case of Bear Stearns, which had a large quantity of these assets on its balance sheets.
C. What has remained the same
The focus on what is new suggests that episodes such as this 2007–2008 unfolding financial turmoil are “black swans”—events so rare and unexpected that there is little that can be done about them. The truth, however, is that (just like real black swans) the underlying causes of the turmoil are, in many ways, familiar.
The early part of this decade saw a long period of unusually easy macroeconomic conditions, with low or negative real interest rates in the major economies and a glut of savings centered in East Asia and the Middle East. In this environment, the global weight of excess savings and excess liquidity fed a steady run-up in asset prices, especially in credit instruments and housing markets, which, in turn, encouraged a build-up in leverage and risktaking, among both regulated and unregulated entities. When the environment turned bad, the overextension of risk resulted in heavy losses and a rush to unwind leverage. As shown in Figure 2 [ PDF 26.4KB | 1 page ], this risk-taking behavior has resulted in a correlation between credit growth, asset prices, and the real economy in what Goodhart (2004) has termed the “excessive procyclicality” of the financial system.
An environment that was increasingly tolerant of risk was evident in the tendency of banks to take on more risk and in a decline in risk premia to low levels, especially in the case of credit instruments. As shown in Figure 3 [ PDF 26.4KB | 1 page ], the Value-at-Risk (VaR) estimates that banks themselves reported show that the banks took on more risk. This is especially striking given that the realized and implied volatilities of most major asset classes, which form a central input to VaR calculations, were falling steadily throughout this period (Bank for International Settlements (BIS) 2006). Even more striking was the steady decline in credit spreads as reflected, for example, in such traded CDS indices as the iTraxx in Europe or the CDX in North America.
But as the upswing in markets gathered pace, there were also important qualitative shifts in financial market participants' attitude towards risk-taking. At many large financial institutions, forward-looking risk assessments were poor, reflecting poor risk measurement and poor governance of risk-taking within those institutions. Misaligned incentives cropped up throughout the financial system, as the penalties for poor decisions were ignored. Subsequently, disclosure weakened, in part because investors slackened in their demand for it. Reviews by supervisors have also made it clear that some banks managed these risks substantially better than others, thanks to closer engagement by senior management and more effective internal controls (Senior Supervisors' Group 2008). This suggests that at least a portion of the subsequent losses suffered by many banks were by no means inevitable.
In this world of ravenous appetites for risk, market participants became increasingly willing to hold rather complex instruments of unproven liquidity and increasingly reluctant to apply sound risk management practices to them. In the end, mistakes in the valuation and risk management of these instruments turned out to be the critical errors that triggered the crisis. Risk management errors with respect to the super-senior tranches of CDOs of sub-primemortgage backed securities evidently had especially significant systemic effects, thanks, in part, to maturity and liquidity mismatches involving these instruments. When investors lost faith in ABCP that had been issued by conduits and SIVs holding these instruments, a credit risk event turned into a liquidity event.
1. What we don't understand
The combination of a credit risk event and a liquidity event seems to have led to the unique depth and duration of the current crisis. Over the past decade or two, financial markets in the developed economies have become pretty good at absorbing large losses, resolving them, and moving on, albeit usually with an altered set of players and altered judgments about risks. In the current turmoil, by contrast, the underlying functioning of the system has come into question. Whole classes of previously abundant assets can no longer find buyers – notably CDOs of ABS, and the instruments based on them, but also seemingly unrelated products such as municipal auction-rate securities.
While stories explaining the turmoil abound, these remain “Just So Stories.” There are many fundamental things we just do not (yet) understand. How, for example, could defaults in a relatively small corner of the US mortgage market lead to such massive losses in broader credit markets and turn into a global turmoil of such proportions and such long duration? Greenlaw et al. (2008) argue that leverage was a major contributing factor. While it is clear that leverage did play a role in the magnification of losses, it is still puzzling how instruments that were designed to spread and diversify risks ended up concentrating the risks.
The sudden evaporation of market liquidity has been even more surprising. Three-month spreads in the international interbank lending markets widened in August 2007 and have remained wide ever since. By late February and early March 2008, investors had seen the near cessation of trading activity in all but the most liquid government securities. Repo markets, supposedly the most robust source of funding liquidity, experienced what can only be described as a run on all “eligible” collateral except for the highest rated government bonds. How could liquidity disappear so suddenly from markets that had not seen any sign of defaults or even credit rating downgrades? Brunnermeier and Pedersen (2007) demonstrated that the interaction of market liquidity and funding liquidity can generate what they describe as a liquidity spiral. Nonetheless, an explanation of the sudden disappearance of funding liquidity during February and March 2008 remains beyond the scope of their paper.
Deleveraging and the hoarding of liquid securities by market makers who also happen to be investors in a broad range of markets played some role in these phenomena. Indeed there were bouts of deleveraging in August 2007, December 2007, and February and March 2008. The first two episodes may have been triggered by disconcerting news about losses in hedge funds or banks. But it is hard to identify a specific trigger for the February-March episode. In conversations with market participants, all they can say is that they just had a “bad feeling about things.” The fact that many market participants seemed to get that bad feeling at about the same time suggests that a common factor was at work. But what that common factor was remains a mystery.
In trying to resolve the liquidity issues, monetary authorities have been mystified by the persistent stigma associated with borrowing from the central bank and the fact that this gets worse at the very time when such borrowing becomes most critical. As discussed below, it is also surprising how the simple mechanism of an auction can make such stigma go away.
Download this Paper [ PDF 182.4KB| 18 pages ].
[previous chapter] [next chapter]
There are  comment(s) for this entry. Post a comment.
|Back to Top|
|© 2014 Asian Development Bank Institute.|