The trends observed from 1980 through to 2002 were gradually making the North Atlantic financial system more brittle. The rapid expansion of the balance sheets of the Euro area mega-banks and erosion in their capital buffers were creating risks for that region, especially against a backdrop of a monetary union that was badly designed for negative shocks. In the United States, the boom in shadow banking involving the nexus of securitization, non-insured deposits, and investment banks was well under way and house price inflation was starting to gain steam. However, these strains were limited and essentially isolated from each other.
The boom of the 2000s saw these risks grow and intermingle (Figure 19). The core Euro area banks were central protagonists in the unsustainable financial boom and bust, in stark contrast to the popular portrayal of them as hapless victims of US financial deregulation. To be sure, US deregulation played a role. In particular, reforms designed to provide greater liquidity in US repurchase (repo) markets supported the explosion of substandard mortgage-backed assets and gave foreign banks easy access to dollars. However, this coincided with a massive boom in lending by the core Euro area banks as a result of inadequate regulation and supervision. As domestic lending opportunities dried up, the core Euro area mega-banks expanded into foreign markets, financing housing booms in the United States and the Euro area periphery. Indeed, this lending was a major catalyst in making these housing and financial bubbles larger and more destructive—the road to the housing crashes went through Basel and Berlin, not Baltimore and Barcelona. The subsequent collapses in house prices created banking and economic crises in the United States and the Euro area periphery that blew back onto the Euro area core, broadening the crisis to the entire North Atlantic region.1 The jaws of the North Atlantic crisis clamped shut.
This is not to say that the banks in the Euro area core all behaved alike, or that the booms in the Euro area periphery did not have their own national flavors. The broad-brush split of the Euro area into a core and a periphery masks significant differences across countries. Within the core, the German banking system included large numbers of small banks whereas larger institutions dominated the French banking system. Less sophisticated German banks bought significant amounts of questionable US assets, including the cash-rich Landesbanken (owned by the German states, or Länder) whose spending spree in the mid-2000s was further boosted by a hump in borrowing before the option to use government-guaranteed bonds expired. Turning to the periphery, the Spanish housing boom was much larger than the one in Italy. This largely reflected the activities of unsophisticated local-government-owned “caja” savings banks that drove much of the Spanish housing boom. Indeed, the cajas followed a similar trajectory to the Landesbanken in terms of investing heavily in real estate, except that the bad investments were national instead of international. By contrast, in Italy banks loans were much more closely linked to local firms. For that matter, there were also major differences across US regions, where house price increases were much larger in the west and east coasts (especially Arizona and Florida) than in the middle of the country. In the end, the similarities between the behavior in the Euro area core (Germany, France, the Netherlands, and Belgium) and in the periphery (the remainder) justify a narrative that splits the Euro area into these two regions, just as the United States is treated as a single entity. These simplifications allow a coherent story to be told without multiple digressions.
* * *
Regulators and the Cross-Atlantic Financial Drift
Pro-market changes in regulation helped propel a cross-Atlantic financial drift in which Euro area banks financed the US housing bubble by buying increasing amounts of securitized US mortgage loans. Regulators focused on updating rules that supported markets and hence investment banking. This motivation was different from the earlier wave of regulatory changes in the 1980s and 1990s, which had aimed at integrating the European financial system and freeing regulated US banks from outdated depression-era rules. These objectives had essentially been completed by 2000 with the introduction of the Euro (helping to create a single European financial market), the repeal of the US Glass–Steagall Act separating commercial and investment banking (integrating the US banking system), and the Basel 1 Capital Accord on international capital buffers (harmonizing rules on capital charges across international banks). The new push was in response to lobbying by banks to make markets more efficient and, by extension, increase their profits. This accession to bank self-interest was to have tragic consequences.
How US Repo Markets Funded the Housing Bubble
The expansion of collateral that US broker-dealers could use in repurchase agreements with clients was the first and most important of the regulatory decisions aimed at enhancing markets. In April 2003, Release 47683 of the US Securities and Exchange Commission (SEC) widened the collateral that US broker-dealers could use with clients from “cash, U.S. Treasury bills or notes, or irrevocable bank letters of credit” to include two major additional types of assets.2 These were, firstly, foreign paper such as sovereign bonds, highly rated non-government securities, certificates of deposit, and bankers’ acceptances, and, secondly, securities associated with the US housing market. Crucially, the second category included mortgage-backed securities.
The SEC release was responding to complaints from the US investment banking industry that the size of the repurchase market was being constrained by limited amounts of eligible collateral. Repurchase agreements were a flexible and convenient instrument to borrow and lend cash on a short-term basis. For example, an investment bank could sell a financial asset to a client in return for cash and an agreement to buy the asset back (repurchase it) the next day at a fixed price, with the difference between the current and future prices representing the return for the depositor. The investment bank then either used the cash to invest in the market or, more typically, lend it to a market trader such as a hedge fund using another repo. Clearly, collateral was central to these arrangements as it provided the client with an assurance that they would get their money back in case anything went wrong.
In a political and intellectual environment that was sympathetic to loosening financial regulation, the SEC assessed that the benefits from deeper repo markets to investors and US investment banks outweighed any risks to clients. The release explained that benefits included “lowered borrowing costs and increased liquidity in the securities lending markets, and greater opportunities for US firms to compete abroad”.3 It observed that broker-dealers (the legal description for the firms at the core of the US investment banking groups) often found it more economical to borrow and lend with each other than with clients. This was because there were no restrictions on the collateral that could be used in repos with fellow broker-dealers, in contrast to the restrictions on the collateral employed for such transactions with clients. Allowing a wider range of assets to be used as collateral would allow clients to lend more to broker-dealers and “earn the fees associated with such transactions and thereby realize greater returns on their securities portfolios”. These benefits, which not incidentally also included solidifying the pre-eminent place of US broker-dealers in global markets, were deemed to outweigh the risks “of customer losses associated with permitting a new category of collateral”.
Two years later, similar motives led the US Congress to give repo markets a further boost with the passage of the Bankruptcy Abuse and Consumer Protection Act in April 2005. This act extended “safe haven” protection to most of the expanded repo collateral granted by the SEC two years earlier.4 Safe haven status meant that if the broker-dealer went bankrupt the lender was able to automatically keep the collateral behind the repo. By contrast, if the broker-dealer borrowed money by (say) issuing money market mutual funds or commercial paper, lenders could only get their money back through a lengthy bankruptcy procedure. Safe haven thus increased the attractiveness of repos over other ways of lending money.
The justification for extending safe haven status to the expanded collateral was to limit the systemic risks of repo transactions. There was a concern that if clients started to worry about the financial stability of a broker-dealer, this could create risks to the wider financial system, given the size of the repo market. But there was a strong element of circularity in this line of reasoning. First, the SEC widened the repo collateral after investment banks complained that a lack of safe collateral was constraining the market. Then the resulting expanded repo market based on less secure collateral was supported through changes to the bankruptcy code. The Congress was trying to minimize a concern about riskier collateral that had been created just two years earlier by the SEC. The ultimate irony is that the collapse of the investment bank Lehman Brothers created exactly the run on the repo market that the safe haven provision was supposed to avoid. This was because over the crisis lenders to the investment banks simultaneously questioned the viability of banks and the value of the underlying collateral. The safe haven provision protected them from the first of these concerns but not the second.
The third and final regulatory lift to the repo market, particularly as regards the participation of foreign banks, came when the Basel Committee loosened its capital requirements for repos in July 2005.5 These changes had been anticipated in the previous year’s agreement on the overall Basel 2 framework (discussed later). At that time, the Committee promised to “maintain its active dialogue with the industry to ensure that the new framework keeps pace with, and can be applied to, ongoing development in the financial sector”.6 One ongoing development where it had indicated that “immediate work should be done” concerned exposures to “certain trading activities” (which was code for repos).
These seemingly technical changes to regulations sparked two trends that drove the US housing bubble and internationalized the spillovers from the eventual bust. The first trend was the increased participation of foreign banks in the US repo market. Before collateral was widened to include foreign paper, overseas banks had almost no presence in this market. By contrast, they were major players by the time of the crisis, using foreign paper to obtain dollars that were then generally invested in US assets. As the SEC predicted, easier access to the repos market further deepened the already substantial role of US assets in international banking, including the booming market for mortgage-backed securities. The deep involvement of European banks in US repos markets explains why the seizing up of these markets after the collapse of Lehman Brothers over the crisis immediately translated into dollar and liquidity shortages at European banks.
The second and even more corrosive trend was to increase the desirability of high-yielding mortgage-backed securities issued by private firms. Before the SEC decision, mortgage-backed securities represented a useful way of pooling risk and selling bundles of loans to investors. With a stroke of the pen, the SEC also made mortgage-backed securities into liquid assets that could be swapped for cash in the repo market. This added a powerful incentive for traders to own, lend, and sell mortgage-backed securities. As these securities increasingly became a vehicle for borrowing and lending cash, purchasers became progressively more focused on the short-term goal of obtaining high returns rather than on the long-term business of providing people housing through sound mortgages.
This search for higher yield boosted the demand for riskier private label securitizations over the safer but lower-yielding assets provided by the GSEs. Private banks were able to offer better yields on mortgage-back securities because they could accept riskier mortgages that by-passed the rules on safety and soundness for “conforming” mortgages that GSEs could accept. In addition, unlike the GSEs, private banks did not provide insurance against the risk that the underlying mortgages would default. This moved the potential losses on defaulting mortgages as a result of a US housing collapse away from the GSEs, with their deep pockets derived from implicit government backing, and toward shadow banks with thin capital cushions.
The nonconforming mortgage loans bundled by private banks in securitized assets included “jumbo” mortgages, whose value exceeded the upper limit on loans that Fannie and Freddie would accept, and where demand was rising along with house prices. In their desire to satisfy the rising demand for high-yielding mortgages, however, private firms increasingly bundled mortgages from borrowers that did not meet the “prime” credit standards required by Fannie and Freddie. Expanding the quantity of “subprime” mortgages was relatively easy as they were being offered to people who wanted to own a home but had been previously locked out of the mortgage market. The expansion met with little opposition from regulators as it coincided with the federal government’s desire to widen home ownership.
The subprime mortgage market exploded after the SEC decision to widen repo collateral to mortgage-backed securities. In 2002, only 7 percent of US mortgages were subprime and only half of those were sold to investors. By 2006, subprime loans were almost half of the market and almost all were sold on to investors; research also finds a major fall in the spread of mortgage-backed securities to treasury notes in the summer of 2003, soon after the SEC decision.7 Because the banks sold the mortgages, they no longer had “skin in the game” and the quality of subprime loans deteriorated to the point where they became known colloquially known as “liar loans” due to the lack of due diligence on the borrowers.8 Due diligence was skimped as banks and investors increasingly saw house price appreciation as the main guarantee against default rather than the ability and willingness of borrowers to repay. This unconventional lending approach was justified by the observation that that US national house prices had never fallen in the postwar period. Unsurprisingly, after this assumption proved incorrect and house prices did start to fall in 2006, the default rate on these loans rapidly escalated. The market collapsed along with its central dogma.
Obviously, changes in rules on repo collateral was not the only driver of the US housing market bubble. Many other factors were involved, most importantly a massive failure of US consumer protection for subprime mortgage borrowers which allowed mortgage lenders to offer increasingly unsafe loans based on miniscule downpayments, low initial charges on the loan, and minimal documentation. This lax attitude largely reflected the Federal Reserve’s overinflated belief that risks taken by banks would be limited by market discipline from investors (the Federal Reserve was the main regulator of mortgage lending standards). However, this deterioration in lending standards only emerged once the subprime market had already been established. The initial takeoff of the subprime market occurred on the heels of the SEC decision to widen repo collateral in 2003. More generally, while the internal dynamics of the US mortgage market boosted the US housing bubble by increasing the supply of mortgage-backed assets, much of the demand for these assets came from the Euro area as expanding northern banks created a cross-Atlantic financial drift into US markets using repos. It is to this story that we now turn.
* * *
How the Basel Committee Encouraged Euro Area Mega-Banks
The most important change to international banking regulations in the early 2000s was the 2004 Basel 2 Accord. In this landmark decision, the Basel Committee agreed to widen the use of risk models to credit risk on commercial loans to firms and households in addition to market risks from trading securities. Since Basel 2 was not implemented until the end of 2007 for the large Europe banks and never really adopted in the United States, the direct impact of these regulations on the North Atlantic banking system was limited. However, in anticipation of lower future capital requirements, European banks reduced the capital that they held in excess of current minimums. In addition, the Basel 2 negotiations illustrate the increasingly cozy relationship between the European mega-banks and their supervisors. This explains why these banks were able to manipulate their internal risk to thin their capital buffers on investment banking without regulators intervening.
The Long and Winding Road to Basel 2
The road to Basel 2 started almost immediately after the January 1996 passage of the market risk amendment, as the Basel Committee refocused on rules for commercial lending to consumers and firms. Most members of the Basel Committee agreed that the existing Basel 1 system that put each loan into a broad risk bucket was relatively crude, with the risk weights often deviating significantly from actual risk, but that it reflected a practical compromise between simplicity and complexity. US Federal Reserve Board officials, however, became increasingly strident advocates of altering this system. As early as May 1996, Chair Greenspan noted that the weaknesses in Basel 1 were becoming ever more evident. By 1998 “calls for change in Basel 1 became a regular theme in public statements by Federal Reserve Officials concerning bank regulation”.9
In addition to a general bias against financial regulation, Fed concerns with Basel 1 stemmed from two trends that were particularly advanced in the US banking system. The first was the expansion in securitizations of mortgages. Fed officials noted that the Basel 1 risk weights encouraged such securitizations since the weight assigned to a mortgage-backed security was lower than the weight assigned to the underlying mortgages.10 Fed officials were correct that securitization was being encouraged by differences in capital charges. However, standardized Basel risk rates were a minor issue. The main driver were differences in required capital coming from the high capital buffers embodied in the US prompt corrective action framework for US regulated banks versus the much lower capital buffers for securitized bundles of such loans required of US investment and major European banks that used internal models to calculate market risks. Fed officials were in the right neighborhood, but looking under the wrong lamppost.
Second, just as they had done for market risks, the Federal Reserve argued that risks from commercial loans were better calculated using internal “value-at-risk” models. The Fed had by now widened its view that market discipline was a better constraint on bank risk-taking than crude government regulations to include commercial banking as well as investment banking. The Fed argued that extending the use of internal models to credit risk would improve the calculation of capital buffers while also providing incentives for the major banks to further refine their risk models. It was therefore little surprise that a month after Federal Reserve Bank of New York President William McDonough became chair of the Basel Committee in June 1998 that its members agreed to review Basel 1.11 While the Basel 1 regime came from a need to unify and beef up international capital rules, Basel 2 was born from a vaguer dissatisfaction with the existing regime. The lack of a clear final objective complicated and confused the process. For example, it was never clear whether the plan was to maintain bank capital buffers at existing levels or to allow some erosion in light of the banks’ use of more sophisticated risk management tools.
Against this background, it is little surprise that the Committee’s conservative initial 1999 proposal to change the capital standards by using credit ratings to assign loans to risk weight buckets was badly received.12 Even the rating agencies opposed the plan, saying that it risked “ratings shopping” by clients, while the large banks (correctly) pointed out that many borrowers did not have external ratings and that the track record of such ratings as predictors of future risk was not encouraging. Instead, the large banks lined up behind the Fed and asked to extend internal models to credit risk. For example, the Institute for International Finance (a trade group for large banks that had already advocated the use of internal risk models) and the International Swaps and Derivatives Association each organized groups of senior bankers to lobby the Basel Committee in favor of using internal models.
The Committee’s second consultative paper, published in January 2001, caved into this pressure and agreed to let large banks use internal bank models to calculate credit risk. This was despite misgivings about the internal ratings approach by some European members of the Committee, most notably the German supervisors. Under the Committee’s proposal, small banks would continue to use the existing “standardized” approach to calculate credit risks while somewhat larger banks would be allowed to use the “foundation” approach, which combined bank estimates of the likelihood of default with supervisors’ estimates of the associated losses. Crucially, the largest and most important banks would be allowed to use the “advanced” approach in which internal risk models were used to calculate the probability of default and the associated losses on commercial loans—i.e., their capital buffers—just as they were already doing for market risks. Depending on their size and sophistication, banks could therefore be under three different capital adequacy regimes—the standardized regime, the foundation internal ratings regime, or the advanced internal ratings regime.
Reactions to this proposal were not tepid. The Committee receiving over a thousand pages of largely negative comments. Many academic and quasi-academic commentators (correctly) questioned the advisability of using internal ratings models given the potential for manipulation. By contrast, the large banks supported the framework but suggested numerous modifications arguing that, based on their initial calculations, internal risk models could lead to a rise in regulatory capital requirements.13 As observed by a knowledgeable commentator at the time (who subsequently helped clean up after the US financial bust) these detailed comments obscured the fact that:
In another sense, though, the banks’ objections to capital levels required by the CP–2 [the second consultative paper] was their only complaint … any feature of CP–2 that boosted capital requirements was unwelcome. On the other hand, the banks were quite rationally prepared to accept regulator y features that they found arbitrary, costly, or even ill-conceived, so long as their capital requirements declined enough to make the changes on net profitable [emphasis in the original].14
After the acceptance of internal risk models had demonstrated the ability of large banks to sway decisions, the Basel Committee engaged in more or less constant negotiations with the banking industry and frequent revisions of its proposals. The Committee was on the defensive, responding to a nearly constant barrage of criticism mainly from large banks, whose focus was on lowering capital charges and hence improving their competitiveness relative to banks that were not able to use the advanced approach.15 As a result:
In many instances, it is not easy to determine if the Committee’s modifications in response to the many objections from banks reflected an enhanced technical understanding of matters not fully grasped in CP–2, on the one hand, or tactical accommodations in the face of broad-based opposition, on the other hand (or some of both).16
The final Basel 2 proposals (released in June 2004) added further layers of complexity and hence more discretion on implementation of rules and standards.17 Reflecting these characteristics, it was agreed that the advanced system should come into force at the end of 2007, a year after the rest of Basel 2. Lobbying by the large banks produced a complex and opaque system using internal models that were easy to manipulate.
Internal Risk Models as the Golden Goose
Astonishingly, the Basel Committee agreed to extend internal models to cover credit risks without detailed information on the consequences for bank capital, just as had occurred earlier with the 1996 market risk amendment. As one commentator put it at the time, this “calls into question the degree to which the Basel Committee understood the implications of its proposal”.18 Eventually, the consequences of the switch from Basel 1 to Basel 2 for banks’ capital buffers were assessed by regulators through a series of Quantitative Impact Studies (QISs). While such studies started early in the negotiating process, because the proposals changed so much over time the only studies using assumptions fairly close to the final rules were the last two (QIS–4 and QIS–5). US regulators and the Basel Committee both published analyses of the results of QIS–4 and QIS–5 in 2006, well after the Basel framework had been agreed (in 2004) and adopted by the European Union in (2005).
The carefully written and highly factual report by four US regulators, which masked considerable differences in opinion across the institutions, found that Basel 2 facilitated a major reduction in bank capital.19 This quietly devastating study focused exclusively on the potential impact of the “advanced” internal ratings approach using the “best-efforts” of twenty-six US institutions. It found that capital requirements for a typical bank would fall by 26 percent, implying that on average banks would be able to add one-third more loans for the same capital base (confusingly, as a result of a dubious weighting procedure, the aggregate reduction was calculated as only 15 percent, but the 26 percent figure is a better estimate).20 Even more strikingly, the US regulators found wide variations in estimated capital charges across institutions. They reported that “much of the observed dispersion [in results] was caused by the underlying source data and methodologies used by institutions to quantify their risks”.21 While this partly reflected the fact that the institutions were at various stages of implementing the framework, the report noted that it also reflected the “flexibility allowed within the framework governing the selection of risk inputs” (emphasis added). They concluded that “portfolio composition was not as important a contributor to dispersion as differences in data and methodologies used”. The models were wildly inconsistent with each other when tested on identical portfolios, just as been found with the 1996 market risk amendment.
The Basel Committee study of the impact of Basel 2 found similar results for a wider range of countries that encompassed all three methods for calculating capital requirements.22 Using the same dubious weighting strategy, the study estimated that capital buffers would fall by 7 percent for banks using the advanced approach, broadly in line with the US results once changes in rules between QIS–4 and QIS–5 are taken into account.23 Like the US report, the Basel study also reported major differences in results across institutions, from a +60 percent to a –60 percent change in capital requirements.
The wide variation in estimates of needed capital buffers coming from different internal risk models is crucial to understanding why the European banks were able to manipulate these models. In the face of such diverse outcomes, it was extremely difficult for supervisors to assess the accuracy of any given model or whether changes to models reflected genuine improvements or a desire to lower costs by minimizing capital buffers. The Basel 2 negotiations had made clear that banks were more than happy to minimize their capital charges. Complex and widely varying internal risk models gave them an easy means to do so.
Another characteristic that became evident is that internal risk models amplify the business cycle. They encourage more lending during booms and less lending during slumps. This is because internal risk models assume that the rise in asset prices during a boom signaled a reduction in risk, which allow banks to reduce the capital support on existing loans and redeploy it to back new lending. As the Euro area was to find out over the crisis, the same mechanism applied during a bust, only with the opposite effect. The US regulators were the most explicit about the pro-cyclical nature of Basel 2, stating that “the relatively benign economic environment prevalent when QIS–4 was conducted resulted in lower minimum risk-based capital requirements than would had been observed had QIS–4 been conducted during a more stressful economic period, since the Basel 2 framework is designed to be more sensitive to expected macroeconomic conditions than current U.S. rules” (emphasis added). The Basel Committee paper broadly agreed, stating “some supervisors noted this [better economic conditions for QIS–4 and QIS–5] likely affected estimates of probabilities of default (PD) through strong credit quality conditions”. By lowering capital charges at the height of the bubble and raising them in the subsequent bust, internal risk models amplified both halves of the 2000s financial cycle.
Implementation of Basel 2 in Europe and Not in the United States
There was never any doubt about the implementation of Basel 2 rules in the European Union. Indeed, one of the reasons that European negotiators wanted to make the final Basel Committee rules so specific was to minimize the need for additional work on implementation via an EU Capital Adequacy Directive. The directive was duly passed in October 2005 without any significant changes compared to the Basel proposals. The “flexibility within the framework” (in the words of the US regulators) provided room for large European banks to take an aggressive approach toward the calculation of capital buffers. The European banking industry was increasingly driven by a small number of bloated national mega-banks who were already using the flexibility provided by internal risk models for investment banking to lower capital buffers and thereby gain a competitive advantage over smaller rivals who were operating under standardized weights. The Basel 2 plan to extend the use of internal risks models from investment banking to the entire operations of a bank provided further incentives to manipulate internal models, particularly as the mega-banks were increasingly competing against each other.
Competition across national supervisors within the European Union undermined incentives to closely examine bank behavior. The lack of engagement in the difficult task of supervising large universal banks with internal risk models was exemplified by the adoption of “light touch” regulation in the United Kingdom in early 2001 (a similar approach was taken in Ireland). The principles-based approach to regulation (the formal title for “light touch”) adopted by the UK Financial Services Agency involved close collaboration between supervisors and regulated banks. In the words of the 2005 Better Regulation Action Plan, the agency committed to “continue to work with the industry where possible, so that they own solutions to market failure problems”.24 Or, in the words of Chancellor Gordon Brown’s accompanying press release “Not just a light touch but a limited touch”. Looking back after the crisis, the Former Chief Executive of the Agency, Hector Sants, commented: “The prevailing climate at the time and indeed right up till the crisis commenced was the market really does know best”.25 UK regulators accepted the US Federal Reserve’s view of the power of market discipline in limiting risks taken by banks.26
There were several reasons why the United Kingdom in particular was happy to publicly embrace light touch regulation. First, UK supervisors had less incentives to be tough on domestic banks as almost half of its banking system was foreign-owned branches and hence supervised by other countries. Second, UK supervisors had a much more active relationship with its banks than was typical on the continent, where the greater reliance on auditors in the supervision process already implied a more hands-off process. Finally, UK regulators were much closer to the Federal Reserve, which espoused the advantages of market discipline, than other European regulators due to similarities in culture, legal systems, and financial market structure.
More generally, the 2000s saw a gradual erosion in capital cushions across the major European Union banks. The process had its own dynamic, as lax supervision in some countries (such as the United Kingdom and Ireland) put pressure on others European supervisors to lighten their own approach as their banks requested a level playing field on the enforcement of regulations. There was also the risk that if supervisors continued to take a tough approach then national banks would divert some of their business to subsidiaries in countries with lighter supervision such as Ireland, thereby reducing the control that national supervisors could exert over their national banks.
The outcome was a massive and unsustainable expansion of European mega-banks. As their domestic markets became saturated, the additional lending went into overseas ventures. The core Euro area banks lent increasing amount to the United States and the Euro area periphery. They also became the dominant players in various other types of loans, such as trade finance or funding projects in developing countries.
It is ironic that the Basel 2 framework that the Federal Reserve supported so strongly was in effect never implemented in the United States.27 The main reason for this was splintered financial regulation that required Basel 2 to be agreed across different banking regulators. While the Comptroller of the Currency was modestly positive about the Basel 2 Accord, the Federal Deposit Insurance Corporation (FDIC) viewed the drop in required bank capital as a result of adopting Basel 2 with concern. Lower capital implied a higher chance of banks insolvencies, and the FDIC was primarily responsible for avoiding such insolvencies and for organizing and paying for any rescues. As a result, the FDIC was not prepared to scrap the leverage ratio that had been incorporated into the US prompt corrective action framework. This was important as the leverage ratio used total assets to calculate capital needs, and hence was not affected by any reduction in risk weights coming from internal risk models. It set a floor below which the capital buffers of the US regulated banks could not fall.
A compromise between the pro-Basel 2 Fed and the more cautious FDIC was eventually reached in July 2007 after long and difficult negotiations. This instituted some relatively aggressive transition floors that limited the short-term erosion of capital under Basel 2. Crucially, it also retained the leverage ratio and prompt corrective action for US regulated banks (the compromise was effectively overtaken by the financial crisis and the decision to revamp bank regulation through Basel 2.5 and Basel 3, discussed in a later chapter).
The experience of the US investment banks vividly illustrates the problems associated with internal risk models. After the European Union decided that investment banking groups operating in Europe had to be subject to safety and soundness supervision, the five major US investment banking groups (called the Consolidated Supervision Entities or CSE) agreed to switch from the Net Capital Rule to the Basel system in 2004. Although this switch has often been portrayed as a weakening of supervision given the problems that the investment banks experienced over the crisis, in actual fact the Basel system was, if anything, more stringent since the Net Capital Rule was only focused on ensuring clients were repaid rather than avoiding bankruptcy. It also had the advantage of unifying the capital rules faced by the US investment banks and the European universal banks. Despite this, however, the outcome of the Consolidated Supervision Entities program was not pretty. As the Congressional Research Service put it:
Whatever the intent of the CSE program, it did not succeed in preventing excessive risk-taking by the participants. By the end of September 2008 all five CSE investment banks had either failed (Lehman Brothers), merged to prevent failure (Merrill Lynch and Bear Stearns), or applied for bank holding company status (Morgan Stanley and Goldman Sachs) [in order to gain access to emergency Federal Reserve funds]. On September 26, 2008, SEC Chairman Cox announced the end of the CSE program, declaring that “[t]he last six months have made it abundantly clear that voluntary regulation does not work”.28
While voluntary regulation may have been part of the problem, this assessment is much too generous to the Basel Committee. The European experience indicates that the internal risk model approach did not provide adequate capital buffers to cover risks from investment banking. The entire Basel system was faulty, not simply its US implementation.
The core US banking system was protected from the excesses of internal risk models by the FDIC’s insistence that the simple leverage ratio be retained. This explains why the crisis had its worst effects on the US investment banks and the European universal banks, while the major US regulated banks were generally financially strong enough to help rescue the system, including by buying troubled investment banks. This is not to say that the Fed-led drive to financial deregulation had no impact on the United States. It did. Lax consumer protection allowed a flood of subprime mortgages without adequate regulatory oversight. The biggest impact of the Federal Reserve, however, came through their championing of internal risk models that allowed a massive expansion in core Euro area bank lending that was ferried back to the United States via the cross-Atlantic financial drift. The ability of the US regulatory process to reject the major reduction in bank capital requirements implied by Basel 2 underlines the advantage of a slow process in which multiple institutions with different viewpoints need to agree on changes. So how did the North Atlantic boom unfold? It is instructive to start with the Euro area banks.
* * *
Charting the Unsustainable North Atlantic Financial Boom
The Unwise Overseas Expansion of Core Euro Area Banks
The Euro area banking system expanded massively over the 2000s boom on the back of thinning capital buffers (Figure 20). Strikingly, the rate of increase tripled after 2004, the year that the passage of Basel 2 ensured that internal risk models would be extended to credit risk by the end of 2008.29 Nor was the process limited to the Euro area. Similar percentage increases in assets were experienced elsewhere in Europe, including the already large UK banking system.
In particular, the core Euro area banks expanded their investment banking operations (Figure 21). Noticeably, very little of the increase in the size of the banks in Germany, France, the Netherlands, and Belgium came from commercial loans (dotted segments).30 By contrast, in the Euro area periphery (Italy, Spain, Portugal, Ireland, Greece, Austria and Finland) the expansion in assets was much more balanced, as the dramatic reduction in interest rates in southern Europe as a result of Euro area membership provided opportunities for more loans to customers.31
Investment banking grew in the Euro area core, expansion was more balanced in the periphery.
Source: ECB Banking Statistics.Investment banking grew in the Euro area core, expansion was more balanced in the periphery.
Source: ECB Banking Statistics.Investment banking grew in the Euro area core, expansion was more balanced in the periphery.
Source: ECB Banking Statistics.Facing limited lending opportunities at home, the core banking system expanded rapidly into foreign ventures, most notably in the United States, United Kingdom and Euro area periphery (Figure 22). The expansion into US and UK investment banking (the black segment) was spearheaded by mega-banks such as Deutsche Bank and BNP Paribas. The expansion to the Euro area periphery (dark gray segment) came through a mix of subsidiaries, branches, loans to local banks, and purchases of government and private paper. By 2007, the assets of core Euro area banks in the US, UK, and periphery were almost equal to core Euro area output. Furthermore, this was a one-way expansion as the corresponding liabilities remained small (light gray segment). Similarly, the core Euro area banks dove into other regions, such as eastern Europe (where areas such as the Baltics also suffered banking crashes) and business lines (such as trade credit, which also suffered over the crisis).
This rapid overseas expansion provided significant parts of the funding for the US and periphery financial bubbles, and ensured that the subsequent busts fed back to the core Euro area banking systems despite limited housing excesses at home. In 2007, on the eve of the US subprime problems, the northern European banks (Euro area core plus UK) owned US assets worth over one-quarter of US output, three times the ratio of 2000. By this time, they also owned assets in the Euro area periphery of around half of the periphery output, double the 2000 value. The result was that when the bubble burst the Euro area suffered a generalized banking crisis.
The twelve mega-banks identified in Chapter 1 (see Figure 3) drove these shifts in the Euro area banking system.32 Indeed, they accounted for the entire increase in bank assets over the 2000s as they expanded from one-third to one-half of the assets of the overall system through a combination of internal expansion, and mergers and acquisitions (Figure 23). By 2008, Deutsche Bank had assets of over 20 percentage points of Euro area output and six other mega-banks had balance sheets greater than 10 percentage points, a threshold than no bank had surpassed in 2000. Their mergers with smaller banks in other EU countries also drove the limited integration of the European banking system. Examples included UniCredit’s 2005 takeover of the German group HBV in 2005, BNP Paribas’s acquisition of Banca Nazionale di Lavoro in 2006, and the massive purchase of the Dutch group ABN AMRO by Royal Bank of Scotland, Fortis, and Santander in 2007. These mega-banks were often not just too big to fail but too big to save. By 2007, many had assets that were over half of national output, while those of ING Group were over double the output of the Netherlands. The delays in acknowledging the true size of the banking problems after the North Atlantic crisis, which slowed the repair of the financial system and lengthened the recession, was more of a necessity than a choice.
Even as the assets of the mega-banks expanded, their capital buffers eroded (Figure 24). Capital buffers were lower and fell more steadily for the universal mega-banks in the Euro area core. By contrast, the larger capital buffers of the commercial mega-banks in the Euro area periphery were more volatile, but fell notably in 2004, when Basel 2 was agreed, and in 2008, when it was adopted.
The Basel measures of capital buffers used by regulators obscured this fall in bank soundness because of the manipulation of risk-weighted assets. Figure 25 plots the ratio of equity to total assets just discussed for individual mega-banks on the vertical axis and the Basel regulatory ratio of core capital to risk-weighted assets on the horizontal axis. Already by 2002 the relationship between these different ways of measuring capital buffers had broken down. In 2008, after the introduction of Basel 2, the relationship had turned negative across the mega-banks. In this topsy-turvy world, Dexia and Deutsche Bank exhibited exemplary Basel regulatory capital ratios despite having capital below 2 percent of total assets while Intesa Sanpaolo, which boasted the highest ratio of capital to assets (8 percent), registered the second lowest Basel regulatory buffer. Put another way, at “well-capitalized” Deutsche Bank a Euro of capital was backing about 60 Euros in assets while at the apparently less well capitalized Intesa the same Euro backed only 12 Euros in assets. In short, over the European banking boom of the 2000s the system became larger, more dominated by mega-banks, less well capitalized, and more exposed to the emerging bubbles in the United States and the Euro area periphery. So what was happening in the United States?
Basel capital ratios became increasingly misleading.
Source: Scientific Committee of the European Systemic Risk Board (2014).Basel capital ratios became increasingly misleading.
Source: Scientific Committee of the European Systemic Risk Board (2014).Basel capital ratios became increasingly misleading.
Source: Scientific Committee of the European Systemic Risk Board (2014).The Changing Face of US Securitization
Investment banking was also booming in the United States. The US banking system retained its dual structure, with regulated banks originating loans and then selling many of them to the less regulated shadow banking system centered around the independent investment banks. Indeed, even in the case of the two US banks that developed a universal banking model involving both commercial and investment banking (Citigroup and JP Morgan Chase), the US Flow of Funds data continued to separate the activities of their “deposit taking institutions” (the regulated banking side) from their “broker-dealers” (the core of the investment banking side)—a residual statistical tic from the Glass–Steagall law separating the two activities. Accordingly, the assets of regulated banks and of broker-dealers should be combined to get a sense of a US “universal” banking system equivalent to that in Europe. While regulated banking remained the dominant segment, it was becoming a smaller part of the system up to the crisis as broker-dealer assets doubled as a ratio of the economy between 2000 and 2008.
The rise in importance of the shadow banking system can be clearly seen in the surge in riskier private-label mortgage-backed securities after the Securities and Exchange Commission (SEC) widened rules on repo collateral in 2003 (Figure 26). With house prices rising faster than inflation, mortgage-backed securities increased rapidly as a proportion of the economy, while the nature of the issuers changed dramatically. Through 2002, the securitization market was dominated by Fannie Mae and Freddie Mac, the two major housing government-sponsored enterprises (GSEs), but after the 2003 SEC decision issuance was dominated by riskier private mortgage-backed asset companies. By the time the housing market soured in 2007, one-third of all mortgaged-backed securities had been issued through private firms that had no credit guarantee. These were the assets whose values collapsed over the crisis.
Private mortgage securitizations surged after repo collateral was widened in 2003.
Source: US Flow of Funds.Private mortgage securitizations surged after repo collateral was widened in 2003.
Source: US Flow of Funds.Private mortgage securitizations surged after repo collateral was widened in 2003.
Source: US Flow of Funds.Foreign banks use of the repo market to borrow dollars also surged after the SEC decision (Figure 27). In 2002, foreign banks were bit players in the repo market (black segment). By 2007, they comprised a quarter of a much larger market (by the eve of the crisis, the repo market was twice the size of that for money market mutual funds, compared to near equality in 2002). Consequently, when the repo market froze over the crisis, Euro area banks had to scramble to replace the loss in dollar wholesale cash that they needed for their market operations.33
Euro area banks provided much of the funding for the US housing bubble. Strikingly, between 2002 and 2007, the increase in US assets owned by northern European banks (the Euro area core plus the United Kingdom, which housed many Euro area investment banking operations) was almost exactly equal to the increase in US private-label mortgage securitizations. While the creation of so many subprime mortgages was a product of faulty consumer protection and US-specific changes in regulation, including for repo collateral, much of the financing came directly or indirectly from European banks using internal risk models (that the Fed had championed) to save on capital buffers. The road to the collapse of the US housing market went from the Fed through the Basel Committee and then core Euro area banks before flowing back to the United States via the cross-Atlantic financial drift.
These changes were aided and abetted by a major increase in the size of the US national regulated and investment banks. Figure 28 shows the assets of the major investment banks (Morgan Stanley, Goldman Sachs, Merrill Lynch, Lehman Brothers, and Bear Stearns) and the three regulated national banks (Citi, JP Morgan Chase, and Bank of America, of which the first two were becoming universal banks that included investment banking operations). At the peak in 2007 the assets of these institutions were almost as large as US output.
Despite rapid expansion, the smaller US investment banks were coming under increasing competitive pressure.34 The five major independent US investment firms remained the most important presence in the sector but their clout was shrinking as they faced increased competition from subsidiaries of universal banks, including JP Morgan, Citi and the northern European mega-banks—Deutsche Bank, Barclays, BNP Paribas, and, from 2006, HSBC. While the leading independent US investment banks, Morgan Stanley and Goldman Sachs, continued to be much larger than these newcomers, the same was not true of Bear Stearns, Lehman Brothers, and Merrill Lynch, all of which floundered in the crisis.
In contrast to Europe, the three regulated national banks never became too big to save because support was centralized at the federal level. While the assets of Citigroup, the largest of the three national banks in 2007 and the one that experienced severe problems over the crisis, were similar in size to the Euro area mega-banks, they represented only about 15 percent of US output a much smaller ratio than in European countries. Consequently, the US government was able to provide adequate support to Citigroup over the crisis, which meant that the US banking system did not suffer the same obfuscation that prolonged banking problems in the Euro area.
The capital buffers of the shadow banks were much thinner. The investment banking groups as well as the two major GSEs (Fannie Mae and Freddie Mac)—the segments that included most of the failures over the crisis—had buffers that were only about half of those of the regulated banks. This dichotomy between well–capitalized regulated banks and thinly capitalized shadow banks in the United States was similar to that between the core universal banks and the periphery commercial banks in the Euro area. However, in the United States the thinly capitalized institutions were a smaller part of the overall system.
Examining the capital buffers of individual institutions in 2007 illuminates the vulnerability of the shadow banking system and Citigroup (Figure 29). Across the twenty-six largest US-owned financial institutions, three of the four lowest capital ratios were associated with investment banks that subsequently failed or had to be rescued, Bear Stearns, Merrill Lynch, and Lehman Brothers. In addition, Freddie Mac and Fannie Mae—the seventh and eighth largest financial institutions in the country that had to be put into receivership after their holdings of private mortgage-backed securities went sour—also had thin capital cushions. Finally, within the regulated banking sector, Citigroup had the lowest capital buffer.
Was the Behavior of the US and European Banking Systems Typical?
An important question is the degree to which the expansion in the Euro area and US banking systems were unique, or part of a more general global trend. This international perspective is useful for explaining the origins of the North Atlantic financial boom and its consequences, as larger banking systems are costlier and more difficult to resolve. Comparing the size of the US and other banking systems is complicated by differences in accounting practices. Assets of financial firms under the US accounting standard (GAAP) do not include holdings of derivative assets such as options and futures contracts that are included in the international accounting standard (IFRB) used by the rest of the world. A second complication is that the broker-dealers whose assets are tracked in the US Flow of Funds data are only one part of the investment banks. Examining pre-crisis balance sheets of investment banks suggests that broker-dealer assets should be doubled to get a sense of the size of the overall US investment banking system on an equivalent basis to the international accounting system (the difference for regulated banks is trivial).
Making this adjustment generates a US banking sector that was larger, expanding faster, and had a larger role for shadow banking than is generally recognized. The US “universal” banking system peaked at 160 percent of output in 2009 with the investment banks making up some 40 percent of the system. The magnified role of investment banks makes it easier to understand how the shadow banking system was able to absorb so many securitized assets and why problems in the US investment banks generated a global financial crisis. Even so, the US banking system remained only about half the size of the European system relative to output. The massive size of the European banks, as well as their presence in the US investment banking sector, explains why the European banking crisis was so generalized and so difficult to fix.
Most other major banking systems did not expand as fast as those in Europe or the United States, suggesting that the Basel rules alone were not the main driver of the North Atlantic debacle. Figure 30 compares bank assets as a ratio to output for European banks (Euro area plus the UK, which includes a lot of Euro area investment banking) and the (adjusted) US banking system since 1999 with the banking systems of Australia, Canada, and Japan, all of which were also members of the Basel Committee. The assets of the Canadian and Japanese banking systems, which were dominated by commercial banks, remained relatively stable as a ratio to output just like the regulated part of the US banks, albeit at vastly different sizes. The Canadian system was the smallest throughout the period while the Japanese banking system started as the largest. The Australian banking system, which unlike the other two countries had a universal bank model, did expand over the boom of the 2000s but remained much smaller than the European system. The Australian banks also suffered a wholesale funding run over the crisis, but this was quickly solved through government guarantees. This suggests that the major drivers of the expansion of the European banks were the universal banking model and overexpansion from lax oversight due to regulatory competition. But what created the demand for loans that the rapidly expanding North Atlantic banking system produced?
The size and increase of the European banks stands out.
Source: National Data Sources from Haver Analytics.The size and increase of the European banks stands out.
Source: National Data Sources from Haver Analytics.The size and increase of the European banks stands out.
Source: National Data Sources from Haver Analytics.* * *
Charting the North Atlantic Macroeconomic Boom and Bust
The booms in the Euro area periphery and the United States that started in the late 1990s were propelled by a potent cycle of rising lending, spending, output, and asset prices. In addition to the “yin” of lax financial regulation, the North Atlantic crisis reflected the “yang” of macroeconomic booms. The US boom that started in the late 1990s exhibited a larger and more persistent imbalance between spending and output than had been seen previously. Figure 31 measures US aggregate spending as a percentage of output. From 1970 through to the mid-1990s there is a clear cycle with spending rising above output in expansions and then falling back over recessions. From the late 1990s on, however, spending rose consistently compared to output. By the eve of the crisis, Americans were spending a record dollar and 6 cents for every dollar that they produced. The flip side to this spending boom was a growing trade gap. As Americans spent more they sucked in more imports and borrowed increasing amounts of money from abroad to pay for them.
The same pattern holds for the Euro area periphery. Figure 32 reports the same ratio of spending to output for Italy, the only country in the group for which data from 1970 is readily available, as well as a shorter series for the periphery as a whole. Through to the mid-1990s, there is again a clear cycle as spending rises and falls compared to output, followed by an extended boom in spending from the late 1990s until the crisis. Indeed, the increase in spending over the pre-crisis boom is similar to that in the United States, albeit from a somewhat lower base. Again, the flip side was a growing trade deficit.35
The Euro area periphery went on a similar spending spree to the United States.
Source: Italian and EU National Accounts.The Euro area periphery went on a similar spending spree to the United States.
Source: Italian and EU National Accounts.The Euro area periphery went on a similar spending spree to the United States.
Source: Italian and EU National Accounts.The macroeconomic “yang” of lower interest rates explains why the Euro area financial boom occurred in the periphery even though much of the additional lending capacity was concentrated in the core. The spending boom in the periphery was sparked by the rapid fall in long-term interest rates as entry into European Monetary Union reduced risk premiums (Figure 33). Yields on ten-year bonds in Italy, Spain, and Portugal were in double digits as late as 1995. By the time the Euro was introduced in 1999, they were below 4 percent and essentially identical to German rates (a pattern that was repeated when Greece joined the Euro area in 2003). While some of this compression in nominal yields reflected the reduction in inflation as a result of joining the Euro area, “real” rates adjusted by inflation also fell markedly. This reduction in real interest rates also partly reflected the pressure on core Euro area banks to find places to lend their rapidly expanding loan portfolio.
Low funding costs boosted by overconfidence in the benefits from European Monetary Union and the need to find new places to make loans led to a boom in investment spending in the Euro area periphery (Figure 34). The left chart below traces how the rise in spending as a proportion of output discussed earlier was apportioned between investment and consumption (with each series measuring changes compared to the situation in early 1995). The entire pre-crisis spending boom in the Euro area periphery came from higher investment, which then collapsed well below its 1995 level after the crisis. Figure 35 shows the same decomposition for the components of investment—machinery, residential housing, and commercial buildings. Over time, the boom in spending became increasingly dominated by residential housing, financed by mortgages to households and loans to property developers.
Investment spending surged in the Euro area periphery before the crisis.
Source: EU National Accounts.Investment spending surged in the Euro area periphery before the crisis.
Source: EU National Accounts.Investment spending surged in the Euro area periphery before the crisis.
Source: EU National Accounts.
Residential spending drove higher investment.
Residential spending drove higher investment.
Residential spending drove higher investment.
A similar pattern was seen in the United States (Figures 36 and 37). While consumption played a more important role in the early stages of the US boom than it did in the Euro area periphery, plausibly reflecting the wider use of credit scores that allowed US banks to automate consumer loans, the latter stages of the US boom were increasingly driven by rapid increases in spending on residential housing.
Higher spending on residential housing was both a cause and a consequence of booming housing prices (Figure 38). Here, again, there are striking similarities between the Euro area periphery and the United States. In both regions house prices rose by about two-thirds compared to the overall price level between the late 1990s and the peak of the boom. The subsequent fall in house prices, however, was much faster in the United States. As will be discussed further below, this largely reflected the fact that the US housing boom was financed by traded assets whose prices fell rapidly while in the Euro area periphery the real estate loans were generally kept on bank balance sheets which made it easier to delay recognition of losses.
US and Euro area periphery experienced similar house price booms.
Source: Europe: National data on All Dwellings from Haver Analytics. US: FHFA Purchase Only Index.US and Euro area periphery experienced similar house price booms.
Source: Europe: National data on All Dwellings from Haver Analytics. US: FHFA Purchase Only Index.US and Euro area periphery experienced similar house price booms.
Source: Europe: National data on All Dwellings from Haver Analytics. US: FHFA Purchase Only Index.What Lowered US Bond Yields in the 2000s?
The US boom was accompanied by a gradual fall in long-term interest rates whose cause is less clear than the European Monetary Union-induced reduction in the Euro area periphery. Three main hypotheses have been put forward to explain the fall in US yields. The first is that it reflected the inflows into US Treasury and Agency assets coming from the build-up in reserves in emerging markets, including China—often referred to as the “global saving glut”.36 The second is that financial inflows from European banks lowered rates.37 The final explanation is that the rise in repos led to a reduction in yields as mortgage-backed assets became more liquid.38 Each explanation highlights a different potential driver, self-insurance by emerging markets, financial deregulation in Europe, and the widening of repo collateral in the United States.
The evidence suggests that all three drivers mattered. Figure 39 provides a sense of the relative importance of these three explanations by comparing the amount of money that was pumped into US markets from each source. For the global saving glut, the chart shows the increase in reserve holdings as a proportion of US output.39 Holdings rose by some 8 percent of US output between 1999 and 2007 (unlike the other explanations, this flow accelerated after the crisis). For European bank inflows, the chart shows the increase in net ownership of US assets by core Euro area and UK banks, which start at a very similar level to reserves but rise twice as fast through 2006.40 Even accepting that capital gains may have played a part in the increase in European bank assets and that not all of these inflows went into the US housing market, it is clear that such inflows played an important role in easing US financial conditions, as it also did in the Euro area periphery. Finally, the expansion in US repos was similar to the increase in European bank assets, partly because the two were interlinked, since Euro area banks used repos to access dollars.
Falling US yields reflected inflows from emerging markets, European banks, and repos.
Source: Treasury International Capital System, BIS, and US Flow of Funds.Falling US yields reflected inflows from emerging markets, European banks, and repos.
Source: Treasury International Capital System, BIS, and US Flow of Funds.Falling US yields reflected inflows from emerging markets, European banks, and repos.
Source: Treasury International Capital System, BIS, and US Flow of Funds.The counterpart to these asset purchases was a rise in the US trade deficit. This was important as the Federal Reserve focused on stabilizing domestic inflation rather than domestic demand. As the rise in the trade deficit limited the amount to which the rise in domestic demand translated into higher output and inflation, the Federal Reserve kept monetary policy relatively expansionary.41 This maintained output close to capacity, but at the cost of supporting the housing bubble.
All in all, it seems reasonable to conclude that the fall in US bond yields reflected a combination of all three drivers—higher reserve asset holdings by emerging markets, European expansion into investment banking, and greater liquidity of safe assets. This suggests that in the United States a significant portion of the fall in yields reflected the internal dynamics of the US financial system, and hence a melding of macroeconomic and financial drivers. This is similar to the Euro area periphery where, in addition to the fall in bond yields linked to entry into monetary union, inflows from core Euro area banks also played a significant role.
Similar Macroeconomic Booms, Dissimilar Busts
The US and Euro area crises were driven by the same underlying dynamics. A steady loosening of financial regulation led to an expansion in bank assets on both sides of the Atlantic, complemented by higher spending coming as a result of lower borrowing costs. Easy borrowing, higher assets prices, and pro-cyclical financial regulation fed on each other to create spending booms that sucked in money, particularly from the core Euro area banks, that artificially supported spending and output.
If the components of the unsustainable asset bubbles were so similar, why then was the timing so different between the financial crises in the United States and Europe? After all, the US crisis started in 2007 with problems in subprime loans and reached a peak in September 2008 when Lehman Brothers went bankrupt. By contrast, the Euro area debt crisis is generally dated from the Greek government’s announcement of much larger government deficits and debt in late 2009, and peaked in mid–2012. The answer to this questions matters because the association of the European crisis with the problems of Greece and unsustainable debt elsewhere in the periphery diverted attention from the more important fact that the real problem in the Euro area was that the European banking system had become dangerously undercapitalized, particularly in the Euro area core.
The answer to why the Euro area crisis lagged that in the United States lies in the type of lending used to finance the bubbles. In the United States, most of the bad loans were sold to investors through mortgage-backed securities. The prices of these securities reacted immediately when problems in subprime mortgages first started to appear in mid-2007 and took a larger dive in the aftermath of the collapse of the investment bank Lehman Brothers, when a loss of confidence in counterparties led to a freeze in repo markets. This fall in market prices led to rapid write-downs of US mortgage-backed securities that affected US and core Euro area banks immediately. By contrast, in the Euro area periphery most of the bad loans were retained on bank balance sheets, which gave banks much more leeway as to when to acknowledge that the borrowers were unable to pay.
In addition to the delayed recognition of bad loans, the much more destructive impact of the crisis on Euro area than on the United States largely reflected the region’s deeper structural weaknesses. In the Euro area, the bloated universal banks went into the crisis with thin capital buffers, which meant that the financial shocks in the US and Euro area periphery whipsawed back and created a banking crises in the core of the banking system. In addition, the design of the Euro area made it difficult to provide the periphery with fiscal or banking support. Weak bank supervision and a deeply flawed monetary union were at the root of why a collapse in housing prices in the United States and European periphery led to a prolonged Euro area economic crisis.
While the crisis in the United States was dramatic, the regulated banks were significantly better capitalized than in Europe. Hence, while large parts of the shadow banking system had to be rescued—especially the smaller investment banks, government-sponsored enterprises, as well as Citigroup—most of the regulated banks were able to support this effort and absorb many of the failed institutions. In addition, US policymakers were not constrained by an incomplete monetary union. The Federal Reserve responded to the crisis by almost immediately funneling money into illiquid markets and institutions (including many Euro area ones). Similarly, the US federal government was able to provide banking support without creating serious doubts about the viability of its finances.
This is not to say that individual decisions taken during the crisis were unimportant. In the United States, the bankruptcy of Lehman Brothers—while possibly understandable given the constraints being faced—was clearly a policy blunder. Similarly, even given its legal mandate, the European Central Bank could have been more proactive as problems in the periphery mounted—after all, the Federal Reserve showed considerable ingenuity in its crisis response, including finding ways to channel funds to distressed investment banks. In the end, however, it was always much more likely that the bursting of housing and financial bubbles in the United States and the Euro area periphery would create bigger shockwaves in the Euro area for the simple reason that the Euro area started with a weaker banking system and a more limited set of policy responses. The next chapter explains why the Euro area was so badly designed.
