In early 2008, policy-makers in the United States needed to deal with the frightening after-effects of what had appeared to be a glorious housing boom. The most immediate problem was a wave of foreclosures, which a Senate report predicted could reach 2 million by the end of 2009. Lawmakers sought to relieve the resulting pain and to preserve the longstanding dream of raising the US homeownership rate. Amidst a sea of lawsuits and recriminations, they needed to figure out where the US system for financing home purchases had gone wrong and how it could be fixed. To do this, lawmakers needed to understand what had happened, particularly because housing had until then seemed like such a bright spot in the US economy. ${}^1$${}^1$^(1){ }^{1}
2008 年初，美國政策制定者需要應對看似輝煌的房市繁榮所帶來的可怕後果。最迫切的問題是一波止贖潮，參議院報告預測到 2009 年底止贖數量可能達到 200 萬戶。立法者試圖緩解由此產生的痛苦，並維護長久以來提升美國房屋擁有率的夢想。在一片訴訟與指責聲中，他們必須找出美國房屋購買融資體系出錯的原因，以及如何修正。為此，立法者需要了解發生了什麼，尤其是因為房市直到那時仍被視為美國經濟中的一個亮點。 ${}^1$${}^1$^(1){ }^{1}

In his 2001 letter to shareholders, Fannie Mae CEO Franklin Raines wrote, “Housing is a safe, leveraged investment - the only leveraged investment available to most families - and it is one of the best returning investments to make… Homes will continue to appreciate in value. Home values are expected to rise even faster in this decade than in the 1990’s.” His optimism was due in part to the importance Americans attributed to owning a home. This importance was reflected in Fannie Mae’s motto, which was “Our Business is the American Dream.”
在 2001 年致股東的信中，房利美執行長富蘭克林·雷恩斯寫道：「住房是一項安全的槓桿投資——大多數家庭唯一可用的槓桿投資——且是回報率最高的投資之一……房屋價值將持續增值。預計本十年房價上漲速度將比 1990 年代更快。」他的樂觀部分源於美國人對擁有住房的重要性。這種重要性反映在房利美的座右銘中，即「我們的事業是美國夢。」

Raines was not alone in touting the advantages of housing as an investment. While house prices in particular regions had suffered temporary declines at various points, average house prices across the United States had risen fairly steadily since at least 1975 (see Exhibit 1). This trend accelerated in 1996, and reached about 12 percent per annum in late 2005 and early 2006. Many observers felt that this rise in prices was due in part to the Federal Reserve’s policy of maintaining low interest rates after the 2001 recession. In the period from 1980 to 2001, the Federal Funds rate (an overnight interest rate that banks charged each other and which the Federal Reserve targeted) had generally tracked economic conditions (see Exhibit 2). After 2001 and until July 2004, however, the Fed kept interest rates low in spite of signs of growth in output and prices. Perhaps fearing a recession that did not materialize, the Federal Funds rate was set to only 1 percent from July 2003 to July 2004. After this, anxiety about inflation seemed to gain the upper hand and interest rates were increased steadily, with the Federal Funds rate reaching $5.25\mathrm{\%}$$5.25\mathrm{\%}$5.25%5.25 \% in September 2006.
雷恩斯並非唯一一個宣傳房地產作為投資優勢的人。雖然特定地區的房價在不同時期曾出現短暫下跌，但自 1975 年以來，美國全國的平均房價一直相當穩定地上升（見圖表 1）。這一趨勢在 1996 年加速，並在 2005 年底至 2006 年初達到約每年 12%的增長率。許多觀察家認為，房價上漲部分原因是聯邦儲備局在 2001 年經濟衰退後維持低利率政策所致。在 1980 年至 2001 年期間，聯邦基金利率（銀行間隔夜利率，聯邦儲備局的目標利率）通常與經濟狀況相符（見圖表 2）。然而，2001 年之後直到 2004 年 7 月，儘管產出和價格顯示增長跡象，聯邦儲備局仍保持低利率。或許是擔心未出現的經濟衰退，聯邦基金利率在 2003 年 7 月至 2004 年 7 月期間僅設定為 1%。此後，對通脹的擔憂似乎佔了上風，利率逐步上升，聯邦基金利率在 2006 年 9 月達到 $5.25\mathrm{\%}$$5.25\mathrm{\%}$5.25%5.25 \% 。

A debate over house prices started around 2004. Some economists, such as Dean Baker, the codirector of the Center for Economic and Policy Research claimed at the time that house prices were like a bubble ready to burst, and that the economy needed to brace itself for a loss of $\mathrm{\$}2$$\mathrm{\$}2$$2\$ 2 to $\mathrm{\$}3$$\mathrm{\$}3$$3\$ 3 trillion in housing wealth. ${}^2$${}^2$^(2){ }^{2} Others felt that, even though increases in housing prices had far outstripped increases in residential rents, this was reasonable in light of the low interest rates.
關於房價的辯論始於 2004 年左右。一些經濟學家，如經濟與政策研究中心聯合主任迪恩·貝克（Dean Baker）當時聲稱，房價就像一個即將破裂的泡沫，經濟需要為損失 $\mathrm{\$}2$$\mathrm{\$}2$$2\$ 2 至 $\mathrm{\$}3$$\mathrm{\$}3$$3\$ 3 兆的住房財富做好準備。 ${}^2$${}^2$^(2){ }^{2} 其他人則認為，儘管房價的增長遠遠超過了住宅租金的增長，但鑑於低利率，這是合理的。

Even in October 2005, when it was common to hear mentions of a housing bubble, developer Bob Toll disagreed and complained “Why can’t real estate just have a boom like every other industry? Why do we have to have a bubble and then a pop?” Meanwhile, several economists pointed out that house price increases were concentrated in particular areas such as San Francisco and New York, where zoning restrictions made it difficult to expand the housing stock. Professor Chris Mayer of Columbia University saw the attraction of these areas coupled with the inability to increase supply as allowing house prices in these areas to remain high “basically forever.” Noting that Tokyo real estate was still more expensive than real estate in Manhattan, he stated: “There’s no natural law that says U.S. housing prices have to stop here. None.” ${}^3$${}^3$^(3){ }^{3}
即使在 2005 年 10 月，當時普遍聽到有關房地產泡沫的討論，開發商 Bob Toll 仍持不同意見並抱怨說：「為什麼房地產不能像其他行業一樣繁榮？為什麼我們一定要有泡沫然後破裂？」同時，幾位經濟學家指出，房價上漲集中在特定地區，如舊金山和紐約，那裡的分區限制使得擴大住房供應變得困難。哥倫比亞大學的 Chris Mayer 教授認為，這些地區的吸引力加上供應無法增加，使得這些地區的房價「基本上永遠」保持高位。他指出，東京的房地產仍比曼哈頓的房地產更昂貴，並表示：「沒有自然法則規定美國房價必須止步於此。沒有。」

While house prices reached eye-popping levels in what Chris Mayer called “superstar cities,” construction was booming elsewhere. Cities like Phoenix, as well as many communities in Florida and around Los Angeles, saw such a torrid pace of construction that builders had difficulty even procuring the cement they needed. New houses in these areas were often snapped up by eager investors and newspapers relished reporting on individuals who managed to resell houses at a gain even before they took possession of them. According to LoanPerformance Inc., more than 12 percent of Phoenix-area mortgages were obtained by investors in 2004, as compared to just 5.8 percent nationwide in 2000. ${}^4$${}^4$^(4){ }^{4}
雖然房價在克里斯·邁耶所稱的「超級明星城市」達到了驚人的水平，但其他地方的建築業卻在蓬勃發展。像鳳凰城這樣的城市，以及佛羅里達和洛杉磯周邊的許多社區，建築速度如此迅猛，以至於建商甚至難以取得所需的水泥。這些地區的新房常常被熱切的投資者搶購，報紙也樂於報導那些甚至在取得房屋所有權之前就成功轉售獲利的個人。根據 LoanPerformance Inc.的數據，2004 年鳳凰城地區超過 12%的房貸是由投資者取得，而 2000 年全國僅有 5.8%。

In the United States, it was common to talk about the “traditional” fixed 30-year mortgage. This instrument required the borrower to make a constant stream of monthly payments during the 30-year term of the loan. These payments were specified in advance; so the interest rate on this loan was fixed. Many of these traditional loans allowed borrowers to “pre-pay” their mortgages without penalty. When interest rates declined, borrowers often took advantage of this feature and refinanced their homes at lower rates.
在美國，人們常談論「傳統」的固定 30 年期抵押貸款。這種工具要求借款人在貸款的 30 年期限內持續支付固定的每月款項。這些付款事先規定，因此該貸款的利率是固定的。許多這類傳統貸款允許借款人提前還款而不受罰款。當利率下降時，借款人經常利用這一特點，以較低的利率重新貸款購房。

Savings and Loan Associations (S&Ls) already offered mortgages with constant payments before the Great Depression, though they were typically less than 12 years long. At the time, other lenders mostly offered short-term mortgages that needed to be refinanced because they had “balloon” payments at the end. During the Great Depression, many households went into default in part because this refinancing became difficult. One government response was to create the Home Owners Loan Corporation (HOLC), which made simultaneous offers to borrowers and lenders. If they both agreed, lenders received HOLC obligations in exchange for their claims against households, although this exchange required banks to recognize a loss on their assets. Households, meanwhile, freed themselves of their previous obligation by accepting new “self-amortizing” mortgages with fixed payments whose terms were based on new assessments of their home’s worth. ${}^5$${}^5$^(5){ }^{5}
儲蓄與貸款協會（S&Ls）在大蕭條之前已經提供固定付款的抵押貸款，儘管這些貸款通常少於 12 年。當時，其他貸款機構大多提供需要再融資的短期抵押貸款，因為這些貸款在期末有「氣球」付款。在大蕭條期間，許多家庭部分因為再融資變得困難而違約。政府的一項回應是成立房主貸款公司（HOLC），該公司同時向借款人和貸款人提出要約。如果雙方同意，貸款人會收到 HOLC 債券以交換其對家庭的債權，儘管這種交換要求銀行承認其資產的損失。與此同時，家庭通過接受新的「自動攤還」抵押貸款，解除先前的債務，這些新貸款具有固定付款，且條款基於對其房屋價值的新評估。

After WWII, banks and S&Ls originated many fixed 30 -year mortgages and held them to maturity. The results were not always happy. When short-term interest rates rose in the early 1980’s, the yield on mortgage assets fell below the cost of paying depositors for their funds. This mismatch was one of the causes for the failure of about half of the 32,234 S&L’s that existed in 1986. Because the government insured the S&L’s depositors, it incurred considerable losses and had to set up a special institution to dispose of the failed S&L’s assets. ${}^6$${}^6$^(6){ }^{6}
二戰後，銀行和儲蓄貸款協會（S&L）發放了許多固定 30 年期的抵押貸款並持有至到期。結果並不總是令人滿意。當 1980 年代初期短期利率上升時，抵押貸款資產的收益率低於支付存款人資金成本。這種不匹配是 1986 年存在的 32,234 家 S&L 中約一半倒閉的原因之一。由於政府為 S&L 的存款人提供保險，因此承擔了相當大的損失，並不得不設立一個專門機構來處理倒閉的 S&L 資產。

The S&L crisis also boosted the securitization of mortgages by two government-sponsored enterprises (GSEs), Fannie Mae and Freddie Mac. Fannie Mae was originally created in 1938 as a government agency. Like Freddie Mac, a twin that Congress chartered in 1970, Fannie Mae eventually became a privately owned publicly traded company. ${}^1$${}^1$^(1){ }^{1} Starting with bundles of mortgages purchased from mortgage originators, the GSEs created and sold mortgage-backed securities (MBSs), which delivered to holders the payments made on these mortgages. In exchange for a fee, the GSEs guaranteed the interest and principal on these loans. This meant that, assuming the GSEs remained solvent (or that the government came to their rescue if they found themselves in financial trouble), the only payment risk faced by the holders of these MBSs was the risk that the underlying mortgages would be repaid before they were due (known as prepayment risk).
儲貸危機也促進了兩家政府支持企業（GSEs）房利美（Fannie Mae）和房地美（Freddie Mac）對抵押貸款的證券化。房利美最初於 1938 年作為政府機構成立。與 1970 年國會授權成立的雙胞胎房地美一樣，房利美最終成為一家私有的上市公司。 ${}^1$${}^1$^(1){ }^{1} 這兩家 GSEs 從抵押貸款發起人處購買抵押貸款組合，創造並銷售抵押貸款支持證券（MBSs），將這些抵押貸款的付款交付給持有人。作為交換，GSEs 收取費用並保證這些貸款的利息和本金。這意味著，假設 GSEs 保持償付能力（或政府在其遇到財務困難時出手相救），這些 MBS 持有者面臨的唯一付款風險是基礎抵押貸款提前償還的風險（稱為提前還款風險）。

Congress capped the size of the loans that GSEs could accept. In 2006, for example, the maximum loan for single-family homes was $\mathrm{\$}417,000$$\mathrm{\$}417,000$$417,000\$ 417,000. To limit their credit risk, the GSEs used standards that were similar to those of traditional originators. To secure sufficient collateral, they took only senior mortgages and generally required the loan-to-value ratio (LTV) to be below $80\mathrm{\%}$$80\mathrm{\%}$80%80 \%. The LTV was computed as the ratio of the mortgage to the property’s market value at the time of origination. Before underwriting loans, the GSEs also looked at the borrower’s income and employment status, level of other assets, and history of foreclosures and bankruptcies. Consistent with the rules of GSEs, home lenders before the 1990 s only lent to borrowers they deemed credit worthy, and generally required documentary evidence on these variables. Until the practice was penalized by a 1977 law, most lenders also denied mortgages to people living in certain “redlined” communities, where these were predominantly inner city neighborhoods with large black populations.
國會限制政府支持企業（GSEs）可接受的貸款規模。例如，2006 年，單戶住宅的最高貸款額為 $\mathrm{\$}417,000$$\mathrm{\$}417,000$$417,000\$ 417,000 。為了限制信用風險，GSEs 採用了與傳統貸款機構類似的標準。為了確保足夠的抵押品，他們只接受優先抵押貸款，且通常要求貸款價值比率（LTV）低於 $80\mathrm{\%}$$80\mathrm{\%}$80%80 \% 。LTV 是指貸款金額與貸款時房產市場價值的比率。在承保貸款之前，GSEs 還會審查借款人的收入和就業狀況、其他資產水平以及止贖和破產記錄。根據 GSEs 的規定，1990 年代之前的房貸機構只向他們認為信用良好的借款人放貸，並通常要求提供這些變數的文件證明。在 1977 年法律禁止此做法之前，大多數貸款機構也會拒絕向居住在某些「紅線區」的居民提供抵押貸款，這些區域多為內城黑人居民較多的社區。

An avenue that remained open to borrowers with problematic credit histories was to apply through conventional lenders for loans insured by the Federal Housing Administration (FHA). The lenders then had to verify that the loan met FHA requirements and the process for doing so was somewhat more time-consuming than in the case non-FHA mortgages. In spite of these standards, about $8\mathrm{\%}$$8\mathrm{\%}$8%8 \% of FHA loans were past-due in 1993, while the delinquency rate on standard mortgages was only $3\mathrm{\%}$$3\mathrm{\%}$3%3 \%. FHA loans were packaged into mortgage-backed securities by Ginnie Mae, a governmentowned corporation that dealt exclusively with federally guaranteed mortgages.?
對於信用記錄有問題的借款人來說，一條仍然開放的途徑是通過傳統貸款機構申請由聯邦住房管理局（FHA）保險的貸款。貸款機構必須核實貸款是否符合 FHA 的要求，這一過程比非 FHA 抵押貸款的審核稍為耗時。儘管有這些標準，1993 年約有 $8\mathrm{\%}$$8\mathrm{\%}$8%8 \% 的 FHA 貸款逾期，而標準抵押貸款的違約率僅為 $3\mathrm{\%}$$3\mathrm{\%}$3%3 \% 。FHA 貸款由政府擁有的公司 Ginnie Mae 打包成抵押貸款支持證券，該公司專門處理聯邦擔保的抵押貸款。

In the 1990s new firms started to lend money to borrowers that did not qualify for “prime” mortgages. Rather than lending directly, many of these firms sought the help of mortgage brokers to whom they paid commissions. The US Department of Housing and Urban Development’s list of lenders who specialized in such “subprime” loans increased from 63 lenders in 1993 to 209 in 2005. Wall Street firms Lehman Brothers, Bear Stearns, Goldman Sachs, Merrill Lynch and Morgan Stanley all acquired such lenders, though all but Lehman Brothers and Bear Stearns did so only in 2006.
在 1990 年代，新公司開始向不符合「優質」抵押貸款資格的借款人放貸。這些公司並非直接放貸，而是尋求抵押貸款經紀人的幫助，並向他們支付佣金。美國住房與城市發展部專門從事此類「次級」貸款的貸款機構名單，從 1993 年的 63 家增加到 2005 年的 209 家。華爾街公司雷曼兄弟、貝爾斯登、高盛、美林和摩根士丹利都收購了這些貸款機構，儘管除了雷曼兄弟和貝爾斯登外，其他公司都是在 2006 年才這樣做的。

One obvious difference between “subprime” and “prime” loans was that the former had higher interest rates and fees. There was, however, no precise dividing line between the two, so that there was no consensus on how to measure the fraction of subprime loans. According to one definition, the value of these loans grew from about $1\mathrm{\%}$$1\mathrm{\%}$1%1 \% of new mortgages in 1993 to $20\mathrm{\%}$$20\mathrm{\%}$20%20 \% in 2006 . At the same time, the FHA share dropped from $11\mathrm{\%}$$11\mathrm{\%}$11%11 \% to $1.9\mathrm{\%}$$1.9\mathrm{\%}$1.9%1.9 \%. An independent analysis by the Wall Street Journal concluded that $29\mathrm{\%}$$29\mathrm{\%}$29%29 \% of the home loans made in 2006 had high interest rates. A large fraction of these loans refinanced existing loans. In many cases, these refinancing loans increased the borrowers’ mortgage debt and thereby made it possible for households to keep some cash for other purposes. ${}^8$${}^8$^(8){ }^{8}
「次級」貸款與「優質」貸款之間一個明顯的差異是前者的利率和費用較高。然而，兩者之間並無明確的分界線，因此對於如何衡量次級貸款的比例並無共識。根據一種定義，這些貸款的價值從 1993 年新抵押貸款的約 $1\mathrm{\%}$$1\mathrm{\%}$1%1 \% 增長到 2006 年的 $20\mathrm{\%}$$20\mathrm{\%}$20%20 \% 。同時，FHA 的比例從 $11\mathrm{\%}$$11\mathrm{\%}$11%11 \% 下降到 $1.9\mathrm{\%}$$1.9\mathrm{\%}$1.9%1.9 \% 。《華爾街日報》的一項獨立分析得出結論，2006 年發放的房屋貸款中有 $29\mathrm{\%}$$29\mathrm{\%}$29%29 \% 的利率較高。這些貸款中有很大一部分是用來再融資現有貸款。在許多情況下，這些再融資貸款增加了借款人的抵押貸款債務，從而使家庭能夠保留一些現金用於其他用途。 ${}^8$${}^8$^(8){ }^{8}

From being virtually unknown in the 1980s, Countrywide Financial became the largest mortgage lender in 2005. A 2003 government report showed that it was also the leading mortgage lender to minority homeowners, as well as one of the largest providers of home loans in low-income communities. When this report was released, Countrywide’s CEO Angelo Mozilo said: “We’re extremely proud of our accomplishments, as they clearly demonstrate our long-standing commitment to provide all Americans with the opportunity to achieve the dream of homeownership. These results underscore our ongoing efforts to discover new approaches to turn individuals and families into homeowners, to develop new loan products that reduce or eliminate the obstacles to homeownership and to make it easier for families to qualify for loans.” ${}^{\prime 9}$${}^{\prime 9}$^('9){ }^{\prime 9}
從 1980 年代幾乎默默無聞，到 2005 年成為最大的房貸貸款機構，Countrywide Financial 的成長令人矚目。2003 年的一份政府報告顯示，它同時是少數族裔房主的主要房貸貸款機構，也是低收入社區住房貸款的最大提供者之一。當該報告發布時，Countrywide 的執行長 Angelo Mozilo 表示：「我們對自己的成就感到非常自豪，因為這清楚展現了我們長期以來致力於為所有美國人提供實現擁有住房夢想的機會。這些成果強調了我們持續努力尋找新方法，將個人和家庭轉變為房主，開發減少或消除購屋障礙的新貸款產品，並使家庭更容易符合貸款資格。」

Contrary to what had been standard practice in the past, lenders such as Countrywide did not offer the same interest rate to all borrowers. This customization was facilitated by the use of automated statistical models that predicted the likelihood of default on the basis of borrower characteristics. Interestingly, the first statistical tools that came into wide use were those developed by Freddie Mac (called Loan Prospector) and Fannie Mae (called Desktop Underwriter). These were introduced to make it easy for mortgage originators to know whether their loans would be acceptable to the GSEs, though their use expanded well beyond this purpose. One variable that played a key role in these models, and which had apparently been absent from previous methods of qualifying borrowers for mortgages, was the borrower’s credit score. While there were several approved commercial credit score formulas (regulators did not allow scores to depend on race, gender, marital status or national origin), the most popular one was the FICO score invented by the Fair Isaac Corporation. This score, which ranged from about 300 for poor credit risks to about 850, appeared to give considerable weight to the punctuality with which borrowers had paid their previous obligations. One reason these scores became important in mortgage applications was that studies by Freddie Mac had shown a strong correlation between FICO scores and defaults on mortgages in the pre-1995 period. ${}^{10}$${}^{10}$^(10){ }^{10}
與過去的標準做法相反，像 Countrywide 這樣的貸款機構並未對所有借款人提供相同的利率。這種客製化是透過使用自動化統計模型來實現的，該模型根據借款人的特徵預測違約的可能性。有趣的是，最早被廣泛使用的統計工具是由 Freddie Mac（稱為 Loan Prospector）和 Fannie Mae（稱為 Desktop Underwriter）開發的。這些工具的引入是為了讓房貸發起人能輕鬆判斷其貸款是否會被政府支持企業（GSEs）接受，儘管其使用範圍遠超過此目的。在這些模型中扮演關鍵角色的一個變數是借款人的信用分數，而這在之前用於評估借款人資格的方法中顯然是缺失的。雖然有多種經批准的商業信用分數公式（監管機構不允許分數依據種族、性別、婚姻狀況或國籍），但最受歡迎的是由 Fair Isaac Corporation 發明的 FICO 分數。 這個分數範圍大約從 300 分（信用風險較差）到 850 分，似乎相當重視借款人過去按時償還債務的情況。這些分數在房貸申請中變得重要的原因之一，是 Freddie Mac 的研究顯示 FICO 分數與 1995 年前房貸違約之間存在強烈的相關性。

One type of mortgage that became popular among subprime lenders was known as a $2/28$$2/28$2//282 / 28 because its rate was fixed for 2 years and then became variable for the remaining 28 years. This mortgage was quite different from adjustable rate mortgages (ARMs) offered to prime borrowers. The introductory rate on $2/28$$2/28$2//282 / 28 s was above the typical rate offered on 30 -year fixed mortgages, whereas ARMs for prime borrowers had initial rates below those on fixed mortgages. Also, rates on $2/28$$2/28$2//282 / 28 s rose considerably when they were “reset” after 2 years. According to the President of the Federal Reserve Bank of Boston Eric Rosengren, the average initial rate for subprime mortgages issued in 2006 was $8.5\mathrm{\%}$$8.5\mathrm{\%}$8.5%8.5 \% (when the conventional 30 -year mortgage rate was below $6.4\mathrm{\%}$$6.4\mathrm{\%}$6.4%6.4 \% ) and reset to 610 basis points above the 6 -month LIBOR rate (which averaged about $5\mathrm{\%}$$5\mathrm{\%}$5%5 \% in 2006) after 2 years. ${}^{11}$${}^{11}$^(11){ }^{11}
一種在次級貸款人中流行的抵押貸款被稱為 $2/28$$2/28$2//282 / 28 ，因為其利率在前兩年是固定的，之後的 28 年則變為浮動利率。這種抵押貸款與提供給優質借款人的可調利率抵押貸款（ARMs）有很大不同。 $2/28$$2/28$2//282 / 28 的初始利率高於典型的 30 年固定抵押貸款利率，而優質借款人的 ARMs 初始利率則低於固定抵押貸款利率。此外， $2/28$$2/28$2//282 / 28 的利率在兩年後“重設”時大幅上升。根據波士頓聯邦儲備銀行行長 Eric Rosengren 的說法，2006 年發放的次級抵押貸款的平均初始利率為 $8.5\mathrm{\%}$$8.5\mathrm{\%}$8.5%8.5 \% （當時傳統 30 年抵押貸款利率低於 $6.4\mathrm{\%}$$6.4\mathrm{\%}$6.4%6.4 \% ），兩年後重設為比 6 個月 LIBOR 利率高出 610 個基點（2006 年 6 個月 LIBOR 平均約為 $5\mathrm{\%}$$5\mathrm{\%}$5%5 \% ）。 ${}^{11}$${}^{11}$^(11){ }^{11}

In the case of reasonable $2/28$$2/28$2//282 / 28 mortgages, there were pre-payment penalties if the mortgage was pre-paid in the first two years but there was no cost associated with pre-paying right before the interest rate was reset. From the point of view of mortgage brokers, this arrangement was attractive because it ensured that many borrowers would refinance after two years, allowing brokers to collect new origination fees. Borrowers were also told that this arrangement was good for them because, if they made timely payments, their FICO score would improve and they would be able to refinance at a lower rate.
在合理 $2/28$$2/28$2//282 / 28 抵押貸款的情況下，如果在前兩年內提前還款，會有提前還款罰金，但在利率重設前立即提前還款則不會產生任何費用。從抵押貸款經紀人的角度來看，這種安排具有吸引力，因為它確保許多借款人在兩年後會進行再融資，從而使經紀人能夠收取新的貸款發起費用。借款人也被告知這種安排對他們有利，因為如果他們按時還款，他們的 FICO 信用分數會提高，並且能夠以較低的利率進行再融資。

There were widespread allegations that some borrowers in this period received home loans on terms that were substantially less favorable than those of conventional or FHA loans for which these borrowers would have qualified. It was also claimed that unsophisticated borrowers had been duped into signing mortgages that continued to have severe pre-payment penalties even after interest rates had been reset to high levels. A lawsuit in Michigan claimed that a mortgage broker working for a unit of Lehman Brothers “confused and pressured” an elderly couple so that they would sign a loan
當時有廣泛指控稱，一些借款人在此期間獲得的房貸條件遠不如他們本可符合資格的傳統貸款或聯邦住房管理局（FHA）貸款條件。還有人聲稱，不懂行的借款人被誘騙簽署了即使在利率重設至高位後仍持續存在嚴重提前還款罰金的抵押貸款。密歇根州的一起訴訟指控，雷曼兄弟旗下單位的一名抵押貸款經紀人“混淆並施壓”一對年長夫婦，使他們簽署了一筆貸款。
whose interest rate would reach $17.5\mathrm{\%}$$17.5\mathrm{\%}$17.5%17.5 \%. Several borrowers told Federal officials that they had simply been lied to regarding their future monthly payments. ${}^{12}$${}^{12}$^(12){ }^{12}
利率將達到 $17.5\mathrm{\%}$$17.5\mathrm{\%}$17.5%17.5 \% 。幾位借款人告訴聯邦官員，他們對未來每月還款金額被明顯欺騙了。 ${}^{12}$${}^{12}$^(12){ }^{12}

What is certain is that some borrowers agreed to make payments that were impossible for them to keep up with over time. A 79-year old retired engineer named Robert Pyle, for example, moved from a $\mathrm{\$}265,000$$\mathrm{\$}265,000$$265,000\$ 265,000 to a $\mathrm{\$}352,000$$\mathrm{\$}352,000$$352,000\$ 352,000 mortgage in 2005 and cleared his credit card debts in the process. Almost immediately after signing the mortgage, which involved over $\mathrm{\$}33,000$$\mathrm{\$}33,000$$33,000\$ 33,000 in fees, he found himself unable to cover the $\mathrm{\$}2200$$\mathrm{\$}2200$$2200\$ 2200 monthly payment. Terry Dyer, the broker who issued Robert Pyle’s mortgage said, "It’s clear he was living beyond his means, and he might not be able to afford this loan. But legally, we don’t have a responsibility to tell him this probably isn’t going to work out. It’s not our obligation to tell them how they should live their lives. ${}^{113}$${}^{113}$^(113){ }^{113}
可以確定的是，有些借款人同意支付他們隨時間無法負擔的款項。例如，一位 79 歲的退休工程師羅伯特·派爾（Robert Pyle）在 2005 年從 $\mathrm{\$}265,000$$\mathrm{\$}265,000$$265,000\$ 265,000 轉為 $\mathrm{\$}352,000$$\mathrm{\$}352,000$$352,000\$ 352,000 抵押貸款，並在此過程中清償了信用卡債務。簽署這筆涉及超過 $\mathrm{\$}33,000$$\mathrm{\$}33,000$$33,000\$ 33,000 費用的抵押貸款後，他幾乎立刻發現自己無法支付 $\mathrm{\$}2200$$\mathrm{\$}2200$$2200\$ 2200 的每月還款。發放羅伯特·派爾抵押貸款的經紀人特里·戴爾（Terry Dyer）說：「很明顯，他的生活超出自己的能力範圍，可能無法負擔這筆貸款。但從法律上講，我們沒有責任告訴他這可能行不通。我們沒有義務告訴他們應該如何生活。 ${}^{113}$${}^{113}$^(113){ }^{113} 」

Some subprime loans required less documentation than was traditionally demanded. Instead of requiring proof of income or independent appraisals of the value of the home, some subprime mortgages were based only on “stated income” or “stated value.” Stated income loans were very convenient for borrowers who had casual jobs that were difficult to document, though they opened the door to fraud by both borrowers and brokers. Another dimension in which some subprime loans departed from traditional ones was in their down-payment requirements. To avoid the need to put any money down, brokers could offer a second “piggyback” mortgage that covered the down payment on the first.
一些次級貸款所需的文件比傳統要求的少。部分次級抵押貸款並不要求收入證明或獨立的房屋價值評估，而僅依據「申報收入」或「申報價值」。申報收入貸款對於那些工作不穩定且難以提供證明的借款人來說非常方便，但這也為借款人和經紀人帶來了欺詐的可能性。次級貸款與傳統貸款的另一個不同之處在於首付款的要求。為了避免需要支付任何首付款，經紀人可以提供第二筆「搭便車」抵押貸款來支付第一筆貸款的首付款。

Some loans originated in this period had a different benefit for borrowers, namely that they required only small payments at the beginning. So called “option ARMs” allowed borrowers to choose how much they wished to pay initially. For those borrowers who took advantage of this option and made low payments at first, later payments were increased so the mortgage would still be repaid in 30 years. Option ARMs were quite popular with certain lenders: 30% of the early 2005 mortgages at California-based IndyMac Bankcorp took this form as did 19% of Countrywide’s. Defenders of these loans argued that they went to credit-worthy customers and that they had low LTVs. Indeed, neither IndyMac nor Quicken Loans, which offered option ARMs as well as loans that had a zero “teaser” interest rate for the first 6 months, was on HUD’s list of subprime lenders. ${}^{14}$${}^{14}$^(14){ }^{14}
在此期間發放的一些貸款對借款人有不同的好處，即它們在開始時只需支付少量款項。所謂的「選擇型可調利率抵押貸款（option ARMs）」允許借款人選擇他們最初希望支付的金額。對於那些利用此選項並在初期支付較低款項的借款人，後續的付款會增加，以確保抵押貸款仍能在 30 年內償還完畢。選擇型可調利率抵押貸款在某些貸款機構中相當受歡迎：2005 年初加州 IndyMac Bankcorp 發放的抵押貸款中有 30%採用此形式，Countrywide 的則有 19%。這些貸款的支持者辯稱，這些貸款是給信用良好的客戶，且貸款價值比率（LTV）較低。事實上，IndyMac 和 Quicken Loans（後者同時提供選擇型可調利率貸款以及前 6 個月利率為零的「誘導利率」貸款）均未列入美國住房與城市發展部（HUD）的次級貸款機構名單。 ${}^{14}$${}^{14}$^(14){ }^{14}

While only $28\mathrm{\%}$$28\mathrm{\%}$28%28 \% of subprime loans were sold in the secondary market in 1995 , this fraction grew rapidly. In $2005,73\mathrm{\%}$$2005,73\mathrm{\%}$2005,73%2005,73 \% of mortgages classified as having a high interest rate were sold. About $\mathrm{\$}500$$\mathrm{\$}500$$500\$ 500 billion of newly minted mortgage-backed securities based on subprime loans were issued in 2005 as well as in 2006. By then, the GSEs faced “private label” competitors in the MBS market. MBS issues by Wall Street firms and mortgage originators like Countrywide and Washington Mutual accounted for $23\mathrm{\%}$$23\mathrm{\%}$23%23 \% of new issues in 2003. By 2005, they issued $\mathrm{\$}1.2$$\mathrm{\$}1.2$$1.2\$ 1.2 trillion of securities, which exceeded the amount issued by the GSEs. In the first half of 2006, the private label share climbed further to $57\mathrm{\%}$$57\mathrm{\%}$57%57 \%. Out of a total U.S. residential mortgage debt of $\mathrm{\$}11$$\mathrm{\$}11$$11\$ 11 trillion at the end of 2006 , the outstanding MBSs of the GSEs still equaled $\mathrm{\$}2.5$$\mathrm{\$}2.5$$2.5\$ 2.5 trillion, but the importance of the GSEs appeared to be waning. ${}^{15}$${}^{15}$^(15){ }^{15}
1995 年，只有 $28\mathrm{\%}$$28\mathrm{\%}$28%28 \% 的次級貸款在二級市場出售，但這一比例迅速增長。 $2005,73\mathrm{\%}$$2005,73\mathrm{\%}$2005,73%2005,73 \% 的高利率抵押貸款被出售。2005 年和 2006 年，基於次級貸款的新發行抵押貸款支持證券約為 $\mathrm{\$}500$$\mathrm{\$}500$$500\$ 500 億美元。到那時，政府支持企業（GSEs）在 MBS 市場面臨“私人標籤”競爭者。華爾街公司和抵押貸款發起者如 Countrywide 和 Washington Mutual 發行的 MBS 佔 2003 年新發行量的 $23\mathrm{\%}$$23\mathrm{\%}$23%23 \% 。到 2005 年，他們發行了 $\mathrm{\$}1.2$$\mathrm{\$}1.2$$1.2\$ 1.2 兆美元的證券，超過了 GSEs 的發行量。2006 年上半年，私人標籤的份額進一步攀升至 $57\mathrm{\%}$$57\mathrm{\%}$57%57 \% 。截至 2006 年底，美國住宅抵押貸款總債務為 $\mathrm{\$}11$$\mathrm{\$}11$$11\$ 11 兆美元，GSEs 未償還的 MBS 仍達 $\mathrm{\$}2.5$$\mathrm{\$}2.5$$2.5\$ 2.5 兆美元，但 GSEs 的重要性似乎正在減弱。 ${}^{15}$${}^{15}$^(15){ }^{15}

The GSEs were major buyers of private label MBSs, and held a combined portfolio of about $1.5 trillion of mortgages and MBSs. The GSEs claimed to be relatively immune to credit risk because they required private mortgage insurance for loans whose LTV was above $80\mathrm{\%}$$80\mathrm{\%}$80%80 \% and because their private label MBSs were highly rated by credit rating agencies. Private label MBSs were obligations of specially created trusts who held ultimate title to packages of mortgages and who distributed the interest, principal, and other payments of the borrowers to different classes (or “tranches”) of bonds according to pre-specified priority formulas. The rating agencies then gave higher ratings to bonds that had higher priority in this distribution of funds. The formulas allocating payments to bonds
GSEs 是私人標籤抵押貸款支持證券（MBS）的主要買家，合計持有約 1.5 兆美元的抵押貸款和 MBS 投資組合。GSEs 聲稱相對免疫於信用風險，因為他們要求貸款的貸款價值比率（LTV）超過 $80\mathrm{\%}$$80\mathrm{\%}$80%80 \% 時必須有私人抵押貸款保險，且他們的私人標籤 MBS 由信用評級機構評為高評級。私人標籤 MBS 是由特別設立的信託發行，該信託持有抵押貸款包的最終所有權，並根據預先指定的優先順序公式，將借款人的利息、本金及其他付款分配給不同等級（或稱「分層」）的債券。評級機構因此給予在資金分配中優先順序較高的債券較高的評級。分配給債券的付款公式
were complex. For example, the share of principal (but not interest) payments going to a particular bond could depend on whether realized losses exceeded a threshold by a particular date.
較為複雜。例如，分配給特定債券的本金（但不包括利息）支付比例，可能取決於實際損失是否在特定日期前超過某個門檻。

In May 2005, for example, Moody’s rated the bonds of a trust set up by Morgan Stanley that held a bundle of geographically diverse subprime mortgages originated by New Century. Moody’s gave its highest Aaa rating to eight bond tranches, representing about $83\mathrm{\%}$$83\mathrm{\%}$83%83 \% of the $\mathrm{\$}1.4$$\mathrm{\$}1.4$$1.4\$ 1.4 billion of securities issued by the trust. An additional 9 tranches were given ratings that ranged from Aa1 to Baa3. Moody’s claimed that, on the basis of the quality of loans, it expected collateral losses between $4.5\mathrm{\%}$$4.5\mathrm{\%}$4.5%4.5 \% and $4.75\mathrm{\%}$$4.75\mathrm{\%}$4.75%4.75 \% for the trust as a whole. It also expressed great confidence in Countrywide and Homeq (a unit of Barclays), who would be in charge of “servicing” the mortgages. The servicing arrangement for this trust was filed at the SEC and required Countrywide and Homeq to “make reasonable efforts” to collect principal and interest. It also entitled the servicers to foreclose on properties when loans became delinquent, and promised that the servicers would be compensated for foreclosure costs, in addition to receiving a fee based on the mortgage pool as a whole. ${}^{16}$${}^{16}$^(16){ }^{16}
例如，2005 年 5 月，穆迪評級摩根士丹利設立的一個信託的債券，該信託持有由新世紀發起的一籃子地理位置多樣的次級抵押貸款。穆迪給予其中八個債券分層最高的 Aaa 評級，約佔該信託發行的 $\mathrm{\$}1.4$$\mathrm{\$}1.4$$1.4\$ 1.4 十億證券的 $83\mathrm{\%}$$83\mathrm{\%}$83%83 \% 。另外九個分層的評級介於 Aa1 至 Baa3 之間。穆迪聲稱，根據貸款質量，預計該信託整體的擔保品損失介於 $4.5\mathrm{\%}$$4.5\mathrm{\%}$4.5%4.5 \% 至 $4.75\mathrm{\%}$$4.75\mathrm{\%}$4.75%4.75 \% 之間。它還對負責“服務”這些抵押貸款的 Countrywide 和 Homeq（巴克萊旗下單位）表達了高度信心。該信託的服務安排已向美國證券交易委員會（SEC）提交，要求 Countrywide 和 Homeq“合理努力”收取本金和利息。該安排還授權服務商在貸款逾期時對房產進行止贖，並承諾服務商將獲得止贖費用的補償，此外還會根據整個抵押貸款池收取費用。 ${}^{16}$${}^{16}$^(16){ }^{16}

While Moody’s, S&P and Fitch all had the capacity to rate MBSs, issuers did not always purchase ratings from all three because, as Diane Wold of GMAC said, “it’s not economical” to do so. The result was that in 2000, when Moody’s was criticized for “being overly tough on issuers and employing stubbornly conservative ratings methodologies that failed to accurately reflect the market’s perception of the deals,” it rated only $39\mathrm{\%}$$39\mathrm{\%}$39%39 \% of all issues. By contrast, the market leader S&P rated $87\mathrm{\%}$$87\mathrm{\%}$87%87 \% of the deals in 2000 . After working hard to patch up its relationship with issuers by, for example, rolling out a new set of ratings guidelines, Moody’s raised its own share of the deals to $64\mathrm{\%}$$64\mathrm{\%}$64%64 \% in 2001 . ${}^{17}$${}^{17}$^(17){ }^{17}
儘管穆迪、標普和惠譽都有能力對抵押貸款支持證券（MBS）進行評級，但發行人並不總是會向三家評級機構全部購買評級，正如 GMAC 的黛安·沃爾德所說，「這樣做並不經濟」。結果是，在 2000 年，當穆迪因「對發行人過於嚴苛，並採用固執保守的評級方法，未能準確反映市場對交易的看法」而受到批評時，它只評級了所有發行量的 $39\mathrm{\%}$$39\mathrm{\%}$39%39 \% 。相比之下，市場領導者標普在 2000 年評級了 $87\mathrm{\%}$$87\mathrm{\%}$87%87 \% 的交易。經過努力修補與發行人的關係，例如推出一套新的評級指導方針後，穆迪在 2001 年將其評級份額提高到 $64\mathrm{\%}$$64\mathrm{\%}$64%64 \% 。 ${}^{17}$${}^{17}$^(17){ }^{17}

A further innovation took place in January 2006, when Markit and a consortium of investment banks active in the MBS market launched a set of indices called ABX-HE (the HE stood for home equity). These indices tracked the price of over-the-counter contracts (called credit default swaps) that acted like insurance on the payments made by borrowers on pre-specified packages of subprime MBSs. One side of these contracts expected to receive a stream of fixed payment and expected to make a stream of “floating” payments in exchange. The floating payment would increase whenever Markit determined that borrowers had fallen behind on the mortgages held by the MBSs or that a write-down of these MBSs was warranted (if this write-down was later reversed, the firm making “fixed” payments was expected to raise its payments to the other party). The standardization provided by these indices made it easy for financial institutions to bet on the evolution of the subprime market. Different ABX-HE indices tracked MBSs of different ratings. ${}^{18}$${}^{18}$^(18){ }^{18}
2006 年 1 月，Markit 與一個活躍於 MBS 市場的投資銀行聯盟推出了一組名為 ABX-HE 的指數（HE 代表房屋淨值），這是一項進一步的創新。這些指數追蹤場外合約（稱為信用違約掉期）的價格，該合約類似於對借款人在預先指定的次級 MBS 組合中所支付款項的保險。這些合約的一方預期收到固定付款流，並預期以浮動付款流作為交換。當 Markit 判定借款人拖欠 MBS 所持有的抵押貸款或這些 MBS 需要減記時，浮動付款將增加（如果後來該減記被撤銷，支付「固定」款項的一方預期會提高對另一方的付款）。這些指數所提供的標準化使金融機構能夠輕鬆押注次級市場的走向。不同的 ABX-HE 指數追蹤不同評級的 MBS。

The role of rating agencies in the secondary mortgage market did not end with the rating of the MBSs themselves. Many MBSs found their way into debt-financed institutions, and the rating agencies were obviously in an ideal position to rate their debt as well. According to a Lehman Brothers report, a substantial fraction of subordinate MBS securities (those rated below AAA) were held by collateralized debt obligation (CDO) entities in 2007. ${}^{19}$${}^{19}$^(19){ }^{19} Legal CDO entities were typically created by investment banks and registered in tax havens like the Cayman Islands. Their obligations (which were also called CDOs) were once again arrayed in tranches, with some tranches receiving the payments from the entity’s assets before others. Many senior tranches of CDOs holding MBSs obtained high ratings, in all likelihood because statistical models determined them to be safe.
評級機構在次級抵押貸款市場中的角色並未隨著對 MBS 本身的評級而結束。許多 MBS 流入了以債務融資的機構，評級機構顯然處於理想位置來評級這些債務。根據雷曼兄弟的一份報告，2007 年有相當大比例的次級 MBS 證券（評級低於 AAA）被擔保債務憑證（CDO）實體持有。 ${}^{19}$${}^{19}$^(19){ }^{19} 法律上的 CDO 實體通常由投資銀行創建，並在開曼群島等避稅天堂註冊。它們的債務（也稱為 CDO）再次被分成多個分層，其中一些分層比其他分層優先從實體資產中獲得付款。許多持有 MBS 的 CDO 高級分層獲得了高評級，很可能是因為統計模型判定它們是安全的。

It was common for CDOs to hold assets of other CDOs. CDOs were also heavily involved in markets for derivative securities. Using trades based on the ABX-HE indices described above, a Merrill Lynch CDO named Norma effectively sold insurance against losses in BBB-rated tranches of mortgage-backed securities. In spite of the apparent risk of this strategy, Fitch, S&P and Moody’s all gave their highest rating to tranches representing 75% of Norma’s value in March 2007. According to a Wall Street Journal article, a hedge fund named Magnetar pushed for Norma’s creation and
CDO 持有其他 CDO 資產是很常見的情況。CDO 也深度參與衍生性金融商品市場。利用上述所述的 ABX-HE 指數進行交易，Merrill Lynch 的一個名為 Norma 的 CDO 實際上出售了針對 BBB 評級抵押貸款支持證券分層損失的保險。儘管這種策略存在明顯風險，Fitch、S&P 和 Moody’s 在 2007 年 3 月仍給予代表 Norma 價值 75%的分層最高評級。根據《華爾街日報》的一篇報導，一個名為 Magnetar 的對沖基金推動了 Norma 的成立並
purchased its riskiest tranches. Magnetar, whose funding sources were unknown, apparently intended to hedge this risk by buying insurance against the fall in AAA tranches of MBS. ${}^{20}$${}^{20}$^(20){ }^{20}
購買了其風險最高的分層。Magnetar 的資金來源不明，顯然打算通過購買針對 MBS AAA 分層下跌的保險來對沖這一風險。 ${}^{20}$${}^{20}$^(20){ }^{20}

CDOs tended to have long-term debt. As an alternative, several commercial banks invented institutions that bought MBSs with funds raised in the short-term commercial paper market. To do so at attractive rates, a favorable rating was obviously essential. The rating agencies obliged. In the case of “conduits,” high ratings were apparently based on lines of credit from the sponsoring institution. Such lines of credit assured the purchasers of the conduit’s commercial paper that the conduit would not have to sell its assets in a “fire sale” if the commercial paper market suddenly dried up.
CDO 通常具有長期債務。作為替代方案，幾家商業銀行創造了利用短期商業票據市場籌集資金購買 MBS 的機構。為了以有吸引力的利率進行操作，獲得良好的評級顯然是必須的。評級機構也予以配合。在「通道」（conduits）的案例中，高評級顯然是基於發起機構提供的信用額度。這些信用額度向通道商業票據的購買者保證，如果商業票據市場突然枯竭，通道不必以「火災拍賣」的方式出售其資產。

Citibank pioneered a related structure, called a SIV (or structured investment vehicle). SIVs also relied on commercial paper and also required high ratings to do so. In the case of SIVs, however, the ering bank’s line of credit was smaller and outside investors were brought in to take an equity position. As a result, regulators allowed sponsoring banks to keep SIVs off their balance sheets.
花旗銀行開創了一種相關結構，稱為 SIV（或結構性投資工具）。SIV 也依賴商業票據，並且同樣需要高評級。然而，在 SIV 的情況下，發起銀行的信用額度較小，並引入外部投資者持有股權。因此，監管機構允許發起銀行將 SIV 排除在其資產負債表之外。

After a borrower became delinquent on his or her loan for a pre-specified period of time, the lender could initiate foreclosure proceedings. Unless the borrower made sufficient payments to become current once again (including fees for late payments) or the borrower reached an agreement with the lender, an official foreclosure auction followed. The borrower could then be evicted from the house. Information on the number of borrowers losing their homes was scant. The Mortgage Bankers Association (MBA) did conduct a survey of loans and reported the fraction of those that were either delinquent or in the foreclosure process. Unfortunately, the coverage of subprime loans in the MBA survey was initially incomplete.
當借款人逾期未還款達到預定時間後，貸款人可以啟動止贖程序。除非借款人支付足夠款項使貸款恢復正常（包括逾期費用）或與貸款人達成協議，否則將進行正式的止贖拍賣。借款人隨後可能被驅逐出房屋。關於失去房屋的借款人數量資訊稀少。抵押貸款銀行協會（MBA）確實進行了貸款調查，並報告了逾期或處於止贖程序中的貸款比例。不幸的是，MBA 調查中對次級貸款的涵蓋最初並不完整。

According to this survey, the overall foreclosure rate fell from $1.23\mathrm{\%}$$1.23\mathrm{\%}$1.23%1.23 \% in the second quarter of 2002 to $1.08\mathrm{\%}$$1.08\mathrm{\%}$1.08%1.08 \% in the first quarter of 2005 . Over the same period, the fraction of loans that was delinquent fell from $4.77\mathrm{\%}$$4.77\mathrm{\%}$4.77%4.77 \% to $4.31\mathrm{\%}$$4.31\mathrm{\%}$4.31%4.31 \%. This puzzled analysts who expected the increase in subprime loans to push this number up. One explanation that was offered was that house price appreciation had allowed many borrowers to refinance loans they could not afford. A related explanation was that when borrowers had some equity in their house, they might agree to sell their homes to extinguish their mortgages and these sales would not be reported as foreclosures. ${}^{21}$${}^{21}$^(21){ }^{21}
根據這項調查，整體止贖率從 2002 年第二季的 $1.23\mathrm{\%}$$1.23\mathrm{\%}$1.23%1.23 \% 下降到 2005 年第一季的 $1.08\mathrm{\%}$$1.08\mathrm{\%}$1.08%1.08 \% 。同期內，逾期貸款比例從 $4.77\mathrm{\%}$$4.77\mathrm{\%}$4.77%4.77 \% 下降到 $4.31\mathrm{\%}$$4.31\mathrm{\%}$4.31%4.31 \% 。這讓分析師感到困惑，因為他們預期次級貸款的增加會推高這個數字。一個解釋是房價上漲使許多借款人能夠重新貸款，償還他們無法負擔的貸款。另一個相關的解釋是，當借款人在房屋中有一些淨值時，他們可能同意出售房屋以清償抵押貸款，而這些銷售不會被報告為止贖。 ${}^{21}$${}^{21}$^(21){ }^{21}

Foreclosures and delinquencies did rise during 2006 and 2007. In statistics released a few months after the fact, MBA reported increases in the delinquency rate to $4.95\mathrm{\%}$$4.95\mathrm{\%}$4.95%4.95 \% in the fourth quarter of 2006, to $5.59\mathrm{\%}$$5.59\mathrm{\%}$5.59%5.59 \% in the third quarter of 2007 and then to $5.62\mathrm{\%}$$5.62\mathrm{\%}$5.62%5.62 \% in the fourth quarter of 2007 . The MBA’s overall foreclosure rate went from $1.19\mathrm{\%}$$1.19\mathrm{\%}$1.19%1.19 \% of all loans outstanding in the fourth quarter of 2006 to a record rate of $2.04\mathrm{\%}$$2.04\mathrm{\%}$2.04%2.04 \% in the fourth quarter of 2007. Meanwhile, RealtyTrac, a company that specialized in helping potential buyers find foreclosed properties, started issuing press releases suggesting that foreclosures were rising dramatically (see Exhibit 3). In December 2007, the MBA announced that, while only $0.79\mathrm{\%}$$0.79\mathrm{\%}$0.79%0.79 \% of prime loans were in foreclosure in the third quarter of 2007, the rate for subprime loans was $6.89\mathrm{\%}$$6.89\mathrm{\%}$6.89%6.89 \%,(see Exhibit 4). A Federal Reserve Bank of Boston study published at the same time concluded that about $20\mathrm{\%}$$20\mathrm{\%}$20%20 \% of loans that were originally financed with a subprime mortgage would end in foreclosure. Observing that foreclosures were more common in neighborhoods where house prices had fallen more, this study predicted even more foreclosures if house prices fell more. ${}^{22}$${}^{22}$^(22){ }^{22}
2006 年和 2007 年間，止贖和拖欠確實有所上升。根據幾個月後發布的統計數據，MBA 報告稱，拖欠率在 2006 年第四季度上升至 $4.95\mathrm{\%}$$4.95\mathrm{\%}$4.95%4.95 \% ，2007 年第三季度上升至 $5.59\mathrm{\%}$$5.59\mathrm{\%}$5.59%5.59 \% ，然後在 2007 年第四季度上升至 $5.62\mathrm{\%}$$5.62\mathrm{\%}$5.62%5.62 \% 。MBA 的整體止贖率從 2006 年第四季度所有未償還貸款的 $1.19\mathrm{\%}$$1.19\mathrm{\%}$1.19%1.19 \% 上升到 2007 年第四季度的創紀錄 $2.04\mathrm{\%}$$2.04\mathrm{\%}$2.04%2.04 \% 。同時，專門幫助潛在買家尋找止贖房產的公司 RealtyTrac 開始發布新聞稿，暗示止贖數量急劇增加（見圖表 3）。2007 年 12 月，MBA 宣布，儘管 2007 年第三季度只有 $0.79\mathrm{\%}$$0.79\mathrm{\%}$0.79%0.79 \% 的優質貸款處於止贖狀態，但次級貸款的止贖率為 $6.89\mathrm{\%}$$6.89\mathrm{\%}$6.89%6.89 \% （見圖表 4）。波士頓聯邦儲備銀行同時發表的一項研究得出結論，約有 $20\mathrm{\%}$$20\mathrm{\%}$20%20 \% 的原本以次級抵押貸款融資的貸款最終會以止贖告終。該研究觀察到，止贖在房價下跌較多的社區更為常見，並預測如果房價進一步下跌，止贖數量將會更多。 ${}^{22}$${}^{22}$^(22){ }^{22}

The economic and human costs of foreclosures were substantial. Borrowers were often dumbfounded to see their debt soar in the foreclosure process as lenders slapped myriad fees on top
歐洲經濟和人力成本的止贖損失相當巨大。借款人常常對於在止贖過程中債務激增感到震驚，因為貸款人會在其上加收各種費用
of the late fees that borrowers incurred when falling behind on their payments. More tragically still, the loss of their home struck many as a devastating personal blow. One borrower said, “I feel like a failure.” Said another, “I feel everything we worked for to that point is totally destroyed”. ${}^{23}$${}^{23}$^(23){ }^{23}
以及借款人因延遲付款而產生的滯納金。更悲慘的是，失去家園對許多人來說是一種毀滅性的個人打擊。一位借款人說：「我覺得自己像個失敗者。」另一位說：「我覺得我們為此努力的一切都完全被摧毀了。」 ${}^{23}$${}^{23}$^(23){ }^{23}

Maintenance on homes that went through foreclosure had often been neglected, resulting in a low market value. In some cases, irate borrowers caused damage to the property before being evicted. Since buyers in foreclosure auctions obtained the property “as-is,” their willingness to pay was modest. As a result, banks would often bid themselves so as to keep their property after foreclosure. According to Realtytrac, banks did this for about 50,000 properties in both October and November 2007. ${}^{24}$${}^{24}$^(24){ }^{24} These properties then sat vacant, attracting vandals and criminals.
維護經過止贖的房屋常常被忽視，導致市場價值低落。在某些情況下，憤怒的借款人在被驅逐前會對房產造成損害。由於止贖拍賣的買家是以「現況」取得房產，他們的出價意願相當有限。因此，銀行經常會自行出價以保留止贖後的房產。根據 Realtytrac 的資料，銀行在 2007 年 10 月和 11 月約有 5 萬處房產採取此舉。 ${}^{24}$${}^{24}$^(24){ }^{24} 這些房產隨後閒置，吸引了破壞者和罪犯。

The devastation hit communities such as Cleveland and Detroit particularly hard. These cities had been declining for some time and the wave of foreclosures converted some of their middle class neighborhoods into boarded up battle zones where foreclosed homes could be obtained for as little as $\mathrm{\$}1,000$$\mathrm{\$}1,000$$1,000\$ 1,000. Minority and immigrant communities were especially affected, as these groups had received a disproportionate share of subprime loans (see Exhibit 5). Nationally, foreclosures were high also in areas with construction booms like Florida, and in California, where Stockton apparently had the highest foreclosure rate in the U.S. Foreclosures affected not only homeowners but also renters, who were routinely evicted, often on short notice. ${}^{25}$${}^{25}$^(25){ }^{25}
這場災難對克里夫蘭和底特律等社區造成了特別嚴重的打擊。這些城市已經衰退了一段時間，止贖潮使得一些中產階級社區變成了封鎖的戰區，止贖房屋的價格低至 $\mathrm{\$}1,000$$\mathrm{\$}1,000$$1,000\$ 1,000 。少數族裔和移民社區受到的影響尤為嚴重，因為這些群體獲得了不成比例的次貸貸款（見圖表 5）。全國範圍內，止贖率在建設熱潮地區如佛羅里達州也很高，加利福尼亞州的斯托克頓顯然擁有全美最高的止贖率。止贖不僅影響了房主，也影響了租戶，租戶經常被驅逐，且通常通知時間很短。 ${}^{25}$${}^{25}$^(25){ }^{25}

Given the substantial costs on both sides, one might imagine that lenders and borrowers would have preferred to come to an alternative arrangement. One possibility was to sell the house amicably before foreclosure with lenders forgiving the shortfall between the amount raised in the sale and the amount of the loan outstanding (a “short sale”). Another was to let the borrower stay in the house and make reduced payments. Lenders often claimed that they sought such negotiations with borrowers, but also acknowledged that the rising volume of foreclosures swamped their negotiation capacity. Borrowers, on the other hand, were often frustrated by the unresponsiveness of lenders. Indeed, a number of borrowers succumbed to scams in which they paid people who fraudulently claimed to be effective at renegotiating mortgage terms with lenders.
鑑於雙方承擔的巨大成本，人們可能會認為貸款人和借款人更願意達成其他安排。一種可能是，在止贖前友好地出售房屋，貸款人免除銷售所得與未償還貸款金額之間的差額（「短售」）。另一種是讓借款人繼續居住在房屋中並支付減少的款項。貸款人經常聲稱他們尋求與借款人進行此類協商，但也承認不斷增加的止贖案件淹沒了他們的協商能力。另一方面，借款人常因貸款人反應遲鈍而感到沮喪。事實上，許多借款人因受騙而向聲稱能有效與貸款人重新協商抵押貸款條款的詐騙者付款。

Some observers blamed the securitization of mortgages for the lack of negotiation between borrowers and lenders. In fact, servicer agreements often gave little leeway for servicers to cut the borrowers’ payments, while compensating servicers for their costs of pursuing foreclosure. It was also alleged that any deal between the borrower and the servicer had the potential to affect different tranches of MBSs differently and could thus lead to litigation. ${}^{26}$${}^{26}$^(26){ }^{26}
一些觀察家將抵押貸款證券化歸咎於借款人與貸款人之間缺乏協商。事實上，服務商協議通常給予服務商很少的空間來減免借款人的還款，同時補償服務商追討止贖的成本。還有人指控，借款人與服務商之間的任何協議都有可能對不同的 MBS（抵押貸款支持證券）分層產生不同影響，從而可能引發訴訟。 ${}^{26}$${}^{26}$^(26){ }^{26}

The first firms to suffer from the increase in foreclosures were mortgage originators. Many of them used short-term funds for financing and experienced cash squeezes when, in early 2007, statistics came to suggest that the delinquency rate on subprime loans was higher than anticipated. Under pressure from its creditors, New Century closed its lending facility in March 2007. It had been the second largest subprime lender and had revealed that $2.5\mathrm{\%}$$2.5\mathrm{\%}$2.5%2.5 \% of its borrowers had defaulted on their first mortgage payment in 2006.
首批因止贖增加而受損的公司是抵押貸款發起機構。許多公司使用短期資金進行融資，當 2007 年初統計數據顯示次級貸款的違約率高於預期時，現金流緊縮。在債權人的壓力下，新世紀金融公司於 2007 年 3 月關閉了其貸款業務。該公司曾是第二大次級貸款機構，並透露其 $2.5\mathrm{\%}$$2.5\mathrm{\%}$2.5%2.5 \% 的借款人在 2006 年未能按時支付首筆抵押貸款。

Financial markets took this in stride, and remained fairly calm until early August 2007 in spite of having witnessed earlier the disintegration of two hedge funds controlled by Bear Stearns. The funds had borrowed extensively and invested in mortgage-backed assets, including un-rated tranches of CDOs that invested in other CDOs. While Bear Stearns may not have been legally obliged to do so, it ended up compensating many of the hedge funds’ creditors.
金融市場對此泰然處之，並保持相當平靜，直到 2007 年 8 月初，儘管早前已見證貝爾斯登控制的兩隻對沖基金解體。這些基金大量借貸並投資於抵押貸款支持資產，包括投資於其他 CDO 的未評級 CDO 分層。儘管貝爾斯登可能沒有法律義務這麼做，但最終還是補償了許多對沖基金的債權人。

By early August, however, many observers were bemoaning the existence of a “credit crunch” and of a “subprime crisis.” Chairman of the Federal Reserve Board Ben Bernanke, meanwhile, had recognized in several 2007 speeches that the housing sector had slowed and that this was likely to reduce the growth of the economy as a whole. Still, his July 18 testimony to Congress emphasized the threat of inflation. He had long said that his “comfort zone” for inflation was between 1 and $2\mathrm{\%}$$2\mathrm{\%}$2%2 \%, and had advocated that the Fed follow the lead of many other central banks (including the European Central Bank) by adopting an explicit inflation target. On August 7, stating that its predominant concern was “the risk that inflation will fail to moderate,” the Fed kept its target Federal Funds rate unchanged at $5.25\mathrm{\%}$$5.25\mathrm{\%}$5.25%5.25 \%.
然而，到八月初，許多觀察家開始哀嘆存在「信貸緊縮」和「次貸危機」。聯邦儲備委員會主席本·伯南克在 2007 年的幾次演講中已經認識到房地產部門放緩，這可能會減緩整體經濟的增長。然而，他在 7 月 18 日向國會的證詞中強調了通脹的威脅。他長期以來一直表示，他對通脹的「舒適區間」是在 1 到 $2\mathrm{\%}$$2\mathrm{\%}$2%2 \% 之間，並主張聯準會應該效法包括歐洲央行在內的許多其他中央銀行，採用明確的通脹目標。8 月 7 日，聯準會表示其主要關注點是「通脹無法適度緩和的風險」，因此將聯邦基金利率目標維持在 $5.25\mathrm{\%}$$5.25\mathrm{\%}$5.25%5.25 \% 不變。

The European Central Bank (ECB) also had a target interest rate, and had last increased it to $4\mathrm{\%}$$4\mathrm{\%}$4%4 \% in June 2007. Unlike the Fed, which intervened in the overnight market every day, the ECB normally only added liquidity to the market in special pre-scheduled auctions. On August 8, however, in response to a spike in its overnight rate to $4.7\mathrm{\%}$$4.7\mathrm{\%}$4.7%4.7 \%, the ECB lent $\mathrm{\$}129$$\mathrm{\$}129$$129\$ 129 billion in an emergency maneuver. This marked the beginning of active official central bank interventions to quell financial market problems that were linked to problematic U.S. mortgages. ${}^{27}$${}^{27}$^(27){ }^{27}
歐洲中央銀行（ECB）也有一個目標利率，並且在 2007 年 6 月最後一次將其提高到 $4\mathrm{\%}$$4\mathrm{\%}$4%4 \% 。與每天在隔夜市場進行干預的聯邦儲備系統（Fed）不同，ECB 通常只在特別預定的拍賣中向市場注入流動性。然而，在 8 月 8 日，因應隔夜利率飆升至 $4.7\mathrm{\%}$$4.7\mathrm{\%}$4.7%4.7 \% ，ECB 進行了一次緊急操作，貸出 $\mathrm{\$}129$$\mathrm{\$}129$$129\$ 129 十億。這標誌著官方中央銀行開始積極干預，以平息與美國問題抵押貸款相關的金融市場問題。 ${}^{27}$${}^{27}$^(27){ }^{27}

It was not, of course, certain why overnight rates had spiked in Europe, nor why the LIBOR (London Interbank Offer Rate) interest rates that banks charged each other for funds (see Exhibit 7) behaved unusually from this point onwards. What was revealed around this time was that, contrary to what their leaders had claimed only a few weeks before, certain European banks were heavily exposed to subprime US mortgages.
當然，當時並不確定為何歐洲的隔夜利率會飆升，也不清楚為何銀行間相互收取資金的倫敦銀行同業拆借利率（LIBOR）（見圖表 7）從此表現異常。當時揭露的是，與幾週前其領導人所聲稱的相反，某些歐洲銀行對美國次級抵押貸款有大量曝險。

The first one to admit this was IKB, a German bank whose major shareholder was a state-owned bank called KfW. IKB had set up an off-balance sheet entity called Rhineland that had evidently taken big bets on US mortgages. When Rhineland found it difficult to raise funds in the commercial paper market, it turned to IKB, which also lacked the necessary funds. By August 3, it became known that the German government had orchestrated a rescue operation involving a further injection of funds from KfW.
第一個承認這一點的是 IKB，一家德國銀行，其主要股東是一家名為 KfW 的國有銀行。IKB 設立了一個名為 Rhineland 的表外實體，顯然在美國抵押貸款上進行了大規模押注。當 Rhineland 在商業票據市場籌集資金遇到困難時，它轉向了同樣缺乏必要資金的 IKB。到了 8 月 3 日，德國政府已經策劃了一場救援行動，涉及 KfW 進一步注資。

The problems of BNP Paribas were revealed on the same day as the ECB intervention. BNP, the bank with the most total assets in the Eurozone, announced that it was suspending redemptions from three funds. The reason it gave was that it was impossible to value these funds as a result of their exposure to U.S. subprime mortgages. After this, rating agencies downgraded large batches of MBSs and CDOs while, one after another, big banks like HSBC and Citibank brought their SIVs back onto their books. Along the way, the great financial houses recognized losses on their investments. The Canadian bank CIBC warned investors of further losses when ACA, a company that had been its counterparty in hedges, found its own bonds downgraded to junk (see Exhibit 6).
BNP 巴黎銀行的問題在歐洲央行干預的同一天被揭露。BNP 是歐元區資產總額最大的銀行，宣布暫停三個基金的贖回。其原因是由於這些基金暴露於美國次貸抵押貸款，無法對其進行估值。隨後，評級機構下調了大量的抵押貸款支持證券（MBS）和擔保債務憑證（CDO）的評級，而像匯豐銀行和花旗銀行這樣的大型銀行也紛紛將其資產支持投資工具（SIV）重新納入賬面。在此過程中，這些大型金融機構承認了其投資的損失。加拿大銀行 CIBC 警告投資者，當其對手方之一的 ACA 公司債券被降級為垃圾債券時，可能會有進一步的損失（見附錄 6）。

Banks that suffered heavy losses needed capital infusions to stay in business. Several obtained help from abroad, particularly from government controlled enterprises that belonged to countries that were running current account surpluses. Thus, the government of Abu Dhabi bought a stake of Citigroup, the Chinese government bought stakes in Bear Stearns and Morgan Stanley, and the Singaporean government invested in Merrill Lynch and UBS.
遭受重大損失的銀行需要資本注入以維持經營。多家銀行獲得了來自海外的援助，特別是來自那些擁有經常帳戶順差國家的政府控股企業。因此，阿布達比政府購買了花旗集團的股份，中國政府購買了貝爾斯登和摩根士丹利的股份，新加坡政府則投資了美林和瑞銀。

Aside from the interbank market, two parts of the financial system showed strains from August onwards. First, of course, was the market for MBSs and CDOs. It was not just that the ABX-HE indices fell in value. It was also claimed that the MBS market was frozen and that it was difficult to trade even AAA MBSs. Expressing his frustration with this, Ben Bernanke said, "I’d like to know what those damn things are worth. ${}^{\prime \prime 2}$${}^{\prime \prime 2}$^(''2){ }^{\prime \prime 2} Second, the volume of asset-backed commercial paper fell by $30\mathrm{\%}$$30\mathrm{\%}$30%30 \% and the volatility of money market yields increased. The difference between the annualized yield on 30-day commercial paper and that on treasury bills, which had hovered between 25 and 50
除了銀行間市場外，從八月開始，金融系統的兩個部分顯示出壓力。首先，當然是 MBS 和 CDO 市場。不僅僅是 ABX-HE 指數價值下跌，還有人聲稱 MBS 市場已經凍結，甚至連 AAA 級的 MBS 也難以交易。對此感到沮喪的本·伯南克說：「我想知道那些該死的東西到底值多少錢。 ${}^{\prime \prime 2}$${}^{\prime \prime 2}$^(''2){ }^{\prime \prime 2} 」其次，資產支持商業票據的交易量下降了 $30\mathrm{\%}$$30\mathrm{\%}$30%30 \% ，而貨幣市場收益率的波動性增加。30 天商業票據的年化收益率與國庫券之間的差距，年初時徘徊在 25 到 50 個基點之間，
basis points earlier in the year, was 155 points on August 28 and had reached 365 basis points on August 20. One UBS trader described the situation by saying, "The market moved beyond mathematical reason and into emotion. ${}^{\prime 29}$${}^{\prime 29}$^('29){ }^{\prime 29}
8 月 28 日達到 155 個基點，8 月 20 日更達到 365 個基點。一位瑞銀交易員形容當時的情況說：「市場已經超越了數學理性，進入了情緒化階段。 ${}^{\prime 29}$${}^{\prime 29}$^('29){ }^{\prime 29} 」

One question that many observers asked was how the relatively small size of the potential losses associated with defaults on subprime mortgages could cause so much turmoil. Estimates of these potential losses varied tremendously, with the Congressional Budget Office pegging them at $\mathrm{\$}100$$\mathrm{\$}100$$100\$ 100 billion at the end of October 2007 and Merrill Lynch claiming at the end of December that they would equal $\mathrm{\$}500$$\mathrm{\$}500$$500\$ 500 billion once option ARMs were factored in. According to a Lehman report of early September, subprime loans amounted to about 1.2 trillion dollars. Assuming constant house prices, this report estimated total credit losses of about $\mathrm{\$}110$$\mathrm{\$}110$$110\$ 110 billion. Supposing that price declines would lead fewer people to repay their loans, a scenario with $30-35\mathrm{\%}$$30-35\mathrm{\%}$30-35%30-35 \% drops in house prices led to losses of $\mathrm{\$}240$$\mathrm{\$}240$$240\$ 240 billion. Some banks, including UBS, were basing their own expected losses on the behavior of ABX-HE indices. In December, the index for AAA-rated MBSs traded at about $70\mathrm{\%}$$70\mathrm{\%}$70%70 \% of par value, which was consistent with a loss of around $50\mathrm{\%}$$50\mathrm{\%}$50%50 \% on all subprime loans. This implied huge losses, though this approach was criticized because the contracts on this index were not heavily traded. ${}^{30}$${}^{30}$^(30){ }^{30}
許多觀察家提出的一個問題是，與次級抵押貸款違約相關的潛在損失規模相對較小，如何能引起如此大的動盪。對這些潛在損失的估計差異極大，國會預算辦公室在 2007 年 10 月底將其定為 $\mathrm{\$}100$$\mathrm{\$}100$$100\$ 100 十億，而美林證券在 12 月底聲稱，一旦考慮選擇權可調利率抵押貸款（option ARMs），損失將達 $\mathrm{\$}500$$\mathrm{\$}500$$500\$ 500 十億。根據 9 月初雷曼兄弟的一份報告，次級貸款約為 1.2 兆美元。假設房價保持不變，該報告估計總信用損失約為 $\mathrm{\$}110$$\mathrm{\$}110$$110\$ 110 十億。假設房價下跌會導致較少人償還貸款，一個房價下跌 $30-35\mathrm{\%}$$30-35\mathrm{\%}$30-35%30-35 \% 的情景導致損失達 $\mathrm{\$}240$$\mathrm{\$}240$$240\$ 240 十億。一些銀行，包括瑞銀，基於 ABX-HE 指數的表現來預測自身的預期損失。12 月，AAA 級抵押支持證券（MBS）的指數交易價格約為面值的 $70\mathrm{\%}$$70\mathrm{\%}$70%70 \% ，這與所有次級貸款約 $50\mathrm{\%}$$50\mathrm{\%}$50%50 \% 的損失相符。這暗示了巨大的損失，儘管這種方法受到批評，因為該指數的合約交易不活躍。 ${}^{30}$${}^{30}$^(30){ }^{30}

One might have imagined that even a loss of $\mathrm{\$}500$$\mathrm{\$}500$$500\$ 500 billion was manageable for the global financial system. One potential source of stress, however, was the fear that other forms of debt like credit cards could suffer a similar fate. Adding to the fear, the financial system was heavily leveraged. According to analyst James Bianco, the net debt of Wall Street firms went from $\mathrm{\$}300$$\mathrm{\$}300$$300\$ 300 billion in 2000 to $\mathrm{\$}1.4$$\mathrm{\$}1.4$$1.4\$ 1.4 trillion in July 2007. ${}^{31}$${}^{31}$^(31){ }^{31}
人們或許會認為，即使損失了 $\mathrm{\$}500$$\mathrm{\$}500$$500\$ 500 十億，對全球金融系統來說仍是可控的。然而，一個潛在的壓力來源是擔心其他形式的債務，如信用卡，可能會遭遇類似的命運。加劇恐慌的是，金融系統槓桿率極高。根據分析師詹姆斯·比安科的說法，華爾街公司的淨債務從 2000 年的 $\mathrm{\$}300$$\mathrm{\$}300$$300\$ 300 十億增加到 2007 年 7 月的 $\mathrm{\$}1.4$$\mathrm{\$}1.4$$1.4\$ 1.4 兆。 ${}^{31}$${}^{31}$^(31){ }^{31}

With high and volatile short term interest rates, the Federal Reserve found itself faced with conflicting pressures. On one hand, Wall Street wanted rate cuts. Traders warned Federal Reserve officials, some of whom they saw as ivory tower academics, that markets would reel otherwise. On the other hand, some people viewed any rate cuts as fostering “moral hazard” by bailing out the people whose ill-founded bets had caused the subprime crisis in the first place. Former vicepresident of the Dallas Federal Reserve Bank Gerald O’Driscoll regarded such an action as a continuation of “a new approach to monetary policy” which he attributed to Alan Greenspan and which he deemed at least partly responsible for the crisis itself. This approach was an asymmetric one, in which monetary policy was never tightened to offset asset price increases due to bubbles (on the ground that it was difficult to know whether a price increase was a bubble) but monetary policy was eased to counteract asset price decreases. According to Driscoll, the resulting monetary policy was biased towards inflation (in both the prices of goods and of assets). ${}^{32}$${}^{32}$^(32){ }^{32}
在短期利率高且波動劇烈的情況下，聯邦儲備委員會面臨著矛盾的壓力。一方面，華爾街希望降息。交易員警告聯邦儲備官員，其中一些人被視為象牙塔中的學者，否則市場將會動盪。另一方面，有些人認為任何降息都是助長「道德風險」，因為這是在救助那些因不當押注而引發次貸危機的人。達拉斯聯邦儲備銀行前副總裁傑拉爾德·奧德里斯科爾認為這種行動是「貨幣政策新方法」的延續，他將此歸因於艾倫·格林斯潘，並認為這至少部分導致了危機。這種方法是不對稱的，貨幣政策從不因泡沫引起的資產價格上漲而收緊（理由是難以判斷價格上漲是否為泡沫），但會放寬以抵消資產價格下跌。根據奧德里斯科爾的說法，所產生的貨幣政策偏向於通脹（包括商品價格和資產價格）。

Those who wanted to cut rates argued that this was not a policy aimed at traders but one that sought to prevent a recession. However, skeptical voices noted that the price of oil was still rising and that the economy was performing well. The debate suggested that people differed on the extent to which the Fed should focus directly on the health of the financial system, and also on how this health ought to be measured. There was further debate on whether the Fed ought to respond to inflation and aggregate activity only after observing changes in these variables, which would have the advantage of making Fed actions predictable. Alternatively, the Fed could seek to be preemptive with respect to these output and inflation, and this seemed to provide another rationale for responding to financial market developments.
那些主張降息的人認為，這不是針對交易者的政策，而是旨在防止經濟衰退。然而，持懷疑態度的聲音指出，油價仍在上漲，經濟表現良好。這場辯論顯示人們對聯準會應該在多大程度上直接關注金融系統的健康狀況，以及應該如何衡量這種健康狀況存在分歧。還有進一步的辯論是，聯準會是否應該僅在觀察到這些變數的變化後才對通脹和總體經濟活動作出反應，這樣做的優點是使聯準會的行動可預測。或者，聯準會可以尋求對這些產出和通脹採取先發制人的措施，這似乎為回應金融市場的發展提供了另一個理由。

In the event, the Federal Reserve maintained its target Federal Funds rate at $5.25\mathrm{\%}$$5.25\mathrm{\%}$5.25%5.25 \% until the scheduled September 18 meeting of its decision-making body. Faced with a chorus of complaints, the Fed announced on August 17 that it would lower the “discount rate,” the rate at which banks could borrow from the Fed, from $6.25\mathrm{\%}$$6.25\mathrm{\%}$6.25%6.25 \% to $5.75\mathrm{\%}$$5.75\mathrm{\%}$5.75%5.75 \%. Typically, the actual Federal Funds rate was quite close to its target. After August 10, this stopped being true and the actual Federal Funds rate fell below 5.00% quite frequently (see Exhibit 7).
事實上，聯邦儲備局將其目標聯邦基金利率維持在 $5.25\mathrm{\%}$$5.25\mathrm{\%}$5.25%5.25 \% ，直到預定於 9 月 18 日召開的決策機構會議。面對一片抱怨聲，聯邦儲備局於 8 月 17 日宣布將「貼現率」——銀行向聯邦儲備局借款的利率，從 $6.25\mathrm{\%}$$6.25\mathrm{\%}$6.25%6.25 \% 降至 $5.75\mathrm{\%}$$5.75\mathrm{\%}$5.75%5.75 \% 。通常，實際聯邦基金利率與其目標利率非常接近。8 月 10 日之後，情況不再如此，實際聯邦基金利率經常低於 5.00%（見圖表 7）。

From then on, the Fed did cut rates at every one of its scheduled meetings both at the end of 2007 and the beginning of 2008. It also cut the Federal Funds rate by an astounding 75 basis points at an unscheduled meeting on January 21, 2008. In addition, and starting on December 17, 2007, the Fed made 28-day loans to commercial banks through the newly created “Term Auction Facility.” According to critics, the collateral backing these 28 day loans was fairly weak. The first auction involved $20\mathrm{\$}$$20\mathrm{\$}$20$20 \$ billion, later auctions involved $50\mathrm{\$}$$50\mathrm{\$}$50$50 \$ billion. Noting that inflation was above its $2\mathrm{\%}$$2\mathrm{\%}$2%2 \% target, the ECB kept its own rate at the $4\mathrm{\%}$$4\mathrm{\%}$4%4 \% level it had chosen in June 2007, though it too found ways of injecting further liquidity into the European banking system.
從那時起，聯邦儲備局在 2007 年底和 2008 年初的每一次預定會議上都降息。2008 年 1 月 21 日的一次臨時會議上，聯邦儲備局更是驚人地將聯邦基金利率下調了 75 個基點。此外，從 2007 年 12 月 17 日開始，聯邦儲備局通過新設立的「定期拍賣機制」向商業銀行提供 28 天貸款。批評者指出，這些 28 天貸款所抵押的擔保品相當薄弱。第一次拍賣涉及 $20\mathrm{\$}$$20\mathrm{\$}$20$20 \$ 十億，後續拍賣涉及 $50\mathrm{\$}$$50\mathrm{\$}$50$50 \$ 十億。歐洲央行注意到通脹率高於其 $2\mathrm{\%}$$2\mathrm{\%}$2%2 \% 目標，於是將利率維持在 2007 年 6 月選定的 $4\mathrm{\%}$$4\mathrm{\%}$4%4 \% 水平，儘管如此，歐洲央行也尋找方法向歐洲銀行系統注入更多流動性。

The Fed’s most dramatic actions took place on March 17, 2008. At that point, it made public an arrangement through which JPMorgan would acquire Bear Stearns at the fire-sale price of $2 a share while the Fed would lend JPMorgan $\mathrm{\$}29$$\mathrm{\$}29$$29\$ 29 billion in a non-recourse loan collateralized with some of Bear Stearns’ assets. The Fed and U.S. Treasury had apparently pushed JPMorgan to offer a low price so as not to be accused of bailing out investors, though JPMorgan did later offer $\mathrm{\$}10$$\mathrm{\$}10$$10\$ 10 a share. The Fed’s loan represented a stark and controversial departure from the Fed’s history of lending only to regulated commercial banks while using only “jawboning” to help arrange financial rescues. The Fed’s simultaneous announcement that it was creating a “Term Securities Lending Facility,” to make 28-day (recourse) loans to investment banks reinforced the impression that it had entered a new era. Among other assets the Fed would accept as collateral for these new loans were AAA mortgage backed securities. ${}^{33}$${}^{33}$^(33){ }^{33}
聯準會最戲劇性的行動發生在 2008 年 3 月 17 日。當時，聯準會公開了一項安排，透過該安排，摩根大通將以每股 2 美元的火災拍賣價格收購貝爾斯登，而聯準會將向摩根大通提供 $\mathrm{\$}29$$\mathrm{\$}29$$29\$ 29 十億美元的無追索權貸款，該貸款以貝爾斯登的一些資產作為抵押。聯準會和美國財政部顯然施壓摩根大通提出低價，以免被指控救助投資者，儘管摩根大通後來確實提出了 $\mathrm{\$}10$$\mathrm{\$}10$$10\$ 10 美元的價格。聯準會的貸款代表了聯準會歷史上僅向受監管的商業銀行貸款並僅透過“口頭施壓”協助安排金融救援的明顯且具爭議的突破。聯準會同時宣布創建“定期證券借貸設施”，向投資銀行提供 28 天（有追索權）貸款，進一步加強了其已進入新時代的印象。聯準會願意接受作為這些新貸款抵押品的資產中包括 AAA 級抵押貸款支持證券。 ${}^{33}$${}^{33}$^(33){ }^{33}

The fear of a recession led many people to call for a fiscal expansion to complement the drop in interest rates. The most burning issue facing policymakers, however, was how to reduce the pain from foreclosures. Many observers claimed that foreclosures would inevitably swell as mortgage rates were reset. The Bush administration thus announced a plan on December 7 under which lenders would voluntarily freeze rates on certain ARM’s. The freeze would apply only to borrowers who were current in their payments and who had less than $3\mathrm{\%}$$3\mathrm{\%}$3%3 \% equity in their home. The plan was immediately criticized both for including too few people and for including too many.
對經濟衰退的恐懼使許多人呼籲擴大財政支出以配合利率的下降。然而，政策制定者面臨的最緊迫問題是如何減輕止贖帶來的痛苦。許多觀察家聲稱，隨著抵押貸款利率的重設，止贖案件將不可避免地增加。因此，布希政府於 12 月 7 日宣布了一項計劃，根據該計劃，貸款機構將自願凍結某些可調利率抵押貸款（ARM）的利率。該凍結僅適用於按時還款且房屋淨值低於 $3\mathrm{\%}$$3\mathrm{\%}$3%3 \% 的借款人。該計劃立即因涵蓋人數過少和過多而受到批評。

Investors were opposed to giving a break to people who could afford the reset, and the plan appeared to offer relief without analysis of the borrower’s income. Alan Reynolds of the Cato Institute complained by saying "It is the antithesis of a free market for the government to … pressure mortgage service companies into renegotiating contracts. ${}^{\prime 34}$${}^{\prime 34}$^('34){ }^{\prime 34} Other opponents objected both to the exclusion of many borrowers and to the voluntary nature of the plan, which they saw as implying that the rate freeze would apply only to those who could have obtained refinancing on their own. Presidential candidate Hillary Clinton favored a much broader freeze on interest rates together with a moratorium on foreclosures for 90 days. Her plan to rewrite existing contracts was positively horrifying to many observers.
投資者反對給予那些能夠承擔重設的人優惠，而該計劃似乎在未分析借款人收入的情況下提供救濟。卡托研究所的艾倫·雷諾茲抱怨說：「政府……施壓抵押貸款服務公司重新談判合約，這與自由市場背道而馳。」其他反對者則反對排除許多借款人以及該計劃的自願性質，他們認為這暗示利率凍結只適用於那些本可自行獲得再融資的人。總統候選人希拉蕊·柯林頓支持更廣泛的利率凍結以及 90 天的止贖暫停。她重寫現有合約的計劃令許多觀察家感到極度震驚。

An alternative supported by many lawmakers was to provide new government loans or grants to people in trouble. Some officials sought to increase the size of the mortgages that the GSEs could purchase and also wished to ease the restrictions faced by the FHA. There was disagreement, however, on whether it was wise to cut the minimum FHA down payment requirement to $1.5\mathrm{\%}$$1.5\mathrm{\%}$1.5%1.5 \% from the current minimum of $3\mathrm{\%}$$3\mathrm{\%}$3%3 \%. Opponents of FHA expansion noted that the agency’s loans had a higher delinquency rate than subprime loans and were also quite likely to end in foreclosure. ${}^{35}$${}^{35}$^(35){ }^{35}
許多立法者支持的另一種方案是向陷入困境的人提供新的政府貸款或補助金。一些官員試圖增加政府支持企業（GSEs）可購買的抵押貸款規模，並希望放寬聯邦住房管理局（FHA）所面臨的限制。然而，對於是否應將 FHA 的最低首付要求從目前的 $3\mathrm{\%}$$3\mathrm{\%}$3%3 \% 降至 $1.5\mathrm{\%}$$1.5\mathrm{\%}$1.5%1.5 \% ，存在分歧。反對擴大 FHA 規模的人指出，該機構的貸款違約率高於次級貸款，且也很可能以止贖告終。 ${}^{35}$${}^{35}$^(35){ }^{35}

The House Judiciary Committee, meanwhile, was working on a bill that would allow bankruptcy judges to reduce mortgage payments when the outstanding value of the loan exceeded the appraised value of the house. This change would lead mortgages to be treated more similarly to other claims in bankruptcy, though only when mortgage debts exceeded the value of the house. Mortgage industry representatives opposed this plan by claiming that it would ultimately increase the cost of borrowing.
與此同時，眾議院司法委員會正在制定一項法案，允許破產法官在貸款未償還金額超過房屋評估價值時，減少抵押貸款的還款額。此變更將使抵押貸款在破產中更類似於其他債權的處理方式，但僅限於抵押貸款債務超過房屋價值的情況。抵押貸款業界代表反對此計劃，聲稱這最終會增加借貸成本。

There were also competing House and Senate bills for regulating the mortgage industry itself. These were aimed squarely at reducing what was seen as an explosion of “predatory lending.” Advocacy groups such as the Center for Responsible Lending had campaigned for strengthened antipredator legislation throughout the subprime boom. While they welcomed some new state laws, including one in North Carolina in 1999 and one in Georgia in 2002, their pleas to Federal officials had fallen on deaf ears. Taking the opposite point of view, Vice President Dick Cheney dismissed rules against predatory lending on the ground that "We don’t want to interfere with the basic, fundamental working of the markets. ${}^{136}$${}^{136}$^(136){ }^{136}
國會眾議院和參議院也有針對抵押貸款行業本身的競爭性法案。這些法案的目標正是減少被視為「掠奪性貸款」爆炸性增長的現象。像負責任貸款中心這樣的倡導團體在次貸熱潮期間一直推動加強反掠奪性立法。雖然他們歡迎一些新的州法律，包括 1999 年北卡羅來納州和 2002 年喬治亞州的法律，但他們向聯邦官員的呼籲卻未被理會。持相反觀點的副總統迪克·切尼則以「我們不想干涉市場的基本、根本運作」為由，駁回了針對掠奪性貸款的規定。

The bills in front of Congress were aimed at strengthening an existing 1994 law called HOEPA (Home Ownership and Equity Protection Act), which forbade certain practices for “high-cost” loans. Loans satisfied the HOEPA definition of high cost if their interest rate was more than $8\mathrm{\%}$$8\mathrm{\%}$8%8 \% above the comparable Treasury rate or if they had combined origination fees above $8\mathrm{\%}$$8\mathrm{\%}$8%8 \%, and most subprime loans did not meet these thresholds. HOEPA barred prepayment penalties as well as negative amortization on high-cost loans. Critics of predatory lending adamantly opposed prepayment penalties, which they viewed as a tool for trapping the unsophisticated borrowers.
國會面前的法案旨在加強一項現有的 1994 年法律，稱為 HOEPA（住房擁有權與股權保護法案），該法禁止對「高成本」貸款採取某些做法。若貸款的利率比可比的國債利率高出 $8\mathrm{\%}$$8\mathrm{\%}$8%8 \% 以上，或其綜合開辦費用超過 $8\mathrm{\%}$$8\mathrm{\%}$8%8 \% ，則該貸款符合 HOEPA 對高成本的定義，而大多數次級貸款未達到這些門檻。HOEPA 禁止對高成本貸款收取提前還款罰金以及負攤銷。掠奪性貸款的批評者堅決反對提前還款罰金，認為這是用來困住不熟悉貸款的借款人的工具。

While the Fed was proposing to tighten HOEPA regulations, the Chairmen of the Finance Committees of the House and Senate, Representative Barney Frank and Senator Christopher Dodd, ridiculed this effort as insufficient and proposed stronger bills. The two bills were similar but the mortgage industry particularly opposed the Senate bill. The latter forced lenders of subprime mortgages to ascertain whether borrowers had the ability to repay their loans and forbade loans that did not meet this criterion. The Senate bill also included a presumption that borrowers were unable to pay if their house payments, including all taxes and condominium fees, exceeded $45\mathrm{\%}$$45\mathrm{\%}$45%45 \% of their income. In the case of refinances, the bill required the lender to ascertain that the borrower would receive a “net benefit” from the new loan. Perhaps the component of the bill that drew the fiercest opposition was the requirement that purchasers of subprime mortgages in the secondary market assume responsibility for any misbehavior by originators including fraud. The bill also required that mortgage brokers act in the interest of borrowers and forbade brokers from “steering” borrowers towards loans that would cost them more.
當聯邦儲備局提議收緊 HOEPA 規定時，眾議院和參議院財務委員會主席巴尼·弗蘭克議員和克里斯多福·多德參議員嘲笑這一努力不足，並提出了更強有力的法案。這兩項法案相似，但抵押貸款行業特別反對參議院的法案。後者要求次級抵押貸款的貸款人必須確定借款人有能力償還貸款，並禁止不符合此標準的貸款。參議院法案還包括一項假定，即如果借款人的房屋付款（包括所有稅款和公寓費用）超過其收入的 $45\mathrm{\%}$$45\mathrm{\%}$45%45 \% ，則假定借款人無法償還。在再融資的情況下，該法案要求貸款人確定借款人將從新貸款中獲得“淨利益”。該法案中可能引起最激烈反對的部分是要求次級抵押貸款的二級市場購買者承擔原始貸款人包括欺詐在內的任何不當行為的責任。 該法案還要求房貸經紀人必須以借款人的利益為優先，並禁止經紀人將借款人「引導」至成本較高的貸款。

Lenders were extremely concerned that “assignee liability” (the assigning of responsibility to purchasers of loans in the secondary market) could cause this market to disappear. Proponents of assignee liability argued instead that the health of the secondary market could be maintained as long as total liability was limited and participants in the market were told what methods for screening borrowers were acceptable. ${}^{37}$${}^{37}$^(37){ }^{37}
貸款人非常擔心「受讓人責任」（將責任轉移給次級市場的貸款購買者）可能導致該市場消失。受讓人責任的支持者則主張，只要總責任有限，且市場參與者被告知可接受的借款人篩選方法，次級市場的健康仍可維持。 ${}^{37}$${}^{37}$^(37){ }^{37}

Some economists were proposing a radically different approach. They wished to move away from the idea that homes should be strictly financed by debt and wanted to add something that looked like equity into the mix. Robert Shiller of Yale University and Andrew Caplin of NYU favored solutions where creditors would receive a fraction of a home’s appreciation when it was sold. One difficulty with crafting such contracts was that creditors would be unwilling to finance homes in this way unless they received a “rent” in addition to receiving the home appreciation on the fraction of the funds that they advanced. One proposal for dealing with this particular problem was to increase the fraction of the home’s appreciation that went to the creditor the longer the borrower stayed in the house. In any event, Shiller and Caplin felt that the subprime crisis provided a great opportunity for the government to help engineer a new wave of innovation in home finance. ${}^{38}$${}^{38}$^(38){ }^{38}
一些經濟學家提出了一種截然不同的方法。他們希望擺脫房屋應該嚴格以債務融資的觀念，並希望在其中加入類似股權的元素。耶魯大學的羅伯特·席勒和紐約大學的安德魯·卡普林支持這樣的解決方案：債權人在房屋出售時可獲得房屋增值的一部分。制定此類合約的一個困難是，債權人不願意以這種方式為房屋提供融資，除非他們除了獲得所提供資金部分的房屋增值外，還能收到「租金」。針對這一特定問題的一個提議是，借款人在房屋中居住時間越長，分配給債權人的房屋增值比例就越高。無論如何，席勒和卡普林認為次貸危機為政府提供了一個推動房屋金融創新新潮流的絕佳機會。

Perhaps the biggest question facing lawmakers was whether they needed to overhaul the way they regulated the financial system as a whole. The role of the rating agencies, in particular, was questioned. Congress had passed a bill in 2006 that prevented these agencies from “selling” high ratings by receiving payments that depended on the ratings that they handed out. Congress had to
也許立法者面臨的最大問題是他們是否需要徹底改革整個金融系統的監管方式。特別是評級機構的角色受到了質疑。國會在 2006 年通過了一項法案，禁止這些機構通過根據其給出的評級收取費用來“銷售”高評級。國會必須
decide whether it wished to exercise more oversight over these businesses or impose further restrictions on their conduct. A related issue was how ratings should be used. Rating agencies had been championed in the past as a good alternative to government supervision of financial institutions. Indeed, their role was set to increase under rules proposed by U.S. financial regulators at the end of 2006. ${}^{39}$${}^{39}$^(39){ }^{39} Consistent with a long series of international negotiations often referred to as Basel II, the proposed rules set minimum capital requirements for commercial banks that depended on “external” ratings of bank assets. ${}^2$${}^2$^(2){ }^{2}
決定是否希望對這些企業進行更多監督或對其行為施加進一步限制。一個相關的問題是評級應該如何被使用。評級機構過去被視為政府監管金融機構的一個良好替代方案。事實上，根據美國金融監管機構在 2006 年底提出的規則，其角色預計將會增加。 ${}^{39}$${}^{39}$^(39){ }^{39} 與一系列通常被稱為巴塞爾協議 II 的國際談判一致，擬議的規則為商業銀行設定了依賴於銀行資產“外部”評級的最低資本要求。 ${}^2$${}^2$^(2){ }^{2}

The rating agencies, for their part, resisted intimations that they had acted wrongly during the subprime boom. In testimony before a Senate panel in September 2007, Moody’s and S&P defended their methodology, their institutional structure and their record. As Vickie Tillman of S&P put it, “Our reputation and our track record are the core of our business.” While the agencies had been criticized for the large proportion of AAA ratings in MBSs, they argued that the historical default rates on highly rated MBSs were lower than those on equally rated corporate bonds. In response to complaints about the conflicts inherent in having issuers pay for ratings, they noted that the committees that actually assigned ratings had no role in negotiating fees with issuers. They also emphasized the benefits of giving all investors free access to ratings, including the additional oversight that this provided. Lastly, they explained the wave of downgrades in the second half of 2007 as being the result of unexpectedly poor performance of subprime mortgages written in 2006. As to their explanation for why this performance changed, Vicki Tillman testified that S&P was “still gathering data to analyze the causes for these inconsistent market dynamics.” In his testimony, Michael Kanef of Moody’s attributed this deterioration to laxer underwriting standards, to declines in house prices and to more restrictive lending standards that prevented refinancing by borrowers. ${}^{40}$${}^{40}$^(40){ }^{40}
評級機構方面，對於在次貸繁榮期間被暗示行為不當的指控表示抗拒。2007 年 9 月，在參議院小組委員會的證詞中，穆迪和標準普爾為其方法論、機構結構及其紀錄進行辯護。正如標準普爾的 Vickie Tillman 所言，「我們的聲譽和業績記錄是我們業務的核心。」儘管評級機構因 MBS 中 AAA 評級比例過高而受到批評，但他們辯稱，高評級 MBS 的歷史違約率低於同等評級的企業債券。針對關於由發行人支付評級費用所固有的利益衝突的抱怨，他們指出，實際分配評級的委員會並不參與與發行人協商費用的過程。他們還強調向所有投資者免費提供評級的好處，包括由此帶來的額外監督。最後，他們解釋 2007 年下半年評級大幅下調，是由於 2006 年簽訂的次貸抵押貸款表現出乎意料地差所致。 至於他們對於這種表現變化的解釋，Vicki Tillman 證稱標準普爾「仍在收集數據以分析這些不一致市場動態的原因」。穆迪的 Michael Kanef 在證詞中將這種惡化歸因於寬鬆的承保標準、房價下跌以及更嚴格的貸款標準，這些標準阻止了借款人進行再融資。

Basel II also allowed banks to use their internal risk-management models to determine their minimum capital. In the case of U.S. investment banks, the SEC codified model-based capital requirements in line with Basel II in 2004. In mid 2007, the SEC allowed Goldman Sachs, Morgan Stanley, Merrill Lynch, Lehman Brothers and Bear Stearns to use these new rules and thereby reduce the amount of capital that they needed to keep in safe assets. A burning question for regulators in early 2008 was thus how to strengthen the supervision of investment banks while taking into account the complexity of their assets and the rapid change in their market values. One proposed solution was to limit the range of securities that highly leveraged firms could hold. Another was to reduce the role of market prices in their regulation on the ground that the prices of financial claims behaved anomalously during crises. ${}^{41}$${}^{41}$^(41){ }^{41}
巴塞爾協議 II 也允許銀行使用其內部風險管理模型來決定最低資本。在美國投資銀行的情況下，證券交易委員會（SEC）於 2004 年將基於模型的資本要求納入法規，與巴塞爾協議 II 保持一致。2007 年中，SEC 允許高盛、摩根士丹利、美林證券、雷曼兄弟和貝爾斯登使用這些新規則，從而減少他們需要保留在安全資產中的資本數量。2008 年初，監管機構面臨的一個燃眉之急是如何在考慮投資銀行資產複雜性及其市場價值快速變動的情況下，加強對投資銀行的監管。一個提議的解決方案是限制高槓桿公司可持有的證券範圍。另一個方案則是減少市場價格在監管中的作用，理由是金融債權的價格在危機期間表現異常。

An even broader issue was how the government should react to the rapid pace of innovation in the financial services industry. This industry had excelled at creating new institutions such as conduits, CDOs, hedge funds and SIVs. Was it necessary to supervise and regulate newly created institutions? Even if these were left alone, did the regulation of banks with government-insured deposits have to change in light of the opportunities offered by new institutions?
一個更廣泛的問題是政府應如何應對金融服務業快速創新的步伐。該行業在創建新機構方面表現出色，如通道、CDO、對沖基金和 SIV。是否有必要監督和規範新創建的機構？即使這些機構不受干預，鑑於新機構帶來的機會，對擁有政府保險存款的銀行的監管是否也必須改變？

Sources: ${}^{\text{a}}$${}^{\text{a }}$^("a "){ }^{\text {a }} Office of Federal Housing Enterprise Oversight (OFHEO); ${}^{\text{b}}$${}^{\text{b }}$^("b "){ }^{\text {b }} S&P/Case-Shiller. (Both the OFHEO and S&P/CaseShiller indices are based on a method proposed by Karl Case and Robert Shiller. This method uses exclusively data on repeat sales to compute how much home prices have changed from their past levels. The indices differ in details such as the sample they use and the weights they use to compute an overall index.); ${}^{\text{c Bureau of Economic Analysis;}}{}^{\text{d Consumer Price Index}}$${}^{\text{c Bureau of Economic Analysis; }}{}^{\text{d Consumer Price Index }}$^("c Bureau of Economic Analysis; ")^("d Consumer Price Index "){ }^{\text {c Bureau of Economic Analysis; }}{ }^{\text {d Consumer Price Index }} Survey; eWest Texas Intermediate; f US Bureau of the Census.
資料來源： ${}^{\text{a}}$${}^{\text{a }}$^("a "){ }^{\text {a }} 聯邦住宅企業監督辦公室（OFHEO）； ${}^{\text{b}}$${}^{\text{b }}$^("b "){ }^{\text {b }} S&P/Case-Shiller。（OFHEO 與 S&P/Case-Shiller 指數均基於 Karl Case 與 Robert Shiller 提出的方法。該方法僅使用重複銷售數據來計算房價相較過去水平的變化。兩指數在樣本選擇及計算整體指數的權重等細節上有所不同。）； ${}^{\text{c Bureau of Economic Analysis;}}{}^{\text{d Consumer Price Index}}$${}^{\text{c Bureau of Economic Analysis; }}{}^{\text{d Consumer Price Index }}$^("c Bureau of Economic Analysis; ")^("d Consumer Price Index "){ }^{\text {c Bureau of Economic Analysis; }}{ }^{\text {d Consumer Price Index }} 調查；e 西德克薩斯中質原油；f 美國人口普查局。

Exhibit 2 The Federal Funds rate and some of its determinants
圖表 2 聯邦基金利率及其部分決定因素

Source: FRED at Federal Reserve Bank of St. Louis
資料來源：聖路易斯聯邦儲備銀行 FRED 資料庫
Exhibit 3 Total foreclosure-related legal filings from RealtyTrac
展示 3 來自 RealtyTrac 的總止贖相關法律申請

Source: Compiled by casewriter with data from RealtyTrac. These filings include the “Notice of Default”, “Lis Pendens” (also a legal notice to borrowers), “Notice of Foreclosure Sale”, “Notice of Trustee Sale”, and “REOs”, which are instances where a bank repurchases its foreclosed property.
資料來源：由案例撰寫者彙整，資料來自 RealtyTrac。這些申報包括「違約通知」、「訴訟留置權通知」（也是對借款人的法律通知）、 「止贖銷售通知」、「受託人銷售通知」及「銀行持有房產」（REOs），後者指銀行回購其止贖房產的情況。

Exhibit 4 Loans on which a servicer is starting foreclosure proceedings in the third quarter of 2007
附件 4 2007 年第三季開始進行止贖程序的貸款案件數量
(\left.\begin{array}{lcc} & & Foreclosure initiations
(\left.\begin{array}{lcc} & & 止贖啟動

Source: Compiled by casewriter with data from Mortgage Bankers Association.
資料來源：由案例撰寫者彙整，資料來自抵押貸款銀行家協會。

Exhibit 5 Homeownership rates by race
附表 5 按種族劃分的擁有住房率

Exhibit 6 Writedowns by major financial institutions announced as of 12/31/07 ($ billions)
附表 6 截至 2007 年 12 月 31 日主要金融機構宣布的減記金額（十億美元）