The 2008 back-test data: everything was visible in 2006 and 2007

Was everything visible in public data before the crisis?

The study answers yes, and it states the constraint it set itself: no privileged information and no insider knowledge. Each data point was published before the crisis, from filings, statistical series and annual reports, so a reader with the right sources could assemble the same picture. That standard makes every figure on this page checkable by a reader.

The study groups the evidence into 3 families: the real-estate market, the derivatives market, and the regulatory framework. A fourth family, the warning signals of 2007 and 2008, shows how the market reacted. The sections below follow that order. Each family includes figures that you can find in the sources listed below.

What did the 2006 housing data show about the 2008 crash?

House prices rose 124% from 1997 to 2006. The price-to-income ratio stood at 4.6×, against a historical average of 3.2×, which the study calls 44% above that average. The subprime share of lending reached 23.5%, up from 8% in 2003, so lending to weaker borrowers tripled in 3 years (EQOS Asia 2008 back-test, 2026).

The default rate on subprime loans reached 14.4% at the end of 2006. Defaults at that level were already a signal, because they rose before the loan resets that the study expected next. The study stresses that FDIC and mortgage-lender data carried these figures in public tables. Together the 4 housing figures describe a market where prices, borrower quality and repayment all pointed the same way.

What did the derivatives data show before the 2008 financial crisis?

CDO issuance reached $520 billion in 2006, 10 times the 2001 level. The notional value of credit default swaps stood at $62 trillion according to the BIS, while world GDP was $54 trillion. AIG alone had sold $440 billion of CDS. Also, 80% of CDO tranches carried an AAA rating (EQOS Asia 2008 back-test, 2026). The study notes that the sources are the BIS Quarterly Review, AIG filings and Inside Mortgage Finance.

The study divides the swap volume by world GDP and obtains 1.15. Bets on risk exceeded the value of everything produced in a year, which the study calls structurally incoherent. The ratio alone does not give a date, yet it shows how much of the system depended on one belief. A ratio above 1 means the contracts cover more value than exists, so a payout would exceed the resources behind it.

Which public sources held the 2006 and 2007 crisis data?

The study names its sources for each family. For housing: the S&P Case-Shiller Index, the FDIC Quarterly Banking Profile and Mortgage Bankers Association data. For derivatives: the BIS Quarterly Review, SEC filings such as the AIG 10-K, and Inside Mortgage Finance. For banks: SEC 10-K filings. The study says all of them are public.

Publicly available does not mean easy to find. The sources sit in different formats and agencies, so the work lies in assembling and comparing them. A repeatable checklist of sources, with a date for each pull, makes the assembly quick and the results comparable between years. Note the retrieval date next to each figure.

Which warning signs of the 2008 crash appeared in 2007 and 2008?

The study lists dated events. In February 2007 HSBC announced $10.6 billion of subprime losses. In June 2007 Bear Stearns closed 2 hedge funds after $1.6 billion of losses on subprime CDOs. In August 2007 BNP Paribas froze 3 funds, and the TED spread moved from 20 to 200 basis points within days. The market ignored the first signal, according to the study, and the interbank market seized in August.

In March 2008 Bear Stearns was bought by JP Morgan at $2 a share, against $172 a year earlier. On 15 September 2008 Lehman Brothers filed for bankruptcy with $619 billion of debt. Each event came after the data had already shown the pattern, which is the point of the study. The Federal Reserve arranged a bailout around the Bear Stearns sale, according to the study.

How did regulation changes shape the 2008 financial crisis data?

The study lists decisions that reduced structural safeguards. In 1999 Congress repealed Glass-Steagall, which had separated commercial and investment banking. In 2000 the Commodity Futures Modernization Act left credit default swaps unregulated. In 2004 the SEC loosened leverage limits for the largest banks, according to the study. Those 3 decisions sit in the study’s regulation table.

Each decision is public, voted or documented, and the study reads them as lowering the integrity of the system. A reader can add regulation as a data family in any market: when safeguards are removed while volumes grow, the pressure builds without any visible alarm. Tracking these changes costs almost nothing and gives context for every other number.

How can you read data like this before a stock market crash today?

Build a short dashboard for the market you follow: price against income or earnings, share of lending to weaker borrowers, default trend, size of derivatives against output, and recent regulatory changes. Date each pull. This page is general information, not investment advice, and a dashboard shows pressure, not timing. Keep the same sources each quarter, so a change in a number reflects the market and not a change of definition.

The EQOS Asia method page explains how these inputs feed a 9-dimension score, and the Λ simulator lets you try it on your own case. The parent page on whether a model would have predicted the crash, and the pillar guide on the stock market crash, give the wider context.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile