FRM Exam Part I · Anatomy of the Great Financial Crisis of 2007-2009

Credit Rating Agencies and Structured Product Failures in the 2007-2009 Crisis

Updated 11 October 2026 · Fact-checked

Rating agencies gave top ratings to many mortgage-backed securities and CDOs because their models assumed low default correlation and rising house prices. The issuer-pays model created conflicts of interest, and investors and regulators over-relied on ratings. When house prices fell, losses were correlated and AAA tranches were downgraded sharply.

Understand Credit Rating Agencies and Structured Product Failures

A credit rating is an opinion on the likelihood of default (and sometimes loss given default) of a borrower or security. Investors, banks and regulators used ratings as a shortcut. Many mandates, and capital rules, were tied to ratings, so a high rating meant cheaper capital and a wider buyer base.

Structured products pool assets such as subprime mortgages and slice the cash flows into tranches by seniority. Losses hit the junior (equity) tranche first, then mezzanine, then senior. A CDO repeated this: it pooled mezzanine tranches of mortgage securities and re-tranched them. Agencies rated the senior tranches AAA even though the underlying pool was mostly BBB-type risk.

Why did the AAA hold up on paper? The models relied on low assumed default correlation across mortgages and on limited historical data, with house prices that had rarely fallen nationally. With low correlation, a pool looks diversified and the senior tranche looks safe. In reality, defaults were driven by a common factor, falling house prices and weak underwriting. Correlation rose, and senior tranches suffered losses. Re-securitization made this worse because mezzanine tranches are very sensitive to correlation and to small errors in the assumed default probability.

The incentive problem also mattered. Under the issuer-pays model, the party that wanted the rating paid for it. Agencies also earned large fees from structuring advice and repeat business. Issuers could rating shop, meaning they asked several agencies and chose the best result. This does not prove agencies knowingly misrated. It shows the pressures that biased judgement and reduced scrutiny.

The final failure was over-reliance. Investors skipped their own due diligence. Regulators hard-wired ratings into capital rules. Originators of loans passed risk on (originate-to-distribute), so they cared less about loan quality. When ratings were cut en masse, prices collapsed, forced selling followed and confidence in the whole market fell. A single ratings scale for corporate bonds and structured products also misled, because structured products have much higher downgrade risk and cliff-like losses.

Key formulas to remember

Tranche loss rule: Tranche loss = min(max(Pool loss − Attachment point, 0), Detachment point − Attachment point); Losses are measured as a share of the pool. Divide by the tranche width to get the loss as a percentage of that tranche.
Tranche loss as % of tranche: Loss % = min(max(L − A, 0), D − A) ÷ (D − A); L = pool loss, A = attachment point, D = detachment point. Thin tranches lose value quickly once L passes A.
Subordination (credit enhancement): Subordination for a tranche = Attachment point; It is the share of pool losses absorbed by junior tranches before this tranche loses anything.
Correlation effect (rule of thumb): Higher default correlation → fatter tail of pool losses → senior tranche riskier, equity tranche often less risky; State this as a general pattern, not a law. It explains why low assumed correlation overrated senior and CDO tranches.

How to solve Credit Rating Agencies and Structured Product Failures questions

Use this method for any question on rating agency failures or structured product ratings.

1Identify what is being asked: cause of the rating error, conflict of interest, a tranche loss calculation, or a lesson or reform.
2If it is a calculation, write the attachment and detachment points and the pool loss, then apply the tranche loss rule.
3If it is conceptual, sort the cause into one of four buckets: model assumptions (correlation, house prices, limited data), incentives (issuer pays, rating shopping, fees), re-securitization (CDO of mezzanine tranches), or over-reliance (investors, regulators, capital rules).
4Match the cause to its consequence: low correlation gives overrated senior tranches, issuer pays gives biased ratings, over-reliance gives no independent due diligence.
5Check each answer option for overstatement, such as saying agencies deliberately committed fraud or that ratings had no information.
6Pick the option that names the specific mechanism, not a vague statement like 'markets were irrational'.

Quickest way: Four-bucket cause sort

When to use it: Use for conceptual multiple-choice questions where you must find the main reason for a rating or structured product failure.

Read the stem and ask: is the issue the model, the incentive, the structure or the user of ratings?
Model: look for correlation, house price assumption, thin data.
Incentive: look for issuer pays, rating shopping, consulting fees.
Structure or use: look for re-securitization, thin tranches, hard-wired ratings, no due diligence.
Eliminate options that are extreme ('always', 'only', 'intentionally').

Common mistakes in Credit Rating Agencies and Structured Product Failures

Saying agencies rated CDOs AAA only because they were paid to.
The issuer-pays story is the most famous one, so students treat it as the sole cause.
Fix: Give model failure (low correlation, rising house price assumption) and incentives together. Both contributed.
Thinking the senior tranche of a CDO has the same risk as an AAA corporate bond.
The same letter grade looks like the same risk.
Fix: Remember structured ratings were far more sensitive to correlation and systematic shocks, so downgrades were sudden and severe.
Claiming diversification across thousands of mortgages made the pool safe.
Diversification works when risks are independent, and students forget the common driver.
Fix: Link losses to a common factor, national house prices and underwriting standards. Correlation rises in stress.
Miscalculating tranche loss by dividing pool loss by tranche size without subtracting the attachment point.
Students forget losses must first exhaust the junior tranches.
Fix: Always compute min(max(L − A, 0), D − A) first, then divide by D − A.
Blaming only the rating agencies.
Questions about agencies prime a one-sided answer.
Fix: Include originators, investors who skipped due diligence, and regulators who tied capital to ratings.

Worked examples

Example 1

A securitization pool suffers a 9% loss. A mezzanine tranche attaches at 4% and detaches at 10%. What percentage of the mezzanine tranche is lost?

Show the solution

Pool loss L = 9%, attachment A = 4%, detachment D = 10%.
Loss absorbed by the tranche = min(max(9% − 4%, 0), 10% − 4%) = min(5%, 6%) = 5% of the pool.
Tranche width = 10% − 4% = 6%.
Loss as % of tranche = 5% ÷ 6% = 83.3%.

Answer: About 83.3% of the mezzanine tranche is lost.

Example 2

Which is the best explanation of why many senior CDO tranches rated AAA suffered large losses in 2007-2009? A) Agencies deliberately ignored all default data. B) Models assumed low default correlation and continued house price growth, so pool losses were underestimated. C) Senior tranches have no subordination. D) CDO ratings were set by regulators.

Show the solution

Test A: it overstates. The issue was assumptions and incentives, not ignoring all data.
Test C: false. Senior tranches have the most subordination, which is why they were rated high.
Test D: false. Agencies, paid by issuers, set the ratings.
Test B: it names the model failure. Low correlation made pools look diversified, and falling house prices produced correlated defaults that reached senior tranches.

Answer: B

Exam tips

Expect conceptual questions that ask for the main cause. Pick the answer that names a specific mechanism such as correlation or issuer pays.
Know the tranche loss calculation. It is quick marks if you subtract the attachment point and divide by the tranche width.
Watch for overstated options. Words like 'solely' or 'deliberately' are usually wrong.
Link this topic to originate-to-distribute and securitization. Questions often combine them.
Remember that ratings were embedded in regulation and mandates, so downgrades forced selling.

Practice questions from Anatomy of the Great Financial Crisis of 2007-2009

Credit Rating Agencies and Structured Product Failures in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

FRM Part IIValuation and Rating of Structured Products

Credit Rating Agencies and Structured Product Failures: frequently asked questions

Why were CDOs rated AAA before the crisis?

Agency models assumed low default correlation among the underlying mortgage securities and relied on short data histories with rising house prices. Subordination and tranching made senior tranches look very safe under those assumptions. The assumptions failed when house prices fell and defaults became highly correlated.

What is the issuer-pays conflict of interest?

The issuer that wants a rating pays the agency for it, so the agency may be tempted to give favorable ratings to keep the business. Issuers could also shop among agencies for the best rating. This is a conflict of incentives, not proof of deliberate misconduct.

How should I explain the rating failures in an FRM answer?

Give the causes in order: flawed models and correlation assumptions, conflicts of interest, complexity from re-securitization, and over-reliance by investors and regulators. Then name the consequence: sudden mass downgrades and loss of confidence in structured products.

Did rating agencies cause the crisis alone?

No. Originators with weak underwriting, investors who did not do their own analysis, and regulators who tied capital to ratings all played a part. Ratings were one link in a longer chain.