Rating Agencies and Conflicts of Interest
Rating Agencies and Conflicts of Interest
Rating agencies face a fundamental structural conflict: issuers pay for ratings, creating an incentive for leniency. A bank issuing $1 billion in bonds shops the rating between Moody's, S&P, and Fitch. The agency that provides the most favorable rating gets hired. All three know this, and all three have revenue pressure to win business. The result is a system where competition drives leniency, not rigor.
Key takeaways
- Issuer-pays funding creates inherent conflict: issuers shop ratings, agencies compete to win business by being lenient
- Government-pays and investor-pays alternatives exist but have administrative and economic drawbacks
- Internal email evidence from 2004-2007 showed agencies rating MBS despite knowing models were flawed
- Post-2008 reforms (Dodd-Frank, SEC oversight) increased transparency and back-testing requirements but didn't eliminate incentives
- Investors must treat agency ratings as one input, not a substitute for independent analysis
The issuer-pays model mechanics
When a corporation issues $2 billion in debt, it hires a rating agency. The corporation pays $500,000 to $2 million depending on complexity. This creates two problems:
First, the issuer controls which agencies to hire. Typically, two or three agencies rate an issuance. If all rate the bond BBB, the market takes that seriously. If one rates it BBB and another rates it B, the market discounts the lenient rater.
Second, rating agencies earn recurring revenue from issuers through "surveillance fees." After the initial rating, the agency charges 10-20 percent of the initial fee annually to maintain and update the rating. This ongoing fee stream creates an incentive not to downgrade — downgrade a major client and you lose not just the initial fee but the stream of surveillance revenue.
A client paying $500,000 upfront and $50,000 annually is generating $500,000+ in total revenue over 10 years if no downgrade occurs. Downgrade to junk status and surveillance revenue often declines as the market reprices the bond. There's no economic gain to rigorous analysis — there's economic pain.
Historical evidence of conflicts: internal agency email
After 2008, regulators obtained internal emails from Moody's, S&P, and Fitch. These revealed explicit awareness of model problems and conflicts:
Moody's VP, 2004: "It could be structured by cows and we would rate it." The comment was dark humor reflecting the felt pressure to rate securitizations without adequate rigor.
S&P analyst, 2006: "This is a very simple and straightforward MBS with [stated income borrowers] and we are rating it as AAA. I have concerns but can't say them on email."
These weren't isolated comments. Dozens of similar emails appeared, showing that analysts knew standards were slipping but compensation and business logic pushed toward lenient ratings.
Moody's email from 2007: "We are rating mortgage securities that should be rated BBB, but we rate them AAA because we've been hired to do so and competitors will do the same."
The transparency created by regulatory discovery made clear that the conflict was structural, not incidental.
Alternatives to issuer-pays: comparison and tradeoffs
Three alternative models exist. None has gained full adoption:
Government-pays model: The SEC or a central bank employs rating analysts and publishes ratings for all securities. This eliminates issuer conflicts. But it's expensive (estimate: $100-200 million annually for comprehensive coverage), and it's politically contentious (who sets rating standards?). The SEC maintains a "Big Three" NRSRO system partly to avoid this, making government ratings a backstop rather than primary source.
Investor-pays model: Mutual funds, pension funds, or index providers pay agencies to rate. This aligns incentives with investor interests. But it has major practical problems. An investor paying for ratings has incentive to seek ratings that justify their portfolio decisions. The rating agency becomes a provider of rationalization, not an objective analyst. Additionally, rating availability declines — indices and funds only rate securities they own or plan to own, leaving many bonds unrated.
Issuer-pays with separation: The issuer pays for ratings, but the rating agency doesn't know which issuer paid for which rating until after publication. This solves the conflict at the cost of administrative complexity and opacity. Rarely used because issuers want to know their rating before market announcement.
By default, the market chose issuer-pays. It's simple, it scales, and it's profitable. The conflicts are accepted as a cost of the system.
Post-2008 reforms: Dodd-Frank and SEC oversight
The Dodd-Frank Act (2010) mandated several reforms:
-
Quarterly back-testing: Agencies must compare historical ratings against actual default rates. If a category rates AAA, it must default at rates consistent with AAA-level risk. Systematic divergence requires methodology review.
-
Methodology disclosure: Agencies must publish how they rate, what variables matter, and how they've changed over time. Previously, methodologies were proprietary black boxes. Transparency reduces ability to shop ratings.
-
Analyst separation: The team rating a security cannot communicate directly with the issuer's investor-relations team. Reduces influence.
-
Compensation restructuring: Analyst bonuses can't be tied directly to issuance volume. This reduces incentive to rate leniently to win business. But it doesn't eliminate it — firm revenue still depends on rating volume.
-
SEC supervision of NRSRO: The SEC now examines rating agencies' compliance with these standards. Agencies can lose NRSRO status (the legal permission to rate) if they fail audits.
These reforms added friction and transparency. But they didn't eliminate the fundamental conflict. An issuer still chooses which agencies to hire. An agency still earns surveillance revenue. The conflict remains — it's just less extreme than 2004-2007.
Do reforms work? Evidence from post-reform ratings
Post-2008, rating agencies have been more conservative. AAA ratings are rarer and demand higher credit quality. BBB-rated corporates are more likely to default than they were in 2000-2005, reflecting that the standard tightened.
However, the conflict re-emerges in cycles. When credit is abundant and issuance high (2017-2019, 2022-2023), rating standards drift. Agencies feel pressure to rate new categories of debt (like CLOs and leveraged loans) despite methodological uncertainty. As they gain experience, standards tighten, and later cohorts get harsher ratings.
A 2019 study by the Financial Industry Regulatory Authority found that bonds rated BBB in high-volume years (2006, 2017) had higher default rates than bonds rated BBB in low-volume years (2009, 2012). The correlation was modest but measurable — competitive pressure to rate leniently pushes during bull markets.
Implicit regulatory reliance despite known conflicts
The irony is that regulators perpetuate the issuer-pays model despite knowing its conflicts. Why? Because alternatives are expensive or politically difficult.
Banks maintain capital ratios based partly on bond ratings. If a bank owns $100 million in BBB bonds, it holds capital based on the BBB default assumption. If ratings become untrustworthy, regulators must impose higher capital requirements regardless of rating.
Insurance companies reserve for defaults based on ratings. Pension funds are mandated to hold investment-grade (BBB+) bonds. These regulations create artificial demand for high ratings.
If regulators moved to government-provided ratings or eliminated the special status of NRSRO ratings in capital requirements, these mandates would evaporate. Bond supply would shift (more unrated debt), and regulatory burden would increase (regulators would have to do the analysis instead of outsourcing it).
It's politically easier to accept known conflicts and oversight than to rebuild the entire infrastructure.
How investors should treat agency ratings
Given these conflicts, what's the rational investor posture?
First, treat ratings as one input, not a substitute for analysis. A AAA rating tells you the issuer is very strong, but it doesn't tell you valuation or near-term risk. Enron had investment-grade ratings days before bankruptcy.
Second, pay attention to rating momentum. An upgrade or downgrade matters more than the current rating — it signals that the agency has changed its view of the credit. But be suspicious of upgrades in bull markets (likely driven by improved sentiment, not credit) and downgrades in bear markets.
Third, use spreads as a cross-check. If BBB-rated corporate bonds yield 2 percent while AAA-rated corporates yield 1.5 percent, the spread (50 basis points) seems too tight for the rating jump. This suggests the market is pricing BBB bonds more cautiously than the rating implies. The market pricing is often correct.
Fourth, compare credit metrics to ratings. A company with 5x leverage, 30 percent margins, and stable cash flow might be rated BBB. If you can point to three similar companies rated BB+, the rating is likely lenient. If you can point to companies rated BB with 6x leverage and declining cash flow, the rating is appropriately strict.
The persistence of conflicts despite reforms
Next
The conflicts in rating agencies highlight why ratings alone are insufficient. Investors need to understand what spreads tell us — when the market prices credit risk differently from the agencies. The next article explores implied ratings derived from spreads, and how bond-market pricing corrects rating agency blindness.