Stripping ‘M’ From CAMELS Undermines Risk Oversight In Banks 

A proposal in the United States to revise the CAMELS rating system — the supervisory tool used by US federal regulators to assess the health and safety of financial institutions — risks diluting one of its most vital components: the quality of management. The so-called HUMPS Act of 2025 aims to curb subjectivity in ratings by eliminating or restricting the "M" in CAMELS — an acronym for Capital adequacy, Asset quality, Management, Earnings, Liquidity, and Sensitivity to market risk. The bill argues that supervisory judgement of management is inherently subjective and lacks transparency. But scrapping it could strip bank supervision of a critical lens: organisational culture.

In the current CAMELS framework, the 'M' may sit alphabetically at the centre, but functionally it is the fulcrum. Management quality influences asset quality, earnings, and even liquidity. The board and management may have different roles — the former sets strategic direction and the latter executes it — but both are assessed together as part of the "M" rating. Eliminating this component might address the issue of examiner discretion, but it risks discarding the very capacity to detect cultural and governance failures.

The controversy over the "M" score has been fuelled by a lack of transparency. In the aftermath of rating leaks from the Office of the Comptroller of the Currency, it emerged that several banks had been rated poorly on "M" despite maintaining strong capital buffers and liquidity metrics. No justification was offered publicly for these scores, reinforcing the perception of opacity. The matter escalated when supervisory authorities treated the leak itself as evidence of governance failure within the affected institutions, which led to further downgrades. This asymmetric accountability has eroded trust in the legitimacy of the supervisory process.
Quantifying Culture

The rationale offered by HUMPS supporters is that poor "M" scores, often based on vaguely defined notions of effectiveness, can trigger severe consequences for banks despite strong fundamentals. While the critique is valid, the answer is not to remove "M" altogether but to improve how it is evaluated. Supervisory subjectivity should not be replaced by blind objectivity. Instead, it should evolve into evidence-based judgement.

In fact, the supervisory process would gain from integrating behavioural analytics and artificial intelligence to assess culture in real time. These tools can spot early signals in leadership conduct, decision-making processes, and internal communications that may not show up in traditional metrics. Rather than being a cost centre, governance frameworks like GRACE (Governance, Risk management, Assurance standards, Conduct risks and Ethical standards) need to be viewed as a vital input for sustainable growth.
The prevailing mindset that views GRACE as inhibiting growth and innovation is flawed. On the contrary, high-quality governance acts as the foundation for long-term performance. A paradigm shift is needed: supervisors must start quantifying the contribution of good governance to profitability. Counterfactual analysis can help. If GRACE standards were missing, would growth and profitability have been sustained beyond a few good quarters?

Enabling Accountability
The issue is not just about measurement but also about accountability. The leaks of "M" scores for US banks exposed supervisory opacity, sparking criticism over regulators' unwillingness to disclose how ratings are determined. If poor scores can limit a bank’s business operations, then regulators must be transparent about their basis. As Fed Governor Michelle Bowman has argued, accountability must apply as much to supervisors as to the supervised. However, “it can be difficult to establish a regulatory culture that includes mechanisms to promote accountability for supervisors and regulators.” Similarly, while she had also argued in favour of some degree of confidentiality in the course of conducting supervision, interest in confidentiality “should not be used to prevent scrutiny and accountability in the assignment of [supervisory] ratings.” 

Supervisory failures are often revealed too late. But when they are, it should be possible to examine the track record of management responsiveness — speed of compliance, transparency, and internal adaptation. These are valid indicators of risk culture. The Indian model of Senior Supervisory Managers, who oversee individual banks on continuous basis from a distance, offers a potential blueprint for improving judgement with proximity.

Other sources of insight can be institutional records. Board minutes should reflect not just compliance discussions but the quality of debate on strategy, risk, and ethics. The nature of whistle-blower complaints and their resolution process can reveal organisational integrity. Relationships among senior management and between CEOs and board members, including independent directors, can influence outcomes in ways that pure financial metrics cannot capture.

A modern rating framework must strike a balance. It cannot rely solely on qualitative impressions nor ignore them in favour of data points. It must identify ways to standardise the evaluation of "M" through indicators like board minutes, whistle-blower mechanisms, executive relationships, and past response patterns. When combined with stress period analytics and case study benchmarks, these can offer a more grounded view of management quality.

The same applies to predictive tools. Deviations in behaviour under stress, inconsistencies between different lines of defence, and misalignment between internal audit findings and senior responses are all signs of deeper cultural weaknesses. These insights, when triangulated with supervisory data and past case studies, can enhance the legitimacy of the "M" score. Though it may still be contestable, it would be defensible.

Above all, culture is not a peripheral issue. The root cause of breakdowns at Wells Fargo, Credit Suisse, and Silicon Valley Bank were not traced to reduced  capital or liquidity — they were essentially failures of culture. If the supervisory process cannot account for that, then it risks failing its purpose.

It is time to move beyond the binary of discretion versus data. The quality of management must remain central to supervision, not discarded in pursuit of artificial objectivity. The challenge is to make "M" measurable — not invisible.