Supervisors on Supervision

“Culture is inherently less tangible than, say, capital ratios or liquidity buffers — so we used substitute words like shared values, beliefs, and behaviors. That makes the work inherently qualitative and, to some extent, subjective.

 If different supervisory teams define culture differently, they will look for different things. 

When I was at MAS, there was no prescriptive definition of ‘good culture’ that could be applied equally across all institutions — given their very different business models, sizes, and operating environments. What’s meaningful for a small community bank in the U.S. may not map to a global institution.

Even if we could have designed a thoughtful methodology, firms understandably perceived it as subjective, which generated resistance — or at least limited buy-in. That, in turn, could undermine the effectiveness of supervisory recommendations. 

So, how do we assess culture without a one-size-fits-all worksheet?”

“Culture isn’t a ready metric; judgment is involved even when anchored to evidence. Conversations about ‘tone from the top’ and ‘walking the talk’ quickly prompt precise questions from boards: ‘What exactly is missing — and why?’ 

We found it easier to embed culture into tangible things: remuneration and performance management systems, and how those might drive good or bad behavior — citing a few incidents without overgeneralizing (especially in banks with large footprints). We relied heavily on audit (internal and external). How management treats auditors tells you a lot. 

So we bundled multiple strands — governance structures, design, incentives — rather than saying ‘your culture is good/bad’ in isolation. A holistic embedded approach worked better than treating culture as a standalone rating. 

What we wanted [at MAS at that time] was something sharper and more consistent — a kind of quality control — so ‘good/bad culture’ judgments weren’t just personal impressions. 

It is important for supervisors to be able to state clearly and to defend their findings or observations with respect to their assessment of a firm’s culture. Board members in financial institutions take their responsibilities and fiduciary duties seriously. They will understandably seek to understand any supervisory concerns — and sometimes, to even challenge or offer other perspectives. So explainability by a supervisor of its findings is important.”

“Think of the supervisory toolkit: upstream are regulations, then supervision, then supervisory actions, culminating in enforcement. We reviewed our enforcement actions and asked: why do some firms fall afoul more often than others — even among peers of similar size in similar environments? Is it about the quality or competence of people? Or is there something about a firm’s culture that incentivizes or drives certain behaviors? 

It is my view that bad culture can be a key contributor to both conduct and prudential risks.”

“Regulators should avoid supervisory assessments based on narrow or limited data points. For instance, over reliance on just employee survey results could lead to subsequent curation by institutions to demonstrate desired outcomes. Regulators should instead rely on a broad set of culture-related observations and indicators.”

“There’s also the question of how involved supervisors should be. Extensive board-level interviews, for example, are resource-intensive for already stretched teams. We needed pragmatic approaches that add value, even if imperfect.

Since leaving my regulatory role, I’m more conscious of another boundary question: where’s the line between effective supervision of culture and regulatory overreach into internal management — where boards and executives should be held responsible? We don’t want to micromanage or dictate specific cultural elements. Principles alone can become too vague; prescriptive rules can devolve into box-ticking. We need to manage that balance.”

“Supervisors must move beyond treating culture as a ‘soft’ topic and link it to hard elements: governance structures, risk-management design and implementation, incentives. We need supervisors who can connect soft and hard qualities.

Culture risk assessment should be about patterns and indicators, not trying to establish strict cause-and-effect. Because culture isn’t physical. But we still need tangible indicators — behaviors, decisions, outcomes. 

Culture should feed into a broader supervisory ‘forensic’ analysis of institutional health and risk culture robustness. For examples: whistleblowing volumes and themes; frequency and type of risk indicators; complaint patterns; speed and quality of remediation; and whether employees feel safe to speak up. We need to appeal to a combination of soft and hard evidence.”

“On the prudential side, MAS ran its annual comprehensive risk assessment of financial institutions — what we called the ‘CRAFT’ assessment. We’d engage boards with supervisory ratings across AML, conduct, risk management, capital, and so on.

Peer institutions often produced different ratings. Why did some fare better or worse than others? Was there a link to the incidence of enforcement? Not always directly — but the propensity to ‘do the wrong thing’ seemed more likely in some places. 

So we asked: what distinguishes a ‘good’ from a ‘not-so-good’ bank? Leadership? Vertical communication? That led us to culture — the organization’s attitudes toward risk, misconduct, fraud, and whether managers pursue KPIs at the expense of other priorities. All the conduct related issues — balanced scorecards, incentives — were attempts to shape behavior and steer culture at both management and working levels. 

Looking back, we were trying to pin what we couldn’t otherwise explain on ‘poor culture,’ without fully defining it. Still, we decided supervisors should engage bank boards on our sense of their culture and risk culture. We had an internal document on supervising for culture — but theory and practice diverge.”

“If enforcement action is based on ‘poor culture,’ is it legally challengeable? Breaching a capital or AML rule is clear. But an action predicated on culture relies on circumstantial evidence and patterns — building a robust case is harder. 

Focusing on board and senior management accountability is key in large, complex organizations where responsibility can be diffuse.

Finally, there are jurisdictional differences — legal frameworks and corporate-governance approaches vary. Achieving broad consistency for global institutions would help.”

“When I was at MAS, we made concerted attempts to supervise for culture — especially in key financial institutions. For us, it was about understanding the human element of risk. We focused on observable behaviors and then decided how (or whether) to intervene, with the goal that culture would function as a line of defense rather than a hidden vulnerability in the financial system. 

That’s also why we brought our Chief Data Officer into the picture — to think about how data and analytics could inform assessment — and we even recruited a behavioral expert from a university to join some supervisory site visits. The idea was to get better at picking up signals.”

“MAS is both conduct and prudential regulator, and also the central bank and AML regulator. During my tenure, conduct issues — behaviors, incentives, remuneration — were constant topics in joint supervisory meetings. We asked whether lessons from conduct could inform prudential reviews.

We also had incidents — our SIBOR-fixing episode (our analogue to LIBOR). Were those traders acting individually, or was there a deeper cultural driver? Why this group, these institutions, and not others? That pushed culture into prudential work. 

At that time, our board would ask: what is the culture at Bank X versus Bank Y? Are they cowboys? Are they careful? Do they try to do right, or only chase the bottom line? As supervisors in and out of institutions, we were expected to have a view. We needed a common ‘song sheet’ internally and when engaging firms. That prompted more structured thinking. 

We had the ‘luxury’ of being conduct regulator, prudential regulator, and central bank — with autonomy — so we could make the case to our MD/Governor and weave culture into supervisory processes. It resonated with leadership and staff. To be honest, teams felt they were already observing cultural signals informally: ‘helpful’ versus ‘evasive’ managers, etc., and sometimes wrote that into supervisory reports.”

“Across regulators and firms, we’re moving to a more holistic, sophisticated approach that balances qualitative insights with data — skills, governance, autonomy, reporting lines, and legal dynamics. That allows better judgment about collaboration norms, stakeholder engagement, including with regulators. 

ESG has also helped. The ‘E, S, and G’ inherently involve culture — respect, trust, integrity, looking beyond dollars and cents. ESG has accelerated focus on culture. 

In insurance, for example, advisors’ conduct, incentives, and reward systems loom large. My general sense is that focus has stepped up — in firms as much as among supervisors.”

“Supervisory skill sets and resourcing is another challenge. [MAS brought in a] behavioral scientist to help with culture assessment, and the involvement of our Chief Data Officer. Each brings value but may not be ‘fit for purpose’ alone. 

Traditional supervisors are strong in financial analysis, modeling, and compliance. But assessing culture demands different capabilities: behavioral economics, sociology, HR management, organizational psychology. 

Building that into already lean teams is hard. Telling supervisors preoccupied with hard metrics, ‘By the way, look out for culture,’ can be unclear in practice — even with internal guidance. Consistency suffered. 

So what we did was to focus on was skill sets. To assess culture, we asked, do supervisors need training in behavioral economics or organizational psychology so they know what to look for — and can ask insightful questions about incentives, groupthink, and challenge mechanisms? We wanted to be more ‘scientific’ in equipping supervisors to do a better job. So, it was less about our own culture, and more about building the right competencies.”

“Because we lacked quantifiable metrics, it was hard to say — firm to firm — ’this is good culture’ versus ‘this is bad culture’ without some subjectivity, even with ‘science’ behind it. 

We tried to collect data and analyze it — so our judgments wouldn’t feel airy-fairy, dependent on whether the supervisor was ‘benevolent’ or ‘harsh.’ We looked for other observable manifestations: HR data, complaints, incidents, audit findings. It becomes a massive data collection exercise — then an integration and analysis challenge — even inside a single large bank. 

And even then, causality is hard: did a cultural trait cause a prudential or conduct failing, or is the observed culture the intermediate outcome of something else — like a control breakdown or leadership failure? 

Global benchmarking would help. Supervisors need reference points for ‘good’ or ‘bad’ cultural indicators to identify outliers at risk.“