Senior control officers may not all support the need for a dedicated supervision head, but when it comes to most-valued controls, most recognise that the all-seeing supervisor is top of the list.
Effective supervision is at the heart of the front office control function and the revolution in non-financial risk management that has occurred across banking in the last five years or so. A bank failing to establish the right supervisory controls, backed up by the right systems and galvanised by the right people in the right jobs who know their responsibilities, is a bank courting calamitous consequences.
So there is a lot of incentive to get it right, and thus it is, perhaps, surprising that more senior control officers do not think banks need an overall head of supervision. While most agree that there is great value in the co-ordination of effort across an organisation, there doesn’t need to be a straight reporting line to a supervision supremo.
It is generally felt that such a person would be too remote from the business to be of real value, and would lack the knowledge to be able to be of value to senior management. “It would be hard to figure out the balance between a functional head of supervision and a knowledge-based head of a business,” says Todd Sullivan, Head of Risk Management, Fixed Income, Americas, at Morgan Stanley in New York. “If you put someone above the head of equities, or the head of FICC, it is challenging to know what they’d add.”
Machines can’t do it all
There are still a large number of controls on which supervisors have to sign off, but, in fact, it has decreased slightly in the last 12 months according to our poll results. Last year, the average minimum number of controls across all respondents was 14 – this has sunk to nine. Meanwhile, the average maximum number of controls has held around the same level (37 in 2018 versus 36 in 2019).
This still seems a great many, but it is difficult to know which ones can be dispensed with. Neither would regulators be happy to learn than banks are ditching supervisory controls. But, in an age of machine learning and AI, a perhaps surprising and certainly refreshing number of supervisors put their faith in the human element.
“If I had to give up every control but one, the one I would keep is the eyes of the supervisor on the person being supervised,” says Sullivan. “It is the physical proximity, the conversations, the looking over the shoulder, engaging with the risk that person is managing, that is so important. I’d choose that over all the controls in the world.”
At the moment, it seems, automated controls, even those galvanised by machine learning and AI, still do not give as meaningful insight into possible breaches of risk as human interaction does.
Rupert Jolley, Chief Control Officer, Global Banking and Markets at HSBC in London, agrees. “What is the one control that is most valuable? It’s to do with people. At the end of the day, you need intervention by people with the right mindset and the willingness to do the right things by clients and shareholders. That’s the critical control,” he says.
This is not to say that automated controls are not highly valuable; in fact, a supervisor could not do their job without them. Of the many automated controls at the disposal of a supervisor, most vital are basic risk metrics – credit risk, market risk, the balance sheet and the P&L – says Joseph Campione, Chief Control Officer at Bank of New York Mellon.
More time please
The poll result to the question asking how many hours a week a manager spends on supervisory controls makes for surprising reading. Some 57% said that they spend between zero and five hours a week in this manner, and of that number 37% say they spend between two and five hours a week.
Those spending between five and eight hours increased from 5% in last year’s poll to 13%, but that’s still only an eighth. Only one in five said they spent between 12 and 15 hours a week.
At the 1LoD New York conference in April, the unanimous opinion of the four-member panel that discussed supervision was that around 25% of a supervisor’s time should ideally be spent on supervision. This would mean in even a 40-hour week, around 10 hours would go to supervision, but, according to this poll, less than 30% of respondents estimated that supervisors spend this amount of time doing so.
Moreover, of course, it is unlikely that many supervisors enjoy a 40-hour week, so the proportion of time on supervision falls even further.
It could be that respondents interpreted the time spent on controls very strictly – that is to say, they thought only about the time spent managing and responding to the data produced by the controls and not the time spent on more informal supervision.
“The possibility is that they’re not counting times looking over shoulders, and not doing reports per se,” says Sullivan, “but this is still very important human stuff.”