The incentive to fly safely

Wayne Byres, Chairman - Regulatory address at the AFR Banking & Wealth Summit 2018, Sydney

Good morning everyone. I’ve had the pleasure of being at each of the past AFR Banking & Wealth Summits, and they have always generated interesting and thought-provoking discussions. As the first speaker this morning, I will do my best to get this year’s proceedings off to a similar start.

Flying the plane safely

My remarks this morning will focus on incentives and accountability in the financial sector: a topic that has attracted quite a deal of attention recently, including in the proceedings of the Royal Commission. Much of that attention is on the fair treatment of customers, but APRA’s interest is from a prudential perspective. I’d like to explain that, borrowing an analogy I’ve used previously, by talking about airline safety.

Flying, like finance, is a fundamentally risky business. As a result, extensive air safety rules and regulations exist to ensure that the incentives that airlines have to minimise costs and maximise profits don’t jeopardise passenger safety. Although regulation imposes costs, and there are sadly still occasional tragedies, on the whole air travel is both the cheapest and the safest that it’s ever been.

Financial regulation is designed to make financial institutions safer, just as aviation regulation is designed to make flying safer. But I envy aviation regulators for one advantage they have over financial regulators: that the desire passengers of planes have for a safe flight is highly aligned with those flying the plane. Pilots have no incentive to take off in an unsound plane. Pilots have no incentive to perform manoeuvres that stress the plane beyond its limits in pursuit of short-term thrills. Pilots, like everyone else on board, have no incentive to do anything other than land safely.

The business of flying might be risky, but pilots have strong personal incentives to minimise the risk. The same alignment of interests, and the natural incentive towards risk aversion, doesn’t exist in the business of finance. A key lesson from the financial crisis was that prudential supervisors need to take a much greater interest in the attitudes towards risk-taking – the risk culture – within the institutions they oversee. What has become clear from this work is that a key driver of risk culture are the formal and informal incentives that individuals face within their organisations, and the accountability (or lack thereof) shown when outcomes are not what they should be.

That recognition is the foundation for APRA’s increasing focus on risk culture, something I have spoken about repeatedly since taking up my role as Chairman. Amongst other things, in 2014 we established a new requirement in our risk management prudential standard (CPS220) – highly contentious at the time – that Boards needed to form a view of the risk culture within their institutions, and ensure action was taken if the risk culture was inconsistent with their strategy or risk appetite. In 2015, we established a small specialist team within APRA to provide a centre of expertise for our supervisors on this issue, and published an information paper in 2016 on the steps financial institutions were taking to meet the standard.[1]

At the time we published that paper, we also flagged our intent to review remuneration practices to examine whether they were reinforcing a sound risk culture. That work was undertaken through the latter part of 2017, and we are releasing the results this morning.[2] I will summarise the main messages shortly, but in brief we see considerable room for improvement.

The role of incentives

From the moment we’re born, incentives influence almost every decision we make. At some stage when you were young, I’m sure you were incentivised to eat your vegetables by the promise of a far more appealing dessert. Incentives continue to influence our decisions into adulthood, and few have more weight than money. It’s natural, therefore, that companies use financial incentives to encourage staff to go above and beyond the bare minimum required to stay employed, whether in the form of sales commissions, annual bonuses or the prospect of promotion.

Those incentives are a response to the fundamental principal-agent problem that exists in corporate life: that the interests of managers and staff are not naturally aligned to those of the owners of corporate entities. Structures are needed to bring them into closer alignment. By and large, that alignment is pursued through remuneration incentives.

When working effectively, incentive arrangements are a win-win: companies generate higher productivity, more sales or better quality output, and staff are financially rewarded for their efforts. But the equation is not risk free. Improperly designed incentives can encourage actions or attitudes that are contrary to the long-run interests of the company itself.

The preeminent case study in getting it wrong is the fall from grace of Wells Fargo, which had designed an incentive system that, as one of the bank’s major investors, Warren Buffett, said, 'incentivised bad behaviour.'[3] Closer to home, it’s no surprise that areas where the financial sector has been dogged by scandal have been areas where the most basic form of incentive – sales- or revenue-based rewards – are prevalent: financial advice, broking, mortgage lending, insurance sales, financial markets trading. Stephen Sedgwick’s review of product-based commissions and payments in retail banking, for example, found that incentives appear to have driven "behaviour that was not in the best interests of customers."[4] Evidence presented during the opening weeks of the Royal Commission has only underlined that point.

Our focus, however, is different. On the prudential front, for example, APRA has been raising concerns for some time that the competitive drive for market share and profits in lending for housing has produced incentives to lower credit standards. Coupled with some borrowers’ own incentive to ‘do whatever it takes’ to get a loan, and the excessive comfort that comes from a period of rapidly rising house prices, there has been inadequate incentive hard-wired into the system to seriously scrutinise applicants’ ability to repay the loans they take out. This obviously has the potential to create adverse outcomes for customers, but also in extremis for the areas of APRA’s interest: the safety of the depositors whose money is being lent out, and financial system stability more broadly.

Accountability

If incentives are the carrot used by companies to boost staff performance, accountability mechanisms are the stick. A robust system of accountability ensures employees, including directors and executives, face consequences for their decisions. Rewards should flow to those who make sound decisions that generate wealth and benefits over the long run. But there should be clear consequences for adverse outcomes too. As with incentives, money – in this case, withholding it – is often the main mechanism by which accountability is imposed.

However, the perception in the community is that in the financial sector, particularly at senior executive level, the carrots are large and the sticks are brittle. Not only are rewards generous, but there are seemingly few repercussions for poor outcomes. It’s in the industry’s interests that this perception changes. Without effective accountability, particularly at the highest levels, companies send a message, to both employees and the public, that they are all care and no responsibility.

From a prudential perspective, this is a major concern. If there appear to be no consequences at senior levels for poor risk outcomes, it undermines sound risk management and the long term financial health of the institution. So, over the past year, APRA has been looking at how well incentives and accountability work at senior levels to support good long-term outcomes, in the context of our broader work on risk culture in financial institutions.

Review of remuneration practices

Incentive and accountability regimes should reinforce one another. Getting incentives right ex ante, so that they promote good commercial outcomes for financial institutions and their stakeholders alike, guides individuals away from behaviours or decisions that produce poor risk-taking and damaging results. Accountability mechanisms operate ex post to ensure good outcomes are recognised, and those who are responsible for bad outcomes face some form of consequences. Designed well, incentive and accountability regimes provide important support for a strong risk culture.

The first international reforms after the global financial crisis were not stronger capital and liquidity requirements, but rather a new set of internationally agreed principles for sound remuneration practices issued by the Financial Stability Board (FSB).[5] Their development reflected a recognition that prior systems of incentives and accountability had failed badly: they had produced little constraint on risk-taking, and little evidence of consequence for those who were ostensibly responsible.

The current Australian remuneration requirements laid out in our prudential standards largely follow the FSB guidance, setting out in a principles-based manner the essential components of a remuneration system designed to deliver a closer alignment between remuneration outcomes and sound risk management. But having the necessary frameworks, policies and procedures documented is not enough. They need to be applied rigorously in practice.

The recent review we undertook therefore involved not just a review of documented systems and procedures, but also the actual outcomes for individual executives. This group included most of the senior executive team at each institution, as well as a selection of high earners outside that group, and selected risk and control personnel. This necessitated collecting details of the performance scorecards for each individual, the corresponding performance assessments, details of Board Remuneration Committee deliberations, and the ultimate remuneration outcomes. We sampled from 12 large institutions, covering 280 senior roles across the ADI, insurance and superannuation sectors. Over a three year period (covering 2014 to 2016 outcomes), we had roughly 800 case studies to examine.[6]

At its heart, our goal was to assess how well the objective of the prudential framework – that remuneration should be aligned to sound risk management and the long-term financial health of the institution – was being met in practice. Or, to use my earlier analogy, we wanted to see whether practices for rewarding pilots were aligned with consistently flying safely.

The overall results of the review are being published this morning. They show considerable room for improvement. The headline message is that the institutions we reviewed by and large had the required frameworks, policies and processes that could provide the basis for a sound system of remuneration. But in many cases their practical application left something to be desired. In other words, they were present in form, but less so in substance.

In particular, while frameworks, policies and processes existed to align remuneration with risk outcomes, their application in practice was often less than robust. Given Australian financial institutions have on the whole been financially successful, executives have been rewarded accordingly. But that financial success has not been universal, nor without a number of missteps. Yet there has been limited evidence of material financial consequences for senior executives when risk outcomes have been poor in their area of responsibility.

Such a situation – where poor outcomes are not reflected in remuneration, even when (or more likely, because) short-term financial targets are hit – can only serve to weaken the risk culture within financial institutions. If variable remuneration is rewarded based primarily on ‘how much’, without sufficient regard to the ‘how’, it creates incentives for short-term risk taking, and a disregard for risk and control frameworks. Fine words and aspirations about the importance of risk and compliance – the tone from the top – will be undermined if it seems that senior leaders are operating in accordance with a different set of rules.

That was compounded by the fact that there were multiple examples where employees below senior executive level were subject to downward adjustments in variable remuneration in response to evidence of poor risk outcomes. However, it was less common to see corresponding adjustments at an executive level recognising overall line or functional accountability. That is not to imply there should be a one-for-one adjustment in each and every case. But overall senior executives seemed somewhat insulated from the consequences of poor risk outcomes.

APRA has to date had a principles-based approach to its remuneration requirements. That remains our preferred modus operandi, as it allows financial institutions of all shapes and sizes to develop remuneration practices that are tailored to their particular circumstances and business model. An international investment bank, a domestic insurer, and a small mutual organisation will rightly want to have quite different remuneration arrangements. Prudential requirements that are too prescriptive will interfere with that sensible outcome.

But our review suggests the current state of affairs is not always delivering a sufficiently strong alignment between remuneration and good risk management. More needs to be done to make risk considerations a genuine and material component of remuneration outcomes, including through greater weighting of risk outcomes and greater use of non-financial measures of performance. Extended periods of deferral could also help provide more ‘skin in the game’ for longer periods of time, allowing the outcomes from risk-taking to be more readily observed. At the most senior levels, Board Risk Committees will need to play a stronger role in overseeing executive remuneration outcomes, and not be afraid to firmly exercise their discretion when events warrant it, regardless of the mechanical application of a scorecard.

On a positive note, some institutions are already making changes to address the issues we have raised. However, I expect that some revisions to our framework, to provide greater clarity as to reasonable expectations for sound practice, will be needed. The new BEAR regime will also necessitate the revision of some aspects of ADIs’ remuneration frameworks to ensure compliance with new minimum remuneration deferral requirements. It will be a pity if banks (and others who can benefit from it) do not use BEAR as a trigger to reconsider their remuneration frameworks more holistically in light of our findings, with a view to considering how to better match remuneration outcomes with long-term performance and good risk management.

Short-termism

Before I wrap up, I want to acknowledge that the remuneration structures and practices of the financial sector – particularly the large listed organisations – are not totally endogenous.

The short-term growth- and profit-based incentive arrangements for senior executives that attract a great deal of community disdain are, at first glance, a product of shareholder demands. Attempts to tie assessments of senior executive performance to a wider set of criteria have been met with investor opposition – and in one notable case in recent years, a ‘first strike’ vote against the remuneration report. That is disappointing.

I’ve been told by one major global investment manager that issues like staff engagement and customer satisfaction are just ‘part of the job’ and don’t warrant performance rewards. I found that interesting, since I thought making a profit and growing the business was a pretty basic ‘part of the job’ too. But it also highlighted that the explanation for growth- and profit-based incentives as being a demand of shareholders may not be entirely true either.

After all, as a current advertising campaign regularly reminds us, everyone in this room is ultimately a shareholder in many large financial institutions, and if this audience is in any way representative of the community, there are likely to be many of you who think that a broader focus on a wider range of performance metrics would be no bad thing. But, of course, it’s not the ultimate shareholders who have a say: it’s the intermediaries (and it may be a chain of intermediaries) who manage and invest on our behalf who most directly influence the types of incentive arrangements that get established. And those intermediaries have their own incentives and short-term performance targets to meet, including maintaining their position on the all-important league tables.

The executives of a financial institution that makes the decision to invest for the long-term, with benefits to a broad set of stakeholders in mind, is unlikely to be rewarded for it. The same for an organisation that decides to dial back its risk settings, at the cost of foregone market share in the short term, in the face of a higher risk environment. That’s evidence of a broader principal-agent problem of misaligned incentives in action: one that unfortunately takes me beyond my brief for today, but certainly is worthy of further attention if we really wish to think about the issue in a holistic manner.

Concluding remarks

In the shadows of one Royal Commission, it’s worth reflecting on the insights of another. Justice Owen, in his report on the failure of HIH Insurance a decade and a half ago, lamented that individuals in that organisation, faced with considerable pressure to sustain profitability, all too often asked ‘is this legal?’ rather than ‘is this right?’.[7] The results of misaligned incentives and a lack of accountability for long-term outcomes, we now know, were disastrous.

The pressure within large financial institutions to sustain their strong financial performance is immense. That is why it is important that more is done to better reflect the extent of risk-taking, and long-term outcomes, in measures of performance. Financial organisations are adept at designing financial incentives for staff to say ‘yes’ to taking risk – after all, taking risk is how profits are generated. Far less incentive exists to say ‘no’, even when it is the right thing to do for the long-term interests of the company itself.

The results of our review point to a range of areas where the financial sector has considerable room for improvement in the way it determines how its executives are rewarded. To return to my earlier analogy, it’s too easy for financial pilots, unlike those who actually take to the sky, to walk away from the scene of an accident unscathed. That can do nothing other than jeopardise the establishment and maintenance of a sound risk culture.

Getting incentives right is not easy. Nor is establishing robust accountability mechanisms. APRA plans to look at strengthening the prudential framework to support both. But it will be disappointing if these issues are just viewed as another exercise in regulatory compliance, or something that gets in the way of good business. Rather, when done well, appropriately structured incentives and accountability mechanisms are a critical component of sustained commercial success. So we are keen for industry participants to take up the challenge of improving themselves, as some are already doing, rather than waiting to be told what to do. Particularly in the current environment, I hope they see a strong incentive to do so.

Footnotes

1. APRA Information Paper (October 2016), Risk Culture.
2. APRA Information Paper (April 2018), Remuneration Practices in Large Financial Institutions.
3. Warren Buffett, Interview with CNBC, 26 February 2018.
4. Retail Banking Remuneration Review (April 2017), available at https://www.betterbanking.net.au/.
5. Financial Stability Board (April 2009), Principles for Sound Compensation Practices.
6. Although this review covered only 12 large institutions, they nevertheless accounted for the bulk of the assets of the financial system.
7. Report into the failure of HIH Insurance (April 2003), volume I, p Ixiii. "In an ideal world the protagonists would begin the process by asking: is this right? That would be the first question, rather than: how far can the prescriptive dictates [of the law] be stretched?"