About Me

This blog carries a series of posts and articles, mostly written by Anthony Fitzsimmons under the aegis of Reputability LLP, a business that is no longer trading as such. Anthony is a thought leader in reputational risk and its root causes, behavioural, organisational and leadership risk. His book 'Rethinking Reputational Risk' was widely acclaimed. Led by Anthony, Reputability helped business leaders to find, understand and deal with these widespread but hidden risks that regularly cause reputational disasters. You can contact Anthony via anthony.fitzsimmons At cranfield dot ac dot uk

Monday, 16 March 2015

De-risking Bonuses

Matching long term incentives to long term risks sounds easy until you consider that long term risks can fester for decades whereas a typical long term incentive has run its course within a few years.

Banks have started to address the issue.  Goldman Sachs requires top management to hold up to 75% of bonuses as share awards until they leave the company, with many senior staff similarly obliged to hold 25% of any bonus award as shares until retirement. This leaves leaders with the temptation to time their departure as they are free to realise their career-long accumulation of shares as soon as they have left.

HSBC has followed Goldman’s example, requiring bonus shares to be held to retirement with a clawback arrangement.  UBS not only has a bonus/malus system but pays most senior bonuses in bonds and shares and requires Executive Board members to hold at least 350,000 shares and the CEO to hold 500,000. At about twenty swiss francs apiece this means holding from 7 to 10 million swiss francs in UBS shares. And Credit Suisse pays part of bonuses in "bail-in-able” bonds that have to be held for 3 years and can be converted to equity or wiped out in the case of trouble.

But none of these arrangements matches the incentive time scale to the period at risk after top management leave, a time when longer term risks can still come home to roost.

In their insightful new book "Risky Rewards", Professor Andrew Hopkins and Sarah Maslen have proposed a solution that includes a 'malus' element that can be set to run for a period after departure to match the duration of post-departure risks.

The core of the scheme is a unitised trust fund into which a proportion of all senior staff bonuses, long or short term, are paid.  On payment in, participants receive units in the fund at their current value.  In the meantime the fund is invested in assets that are not connected with the employer.

The employee or former employee may only cash in their units after the deferral period.  This is done at the unit value prevailing at the time.  However the first charge on the trust fund is to pay the company compensation in the event of defined catastrophic events. 

The authors plausibly suggest that such a scheme would create a highly motivated corps of trusted, knowledgeable and experienced former senior staff keen to provide practical advice to current decision-makers.  Those people would also know where any bodies were buried.

Such a scheme should also reduce sharply any temptation felt by senior executives to focus on short term profit (for example by cutting investment or maintenance or turning a blind eye to unethical sales) at the expense of the long term success of the firm.  This balance between long term success and and short term profit can become even more acute as a leader approaches departure.

In scheme of this kind, the deferral period  can be matched to the influence and seniority of the individual.  Thus a less senior employee with less influence might have a smaller proportion of his bonuses paid into the scheme and the deferral period might be shorter.

In contrast a highly influential individual severely exposed to the temptation to cut corners or save cost to garner short term profit (and a bigger bonus) at the expense of exposing the organisation to long term catastrophic risks might have a substantial portion of her bonuses paid into the scheme and payment out might be deferred for a long period, which could run until several years after his retirement.  The ideal deferral period would depend on the possible latency period of potentially catastrophic risks to which the company might be exposed. 

Another practicality is the issue of compensation for the company if something goes wrong.  This has two parts: defining 'what goes wrong' and quantifying compensation.

As to defining the event, Hopkins and Maslen were writing in the context of the oil industry.  There the major accident they most had in mind was perhaps some kind of fire or explosion.  However, taking that example, oil companies do other things: for example many have trading arms which might give birth to a rogue trader whose ability to operate can be as much due to a systemic risk as any oil accident - perhaps neglecting cuolture or cutting corners on control systems.  This illustrates that the 'event' needs to be defined broadly if it is to capture all potentially catastrophic events driven by systemic forces.

As to compensation, this is perhaps more complicated than Hopkins and Maslen envisage.  BP, with whose travails Hopkins is particularly familiar, is unusual in that it did not buy insurance.  The result was that all losses resulting from the systemic failures at BP at Texas City and in the Gulf of Mexico were paid for by BP and thus suffered by its shareholders.

More typically, the consequences of a major physical accident will be insured at least in part, with only a modest part of the costs falling on the company.  However, reputational damage is a widespread consequence of catastrophic events.  It may not be insurable let alone insured but it is most acutely felt by shareholders in the share price.  If a company comes to be seen as dysfunctional or its management comes to be seen as ineffective, the share price will fall.  The scheme would need to reflect this kind of loss to shareholders.  Making good shareholders' loss through a payment to the company confers a benefit directly to the company and indirectly to shareholders.

Indeed such a fund could come to be seen as the indirect insurer of damage to the reputation of the company resulting from the actions or inactions of senior management, important behavioural and organisational risks with which readers of this blog will be very familiar.  Insurance of reputational risk is something of a 'Holy grail' for many corporate leaders and this could provide a practical solution to which insurers might then be prepared to add greater depth.

This interesting idea needs more thought and development before it, or something like it, becomes a practical proposition. We would welcome readers' comments on how it might be improved and, importantly, made attractive to Chief Executives and Finance Directors.

In the meantime we welcome "Risky Rewards".  It is a readable, research-based analysis and explanation of many risks inherent in incentive and bonus schemes.  It ranges over subjects as diverse as whether and when bonuses and other incentives actually work, how to target them effectively and how to avoid unintended consequences.  And while its roots are in the world of physical accidents, its insights are relevant to all organisations that are run by people. 

Anthony Fitzsimmons
Reputability LLP

Sunday, 15 March 2015

Latency in Systemic Risks

Most behavioural, organisational and process risks typically lie latent for years to decades before they erupt to cause a serious accident or reputational damage.  This is the finding of the latest analysis from Reputability.

We analysed two sets of data to ascertain how long the root cause lay latent before it caused harm or damage.

The first set of data was taken from 24 accidents and crises analysed in 'Roads to Ruin' the Cass business School  report for Airmic.  Almost all the events analysed occurred between 1999 and 2009 and almost all had their origins in what we now call behavioural and organisational risks. 

To this we added a set of 12 major UK accidents analysed by Turner and Pigeon in their seminal book "Man-made Disasters" (1997).  These occurred between 1966 and 1975 and included the Aberfan disaster, the Coldharbour Hospital fire,  the Hixon Level Crossing disaster, the Summerland Night Club disaster, the 1973 London Smallpox outbreak and the Flixborough fire all of which counted behavioural or organisational risks among their root causes.

Since accidents and disasters rarly have a single cause, we and Turner and Pigeon considered when the root causes began to develop and accumulate, un-recognised by those in authority.  This inevitably involves estimates.  Turner and Pigeon analysed their series of accidents into a series of bands ranging from "less than one month" through "3 to 8 years" to "about 80 years".  We did likewise.

The graph below does not show the last two categories, of incidents where the root causes had been at work for more than 20 years.  Thus the graph ends with 94% of events having emerged after 20 years.

Cumulative percentage emerged by N months

The results show:
  • Only 45% of crises had manifested within 3 years;
  • 30% emerged within 3 to 8 years;
  • 25% took longer than 8 years to emerge and
  • 6% had yet to emerge after 20 years' incubation.

These very long incubation periods for what are mainly, possibly all, unrecognised behavioural and organisational risks have three important implications for risk mangers and boards.

First, the long delay before emergence of damage from these slowly incubating risks allows an organisation to appear successful for long periods whilst unknown knowns are harboured in its midst.  The unfortunate truth is that the organisation is suffering from the delusion that all is well because nothing yet appears to have gone badly wrong.  Boards and risk managers should avoid the risk of complacency that our recent research has illustrated by searching systematically for these unknown knowns.

The emergence of child abuse in churches and other respected institutions around the world, countless banking scandals involving banks previously respected as well as despised and the Deepwater Horizon explosion are more recent examples of the delusion that all is well and under control when in fact the organisation is sitting on a powder keg. Since this was originally written, Volkswagen has provided yet another textbook example of understandable but dangerous complacency, with top leaders apparently unaware of longstanding misbehaviour by their staff.

Second, this long latency period has implications for how incentives, particularly so-called long term incentives, are structured.  We have discussed that problem here.

Finally, what of the allegedly widespread leadership thought that 'it won't happen on my watch'?  For a CEO hoping to stay in post for five years, a crude analysis of this data suggests there is a roughly evens chance of having to deal with one of these crises.  About half the risk comes from the current leader and half from the ancien regime.  And if the crisis goes badly, that will be the ignominious end of a career.

Anthony Fitzsimmons
Reputability LLP

You can get a 20% discount when you buy Rethinking Reputational Risk - How to Manage the Risks that can Ruin Your Business, Your Reputation and You through this link by using the code RRRF20