How to read a risk balance sheet
Chief financial officers have become uncomfortably aware of pensions risk and its impact on their companies’ profit and loss. Their knee-jerk response is to remove it hastily from the balance sheet, usually by closing the defined benefit pensions scheme to new entrants.
But is that the best or most efficient option? Is removing one problem simply creating another? Should companies consider their total risk, rather than simply their pensions risk?
Guus Boender, an econometrics expert at Free University Amsterdam and co-founder and partner of the consultancy ORTEC, suggests that these are the questions both Dutch and international companies should be asking before they close or change their pension schemes.
ORTEC, an international consultancy based in Rotterdam, provides decision support in the form of software and consultancy to financial institutions. It is the leading provider of asset liability management (ALM) and risk management modelling for pension funds in the Netherlands.
Recently it developed a model for what Boender calls “enterprise wide risk modelling” to help companies decide about pensions risk. This is at the heart of what ORTEC does, Boender says. He describes the company as “sellers of knowledge”, where the knowledge is encapsulated in the models. A colleague suggested a development of this idea by defining models as “constructive knowledge.”
“That is a very good way of putting it because models are things that contain knowledge that you can really do something with,” he says.
ORTEC sells this knowledge to pension funds, insurance companies and housing associations in two formats – software licensing and consultancy.
Larger pension funds, which have their own ALM teams, are more likely to choose the licensing option. For example, Hermes Investment Management, the independent fund manager that includes its owner the BT pension scheme among its 200 clients, bought ORTEC’s ALM licence last October. The other option is through a consultancy project in which ORTEC will use the software.
ORTEC must listen to clients if it is to keep its ALM models in tune with their needs, he says: “The correct ALM model is not a model that is up to date. It is a model that is five minutes ahead of today. This is because one of the most important things the model does is decide what the future should be.”
In recent years, ORTEC has developed its ALM process into what is called “phased ALM”, a process which supports the four steps of the institutional investor’s decision making - asset /liability management, ALM implementation, portfolio construction and performance attribution.
“ALM is very important, but we also need to look at the decision making process at different levels, both taking decisions and implementing them. This includes risk monitoring and performance attribution to check that things are going to plan and that the managers are performing as expected,” says Boender.
He compares ALM to a super tanker, which cannot alter course quickly in the event of danger. “Performance attribution and risk monitoring ensures you get from harbour to harbour in the most fuel-efficient way.”
One of the Netherlands’ pension super tankers, ABP, recently bought ORTEC’s performance management software PEARL to enable it to analyse its past investment decisions.
The introduction of nFTK next year has stimulated interest in risk monitoring systems. Boender says he is grateful for the stimulus, but believes that the impact of nFTK has been over-rated. “The nFTK has helped in the current times to motivate pension funds. But nFTK should not be too important to funds because they should have risk monitoring and performance attribution in place anyway.”
ORTEC follows developments in the financial industry closely, seeking opportunities to apply systems it may already have developed, he says. The recent revival of interest in commodity investment, spearheaded by ABP and PGGM was one such opportunity. ‘We had modelled commodity indices many years ago, but then few years ago they became relevant again, “ he says.
ORTEC has also created a decision support models for companies considering moving from defined benefit (DB) to collective defined contribution (DC) plans. This was developed in 2004, well before collective DC schemes became a reality, says Boender.
“The model uses data from a spectrum of collectivity, from zero, which is an individual DC plan, to the point where collectivity or solidarity is almost at the level of a DB plan. It then depends how much collectivity or risk sharing the company wants to put in the system.
“We have been able to show clients, from the data of real pension funds, how these different choices would work out, first for the company, since the company has to pay a higher contribution, and for the people in the scheme, because of the risk that they are bearing and sharing.”
The need for pension fund solvency tests has also stimulated interest in risk management models, in this case short-term models. “To satisfy the requirement of the solvency test you have to say what you probability is one year ahead,” says Boender.
To do this ORTEC also uses risk management techniques such as historical simulation, or ‘bootstrapping’, which are applied in the risk management systems of banks. These generate scenarios based on historical price changes.
The challenge for ORTEC is to provide clients with a consistent picture in both the ALM and bootstrapping scenarios, says Boender “This means that if they look at the first two years of the ALM scenarios they get the same answers that they get when they look at the solvency test.”
The most important stimulus to the development of risk management models, however, has been the growing awareness of companies of ‘pension risk.’ largely as a result of the new International Financial Reporting Standards (IFRS)
Over the past year, ORTEC has worked with a large Dutch company-wide pension fund and a Dutch multinational to develop a model-based global pension risk management framework to help multinationals handle pension risk.
The framework enables a multinational to define and quantify its pension risk, consider the relationship of this risk to other financial risks, and identify its own risk appetite; that is, its risk budget in general and for its pension funds in particular and risk management. It can then hedge any unfavourable pension risk that does not fit into the risk budge.
Boender says the new model could slow the lemming-like rush out of DB schemes in Europe and particularly in the UK: “The CEO of a large Dutch company where we did a collective DC project asked me when the project was over. Did we make a good decision?
“I said ‘I do not know. What I do know is that, given the policy bandwidth you have, the decision was excellent. But what I do not know is whether it is more efficient to reduce risks in the company or pension fund risk’.
“What we need to do right now is to look at the risk balance sheet of a company and then we need to identify what is the most efficient way to reduce the risk.”
Corporates should use an enterprise-wide ALM, says Boender. “I hope that enterprise-wide ALMs will happen more frequently, not only here but in the UK. Otherwise it could be the case that companies will make totally wrong decisions about their pension funds.”