Glen Paul: G'day, and welcome to CSIROpod. I'm Glen Paul. In the financial services industry everything comes down to risk management. Whether it’s assessing a borrower’s capacity to repay a loan, determining future prices for commodities, or forecasting exchange rate movements, the financial institutions that have the best mathematical and statistical methods to assess risk have far better potential to reduce their losses and increase their gains compared to their peers.
In an effort to assist the industry CSIRO is applying specialist skills in areas such as quantitative risk, mathematical modelling, probability, statistics, finance, economics, and computer programming, to deliver innovative risk management solutions for financial services businesses.
Joining me on the line to discuss it is Alan Dormer from CSIRO Mathematics, Informatics and Statistics. Last time we spoke Alan, it was about transport logistics, and I didn’t realise you were also involved with the financial industry, but science is everywhere. How long has CSIRO been offering risk management for financial services?
Alan Dormer: Yes, we’ve been in this area for over ten years, working with a variety of institutions, banks, people who provide software for helping the banks, and also insurance companies.
Glen Paul: OK. So what’s involved? Is there one underlying formula for financial risk management, or is every case different?
Alan Dormer: Yeah, pretty much every case is different. There are some themes which run through this, for example banks are particularly interested in providing sufficient capital to be able to ride out risks in terms of market risk for example, credit risk, or operational problems, so we do work helping banks determine how much capital they should be holding to cover these risks. Similarly within the insurance industry, they’re very interested in the most up to date scientific methods for calculating premiums and understanding their exposure to phenomenon such as natural disasters.
Glen Paul: And how does this differ from other financial risk systems on offer?
Alan Dormer: There are a number of areas I think which CSIRO has got some novel approaches. One is we use a technique called real options, and that is a way of dealing with decisions which you’re making now against future uncertainty. It allows you to make investment decisions in a much more sophisticated way than traditional methods, for example discounted cash flow.
We also use a technique called extreme value theory. In simple terms what this allows you to do is to use small events which are quite common to work out the possibility of rare events. And the problem with rare events is that they don’t happen very often, so you don’t really have a very good database. For example, if you’re looking at trying to work out a one in 100 year event, and you only have 20 years worth of data, it’s rather difficult. But using this technique we can calculate the possibility of rare events, and we can also combine this with expert views and data from other sources to get a very robust view of what the risk of say a natural disaster or a financial event is, based on past data.
Glen Paul: So how do you test or validate the models that you’re using to ensure they reflect current market practices?
Alan Dormer: Yeah, there’s two areas. One is we do provide software for a couple of organisations, and we can track the performance of that software against what’s happening in the market for example. And the other area is slightly different, is that we are called in sometimes to validate some of the modelling that the financial institutions do in-house, and that is for some of the regulatory agencies. So basically they would create a model to determine, for example, the risk and the capital that they require to keep... to cover that risk, and a financial regulator requires a third party to validate that, so we provide that as a service to our financial customers.
Glen Paul: Righteo. So is actuarial science the only science used in developing these risk assessments?
Alan Dormer: I would say what we use is more fundamental statistics and mathematical optimisation techniques, rather than just pure actuarial, so it’s using similar techniques we use in different areas. So I would say that the actuarial type techniques are generally specific for the financial sector. What we bring is our experience across a wide range of sectors providing, applying statistics to lots of problems in the world, and therefore I think we can add value to the sort of actuarial techniques that the banks use.
Glen Paul: Hmm, and speaking of problems in the world, many of course are worried about the world economic crisis, and it is deepening, particularly with the situation in Europe. How is this uncertainty dealt with when modelling?
Alan Dormer: That’s an interesting question. In fact we’re talking to a number of clients now about looking at future scenarios, so I think in the past there hasn’t been an awful lot of thought about looking really long term, and I think there are possibilities now to use scenario analysis.
We’ve developed a technique, not within my group in CSIRO but other parts of CSIRO, which is called foresighting, and that is looking at long term trends, and what would happen if this happened. And we can combine that sort of knowledge with expert knowledge of, for example what might happen in the economy in the world, and use the scenario analysis to see if the uncertainty in the future is adequately covered by for example capital reserves, or if an investment portfolio is robust enough for the downside risk.
Instead of just selecting investments on the grounds of fundamentals in a company and maybe the outlook in a particular part of the economy, we can look at a range of outcomes and a range of scenarios, and say well if the worst happened what is the likely impact on a portfolio or an institution.
Glen Paul: So who can use CSIRO’s scientific approach to financial risk management – big players, small players, anybody?
Alan Dormer: I would say in the banking sector we do a lot of work for the larger banks because they are probably the most sophisticated, and they can afford to have departments of people doing this in-house. For the smaller institutions we work with a software company for example that provides software to the banks, so for smaller institutions we would probably prefer to work with people who provide software for them because they don’t have a department themselves.
It is applicable across the whole of the financial sector, for example we have worked with insurance companies, we’ve worked with banks, and we’ve worked with foreign exchange organisations, and there’s probably also potential within the financial services sector as a whole because all financial institutions are required to understand their exposure to risk and make adequate provision for it.
Glen Paul: What about up to EU level, can they contact you for assistance?
Alan Dormer: [Laughs]. I think that would be... we probably don’t do the sort of macroeconomic stuff – that would probably be more of an economics point of view. But in terms of the bank and financial institutions there is information to be gained by looking at these events, because I think if anyone could have predicted the global financial crisis adequately they would have probably become very rich.
Glen Paul: Indeed. Alright, so if people want to get in touch they can do that via the website, or would you like people to email you directly?
Alan Dormer: My email’s on the website anyway.
Glen Paul: OK.
Alan Dormer: I’m more than happy to talk. I think we see a lot of future in applying further science to the financial sector, and another thing that’s become apparent recently with our work with Department of Human Services is that some of these projects are not just about mathematics, there are other sciences that can help. For example behavioural type sciences or even how the customers behave is an interesting angle to add to the research services that we offer to the financial sector.
Glen Paul: Well, another amazing string to the CSIRO bow. Thanks for discussing it with me this morning, Alan.
Alan Dormer: You’re welcome.
Glen Paul: Alan Dormer. For more information find us online at www.csiro.au. You can like us on Facebook, or follow us on Twitter at CSIROnews.