Creative solutions for unique problems

Credit risk management systems are surprisingly sophisticated at many Nordic banks. Ellen Leander examines the changes being made to existing infrastructure to keep that edge.

Their home economies may be small, their loan portfolios not very diversified and their potential customer base somewhat limited, but banks in the Nordic region are easily among the most advanced in Europe in their peer group when it comes to credit risk measurement and management. Indeed, bankers say it is these very limitations that have driven the region’s financial institutions to pursue the development of more sophisticated credit risk practices. And yet, these same banks have not taken credit risk management to the same extremes as many international banks – relationship banking is alive and well, and bank executives there say it is unlikely that the credit derivatives market will ever drive loan pricing in the region.

There are a number of factors that have driven the rapid evolution of credit risk management technology at these banks over the past half-decade. There is no doubt that Nordic banks were badly shaken by the financial crisis that ripped through the sector in the late 1980s and early 1990s – indeed, in some countries the government still owns substantial stakes in the banking sector. Many Nordic banks have turned around and sought out some of the best practices among the international banks and modified them to apply to the specific challenges that their home markets present. The one outstanding exception to this is Sweden’s Svenska Handelsbanken, which has gone down a different route entirely but achieved the same results – an improved credit risk picture. Nearly all the region’s top banks have publicly stated that they intend to apply for the advanced internal ratings-based (IRB) approach under the revised Basel Accord.

These banks have also faced the additional challenge of integrating a number of acquisitions over a relatively short period of time. For example, Finland’s Nordea bank is now a pan-Nordic bank, the result of four major mergers over the past three years. This flurry of mergers has also spurred credit risk management IT development at Nordic institutions, because group executives at the banks’ headquarters desire systems that can give them a better overall picture of credit risk in their new, far-flung empires.

On top of this, many Nordic bank executives claim there is a cultural element driving the development of these systems. Nordic banks “are more on the forefront, regarding systems”, says Holger Otterheim, head of group credit at Nordea, based in Stockholm. “I think the Nordic banks are by tradition very quantitative, always looking into new ways of trying things. They are also very cost-conscious, always looking to see if something can be replaced by a system to make it more efficient. So in comparison with the rest of Europe, we are doing quite well, but compared with the US we are still a little way behind.”

Nordea exemplifies all three of these trends. After its spate of mergers, executives there are concentrating on building a credit risk framework that incorporates the best of ‘best practices’ from the US and Europe. One of the concepts they are adopting from a Norwegian bank, Christiania Bank og Kreditkasse (CBK), acquired in early 2001, is economic capital. CBK had had an economic capital programme in place since 1998. “They were pretty updated,” says Otterheim, “and they knew the pros and cons of the technology available. This helped us when we worked through the process for the whole group.”

Otterheim says the introduction of economic capital measures was driven by the need to be better able to communicate the relationship between risk and profit within the company, and also by the coming implementation of the Basel Accord revisions. But he says the firm is also using economic capital data to communicate with shareholders. The firm is implementing the economic capital model at all its subsidiaries, and expects to finish by the end of 2002. One obstacle that Nordea has had to overcome is the lack of a common platform among the head bank and all its subsidiaries spanning four countries – each bank it has acquired in each country has had a completely different IT system in place. The solution is to apply the same rating methodology across all firms, simply adapting the IT systems within each bank to the new methodology. The single framework that Nordea uses among its divisions sections the loan portfolio into different segments – corporates, small corporates, households and bank & country. Each segment has a credit risk assessment system tailored to its individual features. For example, the internal ratings methodology for corporate customers is built around 18 grades, while banks are rated on the basis of their individual strength, the country risk and their ownership structure.

Once the economic capital system is implemented, Otterheim says he is looking forward to being able to compare returns across the group’s divisions and geographies. “It gives you a golden opportunity to do benchmarking within the group, to learn what you can do better, and what the strengths and weaknesses of the group are,” he says.

A Swedish bank, SEB, has been working with its economic capital measurement system since 1994, according to Ulf Grunnesjo, senior vice-president and head of group risk and capital analysis at the bank’s Stockholm headquarters. Jacob Wallenberg – who is chairman of the bank and a member of Sweden’s prominent industrial Wallenberg family – brought back the concepts of economic capital measurement and management with him after he worked in the US in the early 1990s. SEB then worked with consultants from Bankers Trust’s well-known risk management division (the bank merged with Deutsche Bank in 1998) to implement a portfolio management system for its credit risk team. The advent of the Basel Accord proposals has focused SEB’s energies even more in this area.

Basel focus
“I think the overall focus for the bank right now is Basel,” says Grunnesjo. “We will make sure we qualify for the IRB approach. Whether we use the approach or not is different. It depends on the outcome.” Not surprisingly, the bank’s Swedish operations have amassed a considerable database of credit risk information since it began implementing its portfolio management system in the mid-1990s. Now, the bank is attempting to implement its methodology in its subsidiary banks – which are scattered across the Nordic region, the Baltic states, Poland and Germany. Any credit risk methodology that SEB implements must be approved by local regulators as well as by the Basel Committee – and be of use to the bank’s management. This is a tall order, given the differing economic conditions, banking industry environments and levels of regulatory sophistication in each country.

“What we ended up with on the credit risk side is local methodologies that all need to match a master scale,” says Grunnesjo. For now, judgements will be made using local methodologies, which will then be translated to the master scale on a monthly basis. In terms of IT infrastructure, SEB purchased US-based software firm SunGard Trading and Risk Systems’ Panorama system to organise the data coming in from the individual subsidiaries. “We did that not so much because it has an expanded ability in terms of calculating risk, but because we needed something to focus our efforts on, in order to aggregate our data,” says Grunnesjo. “We thought it would be a good idea to get a system that had all the [risk analytic] capabilities already in place so we could focus on getting the data [from the subsidiaries] into the system.” The credit risk IT systems at the individual bank subsidiaries are being developed or adapted using internal resources.

For SEB, the point of collecting all this data is for its loan officers to use credit risk modelling in the decision-making process in all subsidiaries. “People now understand that risk is something that should be priced,” says Grunnesjo. “It’s so obvious once you get the numbers – the numbers become very important to you.” The bank’s Nordic subsidiaries have been using the credit risk information gathered by the bank to model credit pricing for loan products for about a year.

But Grunnesjo admits there have been a few bumps along the road to understanding. “The problem has been that people have been too overly reliant on the models,” he says. “They thought, ‘Well you have the model and you use the model to evaluate me at the end of the year. So of course I should be acting according to the model.’ But we have developed the model and know its shortfalls, so we say that’s not really true. Use your common sense. Have a look at the model, think about what it tells you, then do what you think is right.”

For example, Grunnesjo says the credit risk models were producing pricing for certain sectors, such as real estate lending, that differed substantially from the prices companies were being offered from competing banks. In the end, he says, loan officers had to make case-by-case decisions about whether to reprice the loans or not, taking the rest of the company’s relationship with SEB into account.

Oslo-based Den Norske Bank is not far behind its Swedish and Finnish counterparts, and the firm’s risk management team has had to overcome the unique challenges that the Norwegian environment presents. The bank’s history of credit risk management dates back to 1995, when it started calculating risk-adjusted return in its corporate and shipping lending divisions on PCs. It brought in consultants Oliver Wyman & Co at the time to install that platform – the bank had been hit badly by skyrocketing non-performing loans in the early 1990s and the Norwegian government had been forced to take a stake in the bank to keep it afloat.

“When we looked into some of the large losses we had made in times of catastrophe, we saw that we were mostly purely opportunistic,” says Trygve Young, group chief credit officer at Den Norske Bank in Oslo. “We were doing deals just to make money that turned out to be too high-risk.”

Developing a portfolio model for credit risk management wasn’t easy, though. “The Norwegian market is different,” says Jacob Laading, vice-president for credit risk models. Indeed, Norway has a population of just 4 million people, and its industry is very concentrated in a handful of sectors, including oil and shipping. As well, there are more than 130 banks, many of which are small savings banks. “You cannot rely a lot on models that have been developed for other economies,” says Laading. “We have an economic structure that is markedly different, and that means we have to rely on our own data.” Luckily for Den Norske Bank, it has a data set that goes back to 1990/91, which is very comprehensive for the Norwegian market.

“Once we had developed an internal model, we could benchmark that against all the other things that had been done abroad,” Laading says. “We have found that it was better to do that, than to use an existing portfolio model based on other data and sort of straitjacket that into our data.” He says when the firm was developing its portfolio model, it looked at RiskMetrics’ CreditMetrics and other modelling products, and found it difficult to identify the right drivers in those models to fit the Norwegian data that Den Norske Bank had to hand.

Today, the concepts of expected default rate, severity, expected loss and economic capital usage are applied at all levels of decision-making, from each individual borrower up to the total lending portfolio. By the end of the year, Young says the firm will have a system in place that will be able to analyse large exposures to specialised risks – particularly useful in the Norwegian environment.

But Young and Laading say their bank has run up against the same challenges that both Nordea and SEB have faced when it comes to translating the pricing that the credit risk model spits out into the actual market-place. “Due to the fact that many of our competitors lack sophisticated tools – especially the savings banks – we do have some demanding tasks with small and medium-sized companies where loan pricing is far away from where we would like to see it,” says Young. “Our colleagues working with the customers are a bit frustrated, because we are imposing on them a system that measures risk-adjusted profitability, and they have some customers that don’t meet those benchmarks. They have to make a strategic decision to see if they can improve the risk-adjusted profitability.”

Clearly, Nordic banks are moving swiftly to implement the kinds of credit risk methodologies and IT systems that will support Basel when it is finalised, but the applicability of all this fancy firepower to the domestic economies that these banks are operating in remains limited in areas. So although executives are very open to using these tools, and remain excited about the ability to use credit modelling to price risk, they remain wary of letting the numbers drive the business completely. “Given the main markets we are into, for us it is all about relationship banking,” says Den Norske Bank’s Young. “I understand that the large US, UK and maybe German banks may have a kind of market approach, buying into credit risk and being able to compose their balance sheet in one way or another. But they have some underlying tasks to hand that we don’t, and we are not in those markets. Coming back to the business concept, in our market our profitability over time is linked to our ability to sell more products to our customers.”Risk