Banks assess new credit risk IT as global accounting change looms

Risk managers in banks are amending models to respond to new accounting rules that require bad debts to be forecast in ...
Risk managers in banks are amending models to respond to new accounting rules that require bad debts to be forecast in advance. Reuters

Australian banks are examining a new technology system for forecasting bad mortgage debts and determining the levels of provisions in an economic downturn, following meetings with the United States' largest privately owned software company, SAS Institute. 

As standard variable interest rates rise, banks' bad and doubtful debts are expected to move higher off their historical lows; as they do, provisions will be determined by a new accounting standard known internationally as IFRS 9 and AASB 9 in Australia, which experts say could make provisions more volatile year-to-year if the economy struggles

SAS has been offering banks around the world a platform consisting of new "expected loss" models which can also be used for modelling risk weighted assets and other credit and operational risk. SAS's senior vice-president of risk research and quantitative solutions, Troy Haines, said while existing risk models have been adequate for an "incurred loss" world, "more granularity is now required and current models may not be fit for purpose". 

While many US banks are adopting their CCAR models used in stress testing for the US version of IFRS 9, four of Canada's big banks have turned to the new SAS system, which is already being used in 35 banks around the globe. 

Canada has an accelerated implementation timetable for IFRS 9, which comes into force in the third quarter of this financial year. Australian banks have to start reporting under the new method in their first financial year after January 2018, with reported profits to be impacted in 2019. NAB is already reporting under the new regime. 

The new standard was part of the global regulatory response to the financial crisis when banks around the world raised provisions by too little, too late. An "expected loss" approach is forward looking, requiring banks to calculate a provision at the time a loan is written. This is carried for 12 months initially, but for the life of the loan – typically 30 years – if credit conditions get worse. This requires more detailed macroeconomic modelling than the current standard, which only requires provisions to be raised when conditions start to deteriorate.

All banks have teams working on IFRS 9 implementation alongside consultants and IT specialists across risk, treasury and finance divisions. The work is considered of high strategic priority given the new standard will flow directly through to reported profits. 

Globally, Mr Haines said IFRS 9 is redefining the relationship between bank chief financial officers and chief risk officers by forcing them to work more closely together. He said the SAS system allows banks to conduct dynamic modelling by changing assumptions in session, to improve risk assessment. The system also audits all the data on which the model relies.

"There has been a big shift over the last year with banks using regulation to drive a better outcome. Regulators want more information from banks about how they manage data. For management, model risk is increasingly important," Mr Haines said during a visit to Sydney last week.  

SAS has provided tools to banks for assessing credit risk and fraud detection for decades; its systems are used in many banks in Australia. But like other US tech companies including IBM, SAS is morphing into a provider of "solutions" that offer an integrated, "end-to-end" service. SAS recorded revenues in its last financial year of $US3.2 billion, a quarter of which was reinvested into R&D.;