In a concerning development for allocators seeking ‘yield’, speakers at a risk management conference in Germany have pointed to distinct deficiencies in risk management models as applied to the very asset classes being used heavily to boost income.
Allocators are increasingly looking beyond debt to credit, property and asset-backed securities, even if recent fund flow research still found bond funds at the top of the sales tables.
Speakers at the annual risk management conference of Union Investment in Mainz, said much more work was required to improve risk models in these asset classes.
Shortcomings related variously to the cleanliness, availability and depth of data used in risk models for these classes.
One academic from Leibniz University in Leibniz said he expected improvements in modelling for the credit asset class, but then he put a timeline of up to 10 years on making such improvements.
Professor Daniel Rösch from the institute of banking and finance at the Gottfried Wilhelm Leibniz University Hannover, said risk models had been commonly employed for traditional asset classes such as bonds and equities, “but only to a lesser extent for real estate, private equity, ABS and hedge funds. There, they are used very sparsely.
“For ABS we do not have advanced risk models and the data availability is very important, because the quality of the risk models depends on it.”
He estimated 40% or more of risk systems are only ‘partially aware’ or ‘unaware’ of the risks of false forecasts.
While he dubbed this low proportion “a very positive outcome”, he warned much work remained to be done on models.
They must be adjusted to forecasting periods – short-term typically for equities and bonds, far longer for credit – and “tailor made models are needed for different asset classes. Also it is important to enhance the models to identify interdependencies of asset classes.”
Roesch said the availability of data to model risk in credit was “fairly poor”.
Joachim Hein, managing director at Union Investment responsible for investment analysis, pointed to findings that two thirds of investors in real estate did not use risk models. He said qualitative evaluation of risk predominated in that field of investment, “but that cannot evaluate, for example, what is the actual potential opportunity of real estate, and what effect will that have on [the allocator's] yield?”
He argued risk management was every bit as important in real estate, as in other asset classes.