One of the first proponents of allocating risk capital to operational risk was Duncan
Wilson in Risk magazine in December 1995, ‘VaR in Operation’. The argument put
forward was that operational risk could be modeled by using Value at Risk (VaR)
techniques in the same way as market and credit risks. The idea put forward in the
article was to build a database of operational loss event data from both internal and
external sources and fit a distribution to the data points. By taking a certain
confidence interval such as 95% a firm could then calculate the operational risk VaR.
Wilson’s conclusions were, however, that any quantitative methods are only as
good as the the application of VaR to market risk and stress testing and scenario
analysis were still required. The use of ‘gut-feel’ was still relevent and the article
advocated a ‘blended’approach’ of quantitative and qualitative methods. It argues for
a strong ‘subjective overlay’ of any quantitative methods so that relative risk ranking
and the application of a ‘panel of experts’ approach is combined with modeling based
on loss event databases.
One of the first banks to describe how they allocated capital to Operational Risk
was Bankers Trust. Douglas Hoffman and Marta Johnson in Risk magazine October
1996 published an article entitled ‘Operating Procedures’.
At Bankers Trust, we view business operational risk as encompassing all dimensions of
the firm’s decentralized resources – client relationships, personnel, the physical plant,
property and assets for which we are responsible, and technology resources. We also
capture certain external areas such as regulatory risk and fraud risk.
It is well known that Bankers Trust is one of the few banks who have successfully
implemented a Risk Adjusted Return on Capital methodology and it was natural for
them to wish to take account of operational risk in that methodology. Their approach
was to develop a database of loss events so that they could produce distributions of
losses to calculate the risk capital. Bankers use a 99% confidence interval to ensure
consistency with their market and credit risk RAROC figures.
The reasons Bankers Trust give for allocating capital to operational risk are as
follows:
Ω To help in strategic business/investment decisions (what is the total risk exposure
of this decision?)
Ω Manage efforts to finance risk more effectively (not just insurance but also
alternative risk transfer within the firm)
Ω To assist in managing business risk
In Risk magazine Bankers Trust describes the following steps to calculating operational
risk capital using external and internal databases of information:
Ω Identify centralized information (from centralized departments such as human
resources) which predicts control risk across the whole firm (e.g. staff turnover)
Ω Categorize which risks the information falls under (e.g. human resources risk)
Ω Compare new centralized information to internal and external loss events database
Ω Match risk categories where the reasons for losses in the using external and
internal databases are the same as the centralized data predictors
To attribute the risk capital Bankers Trust split risk into three risk factors:
Ω Inherent risk factors which are those created by the nature of the business such
as product complexity (exotic derivatives would have a higher operational risk
than spot foreign exchange)
Ω Control risk factors which highlight existing and potential control weaknesses
(age of technology)
Ω Actual losses suffered by each business division
Each of these factors were scored for each division and controllable risks were
weighted higher than non-controllable (inherent) and ‘the core capital figure was
attributed proportionally across the divisions, based upon their total, weighted score
for overall operational risk’.
Hiç yorum yok:
Yorum Gönder