The parametric approach to quantifying operational risk

One of the most common ways of starting to analyze operational risk from a
quantitative perspective is to look at losses arising from the day-to-day operations of
the bank. Caban Thantanamatoo, of Manchester Business School, published a paper
on quantifying mishanding losses which arise in the Operations Department. He
decided to look at mishandling losses as they were quantifiable and there was a
significant amount of data available within the bank and defined a mishandling loss
is a loss arising from a control failure/uncontrolled process. This failure will result
in one of the following:
Ω Compensation being paid to a counterparty for any loss incurred by them
Ω Charges being levied by corresponding banks to cover amendments or cancellations
(usually dependent on the amount and period involved)
Ω Overdraft costs on ‘nostro accounts’
Ω Funding loss – where an account is overdrawn due to incorrect funding and the
overdraft rate is higher than the funding (lending) rate
Ω Penalty fees charged by corresponding banks/central banks for late payments,
amendments/cancellations, late/no preadvice or FNR (Funds Not Received costs
Ω Opportunity loss – where no actual loss is made, but funds are out of the bank
and an opportunity to invest funds at a higher rate was missed. The concept also
applies if an account is left with a long balance which a firm is unable to use in
the market.
A typical frequency versus severity of losses relationship for a given department/
subsidiary is given in Figure 13.1. The figure shows that the frequency of mishandling
losses is inversely proportional to the severity of the payment. This is to be expected
since, on average, the low-penalty cases outnumber the high-penalty cases. Assume
the figure at t0 relates to the current mishandling loss risk distribution, the curve at
t1 is an acceptable risk distribution and that the curve at t2, the minimum Mishandling
Loss distribution that the bank can tolerate. Then a risk management effort
should aim towards shifting the curve at least from t0 to t1 and ideally to t2.
Possible implementation could involve investment projects in IT and personnel or in
personnel only.

The approaches described in this study have evolved since it was published. For
further explanation see Mark Laycock’s excellent chapter ‘Quantifying Mishandling
Losses’ in Risk publication Operational Risk in Financial Institutions.
For a number of banks, such as Bankers Trust and Deutsche Bank, the implementation
of Raroc has been one of the driving forces behind the allocation of capital
for operational risk. It may therefore be an existing commitment to firm-wide capital
allocation and Raroc that causes banks to venture down the operational risk capital
allocation route. However, given the difficulty of quantifying aspects of operational
risk, the reliance on a single number may itself be an operational risk!
Many banks are concerned about the benefits of trying quantifying operational
risk. Quite a number are looking for an approach which combines both a quantitative
and a qualitative approach. Some banks think they have found the solution using a
causal model approach.