between managing it internally or externalizing it to an insurer.
1N ; \PM; ÅZU; LMKQLM[; \W; J]a; QV[]ZIVKM;; \PM; QV[]ZIVKM; KWUXIVa;
_QTT;VMML;\W;IVITabM;\PM;ZQ[S;][QVO;\PM;LI\I;[PW_V;QV;.QO]ZM;;;
.ZWU; Q\[; MUMZOMVKM; WV; \PM; ZQ[S; ZILIZ;; WXMZI\QWVIT; ZQ[S; PI[;
JMMV;KTI[[QÅML;Ja;UIVa;XZWNM[[QWVIT[;I[;I;LQNNMZMV\;IVQUIT;;7VM;
reason for this is the similarity of its quantitative part to non-life
insurance risk, which is not a well-known concept to practitioners in the banking industry. Another reason is the statistical
basis for the quantitative part compared with that of traditional
market risk; operational risk is far weaker, because it is a special
KI[M;I\;MIKP;ÅZU;IVL;JMKI][M; WXMZI\QWVIT; TW[[;M^MV\[;IZM;NM_;
and far between. This is indeed why the operational part of
operational risk management is much more important than the
quantitative part.
There is, however, a more profound basis for the different
nature of operational risk: At heart, analysis of market, credit
IVL; M^MV; TQNM; QV[]ZIVKM; ZQ[S; Q[; JI[ML; X]ZMTa; WV; ÅVIVKQIT; K WV-
tracts.;;;1V;IV;IVITa\QKIT;[a[\MU;QV;_PQKP;ÅVIVKQIT;KWV\ZIK\[;IZM;
XZMKQ[MTa;UWLMTML;;\PMQZ;KWV\ZQJ]\QWV[;\W;\PM;W^MZITT;ZQ[S;XZWÅTM;
can be calculated (for these types of risk) fairly accurately.
7XMZI\QWVIT;ZQ[S;;QV;KWV\ZI[\;;LWM[;VW\;PI^M;I;ÅVIVKQIT;KWV-
tract at its center; rather it has at its core physical activity —
namely, people, premises, machinery and processes. This is one
WN ;\PM;XZQUIZa;ZMI[ WV[;_Pa;\PM;JMVMÅ\[; WN ;QV\MOZI\QVO; WXMZI-
\QWVIT;ZQ[S;IZM;UQVQUIT;;;QVKM;\PMZM;Q[;VW;KWVVMK\QWV;JM\_MMV;I;
ÅZM;LZQTT;IVL;QV\MZM[\;ZI\M[;;NWZ;M`IUXTM;;\PMZM;Q[;VW;JMVMÅ\;\W;QV-
\MOZI\QVO;\PM;ÅZM;M`\QVO]Q[PMZ;_WZSAEW_;QV\W;I;ZQ[S;UIVIOMUMV\;
process. Although it makes every sense to integrate the output
of a loss database into an ERM framework for analyzing the
risk part of operational risk, one should keep in mind that it is
minor compared to the operational part.
Liquidity Risk
Liquidity risk was the preoccupation of bankers up to the ear-
Ta;;;\P;KMV\]Za;;IVL;_I[;ÅVITTa;\IUML;_Q\P;\PM;MUMZOMVKM;WN ;
central banks — the lenders of last resort. By the end of the
;;\P;KMV\]Za;;Q\;_I[;KWV[QLMZML;\W;JM;IV;M`\QVK\;[XMKQM[;;WVTa;
to lash back at the beginning of this century. Analyzing how
\PQ[;ZQ[S;Å\[;QV;I;]VQÅML;ZQ[S;IVITa[Q[;UM\PWLWTWOa;\PMZMNWZM;
merits a discussion.
Liquidity risk comes in two forms: idiosyncratic (also known
as funding liquidity risk) and market. Idiosyncratic liquidity
risk is the classical risk of a bank run, observed so often in the
;!\P;KMV\]Za;;<PQ[;PIXXMV[;_PMV;LMXW[Q\WZ[; TW[M;KWVÅLMVKM;
in the ability of a bank to redeem deposits or, in genera, to
honor its liabilities. This risk is idiosyncratic because an individual bank is responsible for its actions and their effect on
depositors’ trust.
Idiosyncratic liquidity risk must be managed using classical asset and liability techniques. The liquidity gap (shown
QV;.QO]ZM;;;;JMTW_;;Q[; WVM;M`IUXTM; WN ;[]KP;I;\MKPVQY]M;;<PQ[;
IVITa[Q[; IOOZMOI\M[; XZWRMK\ML; KI[P; QVAEW_[; IVL; W]\AEW_[; W^MZ;
[XMKQÅK; \QUM; QV\MZ^IT[;; _PMZM; \PM; ^MZ\QKIT; JIZ[; ZMXZM[MV\; \PM;
VM\;KI[P;QVAEW_;WZ;W]\AEW_;I\;MIKP;QV\MZ^IT;
=`^li\;, 1;DXi^`eXc;Xe[;:ldlcXk`m\;C`hl`[`kp;>Xg
- 50
- 40
- 30
- 20
- 10
0
10
20
30
40
50
Survival time
<PM;[WTQL;TQVM;QV;.QO]ZM;;;[PW_[;\PM;IKK]U]TI\ML;VM\;KI[P;
AEW_[;;)[;TWVO;I[;\PMQZ;[]U;ZMUIQV[;QV;XW[Q\Q^M;\MZZQ\WZa;;\PM;
bank is capable of surviving on its own, assuming no new
business. When this sum becomes negative, the bank requires
external help — for example, a bailout from the central bank
or another government agency. Regulators demand a minimum survival time in order to have the necessary time to consider a course of action in such cases.
Beyond the simple gap analysis, there are additional factors to consider, such as behavioral assumptions for saving
accounts or the quality of liquid assets held for such crises.
The stressing and testing of such assumptions should be an
integral part of liquidity risk analysis.
Even more than operational risk, liquidity risk is a different
beast. Alas, the same urge to integrate appears to be at work
here, making it appear that liquidity risk should be treated
I[; R][\; IVW\PMZ; ZQ[S; KI\MOWZa; ;I[; [PW_V; QV; .QO]ZM;;);;; <PQ[;;
however, is a misconception. Although it is common to think
of risk factors in terms of their value effect, idiosyncratic liquidity is a parallel analytical concept on the same footing as
