.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
American Risk and Insurance Association is collaborating with JSTOR to digitize, preserve and extend access
to The Journal of Risk and Insurance.
http://www.jstor.org
C The Journal of Risk and Insurance, 1990 Vol. LVII, No. 3, 455-470
456
457
458
Performance
'Williams and Heins (1985) state that, "Becauserisk managersmake decisionsunderuncertainty, much of theirperformanceis difficultto evaluatein the short run."
'The problemis discussedby Mehrand Hedges(1987)on page25 as involving"intangiblecosts
and benefitswhose measureis a purelyjudgmentalmatter."
6
field researchconstituted two parts, which provided input from approximately15
insurance,brokerage,and riskmanagementexecutives.Onepartinvolvedinterviewswith brokers
and insurershavingexpertiseregardinginternationalinsuranceprograms.Theseinterviewsoffered
informationconcerningthe financingoptionsavailableto riskmanagersof international-oriented
organizations.The secondpartof the field researchwas conductedthroughRIMS'Researchand
InternationalCommittees.Membersof these committeesoffered reactionsto the surveyinstrument.
459
460
Per unit risk managementcosts are expectedto declineas the risk manager
utilizes increasingly advanced analytical techniques in making decisions.
Performancewill be affected not only by strategicdecisions,but also by the
implementationof those decisions.Risk managershaveavailablea wide range
of tools to make and executedecisions, each of which can be categorizedby
the degreeof analysisrequired.Furthermore,varioustechniquesdecidedupon
are consideredmore sophisticatedthan others.
Methods to implementrisk managementdecisions fall into a number of
categories.Decisionscan be made with little analysis,such as those based on
customor rulesof thumb. Alternatively,actions or decisionsmay be based on
extensiveanalysisconsideringexpectedloss, variance,historicaltrends,and so
on. Multiple measures are available and used in the regression analysis
discussedin the next section to assess the degreeof analysisemployedby the
firm. These include questions regardingthe firm's choice of retentionlevel
based on use of cash flow analysisand effect on earningsper shareversusthe
level of retention customarily used; the use of flow chart and financial
statementanalysis for identificationof loss exposures;and the use of risk
managementinformationsystems,loss triangles,and probablemaximumloss
estimatesfor evaluatingloss exposures.
Informationreducesthe uncertaintyof decision-making(see Tushmanand
Nadler, 1978). One would expect that the more sophisticatedthe analysis
performedby the risk manager,the more completewill be the informationon
whichdecisionsare made. Hence, a decisionbased on more advancedanalysis
would be expectedto yield betterresults.Additionally,the more advancedthe
analysis undertakenby the risk manager, the better should be his or her
position to negotiatewith insurancebrokers,and to managelocal personnel.
Because administrativeexpenses are not included in the measure of risk
management costs used here, performance will not be worsened by the
administrativeburdenof increasedanalysis.
461
462
CAP:
463
(3.39)* (-1.35)**
+ .006CAP
(1.43)**
(.093)
(-.39)
.007SIZE + .023INDCOST
(-3.56)*
(2.78)*
Valuesin parenthesesrepresentt-statistics,
* indicates significanceat the .01 level, ** indicates significanceat the .10
level.
As the t-statistics indicate, four of the six variable coefficients are
significantlydifferent from zero using a one-tailedt-test. The two variables
whose coefficientsare not significantare centralizationand degreeof analysis.
All of the coefficientsexceptthat correspondingto the captivevariableare of
the hypothesizedsign. Tolerancelevelsexceed .86 for all variables,indicating
low collinearity.Further,correlationsare .23 or lowerfor all pairedvariables,
consideredacceptablelevels. Additionally,residualplots show no consistent
pattern, supportingthe regressionassumptionsof linearityof the data and
homogeneityof variance. The adjustedR2 is .19, F = 3.792, significantat
.0026.
Interpretation of Results
464
465
466
467
Appendix A
Survey Results
Identification
(Mean Score Rating)
How important are the following loss exposure identification techniques to your firm?
not
important
1
2
Inspection by local manager
Inspection by corporate risk mgr.
Inspection by outside expert
Risk survey or checklist
Financial statement analysis
Flow chart analysis
Internal communication, such as informal
conversation with employees
3
3.65
3.46
4.17
2.98
2.71
2.73
very
important
4
5
3.75
Evaluation
(Mean Score Rating)
How important are the following loss exposure evaluation techniques for your firm?
not
very
important
important
1
2
4
3
5
Historical loss development (loss triangles)
3.91
Frequency distributions of past losses
3.88
Probable maximum loss estimates
4.05
Maximum possible loss estimates
4.02
Expected loss analysis
3.85
Risk management information systems
3.64
Retention
(Mean Score Rating)
How important is each of the following considerations in deciding upon your firm's
level of retention?
not
very
important
important
1
2
4
3
5
Local management decision
2.20
Level of expected losses
4.42
Level of retention customarily used
2.99
Effect on earnings per share
3.73
3.39
Availability of insurer premium credit
Effect measured by cash flow analysis
3.30
468
How important are the following considerations in your firm's captive decision?
(asked only if have a captive)
not
important
2
1
Better control over insurance program
Broader (or only available coverage)
Lower expenses and/or loss costs
Profit potential
Better loss control or claims service
Improved cash flow
3
4.12
3.93
3.86
2.99
3.23
3.74
very
important
4
5
Organizational Data
Average annual premiums (captive and commercial) paid in fiscal year 1987
(in thousands of U.S. dollars)
$22,540
Average uninsured (self-insured and self-assumed) losses reported in fiscal year 1987
(in thousands of U.S. dollars)
$20.131
Average annual property and casualty premiums paid by parents to captives during
fiscal year 1987? (in thousands of U.S. dollars)
$2.701
Property
$5.705
Casualty
Number of organizations utilizing the following forms of captives
78
Corporate subsidiary
3
Rent-a-captive
30
Industry mutual (pool)
Average corporate retention (including self-insurance but excluding captives) per
occurrence in fiscal year 1987 (in thousands of U.S. dollars)
Fire & Extended Coverages
$2.313
$5.475
Earthquake
$2.564
Flood
$4.218
Premises-Operations Liability
$5,969
Products Liability
Average total assets for fiscal year 1987 (in millions of U.S. dollars).
$10.604
References
1. Baglini, Norman A., 1983, Global Risk Management: How U.S.
Corporations Manage Foreign Risk, (New York: Risk and Insurance
Management Society Publishing, Inc.).
469
470
23.