Cassandra Cole and Kathleen McCullough Co-Editors · 2018-09-04 · Cassandra Cole and Kathleen...
Transcript of Cassandra Cole and Kathleen McCullough Co-Editors · 2018-09-04 · Cassandra Cole and Kathleen...
Journal of Insurance Regulation
Cassandra Cole and Kathleen McCulloughCo-Editors
Vol. 35, No. 3
Optimal Cap on Claim Settlements Based on Social Benefit Maximation
Hong MaoJames M. Carson
Krzysztof M. Ostaszewski, FSA, CFA, CERA, MAAAYuling Wang
JIR-ZA-35-03
Accounting & ReportingInformation about statutory accounting principles and the procedures necessary for fi ling fi nancial annual statements and conducting risk-based capital calculations.
Consumer InformationImportant answers to common questions about auto, home, health and life insurance — as well as buyer’s guides on annuities, long-term care insurance and Medicare supplement plans.
Financial Regulation Useful handbooks, compliance guides and reports on fi nancial analysis, company licensing, state audit requirements and receiverships.
LegalComprehensive collection of NAIC model laws, regulations and guidelines; state laws on insurance topics; and other regulatory guidance on antifraud and consumer privacy.
Market RegulationRegulatory and industry guidance on market-related issues, including antifraud, product fi ling requirements, producer licensing and market analysis.
NAIC ActivitiesNAIC member directories, in-depth reporting of state regulatory activities and offi cial historical records of NAIC national meetings and other activities.
For more information about NAIC publications, view our online catalog at:
http://store.naic.org
Special StudiesStudies, reports, handbooks and regulatory research conducted by NAIC members on a variety of insurance-related topics.
Statistical ReportsValuable and in-demand insurance industry-wide statistical data for various lines of business, including auto, home, health and life insurance.
Supplementary ProductsGuidance manuals, handbooks, surveys and research on a wide variety of issues.
Capital Markets & Investment AnalysisInformation regarding portfolio values and procedures for complying with NAIC reporting requirements.
White Papers Relevant studies, guidance and NAIC policy positions on a variety of insurance topics.
© 2016 National Association of Insurance Commissioners. All rights reserved.
Printed in the United States of America
No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any storage or retrieval system, without written permission from the NAIC.
NAIC Executive Offi ce444 North Capitol Street, NWSuite 700Washington, DC 20001 202.471.3990
NAIC Central Offi ce1100 Walnut StreetSuite 1500Kansas City, MO 64106816.842.3600
NAIC Capital Markets& Investment Analysis Offi ceOne New York Plaza, Suite 4210New York, NY 10004212.398.9000
The NAIC is the authoritative source for insurance industry information. Our expert solutions support the efforts of regulators, insurers and researchers by providing detailed and comprehensive insurance information. The NAIC offers a wide range of publications in the following categories:
The following companion products provide additional information on the same or similar subject matter. Many
customers who purchase the Journal of Insurance Regulation also purchase one or more of the following
products:
Companion Products
Federalism and Insurance Regulation
This publication presents a factual historical account of the development of the
framework for insurance regulation in the United States. It does so in part by
using illustrative early statutes, presenting them chronologically, and in part by
using cases that illustrate the interpretation of the crucial later statutes.
Copyright 1995.
Regulation and the Casualty Actuary
This anthology reprints 20 important papers from past issues of the Journal of
Insurance Regulation that are most relevant for practicing actuaries and state
insurance regulators. It covers a wide range of issues, such as ratemaking,
auto insurance pricing, residual markets, reserving and solvency monitoring.
This invaluable reference explains these complex topics in straightforward,
non-technical language. Copyright 1996.
International orders must be prepaid, including shipping charges. Please contact an NAIC Customer Service Representative, Monday - Friday, 8:30 am - 5 pm CT.
Editorial Staff of the Journal of Insurance Regulation
Co-Editors Case Law Review Editor Cassandra Cole and Kathleen McCullough Jennifer McAdam, J.D.
Florida State University NAIC Legal Counsel II Tallahassee, FL
Editorial Review Board
Cassandra Cole, Florida State University, Tallahassee, FL
Lee Covington, Insured Retirement Institute, Arlington, VA
Brenda Cude, University of Georgia, Athens, GA
Robert Detlefsen, National Association of Mutual Insurance Companies, Indianapolis, IN
Bruce Ferguson, American Council of Life Insurers, Washington, DC
Stephen Fier, University of Mississippi, University, MS
Kevin Fitzgerald, Foley & Lardner, Milwaukee, WI
Robert Hoyt, University of Georgia, Athens, GA
Alessandro Iuppa, Zurich North America, Washington, DC
Robert Klein, Georgia State University, Atlanta, GA
J. Tyler Leverty, University of Iowa, Iowa City, IA
Andre Liebenberg, University of Mississippi, Oxford, MS
David Marlett, Appalachian State University, Boone, NC
Kathleen McCullough, Florida State University, Tallahassee, FL
Charles Nyce, Florida State University, Tallahassee, FL
Mike Pickens, The Goldwater Taplin Group, Little Rock, AR
David Sommer, St. Mary’s University, San Antonio, TX
Sharon Tennyson, Cornell University, Ithaca, NY
Purpose
The Journal of Insurance Regulation is sponsored by the National Association
of Insurance Commissioners. The objectives of the NAIC in sponsoring the
Journal of Insurance Regulation are:
1. To provide a forum for opinion and discussion on major insurance
regulatory issues;
2. To provide wide distribution of rigorous, high-quality research
regarding insurance regulatory issues;
3. To make state insurance departments more aware of insurance
regulatory research efforts;
4. To increase the rigor, quality and quantity of the research efforts on
insurance regulatory issues; and
5. To be an important force for the overall improvement of insurance
regulation.
To meet these objectives, the NAIC will provide an open forum for the
discussion of a broad spectrum of ideas. However, the ideas expressed in the
Journal are not endorsed by the NAIC, the Journal’s editorial staff, or the
Journal’s board.
* Shanghai Second Polytechnic University; [email protected]. ** Daniel P. Amos Distinguished Professor of Insurance, Terry College of Business, University
of Georgia; [email protected]. *** Actuarial Program Director and Professor of Mathematics, Illinois State University;
[email protected]. ****School of Finance, Shanghai University of Finance and Economics, Shanghai, 200433,
China; [email protected].
© 2016 National Association of Insurance Commissioners
Optimal Cap on Claim Settlements
Based on Social Benefit Maximization
Hong Mao* James M. Carson**
Krzysztof M. Ostaszewski, FSA, CFA, CERA, MAAA*** Yuling Wang****
Abstract
Caps on damages constitute one of the proposed reforms of the tort system, especially in the U.S. In this paper, we establish models to capture both the claimant’s and the insurer’s behaviors under a system whereby damages are capped. We also establish an objective function of maximizing the social benefit, expressed as the sum of claimant and insurer benefits. We examine the optimal upper limit that balances both claimant and insurer benefits. Our model implies that the optimal value of the upper limit should be the expected value of the random claim payment independent of the type of probability distribution the claim payment follows. Finally, we apply empirical data to our theory and arrive at an estimate of the optimal damage cap.
1
Journal of Insurance Regulation
© 2016 National Association of Insurance Commissioners
Introduction In the ongoing debate concerning tort reform, caps on damages have been
proposed as a possible solution to address the problem of increasing claims costs. One of the key purposes of such caps is to control insured losses, making insurance more available and affordable (Feldman, 2003; Gfell, 2003; and Liang, 2004). Supporters assert that caps would inhibit fraudulent claims and eventually help to reduce insurance premiums. Many states in the U.S. are considering passing, or have passed, legislation creating caps on damages, especially for non-economic damages. As an example, the crisis in medical malpractice insurance markets can be especially pronounced, and caps have been proposed as a method to help militate against such crises.
There are different views about the efficacy of caps. On one hand, many believe that caps reduce losses for insurers, pointing out reductions in losses in jurisdictions that enacted a cap, when compared with those states that did not. For example, Sloan, Mergenhagen and Bovbjerg (1989) found that, based on closed claim data for 1975 through 1978 (and through 1984), insurers paid 31% less in states with caps on non-economic damages, and 38% less in states with caps on total damages. Kessler and McClellan (1997) found that premiums were reduced by 8.4% in the first three years after a tort reform. Tort reform also reduced the likelihood that a physician would be sued. The U.S. Congressional Budget Office (1998) summarized that caps on non-economic damages “have been found extremely effective in reducing the amount of claims paid and medical liability premiums.” Viscusi and Born (2005) used NAIC data for the period 1984 to 1991 to show that caps on non-economic damages reduced losses by 17% and premiums by 6%. Encinosa and Hellinger (2005) found evidence that states with caps on non-economic damages increased the supply of physicians by 2.2% per capita, as compared to states without caps, after observing the impact of non-economic damage caps from 1985 to 2000.
On the other hand, critics contend that an arbitrary cap amount cannot work for every case. For small damages, the award may still be excessive. For greater damages, the compensation may not be equitable. Also, caps may not necessarily lead to reduction of insurance premium rates, as relatively high insurance premiums on medical malpractice insurance may be caused by other factors such as insurance cycles, weak competition or declines in investment returns earned by insurers (Zuckerman et al., 1990; Richard, 2012; and Paik et al., 2012).
Critics of tort reform often cite the case of California to argue that the non-economic caps do not work to reduce premium rates. When California passed the legislation on caps on non-economic damages at $250,000 in 1975, premium rates were reduced only for a short time and then continued to rise. By 1988, premiums on medical malpractice were up 450% in comparison with 1975. California then enacted California Insurance Code 1861.01, Proposition 103, which required every insurer to reduce its rate by at least 20%. Subsequently, malpractice rates decreased quickly and premium rates decreased by 30% or more within three years
2
Optimal Cap on Claim Settlements
© 2016 National Association of Insurance Commissioners
(Glassman, 2004). Many states have passed legislation to impose caps on claim payments. At the
national level, the Help Efficient, Accessible, Low-Cost, Timely Healthcare (HEALTH) Act of 2003, which would have capped non-economic damage awards in medical malpractice actions at $250,000, was passed by the U.S. House of Representatives in May 2004 but was not passed in the Senate (Govtrack, 2004).
The literature on caps on claims and/or damages is dominated by legal analysis. Crocker and Morgan (1998) and Crocker and Tennyson (2002) were unique contributions by the virtue of creating a quantitative model. They study optimal insurance contracts and indemnification profile, when controlling for fraudulent claims for both first-party and third-party claim settlements from the insurer’s perspective. They optimize the insurer’s response to a potentially exaggerated or fraudulent claim. However, determining an optimal cap should also consider the benefit to consumers. Regulators, of course, should attempt to balance both claimants’ and insurers’ benefits, and consider all relevant risks. Shoben (2005) proposes a societal (and by implication, regulatory) perspective on the issue of caps on claims and/or damages by proposing the following three elements of a tort reform to address the issues of measured degree of wrongfulness, measured severity of harm and connectedness between the defendant’s wrong and the plaintiff’s injury. Shoben (2005) also proposes replacement of contingent attorney’s fee by an hourly rate for work done.
In our opinion, whether the cap on damages helps to decrease the insolvency probability of insurers, protect consumers, promote the demand and supply of liability insurance, and improve the social fairness depends on how to rationally determine the cap on damages by scientific methods and from the perspective of maximizing the social benefit. Caps that are too high, as well as caps that are too low, are not optimally beneficial to the realization of maximizing social welfare.1
In our paper, we quantitatively study the optimal cap on claims and/or damages by analyzing both the behavior of claimants and insurers under tort reform law that puts caps on claims and/or damages paid by insurance. We analyze the problem from the regulators’ perspective. We study the following problem: How do we set up an optimal upper limit on claim damages if we want to balance benefits between insurers and claimants? Consumer protection regulators aim for damage awards large enough to properly compensate victims of negligence or of malfeasance. However, insurance regulators also have genuine interest in limiting the awards to levels that would not endanger insurer solvency. Hence, the issue at hand must be cognizant of consumer protection and the prudential regulatory perspectives.
1. In the notes from Gfell (2004), he points out that while a national cap on non-economic
damages in medical malpractice actions may pass constitutional muster, Congress would be wise to attempt less invasive limitations first. That is, Gfell suggests that Congress should seek better economic data upon which to base its decisions, and thus help to assure the crisis would be covered with minimal negative consequences.
3
Journal of Insurance Regulation
© 2016 National Association of Insurance Commissioners
As a preview to our results, we get the intuitive theoretical result that the optimal cap on damages is the expected value of random loss. It implies that a cap on damages is not exogenous to the distribution of random loss. The frequency and severity distribution of liability injuries does affect the caps on damages.2 Therefore, the injured claimant can obtain fair claim payment. Based on our empirical data, our analysis indicates that the optimal cap on damages is far less than caps found in reality, such as caps for damages related to general liability insurance. For some other types of liability insurance (e.g., medical malpractice liability insurance), our analysis suggests that the insurer’s total claim payment without caps is much greater than that with caps. As more than one-half of U.S. states did not have caps on damages, the absence of caps may be one reason for the medical malpractice insurance crisis (besides long-tail liability and exaggerated losses to obtain higher judgment awards).
Our study is organized as follows: In the next section, we establish models used for the third-party claimant. Then, in the third section, we discuss the determination of optimal caps, based on the criterion of maximizing the social benefit. In the fourth section, we provide two important examples to illustrate the application of our models to the problem. We then present sensitivity analysis, and in the final section, we provide our conclusions.
Models of Third-Party Claims Settlements
We consider third-party liability insurance in which we assume the claimant and insurer to be risk-neutral. The claimant has a real loss of X , which is only fully known to the claimant. However, the claimant files a claim of s (≥ X ). The jurisdiction of the claim has an upper limit of claim damage of θ. The claim amount of s may still exceed the amount of θ, but the indemnification amount of I(s) paid by insurer must be less than or equal to θ. If insurer can identify the real amount of damage, then I(s) = X when X θ. When the indemnification amount is greater than the real amount of loss, I(s) – X is the realized exaggerated claim amount for claimant. When the real loss is greater than θ, the highest claim possible is the amount of θ, and X – θ is the loss for claimant after compensation.
Let E Y1
be the expected terminal wealth of the claimant, and E Y
2 be
the expected terminal wealth of the insurer.3 Using that notation, we can write:
2. Gfell (2004) points out (in the notes to the article) that reforms that protect the public are
what are needed, not reforms that blame the injured, the disabled, and victims of medical ineptitude and neglect. It is in this light that caps, based on the expected loss incurred to the injured, may help maximize social benefit.
3. When claimant and insurer are risk-neutral, their expected benefit functions are exactly the same as their utility functions. In the context of the theory of the firm, a risk-neutral firm facing risk about the market price of its product, and caring only about profit, would maximize
4
Optimal
© 2016 Natio
and
Where
WY
Also, C
1
indirect ex
x, and 1
assume thaare strictlythese costs
Whilecosts, not nthat it is a create a sitclose to zeexaggeratioupper limiexpenditurreal loss (t
the expectedproduced, et
Cap on Clai
onal Association o
Y1is the claiman
1,x, represe
xpenditures to
1, with
1 0
at the claim coy increasing wis with the follow
this may seemnecessarily jusreasonably go
tuation where ero, the claimanon or misinforit cannot beres of setting uthe real loss b
d value of its pc.).
im Settlemen
f Insurance Comm
nt’s initial wea
ents the costs
set up claims.
0, where 1 is
osts for claimanith the deviatiowing specific f
m like a simplifst accounting cood model of eas the real los
nt is less likelyrmation since e compensated
up claims will ibeing far below
profit (with resp
nts
missioners
alth, and WY2
i
that the claim
We model tho
s the cost para
nt under a region of real lossfunction:
fication of the costs) incurredeconomic realits gets closer to
y to increase ththe exaggerat
d by the insurincrease. On th
w the damage
pect to its choic
is the insurer’s
mant incurs for
ose costs as a
ameter for the
ime of a cap ons away from θ
reality of costd by the claimaty. On one hano the upper lim
he costs throughtion of claim rer; and, on thhe other hand, cap), the more
ces of labor inp
s initial wealth
r the direct and
function of e claimant. W
n damages of θ, and represen
ts (all economiant, we suggesnd, it intends tomit, or x ih some form oloss above th
he contrary, ththe smaller th
e willing is th
put usage, outpu
h.
d
,
We
θ nt
c st o is of he he he he
ut
5
claimsettinclaimdama
S
falsifi
2
the dimposreducpartieintrodfuncti
T
insurecap re
Deby Thi The vinsurethe clwill lthe otclaimbut ththe poptim
mant to exaggerng up claim. Thm cost using thages and real loSimilarly, C
2
fication of the c
0. We also as
eviation of thesed on the po
ced investigatioes. As we stateduce a similar pion under a reg
The key idea oer cost functioesult in costs in
terminat Maximiird-Party
values of the cer. On one hanlaimant may telose money (pather hand, wh
m payment obtahe insurer willperspective of mization proble
rate the claim lherefore, it is nhe quadratic f
oss.
2,x, is
claim by the cl
ssume that the
e real loss awolicy payout, ton cost of insd above, 1 is parameter: 2 tgime of a cap o
f this simplifieons are designencreasing at an
tion of Oizing Socy Claims
ap on claim pand, when real dend to claim may a larger cla
hen real damagained by the cla benefit becauthe insurer,
m:
Jou
© 2016 Nation
loss, even at thnecessary to imfunction of th
the insurer’s
laimant, set up
associated cos
way from . Wthe falsificatiosurer. Cost parthe cost param
to capture the ion claims and/o
ed costs modeled so that largn increasing rat
Optimal cial Bens
ayments affectdamage X is lmore payment aim) because oge is larger thaimant will be use of less mon
its benefits a
urnal of Insu
nal Association of
he expense of inmpose a measuhe difference b
s costs to i
p a function of
st to the insure
When the loss on cost is redurameters are d
meter for the clinsurer’s effector damages pay
l is that both thge departures fte.
Cap efit Und
t the benefits oless than the cathan real dam
of exaggeratedhan the cap of
reduced in relaney paid than are maximized
rance Regul
Insurance Commi
ncreasing the cure of control between the c
investigate po
f , x, and 2
er is increasing
approaches thuced along widifferent for thlaimant, and nots. The insureryout is express
he consumer afrom the level
der
of both claimaap of claim pay
mage, and the id claim paymenf claim paymenation to the reareal damages.
d via the foll
lation
ssioners
cost of on the
cap on
ossible
, with
g with
he cap ith the he two ow we ’s cost
sed as:
and the of the
ant and yment, insurer nt. On nt, the al loss, From
lowing
6
Optimal
© 2016 Natio
The o
benefit. Th
Generally, benefits ofsocial fairnsocial costto determinexpressed maximizin
We fin
damages a
wealth of bfunction (7claimant (t
This oshown that
4. The
Thomson (1determine ththe insured.
Cap on Clai
onal Association o
optimization phat is,
these two of insurers and ness, balancingts in determininne optimal capas the sum of
ng
nd from objectccording to the
both insurer an7)), and at thethe second termoptimization prt the solution is
Nash (1950) sol981) solution ma
he optimal partici
im Settlemen
f Insurance Comm
roblem of the
objectives cannthe claimants
g the benefits ong the cap of t
p based on the f the benefit o
tive function (e objective fun
nd the claimante same time, mm of the right sroblem is solvs surprisingly s
lution maximizesaximizes the sumpating rate by m
nts
missioners
e claimant is
not be realizeare in conflic
of both insurer the claim. Whobjective of m
of the insurers
(7) that if we dnction of max
t (the first termminimize the ide of objectiv
ved in the Tecsimple:
s the value of them of the two utilit
aximizing the sum
maximizing h
ed simultaneoct. A regulatorand claimant, at is optimal f
maximizing theand the claim
determine the
E Y , we can
m of the right sicost of both i
ve function (7))hnical Append
e product of the ties. Mao and Osm of the benefit
his (her) claim
usly since thr must consideand decreasing
for regulators ie social benefitmants4—that is
optimal cap onn maximize th
ide of objectivinsurer and th). dix, where it i
two utilities. Thstaszewski (2007of the insurer an
m
he er g is t, s,
n he
ve he
is
he 7) nd
7
Tlo
It
easily*( )t
param
T
of losof ranoptimBasedreason
F
It
paramriskinclaimNotabnegatparamof the
Lthe mis, of settinregulathe pupaym
Fsituatmomedata a
The optimal levoss:
t is also impory be dynamic, ) ( )tE X . For
meters of t an
Therefore, we ss distribution ndom loss to th
mal cap. Thered on the ananable and intui
Further, the opt
t is the sum ometers multipliness of the rea
mants and the bly, the costs tive effect onmeters may varem at the expenLet us note thamodel is the exf course, a keyng premiums aators. Furthermublic in genera
ments. Finally, there etion when datents or a maxiare censored o
vel of a cap on
rtant to note theif the type of
r example, if
nd t , then th
can dynamicalchange. Two, the injured and fore, the injur
alysis above, itively appealintimal level of t
of initial wealtied by the varial economic lo
insurance coto the claiman
n the final sory between thense of the otherat the key parapected value o
y parameter soand is likely tomore, it is a real, and to juries
exist many stata are limited.imum likeliho
or grouped, cer
Jou
© 2016 Nation
n damages is th
e following twf distribution o
the loss foll
he optimal cap
lly change thethe optimal cathe frequency
red claimant cour theoretic
ng result. terminal social
th from whichiance of the re
oss, representedompanies, affents and to theocial outcomee two, but the kr, the net effecameter estimatof the real econought in actuaro be studied b
elatively simples in particular,
atistical proced. If data are
ood estimator crtain statistical
urnal of Insu
nal Association of
he expected va
wo points. One,of loss remainlows lognorm
is:
e optimal cap wap is endogenou
and severity dan obtain the al model app
benefit is:
h we subtract eal loss. This d by its varianect the optime insurance come (although, okey point is thact on the societyted for the pranomic losses trial analysis byby both insurae concept that to justify the e
dures for estimperfectly avai
can be easily l methodologie
rance Regul
Insurance Commi
lue of real eco
, the optimal cans the same. T
mal distribution
when the paramus to the distri
do affect the vafair claim pay
pears to prod
the sum of thmeans that bo
nce, and costs mal terminal w
mpany have eof course, theat if we increay is zero). actical applicatto the claimanty insurance firance companiecan be presen
existence of a c
mating it, even ilable, a methapplied. But ees can be appl
lation
ssioners
onomic
ap can That is, n with
meters ibution alue of yment. duce a
he cost oth the to the
wealth. equally e cost
ase one
tion of t. This rms in es and nted to cap on
in the hod of even if lied. If
8
Optimal Cap on Claim Settlements
© 2016 National Association of Insurance Commissioners
only data for cases with caps already applied are available, the data are right-censored. Datta (2005) shows a possible statistical methodology to be applied in such a situation. Below we give an example of estimation of the expected value parameter for grouped data, when only ranges of data are available, and not all individual cases. (This situation may be encountered in practice by both insurance actuaries and insurance regulators.)
Examples to Illustrate Our Results on Optimal Caps
Example 1: General Liability Insurance
Table 1 lists the average claim payment of 217 general liability insurance
policies. The claim payment is divided into several intervals (the first column) within the minimum value zero and the maximum value $300,000, which is the policy upper limit. The second and third column, respectively, list the number of claims and the average claim payment in which the claim payment for each time falling in each interval (the data are used by Xiao, 2008, and originates from Klugman et al., 1998).5
We will now illustrate our result by calculating an estimate of the parameter of the mean economic loss, which in the model proposed here is equal to the optimal level of damages cap. We could estimate it from the sample mean, but given that the data are grouped, it is more reasonable to attempt to fit a probability distribution and then use an estimation of that distribution’s mean. By examining the data, we hypothesize that the lognormal distribution is a good fit. In the Appendix, we perform a test of this hypothesis (the chi-square goodness of fit test) and find that the distribution can be used to model this data. We also find (in the technical calculation) the estimates of the distribution’s parameters, and we use them in the analysis that follows.
5. We use general liability insurance as an example since firms generally purchase this
coverage as one of the first steps to protect their assets. This safety net is critical in a society in which the number of lawsuits and the value of judgment awards have increased over time (Howell, Commercial General Liability Insurance Definition |: www.ehow.com/about_5184654_ commercial-general-liability-insurance-definition.html).
9
* Here$300,0
T
expec
and th
S
equatFigurusingnumbapprorestricperspsolvendepenoptim
e the claimants w000 from the insure
Therefore, the ected value of th
he optimal exp
Substituting th
tion (11), we ore 1 is the em the data in Ta
ber of policieoximately 25%ction, but asseective on the ncy and overandent on the
mization proces
Average C
with real loss in eer due to that being
estimated optimhat lognormal d
pected value of
he estimates o
obtain an estimmpirical functioable 1. It shouls/claims affec
% of all claimsessment of howeffect of the c
all societal wewealth levels
ss itself is indep
Jou
© 2016 Nation
Table 1: Claim Paymen
excess of $300,00g the upper limit s
mal value of thdistribution:
f social benefit
of the parame
mate of the opton of cumulatld be noted thacted by the cs; this does now significant tcap on consumelfare. The ops and risk papendent of tho
urnal of Insu
nal Association of
nt in Dollars
00 can only obtaset by the insurer.
he cap on the d
is:
eters 9.2937timal damagestive probabilitat while we docap, the numbot appear to bthe restriction mers’ rights, intimal social b
arameters. (Buose quantities,
rance Regul
Insurance Commi
ain the claim paym
damages award
76, 1.6271s cap * 408ty distribution o not know theber representsbe an overwheis depends on
nsurance compbenefit is, of cut, let us notas the optimal
lation
ssioners
ment of
d is the
3 into
844.85. fitted
e exact s only elming
n one’s panies’ course, te, the cap is
10
Optimal
© 2016 Natio
set withoutrandom va
Em
It is im
the expecteshows thatdeterminedappear in wclaims is aTherefore, insurers. Oexperiencin Example
The pmechanism
6. Elin
(1998) (whiInsurance Sgoodness-fitto fit the indupper limit othat the twosimilar. Howexample, whrelatively hig
Cap on Clai
onal Association o
t reference to tariable.)
mpirical FuncClaim Pa
mportant to noed value of insut the cap in td by us (aboutwhich the realas large as 217
high values oOn the contraryng the true cos
2: Medical M
professional liam for defending
g (2012) used thich was randomervices Office) t-test. Those resul
demnity loss. Theof policies is $509o estimates of thewever, the averaghich further illustgh.
im Settlemen
f Insurance Comm
them; it is deri
Figction of Cumuayment Fitted
te the contrastured loss and che real world t $41,000). In l loss is greate7. Thus, only of caps play l
y, high caps wists associated w
Malpractice
ability insurang medical malp
he indemnity los
mly chosen from to fit the empirlts show that the e expected indem9,598. Comparinge expected indemge upper limit oftrates that the upp
nts
missioners
ived only as th
gure 1:
ulative Probabd Using the Da
between the ccaps in reality.
($300,000) isfact, Table 1 r than $300,001.4% of claimlittle role in dll encourage cl
with these risks
Insurance
nce (PLI) systpractice claims
ss data of generalate settlement
ical distribution log-normal distr
mnity loss they estg the result from
mnity loss for gef policies for Elinper limits for gen
he mean of the
bility Distributata in Table 1
cap determinedThe example d
s much greateshows that on00, while the t
ms exceed the $decreasing clailaimants to taks, thus inviting
tem is the prs and indemnify
al liability from lags and were of the indemnit
ribution is one of timate is $41,210Eling (2012) to o
eneral liability inng is even largerneral liability insu
e economic los
tion of
d here based ondiscussed by u
er than the capnly three claimtotal number o$300,000 limitim payment byke risks withoumoral hazard.
rimary fundingfying successfu
Free and Valdeprovided by th
ty loss using thf best distribution0, and the averagour result, we finnsurance are quitr than that for ouurance in U.S. ar
ss
n us p
ms of t. y
ut 6
g ul
ez he he ns ge nd te ur re
11
negligmedicinsuramarkeKane insurathe opnumbin Co
* Sincenumbethis pointerna
SourceMarch
Fdistribdistribdistrib
1 statist
gence lawsuitscal liability insance crisis. Seets (e.g., Karl and Emmons,
ance from a diptimal cap wit
ber of claims bonnecticut from
Siz
e the range of $1 mr and percentages rtion of the data in
als.
e: Connecticut In2008
Figure 1 desbution (CDF) bution quite wbution by m
.4 7.57 We usetic STAT=0.26
s (Amon and surance by inseveral empiricand Nyce, 201, 2006). In thisfferent perspecth the empiricaby size of indemm the 4th quarte
ze of Indemnit
million and over iof claims falling i
nto four intervals a
nsurance Departm
cribes the eusing the datawell. We calc
maximum like
e the Kolmogor682. Since the
Jou
© 2016 Nation
Winn, 2004).surers may be cal papers stud14; Born et al.,s second examctive. We use al data from thmnity paymenter of 2005 thro
Table 2: ty Payments (
in the last line of Tin this interval. Weand assume the nu
ment, Connecticu
empirical funca in Table 2. Tculate the valelihood estim
rov-Smirnov te critical value
urnal of Insu
nal Association of
. Thus, large one importan
dy medical m, 2009; Lei and
mple, we study our theoretical
he “real world.t of medical m
ough 2007.
(2005 4th Q–20
Table 2 is so largee address this aspe
umber of claims un
ut Medical Malp
ction of cumThe data clearlues of param
mation and o
est and obtain for determini
rance Regul
Insurance Commi
claim paymennt reason for limalpractice insu
d Schimit, 200medical malpr
l model to dete” Table 2 show
malpractice insu
007)
e, it results in a veect of the data by dniformly falls in th
practice Annual
mulative probrly fit the logn
meters of lognobtain ˆ 12
the value of thng whether to
lation
ssioners
nts on iability urance
06; and ractice ermine ws the urance
ery large dividing
hese four
Report,
ability normal normal .1015,he test reject
12
Optimal
© 2016 Natio
the null hywe do not r
E
Thus,
damages c
We kn
on damageare relative$521,230, payment w
In Apfrom 2005 analysis hNorth Dakand it is Georgia, IdVirginia. Wthis may bin the U.Slosses to ob
Cap on Clai
onal Association o
ypothesis is 0.3reject the null
Empirical Prob
our theoreticalap should equa
now from Avrae for medical mely large. Settinwould reduce
with this cap is pendix Table to 2007. Com
ere is close tokota and Utah. higher than cdaho, Kansas,
We note that mbe one of the imS., along with btain higher ju
im Settlemen
f Insurance Comm
3912, which ishypothesis.
Figbability Distri
Using the D
l model calculaal:
aham (2014) thmalpractice insung the damagee the total claabout 24% of tA2, we list thparing the datao those set by The cap we c
caps set by thMissouri, Nev
more than one-hmportant reasoother reasons
udgment award
nts
missioners
s greater than
gure 2: ibution of ClaData in Table
ates that, based
hat the state of urance. Thus, t
e cap at the amim loss paid that without th
he caps of 20 a in Table A2, y five states: Fcalculate is lowhe states of Avada, Oklahomhalf of the U.S
ons related to tsuch as long-s.
the value of th
aim Payment F2
d on the data, th
Connecticut hthe total claimount suggestedby insurers. T
he cap. states set by sthe cap of $52Florida, Illinoi
wer than that seAlaska, Califorma, South CaroS. states did nothe medical matail liability an
he test statistic
Fitted
he optimal
has not set a capm payments herd by our modelThe total claim
state legislation21,230 from ouis, Mississippi
et by Marylandrnia, Coloradoolina and Wesot set caps, andalpractice crisind exaggerated
c,
p re l, m
n ur i, d, o, st d is d
13
SenMo
T
and
F
increa
assumequat
E
to the
oppos eq
distribvariabmodedistribthe clparam
7
of asymselect distribindivid
nsitivityodel of C
Taking the part
of the logno
From equationsasing function
mption that thetion (12) by equ
Equation (14) se parameter site is true whe
quals exactly bution obtaineble. Since lognels of claims dbutions, caps olaim distributiometer will do
7. The distributionmmetric, non-negmodels with two ution (Xiao, 2008dual claim distrib
y AnalysiClaims D
tial derivative o
ormal distributi
s of (12) and (1of both and
e claim loss fuation (13), we
shows that whe
is smaller th
en 1. In f , which is, le
ed by taking tnormal distrib
distributions,7 oon claims will on, while for mominate.
ns of loss often usgative and with faparameters, such
8). We have testebution in our exam
Jou
© 2016 Nation
is in theDistribut
of equation (11
ion used in that
13), we see thad parameter
follows the loge also obtain:
en 1, the han its sensitiv
fact, the ratio oet us recall, th
the natural logbution is a reaour work sugg be more sens
more stable cla
sed in actuarial scat tail, etc. Based h as: lognormal died that we can usemple.
urnal of Insu
nal Association of
e Lognortion
1) with respect
t model, we ob
at the optimal vrs of the loss d
gnormal distri
sensitivity of tvity to the par
of sensitivity tohe standard dev
garithm of theasonable modegests that for msitive to the voaims distributio
cience generally hon these characteistribution, Weibe lognormal distr
rance Regul
Insurance Commi
mal
to the paramet
btain:
value of the capdistribution und
ibution. By di
the cap on payrameter , an
o to sensitivviation of the n
e lognormal rael for many remore volatile olatility parameons, sensitivity
have the characteeristics, we usualull distribution, Gibution to describ
lation
ssioners
ters
p is an der the
ividing
yments nd the
vity to normal
andom ealistic claims eter of to the
eristics lly Gamma be the
14
Optimal Cap on Claim Settlements
© 2016 National Association of Insurance Commissioners
Implications and Conclusion This paper develops models to analyze both claimant and insurer behaviors
with caps on claim damage for third-party insurance. We establish the optimal strategy for regulators to set the cap to balance both the claimant and insurer’s benefits, considering different risks, based on the criterion of maximizing social benefit. The cap and the cost parameter can be used to control the claimant’s claim inflation. As there is information asymmetry for the amount of damage, the regulator and insurer hardly capture the actual loss amount, especially for the non-economic loss. Our analysis suggests that the optimal value of the cap, considering the maximization of social benefit, should be the expected value of random loss amount independent of the kind of probability distribution it follows.
For losses that are relatively easy to evaluate in terms of their real amount (e.g., economic losses), regulators may consider capping the damage awards on a case-by-case basis, following the principles of equitable justice. But the principle developed here is a simple and practical guide even for such case-by-case evaluations. We acknowledge that our paper utilizes a static equilibrium model, and this is undoubtedly its limitation. There are significant dynamic processes involved in consideration of incentives of all parties involved in legal damages caps. But, such dynamic processes require a more complex multi-period model, and must include incentives faced by the injured party, insurance company and political decision-makers. We consider our work to be a first step toward such a complex model, and creation of such a model is reserved for a far larger and longer-term research project.
Our calculation with empirical data indicates that the cap determined by us in general liability policies is much less than caps observed in the real world. Our other example for medical malpractice insurance indicates that the insurer's total claim payment is much higher without caps than that with caps determining based on the expected loss. The fact that many states have not set caps on damage might be one of the main reasons for the crisis of medical practice insurance in the U.S. We believe that our work may be a valuable contribution in the ongoing debate on managing the cost of claims and the issue of imposing a legal cap on them.
Because insurance regulators are focused on the solvency of insurers, regulators may tend to favor damage award caps in order to reduce the likelihood of insurer insolvency. Similarly, insurance regulators are concerned with insurance pricing, and regulators also may tend to favor caps on damage awards as a way to reduce the expected costs of insurance. Our analysis on optimal damage caps provides insurance regulators with results that shed light on optimum cap levels. The analysis, thus, can help insurance regulators to better achieve the goals of insurer solvency and consumer protection. Of course, others may also need to consider factors such as trying to limit frivolous litigation, decreasing moral hazard and holding parties responsible for their behavior, which might suggest increases in the level of caps or even the removal of caps on damage awards.
In future work, we hope to focus on the reasons that for some liability
15
Journal of Insurance Regulation
© 2016 National Association of Insurance Commissioners
insurance (e.g., general liability insurance) the value of optimal cap observed in reality, while for other types of liability insurance (e.g., medical malpractice), the value of the optimal cap is close to that observed in reality. Another further study may focus on determining optimal cap on damages and corresponding premium considering dynamic incentives of the potential injurers and corresponding injury phase of the environment.8
8. We thank an anonymous reviewer for this suggestion.
16
Optimal
© 2016 Natio
Techn 1. DerivaEquation
In ordwith equat
Setting
Note t
i.e., when tvalue obtaiequation (Abenefit. Fr
condition E
From
parameters
on the expequation (7
Cap on Clai
onal Association o
nical App
ation of the On (7)
der to maximizeion (7) and obt
g equation (A1
that since
the first derivained by solvingA2), we can orom equation (
E X * 0
the equation s of claim cost
pected value o7), we obtain th
im Settlemen
f Insurance Comm
pendix
Optimal Solu
e the quantity tain the follow
1) equal to zero
ative is zero, thg the equation btain the optim(A2), we see
0 so that
(A3), we see and settlement
of the loss dishe optimal soci
nts
missioners
ution to the
in equation (7)ing equation:
o, we obtain:
he second deriv(A2) will yield
mal cap of claithat the optim
that the cap t cost 1 and stribution. Subial benefit:
Problem Sta
), we use the L
vative is negatid a maximum. im * maximi
mal solution
of claim is un
2 , while it is
bstituting equa
ated in
Leibniz formul
ive, the optima By solving thizing the socia* satisfies th
unrelated to thonly dependen
ation (A3) into
a
al he al he
he nt
o
17
2. Ch(For
(
paymfuncti
(
H0 : F
cumu
paramof thvariab
(
hi-Square G this append
1) Use 2 goo
ment can be takeion is:
(2) Perform
F x F0 x,ulative distribut
meters. (They ahe normal disble.)
3) Use the dist
oodness of Fdix, we refer
dness of fit tes
en to originate
m a hypot
, and H
tion function o
are not the meatribution equa
tribution functi
Jou
© 2016 Nation
Fit Test for th to Xiao, 200
st and based on
from a lognor
thesis test
H1 : F x F0 of the lognorm
an and varianceal to the natu
ion chosen to c
urnal of Insu
nal Association of
he Lognorma08.)
n its outcome a
rmal distributio
with the
x,, , wh
mal distribution
e of the lognorural logarithm
calculate the ex
rance Regul
Insurance Commi
al Distributi
assuming that c
on. Its density
null hypo
here ( )F x i
n, and ,
rmal distributiom of the logn
xpected values
lation
ssioners
ion
claim
othesis
is the
are its
on, but normal
:
18
Optimal
© 2016 Natio
Where
likelihood
where ~Q
Q
2 r
reject the
distribution
(4) Caobtain: following r
and simila
are listed b
Cap on Clai
onal Association o
e n is sampl
estimation and
2~ ( 1)r k
k 1,
hypothesis o
n for the data.
alculate the val 9.29376, results are obta
arly E1
49.17below in Table
im Settlemen
f Insurance Comm
le size, and d consider the f
and k is the
f 0H and th
lues of parame1.62713.Usin
ained:
7, E2 27.00
A1.
nts
missioners
, are valu
following statis
number of unk
he distribution
ters by maximng the equation
, ..., etc. can a
ues obtained fr
stic:
known paramet
n assumed is
mum likelihood ns (A6), (A7)
also be obtain
from maximum
ters. If
not a suitabl
d estimation and) and (A8), th
ned. The result
m
le
d he
ts
19
The CCalculation Re
Jou
© 2016 Nation
Table A1:
esults of 2 G
urnal of Insu
nal Association of
Goodness of Fi
rance Regul
Insurance Commi
it Test
lation
ssioners
20
Optimal
© 2016 Natio
Since
Q 4
it is failed
takes the lclaims in th
The Val
9. The
(May 2014).
Cap on Clai
onal Association o
4.51 0.052 12
d to reject the
ognormal distrhis context.
lues of Caps oU.S. in 2
data in this table The states that a
im Settlemen
f Insurance Comm
21 16.92
hypothesis H0
ribution as a su
Taof Medical Ma2005 4th Q to
is selected from re not listed in th
nts
missioners
2,
,and thus we
uitable one for
ble A2: alpractice Insu
20079(Thousa
the database (in the table have not s
proceed with
r representing
urance for 20 Sand Dollars)
the clever file) ofset caps during th
the model tha
the amounts o
States in the
f Avraham, Ronehese three years.
at
of
en
21
Journal of Insurance Regulation
© 2016 National Association of Insurance Commissioners
References Amon, E., and H.N. Winn, 2004. “Review of the Professional Medical Liability
Insurance Crisis: Lesson from Missouri,” American Journal of Obstetrics and Gynecology, 190: 1534–1540.
Avraham, R., 2014. Database of State Tort Law Reforms (DSTLR 5th), University of Texas Law, Law and Econ Research Paper No.e555, Social Science Research Network, accessed at http://ssrn.com/abstract=902711.
Crocker, K.J., and J. Morgan, 1998. “Is Honesty the Best Policy? Curtailing Insurance Fraud Through Optimal Incentive Contracts,” Journal of Political Economy, 106: 355–375.
Crocker, K.J., and S. Tennyson, 2002. “Insurance Fraud and Optimal Claims Settlement Strategies,” Journal of Law and Economics, 45: 469–507.
Born, P.H., M.M. Boyer, and M.M. Barth, 2009. “Risk Retention Groups in Medical Malpractice Insurance: A Test of the National Chartering Option,” Journal of Insurance Regulation, 29: 3–33.
Datta, S., 2005. “Estimating the Mean Life Time Using Right Censored Data,” Statistical Methodology, 65–69.
Eling, M, 2012. “Fitting Insurance Claims to Skewed Distributions: Are the Skew-Normal and Skew-Student Good Models?” Insurance: Mathematics and Economics, 51: 239–248.
Encinosa, W.E., and F.J. Hellinger, 2005. “Have State Caps on Malpractice Awards Increased the Supply of Physicians?” Health Affairs: W5-317-W5-325.
Frees, E.W., and E.A. Valdez, 1998. “Understanding Relationships Using Copulas,” North American Actuarial Journal, 2: 1–25.
Feldman, H.L., 1996. “Harm and Money: Against the Insurance Theory of Tort Compensation,” Texas Law Review, 75: 1567.
Gfell, K.J., 2004, “Constitutional and Economic Implications of a National Cap on Non-Economic Damages in Medical Malpractice Actions,” The Indiana Law Review, 37: 773.
Howell, R. Commercial General Liability Insurance Definition, eHow, accessed at www.ehow.com/about_5184654_commercial-general-liability-insurance-definition.html.
Gaskin, D.J., and P. Richard, 2012. “The Economic Costs of Pain in the United States,” The Journal of Pain, 13: 715–724.
Gfell K.J., 2003. “Constitutional and Economic Implications of a National Cap on Non-Economic Damages in Medical Malpractice Actions,” Indiana Law Review, 37: 773.
Govtrack, 2004. See https://www.govtrack.us/congress/bills/108/hr4280. Kane, C.K., and D.W. Emmons, 2006. “The Impact of State Tort Reforms on Medical
Liability Payment,” Interim Report to The Physician Insurers Association of America, December.
Karl, J.B., and C. Nyce, 2014. “Determinants of Market Entry: Evidence from Medical Malpractice Insurance,” Journal of Insurance Regulation, 33: 1–27.
22
Optimal Cap on Claim Settlements
© 2016 National Association of Insurance Commissioners
Kesser, D. and M. McClellan, 1997. “The Effects of Malpractice Pressure and Liability Reforms on Physician Services,” Journal of the American Medical Association, 293: 2618–2625.
Klugman, S.A., H.H. Panjer, and G.E. Willmot, 1998. “Loss Model from Data to Decisions,” John Wiley & Sons.
Lei, Y. and J.T. Schmit, 2006. “Factors Influencing the Demand for Reinsurance in the Medical Malpractice Insurance Market: A Focus on Organization Form,” Journal of Insurance Regulation, 26: 1–8.
Liang, B.A., and L.L.Ren, 2004. “Medical Liability Insurance and Damage Caps: Getting Beyond Band-Aids to Substantive Systems Treatment to Improve Quality and Safety in Healthcare,” American Journal of Law & Medicine, 30: 501–541.
Mao, H. and K.M. Ostaszewski, 2007. “Application of Game Theory to Pricing of Participating Deferred Annuity,” Journal of Insurance Issues, 30: 102–122.
Nash, J.C., 1950. “The Bargaining Problem,” Econometrica, 18: 151–162. Paik, M., B.S. Black, D.A. Hyman, and C. Silver, 2012. “Will Tort Reform Bend the
Cost Curve? Evidence from Texas,” Journal of Empirical Legal Studies, 9: 173–216.
Shoben, E.W., 2005. “Let the Damages Fit the Wrong: An Immodest Proposal for Reforming Personal Injury Damages,” Akron Law Review: 39(4), Article 9, accessed at http://ideaexchange.uakron.edu/akronlawreview/vol39/iss4/9.
Sloan, F.A., P.M. Mergenhagen, and R.R. Bovbjerg, 1989. “Effects of Tort Reforms on the Value of Closed Medical Malpractice Claims: A Microanalysis,” Journal of Health Politics, Policy and Law, 14: 663–669.
S.A. Klugman, H.H.Panjer, and G.E. Willmot, 1998. Loss Models: From Data to Decisions, John Wiley & Sons.
Thomson, W., 1981. “Nash Bargaining Solution and Utilitarian Choice Rules,” Econometrica, 49: 535–538.
United States Congress, Congressional Budget Office, 1998. Preliminary Cost Estimate, H.R. 4250 Patient Protection Act of 1998, accessed at https://www.cbo.gov/sites/default/files/105th-congress-1997-1998/costestimate/ hr425000.pdf.
Viscusi, W.K., and P. Born, 2005. “Damage Caps, Insurability and the Performance of Medical Malpractice Insurance,” Journal of Risk and Insurance, 72: 23–43.
Xiao, Z.Y, 2008. Risk Theory, (in Chinese), Publishing House of People’s University, Beijing, China.
Zuckerman, S., R.R. Bovbjerg, and F. Sloan, 1990. “Effects of Tort Reform and Other Factors on Medical Malpractice Insurance Premiums,” Inquiry, 167–182.
23
Journal of Insurance Regulation
Guidelines for Authors
Submissions should relate to the regulation of insurance. They may include
empirical work, theory, and institutional or policy analysis. We seek papers that advance research or analytical techniques, particularly papers that make new research more understandable to regulators.
Submissions must be original work and not being considered for publication elsewhere; papers from presentations should note the meeting. Discussion, opinions, and controversial matters are welcome, provided the paper clearly documents the sources of information and distinguishes opinions or judgment from empirical or factual information. The paper should recognize contrary views, rebuttals, and opposing positions.
References to published literature should be inserted into the text using the “author, date” format. Examples are: (1) “Manders et al. (1994) have shown. . .” and (2) “Interstate compacts have been researched extensively (Manders et al., 1994).” Cited literature should be shown in a “References” section, containing an alphabetical list of authors as shown below.
Cummins, J. David and Richard A. Derrig, eds., 1989. Financial Models of
Insurance Solvency, Norwell, Mass.: Kluwer Academic Publishers. Manders, John M., Therese M. Vaughan and Robert H. Myers, Jr., 1994.
“Insurance Regulation in the Public Interest: Where Do We Go from Here?” Journal of Insurance Regulation, 12: 285.
National Association of Insurance Commissioners, 1992. An Update of the NAIC
Solvency Agenda, Jan. 7, Kansas City, Mo.: NAIC. “Spreading Disaster Risk,” 1994. Business Insurance, Feb. 28, p. 1.
Footnotes should be used to supply useful background or technical information that might distract or disinterest the general readership of insurance professionals. Footnotes should not simply cite published literature — use instead the “author, date” format above.
Tables and charts should be used only if needed to directly support the thesis of the paper. They should have descriptive titles and helpful explanatory notes included at the foot of the exhibit.
Journal of Insurance Regulation
Papers, including exhibits and appendices, should be limited to 45 double-spaced pages. Manuscripts are sent to reviewers anonymously; author(s) and affiliation(s) should appear only on a separate title page. The first page should include an abstract of no more than 200 words. Manuscripts should be sent by email in a Microsoft Word file to:
Cassandra Cole and Kathleen McCullough [email protected]
The first named author will receive acknowledgement of receipt and the
editor’s decision on whether the document will be accepted for further review. If declined for review, the manuscript will be destroyed. For reviewed manuscripts, the process will generally be completed and the first named author notified in eight to 10 weeks of receipt.
Published papers will become the copyrighted property of the Journal of Insurance Regulation. It is the author’s responsibility to secure permission to reprint copyrighted material contained in the manuscript and make the proper acknowledgement.
NAIC publications are subject to copyright protection. If you would like to reprint an NAIC publication, please submit a request for permission via the NAIC Web site at www.naic.org. (Click on the “Copyright & Reprint Info” link at the bottom of the home page.) The NAIC will review your request.