Rei , Simon; Schube , Sab ina; Wambach, Achim
A icle — Published Ve sion
Se ing incen i es igh wi h long- e m isk adjus men
The Eu opean Jou nal o Heal h Economics
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Rei , Simon; Schube , Sab ina; Wambach, Achim (2024) : Se ing incen i es
igh wi h long- e m isk adjus men , The Eu opean Jou nal o Heal h Economics, ISSN 1618-7601,
Sp inge , Be lin, Heidelbe g, Vol. 26, Iss. 2, pp. 147-151,
h ps://doi.o g/10.1007/s10198-024-01751-6
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323290
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
EDITORIAL
The Eu opean Jou nal o Heal h Economics (2025) 26:147–151
h ps://doi.o g/10.1007/s10198-024-01751-6
ha e s a ed wi h baseline se ups whe e only a small se o
a iables such as egion, gende and age a e used as isk
adjus e s [8]. Today, mos schemes a e mo e elabo a ed
and in pa icula also ake mo bidi y da a in o accoun .
These ad ances ha e inc eased he accu acy o de e mining
expec ed heal h cos s. The ocus o many academic con i-
bu ions in he a ea o isk adjus men has been he echnical
design o op imize he accu acy o isk adjus ed paymen s
using classical s a is ics [5, 7, 10, 13], and mo e ecen ly
machine lea ning [3, 4, 9, 14].
Incen i es o in es ing in en ollee heal h
Risk adjus men , which educes incen i es o he insu ance
o disc imina e agains high-cos indi iduals, mimics he
p emiums in a compe i i e insu ance ma ke : The o e all
paymen s ecei ed by insu ances, i.e. he expec ed cos s o
he insu ed, coincide wi h he ai p emium o one-yea
insu ance con ac s. Howe e , Eggles on e al. [1] show in
a heo e ical model ha isk adjus men educes in es men
in p e en i e ac i i ies. The in ui ion behind he model is
ha when insu ances ecei e expec ed heal hca e cos s es i-
ma ed om mo bidi y da a, hey lack incen i es o mi iga e
mo bidi y ela ed cos inc eases h ough (cos ly) p e en ion.
The gene al pa e n ha insu ances in es in he heal h
o hei en ollees only when i pays o inancially can be
obse ed in he p i a e employe -sponso ed US heal h
insu ance ma ke wi hou isk adjus men . An empi ical
example highligh ing he nega i e e ec o isk adjus men
on p e en ion comes om a e o m o he Medica e Ad an-
age scheme in he US. Lissenden & Balk ishnan [6] show
ha when diagnoses (including hose p e en able by ac-
cina ions) a e added o isk adjus men , accina ion a es go
down. Weinhold e al. [15] show ha in he Ge man social
heal h insu ance ma ke , expendi u es on p e en ion esul
in a ne loss o insu ances. When conside ing he change
in isk adjus men paymen o he hen heal hie indi idual
in his se ing, he cos s o in es ing in success ul p e en-
ion ou weigh he sa ings. In hei analysis o US heal h
Cu en isk adjus men schemes in social heal h
insu ance p o ide li le incen i es o heal h insu -
e s o engage in in es men s in en ollee heal h and
e iciency enhancing inno a ions. We p opose a long-
e m isk adjus men scheme, which s ill p e en s isk
selec ion, bu makes i inancially a ac i e o insu -
e s o ca e abou long- e m heal h o hei en ollees
and inc eases e iciency.
The ole o isk adjus men in social heal h
insu ance
A key challenge o heal h policy is o o e a o dable
heal h insu ance p emiums o a popula ion cha ac e ized by
la ge he e ogenei ies in isks and cos s. Many heal h sys-
ems ha e decided no o se p emiums based on he indi id-
ual en ollee isk, bu ins ead aim o p emium equaliza ion
ac oss he popula ion. This equaliza ion c ea es incen i es
o selec low- isk (and he e o e low cos ) indi iduals. To
de e insu ances om disc imina ing agains indi iduals
wi h high expec ed heal hca e cos s, many social heal h
insu ance sys ems (e.g. Belgium, Ge many, Swi ze land,
he Ne he lands o he US Medica e and Medicaid) ha e
in oduced mo bidi y-based isk adjus men schemes. Unde
hese schemes, insu ances ecei e he expec ed heal hca e
cos s o each insu ed indi idual.
In he ideal case, he isk adjus men scheme makes e e y
en ollee equally a ac i e o he insu ance, allowing hem
o ocus on cos managemen and se ice quali y ins ead
o selec ing good isks. Mos isk adjus men schemes
Simon Rei
simon. ei @zew.de
1 ZEW – Cen e o Eu opean Economic Resea ch,
L7 1, 68161 Mannheim, Ge many
2 FAU E langen-Nü nbe g, Nü nbe g, Ge many
3 Uni e si y o Mannheim, Mannheim, Ge many
Accep ed: 23 Oc obe 2024 / Published online: 20 Decembe 2024
© The Au ho (s) 2024
Se ing incen i es igh wi h long- e m isk adjus men
SimonRei 1,2,3· Sab inaSchube 1,2· AchimWambach3
1 3
S. Rei e al.
insu ances, Fang & Ga azza [2] do howe e show ha
heal hca e in es men is highe when he expec ed enu e
o he employee (a he i m and he e o e also a he heal h
insu ance) is longe . This indica es ha – gi en he igh
long- e m incen i es – insu ances a e willing o in es in
long- e m heal h.
While empi ical e idence on nega i e incen i es o
heal h insu ances ocuses on p e en i e ac i i ies, he gen-
e al pa e n also applies o o he se ings. Decisions ega d-
ing in es men in en ollee heal h also ma e when cu a i e
he apies (e.g. Di ec ly Ac ing An i i als o Hepa i is C o
CFTR modula o s o Cys ic Fib osis) a e mo e a o able
in he long un han main enance he apies. O he examples
conce n he in es men in cos ly in as uc u e (e.g. pa ien
moni o ing o ehabili a ion acili ies) which inc ease imme-
dia e cos s and will only yield bene i s by educed expec ed
cos s in he u u e.
A simple example o incen i es in isk
adjus men schemes
As explained abo e, he inancial s eam o he heal h insu -
ance ollowing isk adjus men schemes ollows he same
p inciples as a one-yea con ac in a compe i i e insu ance
ma ke : Each insu ance ecei es he expec ed expendi-
u es o hei en ollees o he ollowing yea . These annual
paymen s acili a e e icien esou ce managemen o he
upcoming yea . Howe e , annual paymen s om he isk
adjus men scheme imply ha insu ances ha e a inancial
incen i e o in es in en ollee heal h only i he in es men
pays o wi hin a yea .
Figu e 1 illus a es hese incen i es. The calcula ions
a e based on he a e age expendi u es in he Ge man social
heal h insu ance sys em om he yea 2021. We assume
ha he alloca ion and he expec ed expendi u es o a
40-yea -old man a e 1,720 eu os in he i s yea . The allo-
ca ion in he ollowing yea s inc eases due o wo ac o s:
Fi s , an assumed heal h cos in la ion a e o 3% pe yea
(o e all in la ion is igno ed o make paymen s compa a-
ble ac oss ime). Second, he expec ed cos s in he second
yea a e hose o a 41-yea -old, in he hi d yea hose o a
42-yea -old, e c. The g een line depic s his cos p og es-
sion, along wi h he co esponding annual paymen s om
he isk adjus men scheme o he insu ance. The blue line
ep esen s a po en ial expendi u e cu e when he insu -
ance inances an in e en ion in T = 1 ha cos s wice he
expec ed expendi u es in he in e en ion yea compa ed o
no mal ea men , bu sa es 30% o he o iginally expec ed
cos s in he long un. No e, ha hese cos sa ings will be
e lec ed in he annual paymen o he insu ance, shown by
he do ed black line: The in e en ion imp o ed he heal h
o en ollee and he isk adjus men scheme educes he pay-
men s o he insu ance. Wi hou he in e en ion, he insu -
ance ecei es o al paymen s o 23,211 eu os o e en yea s
and incu s expendi u es o he same amoun . I he in e en-
ion is inanced, o al cos s o e a en-yea ho izon a e lowe
a 19,132 eu os (due o educed expendi u es om yea 2
onwa ds), bu he o al paymen s om he scheme a e only
17,896 eu os (since he expec ed cos s and he e o e pay-
men s o he insu ance a e educed om yea 3 onwa ds).
The e o e, he insu ance has no incen i e o implemen
measu es ha only yield bene i s in subsequen yea s.
P oposed measu es o incen i ize
p e en ion unde isk adjus men
The lack o incen i es o p e en ion has ini ia ed a deba e
on possible heal h sys em e o ms. Cu en ly, he e a e h ee
app oaches o modi y he incen i e s uc u e o in es ing
in en ollee heal h in heal h insu ance sys ems. Fi s , one
op ion is e e ing o simple isk adjus men (pa ly) igno -
ing diagnoses: While his would inc ease he incen i e o
in es in p e en ion, i can also lead o s onge incen i es
o isk selec ion based on diagnoses. Second, policymake s
could manda e insu ances o in es di ec ly in p e en ion.
This app oach is o example applied in he Ge man social
heal h insu ance. While such manda es inc ease p e en i e
ac i i ies, hey a e unlikely o esul in op imal p e en ion
as he e is s ill no incen i e o p o ide he mos e ec i e
in e en ions. Addi ionally, his s a egy does no conside
u he in es men s in en ollee heal h, as discussed abo e.
A hi d way p oposed by Eggles on e al. [1], is o apply
pay- o -pe o mance o p e en i e ac i i ies. The idea is o
inancially ewa d insu ances when he heal h s a e o hei
en ollees de eloped be e han p edic ed, se ing a di ec
incen i e o in es in hei heal h. While such a scheme
would add ess he incen i es o p e en ion, ope a ionaliz-
ing i is challenging due o he possibili y o en ollees ha -
ing e y speci ic combina ions o diagnoses ha in e ac in
hei e ec on heal h s a us.
New p oposal: Se ing long- e m incen i es
in isk adjus men
We p opose a no el app oach o ewa d insu ances o suc-
cess ul in es men in en ollee heal h. Building on he idea
om Eggles on e al. [1], we p opose o b eak he nexus
be ween heal h imp o emen s and lowe alloca ions. We
sugges ha longe ime pe iods in he isk adjus men dis-
ibu ion mechanism make in es men s in p e en ion and
inno a ion a ac i e o heal h insu ances. The long- e m
1 3
148
Se ing incen i es igh wi h long- e m isk adjus men
na u e can be achie ed by changing he ime ame o he
paymen calcula ions. Ins ead o annual calcula ions, he
expec ed expendi u es o he en ollees should be compu ed
o he ollowing en yea s based on cu en en ollee cha -
ac e is ics. Each yea , he insu ance would hen ecei e he
ini ially calcula ed annual amoun o he en ollee o he
espec i e yea . F om an insu ance s andpoin , his co e-
sponds o a shi om a one-yea -con ac owa ds a en-yea
con ac . Th ough his longe - e m calcula ion o paymen s,
heal h insu ances can bene i om cos sa ings esul ing
om imp o ed ea men . A heal hie en ollee incu s lowe
expendi u es in he ollowing yea s, while he paymen s o
he insu ance emain unchanged – he eby, inancially ben-
e i ing he insu ance. On he o he side, a sicke en ollee
causes highe expenses han he paymen s om he scheme,
causing a inancial loss. The insu ance is he e o e incen i -
ized o in es in he long- e m heal h o i s en ollees. A he
same ime, he en-yea calcula ion o alloca ions ensu es
ha , on a e age, an insu ance does no incu losses. The
alloca ions exac ly co espond o he “ ai p emium” ha
would be demanded in a compe i i e ma ke wi h en-yea
con ac s. The mechanism o such as scheme is shown in
Fig. 2.
In con as o Fig. 1, paymen s om he isk adjus -
men schemes a e no adjus ed a e a success ul in e en-
ion. The e o e, he insu ance now has o al expendi u es
o 19,132 eu os o e he en-yea pe iod bu ecei es a
o al a su plus o 4,080 eu os since he cos sa ings emain
Fig. 2 Illus a ion o en-yea isk adjus men scheme
Fig. 1 Illus a ion o one-yea isk adjus men scheme
1 3
149
S. Rei e al.
heal h bene i s and inancial bene i s migh di e ge. To
a oid p oblema ic incen i es (beyond he classic isk
adjus men scheme), he calcula ion pe iod o he p o-
posed isk adjus men scheme could be educed om
en o i e (one) yea s o indi iduals olde han 70 (80).
(i ) Challenge o long- e m cos p edic ion Much o he
ecen echnical li e a u e on isk adjus men ocuses on
me hods o imp o e he p edic ions o he isk adjus -
men models. While p ecisely p edic ing heal h ca e
cos s is challenging o he cu en ly common one-yea
schemes, ex ending he p edic ion pe iod will cp. wo sen
he i . Howe e , as long as he p edic ion e o is no
co ela ed wi h in o ma ion obse able o he insu ance,
a en-yea p edic ion pe iod does no inc ease selec ion
incen i es. Be e da a and mo e ad anced s a is ical
models can also alle ia e his conce n in he u u e. The
h eads om p edic ion e o can be u he educed
by sepa a e inancing mechanisms o ex eme-cos
pa ien s and he adjus men o he calcula ion pe iod o
ewe yea s o olde indi iduals which ha e a highe
a iance in cos s.
Conclusion
While mo bidi y-based isk adjus men s ongly dec eases
he incen i e o insu ance o disc imina e agains en ollees
wi h high expec ed cos s, he cu en one-yea ime ame
o paymen calcula ion and alloca ion binds insu ances
o sho - e m decision-making ega ding en ollee heal h
in es men s. Consequen ly, insu ances ha e weak incen-
i es o in es in p e en ion and in inno a i e echnologies,
especially cos ly in e en ions ha only inancially pay o
in la e yea s. We p opose ex ending he isk adjus men
alloca ion pe iod o span o e en yea s, enabling insu ances
o bene i om in es men in en ollee heal h. In pa icula
in inc easingly common managed ca e plans, insu ances can
be e use long- e m in es men s o bene i hei inances,
en ollee heal h and ul ima ely inc ease he inancial sus ain-
abili y o heal hca e sys ems.
Funding Open Access unding enabled and o ganized by P ojek
DEAL.
Open Access This a icle is licensed unde a C ea i e Commons
A ibu ion 4.0 In e na ional License, which pe mi s use, sha ing,
adap a ion, dis ibu ion and ep oduc ion in any medium o o ma ,
as long as you gi e app op ia e c edi o he o iginal au ho (s) and he
sou ce, p o ide a link o he C ea i e Commons licence, and indica e
i changes we e made. The images o o he hi d pa y ma e ial in his
a icle a e included in he a icle’s C ea i e Commons licence, unless
indica ed o he wise in a c edi line o he ma e ial. I ma e ial is no
included in he a icle’s C ea i e Commons licence and you in ended
wi h he insu ance. This su plus comes wi h wo posi i e
e ec s. Fi s ly, he su plus can be used o implemen ca e
inno a ions ha u he educe u u e heal h expendi u es.
Secondly, he alloca ions, based on a e age cos s o all
insu ances, dec ease – he eby educing he o e all cos o
he heal hca e sys em.
P ac ical conside a ions o a long- e m isk
adjus men scheme
(i) Swi ching insu ance One big ad an age o he cu -
en isk-based adjus men scheme is ha i allows
en ollees o swi ch insu ance on an annual basis. This
ad an age emains wi h he long- e m isk adjus men
scheme, i swi ches happen acco ding o he ollowing
ules. Suppose an en ollee decides o swi ch om insu -
ance A o insu ance B a e h ee yea s, so he ini ial
isk assessmen a insu ance A was made h ee yea s
ago. Upon becoming a policyholde o insu ance B,
she will ecei e a new en-yea assessmen o ming he
basis o he paymen s o insu ance B. I he expec ed
cos s ha e inc eased in his new assessmen , he depa -
ing insu ance A mus compensa e he di e ence o
he isk adjus men scheme o he yea s he en ollee
would ha e o he wise emained wi h i , in he example
o se en yea s. I he expec ed cos s ha e meanwhile
allen, insu ance A ecei es he di e ence o hese
se en yea s om he scheme. Thus, he swi ch o an
en ollee o ano he insu ance is in expec a ion inan-
cially neu al o bo h he depa ing insu ance A and he
ecei ing insu ance B.
(ii) High-cos pa ien s The eassessmen o isk o
indi iduals swi ching be ween insu ances also alle i-
a es he h ea o disc imina ion agains indi iduals wi h
ad e se heal h e en s in he i s yea s o he en-yea
pe iod. Al hough such an indi idual is hen loss-making
o he insu ance, he insu ance mus decide be ween
co e ing he cos s o ea men o compensa e he
loss a e eassessmen when he indi idual swi ches
insu ance. Concep ually, such a disc imina ion isk
can be u he educed wi h dedica ed mechanisms o
ex eme-cos pa ien s such as isk sha ing [12] o high-
isk pooling [11].
(iii) End o li e The p oposed mechanism elies on he
assump ion ha en ollees wi h be e heal h a e less
cos ly, as in es men s in heal h lead o lowe expendi-
u es in he u u e. Howe e , his assump ion may no
always hold ue. Pa icula ly o end-o -li e spending,
1 3
150
Se ing incen i es igh wi h long- e m isk adjus men
8 Rice, N., Smi h, P.C.: Capi a ion and Risk Adjus men in Heal h
Ca e Financing: An In e na ional P og ess Repo . Milbank Q.
79(1), 81–113 (2001). h ps:/ /doi.o g/10.11 11/14 68-0009.00197
9 Rose, S.: A machine lea ning F amewo k o Plan Paymen Risk
Adjus men . Heal h Se . Res. 51(6), 2358–2374 (2016). h p s : / / d
o i . o g / 1 0 . 1 1 1 1 / 1 4 7 5 - 6 7 7 3 . 1 2 4 6 4
10 Schokkae , E., Van De Voo de, C.: Risk selec ion and he speci-
ica ion o he con en ional isk adjus men o mula. J. Heal h.
Econ. 23(6), 1237–1259 (2004). h p s : / / d o i . o g / 1 0 . 1 0 1 6 / S 0 1 6 7 - 6
2 9 6 ( 0 3 ) 0 0 0 4 0 - 7
11 Van Ba ne eld, E.M., Lame s, L.M., Van Vlie , R.C.J.A., Van
De Ven, W.P.M.M.: Manda o y pooling as a supplemen o isk-
adjus ed capi a ion paymen s in a compe i i e heal h insu ance
ma ke . Soc. Sci. Med. 47(2), 223–232 (1998). h p s : / / d o i . o g / 1 0 .
1 0 1 6 / S 0 2 7 7 - 9 5 3 6 ( 9 8 ) 0 0 0 5 6 - 2
12 Van Ba ne eld, E.M., Lame s, L.M., Van Vlie , R.C.J.A., Van De
Ven, W.P.M.M.: Risk sha ing as a supplemen o impe ec capi a-
ion: A adeo be ween selec ion and e iciency. J. Heal h. Econ.
20(2), 147–168 (2001). h p s : / / d o i . o g / 1 0 . 1 0 1 6 / S 0 1 6 7 - 6 2 9 6 ( 0 0 ) 0
0 0 7 7 - 1
13 Van Klee , R.C., McGui e, T.G., Van Vlie , R.C.J.A., Van De Ven,
W.P.P.M.: Imp o ing isk equaliza ion wi h cons ained eg es-
sion. Eu . J. Heal h Econ. 18(9), 1137–1156 (2017). h p s : / / d o i . o
g / 1 0 . 1 0 0 7 / s 1 0 1 9 8 - 0 1 6 - 0 8 5 9 - 1
14 Van Veen, S.H.C.M., Van Klee , R.C., Van De Ven, W.P.M.M.,
Van Vlie , R.C.J.A.: Explo ing he p edic i e powe o in e ac ion
e ms in a sophis ica ed isk equaliza ion model using eg ession
ees. Heal h Econ. 27(2) (2018). h p s : / / d o i . o g / 1 0 . 1 0 0 2 / h e c . 3 5 2
3
15 Weinhold, I., Schindle , C., Kossack, N., Be nd , B., Häckl, D.:
Economic impac o disease p e en ion in a mo bidi y-based
inancing sys em: Does p e en ion pay o o a s a u o y heal h
insu ance und in Ge many? Eu . J. Heal h Econ. 20(8), 1181–
1193 (2019). h ps: //d oi. o g/10 .100 7/ s10 198-0 19-01086-7
Publishe ’s no e Sp inge Na u e emains neu al wi h ega d o ju is-
dic ional claims in published maps and ins i u ional a ilia ions.
use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed
use, you will need o ob ain pe mission di ec ly om he copy igh
holde . To iew a copy o his licence, isi h p : / / c e a i e c o m m o n s . o
g / l i c e n s e s / b y / 4 . 0 / .
Re e ences
1 Eggles on, K., Ellis, R.P., Lu, M.: Risk adjus men and p e en-
ion. Can. J. Economics/Re ue Canadienne d’économique.
45(4), 1586–1607 (2012). h ps:/ /doi.o g/10.11 11/j.
1540-5982.2012.01747.x
2 Fang, H., Ga azza, A.: Dynamic ine iciencies in an employmen -
based Heal h Insu ance Sys em: Theo y and e idence. Am. Econ.
Re . 101(7), 3047–3077 (2011). h p s : / / d o i . o g / 1 0 . 1 2 5 7 / a e . 1 0 1 .
7 . 3 0 4 7
3 Hols e , T., Ji, S., Ma inen, P.: Risk adjus men o egional
heal hca e unding alloca ions wi h ensemble me hods: An
empi ical s udy and in e p e a ion. Eu . J. Heal h Econ. (2024).
h ps:/ /doi.o g/10.10 07/s1 0198-023-01656-w
4 Ismail, I., S am, P.J.A., Po ai , F.R.M., Van Wi eloos uijn, A.,
Koolman, X.: Add essing unan icipa ed in e ac ions in isk equal-
iza ion: A machine lea ning app oach o modeling medical expen-
di u e isk. Econ. Model. 130, 106564 (2024). h p s : / / d o i . o g / 1 0 . 1
0 1 6 / j . e c o n m o d . 2 0 2 3 . 1 0 6 5 6 4
5 Lay on, T.J., McGui e, T.G., Van Klee , R.C.: De i ing isk
adjus men paymen weigh s o maximize e iciency o heal h
insu ance ma ke s. J. Heal h. Econ. 61, 93–110 (2018). h p s : / / d o
i . o g / 1 0 . 1 0 1 6 / j . j h e a l e c o . 2 0 1 8 . 0 7 . 0 0 1
6 Lissenden, B., Balk ishnan, R.: Does Risk Adjus men Reduce
Vaccina ion in he Elde ly? E idence om Medica e ad an age.
Med. Ca e Res. Re . 77(2), 176–186 (2020). h p s : / / d o i . o g / 1 0 . 1 1
7 7 / 1 0 7 7 5 5 8 7 1 8 7 8 5 5 5 9
7 McGui e, T., Zink, A., Rose, S.: Simpli ying and Imp o ing he
Pe o mance o Risk Adjus men Sys ems (w26736; S. w26736).
Na ional Bu eau o Economic Resea ch. (2020). h p s : / / d o i . o g / 1
0 . 3 3 8 6 / w 2 6 7 3 6
1 3
151
