PHN-8722
PBN_ Tecffnica, aNrte-z
FINANCING HEALTH SERVICES: SOME SPECULATIONS
AND AN EXAMPLE FROM INDONESIA
by
Thomas Helminiak, Ph.D.
Ralph Andreano, Ph.D.
December 1987
PoDtlation and Hiumn Resources DeDartment
World Bank
The World Bank does not accept responsibility for the views expressed herein
which are those of the author's) and should not be attri.but.ed to the World
Bank or to its affiliated organizations. The findings, intervretations, anj
conclusions are the results of research supported by the Bank; they do not
necessarily represent official policy of the Bank. The designations
employed, the presentation of imtaterial, and any maps used in L`is documient
are solely for the convenience of the reader and do not imply the expression
of* any opinion whatsoever on the part of the Wor]d Bank or its affiliates
concerning the legal status of any country, territory, city area, or of its
authorities, or concerning the delimitations of its boundaries, or national
affiliation.



PHN Technical Note 87-22
FINANCING HEALTH SERVICES: SOME SPECULATIONS
AND AN EXAMPLE FROM INDONESIA
A B S T R A C T
The pE-er is intended to supplement the existing literature on health
service finaincin  schemes and the choice of government, versus privae sector
roviso S O' h     tii services. While this iterature lns now           laroe
;i    no   a       t ýnd the cmparative and analytical commnle; i
liru     n  es   rv  fl flIor- makitig generliations useful to po
1":UG i     'i iioi.' Ur  interit 10n  Ln  t-sil  paper  iS  t'v p ' 1  1
he I ~issuets frm ithe perspective of traditionai economics, namejY
efficienc  and equity. The authors also comment on some observed trends
throughout the world regarding financing schemes and public-private delivery
st lruc ture. This speculajton and broad review is parti.ally reinrced wich
some fíeid observat.'ons on the same issues from Indonesia. fr    rd
possible 1cissons for poi icy formulation, the need is empaifzed for miorc
systematic measurements, on a comparable basis, of the experieice Of a
number of countries
* * * * * * * *
Prepared by:    Thomas Hfelminiak, Ph. D. and Ra iph Andreano, hi .
U. ni.vers t1v  of' Wiscons;n
December LJ87



Overview
This paper is intended to supplement the existing literature on financing
scheme and their differential impacts on the production of health services.
This literature has now become large but it is not easy to find the kind of
comparative and analytical commonality in it that makes for generalizations
useful for policy. Our own intention in this paper is modest: we want to
speculate, in ways the current literature makes difficult to do, about the
financing system issues from the perspective of traditional economics, namely
efficiency and equity. This speculation is partially reinforced with some
field observations on the same issues from Indonesia. We want to emphasize
that a single case does not constitute a definitive piece of evidence on issues
of such complexity. Yet it is our hope that as research by economists on
health care finance issues develops our speculations and applications to a
major country can help guide such research. For policy formulation some
generalization from collective country experience will be necessary; to date,
intercountry comparisons on issues of efficiency and equity are hampered by
a coherent analytical framework as much as by a lack of comparable and usable
quantitative information.



General Cverview
1.0       Introduction: Brief Survey of Financing Systems and Delivery Structures
2.0       Issues in Financing and Delivery Organization Choice
2.1   Alternative Arrangements
2.2   Evaluation Criteria
(i) Aggregate collection
(ii) Funding efficiency
(iii) Funding incidence
(iv) Efficiency of health services production
(v) Distributional equity
(vi) Utilization efficiency
2.3   Summary
3.0       General Options for Countries
3.1   Government vs. Private Provision
3-2   FFS Financing
3.3   Government Tax Revenues
3.4   Commercial Health Insurance
3.5   Private Firm Arrangements
3.6   Social Health Insurance
3.7   Community and Cooperative Schemes
3.8   Summary
4.0       Indonesian Case Study
4.1   General Introduction
4.2   Government vs. Private Sector Provision
4.3   FFS Financing


4.0       Indonesian Case Study (continued)
4.4   Government General Revenues
4.5   Private Health Insurance
(i) Introduction
(ii) Commercial Third Party
(iii) RMO's
(iv) Capitation - Outpatient
4.6   Social Health Insurance for Civil Servants (ASKES)
4.7   Community Financing
4.8   Cooperative Arrangements
4.9   Private Firm Schemes
a.10 DUKTM
4.11 Summary
5.0       Concluding Notes


1.0 Introduction
It is increasingly recognized that the given (pattern of) health care
financing system(s) of a country, as well as the mix of public-private authority
in the provision of services, significantly impact on numerous factors governing
the aggregate production of health. These include the magnitude of total
resources devoted to health services, particular delivery mechanisms, the
health problems focused on, the incidence of resource funding and services
distribution and various other aspects of efficiency concerning thE conversion
of contributed resources into improved health status.
A variety of elements account for the specific financing systEms and delivery
structures found in individual countries. These may include the following (as a
suggestive, not necessarily definitive, listing) with varying emphases depending
on the country:
a)   historical inertia - possibly representing decisions c: previous
authorities, which may not well represent the 'references of
current authorities; but the pattern persists beca,se of :ercei:ed
difficulty of effect:ng a change, insufficient convizziCn o7
consensus on what the change should be, etc.
1)   political pressure grou;s associated with the sausue. 7e
a:ay include provider grcups benefiting from the c
and concezued tha: another pattern may result in eccncmic narm.
c)   socio-cultural factors - influencing actitudes about health and
appropriate interventions.
d)   political dogma - broad dictates about permissable arriagements.


e)   level of econom    developmen: -   cludin   evelopnta     attributes
such as levels cf personaliA
-)   act.ve political interpretatior.   the per:eived delivery of
aggregate, effi,-c-                             able 'ro  a
given financing/structure pattern, relative to other pocssible
Patterns, according to seme exression of national goals.
Economists, of course, may hope to Cflr :orthwhile assistance in t,e last
ca:eory of dterminancs, witn     .e c4nstrints imosec by :h e ce       c,aegr ie,
as wel as the constraints of available data and theory.   Depending on the
strength and conviczion of the analytical advice otfered bv econcmists, t-ere
may be opportunities for relaxing the constraining influence of some of the
other determinants. Further, certain of the other determinant categories may be
expected to change over time. An obvious candidate here is that of the level of
economic development. It may be hypothesized that with changes in the level of
economeic development (and with changes too -- concomitantly or independently --
t.e other categories) natural evolutionary changes in financing pattern and struc-
ture may also tend to occur in countries.  Particularly insofar as these natural
evolutionary changes might not precisely conform to changes that would be suggested
y the final category of active interpretation of efficiency and equity criteria,
then the ability to predict these changes and tieir implications would be of va!Ue
-ere again, economists might be useful in attempting through available theorv,
assisted by country experiences, to form generalizable principles of change.


-3-
Two researcaers Iave rioneered tIe 7iterature on heal.r, care fInanC-.-2
rends:   rian Abe.-S it, egina ni t. t::e 1.6 's, has reviewed iernaz-.na
patterns of realth SeCt3r e-=e ndiron Roe=er, over Several decadesl
numercus stuaes jE a           ealth care Syste=S    a,.d 1s-
generalizations a'out dinacin   sy,scem  rCnds,
ne ee--S:ih e:-edi-.:re studv, which was cnduced du=i~ a :en year
7er'oc anc iunsea      1:967 ,..uce    :-:i ceveloped and devel.=ping cur.
:ough haperc -v cons aera-e urcble=s cr deritiona! comDaraD-4
[3
available iate   ana   . data, the st:udy suggested thaa:   (1) :here was a   ry
=ugi posi Ive ccrret---a  n zezwee.. per cap--  -cc.e and t'l
[4]
nat-onal -ncome spent or. '-ealtn;          positiv1e correlaion peeeen   r
c a Di ta i - c  e an. tre :er za nr-a v-e cZ na nal zvm  s -jen:  :ar c esc cm
-a                           -ee
Mo r   .aua) w as    na uaec        tazre rltcrion c, : ei.:'r *ne  o efen, consrEez
a esendr                                                 2oc-s                 -
-   sl=re  of  t  ue      .. '   c d  to  .e   '.1  f C
seen ro depend on the level o   -:aion c ected;  (5) ased on da:a for a
ed nuber of developed countries only, there was a secular rise in the
Dro=ortion cf national incoe spent or. health care. Ipliit in these finJings,


-4-
,as the ccnclusion that health needs were not a determinant of national healt.
expenditures.  Abel-Saith hv-othesized that the form of fnancing system     ig
eone dter-Lant of ex=encditure,  cuna     ab±  evan2-~~~VSOO_
the finding that the countries which spend the mcst on health were those which
had several major sources of finance; those that nrecminantly employed gove-
n -iancg spen         relatively less  f  neir nazionalI inccme.
The Abel-Smith study was hampered by the frequent lack of data for cther
than Ministry of Health (MOH) and social insurance health expenditures in deve-
loping countries. This continues to be a problem today, though developing coun-
tries are being widely encouraged to gather financing data which includes all
sources of health care expenditures. On the other hand, for a growing number of
developing countries, health sector studies are being individually done, by a number
of external agencies and research groups, providing updated and more detailed health
expenditure data. However, it largely remains for the results of these studies to
be synthesized for potentially generalizable lessons in the spirit of the Abel-Snith
study.
Milton Roemer has emphasized an observed long-term trend throughout the world
for increasing social organization of the health sector - both in terms or
-nancing systems and delivery systems, though the particular pat:arn varies Mich
the political background of a country.
Paraphrasing the observations of Roemer and others, countries may be
broadly grouped into: a) free enterprise systems, where services are mainly


e 4vere in a   :ee -ar".et an: where, in =ore develcpec cou ntries, ..:ina nia
c. e_ctiv:.za:ion has occurred  - the or  or vclunta r  ccmercial Lnsurance;
b) wel"are stats systes, where industrialized countries have mainly colla-ci-
vizec -inancing systes t.rough social insurance, t.houln delivery syst:ems are
usuallv still private; c) eecnonically under-developed countries (not clearly
developing), wnose nealth care aelivery systems are a"so severely under-develc:e i;
d) "transitionalt' systems, where ccuntr±es are experier.cing signiricant exran-
s.on i:i health care services, with at least beginnings of social insurance systens
ror health care for workers; and e) socialist systems where both delivery and
inancing sysze.s are handled by some level of governmenz.6
hough there is certainly w-ie variatior. in relevant contributory fac:ors
among countries, a ';potnesis o` broad secular pattern would be as follows.
Cour.tries C Lcwer levels or dave-cpment wil have health services prov-,  ::7a
YOH facilities, perceived as greatly under:unded and effectively serving-nly
a small segment c the population. The rest of t'ne pouulaicn of :hese cu:i
-il l r-l- (-o  - ,-ent  h  receive an-,r re,--) r. -radi-a     rac     -
and her*aliszs, sel:-mec:caticn, and a mocer private medical sector whic      y
correl ace somew-at with the level of prcsperity of the urban poulation.
modest levels of idustrializacion, some employees (represencing a very 5mal1
:erzen:age of the populanion) will receive emplover-based heal:h care coverae -
purcr.a sec :v tne e=p`cver :rom te r.vate sector. ±ne :nc:v:cual emolo-er
priv7ate financing arrangements - especially as they become more widespread in
a developing country - may presage the introduction o  a social insurance


arrange=ent for e=plovee *health care. 7- social insurance heath care arrange-
-ent wil  raduall  exard as rore emnolvees e-ter the wave earn   sector;
however, there will also co=monly be a tendency for exparsion of benefit levels.
a ile the social health insurance sche=e ray 'ntially be suported, at least
in Drincile (indirect public subsidies ray be present) , by employee/employer
rs, as the svster becoes more strongy established politically, there
will b e n cresn s  ressures for support frem genera! revenue financing.
Often, however, the national social insurance health progra=s in developing
ccuntries have coveraga limited to small proporticns of rhe population.
Table 1 shows the nurber of countries, for various years between 1941 and
1933, which had anv social security progra= and the nu=ber which had specifi
D02o   components. The table reveals that over the entire period, 1940-93,
as wel! as in the more recent perioc since 17, "sickness and maternity"
co=ponents of social security progrars have grown more rapidly than has the
nuber of countries with any type of social securit'y program.  -us, wor~:-wide,
it appears that social healt'h insurance eas  eco-me a relativey more ioo-rta:
element in social security generally, in addition to considerable absolute
ex: ans on.
Separare rural-oriented social insurance arrange=en:s, both cCmunity-yased
ann tnrougn coodity cooperatives, have so far been ;xtremelv modest factors
in extending health services co larrely unserved rural populations. Pilot
projects, frequently en.couraged by external sources which have established
dedicated, enthusiastic administrations in the pilot areas, have apparently
tended to have less satisfactory results in most replication attempts.
Because collected expenditure data for developing countries has especially
neglected private sector financing, it is rather difficult to describe long-term


trends in. publi versus Drivate sector d2livery organization. :t is our
presu=rtion, however, t*nat despite the tre.d toward collectivist fi.ancing,
:rivate sector service  rovision has tendeC to grow, sigi:cantly, 'coth
asc-utely and ralacJvely, J- many -    not =ost - developing councries.
Paraphrasing Bickaell and Lebowitz, it zight be debated whether this trend
is nainlv determLied by an inexcrable market rechanism, an indic-tment of
Dublic sector delivery, the in`!uence of physicians and their behavior, a
-or= cf conspicuous consumption, or a natural desire for choice.
Constrained public sector health budgets, where they have not kept pace
with the increase in demand for more and better qua.ity nealth services, have
ur.doubtedly often defaulted to the advanceent of private sector delIverv.   en
relatively poor patients are often believed to turn away frcm the long waitin
-ines and crowee  concit-ons of Du.lic            ifcilities f ravor of private services.
Various commentators have referred to the vicious cycle which operaces in many
such countries - especiallv, perhazs, where docors (almost always because z'eir
jovernment Day is se low) perform dual service in boch the public and private
sectcrs. As patiencs turn to the priva:e sector to escape the crowded cdi:
and inattention of public sector practitioners, the expand ing Dr4iate oractices
Zf t`ese orac:itioners lead to furt:er decreases intheir devotion to publ
sector dut-I.
:is has been a highly impress:onis:c,  broac-orusn sketch   of some hea.th
secor trends. Obviously, there is room for a great deal more detail and specifi-
city (qualitative and cuantitative) in the further development cf this subject.


-8 -
2.0 D  ssues in Financing and DeliereOr;aizaio      Choice~
.conomists nava, increasingy during the past ceca de, rered est-';
attempts at sheddin light on the eficiency and ecuitym inacts o     ndividual
:Inancing and delivery ar ranents. ..ne ear'y work has stressed various
taxonomies or -inancing arrangements and suggestions regaraing some or tae
fi.jciency and ecuity issues associateA with1 the aranvemens.   Reports or
actual cuntry excerience  n these areas has tended to te zore heavivyzores-
sicoistic than cuantitatively analytica..   On t,e other hand,  any care,,i
=icro studies have been an are 'eing Aone pertaining to certa-* e..ements of
:inancing and delivery structure for local areas of various countries. A
challenain   ocortuni ty may exist for synthesizin these studies alon  wit the
broader observations to learn what results or tendencies may be nationally and
international.y extraoolated and what local parameters and variables =av te
critical.   hen 'he Priorities or c e reza:ning researca a-2nda could 'e artic2-
lated. Accitiona- :acro stucies or national ex=eriences may be va'ae, eze-
Cialiv Where it is oossib7e to   arsna17 time series evcence on ec:cienc -
ecuity impac:s associated with actual shifts in financig and/or Aelivery
structure.
Jntil now, apr:riate data for analvsis of :hese issues -as not ean
c'ui    y     ecedn     velopng countries.     ·hile collection of such d:a  is-
now ceingwide-y encouraged, tera ts uncou_tezy opportunity for progress
tots aifficult area by reviewing existing incor=ation on financing arrangenents
4ntensively, with particular attention to countries which recently have
e::perienced a shift in financing or public-private delivery struacture.


2.1 Alterati_Ve Arranvemen s
The f7undmental dce-y i        = fiarcing arrangements -is 'et-een
taxatiCn ad -ersonal oayent. Gvern_en: taxaticn =av 'ce deri;ed from senera
tax revenues, or Prom ear-=arked tazes (designated 'or the health sector at
tie of co..lection), either of wnica  ay ce obtainein via nationa, p rovincia .,
or distric: govern-ent level. Gover=ment tax revenues =ay be supplemented by
eficit fiancfing and international assis`ance.  Soal health insurance revenues
(though often supple=ented by general tax revenues) , are, in pri.nciple, derivec
from e=plcyee and employer contributions of those workers who are scatutorily
(someties voluntarily) covered by the arrangement.
Personal pay=ent arrangements include fee-for-service (paymr.ent by a ser-::ac
recipient fr the service) , ccimmercial trd-party insurance, reayent caci:a-
tion arrange-ents (isurance arrance=ents where specified p rovicers assume r
under ccrractual arrangecencs) , cc-zzunit financing (eften with in-kind or lacor
cntributions), coop erative arrangemets (frequently also opera:i-g witi
ccllections), private firc arrange=ents cffering health care services co their
a=plcyees and often to empIoyees   fa=iies also, general charity and orvanized
vcluntary avencies.
The isSue of ser~;ice provisin under vulc versus pr-ivate:authori:n
a spectruc ransing frm facilities cperated; by gcvernmen: depar:zental 'cudge:s
(such as '-:- facilities), to zhose operated  y guasi-overn:en agencies
(soc2eties exe`Dlified by seni-autonomous social security agencies) , to private
providers which =av be subsidized, regulated, or left to operate with virtual
independence.
Varicus combinations of finarcing arrangements may serve a single coun=ry -
either operating in.ivi7ualy to serve different groups within the country


-10-
'lor ser-;`n  :.-e sam~e vr=~                :~ unda"s      , r Zre-at:In;  cr::
e a-m:,eents c-_-r ::-an f-e--t-a :
tcat   u:e to ser;ices coverage (risk-s ring) for a sa rca-accessa 'D a
e -a Ivenif-1l Do Du Iat ior,
,so, sources of perscnal payment are often chamne.     prouga pu*lic secto=
ac:-c4  -s (usuaily whaen thesa pub ic raC-..-..Ies recuire recipit ca,y=ent)
as well as to private providers; and some gevrnment paymen: scurces are
to orivate seetor oravicers                           4=icil: as exbc-.  r :.i. Sor
when social security sc*e~es allow the use of private  roviders)
' basic rule intne selection of a nealth care -:'an-.nz svstaen1 - regardless
or tre pub-ic-private dalivery structre m'x - is that (disregaringcontributions
:rom international sources, which cannot be counted on for sustan:ia- :asion)
zre calivery or health care serzices to the -ooulation or a cunrrY will, ir. any
case, ultimarelv be paid for oy the pCpulation or tae country. -e caoce or
inancir.g sysrez, however, as indicated  nrevi2usl, cn         -
(and often cbscure) the ultimate i-.cidence of pa-y-menz within the coury, ese
affeccinv the eff`icienc  of delivery and utilization of services and the di st  -
out -onal ecuit7 of the ser-ices.
2.2 Ealuatiocn Crieria
Beow  w  list si  crierf h~-Jc   -a, 'De useful f- eva-jat_4== - - --
financing arrang-ents. Some o -hese criteria, especia-l.   the aotter three,
may also be useful for evaluating public-private crganizational arrangements
ror services delivery. 3reaking uD the choice analysis into these separate
criteria may be heliful in (1) assisting in a clearer sccietal recoznition/
articulation of the goals it wishes to pursue, relative to potential system-
choice impacts; and (2) enabling a greater sense of tradecffs -    ,ang separace


-11-
:-.ancn- anacelivery system i-jacts - which may be ootainei.     W n'..e a partiv-
iar cri:terin or goa! may ternd to be dominating, the possible cost cf exclusi:
attention to a single goal may be soberi.
(i) Aggregate collection.
In an ."econcmically ideal" world, the amount cf funding directed to Inea"t
sector services would be determined on the basis of allocative efriciency bet:een
sectors. As a practical matter, market imperfections and grossly inadecuate
information about marginal intersectoral benefits leave a large interval or
uncer:ainty regarding desirable funding of the health sector.
Given the existence cf large populacion sub-groups who are unserved or
seriously underserved, government health authorities have made emDhatic arguw.:s
that the health sector is "underfunded" - at least in absolute terms. Lut with
absolute resource limits, allocative tradeoffs, including other critical human
needs (food, clothing, shelter) , limit the resource funding of the health sert_:.
Neverheess, it is apparent that the (       ix of) health care financing s
chosen by a society can i.fluence the aggregate resource allocation to te sec:Dr.
hie governmental '-ealth authorities =ay find their allocation from general
revenues - for -ublic ser-ices and subsidies - to be substantialyr
at any  oin: in time, they may recognize thar advccacy of new/'aL:te-a:i:e
nan-cin  svstems will likelyr add to the aggregate resources devoted to the
aeat   sector. -  (special-y w1 re the new source adds to a mix or syscems,
substitution and complementary funding effects may lend uncertainty to the net
resource increment deter-mined by a funding source. Whether or not there wil


be o :.it-a-ou1a rary  ispacemen: in raso cse to alte:r-nive funcd1- :s an.
e eent or this concern.)
Eo r-s to e-xand the financial funing of the sectcr n eed, o course, to
consider the ab sor re caDacity of the seccr - to exDand i:s phvsica  resources
otherwise the addrtional fun2g may 'e larve y Cissipated :y sector ita:.cn.
(i) FundL-g effic.ency.
There are adcnstrative coss in the co2cn            transfer   f financ
resources to the health sector and these costs may differ si'nificantly between
i.ancing syste=s. .nere collection costs are n:gn, on-y a medast
collected resources =ay be available to _und health care services. In certain
rural coverage arrangaents, collection costs conceiably could ap=roach or
exceed the azounts colleced. Various administrative procedures have been
o.served. to influence runding efficiency. For exapDe, it- has been noticed :hat
.rcuentlv when usar charres are retained 'v a health ser-ices organization     the
collection of these charges tends to e  uca c igher tan wnen they are
to a central treasury. (Oppor:unities fo graft or corrupton in collectad
charges =ay also da4nish resources available for sector use.)
( iii  F din  in cid ee nre .
--e issue cf wnica =e-oers o a country    D ouia:-.on are recu:rec to 3a1
to t.e health ser-ices delivered in t'e country ay firstoe searated accorin
tosysves that w-holly or   arcially charge cnIy those who utilize the services
and those that charge persons in population groups independent of their actual
utilization of services. The latter funding category -ay be associated with
exlicit coverage arrangements, such as social2 insurance schemes, or with general
taxation - where there may be a low correlation between funding source and
effective coverage. With considerable iecuality in the incidence of sickness


in populations, vovern=ents have widely regarded cz,7erage arrangements - where
the czsts are shared by both sick ard poor - O hold ecuity apDea. But sociecal
cer:.niticns oEf fun.4 equ:.t7 may vary: Should it be totaL.y indeDendent or -
or somehow related to - the right to receive serices? Should funding equitv
be horizontal or vertical? If it demands progressivity, how vuch?
Again, alternative financing systems will dier with regard ti their
a bsityz' to -mple=ent a givWen t un.d4  ecuity....  Fur her, i.: may be diicu-
to esti'ate where the ultimate incidence of a funding scurce lies. Where, fir
exazple, special taxes are levied on productive factors, is the tax predominantly
paid by the factor owners or by the co=unity at large, which =urchases t-e out:ut
of the factor?
(iv) Eficienc of health serices production.
This criterion 4.s orten referred to as cperator.al efficiency - the unit
cosc of production of given health services. The copetence of managemen,-
wei_t ana comPlex:ty of ureaucracies, the a-equacv of accounting systems an
incentives for diligent work~. by deliver- system personnel :0ay all contribue
s:~.ircanti  to Jroucing health services =ost efficient  under alterativt
:inancing systems, or ubi     rvc-riatde deliver- structures. The metico of rezu.era-
:n. 3or 5er-7c prvidear s ise c ony considr5     an imor:ar.t incen:-ve fcr
influenc-g procuctivity. (Re-uneration inOcntves 7ay aso affect utilization
er::c-ency, as discusseA oelow.) Poor financial planning and bucgeting can also
contribute to severelv dizinished operational efficiency. Realth sector facili-
ties =ay be constructed, but left largely unused due to inadecuate recurrent
budget funds fcr the facilities.
An important issue is whether private sector delivery of services is more
operationallv efficient than delivery by the public sector; and, iI so:


(I unica areas or: se:-n.ce ce'iery snocu. a te naca cy tae~ pr.vte sec tor
(:a".1ý'ng in~to  zt  :siietral2a0f15 for Ot"her crizarla)?     (2) Ca. Z-actors
producin'g greater                                  '-eer.cy i he prate sector C iCnt:.:oanc, zeZ- aps
t:a-is.ated :.nto lessons zoz improving :ublic sector er sciencv (es:ecial
whýere traaeorrs :o  other criter4aarge ror puolic sector rEten-ion)?
(v)  Disriutional equity.
ý.ne unacual access to neaa.th serrices is one of th.e most :ro:iatit-
of the health sectors af most courtries. The ipac- of alzerna:ive financi=7
syste=s and delivery strutres cn the      ctive distribution of heal    se--rvce
aDCong th.e population :-s one of tne =reatest cOncerns of countes.
eco d'd tnat, to the extent tha- dstr-DutDona -cu-Z    , 
ser;ies is a reflection oF -c=e distribution concers, s-uZn        -       ca
be aconlished  ore e ffI  intly thrzugh transfers o ene-al purcasing  o   r
ransfers in kind (or specirEi suosid:es)   hizh radiistr..ue h.ealth serrices
..r..  .aef e..  - -...  ..- :n -..  o.. at  ete - :...  - ... - -. ,.  - . s  -..tn, .. .tcD tD t t
udage wic` goods and ser,ices wi   axie his satisfatin. soaever sca et:cs
(which dicate the for= of transfar) have ben rcgnized as preerring Do att
riros a  consutption zenav±or;  .e. they exer   nce lass o an e -:erna
D:sec-,o,.  o- to  ~o e,'  ,o-a    t  t     --  ,     - S---    -a- 5
[10]
suzn as i-l health in segzents or the society.


7-Droving distributional equity may include any, or all of (1) diminishing
direct financail barriers to obtaining services; (2) diminishing indirect coszs -
suc. as travel tine (and, perhaps, waiting tie) to obtain health services;
(3) socio-cultural barriers to access, including feelings about modern medicin
relative to traditional alternatives. In rural areas of developing countries,
indirect costs associated with the physical distance from health services may
,el! be the domi4ant factor in distributional equity. Numerous studies hava
shown the highly restricted geograDhic radi4 for effective coverage of scatterad
rural health centers. Consequently, much attention has 1focused on ways of bringin
health services closer to people.
The effect of financial barriers to health services has been the subject of
considerable research, though the area remains cuite =urky and in need o addi-
tional careful studies. The issue involves mora Drecise understanding of ie
deternce effect on utilization of services which is deterinedI by various for:s
of cost-sharing by health service consumers. Estinations of demand elas:izitieS
r di---rent zroups in diffarent                  countries have providaed greazy
varying interpretations o- service utilization resonse to price. Obviously 7anv
actrSay        exected to d      -ferentially izpact or different grouDs (who have
di rent degrees of poverzy, heaet perceptions, etc.) in different areas, :ic:
resu_:ant imp:acts also on estimated demand elasticities - -whic of course will
vary t~rougI Zeren    rice ranges. Existing studies differ as to whether or
not prica has a major inhibiting effect on utilization - even at highly subsi-
dized pricing levels - and even offer findings that, in some instances, th-ere
is a positiva correlation (even for the poor) between price and services utili-
zation. Again, this is an area requiring mora studies, along with efforts to
synthesize the inplications of the varicus findings. To the ext=nt that


Sul-s ... ..z .-ea....  ar.  ..- ,  -  -  ..t - -,...--,.  Dr-c  -   a.-ts  as
a dterrenc to t-ne tarvet vr3ops of the subsi,> t`e subs`Id 7ay 'er
captured 'y groups other than those intende to rezeive it.
Scme research artention has also been, v-ven  one     ss:.e erects of sic
c.tural hindrances to custrizuional     ty,   :ougn vucn -pvrovec uncerstanii=-
is required nere, as wen.
(vi)  Utilization errfcenrcy.
This criterion "ight also be rederred to as allocative e:ficiency wihin the
neaah sector. zkcng the different possib.le aspects of u:
3 r nea th services, are the riowIng:
. re the most ericient noces o     resource converston to sea:er
status improvement elyed? This is separate fro the issue
earlierdiscussed of whether the indiVidual ser7i2   deliVary
modes achieve :axiu: operazional erc-.ncy, i.e. dei;e.
given health services at lowest unit cost. Gi4ven t'at lowest
cost per unit efficiency or `ealth serrices is obtaned, e
present criterion leads us to the more perti-'en, but also
-ore di..icu.t issue or the unit cost efr:ciency or oneat-n
status i:=Drovezent.
2.     e given modes of health services delivery reingutii::ed by
indi.vidual members of the "o-ulation at ota      evels - -.
not    erutilized or overutili::ed below (beyond) the oCi-t
where the incre=ental izprovement to health status exceecs
(is exceeded by) the incremental cosz level by some interal
related to overall resource availability to th.e sector?
3.   Are -odes of ser-ices and dartcu       elivery ssems direc:ed
to disease groups (population groups) where they can 'est
diminish tne rastraininv effect of disease on cne econc=ic
development cf the country?
In developing countries (and to a large extent in developed countries also)
the zajor focus of attention has been on the first of these issues - utizio


cf the most efficient deliv-ry =odes. Athough ccr.versions between health
services delivery and heal: status i-.act are exzremey diffcut to esimate
t`lere has beer substantial agreeren, for several decades, that utilization
e:r:ciency can be imprwoed ov greater em--asis on preventive care methods.
Scmewhat more controversial and subject to contLuing exjloration, is the arvu-
=ent that efficiency can be enhanced by increasing the proportion of certain
zaanower resources in healt- servies celivery which are less costy to train.
Attention to the second issue - individual utilization level.s of services -
has mainly centered on the appropriate level of consumer cost-sharing, which
(as cescribec above) is a1so a key topic of cistributional equity. It 4s
argued that, besides assist:ing in consuer recognition of health service resource
ccsts, cost-sharing romotes rescurce cost cvnsciousness among roviders, dis-
coura2ig ":rovider-induced demand."   iceaily, the cosc sharing pro-portions
should reflect the relative resource costs to society cf different services
(adusted for any -enecit externalfies). The obvious problem, however, is the
confict with the distributioral ecuity critericr. and tIhe comlexity of
trfeoffs for differently situated members of the population.
Te thr.  issue - :he relative impacts on economic developcent of servi:ces
delieredto ddiernc dseas  ares an/or  idieentDopulation 7731upSisas
te subjec: o  continuing analytical researc., th.ougn, cnce more, results are
s   a  mix. :.-ements of the issue concern both the extent of irovement
in worker productivity (always regarded as favorable) and the possible expansion
in poDulation growth (most often considered unfavorable) resulting from disease
reduction.


2.3 Su    r-
,e iacts of t:hese critaria, -eightd - i in st ances of tradeoffs - cv
s:cietal goals (int=ereted, as :he case may be, y na`icnal authorities sr
ex:ea. runding agencies), can assisz   :nhe choice or .eal.h care  ining
systeM and de.ver          r      Obvious_V the izormationa. neecs (ana
1erhaDs analyti:al capabijities to) for   aluaig these cri:eri~ are sericus
deficient. However, as suggested at the cztset of this iIscussion, even at
present inforation -evels, it is probably useful for countries to attempt to
e-;aluate the cotential tradeoffs associated with Darticular choices.
t would appear tnat, perhaps conco=itant to (anc presumed excused 'y)
`-e general lack of infosation - secially in the e"ficincy. areas - countries
(deve-oped and developing) have been giving predominant atte*:ion (at least
r*~etorica-lv i noz always in reality) to the distri^utional ecuity issue.
Je hvothesize that: `1) Concerns with c:.striutionai ecuity have frequentLy
( inappropriat ely) over;neme,.. not:.ons c allocative and operaticral e:::::ncy
(2) the approach to distributional ecuity has often been haphazard - with
surs:cies frequent-y predomi--antly favoring groups not targetted for the;
(3) the efficincy Costs of achieving distributional q      (-h er e th ere has been
susatial progress in this area - most Oft=n in the more industrialized coun- -
:riesì have been severely nisunderstood; and (4) the possibilfti-s for substan-
:,a-'y increasing efficiency under given (targettel) Levels of distributional
euity are considerable, though largely unexplored.


Regarding both the effiien c y and equity issues discussed, there is a
s:riki4n  dispar-t t etween co ncrIly prcc laied national _oals/ideal_s i  eve' -
ing countries and actual  vezen: of the health sector in the proclai edd     -
:ons. Is this a greater problem in the health sector than in other sectors;
and, if so, whv?  In scme cases the problem may be that of rhetoric about wha:
is desirable outpacing what is financially possible, at existing stages of deve-
Ioment.  In other instances the influence of some of the elements citec at t'e
outse: of this paper (e.g. political pressure groups associated with the status
quo, socio-cultural factors) ay be underappreciated. Often various practical
difficulties certainy sow the process - e.g. converting existing invest.ents
in sophisticated curative facilities to invest=ents in rural pri=ary case.
Äg_i   one hopes  h   improjec econCnc analvsis of how to proceed can contri-
bute; and perhaps econonists can also play a greater role in transferring
naticnal political rhetoric about health sector goals into public recognition
of the hard c.oices recuired `or impleentating real progress towar- t"e gcas.
3.0 General Cotions for Ccutries
. Gever--=en t vs . Prat Provsio,ýn
cst dev:elcpi   countries nave govern=ent operated ser.ices and orcamrs,
variously  rovid through na:ional, provincial and district levels of govern=enz.
The services operate through numerous vehicle formats, including special cam-
pagns, programs and netwcrks (in principal) of facilities and personnel operating
government hospitals, health centers, sub-health centers and health posts.
The operation of these services may take place through a complex web of finan-
cial transfers, involving funding from external sources, national and sub-
national gcovernment revenues, and user charges.


list of these courtries alsz have subsa::fal prva e sectors for services
provisicn - for *oth  odern =ecici`e and :practice. "'sc Of t`e
riate sector provision is through fee-J.r-ser:ize.   urz-ases (frequent
re=resencing se-c-edication)a         for a maor share. of private sectcr
provision.
As stated earlier, the co-untry options f   sr ser:ices -rovision `ncluCe
government (or Cuasi-government) provision, governt-r       a:ed and/or
-overnment-subsidized private sector provision, and non-interfered private
sector provision.
Selection from these options will tend to center on the latter threeia
1isted in the procedig section:  efficEiency o   ealth services producncr,
strioutional equity of health services, and utilization efficiency of health
services.
Generally, there are broad suspiicns that, lacking certain market incentives
f3r efficifency, govern-menýt services t-End tc 1:g ina te=mS Cf SeJIZe Dprod'UC:i_;-=
efficiecy and in soe aspects of utiliza:ion efficiency.   Cn the other band,
services provided by the non-nterfered private sector are recognized as
i societaly unaceptable     isiu      a   qauitv =nd societallyineicien
reswurce allocation in areas affected by externalities.
hepec issues in        he apprcDriate public-private-ra-gulator-
zix of services provision are nu=erous. Broadly, it is a matter of esixatins,
in each country, whether it is simpler (less costly) to more nearly approacä
goals of distributional equity and operational and allocative erficiency by
attempting to introduce missing incentives (by changes in procedures of adminis-
tration, orgar.ization, or remuneration) into pubic sector provision, or through
governnent introduction of regulation or subsidizatior into private sector
prcvision.


Ve2 Fa- 20--sen-, e  g T'an.z~z
u      nficiency _aY iarY wit     t`e size and type of fee-f:r-Iservce
provider arrngement. As noted earlier, in the case of governimant facilities
wnere funds =ust be remitted to a higher level treasury, or where there are
oDDortuniities for embezzie=ent or corruption, funding e...ficiency may be poor.
urning equity under fee-rcr-service is often considered pcor in that costs
=ust be fully borne by those unlucky encuvgh to become il and recuire health
services. Funding equity is further diinished to the extent that fees - as is
nost often the Case - are not graduated accrdigng to ability to pay. The
sericusness of this inequity will dend on the =agr.itude of the fees (wich
=ight be reduced by subsidization - in whic case the funding equity of the
subsidy sourze =ust be also cosir     , rl ative to disosable hoseho       incoe.
In the case of full (unsubsidized) fee-for-service pricing, distributicnal
equitv - at least within the =odern medicna  sector - will no ally be exr     Ie
poor, with the substan:i.al eliization of a large portion of the populaion
o  inar.cia  access to services.
Service fees, as discussed earlier, are a zajor factor in allocative effi-
cinncy at the individual level, though the i=at nay be ither favorable (to
the extent that overutilization is avoided) or unfavorable (to the ex:ent
i   i     e      , deending on the actual fee level confronte   y
inividuals and its contribution to rea' ccnsumer cost (inluding indirect, as
well as direct costs). The possible impact of fees on reducing provider-
induced demand may alsc be considered. De Ferranti has discussed the types of
health services which are candidates for exemption fror private marginal cost
pricing on social efficiency grounds. (However, those services which do not


-22-
:ualil y2- or exe-ptzion frz= full arvginal co st Dpricing, will1, as de 7-erran1z2 a-so
:ei,- e      i - ~e ur.--u:ilied acc:dL,2 to --i-y c:     eL    '
3.3 Gcvermment l-ax Revenues
eovernment financing, frc  general tax revenues, efricit financ-ng, cr
en=arked taxes, in theory, offers the opport=nity for expanded avgragatefundi
of 'ealth services. Avenues include expanding the tax   ase, increas-.in  tae ave:
of taxaticron .,a raising the proportional alliocation to the     sector.    1*ile
sc=e countries nave experienced signicant funding incr&ases from government
revenues, ror most developng countries it see-s unlikal- tInat suostant:al r_~1
Der caDita increases can be ccunted onr from this source. Government tax revenue
~inancing is someties unfavorably regarded as scew-at unrehiaD,e. Actu--
-=z =ay :recuently fall short ofr Dannec allocations.
inese various forms of govern-ent revernue financing, in ricip-e, al-ow
broad flexibility for funding equity. (:ncome taxes ma-y offer the best op:czn
or ver:ical equity; sales and anrculural product taxes - deDencing on -he cCn-
suner destinaticns - Da be ragrescive.)
To the extent that a tax collection structura already, exists, thera may' -e
sozfun~geficer.:; n  inoy icluingthe  --:n sez~or~     _u:!c~Le
a::rangeents :or disse=inatinv funds -hrouh ultiole .evels c: government
a- arious agnc-s zay *b costly.
Evaluation of service production efficiency, distributional equity and
tilization efficiency will depend on which of a possible range of delivery


Mec*anisms is support:ed by govern=enr revenue fund s.7 ou,h most cormoni
associated with MOE Zacilities and progr. support, ovretrevenuesma
also subsidize other financ-ng sechanisws, as wei- as elements ef private sez:2r
deli-very. These eficency and equity cri:eria are perhaps best appliea in a
discussion of oub!ic versus private sector delivery structure. (Most broadl-
the subsidies from government ta= revenues offer the ocortunity for improved
distributional equity, but - insofar as pricing ientives are interfered with,
production and utilization efficiency may suffer.)
. 3.4 Commercial Health Insurance
Comercial insurance arrangements canstitute risk-sharing procedures,
provid-g coverage to Dersons (enrolled individually or in grou-p)s) able and  i n
to pay an actuarial premiu=, for a cesjgnate.. set of oenefits available an
event of illness. Coercial insurance plans may, in some countries, be pur-
c'ased 'y Drivate firs for ezployees, or even (less frequentlv) uti1:Zed by
gover,.=ents for social health insurance programs.
Though a potental source for aug-en.ing funding of: the sector, c
insurance Dlans seem to represent a ver- low percentage of financing arrn-en
in developing coun:ries. Whether tIis   the resul t of socio-cultural
abou: risk, high transaction costs relat ve to the smal  economic marke', or
other fac:ors rel aced to the str,cture 0f develoing country health sectors,
is not clear. While it is plausible that commercial insurance would flourisn
according to the magnitude of personal Dayent financing arrangements generally,
there is also  the opportunity to cover "buy-up" provisions, where allowed by
the standard service coverage of government and social insurance schemes.


u-dZ,S e              e3 e rfy                   n-as, --7 be -e     e
=%...l  Wp       -             oh) as a covera. a        .r-e.e.,  o. . e   -
and the wei- who enro     share the nealth services costs; and (2) the preDiuzs
are cl.leced frC  a cetter-Cfr seåent o the cDu 1aticn who mostl     eect to
a   tese coszs -' addition to their share o governent revenue CCIlecti-ns
The efrect cf con=ercial insurance on service production erriency, wal e
ncertain is poteztally positive ir. that (1) in a beter osition than indi-
vidual consuers to zarsha. relevant       :-a_t-    and :.ruence ce"ana ror spec_ --
services, insurers can encourage provider productive er::leZy; and 2.) tre
4,cr=e:e   choice r ccnsuers - bet-ween Dubi-i    nd  ri ae setor -rv-'
een individual nrivate sec:or providers - can a..so encoura-e product
....... o.. a..  e uity  -s  o-ten  vie .ed  as  u ..c.o- i  a.... d b.v
su-earr --     =-an  -   Mez'er- or these blans ar    -en to c 2*-
4is-roortionate shares o health sector resources.
_tiliza-.on -une               cval :nsuanc           (as wel:. as under othrer
ee       -arra--ns) -s ne-atvev ---ctad -y the o:rcben o-    ,_oral hazar ,
i. - v Fau..v as c2nanes -n te  0e-a5or or consu=ers or eatr     serv:.ces
wnic  are incuced cy their healt' -insurance coverage.       tnese behavior cha.ges
will have consu=ers deandine health care services in -reater quantity and
qu:ality than is allocatively eff:icient.


P5rivate firs in =any developing countries individualy provide health
[£15 1
cara Service arrancoments to their eployees.    These arrangements may or may
not also cover employce dependents. Services may be offered under a direct
arrangezent of e=ployed health personnel and owned facilities, through contrac-
tual arrargeents with Drivate providers, through rei:bursed fee-for-service,
chrough granting cf a fixec sum, or through Soe coMbinatior or anlv of t'ese.
The private fi= arrangements are probably =ainiy the resu: or empoyer initia-
tive, being deter-ied less frecuen tly by govern=ent regulation oertaining to
emloyer arra.gements for services. Where the government decides to legislate
enployee health services coverage, i: is zore often done through social insuranca
sc'emes.
While private firm arrangements :ay add someAhat to the aggregate resource
aliccation to the health sector, the imDact for deveio:ino countries see-s likcly
to be =odest. An aroreciable -ortion or the -rivate firz resource contriouz::n
(probably a s=all =ercentage of the national allccation,) likely disDIaces prz-
vate fee-for-service or government-supplied services.
Fundin  efficiency should be very high for private fi:. arrargements, since
ere usualy -s no collection ot ir= ius or ta-:es, nor cmple:-: inter-egencv
transfers. Such arrangements should also represent good funding ecuity, givn-.


-6
the implicic risk-sharing among members; and given that the fund:ng incidence
is borne by the emplDyees, rather than shifted forward to the purchasers of the
firms' produzts. (The issue of incidence is further discussed in the following
section on social health insurance.)
?roducticrn efficiency =ay vary witn the particular arran~ent o4 t,e 2r4r
Cr se-rices provision -          direct or -direc  - and  he toe      s
t'e fri.  -n a -ater section on Indonesia, cata is reviewed which suggests that
t-e separate arrangements o   d    al    - ir's tend to `ave signii::-antlv hi
prvductiorn costs than can be obtained uncer socia insurance (with services
purchased at unsubsidized nrices fro= the public sector) . Dis trbutional eczfv
and utilization efficiency will also var between cunis wimh :he zean level
and sophistication of benefits and the particular benefit transfer mec.anism.
Generally  iz =ight be expected that distributional ecuity will be izpacted in
a similar fashion as under fee-for-service and ccmer-iial insurance. Regarding
utilization efficiency, the e'eed   act is less cer:ain. Under both
serv:ces Drovision by co=panies and fixed zonetarv allotments utilization effi-
c:nc zy e se-_e-j,at nigner than uncer other cov;~:rage f=nanci=v arr=zgeze-s.
2.0 Socia. 3ealth nsurance
ealth insurance offered as a Dart of nazional social security prograzs
appears to be increasingly adopted by developing countries. A variety of
-atterns are followed. Programs =ay be administered by the Ministry of Labor,
by the Ministr-y of Health, or, perhaps, jointly. Under the more common example
of Ministry of Labor administration (probably through a semi-autonomous agency)
health care services may be provided directly through the organization's own


Cacii .ties anp personnel (usually ex=lusive-y ut..zed by covered members) or
th.ey nay ce purchased from other existsig service pr-viders. The governnent
act as its own insuror or (less co:aon..y) enploy private insurors.
Social health insurance is believed to carry a particular appeal in that its
benefits tend to be regarded as a "right" deter:ained by the premium contribu:cns
of t`e merbers - rather than as a paternalistic charity, supposClyi associaced
 the ,  amily of services.
The increase in aggregate funding of the health sector, plus the relative
stability7 of the funding, are usua-y cited among the primary advantages of
social health insurance by the countries wnicn nave adopted it. So=e observers
have sugested that social health insurance programs tend to weaken MOI's, possibIy
cisDlac4:g .. ..unding. Roemer, however, concludes that tne evidence dces no:
surjort this hvothesis and cites the Abel-Sith studies as sDuorting the view
that countries wit social health -nsurance progras spend a greater propor=ion
f 16]
of GNP on health than those without
-hough little data has been reviewed regarding the fund-ig er:cienc; or
these progr=s, the efficiency of the incre=ental fundin  for health insurance
acc   to existing social security 7rogra_s woula seen ravorable. Wowever, -her-
=7     varizus practical pro:le*s, related to the calculation of thne ic     -
premiu, which zay add to the costs of :reniu: collection. (This night result
for exaple, wnen a d.rferent premiu caIculation :aethoco-ogy - pernaps 4nvv'-
nuber of dependents - is necessary for the health insurance portion.) Where
social health insurance pre2iums are collected from members not covered by the


-28-
exsti' social secur:t -rz ra, spec:.a 17 -nere          Dl z yer) co...-ecin
units are s=all (with =inial account.ng Systems) and geographicaly sc     r
d-:Mizer.4cv -a s:aro.. diD, i - votertiall y r' -- ,e
-er some cont riutory un is. 7ven wn te socal health inurarce Dr:cra-
joins an existing social secu:rity progra, however, mucI wil depenc on t-e a:_i~
trative fundi=g eficiecy of the existing prcram.     n   rici'., relative to
private insurance (isnecL.lly for-croi    insurance), with administrativ    a
costs, social hea1ta insurance snould be capable cf higher net yie-cs.
n:ie level or :uncing ecuity wil, vary depending on the particular rincIs
arrane=ents or social insurance scne=es.   .s in the case of ocher covara2e
arrange=ents, ecuity is favored by a sharing ol costs .etween bth the healtny
and the sick. Vertical equity is enhanced by sche=es which collectre:uzs as a
percentage o   arnings.   (3ut tere =av be a cao an earninas subjec:ac to     n
collection.)  Cost-sharing, w     t its uncertain imaocts en utilization, has  e
to be in requen    e=up,clyed In social nsurance progra s c: Jevelopin  cun!ries;
'cwever, as lremi levels co=e under increasi-g pressure due to rising prograz
costs, cost-sharing =ay be increasing`y turned to in tne ruture. run:. s ecu-.tv
see-s Doorly sered when progras - which tend to serve the better-of± segents
ad the population - suzplement: collected premiums with general revenue inancingz.
Regarding the funding deriveidfrom erployee/ezployer prerciums, the ultimate
c:cence a: the Cre=ius - whetaier )redo=inantly on tne ccntriu"ers or tne
consuners or tne rms' outputs - has been argued. The incidence will depend or -
-ne competitiveness of the labor and product markets and or. the elasticity a
the supply of labor. Assuming the premiums cannaot be readily avoided by switching
jobs and that we are looking at a relatively inelastic labor supply (the expected


-90-1
the worker premi-m will be predc-iantl'y absor'e- 7abor '17   ith
per.ec- markets  the division or uhe prem2u be:ween the e:r,iovee and eplyer
will have no signifcance. The e=Dloyer will regard both seg=ents C: t'he prezis
as equivalent aspec:s of worker wages and - at least beyond the short-run -
or-set these costs with decreased wages. However, Musgrave suggests that in
unionized labor narkets, ur.icr.s wil find it more dif-icult to aczept wage
offsets to e=plcyer contributions than emplovee contributions. In this case,
there is greacer likelihood of the employer contribution being passed or. to t-e
product consumers. {17, pp. 279, 495-497.]
Wich a variety c: delivery mechanisms and provider efficiency, incentives
possi'le under social insurance, generalizacions about productive efficiency are
d ifficut-, . As indicated in an earlier section, however, it might be argued
there may be significant =anagerial efficiencies (with, perhaps, =ore skillful
application of provider efficiency incentives) under a single agency resDonsible
for e=ployee healtl care coverage than under the mulciple inivicual sehees c
priVate ezployers.
he ex-erience of those social insurance progrars which have existec rr
=any years (predominantly those of Latin Arerica) suggest rather pcor perr-ance
t-e area of diStri-utinal equity. Exzect for sysvezs wichave attenpted
to exand too ra(i7 (erely prov-.i.n " paper rients  wit.out acecuate
access for some), distributive ecui:7 within the Drocra. membership s enerally
considered favorable. However, many progra-s are confined to modest proporzions


-30-
of the population w-- cnsu=e a disp cpor:ionate share c  national health
resourcas.
Especially  hr    rogra-s cffer generous leefit leve.s (         sophis-
ticate.expensive orzcedures), officia_s wishing to e:cpand the program beyon..
t-e =ore arfluent industrializer sect)r must confrontcrric-t choices. ½ey
can wait (perhaps ny decades) for the =odern industrialized sector to ex:and
to include ~uch hihe r proportions cf the opulation; they can seek to eoanod
population coverage t.hrough support from general tax revenues (problematicai);
thev can exDan via differentiatea service bmefit  eis (ossibne erc
com-rcmised equity -,prove=ent); tha zan try to extend More or less uniform s;r-
v,.ce benefits to an expanded group by having arope who contribute  '-er Dra='
aCrunCs subsidize the benefits of those who, at best, are able to contribute muh
saller pre=iuts.  Insofar as the Iatter choice represents (as seems likely) a
diinution of benefit levels for the oriinal enrollees  the original enrcIlee
group may 'e ex:ected to offer resistance to thjis route.  Ooruniies      r su
expansion wculd appear to be strongesc wlere it has Deer possible to rescrain
he level of beneits~ inclu :.ng the use oZ sopnisticated procedures and the
csznstruction of highly e:=ensive facilities.
A pro-ecuit:y argu.ent for social health insurance (which iplficitly assu-es -
there is no dislaceme nt of 'O1 Funding), is that the O   : exenditure  which was
dirce-d to those urban workars who are subseauently cov,ered by the social
insurance program, can then be zade available to better serve both the rural
population and the less well-off urban Dopulation who remain cutside the social
insurance program.
..e pDer:or=ance of social insurance programs in utilization efficiency
probcably closely rirror their perfor=ance in distributional equity, due to


-31-
ec=en contribuory faz:ors. The experience of most social insurance prograns
has sho=n a s:.    tendeny tzwaras urban-based curative medieine, with minimal
attention to preventive care.   specially, again, where services are provided
under the direct pattern by the social insurance organization, faZilities and
equipmen: have tended to be sophistica-ted and e,pensive. Advocates of the
social insurance route often mae t.he point that services developed under tese
sz`emes offer high quality standaris to be emulated by other areas of the - e
sector. However, in many instances the developed standards may exceed what t`e
country can aziord to emulate. .r4:cencyv of inaiviaual utilization of sexces
u-der sccial insurance (given the infrequent use of cost-sharing) may tend to be
affeted by the sa=e mora- hazara issues cited for ccmmercial insurance.
3.7 Co=unity and Cooperative Sc'-es
Eoth community and cooperati-e arranceents represent attempts to marsha.l
unta=pea resources 4in larvely unser-ed rural areas to exand the meager cr non-
existent health care services available to this population. The major exressed
conce:- (which is ncc compe1.n) 1-s that comun=ity and coDerative ef:or  -,
tend to mitiate the devotior. of government (little as it may effect:ively be)
:o c:hanneresourzes to t*.ese gros. Some also feel that fundinv ecuity, i
rv, where the most ipoverished groups are asked o contribute resources
health care. uevernhless until development levels grow substantiallv and wa-zs
are found to extend health services to rural areas from other financing sources,
efforts should be made to solve the proble=s of management and direction of couni
and coerative schemes, attempting to emulate reported cases of success in t-ese
areas.


3.3 Su:m'ar-;
As sta:ed Dr aviusIy, evaluat ion of alterative arr anements according to
tnese crteria ana :e 'aanicirn - tracers wil obviousLy be            ~,spe-
cia7ll  with current infmorat iin dei ciencles. Nevertheless, at::pts shCuld
-ada to a rticu1ata, as wel as possil'e, exec:ed mDacts a ccordin  to these
C=2--er4-a w~nalternativie aiir=ir..   :    -zeers are bein~gcr.iei (or
rcg     e     be indepenenty        vin ). Reviews of country exerience,
of these criteria, where such changes have receitly taken place may provide usaf-`
1lessons.
aations of course, will normally   nvolve incremental changes - rather
an --ediate t-tal replace=ent of entire sector systems and structures.
valation o    changes wil need to consider the approriat e counterfaz:ua      o
iv-en choice. Estimating to what extent given alternative systems compete.
auz=ent existinr systems (probably rarely) versus displacing some :crtion
eiSting svstems and, especially, estiating the degree of dipa         nf
suecific systems will be key elements of the evaluation.
-.   :ndonesian Case Studv
.   a Gnera! Intr=duction
A single brief field study was conducted in Ircor.esia, to enlarge our appre-
ciation of sone of the real issues (and their perceived priorities) being cen-
fronted in a najor developing country, with a broad mixture of financing systems
and a divided public-private sector delivery structure.
Indonesia also offered the advantage of reDresenting a country enterinv a
significant health care inancing svste= change. Indonesia formally initiated,
during the period of our visit, a social health insurance program :for private


-33-
firs employees (or a Dilot basis in Jakarza), addi`-g to an existing, separate
social insurance rogran     *r cv: servanZs.
Given the very iited :i-e available during our visit, we concentrated our
attention on t-e plans anJ issues concering the new social health insurance
arrangement, known as DMCK (for "Dana Upaa Kesehatan Masyarakat," or Fund for
the 'Health of tne People).  tie :he discussion which tollows will rerlect
concentratIon, we will, for perspective and whatever lessons can be learmed,
aso briefiv review the status of Indc.nesia with regard to other financing sys s 
anc t-e pub.ic-Drivate delivery structure. '.he order of discussion of systems
will be sozewhat altered fro that of the previous section. The discussions cf
private firm arrange.ents and the proposed DUK4 scheme of social health insura.e
tor vrivate :irms are lef: to the end.
4.2 Government vs. Private Sector Provision
1nconesia. health care services are -rovided tnroush faci....t ies operatec
under the authcority of the MC, as well as that ol orovincia7 anc ocal gover-rn:
and also by certain other gover.ment departments (including the armed forces)
priarily for the personnel of these depart=ents. Services are also providedi
substarial voluze, throug  a =odern Drivare sector, as well as through traditiona
oraczitionaars. Private service jrov:sion Ias been reoorted to be larve and
World Bank estinates, for FY 1982/83, indicate that, private sector
providers accounted for 58.6 percent of total expenditure and services provided
through the Dublic sector accounted for 41.4 percent of total expenditure.


-ndonsians predomiatly obtair. zcsi:r ca.re services either frc a pu"li
network of govern.ent hospit als, ieal  centers, sub-KCs, and health posts -
=lostIy az hg1,y suOsdizeafee levels, .s we12 as trouch private practii*~
and private nosit:als - at zucn 1iger                signi:es 'cant volume of
services ifs also believed to be deliverei bt :raditiona! Dractitioners - otten
with pavent-fi-kLnd, though we have srtn i:t e data on this. Retail druig
sales account for a quite high pr:porti:n cf ex:eitures.
Revorts suggest thaz the -rivate role has been expanding for at least the
-ast cecade. 'able 2 , showing the cistribution of nospita7s and hospital heds
ror various government authorities and the private sector, during the perfod
1978/79 to 1983/84, supports the belieJf tat !he private sector role is continz
to grow. The table indicates that dur-'g the five year Deriod, 1978/79 to i9 3
privata sector hospita-l beds ex--anded by 2j ercent, w-lle  bosial oeds o a
=cvern=ent authorities increased by just under á Dercent.  (At the 1983/84 le
private sector oeAs accourted for 30 percent of a11 ceds, and 28 percent ot
total of .0- and local overnent - excluding zilitary plus other departmen: -
Ius private Deds.)
Most prcticing physicians in Indonesia aremDoy ed      t-e governzent and
also wcrk tart-ti-e in mrrvate tractice. Gove-ment :av for enysicians is vide
ac:ow-ecged to be extremely low. As a result, the government accepts the fac:
that pnysicians only work a few hours a day in their goverment job, spending the
re-ainder of the day in private sector employ--ent.  Tis private sector suppe1ment_
to physician salary =ay involve fee-for-service practice with individual clients,
some contractual arrangement with a private fir for delivering care to its
e=ployees, or both of these. Some gover=ment physicians work for the social
insurance progran for civil servants, ASKES, and see ASKES patients at the health


centers dur n- af:ernoon sessions. 2eportelv, the rate paid to these .SKS
pnysicia.s is coubie the ;oveien.t race for regular health center work by
[19]
hysicians.    Many of these o2siciaZs also maintain a privace practice.
GCovernment law for licensing pIysicians requires that they work for the
govern=ent or for a government-rezognizad institution. ·New physicians register
with the Department of Eealth, which then tries to find a job for them. Earli  '
this wasn't a problem; however, the gcverzent sector is now absorbing only abc,
500 new doctors a year of an annua' output of 1,500 to 1,800 doctors per year
[20]
Indonesia.    Wnile waiting for the Department of Health to find a job for the
a phy sician may looI for a j obo n his own - outside of government. If success-
in finding a job outside of government, the physician may seek the agreement cc
the Departzent of ealth - which usually avrees. Hcwever, doctors are said to
Dre:er to obtain a government position, because: (1) Otherwise, their future
c.nances o oefrgn accented for specialist education will be severely Iminisnei
(2) thare is greater security in govern,ent employmen:; and (3) the government
pension is valued.
E..`ective eligiccity to obtain a soecialist education mav be obtaineC vy
working n a goverment healtr center ijakarta for five years, for t-ree years
l cen  elsewhere in Java, for :wo years outside of Java. or for one
year in East Tifor.  (The 'dnesian Medical Association wcul CIke to see ot`er
Incentives emploved for obtaining iysicians in these areas - -ncluding better
salary and housing.) At present, there is said to be a much higher demand for
sp ecialists than for general practitioners in the private sector.


:.e '0, at its highest 'nnzation level, has separate authorities (in
and ounity; and health center care. T-is is -erceived by some as an i     edi-
=ent -or efficient, integrated health care delivery, es=ecially with regard ta
se,rices de 4-ery via the ASK:.s and DC  progr=~s and possible public and priva:e -
sector c orainat .on urc-er tnese 7rogras,   nen-ves    r de...-.very C care at :te
ao-ropriate level, without un.necessary' referals, nay be zore cirr:.cut to obtain.
is reorted that most of the roughly 5,000 health centers in Indonesia are
ad=inistered b'y district level government, thoug  so:e are acnistere: Dy
provincial govern-zent.
.."rether or not health centers are officially e:-ected to 'be cpen to :rovie
se:-r4ces to z-e pub1ic during fuI1 gcvernent          ourking hours (as soe have stated)
was not clear to us. However, mos- persons we talked to acknowledged that -
outsice or the health centers that are coen in the afte-oon for A:KS Kbers -
altr. centers are ore and s ta -rd only in the zornin, erhaps wi:h doctors
-resent for as  ittle as two ocurs.   (We were also toD that doctors na: onl  e
present at the health centers for three days a week. Waiting patients are rezer:ed
to 'e                    ':c)l  cealt with av the doctors, at rates or 0 sa-.ents :. two  ours or
ass.r d)
M'any persons e:coressed the view that sor-rices delivered by public health
centers were regarded as inferior to tnose that igne ce obtaine fron a riva-
physician. The need to wait in line, the very brief tie spent with patients by
the health center doctors, and the lizited hours of doctor availability at the
health centers were suggested factors. It was also reiarked that patients tended
to feel that higher cost services =ust be better and that there is status in
taking drugs - which can be zore readily obtained via private practitioners.


-37-
The Directorate Gareral of !-edical Care exercises authority over both
Covern-ent and private hospitals. The LG of Medical Care issues direactives
applying to quality of care for both the public and private sectors and issues
per=ission to build hospitals. Gevern=ent hospital rates are fixed by local
5overnment. There is no rate regulazion for private hospitals. We were told
that the room and board rate fcr private hospirals ranged from five to ten ti=es
the maximuz for govern-ent Iospitals.
Apparently, the most signicicant hospital regulation is the requirement
that 25 Dercent of all beds should be set aside for the poor. This appears to
nean t-at 25 percent of all beds must be avai1lable at a third class or fourth
class accomodation level. Patients using these beds generally are expected to
pay, though a? rates - established by local (kabupaten) governrent) - whi.c are
elow cost. A governrent fund, Pelayanan Social (Social Assistance Services) =a-
ce usad to provide hospital services free of charge to those who are unable to ray.
lanrous reports.ave  indicated that govene ntospitals are    d-utilized.
Te -ed occupancy rates (BOR's) show  in Table 3 , suggest that utiization
declines with lower levels o govenent control: MOH had a .845 0R in 1985,
corpared to .759 for zrovicial governent hospitals, and about .59 `or the
Z sr c a. na nce'-c ospýDt s.
.:rvate casoi.als seem to have a ver-, his.h deand for their too cass becs
with substantial waiting lists reported for rhese beds. The private hospital
occupancy rate for the lower class beds is apparently much lower, however. The
overall BOR for private hospitals is shown in Table 3  to be .584.
A  private hospital which we visited in Jakarta (Sint Carolus, a non-
profit hospital), however, had an unusually high percentage of beds at third


cial r ecuir=men.  This hospital nevertheless had a highi CVera:  EOR -3 prcen.
Tis hoSDital is one c the private hcs itals conractin   to receie 2 ASK7
Dat ients.  (The ASKES patients represerted 13 pe"ernt of the hospital's Da i2n:
icad,   ring April to December, 1984.) The hospital direczor stated that -Is
costs for third and fourth class beds - cc=zuted for ASKES rainburseren negoi
ticn - were higher CIhan e:xpected costs `or o:er private 'ospitals wi:r lower
percentages of such beds, since he had fewer high class beds with which to subsi-
dize the third and fourth class- beds.
4.3  Fee-for-service FinancL-i
The dc-nant share of private sect:r ser-,ice .rovisin in  ndonesia        s
to be on a fee-for-service basis. A relatively smal. pr oportion of pr-vate sect:
services are '.elieved to be supplied through the coverage arrangements 0 inc.v:-
dual privace firs; and an extre=el sa        oortn is understood to be provizod
through Drva"e ins-urance.  (Bcth of these latter two financig arrangecetsare
discussed later.)
Public sector serVIce coDayent re.uirement s account for a rat-er S=all  D-
tortion of Cublic secor expenditures. The health center fee, called a
tion fee and which officially ent4itles the atien:t to ali seic-,es associated i:h
a health center visit, has for a consieraie time been set at Rus 150 (abcu=
y0.15) . An incentive to collect the fees is given by the entitlenent of health
center staff to retain 10 percent of all fees as honoraria. (Some reports suggest
that this incentive has tended to diminish application of the rule that services
are to be given free to those too poor to pay.)  Besides the official Rus 150 fee,
various other unofficial fees (altogether increasing the total fee by as =uch as


-39-
five times) are said to be frequently collected by the health centers. These
u=orcla-  ees a7 ce re aned 2v health center staf,1.
heeller fund that the cilection or cricia- fees served as a net generc::.:
of revenue by the health centers to district government operation (i.e. the ne:
transfer fro= second level government to the health centers was n.e£ative).
The rea- cost per public nealch center visit has been various-y estimated
for us at RDs 1,000 ($0.90) to RDs 1,700 (n1.55). There are a few private hea
centers in Jakarta, however, where charges are on a component basis - requirin2
some anayis to establish cost comparisons with the official, per visit rates cr
the public health centers.
Wheeler believes that public health center fees conribute to inecuities
in the distribution of healt. serv-ce benefits in that, while representing a
substantial subsidy to users, the fees disccurage use by poorer members of tIe
cocnunity. Ccnsequently, the subsidy is mainly transferred to the relat-.'e-I
r 221
Ber=an'5 study (in central Java) see=s generall- su-por`ive tf this adverse
ecuity- findig. Ber:an found that low inceme Dersons reporting illness werna
-ess likel% to use -ealth center ser~ices than higher incone Dersons. (Low incone
persons were fcund =ore likely to use subcenzer and ealth post services thar
,;ere gher inccme -ersons.) Eeran considered the f0nains to be partia.l;


-40-
related to the greater effect of price on the poor, but also feels that more
decentralized types of service organization, offering better staff familiarity
[23]
and lack of formality may be more acceptable to low-income persons.
According to World Bank estimates, fees paid for private services accounted
for 35.2 percent of all health sector financing in FY 1982-83. Fees paid for
public services accounted for 3.6 percent; and drug purchases from comercial
outlets accounted for 23.1 percent of the total sector financing. [18, p. 5,
table 2.]
It has been remaked that while there is a perception of an insufficient
supply of drugs by managers and providers in the government health sector, tere
[241
appears to be appreciable inappropriate utilization of drugs in Indonesia.
4.4 Government General Revenues
According to World Bank data, "public funds" accounted for 32.6 percent
of all financing in Indonesia during Y 198-33. [18, p. 5, table 2.1 Ve have
not investigated the components or the collection methodology of this funding.
While most government revenues allocated to the health sector are raised at the
national level, much of the actual expenditure takes place through provincial and
district levels of government. The actual allocation of the public funds within
the sector involves a considerable number of separate budgets and funding channels.


The complexity of funding arrangements for public sector services in Indonesia
has been well documented by Wheeler [22] (and raiterated by others). Some of the
za- ;oýints are:   (1) Al-ost all funcions are:fnaced via multiple budets
deriving from =ore t*han one leve! of Jover=ent, with operational and policy
contrel exerted by each; (2) the =ultiple-source funding contributes to poor flexi-
....ity in allocation or runcs between functizns, Door accountabL7 in ex:encitzra
excessive ad=inistrative cost, and aliccation of expenditure which mav not wel.
reElec: stated priorities; (3) finance for   vr enW. ve and promotive activitis of
the jrovincia /local level-ac inisterec nea t centers is obtained from natisnal
sources.  nowever, 'ealth center :ees, =ainly derived frm curative ser-ices, are
recuently a net revenue source to the provincial governent. Consecuently,
inzentives tor prevenzive/oromotive efforts tend to be poor.
While we di  not undertake an-y budgetary investigation, it is our impression
t`at cnly a very s:zalt fract:on of vcve=ent revenues are eaployed as direct
su  -s - z.'- e.r.va- se,tz.-  anc t`.a  ses  ar  nI  d ie..-- c e.ri.a.e
[251
-oszitals.


-42-
4.5  ?riv?ate F< i'nsurre
(i) tro-uction.
:..uh private insurance, ou:size of priva:e fir. arrangeneng-   (wnzc
usuall: do not fnvolve a financial risk interedary) apparently reDresents a-
e:tremly s=all pC-rt-on of total fina.cing arrangements -i indonesia, -se were
intrigued to learn that there were sone efforts to develop 50 type caitation
arrangneents in addition to regular third parzy cc=ercial ins.urance.
(ii) Coercial Third Party.
=xis ti     = cocmeraial third-party healzh insurance plans in 1ndnasia recrtet-
of:er very restricted benefits. Consecuently, providers are reluctant to bil
insurance companies on behalf of the apparent small ercentage of persons holdin
such polcis.    Providers prefer to bill individua1 patients, in order to a-oI-
anticipated disputes arising from partial pay,ent associated with the bei
1i'itations. This, in turn, 1ure        .its the appeal of tiese co=ercia' poli-
cies. Opini'fons offered in explanation o: the ver-; limited :opulaticn coverge `v
co-ercial insurance included:   (1) Few cDoanies offer cr ercial insur ance Vue
to bureaucratic obstacles; (2) the (uncertain?) fee-for-service structure of the
Private sector makes premium computation difficult; (3) the insurance concept in
general is not well appreciated in Indonesia and, in the case of health, Deople
don't consider the possibility of sickness when healthy. We were not able to
investigate the legitimacy of any of these reasons. We did hear of a coupla of
commercial companies that were attempting to expand their health insurance
business in Indonesia. (Attempts to arrange interviews with these companies,
before our departure, were uýnsuccessful.)


-3-
(iii)   R{MO's.
We interviewed directors of two limited membership prepayment capitation H.XO
arrangements in Jakarta. The first arrangement has been limited, by its organizer-
director, to the employees (approximately 5,000) of four related ccmpanies,
resulting in a total membership, including dependents, cf roughly 14,000. The
director contracts with health providers, including private hospitals, which are
paid fee-for-service by the EMO. The capitation fee for members is Rps 10,500 per
month ($9.55) which is paid by the employer for the employee, spouse and up to
three children.  (Additional family members - including those with extended family
relationship to the employee - are allowed to join additionally by each individually
paying Rps 8,000 ($7.25) per month.)
The services are comprehensive, including preventive and promotive care,
health education, immunization, treatment for diseases, glasses, and hearing aids.
?roviders contracting with the HMO bill the HMO directly for services to its
members. There is no cost-sharing, except for drugs, which have a Rps 400 ($0.36)
cooament.
The organizer-director of this RM0 expressed uncertainty about his legal
authority for extending this £!0 arrangement beyond the employees and families
of the currently enrolled firms.
The second arrangement which we contacted was being conducted by a private
non-profit hospital in Jakarta which also has six community health centers in the
area around the hospital. This hospital was operating its HMO experimentally, to
obtain cost and operational experience. Accordingly, membership had, so far,
only been offered to the hospital employees (about 2,000) and their families,
together totalling about 5,000 persons.


-.44-
This HMO concept includes community he lth care, preventive and promotive care,
as well as curative and inpatient care. Patients are screened by the hosoital'S
health centers, in first-line care, prior to hospital referral. H-3spitalization
is at third class level and "sohpisticated measures" are to be excluded.
There have been exploratory talks between the hospital director and at least
one private firm reportedly interested in joining the HMO. ?relfiinary discussions
have also been held with a commercial insurance company which is interested in
marketing the arrangement and collecting premiums on behalf of this M O. Thought
is also being given to the possibility of offering the EYO plan to the new soci:l
insurance health scheme for non-civil servants, DUIKM. (The tentative capitation
cost estimate of the plan of Rps 225,000 annual (Ros 18,750 monthly) seems somewha:
high, however.)
These HO arrangements may well differ from those in the West because the
degree of financial risk assumed by Indonesian doctors appears minimal. None-
theless, the HMO concept is at least evolving in Indonesia.
(iv) Capitation - Outpatient.
A capitation-basis outpatient clinic, operating in Jakarta was visited.
Founded originally to serve expatriates, the clinic is now open to Indonesians
as well. It presently has about 5,700 members, most of whom have been enrolled
by their employers. A fairly sophisticated level of care, which includes preven-
tive medicine, is offered. Operated on a non-profit basis, the cost of membership
includes an entrance fee of $100 (about Rps 110,000) and a fluctuating (depending
on actual annual costs) annual fee of $200 (Rps 192-220,000). Hospital care is,
not included. There were reported to be two more such arrangements, operating in


Jakarta (perhaps catering somewhat more hea:vi1 towards middle-cLass indonesians ),
but we were not able to visic them, or to confirm that they were, in fact, also
operating on a prepayment capitation basis.
The director of the visited clinic expressed an interest in extending the
arrangement to other cities in Indonesia, but felt that government restrictions
regarding the establishment/extension of the arrangement were ratner onerous.
Conflicting statements were offered by cthers interviewed regarding the
willingness of the government to have extensions of such relatively sophisticated
care arrangements. Some understood that the government wishes to prohibit
additional such arrangements, while others said the government concern with
these sophisticated clinics merely involved the requirement that their staffs
be Indonesianized. It was suggested that the client market for these clinics
largely included persons who otherwise would go to Singapore for their health
care.
4.6 Social Health insurance for Civil Servants (ASKES)
The insurance scheme for civil servants is known as ASKES, short for
Assurransi Kesehatan (Iealth Insurance) . The ASKES title for the program was
replaced b7 "Faden Penyelenggara Dana Femeliharaan Resenatan," cr 3PD?K (which
=eans something like Ecard ::pla=enting Health Care Financing), in accordance
with a depar:ure fro the previous concept of the program which we were tcl
involved "operation like a commercial health insurance scheme." A further revi-
sion in the title of the program (relating to further evolution in its organiza-
tional concept), though not yet final, is to "Perum Husada Bhakti." "eru"
identifies the arrangement as operating through a government enterprise with some
autonomy in financing arrangements. That is, it does not operate through a


gov=arzent deart=en:al bu      and can use its revenue to upgrade :eself.    (On
the o t    and, it does not have t'ne ful   autvony of a Vovernmantr enter7rs
oaratad on ror:.t-or. tiza zrceAuras.    `:usada Bhakti" sizOi- zearns
Serices.  Since, newever, tne ÄSKES term is still oopularly usea by govern--
crfcia.s, we wil11 continue to use it here in describing the program.
--   :s cofusory    or vovern=ent =    cployees worki-g at 1 laves c:
government. ° ASKES ~egan in 1963, based on a presidential decrae.     Pri*r :c
1978, the program reaportedly ran deficits. Up to 1978, participants wera aloe
to seek health care from private ractitioners and private hospitals, as       l as
=ro  the governnent puskes=as (health centers) and 'osDtals.    -n 1978, various
restrictions were implemented   includin the necessity to entr tr    systa-
[27]
th-e -uskes=as only and to use only goverment hcs:itals, suppemented cy
cartain orivate hosoitals which  ave been enrolled in the Drocra=.    ,1leven -r va:e
hosoitals in Jakarta are enroll d in AJK-S.)   While AS-"S est-bished a:i
reibursement rates to private practitioners (Rus 1,000 in 1978 - about 32.1C
-hic- wera ca_ow th-ir usual privata rat-s (around Rps 2,000 - about _4.50 in 9L5
these ratas were stiI well above the su`sicized gaovernment Duskesras charve 3c
Rs 130 (about S0.25 ten; S0.14 in Apr-i,   935) - pa:. oy ,1e ceneral uctic anu
Äy S.EKS for visits of its =e:bers to the puskesnas - though :erhas not
above the real o,erating cost of the Zover-rent usesnas.


Origina1 y, ASKES withhed 5 perer.ct of e-jloyee salaries; however the race
was    cd in several steDs and now is at 2 percent of salary.    'h the curr-n:
restriction  to use cf the govern~ent puskesmas anc hospitais `p`us the designar:d
private hospitals with negotiated rates) and the paymenz by ASKES of the subsidized
gcver-nent rate    the ASKES program - even at the 2 percent contribution level -
is running surjluses.
Care in both government hospitals and the designated private hcspita"s is
at third class level. Patients can, if they chocse, buy up to a higher care
level, at their own expense.
ASKES currentIv has 3.5 million employ,ees enrolled, with a total coverage
cf 12.5 millicn including dependents. (This is about 8 percent of the populatirn.)
ASKES is presently, as a "pilot project" excerinenting with allowing higher
ranking e=3loVees to obtain first class hospital acco   dation, without additional
:ay=enz by theb .  (This apparentlv is in response to some household survevs done
iy ASKES which indicated dissatisfaction with the program aong the higher r
e p1oyees due to the third cIass acco=odation restriction.)·
4..7 Cozunity Financin
re operafional of of priary heal:h care in Indonesia is known as
Penger2an2an Keseha:an Masyarakat Desa, PCmD, or Village Co=munity Health
Development. PD represents the effort to Drovide simple promotive ana
curative services in. each village through health posts staffed via community
orzanizations (kaders) which can call upon people for volunteer labor.


-48-
(7he volunteers ar r o be rained and supjorted 'y the Duskesmas.) '"he idv
zua± .Caoers reaca. cec-.sions on tne ne.t r.eeos o: tnae DecDle zr tr.e comu::
a were advised that, acicrdiZg to l,a:est 'OK reports (:btained ChrZughI t-I
puskesmas), PIUD exists i all 27 provinces of incesia, n   69 di fs:ricts,
1673 sub districts and 7,693 villages.  (There are about "65,0C0 villages  n al
oI Ind onesia.)  Since only 60 percent of :he puskesmas reported, it was sugst
ha t the number of f:'D viages -hz be sig3ifcan ly greaer -  erhaDs a t
higher.
An ele=.ent of P:D is Dana Sehat (Health zunc), which involves the dole   :
of runds - t*roul2n =ney or in-kind contr.'Dutions - within the co=un.ity fo:
~ealth care. Ärrange=ents regarding the collection of 2un-s and the usas to which
the funds are put reportedly vary widely. Dana Sehac fu.ds may   e used to pay
health center fees, pay hospital fees, assist travel to a EC, rai=burse health
post volunteers, pay for the space occupied by the health post, etc.
ach co=unity kader deterin. es the aount which is to be collected.
P_cletar; azouns collectad, a co=unit- i? Irian Java which ccllects     32s C0O
(SO.45) per faDily Der month was cited (presumably as a relatively ?rosperous
arrange=ent) ~t--kind contributions also aouear   z '-e connon. Aricultural
roduct     :      o~nmay only De 7iven in the har-7est season.
No szatistics were cobcained raearding the aueer of v4i1ages nich have a
:zctioning Dana Sehat. "e were told that i many villages whera Dana Se-at was
established, it collapsed within a few months. Apparently there have been
greatly varying levels of success in both individual P:"D'and Dana Sehat experience
ano.g the indonesian villages where these have been imiplemented. Individual
successes are eagerly reported. (The in-kind contribution arrangement of a
'illage in Bali, where households produced roof tiles and contributed


tiles to the Dana Sehbt, was reported by numerous interviewed persons.) But
there are also recognized to be many failures.
An Ind3nesian economis: who has reviewed ?KHL/Dana Sehat arrangements, :ld
us that it was difficult to find a general formula for why some arrangements -. re
successful and others not. he agreed with the view of many observers, however,
that a dominant factor was the motivation (some say the charisma) of the local
leadership.
Another, quite pessimistic view is that the kader involves a simplistic
concept of the village, with the incorrect assumption that the villages represent
cohesive systems. The observer expressing this view felt that PK/Dana Sehat
was floundering and was no longer being strongly pushed in Indonesia.
The most optimistic views obtained regarding PKD and Dana Sehat suggested
that these arrangements could provide some resources for the health care needs
of rural communities. But most also seemed to feel that, generally, these arrangE
ments were not likely to make more than a modest contribution to the total rescure
needs of these communities for health care.
4.8 Coocerative Arrangements
We were not able to learn of any cooperative-based financing arrangements
zrom persons interviewed fin jakarta. It was suggested that some of the in-knd
community arrangements, involving contribution of a fraction of a commodity
produced by members of the community, have approached the cooperative form -
with the contribution being collected by the marketing agency for the product.
We did not learn of any of these arrangements which extended beyond individual
communities, however.


-50-
4.9 Private   ir   rrangements
Private employers in Inconesia, uncil now appear to have generally made
individual arrangements for the provision of health care services to their
e=ployees. A sample survey was done of Jakarta fir=s' employee health care
arrangements, to provide information for the planned non-civil servanus 1-l:h
[29 ]
insurance scheme, DUKI.    (The sample of firns was drawn from the records ct e
social security agency for the private sector - ASTEK.) Acoording tothissrvey
171 of 173 responding firms had health care arrange-nments for their emplovees.
Table 4 , which describes the existence of healch care arrangements according
to the number of persons employed by the responding firms, shows that only in
the category of fewer than 100 persons employed were fi=s found which
have health care arrangemen:s for their e=ployees.  (Even in this, lowest defne
category of less than 100 persons employed, more than 95 percent of the resocndcn
irns had arrangements for health care. It would, of course, '-e of interest :o
Obtain this information both for a larger sample of small firms of lass tan
100 employees and for firms outside of Jakarta.)


-51-
The DU2 planning surv7ey also indicated that 80 percent of the respcndin;
firms which had health care arrangements for employees also included employee
families (with some variation in number of family members covered).   The sur 
further indicated that of the firms with health care arrangements, 4 percent ,sad
only facilities owned by the firm, 45 percent used both facilities owned by ta
firm and outside facilities, and about 51 percent used outside facilities nm   :
Table 5 describes the methods of reimbursement used by the fir-s for utilizatm'n
of general practitioners, specialists and hospitals. While tariffs established
by the provider predominate, in many instances rates negotiated by the firm
(service fees, lump sum, per capita enrolled) account for a sizeable fracticn of
those firms providing information. Employee copayment of some sort was used
in only about 9 percent of the plans [30, p. 13, table 23.] The average expen-
diture per employee per year, according to number of employees in the firms is
described in Table 6. The mean expenditure for all firms which provided informa-
tion is shown to be about Rps 133,000 (about $120). Though there is no clear
pattern according to employee size of the firm, it is interesting that the firms
of less than 100 employees had the highest mean expenditure - over Rps 184,000
(about $165).


-52-
(i)   Employer interviews.
In a highly non-random selection of firms, we encountered the following
arrangements - as part of an effort to get some "feel" for what individual
enterprises were doing about health care and their attitude towards the new
DM social health insurance scheme. Constrained by contacts which could be
quickly arranged, we interviewed officials of one private sector  ir and two
government enterprises (not covered by ASKES).
The private firm whose officials we interviewed has 4,950 employees;
including family members, the coverage of their health plan is about 24,750
persons. The company has seven doctors on staff. The doctors are on salary
and each works about two hours a day. Four nurses are also employed. If
referrals outside of the company clinic are necessary, the patient will pay
the doctor and be reimbursed by the company for the full amount - if the referral
was authorized by a company physician. If the patient independently seeks care
from a non-company physician, the company will only reimburse fixed amounts
(Rps 1,500 - $1.35 - for a general practitioner and RDs 3,000 - $2.70 - for a
specialist). The firm has contracts with private hospitals for inpatient care
of its employees and dependents. (They get a 20 percent discount from the hospi-
tals.) The firm's current annual budget for health care is Rps 1,663,200,000
($1,512,000), which is Rps 28,000 ($25.45) per month per employee (or Rps 5,600 -
$5.10 - per month per person covered). This health care expenditure was calcu-
lated by a company official to amount to 16.6 percent of their "personnel budget."
The company president said he was not concerned about the level of health care
expenditure and indicated skepticism about the ability of DUKM to provide adequate
(comparable) benefits.


-53-
The first state enterprise contacted has 7,100 employees; counting
family members, 17,800 persons were covered by its health plan during 1984.
The enterprise employs three doctors who staff company clinics which are used
by head office employees and family. Away from the head office, employees may
go to any private physician and be reimbursed in full by the company. The plan
does not have contracts with any hospitals, nor does it receive any discCunts.
(The enterprise director indicated that because the better - private - hospitals
did such a good business, it was difficult to get discounts.) Employees may go
to any hospital they choose and be reimbursed in full. (The level of accommoda-
tion they may be reimbursed for depends on their company rank.) There is some
cost sharing for drugs - depending on the particular drug (how essential?). This
enterprise's expenditure on health care in 1984 was Rps 2.2 billion ($2.0 million)
which is about Rps 25,500 ($23.20) per month per employee (or Rps 10,300 per
month - $9.35 - per person covered). This expenditure was calculated by
the enterprise as representing 6.33 percent of their personnel expenses.
(Average salary levels appear to be relatively high at this enterprise.) The
director stated that they are quite concerned about the health care expenditure
level and had been considering alternative plans - including only paying 75
percent of employees' health bills, or giving employees a lump sum for health
care.
The second state enterprise contacted has 1,500 employees. Though exact
statistics on number of covered dependents were not readily available, it


-54-
was estimated that total coverage of the health plan - including dependents is
about 6,000. Employees are given a lump sum, equal to about 10 percent of salary,
to cover their outpatient care. Employees can go to any practitioner and pay
the bills themselves out of the lump sum, which they receive whether or not they
have outpatient expenses. Inpatient hospital care, dental care, glasses, and
prostheses are separately paid by the organization, according to the actual expen-
diture incurred, though maximums - related to salary - are set. Health expenditures
for this enterprise during 1984 averaged Rps 37,923 ($34.50) per month per employee
for the 10 percent lump sum payment for outpatient care, plus an average of Rps
1,390 ($1.25) per month per employee for the expenditure in the separately covered
areas. The total expenditure thus amounts to Rps 39,319 ($35.75) per month per
employee, or about Rps 9,830 ($8.95) per month per person covered. This expend-
[311
ture is equivalent to about 10.3 percent of earnings.
4.10 Social Health Insurance for Non-civil Servant Emplovees (D=KM)
Dana Upaya Kesehatan Masyarakat (Fund for the Health of the People), or DUKM.,
[32]
beG:an operation on a pilot scheme basis in Jakr-arta on April 1, 1985.  DUK-M i s


-55-
generally viewed as an extension of the ASKES health insurance scheme for
civil servants to non-civil servant employees. The management of the DUKM1
ASKS
scheme,however, will be separate from the management of VRO. The stated
A
DUKM objectives include improving the health of the Indonesian people
as much as possible, lowering financial risk to individuals by pooling
funds for health care, and improving the efficiency and effectiveness of the
health care system. DUKM will be part of the national health system. Inii.
enrollees, in the Jakarta pilot scheme, will include (by Decree 3318 issued
the Governor of Jakarta) all local government enterprise employees on compul: vy
basis, plus the employees of private firms which join on a voluntary basis -:.e
enrollment decision to be made by the management of the firms. If and when £ K
membership will be compulsory for private firms is yet to be determined. (YcsZ
officials interviewed seemed to think that membership for private firm employees
would ultimately be compulsory.)
DUM is to have the same benefits as the civil servants scheme, ASKES.
Patients will enter the system through the health centers, with referral, as
necessary for specialist or inpatient care at hospitals. A proposal to allow
patients to choose between a government health center or a private practitioner
has been deferred. Patients will be asked to enroll with a specific health cen=e
of their choosing; and will be allowed to shift to another health center at six
month intervals. As under ASKES, hospital accommodation will be at third class
level.
Unlike ASKES, where provider reimbursement is according to the subsidized
fee levels of the government facilities (except for a few enrolled private
hospitals where negotiated fees are paid), the intent of DUKM is to pay full,


-56--
nonsubsidized costs for services. Health center reimbursement will be on an
enrolled capitation basis. 1ospital reimbursement will principally be on a -er
day (package tariff) basis, with certain proce'ures included and others excluded.
(Specific excluded procedures will pres!mably be paid at agreed rates for the
procedures.) Actual reimbursement levels are still being determined.
The division of administrative authority for DUKX, between the Ministr-y
of Health and the Ministry of Labor, was finally resolved - at least in principle
after a lengthy period of confusion and uncertainty, on March 8, 1985 (officially
March 18, 1985) with the signing by the two Ministers of the Surat Keputusan
Bersama (Statement of Joint Agreement), or SKB. The SK3, worded somewhat generall
says that: (1) Health care for workers will be based on the principles of the
national health system and on principles of worker health; (2) a special fund for
health services will be needed which will come from the people for the health of
the workers. The contribution amount will be decided upon by both the inister
of Health and the Minister of Labor; (3) the Ministry of Labor will be resoCnsi`le
:or collecting the funds and guarantees the allocation of the funds according to
a plan mutually agreed on; (4) the Ministry of Health will be resoonsible for
providing the health services for the workers and guarantees that services wi4l
3e gien according to range an cuality mutually agreed upon; (5) policfes and
operational instructions will be further elaborated by a body created by ana
[331
responsible to both Ministers.    (MOH officials freely interpreted all this to.
us as meaning the MOL will collect the money and the MOH will spend it.)


The Ministry of Labor, under its existing social insurance ;lan for the
private sector, ASTEK, collects 4 percent of employee wages (3 percent from
employers, 1 percent from employee). The DUIL premium will be added to the
existing ASTEK contribution. With the SK3 agreement between the two Ministers,
the DUM premium was determined to be 7 percent of earnings.34] (The 7 percen:
contribution applies to the range of salary Rps 70-20,000 ($63.65 - $181-82)
i.e. the lowest premium will be Rps 3,5C0 - ($3.18) - and the highest will be
Rps 14,000 ($12.73). The administrative fee of ASTEK, for collecting the premium,
was not yet known. Presumably, this would be deducted from the 7 percent availabL
to DUKM.
The DUKM premium of 7 percent of salary - compared to only 2 percent for
ASKES - is expected to both pay the full cost of services received by DUKM
participants and also to allow the health centers to upgrade the quality of their
facilities - thereby extending benefits to the general public outside of the DTKM
program who utilize the government health centers. It is further argued that the
substitution of full cost reimbursement for previously subsidized services in the
government facilities, will allow substantial government health budget funds,
presently devoted to these largely urban curative services, to be released for
rural primary care services.


-48-
Since government health care services are considered to be underutilized
at present, TD, officials expect that the extra demand for services under DUK
can be accommodated without stress. Health centers which are now only open in
the morning, will be kept open for an additional two hours for DUKM members.
Depending on various benefit restrictions of DUKM (such as whether or not a
choice of private practitioners will be allowed), some shifts of resource employ-
ment within the health sector may be necessary.
Based on the DUKI planning survey previously described, DUKM is expected zo
offer employers significant cost reductions for providing health service benefit
coverage to their employees. The survey found that employers on average were
presently paying 22 percent of employee earnings for health care benefits, compared
to the scheduled 7 percent DUKM premium. It is presumed (by program planners)
that a substantial part of the cost difference reflects the inefficiency of each
employer separately organizing health services for its employees. To the exten-
that higher benefit levels are represented by the employer plans, the employers
supposedly can compensate the employee for benefit reductions under DUKM. (There
may also be some dislocation costs, especially where employers are providing
services through wholly-owned health care facilities.)
Recent :ax legislation -aking health care benefits provided by an employer
taxable unless provided through insurance, may further encourage employers to
 oin  DUK1M. On the other hand, the present threshold income levels, below which
there is no tax obligation, are fairly high: Rps 240,000 ($218) per month for
a faily of five, Rps 80,000 ($73) per month for a single person. According to
the mean income level of Rps 110,000 per month estimated by ASTEK for Jakarta
(much lower elsewhere in Indonesia), a large proportion of employees would seem
to be unaffected by this tax change.


-59-
The intention is to extend DUMf :o employers who are enrolled in ASTEK,
initially in Jakarta, then in ten other large citi:'es of Indonesia. The pilct
scheme coverage for Jakarta is estimated to be 0.5 million employees, plus
1.5 million dependents, or about 2.0 million total. (This assumes that all of
the private employers are convinced, or compelled, to join.) According to ASTE.
officials interviewed, current ASTEK membership includes 3.1 million employees
from 7,000 enterprises. (About 50 percent of these are on Java Island.) it
was indicated that the present number of employees who could, technically, come
(35]
under ASTEK resnonsibility   is a'out 14.3 million. According to ASTEK's eszimate
of four persons per family, the total coverage with families could amount to about
57 million persons. Adding present ASKES coverage (12.5 million L'icluding depen-
dents), plus the military, would result in about 45 percent of the population
which might be covered at current employment levels. On the other hand, it is
at least doubtful whether ASTEK coverage can approach the indicated potential
responsibility level for some time. We were told that the 1989 ASTEK target
coverage was 5 million employees, or about 20 million with deoendents. Reports of
another observer of the ASTEK operation suggest that the administrative records
for both enrollees and claims are extregely deficient. Enrollment records are
reportedly much out of date and processing of ASTEK claims is understood to be
very late. Assuming this to be true (not confirmed by us) imprpvements in the
records maintenance procedures of ASTEK may be an advisable pre-requisite for
either the extension of ASTEK coverage, or the addition of DUKY premium collec-
tion to the present ASTEK premium.


-60-
Officials involved with the design of DUK  indicated that it is intended
to later extend the program to include unemployed persons in Jakarta, as well az
others, in rural areas, who are not under the potential coverage of ASTEK. it
was suggested that these persons might be enrolled by the health centers (rather
than by ASTEK - which would not be involved in collection of premiums either).
Certain existing subsidies - assisting poor persons to receive care might be
channeled into DUK.M. Also, along this line of thought, the idea was expressad
that the Dana Sehat arrangements might be integrated. These ideas, however, are
evidently in a rather nebulous stage.
1   Summary
Our brief review of the fLinancing arrangements and public-private deliver-.
structure in Indonesia does not pemit us to draw any solid conclusions in termS
of the evaluation criteria OUtlined earlier. Many of the observations that we
have made for Indonesia, however, appear to be consistent with impressions/judg-
ments made regarding these issues in other countries.
Very briefly, public sector deliver: in Indonesia appears to have poor
ooerational and utilization efficiency; in addition, some aspects of funding
efficiency may be unfavorable. Efforts regarding distriDutional equity ara, at
least, imperfect. Ae are uncertain about the status of funding equity.
Private sector delivery gives the impression of somewhat better operational
efficiency, though it is probably also associated with lower distributional equicy
and some utilization inefficiency.
With respect to financing systems for health care in Indonesia, the key issue
at present, is the probable impact of the newly initiated social health insurance
scheme for non-civil servants, DUKM. This change for Indonesia reminds us that


most relevant changes will be incramental, rat'ner than total replacements of
aggregate financing structures. Newly in:roduced systems will, to varying degrees,
augment, partially displace, or produce ccmplementary effects in existing systems.
Thus the evaluation obviously depends not merely on how a given system or struc-
ture ranks absolutely -- relative to all possible systems or structures, but how
it ranks relative to the counterfactual which it augments, displaces, or other-
wise impacts upon.
The immediate question, for Indcnesia, thus is - what will be the different
quantitative impacts of DUKM (regarding augmentation, displacement, or comple-
mentarity) on private firm arrangements, fee-for-service provision and government
general revenue financing, (plus, perhaps public versus private sector delivery)
in terms of the evaluation criteria? For the displacement that occurs for each
of the other systems, can/will DUKM provide net increased funding, funding effi-
ciency, funding equity, operational efficiency and utilization efficiency?
Where there are tradeoffs among criteria, what is the net assessment of the
tradeoffs? In the case of augmentation, is the augmented funding inter-sectorally
allocatively efficient?  (Normally, health sector decision-makers will leave it
to other authorities to worry about this.) For the augmented funding, is there
anything absolutel offensive, in terms of fundig or distributional equity,
sufficient to offset the supposed health status benefits associated with the
augmentation?  (Alternatively, do positive fund-in  or distributional equity and/or
health status benefits due to the augmented funding offset negative net displace-
ment assessment?)
The estimations/assumptions which we learned about DUKM involve an expecta-
tion that it will eventually replace most or all private firm arrangements; and
that some amount of government service costs will be replaced by DUKM and become
available for use elsewhere. We are not aware of any estimate of the amount of


government services or of private fee--or-service provision which might be affected.
,he rationale for the change appears to mainly rest on expectations that relative
to private firm financing displacement, (1) significant operational efficiency
will be obtained (based on data from the planning survey) ; (2) there will not be
significant offsets in funding efficiency; (3) utilization efficiency - in terms
of appropriate health service production modes - can be improved, or, at least,
not allowed to deteriorate as has occurred under other social insurance programs;
(4) savings from operational efficiency enhancement can be used to improve distri-
butional equity. Procedures to deal with individual utilization efficiency appeare
yet undeveloped.
7hile the available data appear greatly insufficient to lend much confidence
to the expectations, hopefully, the "pilot" basis initiation of DULKM in Jakarta
will provide appropriate information in these areas (as well as lessons on how to
-orove performance relative to the evaluation criteria generally), prior to a
decision to extend the program.
Careful collection of data for the DUKM program should also be encouraged and
monitored for possible generalizable lessons for other developing countries.
5.0 Concluding Notes
(1) Advice on how countries may proceed in this area can probably be best
produced by carefully exam4ing the experience of both developing and developed-
countries with respect to the issues of the impact of (incremental) changes in
financing systems and public-private delivery structures. This experience may
be obtained by (a) carefully reviewing and attempting to synthesize existing
reports on country ex-perience; and (b) monitoring the experiments of countries
like Indonesia which are launching major efforts in this area.


-63-
As we have noted several times, there is a large amount of literature now
on the subject but to date we have not found it possible to make the kind of
policy linkages between country studies and reports. Any systematic effort of
the future should try to make measurement issues comparable across countries as
well as to use some systematic analytical framework of the sort developed in
this paper. Also, more country visits by experienced researchers would help
expand the knowledge base of what is going on now in countries.
(2) Somewhat greater research attention should be given to the notion
of evolutionary and developmental trends regarding movements and changes in rina-
cing systems and public-private delivery structures, to develop generaliznble lesso
for countries.
We are struck by how much of a nation's current health care delivery and finan
systems are linked to these non-econcmic quasi-evolutionary notions we discussed in
Section (1.0). It is difficult to make cross country generalizations without some
detailed appreciation for the role played by evolutionary and historical forces
in countries; these tell the researcher much about why something is done and 'ow
likely it is to succeed.
(3) The constraining factors contributing to the disparity between naiocnal
-oals and rhetoric and improved movement towards these goals might be better
investigated, with regard to the possibility that economists might play a greater
role in reducing the disparity.
Economists have for too long neglected the health sector and the tools of
modern economic analysis have a power to contribute to public policy far beyond
the current levels of appreciation in countries. But economists, too, must


-64-
reco,nize that the power of analysis is limited by constraints such as those
noted in (2) above.
(4) Experiments should be encouragec for introducing efficiency incenti, £
(for both operational and allocative efficiency improvement) into areas prese. -2
lacking such incentives. in areas where efficiency incentives have been ignored
because equity was a dominant concern, the tradeoff in lowered equity may be
found to be slight and/or the imorovement in efficiency may be sufficiently large
to allow alternative (less costly) means of equity improvement.
These areas offer potentially important ones for funding agencies. We need
to have real live experiments before it is possible to formulate public policy
in this area.  7unding agencies could well profitably invest in research of this
kind in a wide variety of country settings.
(5) While there may be ooortunities for introducing improved efficiency
incentives into government services provision, the real need for provision of
services by government should be critically examined in each individual area of
such provision. The possibility of less costly private sector provision with,
?ernaos, appropriate subsidies where necessary to correct equity or externalt-y
related allocative efficiency problems should be stressed.
Our survey clearly shows the power of private sector delivery and finance.
Scme exoeriments under (4) above could also shed lighc on where best to ntroduce
economic factors in public provision. Governments fear private sector develop-
ment. And as we have noted, with some legitimate reason. But it is equally
true that private sector arrangements will grow despite what governments attempc
to do. Therefore, both efficiency and equity concerns suggest that a private-
public mix will have different optimal properties in countries. It is an issue


-65-
trat will not go away and it is one where again, lendinz agencies could have a
role to play in funding evaluation research and live experimfents.
It is often easy to say "further research" is needed on a subject as a way
of getting present researchers off the hoot. We say that here but with a
different intent.  Research is going on in country after country but the problem
is i: doesn't add up to anything that can guide policymakers. We need to find
out the details of things going on in countries and why they are done n such
a way. So while much research has been done on the subjects discussed in this
paper, very little of it has applicability across countries and only one study
we know of [131 has attempted to put an analytical cloak around the subject.
What we hope we have accomplished in this paper is to expand both horizontally
and vertically, the cloak of analysis economists have to offer tempered by the
need to get inside a countries health finance system to understand why things
are done as they are. It is this kind of research that is needed.


Table 1
Nu=ber of Countries with Social Security Progras, According to Progra= Types
Represented, Various Years, 1940-1983
i '9CO  1949     19-3    '_'7    1927     '^
Any type of program                        57      58      80      120      129     140
Old age, invalidity, survivors            33       44      58       92      ii4     130
Sickness and maternity                    24       36      59       65       72
ork injury                                57       57      77      117      29      136
21      22       26       31      38       40
Faziy1 a1owanes                           7       27      38       62       65      67
ource:   Social Security Pr ograms 7rugort 0     he vord - 1983, Ri`rc       : e
No. 59, U.S. Department of Health and Hu=an Services, Social Securi'y         -
Admnisraton,April 1984, p. x.


Table 2
Distribution of Hospitals and Beds, According to Government Department and Private Ownership,
1978/79 - 1983/84
1978/79          1979/80         1980/81.         1981/82          1982/83          1983/84
vne rsh I     p      Hospitals/Bed  1ospitals/Bed,   Hospaly/Beds   losp tals/Beds  liosEitals[Beds  Hospita hleds
Inistry of Health       43/13,103        45/14,167        48/14,303        50/ 14,165      53/ 15,232      53/ 15,232
rovinces/Regencies     310/32,705       323/33,022       333/33,722       340/ 34,400     339/ 35,440     339/ 35,769
Municipalit les
inistry of Defense     183/14,370       183/13,638       171/13,258       164/ 13,833     141/ 12,543      140/ 12,515
ither Departments        86/ 9,665       88/ 9,637        88/ 9,531        85/  9,566       85/  9,021      86/ 9,046
otal Government        622/69,843       639/70,464       640/70,814       639/ 71,964      618/ 72,236     618/ 72,562
rivate                 547/24,788       551/26,275       568/27,729       600/ 29,045      628/ 30,676     655/ 31,119
otal Government
nd Private            1,169/94,631    1,190/96,739     1,208/98,543     1,239/101,009    1,246/102,912   1,273/103,681
ource: Adapted from data in table, "Perkembangan Jumlah Rumah Sakft Dan Teinpat Tidur, 1978/79 - 1983/84,"
supplied by Directorate General of Medical Care, Department of Health, Republic of Indonesia.


Tabe 3
Bed Occupancy Ratio and Average Length of Stay in 'ospitals According to
Government Department and Private Ow-nership, 1983
Bed Occu-ancy                  Average Length
Cnershb                        Ratio (%)                    of Stay (Das)
inis try of 1ealth                 84.5                           8
Provinces                          75.9                            7
M-unicIalities                     59.4                            6
Regencies                          53.5                            6
Yinistrv of Defense                50.3                            3
Othcer Depart=ents                 49.4                            7
Zrivate .8.                                                        7
Averaze for All                    65.5
Source: Extracted from table, "Perkiraan Kegiatan Rawat Jalan and Rawat Tinggal
RS Menurut Pengelola, indonesia Tahun 1983," supplied by Directorate
General of Medical Care, Department of Fealth, Republic of Indonesia.


Table L
istribution of Firms, With or Without a Health Se-ices Arrangement 'or their
Employees, Aczording to No. of Employees in the Fi=
o. of Em:ovees                  umber of Comani4es
With M..        without H.S.
Arrangement        Arran2emen:          Total
2,000 +                           6                  0                  6
1,000 -                          11                  0                 11
5C -                          24                  0                 24
100 -                          37                  0                 87
Less than 100                    43                  2                 45
Total                           171                  2                173
Source: "Lapcran Sementara Penelitian Pelayanan Kesehatan Bagi Karyawan Perusahaar
Di DKI-Jaya," (Temporary Report Study for the Health Services for the
Workers of the Private Companies of Jakarta) p. 2, Table 1.


abl   5
System of Reimbursement Used by Fv     s for Gneral 7ractittiers, Speciaists      2
z:spitals
Numbrer of Firs
Ri:busement Syt_                    G.P.'s              S___ali_s_ s                -
Lun-su    per  onth                    13                     4
Rates set by firms                     28                     24                20
Rates se: bv rovider                   96                   108
Fer head enrolled                        0                     0
Mo inf rza ion                          27                    28                 i
Total                                 164                    -64               158
Source: Laporan Sementara Penelitian Pelayanan Ke.sehatan Bagi Karyawan Perusahaan
Di DKI-Jaya," (Temporary Report Study for the Health Services for the
Workers of the Private Companies of Jakarta) p. 6, Table 9.


Table 6
Mean A=cunt of    penditure f   eath, Per Eploee, Per Year, According to
No. of E-aployees in the Firm
No. of Firas
No. cf E   oyees          in Sam:e           Mean Ex--endiure        Standard
2,000 ÷                        2            Rps 106,000 (96.35)     R?ps 94,000 ( 3.
1,000 -                        6             Rps 162,600 ($147.80)   R=s 77,695 (J0.
SCO -                       11            Ros 88,600 ($80.55)     R3s 25,933 (S2.
D -                        44            R       ,234 (S96.5)    Rs 17,      (
Less than 100                 30             Rps 18,210 (3167.45) Rs 41,A02    37 .
To tal                        93            Rps 132,933 ($120.35)    Rps 16,997 (515.
Source: "Laporan Sementara Penelitian Pelayanan Keselatan Bagi Karyawan Perusa-aan
Di DKI-Jaya," (Temporary Report Study for the Health Services for th
Workers of the Private Companies of Jakarta) p. 13, Table 24.


Acknowledgements - The authors wish to ispecially thank David de Ferranti
(world Bank) and are appreciative of asSistance and suggestions offered by
C. Tamburi (I.L.O.), Brian Abel-Smith (London School of Economics) and numerous
persons interviewed in Indonesia and others at W.H.0., Geneva. The authors alone
are responsible for any errors.


References
1.   By natural evc,ltinary7 changes we refer to changes tending tc be fostere-
by economic development, or other societal impacts - but especially change
that are not fostered by the final category of factors, i.e. active inter-
pretation of efficiency and equity (et. al.) criteria.
2.   Just as existing systems of financing and delivery structure may be diffic
to change, due to political and institutional inertia factors - as well as
entrenched lobbies associated with these systems; system changes, when the"
do occur, may develop "traps" - determined by the same sort of factors - w-:
make reversals of the change extremely difficult to obtain.  Forewarned e1.t:S
at avoiding unfavorable evolutionary changes (or guiding the changes along
more desirable paths) may be much more feasible.
3.   Besides clear problems in comparability of aggregate expenditures between
countries, the correlation between these expenditures and delivered health
services was uncertain.
4.   At the time of the study, the proportion of GNP spent on health in the
countries studied varied from 2.5 percent to 6.3 percent.
5.   Abel-Smith, B., An International Study of Health Exnenditures, and Its
Relevance for Health Planning, World Health Organization, Public Health Paper
No. 32, Geneva, 1967. The findings described here are only a selected few
of the study results.
6.   This classification is drawn from: Roemer, 'M.I., Comoarative National
Policies on Health Care, Marcel Dekker Inc., New York, 1977, pp. 13-20.
7.   This exransion  is undoubtedly more pronounced in terms of expenditure propor-
tion than proportion of the population served.
8.   Bicknell,W.J.and Lebowitz,A.G.,"The Public's Health and Private Care: The
Private Sector in Developing Countries," Health Policy Institute, Boston
University, Boston, mimeo, n.d.
9.   Authorities may wish to consider the stability of a funding source in addition
to its anticipated incremental funding contribution.
10.   This kind of externality, involving offense to people's sensibilities because
others are ill, is separate from the more allocative-oriented externalities
associated with public health disease transmission issues. While the
societal concern for distributional equity presumably is directed to equity
of (realizable increment in) health status, the concern is normally opera-
tionally expressed in terms of equitable access to health services.
11.   One study giving some attention to this aspect was done by Peter Berman in
Central Java, Indonesia. Berman's findings are described in a later section
on Indonesia. See reference 23.


References (crr.tinued)
12.   A seventh criterion, "macroeconomic impact," covering effects of arrangements
which extend beyond the health sector (such as, impacts on aggregate savings
or on general inflation) might have been added. However, it was felt that,
given present analytic capabilities in this area, attempted evaluations
according to this criterion were not likely to be very useful for developing
countries.
13.   de Ferranti, D., "Paying for Health Services in Developing Countries,"
Population, Health and Nutrition Department, World Bank, September, 1984.
14.   Pauly observes that these changes represent rational economic behavior - not
moral perfidy. Pauly, M.V., "The Economics of Moral Hazzard: Comment,"
American Economic Review 53 (June 1968), 531-537.
15.   Unlike the pattern of some industrialized countries, firms in developing
countries less commonly provide commercial health insurance policies to
their employees. The individual arrangements of firms may either (according
to contrary arguments offered in one country - Indonesia) be (1) the result
of the absence of a significant commercial market; or (2) a contributory
cause of the lack of a viable commercial insurance market.
16.   Roemer,M.I.,"Social Security for Medical Care: Is It Justified in Developing
Countries?" International Journal of Health Services, Vol. 1, No. 4,
November 1971, pp. 354-361.
17.   Musgrave, R.A. and Musgrave, P.B., Public Finance in Theory and Practice,
Fourth Edition, McGraw-Hill, New York, 1984, p. 278. Musgrave argues that
the result will tend to be the same for competitive and non-competitive labor
markets, assuming all parties were maximizers before imposition of the payroll
premium.
18.   "Expenditure and Financing Issues in the Health Sector in Indonesia," Draft,
World Bank, December 1983, p. 7.
19.   Kraushaar, D.L., "Estimating Ambulatory Care Costs Under a Proposed National
Health Insurance Program in Jakarta," Indonesia, Thesis, Johns Hopkins Univer-
sity, Baltimore, 1984, p. 258.
20.   Personal communication with Dr. Azrul Azwar, MPH, Secretary General, Indonesian
Medical Association.
21.   Personal communication with Daniel L. Kraushaar.
22.   Wheeler, M., Financing Health Services, Institute of Local Government Studies,
University of Birmingham, England, December 1980.
23.   Berman, P.A., "Village Health Workers in Java, Indonesia:  Coverage and Equity,"
Social Science and Medicine, Vol. 19, No. 4, 1984, pp. 411-422.


References (continued)
24.   About two years ago the number of separate drugs supplied to the gcvernment
health centers was reduced, in accordance with an essential drug list. It
is suspected that this has resulted in a significant increase in the number
of referrals to hospitals, due to the wider number of drugs available from
the hospitals. Consequently, there is consideration of returning to the
earlier, wider availability of drugs in the health centers.
Kraushaar has observed that there seems to be a certain status associated
with taking drugs and suggests that the wider availability of drugs from
private practitioners may contribute to the preference for treatment from
private practitioners. (Personal communication from Daniel L. Kraushaar.)
25.   The director of a private hospital visited in Jakarta, estimated that the
present annual subsidy received from government of Rps 13.5 million
($12,300) represented less than two-tenths of one percent of the hospital
budget.
26.   Various government agencies (apparently those with some autonomy) are not
included in ASKES, however.
27.   Patients can use private practitioners, if willing to make direct out-of-
pocket payments to the private practitioners and seek the fixed amount
reimbursement from the government. (There may be further limitation on
this.)
28.   See reference 18, pp. 62-67 for a discussion of the probable magnitude of
the subsidy obtained by ASKES by paying the standard rates.
29.   The results of this survey appear in the unpublished document, "Laporan
Sementara Penelitian Pelayanan Kesehatan Bagi Karyawan Perusahaan Di DKI-Jaya"
(Temporary Report Study for the Health Services for the Workers of the Private
Companies of Jakarta). We believe this survey was conducted by a team
representing the Department of Public Health of the Medical School of the
University of Indonesia, ASKES (BPDPK) and ASTEK. The survey information is
for 1983. The number of firms contacted by the survey was 250, however
responses were only obtained from 173 firms. We did not have the opportunity
to review the methodology according to which this survey was conducted -
including what procedures, if any, were employed to verify the validity of
the reported data. (The data apparently was supplied by firn managements in
response to questionnaires.) One Indonesian (of non-official status) expressed
great skepticism to us about the validity of the reported information, based
on (1) the belief that firm managements would probably not respond candidly
on this subject; and (2) the opinion that reported benefit levels were much
higher than believed to be the case.


References continued)
30.   "Laporan Sementara Penelitian elayanan Kesehatan Bagi Karyawan Perusahaan
Di DKI-Jaya," (Temporary Report Study for the Health Services for the Workers -
of the Private Companies of Jakarta) p. 5, Table 7.  (See reference 29.)
31.   For both the private firm and the first state enterprise, health care
expenditure amounts were calculated by these organizations with reference
(apparently) to their total personnel budgets, which presumably include the
health expenditures and other perSoanel welIare costs. Consequently, the
comparable expenditure percentage for tc.e second state enterprise personnel
would be lower - probably under 9 percent.
32.   While officials associated with the MOH stated that the DTIKM scheme had begun
on April 1st, it appeared to us that effective operation was delayed, pendinc
resolution of certain issues with the Ministry of Labor and final decisions
on a number of operational elements - including specific remuneration levels
for service facilities and practitioners. Uncertainty about the specifics
of DUK-M operation was also delaying the decision of private firms to join
(according to the present voluntary basis of the scheme for private firms).
Earlier 104 firms had indicated willingness to join; however, all except
eight of these had withdrawn their intention to participate (as of March 29th)
pending clearer definition of the DUK'M arrangement.
33.   "Surat Keputusan Bersama Menteri Tenaga Kerja Dan Menteri Kesehatan,"
March 18, 1985.
34.   The contribution previously was intended to be set at 7ps 2,500 ($2.27) per
person per month. There was some debate whether the amount of actual income -
including incentives and bonuses - or the number of dependents might be more
likely to be understated. We were told that the two methods of premium collac-
tion were expected to yield similar revenue amounts, based on assumptions of
average income in Jakarta of Rps 110,000 per month (.07 x Rps 110,000 = Rps
7,700) and average family size of 3 persons (Rps 2,500 x 3 = Rps 7,500).
(However, we've also received estimates of both larger average family size
and lower average income.)
35.   ASTEK social insurance is compulsory for firms that have at least 25 employees
or a payroll of at least Rps 1 million/month. ASTEK expects that firms smaller
than this might not maintain accounting records sufficient to handle the ASTEK
premium.