UBC Faculty Research and Publications

Charging Peter to pay Paul : accounting for the financial effects of user charges Evans, Robert G.; Barer, Morris Lionel, 1951-; Stoddart, Greg L. 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Evans_R_et_al_Charging_Peter.pdf [ 1.68MB ]
JSON: 52383-1.0048422.json
JSON-LD: 52383-1.0048422-ld.json
RDF/XML (Pretty): 52383-1.0048422-rdf.xml
RDF/JSON: 52383-1.0048422-rdf.json
Turtle: 52383-1.0048422-turtle.txt
N-Triples: 52383-1.0048422-rdf-ntriples.txt
Original Record: 52383-1.0048422-source.json
Full Text

Full Text

Charging Peter to Pay Paul:Accounting for the Financial Effects of User ChargesRobert G. EvansMorris L. BarerGreg L. StoddartHPRU 93:170 DECEMBER, 1993HEALTH POLICY RESEARCH UNITCENTRE FOR HEALTH SERVICES AND POUCYRESEARCH429· 2194 HEALTH SCIENCES MALLUNIVERSITY OF BRITISH COLUMBlAVANCOUVER, B.C. CANADAV6T1Z3The Centre for Health Services and Policy Research was established by the Board ofGovernors of the University of British Columbia in December 1990. It was officiallyopened in July 1991. The Centre's primary objective is to co-ordinate, facilitate, andundertake multiclisciplinary research in the areas of health policy, health services research,population health, and health human resources. It brings together researchers in a varietyof disciplines who are committed to a multidisciplinary approach to research, and topromoting wide dissemination and discussion of research results, in these areas. TheCentre aims to contribute to the improvement of population health by being responsive tothe research needs of those responsible for health policy. To this end, it provides aresearch resource for graduate students; develops and facilitates access to health and healthcare databases; sponsors seminars, workshops, conferences and policy consultations; anddistributes Discussion papers, Research Reports and publication reprints resulting from theresearch programs of Centre faculty.The Centre's Health Policy Research Unit Discussion Paper series provides a vehicle forthe circulation of preliminary (pre-publication) work of Centre Faculty and associates. It isintended to promote discussion and to elicit comments and suggestions that might beincorporated within the work prior to publication. While the Centre prints and distributesthese papers for this purpose, the views in the papers are those of the authorts).A complete list of available Health' Policy Research Unit Discussion Papers and Reprints,along with an address to which requests for copies should be sent, appears at the back ofeach paper.UBC CENTRE FOR HEALTH SERVICES AND POLICY RESEARCHDISCUSSION PAPER HPRU 93:170CHARGING PETER TO PAY PAUL:Accounting for the Financial Effects of User ChargesRobert G. EvansDepartment of Economics, andCentre for Health Services and Policy ResearchUniversity of British ColumbiaMorris L. BarerCentre for Health Services and Policy Research, andDepartment of Health Care and EpidemiologyUniversity of British ColumbiaGreg L. StoddartCentre for Health Economics and Policy Analysis, andDepartment of Clinical Epidemiology and BiostatisticsMcMaster UniversitySeptember, 1993This work was funded by the Ontario Premier's Council on Health, Well.belng andSocial Justice. Responsibility for the views expressed herein, and any errors oromissions, rests solely with the authors. R.G. Evans is supported by a National HealthScientist award from Health and Welfare Canada, and is a Fellow of the CanadianInstitute for Advanced Research. M.L. Barer and G.L. Stoddart are an Associate and aFellow, respectively, of the Institute. The authors wish to thank the many individualsboth Inside and outside the health care system who have taken the time to share theirviews on user charges.PrefaceTbis is one in a series of articles by the authors about the ongoing debate overuser charges in the Canadian health care system.In this paper we outline a formal and comprehensive analytic framework inwhich income trausfers • the principal effects of user charges - can be traced betweengroups in the population (e.g. the healthy and the sick, the rich and the poor), betweenpayers and health care providers, and among providers. The accounting relationships inthis framework, expressed as equatins, are a sub-set of those underlying the economy­wide national income and product accounts. They make it clear that "it is impossible todo ouly one thing" when charges are imposed -- or removed. The framework is used toanalyze the patterns of income transfers associated with different types of user charges.Other papers in this series focus on other aspects of the user charge debate,including the popular arguments in support of, the common rationales for, the keysupporters of, and the principal gainers and losers from such charges. A briefdescription of each paper follows."The Remarkable Tenacity of User Charges" documents the history of the usercharge debate in Canada. It reviews the participation, positions and rationales ofCanadian interest groups in debates over "patient participation" in health carefinancing."Who Are the Zombie Masters, and What Do They Want?" likens the recurringproposals for user charges to zombies - the so-called 'walking dead' - because althoughthey have been repeatedly rejected by policy-makers and the general public (and thesubstantive claims of their supporters refuted by analyses of the effects of such charges),these proposals refuse to remain buried. This paper examines why that is the case, andwho stands to benefit from the introduction of user charges."User Charges, Snares and Delusions: Another Look at the Literature" reviewsand extends an earlier in-depth analysis of the effects of user charges which three of theauthors published in 1979. The paper assesses whether experience and publishedliterature in the years since then alter any of the (largely negative) conclusions of theearlier study concerning the ability of direct charges to patients to achieve importantpublic policy objectives, including controlling health care costs.It's not the Money, It's the Principle" examines why user charges exist for somehealth care services and not for others. The paper analyzes the characteristics ofservices which (do or should) underlie decisions to charge in part or in whole for specifictypes of services. We propose a framework for evaluating the justification for, andfeasibility of, user charges for particular types of services in particular situations."Why Not User Charges? The Real Issues" examines some of the most frequentlyheard arguments for user charges and looks at what evidence there is for claims and2counter-claims that are often made. Because statements in the "popular" debatesometimes seem incousistent with each other, or unrelated to or at odds with the facts,we explore the statements more carefully, asking what they really mean, what valuesthey are based on, and what fundamental issues are at the heart of the user chargecontroversy. This paper in particular is intended for a wide general audience andassumes that most readers will have heard • or perhaps made • the arguments described,but will have little detailed or technical knowledge of the issues involved.In addition, a bibliography entitled "User Charges in Health Care" provides anexteusive set of references to articles of relevance to the user charge debate in canada,drawn from diverse sources including academic research and policy analysis literature,the popular press, government documents and reports, and the publications and reportsof non-governmental organizations including the professional associations representing avariety of health care providers.Gainers and Losers: Who Pays, Who Is Paid, and Who Is Cared For?User charges are one of the ways In which revenues can be raised for the healthcare sector. Whatever other effects they may have -- encouraging people to be moreresponsible In their use of care, preventing people from getting access to the care theyneed, or any other rhetorical claim -- they always trausfer funds from those who usecare, either to those who provide It, or to those who pay those who provide it. Thus anyunderstanding of their impact must include an awareness of their relative effects ondifferent people in the society. Most of the political debate over user charges includes,implicitly or explicitly, a concern for their distributional effects -- who gains, who loses?This immediately raises the question, "Gains or loses relative to what?" to whichthe answer is, "Relative to some other way of financing health care." No one seriouslybelieves that health care, or any other resource-using activity, is "free". Abstractingfrom minor revenue sources such as charities or lotteries (both of which in fact alsoinclude an indirect diversion from government revenue), there are three principalchannels through which funds can be raised to pay for health care: taxes, user charges,and insurance premiums.Modern health care systems are financed primarily from tax revenue. Forpolitical palatability certain taxes are sometimes labelled "social insurance premiums",and may be collected either by governments or by agencies which are at half arm'slength from governments - in essence quangos.' But there is relatively little trneinsurance in health care finance. The one country which appears to rely primarily onprivate insurance, the United States, actually provides large tax expenditures (currentlyestimated at about $40 billion) from federal government revenues to the privateIndustry.' Table 1 shows the percentage of health expenditures from public sources forthe twenty-four countries of the OECD [2]. This percentage is over three quarters forWhen payments are compulsory, unrelated to the riskstatus of the "insured", and not in fact a condition ofentitlement to care, then you may call them what you will butthey remain taxes. They may, however, be a particularlyregressive form of tax.2 This public subsidy is probably critical to the survivalof the 'private" American health insurance industry at least inits present form. Even with the subsidy, while private insurancecovers more people, public programs pay more of the bills.Private health insurance covers only 32.5% of U.S. healthexpenditures, compared with 43.9% from public sources [1]. Ifone transfers the $40 billion in tax expenditures from privateinsurance to public expenditure, the proportions change to 27.2%private insurance, 49.2% public spending. (The percentagesdiffer from those reported in Table 1 because the OECD usesslightly different spending definitions.)TABLE 1International Comparison of Public Shareof HealthCare SpendingAustraliaAusuiaBelgiumCanadaDenmarkFinlandFranceGermanyGreeceIcelandIrelandItalyJapanLuxembourgNetherlandsNew ZealandNorwayPortugalSpainSwedenSwitzerlandTurkeyUnited KingdomUnited States198063%69%83%75%85%79%79%75%82%88%82%81%71%93%75%84%98%72%80%93%68%27%90%42% (7)199068%67%83%73%83%81%74%73%76%87%75%76%72%91%71%82%95%62%78%90%68%36%84%42% (48%)** adjusted to include tax subsidySOURCE: GJ. Schieber, r.r. Poullier, and L.M.Greenwald "U.S.health expenditure performance: An international comparison and dataupdate" HealthCare Financing Review 13:4(Summer, 1992)2thirteen countries, and between two-thirds and three quarters for another eight. Onlythe United States and Turkey are under one half.However it is raised, the total amount of revenue provided for health care in anysociety is linked by a fundamental accounting identity to two other importantaggregates. It must always equal total expenditures on care -- the total cost of the carereceived and provided •• and it must equal the total of all the incomes earned from theprovision of care by those who directly or indirectly participate in -- supply reimbursedresources for •• Its provision. Schematically:TOTALREVENUETOTALequals EXPENDITURETOTALequals INCOMESEach dollar (pound, franc, mark...) spent on health care must simultaneouslyhave come from someone, and been paid to someone. The relationship need not, ofcourse, hold for anyone of the Individuals In the society. Most people will contributeeither more or less on the revenue side than the cost of the services they use; on theother side of the equation the amount that people earn from providing health care willtypically be either much more, or much less, than they contribute to pay for It. But inaggregate, summed over all the individuals making up the society as a whole, theequation must hold as an Identity.'Within this relationship, Individual persons may playa number of different roles,using or providing, and paying or being paid for, many different forms of health care Inmany different ways. In delineating the various redistrlbutional effects of user charges,we identify the reallocation of costs and benefits among the members of society underfive heads, as changes In:1) The share of costs borne by different payers,2) The share of services used by different persons or groups,3) The incomes of providers, relative to the rest of society,3 This presupposes that the society in question is a·closed" economy, at least with respect to health care. Thepeople who live in it neither purchase health care fromoutsiders, nor sell it to them. This is ·almost· true forhospital and medical care in Canada; the cross-border flows ineither direction are minuscule (though they receive adisproportionate amount of attention, both in and outside Canada,for political reasons) [3]. For medical equipment and drugs, ofcourse , there is a substantial international trade. This can berepresented in the framework below by adding a "Rest-of-the­World" sector as an additional (large) ·person" or set of personsin the economy.34) The relative incomes of different providers or provider groups, and5) The numbers and types of 'providers' reimbursable, directly or indirectly, fromhealth care budgets.In each case, we can show the characteristics of persons which determine whetherthey will gain or lose from user charges. What is notable is the number of differentways in which gainers and losers can emerge. This in itself is indicative of the breadthand diversity of the potential pro-user charge constituency; there are many potentialgainers, in many different roles. On the other hand, the framework which is implicit inthe relationship above, and which we will develop more formally below, shows that allthese gains are at someone else's expense •• and we can keep track of whose.4People and Services: The Algebra of UtilizationTo represent the different individuals in the society of interest, we can assign eachof them a unique integer number i between 1 and N (inclusive), where N is the totalnumber of people in the society (for Canada, about 28 million). A randomly selectedperson from this society can then be referred to as person i.Out of all the commodities, goods and services, produced and used in this society,a certain subset are designated as "health care". How these commodities come to beidentified as "health care", while others are excluded from the designation, is animportant question. It lies behind sometimes very heated arguments over the "correct"definition of health, and over which 'providers' should therefore be entitled toreimbursement from public programs. "Entitlement" means different things to peopleon opposite sides of the health care transaction. "Health care", and its providers,receive very special regulatory and financial treatment, so the designation carriesimportant and coveted advantages.But for the purposes of this discussion we can bypass this issue. (We address itelsewhere, see [4].) Whatever the current definitions happen to be, we let the number ofdifferent commodities which they include -- such things as doctor visits, surgicalprocedures, days in hospital, prescriptions filled, and so on -- be represented by M. Wecan make the descriptions of these commodities as specific or as general as we wish, bymaking M large or small.For a realistic description of a modern health care system M would have to bequite large to distinguish different types of professional contacts, procedures, etc. Amedical fee schedule, for example, or a list of diagnosis-related groups of hospitaladmissions, or a catalogue of prescription drugs, each represents a sub-set of M; wewould have to add the numbers of different items in each of these lists together to beginto approach a value for M. On the other hand, for some rather crude representationswe may use a very short list, in which a contact with a physician is treated as a singlekind of commodity, a filled prescription is another, and a hospital bed-day a third.Small values of M correspond to highly aggregated and heterogeneous "items" of'service.As we did for the persons, we label each of these commodities with a uniqueinteger j between 1 and M, and we designate a randomly chosen commodity from this(long or short) list as health care of type j.We can then designate by <ilJ' the total quantity of health care of type j receivedby person i (during a particular time period, say the year 1993 -- one could add anothersubscript t to designate this period). The total quantity of health care of type jproduced in this society in this time period will then be the sum of the amountsprovided to each of its N members. If we call this total QJ' then:5or in a more compact notation:In any particular year, many of the IJJJ will be zero •• if person i did not use anyof service j. If she were really healthy IJJ.J might be zero for all values of j .- person ireceived/used no care of any kind. But the total volume of services of type j providedand received in this society will be represented by this summation.Mathematically we can carry out a corresponding summation to represent thetotal volume of services received by person I, as by:But this procedure is Invalid because we are now adding up apples and oranges -- ormuch worse. As the Index j takes on different values, it counts numbers of office visitsto doctors, laboratory tests, days in hospital, open heart procedures... The meaulng of asimple total of "services" is not obvious. Two doctor visits, three lab tests, and twoprescriptions in the course of a year equals seven "services", as does a triple by-passoperation and six days in intensive care. But we would not want to treat them asequivalent "packages" of services, in any sense. In this case aggregation leads tostatistical nonsense; it can also be a potent source of conceptual nonseuse.Nevertheless it is quite common for payment agencies to report the "averagenumber of (fee-for-service) procedures received per beneficiary" in a particular year,and for newspapers to report this as if it were a meaningful number. And it may be,provided the mix of services has not changed -- but one cannot tell from the servicecount itself.The legitimate calculation is to assign each procedure a 'weight' to make themcomparable with each other. For surgical procedures, for example, a "hernia­equivalent" scale has been developed such that each operation can be assigned a valuereflecting its time and complexity relative to the repair of a hernia. Hernia repairsthemselves are given a value of 1.0, More complex operations are multiplied by weightsgreater than 1.0, and less complex ones by lower weights, before adding up a total of"hernia-equivalent" procedures. This set of weights is an example of a "Relative ValueScale" or RVS, which converts items measured in different units Into a common andtherefore summable unit.A more common source of weights is prices, which have the dimension "dollarsper" whatever the unit may be in which a good or service is measured. If each of theM different health care services in this society can be assigned a price, say PJ, then the6product PJ*<L.J is the value, In whatever currency the price Is defined in, of the services oftype j received by person I, In this society (during the time period under consideration).This leaves open, of course, the question of the underlying basis for the fee schedule orprice list itself; do the fees reflect any meaningful characteristics of the differentcommodities or services themselves? The current Interest In "Resource-Based RelativeValue Scales" (RBRVS), which attempt to develop weighting systems for physicians'services correlating better with the actual resources -- time, effort, skills, andmaterials/equipment - used up in their provision, arises out of this concern.'The summation: LJ[PJ*llt~now has a perfectly valid meaning; it is the total value in dollars of all the healthservices of whatever type received by person I, the total cost of her care. And we cansum these expressions across all individuals, thus:L1J[PJ*<L.Jl = LJ[PJ*LJlltJl = LJPJ*QJ ...(1)The total costs of health care in this society are equal to the costs of care of eachtype j -- doctors' services, hospital care, prescription drugs... -- added across all M types.And the cost of each type of care j is defined by the total volume of that form of care,QJ' measured In physical units, multiplied by Its price (cost, value) per unit, p/4 Of course this process itself requires the use of animplicit or explicit weighting of the relative values of thedifferent resources used in providing each type of care or 'feeitem'. And these, too, may be based on prices which do notnecessarily reflect the relative values that might emerge in"free and perfect" markets. Nor would we necessarily want themto.5. This of course assumes that the Pj used in the summationactually measure the unit costs of the corresponding services.For some services, those of physicians in a fee-for-servicereimbursement system, for example, the selection of theappropriate Pj is straight-forward -- they are the fees paid (bywhomever) to the service provider. But for organizationssupported by budgetary allocations, such as Canadian hospitals ormedical clinics in the British or Swedish systems, the unitprices/costs of the many different types of service provided willnot be defined. Nevertheless, in principle these implicit"prices" exist, and can be extracted with a sufficientlysophisticated accounting system. (But perhaps not tQQ .sophisticated. We are aware of the conceptual problem ofallocating inherently joint costs, but in the real world thoseco-exist with explicit pricing systems and few analysts balk atusing the associated prices.)7People and Payments: The Algebra of RevenuesAs pointed out at the outset, however, this total will also and by definition beidentically equal to the product of two other ways of aggregating up to the total costs ofhealth care. It will be equal on the one hand to the total revenues which are raised topay for health care, and on the other hand to the total incomes earned by those whosupply the services used in health care production.Taking the revenues first, we pointed out above that the most common way ofpaying for health care is through taxation. Funding may be drawn from generalrevenue •• e.g. Canada, the U.K., or Sweden •• or there may be a separate "earmarked"tax system with its own base and rates, often referred to as "sodal insurance" •• e.g.Germany, France, the United States •. which is then supplemented out of generalrevenue.(Quite) roughly speaking, the total burden of taxation (not just the personalincome tax) can be thought of as falling upon individuals more or less in proportion totheir Incomes," If we consider again person i in this society, and refer to her income asYI' and to the components of the general average tax rate which go to pay for healthcare as t, then person i pays tYI for health care, through the tax system. Summingacross all persons will then yield the total of public spending on health care.'The proportion of health care revenue which is raised through direct charges tousers •• in Canada, this includes much of dental care and of prescription drugs inambulatory care, and most non-prescription drugs •• can then be represented byattaching a parameter, CJ, to person l's use of health care of type j, equal to the out-of­pocket user charge per unit for goods or services of type j. For services which are fullycovered by government or private insurance, this CJ will be zero; for those which arewholly unsubsidlzed or uuinsured it will be equal to PJ' Partial "coinsurance" of, say,20% would be represented by a value of CJ equal to 0.2P/6 Much more could be, has been, and will be said on thistopic -- but not here, and not by us.7. But what about the deficit? Suppose governments borrowto meet part of their expenses? Well, obviously health care isan intra-marginal public expenditure .... Or perhaps t is thediscounted future value -- with perfect foresight -- of the taxrate required to repay future borrowing, with interest .... Theseconsiderations can be consistently represented by complicationsto the algebra which do not add enlightenment.8. Things can of course get quite complicated withdeductibles. Moreover, the value of Cj can be different acrosspersons, if some (those below a particular income level, for8Finally there may in fact be some component of true insurance, usually private, .which charges a fixed premium based upon some estimate of the risk level of the groupto which person I belongs. This will relate neither to her Income, nor to her actual useof services •• at least not in the current period. Insurance premiums are based on theexpected use and cost of services, and this in turn may be affected by past use, orincome level, or any other correlate of cost. But for each period taken In Isolation,premiums are a constant. They vary from person to person, of course, both becausedifferent persons have different levels of coverage, and because the same coverage willbe priced differently for different persons. We label this premium ~.Person I's total contribution to the financing of health care will thus be the sumof these three components •• taxes, user charges, and true Insurance premiums. H weadd across all persons In the society, we get total revenues raised to pay for health care:•••(2)This sum Is Identically equal to the total cost of health care, expression (1) above.example, or above or below a certain age) face different charges.This would be reflected in a larger set of parameters, Cij' butagain, complexity without enlightenment.9People and Earnings: The Algebra of IncomesSo where does the money go? To pay for health care, of course, but that is aprocess, not a destination. Revenues are drawn from someone's pocket, or bankaccount, and must end up in someone else's. Whose? Since there are N people in thissociety the payments must go, in aggregate, to those same people - but not in the sameproportions.People are paid for providing health care, in amounts which bear some relationto the time, effort and skill which they put into providing it, or the other resources ••materiais and supplies, services of physical capital such as buildings and equipment, or"intellectuai property" which they provide. Those inputs of time, effort and skill, andother resources, are not in generai the goods and services themselves, though we oftenspeak loosely (and collect data) as if they were. Any particular heaith care service willusually involve, in its production, the services of severai different types of people, as wellas various forms of supplies, and capital services. In the economist's jargon, these arethe "factors of production", and the price/cost of heaith care services is the source of thepayments to the owners of those factors.As with the different types of commodities included as health care services, wecan identify a large number of different factors which are required for and used in theirproduction. Some are quite specific to health care •• the services of doctors, dentists,and nurses, for example. Others are quite generai •• the construction company whichbuilds a hospital building or the pension fund which invests in the shares of a drugcompany or a commercial laboratory. Still others are in between •• the retailpharmacist who spends part of her time dispensing prescription drugs, and part runninga general retail store which is in no sense part of the health care sector (and may evensell tobacco). Such a pharmacist is providing services specialized to health care, buteven if working full-time, is ouly employed part-time in health care.If we let L represent the total number of different types of inputs or factors ofproduction which are used in producing health care services, then we can designate thetotal quantity of a randomly selected type of input used by Zk' where k can take on anyinteger value between 1 and L. Since these inputs (factors of production, resources) areeach owned by someone, who is paid for its use, we can designate by ~k the quantity ofinput of type k supplied by person I, There will be N*L such values, but most of themwill be zero, because most of the people in this society do not work in the health caresector or supply services to it, and of those who do, most provide a very specific rangeof services.The ~k refer to some physical measure of input •• person-hours, or furnishedsquare feet per year, or kwh, for example •• and will be reimbursed at some rate ofpayment per the relevant unit -- which may be explicit~ salaried persons) or implicit~ fee-for-service practitioners). This rate of payment can be labelled as Wk, the going10average rate of payment per unit for input services of type k. As in the case of the PJabove this rate will not always be identified specifically by the prevailing institutionalarrangements, but it will be there,"Accordingly we can set up the sum:which is the total income received by person i from the supply of all types of inputs usedin the production of health care services. Person i might be a hospital-based nurse,earning all her income from providing a single form of input •• for her, only one valueof 1.,. will be non-zero. Or she might be a physician, owning a share in the building inwhich she practices. Her income would include a share of the return to the nse of thisform of capital. But she might also be a pensioner, whose pension fund owns shares in acommercial laboratory.Income earned from providing inputs to health care production is thns notexactly the same as the income of those generally regarded as health care workers.Some share of their income is in fact from other sources -- think of the retail pharmacistabove, or the physician with investments in non-medical real estate •• and someproportion of the incomes earned (directly or indirectly) from providing health care goto people not commonly identified as directly connected with this sector.When we add up the incomes earned from providing health care, across all the Npersons in this society, thns:...(3)we arrive by a third route at an exact measure of the total cost of health care.We would in fact find that the values of i which corresponded to personsgenerally recognized as working in health care, do account for most of this total ofincomes earned from its production. But the issue of who draws income from this sectoris not a trivial one. For example, most of the problem of excessive "overhead" costs inthe U.S. health care system •• and they are monumentally excessive, estimated to be asmuch as $100 billion greater than in a more conventional system [5) •• corresponds to9 As a generalization, the more clearly the Wk areidentified, the more obscure are the Pi' and conversely. Thus inCanada, the prices of physicians' services -- fees -- aregenerally well defined, but the payment per hour of physicians'time and skill is implicit in their net incomes. On the otherhand, while the prices/costs of the different services providedby Canadian hospitals are not explicitly identified, the amountsthey pay for each of their inputs are.11the growth in the numbers of non-ellnlcal personnel •• administrators, lawyers,acconntants, insurance salesmen, health care researchers, economists •• who havemanaged to insert their services into this sector, as "necessary" for the production ofhealth care, and thns to draw part or all of their incomes from this source. Thissignificantly increases the range of values of i and k for which the Zak are non-zero,These three alternative ways of aggregating up to the society-wide total of healthcare expenditures are different ways of expressing the same thing, not merely differentways of estimating it. They mnst therefore be equal in principle, even if theinadequacies of real-life data systems may lead to different values from differentapproaches. And since they are identical, any change in any of the variables in therelationship must be matched or offset by corresponding changes in some of the others.In particular the conversion of some of the Zak from zero to positive •• new inputsbeing used and paid for •• must correspond to either a fall in the incomes of some ofthose currently drawing their incomes from health care, or an increase in the total costof care (or both). The first might be observed if, for example, more or more highly paidadministrators, or better health care research, led either to "greater efficiency" •• asaving in inputs used which would be reflected in a fall in some of the other values ofz.., and of the incomes of those who supplied those inputs •• or more effective bargainingto lower the values of the Wk' In concrete terms, some other people lose their jobs orhave their wages cut, and/or some suppliers lose part of their markets, or receive lowerprices for their products (with implications for their workers or shareholders).If, however, new income claims are added without reducing the establishedclaims, then total costs must rise. This has been the U.S. experience; increases inoverhead costs have been translated into increased unit costs of clinical services •• higherPJ' Efficiency gains, regularly predicted on the basis of local experience, have yet toappear in aggregate. Alternatively some concede that costs~ increased, but claim thatthere are significant benefits, very real though difficult to measure (and in any case notcurrently measured) which correspond to the increase in costs [6,7]. The nature of theservices has changed so that the QJ are different and if properly accounted, wouldrepresent "more" output, not just higher prices. This is also a common argument byproviders of clinical services. Whatever the pathway, any increase on the income sidemust correspond to an increase in total expenditures •• a rise in P or Q or both •• andan increase in revenues to support them.On the outlay side, if the YI and ILJ are not changed then some or all of the t, eJ,and ~ must increase. Rising costs place upward pressure on tax rates and insurancepremiums, and feed calls for increased user charges (supported, for example, by therhetoric of the need "to encourage consumer responsibility"). Recall that the parametert in this case is the share of total tax revenue going to support health care. It canincrease, while holding total tax rates constant, either by reducing other components ofpublic expenditure or by deferring an increase in overall rates (or a reduction in other12expenditures) through borrowing.Putting the pieces together:L,{tY, + LiCj*qlj) + R,} =LJPj*L,q,j] =L'k{Wk*Z,k} ..•(4)Total revenues raised equal total health care costs equal total incomes earned from theprovision of health care. Thus a change in anyone of these variables must beaccompanied by a change in one or more of the others, in such a way as to maintain thisrelationship.13Introducing User Fees: What Else Happens?As hospital and medical care are currently funded in Canada, CJ and R, areeffectively zero. One can pay extra for preferred accommodation in hospitals, and canbuy Extended Health Benefit insurance to cover this, but the amounts involved aretrivial. We can thus treat such services as entirely tax-financed without any significantdistortion. If user charges were then to be introduced for such services, CJ wouldincrease from zero to some positive amount.It follows then that at least one and perhaps several other variables in equation(4) would have to change. There is room for debate about which variables would in factchange, and also over whether such changes would be desirable in some general sense.But there can be no debate about whether some other variables would change -- as theecologists point out, it is impossible to do only one thing.Much of the debate focuses on the relation between user charges and theutilization of health care. Treatments of this topic in the economic literature, workingwith the elementary "supply-and-demand" framework of the textbooks, are virtuallyunanimous in assuming that any increase in user charges will result in a decrease inoverall utilization, a fall in Q.IO They are not always specific as to the expected effectson P, but most applications of the "supply-and-demand" framework "predict" a fall (orat least no change) in P. Hence the standard story -- user charges will bring downhealth care costs,"Other commentators argue for user charges as ways of moving more money intohealth care, of raising expenditures in a system which they claim to be "underfunded"[8]. The implicit "theoretical model" underlying this argument is not spelled out in10 Many versions of this theoretical framework have beenoffered which appear to be more sophisticated, but so long asthese preserve the central assumption of exogenous "consumerdemand", they are simply variations on the same elementary theme.II This framework is also used as the basis for anargument that health care costs are escalating because care is"free" to the user, but in fact this extension is illegitimate.The "supply-and-demand" framework is a description, (right orwrong) of a static equilibrium process, and says nothing at allabout why the values of variables should be changing over time.There are several examples in the economic literature of writersmaking up for this silence by assuming that "demand" is simplyincreasing over time -- "consumers' tastes are changing" -- whichwould seem to be a classic example of the logical fallacy ofpetitio principii or begging the question. 'Changing tastes" nototherwise qualified, is simply a label for the ignorance of theanalyst.14detail, and there are several different lines of argument (all with somewhat moresensitivity to the institutional realities of health care delivery than the naive "supply­and-demand" framework) which support the "prediction" that user charges will raisethe overall costs of care.Each of these different lines of argument can be represented by different patternsof movement of the variables in equation (4). We will refer to some of them below. Butin the end, since different theories are available to "predict" that Q will move up, ordown (or remain unchanged) in response to a change in CJ, the question is ultimately anempirical one. We review some of this empirical evidence elsewhere [9].What is not ambiguous, however, is that any changes in the CJ will haveredistributive effects. User charges will take money from some, and give it to others.They may also have significant effects upon the distribution of health care, more forsome and less for others, whatever their effects upon the overall level of care utilized.Within the framework of equation (4), which must hold for all health care systems, wecan represent the distributional effects of alternative ways of financing health care, andshow how the burden of payment would be redistributed by various proposed forms ofuser charges.15User Fees or Taxes: Redistribution among PayersWhen there are no user charges or insurance premiums, people contribute tofinancing health care (roughly) in proportion to their incomes. H user charges do notaffect the overall volume of care used," then (holding all prices PJ constant) anyintroduction of such charges must result in a fall in t, Increased user charges lead tolower tax payments,"Whether person i gains or loses from this change will depend upon therelationship between her income and her use of the various forms of health care, .. Y,and qlJ" It should be obvious that in general those who are healthy and wealthy willgain, and those who are poor and sick wlIl lose, while the effects upon those who arewealthy and sick or healthy and poor are more ambiguous. But from equation (4) wecan derive a somewhat more precise description."12. As discussed elsewhere [9], this is, in fact, the mostplausible assumption for hospital care, and probably forphysicians' services as well. This is not the same as assumingthat people's decisions to seek care are insensitive to price, orin the economists' jargon, a zero elasticity of demand. On thecontrary, the empirical evidence is quite clear that prices doaffect these decisions, and negatively, just as simple "supply­and-demand" models of behaviour would predict. But the evidenceis equally clear that individuals' decisions to seek care do notin themselves determine the overall level of care provided, whichdepends rather on the capacity and objectives of providers, andthe advice they give patients. So user charges, at least withinthe ranges observed in modern health care systems, may influencewhich people receive care, but do not appear to influence overalllevels of care provided. No doubt some level of charges would behigh enough to limit overall use, but if these are politicallyintolerable, and in any case are beyond any yet seen, then thistheoretical possibility is of little practical. relevance.13. That is, lower tax payments for health care, and lowerthan they would otherwise have been. Recall that in equation (4)t is that portion of the overall tax rate which goes to financehealth care. It can be raised or lowered, without a change inoverall tax rates, if tax revenues are diverted from or to otherpurposes. In particular a reduction in any public borrowingwhich would otherwise have occurred, amounts to a reduction infuture, rather than present, tax rates.14. We are here and throughout referring only to ex postgains and losses in financial position or in some cases in accessto services. There will also be ex ante losses of utilityresulting from the increased exposure to risk associated with anyshift from tax to user charge financing. Insofar as ex ante risk16Holding total expenditures, and therefore total revenues, constant (and assumingno expansion of private insurance), and letting the changes in C and t be designated byL'1C and L'1t, we can see that:...(5)That is, individual i wJ1l pay more in user charges and less in taxes, in amountswhich are determined by the increases in the various CJ and the decline in t, applied toher income and pattern of use. H we sum over all individuals, the total revenues fromuser charges will exactly equal the reduction in payments to health care from taxrevenue. From this equation we can derive the reduction in tax rates that is madepossible by the imposition of user charges. A little bit of algebraic fiddling shows this tobe:...(6)Not very surprisingly it is equal to the total revenue from user charges divided by thetotal across all individuals of taxable income. But it has a negative sign, because ifcharges go up, tax rates go down. Using this measure of the change in tax rates, we canthen represent the change in person i's financial position resulting from the introduction.of user charges. Her net gain is:1sThe first term is the amount by which her taxes used for health care will go down, andthe second term is the additional amount she must now pay in user charges. Whilethese amounts cancel out over the population as a whole, it is quite obvious that formost people, expression (7) will be either positive or negative, and can be quite large.Furthermore a simple rearrangement, extracting individual i's income, YI' from bothsides of (7), yields:in which it may be seen that the first term is the ratio of total user charge payments,status is correlated with actual use, and uncorrelated ornegatively correlated with income, this loss of utility will alsobear more heavily on those at lower income levels.15. In representing person i's net gain, we change the signof the L'1t from (6). A fall in tax rates, a negative value of L'1t,translates into a positive change in person i's position -- shepays less in taxes -- which must be set against any increase inthe user charges she must pay.17society-wide, to total taxable income, and the second is the same ratio for person i. Thusthe algebraic fiddling leads to a result which may in fact be intuitively obvious -- personi gains from the introduction of user charges if the proportion of her income which shespends on such charges is less than the proportion for the community as a whole.Otherwise she loses.Yet another way of putting the point, by rearranging terms in (7):[Y/E.y. -L/ACj*lLj)/{L'j(AqllJ.J)}]*{Ly(ACj*lLj)} •••(9)The first term is her share of total taxable income, the second is her share of theincrease in user charges for health care. Thus the gainers from user charges are thosewhose share of taxable income exceeds their share of out-of-pocket expenditures forcare. Obviously this includes the "healthy and wealthy", with above average shares ofincome and below average use of health care (weighted by the imposed charges), whilelow income, high use persons lose financially. But if one is wealthy enough, one can bean above-average user of health care, and still gain financially from a user chargepolicy; conversely if one is poor enough, one can be a below-average user and still lose(unless of course one uses no care at all, in which case one is always made better off byincreasing user charges and lowering taxes.)Expression (9) also indicates that the size of ones' gain or loss is proportionate tothe total amount collected, society-wide, in user charges. H one gains (loses) from suchcharges, increasing them simply increases the gain 00SS).16As is well known, health and wealth are correlated. To the extent that utilizationof health care also correlates with ill health, Y. and lLj will be negatively correlated.Going back to expression (7), this implies that over the population as a whole, the firstand second terms will be negatively correlated. Persons with high incomes -- largevalues of Y. -- will tend to have lower levels of health care use, and hence lower outlayson user charges. And the reverse will be true for people at lower incomes. Thus thecloser the correlation between health and wealth, the greater will be the financial gainsof the gainers, and the losses of the losers. The redistributive effects of user charges arethus accentuated."16 It follows immediately that those who advocate theintroduction of user fees in the expectation of personal gain[10] could be expected to support increases in their magnitudeonce they had been introduced.17 Illness is correlated with age, and there are, amongthe elderly, people with a good deal of wealth even though theircurrent incomes may be relatively low. But these "whoopies" - ­well-heeled older persons -- do not represent a large share of18the age groups -- the "oldest old" -- that make heaviest use ofhealth care.19Relaxing the Assumptions, Strengthening the ConclusionsThese results follow logically from the assumptions of the model, the critical onesbeing that health care is funded either from taxes or from user charges, and that overallcosts of health care do not change. But they also depend on the construction of equation(4), in which taxes have been specified as proportionate to income, and user charges asfixed amounts of $CJ per unit of health care of type j. Since the relevance of the resultsfrom a model depends upon the validity of that model's underlying assumptions, onemight reasonably ask what happens if these assumptions and specifications werechanged.If we relax the assumption that overall utilization does not change, then theimpact of user charges on the distribution of the financial burden will be different. Butif one were to make the assumption which economists commonly make (on a priorirather than empirical grounds), that overall utilization and costs will go down, then theright hand side of equation (5) will be, not zero, but some negative number equal to thehypothesized fall in total health care expenditures. Carrying this into equation (6), then,tax rates will fall even further. If~e replaced the zero with -K, then in equation (6)tax rates fall by an additional KI{LIY,}, and in equation (7) this generates an additionalpositive term proportionate to person i's income. The financial savings from reducedutilization are translated into larger gains for higher income people.Thus assuming a negative response of utilization to user charges reinforces the(financially) redistributive impact of user charges, from lower to higher income people.Only if the introduction of such charges were to increase the total costs of health care,and if these increases resulted in tax increases, rather than in further increases in usercharges, could one change this conclusion. And even then, if the overall tax systemremained roughly proportionate, the result of an increase in both taxes and user chargeswonld be that while people at all incomes would pay more, the greatest relative increaseswould be borne by those at lower incomes, because being less healthy, they are morelikely to use care.I. .Allowing for private insurance has equally straiglUf!>rward results. If overall useand cost do not change, then one simply adds a positive 2.,;1(L\~) term to the left handside of equation (5), while the right hand side remains at zero. The fall in tax rates will18 If they can afford it. These comments assume similarchanges in use by all individuals. But user charges which werehigh enough to force lower income people to reduce theirutilization significantly, relative to those at higher incomes,or to push them out of the system entirely, could reverse theredistributional conclusions here. After all, the burden ofpaying for Mercedes-Benzes falls disproportionately on wealthypeople. Changes in the pattern of the qij are considered furtherbelow.20be matched by some combination of increases in user charges and private premiums,with the mix depending upon the coverage and comprehensiveness of the private plans.To the extent that private insurance substitutes for user charges, of course, theprobability of any (negative) utilization response is reduced. And since experienceindicates very strongly that private insurers are incapable of imposing effective costcontrols on health care systems, the consequences of a shift to such coverage are likely tobe an increase in both the quantity of care used, and the prices paid for it.But private insurers price coverage according to the risk status, the expected cost,of the person or more commonly the group covered. Thus any increase in privateinsurance premiums and corresponding reduction in tax rates will transfer financialburdens away from those at high income, and toward those at high risk of care. SincelIIness, and care use, tend to persist over time, past use is one of the better predictors offuture use. Thus private insurers tend to link premiums charged to past use ••experience rating •• and either charge the highest premiums to the least healthy, orsimply refuse to insure them. Allowing for the reintroduction of private insurance alongwith user charges thus reinforces the redistributive conclusions above.21Integrating User Charges with Taxes: More Algebra, Same ResultThere are many forms of user charges, and many forms of taxes, and the analysisabove concentrates on the simplest form of each. One could impose charges whichexempted those at lowest incomes, for example, in which case the redistribution fromless to more healthy and wealthy would operate only over part of the population. Anoften suggested alternative would not impose user charges as such, but would add (all orsome portion of) the outlays made on behalf of person i to her taxable income. The lefthand side of equation (4) would then become:L,{t(Y, + rLJ[P/'lJJJ])}The proportion of the cost of care of type j which is added to taxable income isrepresented by r, whose value would lie between zero and unity. Health care would stillbe fully tax financed, but ones' tax payments would depend upon use of care.With this expanded tax base, of course, t must fall to hold revenues constant.Once again, some gain and some lose. H the health expenditures made on your behalfrepresent a proportion of your taxable income which is less than the initial (i.e. pre­scheme) tax rate to finance health care, your taxes fall, regardless of the value chosenfor r, The size of your gain will, however, be greater, the larger is r; moreover yourgain will also be larger, the larger is your taxable income, and the smaller is your outlayon health care.In the simplest terms, let Y be total taxable income in this society, and X be totaloutlays on health care. Prior to including them in the tax base, these outlays were fullyfunded at a tax rate of t, so tY = X. H some proportion r of these outlays is included inthe tax base, then t'(Y + rX) = X, where t'< t. Health care outlays can now be fundedby a lower tax rate on this larger base. Substituting tY for X in the second expression,we get t'(Y + rtY) =tV, or t' =tY/(Y + rtY) =t/(l + rt). The new tax rate to supporthealth care is lower than the previous rate, by an amount which increases with r.Now cousider person I, whose gain or loss from including a share of health careoutlays in taxable income will be the change in her taxes. These will be:t'(y, + rX,) - tY, =[t/(1 + rt)][Y, + r~] - tY"or regrouping:{trY, + rX,] - (l + rt)[tY,]}/(1 + rt).Cancelling the terms in tY, and taking out the common factor rt, her gain becomes:[rt/(1 + rt)][~ - tY,]22Thus her taxes go up If outlays on health care on her behalf exceed her tax paymentsprior to the Inclusion of such outlays as a taxable benefit, and go down otherwise. Theamount of her gain or loss will be greater depending on the value of r, the proportion ofoutlays so Included. But for any given values of r and t, the net gain will be larger, thelarger Is YI and the smaller Is ~'As a numericallJlustration, consider the hypothetical data In Table 2. Persons Aand B have Incomes of $20,000 and $80,000 respectively, and each of them has Illnesswhich results In health care costs of $10,000 during the year. Thus taxes of all typesmust raise $20,000 In total to cover the costs of care. Assuming proportionate tax rates(all types of taxes taken together), person A pays $4,000, person B $16,000. The overalltax rate (20%) Is lower than the proportion of person A's Income represented by healthcare, and higher than person B's. If we add health care to each person's tax base as ataxable benefit, then the overall tax rate needed to cover the $20,000 cost of care Is16.67%. Person A's tax burden rises; person B's falls."Thus, even though such a proposal appears to respond to distributional Issues, Itstill transfers funds to the healthy and wealthy, at the expense of the poor and III. Andagain, the fact that In the Canadian population (and most, probably all, others) healthand wealth are positively correlated accentuates the degree of redlstrlbutional Impact.What Is rather Ironic, however, Is that proposals for Integrating user charges Intothe tax system originated In the United States, and continue to be discussed there, asways of redistributing Income In the other direction -- from higher to lower incomeusers of care [11,12]. And indeed they would have this effect -- In a system whichpresently has very substantial user charges. But If one starts from a tax-financedsystem, as In Canada or most of the rest of the developed world, such proposals havedistributional effects which are the exact opposite of those Intended by their Americandesigners. It is an unfortunate aspect of the Canadian mentality that Ideas and policiesare so often Imported uncritically from the United States, even when the Canadianenvironment Is obviously and radically different.19 This result holds even when different individuals areassumed to have different tax rates t, so long as all tax ratesare adjusted proportionately when the base is increased.TABLE 2: Health Care as a Taxable Benefit: The Rich Get RicherPerson A Person B Total[A] Taxable Income $20,000 $ 80,000 $100,000[B] Health Care Costs 10,000 10,000 20,000[C] BfA 50% 12.5% 20%[D] Total Taxes Payable 4,000 16,000[E] Income + 'Benefits' 30,000 90,000[F] New Tax Rate 16.67%[G] Total Taxes Payable 5,000 15,00023What About Tax Credits? A User Charge Is a User Charge Is...One can also show the impact of another proposed "alternative" to user charges,in which it is suggested that each person be assigned some dollar value of health care"credits" for each year, and required to pay some proportion of all expenditures beyondthat level. Those who did not exhaust their credit level, would get a rebate of a portionof the unused credit.Letting X, represent person i's credit level (and thus allowing for the scheme torecognize that different people have very different levels of risk), let s represent theproportion that people must pay of the amount by which their actual expenses exceedthe credited amount, and also the proportion of the "unused" credit which is rebated.Then the left hand side of equation (4) becomes:L,{tY, + s(LJ[PJ*q,J] - ~) - s(X,- LJPJ*q,J])}The first term multiplied by s is the amount by which health care outlays for person iexceed her credit, and is set at zero if they do not. The second term multiplied by s isthe amount of her "unused" credit, and again is set at zero if there is none. But thenthe two terms taken together reduce to:L,{tY, + s(LJPJ*q,J] - X,)}where the second term can be either positive or negative, or setting sPJ = CJ:L,{tY, + LJ[CJ*q,J] - sX,}So one is back to a plaln, common or garden-type user charge with only the addition ofa flat-rate credit of sX,. This is unrelated to either income or actual use of care, butmight be based on risk or expected use.If the total of the X, is set equal to overall health expenditures, then user chargerevenues and rebates will cancel out over the society as a whole, and tax rates willremain unchanged. The scheme will simply redistribute income among people accordingto whether their actual use of care exceeds or falls short of the "expected" amountsembodied In X; Those who have "unexpected" accidents (the usual kind), or thosewhose risk status is incompletely reflected in their assigned credit levels -- who havesome form of chronic illness, for example, which is not recognized in whatever formulais used to set these individual values -- will transfer funds to those who are healthierthan "expected". The more (less) sensitive and sophisticated the process of setting the24individual credit levels, the smaller (larger) wlll be these gains and losses."Like any other nser charge scheme, this proposal transfers funds from the moreto the less 1lI. Whether it also transfers from lower to higher income people dependsupon (a) whether the rebates do in fact match the charge revenues, and (b) whether theprocess of setting the individual credit levels provides higher levels to those at lowerincomes, sufficient to reflect their higher risk status. If the total of charges exceedsrebates, so that tax rates fall, or if individual credit levels incompletely reflect risk, thenthere wilI also be a net transfer from lower to higher income people,"20 Advocates of such a scheme may believe that totalutilization and costs of care will fall, which is in fact ratherunlikely [9]. Alternatively they may predict, quite plausibly,that even after the credits the scheme will income redistributefrom high to low users, and they expect to be among the latter.21 There would also, of course, be administrativeproblems, and costs, involved in determining individual risklevels. If one relied only on such measures as age and sex, theprocess would be easy and cheap, and also very incomplete. Theresult would be large transfers from the chronically ill to thechronically healthy. But if one tried to establish trulyindividualized risk and credit levels, one would in fact be doingexactly what insurance companies do -- underwriting -- only inreverse. This process is difficult and costly, and providesopportunities and incentives for "false signalling" by bothclients and underwriters. In a public system it would also drawconsiderable litigation and raise some very awkward issues ofdiscrimination. One might predict fairly confidently that the Xivalues would not be individualized, and thus that theinterpersonal transfers would be considerable.25Recommendations Imply Values •• But Whose?The principal impact of introducing user charges into a tax-financed system, orfor that matter of removing them, is the redistribution of financial resources from somepersons to others. One may approve or disapprove of these transfers; that is where thevalues come in. The expert, economist or otherwise, has no legitimate claim to have hisor her values in such matters accorded any higher status than those of any othercitizen/voter. But he can, and we do, insist that the fact of this redistribution, and itsdirection, be recognized.Anyone who advocates a change in the financing structure of health care, forwhatever reason and in whatever direction, is simultaneously advocating an extensivebut quite predictable process of income transfer from some members of society toothers. It seems only honest that such advocates be prepared to admit, and defend, thenecessary cousequences of their proposed policies. "A man must be presumed to willthe consequences of his own acts" •• this, we believe, extends to "woman", and torecommendations as well as acts.In particular, those economists who claim that economics as a "science" is value­neutral, so does not make comparisons of utility or well-being among persons, yet whostate baldly that user charges are good policy because they "improve allocatlveefficiency" are in fact advocating such transfers, and therefore implicitly makingprecisely the interpersonal comparisons which they would have others believe they donot and cannot make. There is nothing wrong with having values •• how else could onerecommend anything? But there is something wrong with trying to deceive others (andperhaps oneself), as to what those values are, and trying to cover with the white cloak of"scientific" objectivity, value judgements which one knows or strongly suspects wouldnot be shared by ones' fellow citizens if left naked for their tnspecnon."22 Interestingly, there is empirical evidence indicatingthat the personal values of economists do differ, on average,from those of the general public [13], making any value-basedpolicy recommendations, or claims of "scientific objectivity",doubly suspect.26Redistribution of Access to Health Care ServicesWho pays for health care, and how much, is an important part of health carepoliey, but obviously not the whole of it. The principal concern of public debate has, infact, been who gets care, and of what kinds. [One might feel that, since most peoplewant health care, not for its own sake, but because they believe that it will benefit theirhealth, the focus of discussion and of policy should rather be the ultimate objective ofhealth rather than a concentration on one particular pathway to it, through health care,but this idea, though very old, is at best a very new part of health policy~ 14, 15].A standard argument against user charges is that people will be deterred fromseeking care, and that their health will suffer as a result. If we let H, represent somemeasure of the health status of person i, then in terms of the variables in equation (4)this argument can be represented as:and:ILJ = Fi9dB, =G(ILJ' H,)...(10)...(11)where F and G stand for some functional relationships.The use of care by person i will be affected by the price she must pay for it, andher health may be improved (depending upon what it was like to start with) by her useof some forms of care. (There are in fact M different relations (10), one for each type ofcare, and the influence of price on use may be quite different depending upon which isreferred to -- strictly speaking there is no such thing as "health care'") Thus usercharges, if they deter people from seeking needed care, may be harmful to health, and totheir opponents that in itself is a powerful counter-argument. After all, the majorpurpose of introducing public programs to finance health care in the first place wasprecisely to enable people to get the care they "needed".Advocates of user charges, on the other hand, tell a more complex mix of stories.Some, principally those influenced by the mechanical "supply-and-demand" apparatusof elementary economics, accept and indeed celebrate the relationships FJ• (These theyidentify as "demand curves", which are in reality only a particular special case of amore general class of possible relationships.) But they then ignore completely therelationships (11), simply failing even to meet the arguments of those who are concernedabout health consequences. In so doing they are in effect advocating that health care betreated as no different from any other commodity, and implicitly assuming that anyinfluence of health care on health is irrelevant.As for the impact of user charges on the distribution of health services among thepopulation served, opponents will often assert that charges are more likely to deterpeople with low incomes, reducing their share of the care provided and accentuating27their health disadvantages. On the other hand those who believe either that health carehas nothing to do with health, or that if it does, that should not affect how it is paid for,find such redistributional effects uninteresting. No one worries about access toMercedes-Benzes, after all. But both lines of argument suggest that by reducing theutilization of particular individuals, user charges will lower overall utilization, andthereby costs, of health care. Each component of equation (4) •• revenues, expenditures,and incomes •• will fall.Many of the arguments for user charges, however, and particularly those putforward by providers, contemplate an increase in total expenditures [8]. Some aresimply trying to add on a user charge to the amounts they are reimbursed by the publicplans so as to raise prices of services and their own incomes •• a simultaneous increasein Cl , Pl , and certain of the W1k, with no necessary impact on quantities used. Butothers tell a more nuanced story, in which the principal impact of user charges is onwhich services are provided, to whom, rather than on aggregate use or expenditure,"The argument is well illustrated by remarks attributedto the premier of Alberta, Ralph Klein, and his Minister of Municipal Affairs, SteveWest [16]. The former was said to be proposing changes to the Canada Health Act topermit provinces to impose user charges, without financial penalty, and was quoted:"...steps have to be taken to cut down on abuse and perhaps a small user fee for thosewho can afford it might be a way to do that." On the other hand Mr. West, insupporting the call for user fees, was more concerned about access to particularexpensive services for those willing and able to pay for treatment: "I don't want toretire with half a million dollars in the bank, be 92nd. on the list for heart surgery, anddie with all that money in there."We suspect that these two comments, taken together, represent a much morecommon view among advocates of user fees than the rather peculiar (to us) position ofthe neo-classical economist who believes that health care has nothing to do with health,or at least that we should all pretend that it does not. And we single them out, not inany way to target Mr. Klein and Mr. West, but because we believe that they have doneus the service of representing this view succinctly and authoritatively. Their rationalesdo not, however, necessarily add up to a less expensive health care system, only adifferent one.23 The steady accumulation of evidence may also be takingits toll. For those who asserted confidently twenty years ago[8] that user charges were the way, and the only way, to limitcost escalation, the subsequent direct refutation frominternational experience [9] is at least inconvenient. Truebelievers continue to repeat the litany, but others are modifyingthe story somewhat to fit the facts.28Those who pursue this line of argument quite clearly believe that health care doesmatter for health, at least for the health of some individuals in some situations. Butthey also believe that at present there is a lot of care, and associated expense, that is noteffective, and that there is.also a lot of care going unprovided, because of capacityconstrainis in the Canadian system, which would be effective and should be provided.They propose a rearrangement of service use; fewer services for "abusers" (whopresumably do not have real needs, or at least not for the care they are seeking) andmore for those who really need care (and who might die without it). In terms of theformalism we have been using, they believe that an increase in some of the CIt.J matchedby a decrease in others •• different people, different services •• could lead to a moreeffective system overall, and that the introduction of user charges would help to bringabout this rearrangement.This leaves open the question of whether overall use and cost would go up ordown. Mr. Klein is quoted as saying that the user fee would help offset the growing costof Medicare, and would only be charged to "those who can afford it", which seems toimply that the primary intention would be to raise revenue rather than to reduce cost.And Mr. West seems to be anticipating an increase in the rate of cardiac surgery,supported by an inflow of private funds. This suggests that the ultimate objective is ahealth care system with higher overall costs •• which is certainly cousistent with thearguments for user charges made by providers •• but more costs borne by users and lessout of taxes. A larger number of "needed" complex diagnostic and surgical procedureswould be provided for those willing to pay for them, and perhaps less care, hopefullyless "unnecessary" care, would be provided to those who respond to charges.This seems to correspond, point by point, to a move in the direction of theAmerican system. If that is Mr. Klein's objective, then his recommendation seems tomake perfect sense •• up to a point. The empirical evidence [9] indicates that usercharges do influence peoples' decisions to seek care, in the expected direction, but thatthey do not appear to influence overall levels of use through this route." Rather they24 It is remarkable how many economists routinely committhe "fallacy of composition" against which we warn students inthe first year courses. Relations (10) assert that an individualfaced with an increase in the price of a commodity will,everything else being egual, reduce her use of it. And theempirical evidence confirms this. But they do not say that ifeveryone faces an increase in prices, for all the differentcommodities, that total use will fall, because in that caseeverything else does not remain equal for each individual. Theaggregate relationship is not necessarily the sum of theindividual ones, or, in one of the examples commonly used infirst year courses, "If you are in a crowd watching a parade, andyou stand on tiptoe, you will see the parade better. Thereforeeveryone should stand on tiptoe .... " The empirical evidence29redirect care from those who are more sensitive to charges, to those who are less.Their effect on overall use and cost depends upon the base line for comparison.If user charges are a substitute for capacity constraints, price controls, and globalexpenditure caps (as many in the U.S. have recommended) then they lead to increasesin total expenditures, relative to alternative, more effective policies. (As the U.S. hasexperienced •• but not learned. They continue to rely heavlly on user fees.) On theother hand, if user charges are levied along with various forms of direct budgetarycontrol, as in some of the European systems, they have no clear effect one way oranother on total costs," They do, however, enable the well-to-do to buy their way tothe front of any queues that may develop, which seems to meet Mr. West's concern.Where Mr. Klein's argument leaves the evidence behind and strikes out on itsown, however, is in the assumption that the redistribution of services which it implieswould represent a more effective health care system. One may be concerned about boththe over-provision of ineffective care and the under-provision of effective care, withoutany presumption that this situation would be improved by user fees. The argument thatsuch charges reduce "abuse" founders on the observation that no one, doctor or patient,seems to be able to define "abuse" [17], although like pornography, they can all think ofone or two egregious examples. It should not then be surprising that the availableevidence in both Canada and the United States does not support the argument that"free" care leads to more "unnecessary" use, or that user fees lead to less [9].At the other end of the scale, it should not be surprising that people who havebeen told that they 'need' heart surgery are willing to pay whatever it costs -- if theyhave the money. It does not follow that they will necessarily benefit as a result. TheU.S. experience with unlimited access -- for those with the money -- is well documented.A lot of people receive services which, in the judgement of external experts, areinappropriate, unnecessary, in some cases harmtul," But is this "abuse"? Presumablysuggests that overall rates of health care use are primarilydetermined by the capacity and the objectives of the suppliers ofcare, not (at least within observed ranges) the levels of chargeslevied.25. The only European nation that seems rhetoricallycommitted to the importance of the 'ticket moderateur" as amechanism for cost control, France, is also the country in Europeleast successful in overall cost control -- but the direction ofcausality is ambiguous.26. Chassin et al. [18) review the relevant literature.Their conclusion is supported by a recent comparison of patternsof management of patients with myocardial infarction (heartattack) in Canada and in the United States. Rouleau et al. [19)found that while such patients in the United States are nearly30all those who underwent such Inappropriate procedures believed that they were going tobenefit, and some died In the comfort of this belief.On the other hand, documentation Is Increasingly emerging to confirm what onemight have expected, that those who do not have the money, or the Insurance, receivesignificantly fewer services of specific and Identifiable types, and that their health suffersas a result. As supporters of the U.S. health care system often point out, It Is not truethat people with no Insurance receive no care. It Is however the case that they receiveless care, enough less to make a difference. And so do some who have Insurance -- thosecovered by the Medicaid program for people In poverty.The relationships between health status and health care are sufficiently diverseand complex to keep busy a large proportion of the International health servicesresearch establishment. The point we want to emphasize here, however, Is that theImpact of user charges on who gets what kind of services Is much more significant thantheir effect on the overall level of use and expenditure. The more a system Is fundedfrom user charges, the more the mix of services provided Is determined by willingness topay. But "willingness" Is a product of desire and ability.There Is a considerable amount of evidence to support the (rather Intuitivelyobvious) point that user charges reallocate services from those with fewer resources tothose with more -- the greater willingness to pay of the latter being a naturalconsequence of their greater ability to pay. But there Is none to Indicate that thiscorresponds to a more effective mix of services, because needs are Inversely correlatedwith ability to pay, and because In any case people do not generally know their ownneeds. That Is after all why they seek professional advice. And while there are goodgrounds for believing that the quality of the resulting decision-making process could beconsiderably Improved, there Is neither a priori logic nor empirical evidence to supportthe proposition that this would be assisted by user charges.The rhetoric of "shortages", "cutbacks", and "rationing" which surrounds theprocess of adjustment of the Canadian health care system to a more slowly growingeconomy may lead Increasing numbers of people to fear that they personally will not beable to get care when they need It. User charges may then come to be seen by morepeople as a desirable policy precisely because they serve to reallocate care from someoneelse, to those of us who can afford to pay.ThIs seems to be Mr. West's point. If there must be a shortage, let It fallontwice as likely to undergo coronary artery bypass grafting(CABG), there is no difference in mortality rates. Overall CABGrates are about twice as high in the United States, but cardiacmortality is lower in Canada, as it is for most causes of death[20] •31someone else. Those willing and able to pay should be served. If many of us do notwish to admit to ourselves what our fears and motivatious really are, then the rhetoricof "personal respousibility", "control of abuse", and perhaps a bit of "mutual sacrificein difficult economic times" will provide a comfortable cover for what might otherwisebe a rather embarrassing attempt to get more for "us" at the expeuse of "them".In the past, the Canadian community decided, through its political processes, thatwe wished to redistribute the use of health care from the more to the less healthy andwealthy, and to redistribute costs in the opposite direction. We put in place a financingsystem to do this. It worked. If now we as a community decide that we want toincrease the proportion of health care used by those with relatively more resources, andmove the payment burden in the opposite direction, that is obviously also a legitimatepolitical decision. User charges are one way of achieving that result.There is, of course, nothing sacred or even scientific about the present level orpattern of health care services provided in Canada. Experts of various persuasions haveassembled a great deal of evidence to support the general proposition that a reallocationof services -- more of some kinds for some people, less of others for others -- couldsignificantly improve the efficiency and the effectiveness of the Canadian health caresystem without any increase in overall costs.Recent Royal Commissions and other provincial public enquiries have reachedthe same conclusions, recommending in particular a shift in emphasis from acute carehospitals to alternative community-based services. They have also stressed the need formore attention to the outcomes resulting from particular services; much of presentactivity is inappropriate or unevaluated. Thns trying to rearrange the lLJ to improvehealth outcomes seems widely accepted as a desirable social objective. A number ofpolicies are currently under discussion or in implementation to try to do this.But there is no basis for any notion that the rearrangement that would followfrom nser charges would move us in this direction; quite the contrary. Willingness topay does not correlate with capacltv to benefit.32Redistribution between Payers and ProvidersWorking our way through equation (4), we have considered the redistributionaleffects of user charges on who pays for health care, and on who receives which services.But any rearrangement of the types of health care provided, or in how they are funded,will also have implicatious for whose services -- inputs, or factors of production -- will beemployed in and paid for providing them. Who gets what jobs, and how much will theybe paid?In Reinhardt's [21] graphic analogy, each person who is paid from totalexpenditures on health care, represents a "place setting at the health care feast". In ourmore pedestrian terms the number of places set is the number of Z,k with positive values,and the amounts put on each plate are represented by the size of the W1k and the Z,k'Any change in health policy will influence the number of settings, the size of theportious, and who is invited to sit.For example [only], the simple "supply-and-demand" story of the economicstextbooks would have us believe that user charges lead to lower use of care overall,which implies fewer resources employed in providing it. The process of adjustment mayinvolve a "short-run" trausition stage [whose length is not specified, it could be as longas a working lifetime for some professionals] in which some people find themselvesunder- or unemployed, some suppliers find their markets shrinking, and competitionmay lead to a fall in fees, prices, and wages. Average portions, incomes, shrink. Buteventually some people leave the health sector entirely -- Z,k for them goes to zero -­and/or others choose not to enter. Jobs are lost. The levels of employment and incomefor the remainder go back to their "equilibrium" level. Some producers of materialsand supplies for the health care sector are "shaken out", and move into other lines ofbnsiness. Portion sizes may be restored, but fewer places are set.Or so the economics textbooks confidently assert. The reality is usually quitedifferent, and rather more complex. But the general point is valid; any change in thelevel and mix of output, or in prices, must have balancing effects on the income side.Those who, in defiance of the textbooks, argue for user charges as a "cure" for aperceived problem of underfunding are simultaneously arguing for increased incomesand/or jobs in the health care sector.The simplest example of a cost-expanding user charge is extra-billing byphysicians. When physicians' fees are set by negotiation between physicians andprovincial governments, and fully paid by those governments, the introduction of extra­billing would raise both PJ and CJ by the same amount. If, as its physician advocatesclaim, the process of extra-billing were moderated to ensure that no one was denied care-- only those who "can afford it" being billed -- then the <LJ would not change and thewhole of the increase in prices would go into increases in Wk -- payments for the33services of physlclans," This particular form of user charge simply transfers incomefrom patients to physicians.But this may not be the whole story. Practice overheads may have to increase tocover the extra administrative costs of bllllng." This implies more use of the servicesof medical office assistants •• an increase in some of the z... The increase in prices anduser charges wonid then be partly absorbed in additional resources used in medicalpractice, for the same level of service output.A similar point has been made by hospital administrators in B.C., who havesuggested that the small "nuisance" user charges prior to 1986 cost more to collect andaccount for than they brought in as revenue. Yet larger charges, which might pay fortheir own collection costs, would in the Canadian context have required the exemption ofthose who could not "afford" them •• hospital users tend to be elderly, elderly peopletend to be poor, poor people tend to be sicker •• and this adds further administrativeexpense to determine who these are. The ambiguity of hospital administrators on thisissue is understandable.Such increases in input use represent a reduction in the technical efficiency ofmedical practice. In economic jargon, one refers to a "production function" which linksthe inputs or resources used in production, with the outputs of goods and services thatare of value to users:...(12)Thus the total output of each type of service j is dependent upon the amounts of thedifferent inputs Zk used in its production. This is in one sense a "technical" relation,like (11) above which was a "production function" for health. Both are based upon thepossibilities, in presently known technologies, for combining physical inputs, time, andknow-how to produce certain outputs. Some minimal set of inputs is required toproduce any specified level of output: one can use more, but not less.Equation (10), by contrast, purports to describe how people will behave inrespouse to certain circumstances. In fact, however, both (11) and (12) have a27 Most economists would probably take a less charitableview of extra-billing ·only those who can afford it·, regardingthis as price discrimination according to the elasticity ofdemand, and recognizing the profitability as well as thepolitical appeal of the tactic.28. In February 1993, certain physicians in North Vancouverwho had opted out of the provincial plan pursuant to a feedispute, began extra-billing their patients on the grounds thatopting out had added to their administrative costs.34significant behavioural component as well. An exact relationship between a set of inputsand a corresponding level of output, (12) written as an equation, is commonly assumedin theoretical economics. But this corresponds to perfect (technical) efficiency, no waste,and state-of-the-art technology. It is an abstraction rarely if ever observed in practice,in any industry. Rather (12) holds as an inequality, a boundary condition, with theextent of departure from potential equality representing the degree of technicalinefficiency, wasted resources, in the production process. Behavioural considerationsobvionsly enter into the determination of this degree of inefficiency.This behavioural component is well illustrated in the extra-bllllng case. If theinputs used increase -- more office overhead -- and the service output does not, then theright-hand side of relation (12) increases while the left-hand side does not. Theproduction process becomes less efficient. But the costs of this decreased efficiency areborne by users of care, in the form of out-of-pocket costs, while the practitioner gainsthrough fee increases which are large enough to pay for the increased overhead costs,and still yield a higher net Income."Other adjustments are possible. Physicians sometimes argue that the feesnegotiated with provincial payment agencies are too low to permit them to spendadequate time with their patients. To make an adequate income, they must speedpatients through too fast to provide adequate care," By extra-billing, they are able to29 In elementary "supply-and-demand" models of economicbehaviour, the possibility of the supplier gaining from reducedefficiency is never considered, because it is assumed (usuallyimplicitly and sometimes unconsciously) that suppliers aremotivated strictly by profit considerations, and sell theirproducts to informed buyers in competitive markets withoutregulation or collusion. If all these assumptions held, thenindeed such behaviour would simply result in patients moving awayfrom the extra-billing physician. But these textbookassumptions, ,while analytically convenient, are very far fromreality. Moreover physicians are acutely aware of thispossibility, and take steps collectively to minimize such"market" responses -- as any rational suppliers would. Norshould one focus only on physicians; when "lean and mean" privatesector industries are studied, they too show significantdepartures from equality in relation (12), because the textbookassumptions are at best an approximation there too.30. Such an argument is most interesting. If takenliterally it implies both that physicians can exerciseconsiderable influence over their own patient flow, thuscontradicting the standard economists' assumption of exogenousdemand, and that they are prepared to use that influence, even tothe detriment of their patients, in order to meet self-determinedincome targets.35spend more time with each patient. If this argument Is valid, then one would expectextra-billing to result in a fall In the total volume of patient contacts -- though not as aresult of patient decisions. [As noted above, not every inverse relation between pricesand quantities represents a demand curve.} To the extent that Increased prices wereoffset by decreased quantities, the increased costs borne by users of care would be partlyoffset by reduced costs to taxpayers, while physicians would receive their gains partly ingreater incomes, partly In reduced patient loads.But healthy Canadian taxpayers might be quite short-sighted to conclude thatthey would gain from a return to extra-billing. Physicians' representatives were quiteexplicit, in their opposition to the Canada Health Act, In arguing that the right to extra­bill, whether or not exercised, was a "safety valve" to protect them against unfairlyrestrictive government controls on their fees. This implies that the political pressurecreated by increased extra-billing could force up negotiated fees paid by the publicplans. In this case the prices of services, the PJ, would rise by~ than the CJ,. Implying an income transfer from taxpayers to physicians levered out by thediscretionary nse of user charges.This seems to be exactly what happened in Ontario in the late 1970s and early1980s. Nor did Canadian physicians fail to notice that Quebec, the one province inwhich extra-billing was effectively banned during the 1970s, also had by far the lowestincrease In tees,"Physician extra-billing makes the linkage from user charges to the relativeIncomes of providers quite simple and direct. But the causality may be more indirect.Consider the process of wage negotiation In a public hospital system. Wages forworkers translate directly into Implicit prices and then Into tax rates -- W, P, and t riseor fall together. But suppose that a hard-nosed or hard-pressed government simplyrefuses to adjust hospital budgets to cover a negotiated wage settlement, so that totalcosts do not change. Higher wages now mean fewer positions -- W up and Z down.31 Dentists in Canada illustrate another connectionbetween user charges and overall fee/price levels. Dentalinsurance is predominantly private, with public subsidies, andincludes some degree of coinsurance, allegedly to hold downoverall costs. But patients, when they are asked to pay theirshare of a dental bill, are generally not aware of the relationbetween the reimbursed amount, the coinsurance rate, and anyadditional amount which the dentist may have added on. Thusextra-billing is much less noticeable than it is in thephysician's office where the patient does not normally expect topay at all. Requiring the patient to pay some small amount mayor may not reduce utilization, but it probably on balanceincreases prices.36Hospitals may try to do the same work with fewer people; but If they cannot or will not,then P rises and Q falls.Depending upon the way In which "demands" for service are generated, the fallIn Q can lead to perceived "shortages", queuing, and rationing, and pressure for a spill­out of workload Into a private, self-paying sector. This may be, In effect, what Mr.West was advocating In his remarks quoted above, and It Is a common argument. Letthose who are willing to pay for extra services, pay for them, and thus expand thecapacity (and cost) of the whole system. In the process, one would see an Increase In theIncomes of the fee for service practitioners -- surgeons and anaesthetlsts, for example, orradiologists •• who provide the additional services privately, quite possibly at fees abovethe provincial schedule, and more employment for hospital support workers In theprivately supported facilitles.3ZIn this case, the development of a private service would In essence be "ratifying"the wage Increases In the public sector, converting their effects from a loss of output andjobs, Into an Increase In total system costs, and a shift In both cost bearing and access.Back at the health care feast, those stili at the public table would have succeeded Ingetting larger portions while those displaced have moved to a newly opened privatetable.ThIs form of cost escalation supported by user charges has not been an issue todate in Canada, unless one considers the occasional political theatrics over patients goingto the U.S. a nascent "private sector". But It has been a serious problem In the U.K.,and a threat in Sweden. The problem Is particularly severe If the same persons, usuallyphysicians, are able to work In both the public and the private sectors; they are thenable to manipulate access to the former to steer patients to the latter, where prices arehigher. The best of all worlds Is to be served at both tables.32. Conversely, fee for service practitioners who aredependent upon hospital services lose income from anyrestrictions in throughput. This is most obvious during ahospital strike, but any limitation on the levels of diagnosticor therapeutic activity in hospitals will have a negative effecton the billable workloads of some physicians. Fee-for-servicephysicians are thus economically at risk in negotiations overhospital budgets and capacity levels, so one should not besurprised to find some of them supporting the development ofprivately funded facilities.37Redistribution among Providers: Galners and Losers on the Supply SideSalaried workers in the Canadian hospital sector appear collectively firmlyopposed to user charges or the development of a privately funded delivery system. Buttheir economic interests may be somewhat ambiguous. The emergence of such a systemmight in fact provide powerful support for the contention that the public sector is notadequately funded," On the other hand, even if more private funding did bring inmore resources, ralsing total incomes and employment in the health sector, thedistribution of these benefits among different suppliers might be very unequal and fewmay "trickle down" to the salaried hospital worker. When the mix and volume of careis determined by willingness to pay rather than professionally judged need, the patternof inputs used, and their rates of pay, will change. Some providers lose, even if totalcostslincomes rise.Agaln the U.S. is instructive, and in general the changes are rather what onewould expect. Willingness to pay is highest among those who have most resources, andfeel themselves most at risk and least able to judge the value of the service. Americanphysicians in general have the highest incomes in the world, relative to the generalincome levels in their society, while those physicians who provide services on referral(sometimes from themselves, self-dealing), and who serve the well-endowed or well­insured .- the diagnostic specialists, cardiologists, thoracic and ophthalmologicalsurgeons •• are particularly generously paid.Primary care specialists, on the other hand, who depend more on patient-inltiateddemand and have less scope for recommending mysterious and expensive interventions,have lost a lot of ground economically. Not surprisingly, young physicians areabandoning these areas and specializing where the money is. Those who fear a resulting"shortage" of primary care physicians are applying a professional standard forappropriate levels of care -- patient needs. But the marketplace is instead respondingpredictably and powerfully to the criterion of willingness to pay. The effects onproviders, the pattern of galners and losers, are the invisible hand's way of encouragingresource owners to redirect their resources to the "most valued" forms of output •• onthat criterion.At the same time, efforts to control costs in the U.S. have led to very substantial33. Much of the pressure for major reform in the U.K.National Health Service has arisen from the perceptions, right orwrong, of politicians that physicians and other hospital workershave shared a common interest in low productivity in publichospitals. Waiting lists place political pressure on governmentsto increase health service budgets, while steering patients tothe private side of hospital consultants' practices.38decreases in the use of inpatient care. This has not affected the trend in overall costs,which keep climbing, but jobs and incomes have moved out into the ambulatory caresector. More of the income goes, not only to physicians, but also to technicians, and tothe corporate suppliers of equipment, reagents, and drugs -- those who support the"mysterious and expensive" services. The more traditional inpatient roles - wardnursing, housekeeping, dietary -- are shrinking. There are thus economic as well asideological grounds for hospital unions in Canada to be very suspicious of user charges;their members may not be the ones who gain. Portions at the private table tend to bemuch more unequal than at the public, and some of those displaced from the publictable find no place there."Redistribution to New "Providers"-- Setting New Places at the Health Care FeastBut many others do, who would not be seated at the public table. H extra-billingis "necessary" to cover the extra administrative costs of opting out of provincial medicalcare programs, then a general re-introduction of physician- or hospital-administereduser charges would obviously have a more significant impact on total resources used inthis sector. User charges administered by governments could require the setting up ofadministrative systems to determine who was to be charged, and how much; theenforcement process would also be rather interesting. On the whole, feasibilityconsiderations suggest that any user charge system would be either de facto provideradministered, or integrated with the income tax collection process.But if the prevalence of user charges became so widespread as to lead to the re­introduction of private insurance, then one would also see an Increase in L, the numbersof~ as well as amounts of inputs Z used in the production of health care. Theoverhead costs of financing and managing the delivery of health care which are borne byproviders, and which are significantly increased when funding comes from multipleprivate sources, are embodied in the PJ, the unit prices of the services actually providedto patients. But a private insurance system, in which the R. are increased to take upsome of the increases which would otherwise occur in the CJ, could as described abovebe represented either as adding to the costs of health care services themselves or asadding additional types of services, such as "risk bearing".These "services" require resources for underwriting, marketing, claims34. Fuchs and his colleagues have found that the incomes ofphysicians (relative to the rest of the workforce) are higher inthe United States than in Canada, although their workloads arelower [22]. Hospital workers as a group do not have higher(relative) wages in the United States, but this is because higherwages for the higher-skilled in the U.S. are offset by higherwages for the lower-skilled in Canada [23].39administration, etc., none of which are needed in a universal public system. In equation(4) above these activities can be represented either as new QJ' with their own P and Zcounterparts (increasing M as well as L), or as increased prices, the PJ• and resourceinputs for the existing pattern of services. In the former representation, the additionalcosts correspond to additional services, but these do not enter into any of the relations(10). Insurance salesmen do not in themselves make much contribution to anyone'shealth, separately from the care which their policies may reimburse [nor, for thatmatter, do economlsts]."Alternatively, if the cost of insurance administration is included in the prices ofhealth care services themselves, then it is recorded as a decrease in the efficiency ofproduction of such services. Either way, relative to a universal public system, theseactivities represent "cost without (health) benefit" [5].3635 As noted above, advocates of private coverage have on apriori grounds alleged indirect benefits corresponding to theextra cost, in the form of either improved efficiency oreffectiveness of health services themselves, or increases inconsumer satisfaction. But these "benefits" are seen only withthe eye of faith; actual experience is that private insurancecoverage is associated in aggregate with high costs, questionableeffectiveness, and low expressed satisfaction.36 So why does anyone ever buy insurance? Recall that atthe outset, we pointed out that this paper analyses distributionex post, at the end of the day or year, when all events areknown. And in fact no one does buy or sell insurance for lastyear. In this sense, we may appear to have rigged the gameagainst private insurance. But the baseline for comparison iscritical.Insurance is llQt an efficient way of buying health careservices, in a world of certainty. What insurance buys, in anuncertain world, is risk reduction for the individual, at the(certain) price of higher overall costs for the group. If ourbaseline is a world where t = 0 and Cj = Pj' for all j, then theindividual may be subject to a high degree of financialuncertainly, and insurance may be a good buy even if it raisestotal costs of care.Indeed, it might seem as if we knew that this was so,because people did voluntarily buy such insurance prior to thepublic plans, and in the U.S. they still do. Unfortunately wecannot draw this conclusion with complete certainty, because inboth cases there were large public subsidies through the incometax system to encourage such purchases. We do not know how muchprivate insurance would have been (would be) purchased, if buyershad to pay its full cost.But if instead ones' baseline is a world of universalcomprehensive public coverage, then the (financial) risk which40The massive and rapidly increasing overhead costs of the American healthinsurance system are a natural consequence of the modes of finance chosen. They donot all follow directly from the important role played by nser charges in that system, inthe sense that overheads in the American system pay for much more than simply thelevying and collection of direct charges to patients. But such charges probably !ttpartly responsible, indirectly, for the extraordinary level of administrative waste in thatsystem.In the first place a substantial share of private insurance -- the so-called"Medigap" market, is for coverage against large nser charges for the care of the elderly.The public system imposes such charges, allegedly to discourage excessive nse, and thenthe tax system subsidizes the purchase of private insurance to cover those charges, atconsiderable administrative cost. Each component of the policy looks rational: their sumis transparent lunacy. H user charges were eliminated from the public system, theprivate coverage would disappear, at considerable savings all round. But of course taxeswould be increased....In addition to inducing this inefficient mix of public and private coverage,however, nser fees contribute to the relentless escalation of costs in the U.S. health caresystem, which is unmatched in the rest of the world. This escalation has spawned anextraordinarily wide, and expensive, range of institutions and mechanisms in response.Some are intended to limit overall expenditures; these have failed. Others are intendedsimply to push the costs onto someone else; some of these have been more successful.But each such effort either to control overall costs, or to push them onto someoneelse, against predictable resistance in each case, has added to the overall cost of thesystem. The struggle draws in more administrators, lawyers, marketers, actuaries,economists -- more and more places, and well-fed ones, being set at the health care feast.These additional incomes added on the right-hand side of eqnation (4) become reflected,as described above, in increasing values of the PJ and/or the QJ'the individual would otherwise bear is already being pooled atthe highest collective level. In this context, the introductionof user charges and private insurance coverage corresponds to anincrease in individual risk-bearing (since private coverage israrely as comprehensive as public, and can never be universal) .The administrative costs of such a private system then representnot only "cost without benefit", but ·cost with loss" for thecommunity as a whole. For those individuals whose taxes fall bymore than the increase in their private premiums and usercharges, the healthy and wealthy again, the private system may bepreferable. As it may be for the wealthy and unhealthy (so longas they are not too unhealthy) who want to be sure that anylimitations on access do not bear on them.41User charges play a central role In this cost-shifting process, both because theimposition of such charges is a major component of that process, and because theadoption of more effective and less wasteful iustltutious for cost control has been heldoff, politically, by coustant repetition of the claim that user charges rather than directpolitical action will deal with the problem. False hopes lead to false remedies; and theadministrative "therapies" have become more and more expensive as the situationdeteriorates.Canada is not the United States; the point is not that "It can happen here!"[although in fact we believe that it could]," Rather the U.S. enviromnent provides thebest illustration of how user charges can contribute to redistributing incomes within thehealth care sector itself, by providing opportunities for a number of other "factorsuppliers" -- providers of administrative and financial services -- to increase the incomesthey draw from the health care sector. As they do so, either total costs go up, or theincomes of those who provide clinical services must fall.Again, it is not accidental that the American College of Physicians, the secondlargest physicians' organization in the U.S. after the A.M.A., came out several years agoin favour of a national health insurance plan. Its members have noticed their portionsbeing squeezed, despite the continuous increase in overall costs, by the rapid increase inthe numbers of non-clinicians being served, as well as by the overflowing plates of thesurgeons and radiologists.Nor are sellers of administrative and financial services the only "health carewannabees" who might quite reasonably hope to gain from some form of privatefunding for health care. There are also "wannabees" whose ambitions focus morespecifically on the QJ' as they offer services which are on the fuzzy boundary of thehealth care system.Chiropractors and naturopaths, for example, have no doubts that they areproviders of health care, and these occupations display many of the features of thehealth professions. But their services are not within the scope of the Canada HealthAct; provinces need not cover them to meet the federal standard of comprehensiveness.Many physicians do not recognize them as "members of the health care team"; quite thecontrary. And the public are split; while some use such services routinely, others do notcontemplate any circumstances under which they would visit a chiropractor ornaturopath.Correspondingly, the services of such practitioners are reimbursable at provincial37 And we note that there have been, from time to time,statements from people in the Canadian insurance industry insupport of a re-introduction of private funding, both user feesand private insurance.42option, on whatever terms provinces choose. Provinces can impose user charges andlimits on the number of reimbursable visits and procedures, and do. These practitionersregard such restrictions as unfair, placing them at a competitive disadvantage relative tophysicians. They argue that either their services should be fully and comprehensivelyreimbursed, or - given the fiscal realities of the times, user fees should be imposed onphysicians' services as well "to level the playing field". The explicit intent of the userfee in this case is to increase the flow of income to this group of practitioners -- factorsuppliers -- at the expense of physicians and/or their patients.Such practitioners are at the edge of health care as currently defined by thereimbursement system; many others who claim to offer health-enhancing services areoutside that definition. They deal with their clients on a straight-forward market basis.All would benefit from inclusion within the public reimbursement system, and have aneconomic interest in seeing that system made more expensive, relative to their ownservices. While very diverse, this group is not large. But it represents still another formof interest in the expansion of user charges for "conventional" hospital and medicalcare.ConclusionThe principal effects of user charges are straightforward income transfers -- whatPeter loses, Paul gains -- and whether one thinks this is a good or a bad policy dependsultimately on what one believes about the relative deservingness of Peter and Paul.[This in turn will be in part correlated with whether one is Peter or Paul.] In this paperwe have developed a comprehensive accounting framework within which one can keeptrack of all of these transfers, and have used that framework to trace out the effects ofseveral suggested forms of user charges. One can demonstrate quite rigorously thatsuch charges consistently transfer income (net of health costs) not only from the less tothe more healthy, but also from those with lower to those with higher incomes.Arguments to the contrary, based on a partial or incomplete specification of the impactof charges, are shown to be erroneous. In effect they fail to address the question "Whathappens next?' or "How do the accounting identities stay balanced?"The accounting identities also force one to be explicit about the relationshipbetween user charges, and the incomes of and opportunities for different types of serviceproviders. To the extent that such charges change either the volume or the mix ofservices used/provided, or the prices received, they must simultaneously change theincomes and/or work opportunities of providers. Who gets paid, and how much, forsupplying their time, their skills, or their capital to the health care sector? If user .charges are the first step towards an environment with more private fonns of financingfor health care, this will have significant effects on the mix of inputs used in itsprovision, and particularly its administration and financing, and on the rates ofreimbursement received by those suppliers.43Here the issues are in part distributional •• providers on average will gain relativeto the rest of the community, but some will gain and others may lose. But the mode ofpayment can also have a significant bearing on the total volume and cost of resourcesused up to provide a given amount of health care - on the technical efficiency of thesystem. Some systems cost a lot more to run than others, paying high rewards fordispensing paper as well as pills. And paying people to push either paper or pills, whenthere is no corresponding improvement in anyone's health, is waste pure and simple.Enhanced market opportunities in the financial sector come at the expense of overalllosses, on average, for everyone.The effects of user charges on access to care, and ultimately on health, cannot berepresented within the accounting identities linking financial flows. As noted above,however, evidence seems to support quite strongly the intuitively obvious conclusion,that user charges will increase the access of those with greatest willingness/ability to pay,at the expense of those with less [9]. Some of their advocates are explicit that this is theobjective. Such changes in the types of services provided, and to whom, may haveeffects on the health of recipients .- and non-recipients •• raising questions about theway in which different payment systems influence the translation of health care servicesinto health outcomes. To what extent do the people who "need" care, in the sense ofhaving the capacity to benefit, actually get the care they need, and how might this beaffected by nser charges?Again other evidence _. principally but not entirely from the United States -­shows that willingness or ability to pay is not correlated with (expert-judged) need forcare [9]. People who are deterred from care-seeking by user charges are as likely toforego "needed" as "unneeded" services. And (again in the United States) the negativehealth consequences for those "deterred" are increasingly documented. In the Canadiansystem (unlike the American) roughly ninety percent of all Medicare expenditure is forservices requiring the explicit recommendation of a physician! It is thus difficult toargue that patients making "frivolous" demands, whether knowingly or otherwise, cangenerate more than one or two percent of overall costs without the support or at leastthe acquiescence of a physician.But these considerations, though very important, take us beyond the scope of thefinancial relationships. What those relationships demonstrate quite clearly, by takingfull account of all the necessary responses and adjustments (necessary in a logical, not anormative, sense), is that whatever eise they do, user fees redistribute income. And theyconsistently redistribute, not only from the sicker to the healthier, but from those withlower to those with higher incomes. In this light, the concentration of advocacy amongthe well-to-do and their representatives makes perfect sense.44REFERENCES[1] Letsch, S.W. (1993) "National Health Care Spending in 1991" Health Affairs 12:1(Spring) pp. 94-110.[2] Schieber, G,J., J.-P. Poullier, and L.M. Greenwald (1992) "U.S. healthexpenditure performance: An international comparison and data update" HealthCare Financing Review 13:4 (Summer, 1992) 1-87.[3] Bardin, S., I. Morrison, R Mittman, A. Saveri, G. Schmid, and J. Wayne (1992)Health. Health Care. and Health Policy: An International Perspective (SpecialReport SR-472) Menlo Park, CA.: Institute for the Future (July 31).[4] Evans, R.G., M.L. Barer and G.L. Stoddart (1993) "It's Not the Money, It's thePrinciple" Toronto: The Premier's Council on Health, Well-being and SocialJustice Discussion Paper (forthcoming).[5] Woolhandler, S., and D.U. Himmelstein (1991) "The Deteriorating AdministrativeEfficiency of the U.S. Health Care System" New England Journal of Medicine,Vol. 324, no. 18 (May 2) pp. 1253-8.[6] Danzon, P. (1992) "Hidden Overhead Costs: Is Canada's System Really LessExpensive?" Health Affairs 11:1 (Spring) pp. 21-43.[7] Thorpe, K.E. (1992) "Inside the Black Box of Administrative Costs" HealthAffairs 11:2 (Summer) pp. 41-55.[8] Barer, M.L., V. Bhatia, G.L. Stoddart and RG. Evans (1993) "The RemarkableTenacity of User Charges" Toronto: The Premier's Council on Health, Well­being and Social Justice Discussion Paper (forthcoming).[9] Stoddart, G.L., M.L. Barer and RG. Evans (1993) "User Charges, Snares andDelusions: Another Look at the Literature" Toronto: The Premier's Council onHealth, Well-being and Social Justice Discussion Paper (forthcoming).[10] Evans, R.G., M.L. Barer, G.L. Stoddart and V. Bhatia (1993) "Who Are theZombie Masters, and What Do They Want?" Toronto: The Premier's Council onHealth, Well-being and Social Justice Discussion Paper (forthcoming).[11] Feldstein, M.S. (1971) "A New Approach to National Health Insurance: ThePublic Interest (Spring) pp. 93-105.[12] Rice, T. and K.E. Thorpe (1993) "Income-Related Cost Sharing in HealthInsurance" Health Affairs 12:1 (Spring) pp. 21-39.45[13] Frank, R.H., T. Gilovich and D.T. Regan (1993) "Does Studying EconomicsInhibit Cooperation?" Journal of Economic Perspectives 7:2 (Spring) pp. 159·71.[14] Evans, R.G., and G.L. Stoddart (1990) "Producing Health, Consuming HealthCare" Social Science and Medicine, 31:12 pp. 1347·63.[15] Barer, M.L., R.G. Evans and T.R Marmor (1993) Why are Some People Healthyand Others Not? Hawthorne, N.Y.: Aldine-de Gruyter, forthcoming.[16] The Toronto Globe and Mail (1993) "Alberta Premier to Press for Medicare UserFees" Wednesday, January 13, p. A3.[17] Woodward, C.A., J.R. Gilbert, RS. Roberts et al. (1983) "When is a Patient's Useof Primary Care Services Unwarranted?" Canadian Medical Association Journal129 pp.822·7.[18] Chassin, M.R, RE. Park, A. Fink et al. (1986) "Indications for Selected Medicaland Surgical Procedures •• A Literature Review and Ratings of Appropriateness(R·3204/2-CWFIHFIHCFAlPMTIRWJ) (May) The Rand Corporation, SantaMonica, CA.[19] Rouleau, J.L., L.A. Moye, M.A. Pfeffer et al. (1993) "A Comparison ofManagement Patterns after Acute Myocardial Infarction in Canada and theUnited States" New England Journal of Medicine Vol. 328, no. 11 (March 18) pp,779·84.[20] Nair, C., R Karim, and C. Nyers (1992) "Health Care and Health Statns: ACanada-United States Statistical Comparison" Health Reports, Vol. 4, no. 2(October) Ottawa: Statistics Canada (cat. no. 82·003), pp. 175·183.[21] Reinhardt, U.E. (1982) "Table Manners at the Health Care Feast" in D. Yaggyand W. Anlyan, eds, Financing Health Care: Competition vs. RegulationCambridge, Mass.: Ballinger, pp. 13-34.[22] Fuchs, V.R and J.S. Hahn (1990), "How Does Canada Do It? A Comparison ofExpenditures for Physicians' Services in the United States and Canada", NewEngland Journal of Medicine 323(13):884·890.[23] Redelmeier, D.A., and V.R Fuchs (1993), "Hospital Costs in the United Statesand Canada" New England Journal of Medicine Vol. 328, no. 11 (March 18) pp.772-8.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items