Policy Research WOFkKING-PAPERS CountryOperations CountryDepartment Public Disclosure Authorized TheWorld Bank February1993 WPS 1090 GovernmentExpenditures as a Citizens' Evaluation of Public Output Public Disclosure Authorized Public Choice and the Benefit Principleof Taxation Public Disclosure Authorized ThanosCatsambas Elementsfrom the theoriesof publicchoice and benefit taxation are combinedto form a conceptualframework for quantifying citizens' preferences about privately available and publicly providedgoods and services.When the theoreticalresults were Public Disclosure Authorized applied to four Central European countries, under certain behavioraland stabilityassumptions, both the evaluationof and desired level of public output by individuals was found to be lower than that actuallyprovided by government. Poicy RewhWokingPapssdiinaMthefindi&nofwork in pmg nda conangethcexchangeofideas among Bank staffand aUrothmutestdindivecopmentissuesepapeamdistnbutedbytiReeachAdvis5 yStaff,cary thename ofthe uthors,reZflet onlytheirviws,andshoud beusd andcitedaccoringly.Tlhefindings,i ons.andconclusionsarthe awhom'own. heyshould not be sttlbuted to the Wodd Bank.its Boardof Directors,its management,or anyof its manber countries PolicyResearch CountryOperations WPS 1090 This paper-a product ofthe CountryOperations Division, Europe and CentralAsia, Cour.try Department 11-is part of a larger effort in the departmentto addressissues of economicpolicy in CentralEuropean countries through analytical techniquesbased on marketeconomic principles.Copies of the paper are available free from the World Bank, 1818H Street NW, Washington,DC 20433. Please contact Anita Correa, room H11-107, extension 38549 (February 1993, 16 pages). Combiningelements from the theories of public Interpretingindirect evidence,he shows that the choice and benefittaxation, Catsambasdevelops privaZasector wouldprefer less government a frameworkin whichprivate cidzens can acdvity in all countries,from a low of 5 percent evaluatepublic activities. less public spending(in Poland)to a high of one- third less (in Slovenia).If those governments Why, and under what circumstances,do were to followthose guidelines,their spending- "bureaucrats"increase the size of the public to-GDPratios wouldmore closelyresemble the sector and the amountof public spendingin their 1987-89average for a selectedgroup of own self interest? Europeanmarket economies. What does the private sector think public Catsambasalso introducesa more rigorous, output shouldbe, what is actual publicoutput, if not necessarilymore objective, approach to and how does the private sector evaluatethat determining"optimal" government spending. output? This approachrequires little information,but uses i static modeland requires faith in the Catsamrbasapplies the theoreticalresults of direcdon of causality for some key variables.To an attemptto answer these questionsin four the extent that one can accept those limitations, CentralEuropean countries(Czechoslowakia, the model may be a useful operationaltool in Hungary,Poland, and Slovenia),using actual public spendingevaluation. data for 1989-91and projectionsfor 1992. ThePolicy Research Working Paper Series disseminates the fmdings of workunder way in theBank. An objective of the series is to get these findings out quickly, even if presentationsare less than fully polished.The findings,interpretations, and conclusionsin thesepapers do not necessarilyrepresent official Bank policy. Producedby the PolicyResearch Dissemineion Center GOVERNMENT EXPENDITURES AS A CITIZENS' EVALUATION OF PUBLIC OUTPUT: PUBLIC CHOICE AND THE BENEFIT PRINCIPLE OF TAXATION THANOS CATSAMBAS CENTRALEUROPE DEPARTMENT THE WORLD BANK' I. INTRODUCTION The questionof the optimal level of public expenditureshas long concernedstudents of public finance. In his classical analysis of the efficient level of public expenditures,Samuelson [1954] introduceda reference private good and developedthe well-knowncondition that an optimal level of expenditurerequires that the sum of the marginal rates of substitutionbetween the public good and the private good must equal the marginal rate of transformationbetween the two goods. Samuelson's theoreticalresult assumes that all of the revenue needed to financepublic expenditurescan be raised with lump-sumtaxes. Since this is not generally possible, the result must be modified to accountfor the distortionaryeffects of an actual tax system. Ballardand Fullerton [1992] call this modification the "marginal cost of public funds" (MCF), and provide an interestingexplanation of how several earlier attempts to address this problem can be seen under the unifying prism of this concept. They argue that many importantcontributions, including the early Harberger [19641approach to the excess burden of taxation, and the subsequentcontributions by Stiglitz and Dasgupta [1971] and by Atkinson and Stern [1974], can all be interpreted along the lines of the MCF approach. Wilson [1991] persuasively argues that the optimal level of public good provision must take into account not only efficiency but also distributionalconsiderations. Along these lines, the relationshipbetween taxes and expenditures,and its implicationfbr the optimal provision of public output, can be seen from yet another angle, through the postulates of the "benefitprinciple of taxation". According to this approach, an equivalenceis drawn between the market mechanismand the provision of public expendituresthrough the budget. If tax payments are regarded as a "price" for the provision of public expendituresby Government, then taxes are related to expendituresthrough Samuelson's conditionof efficiency, namelythat the sum of unit tax shares must equal the stumof the marginal rates of substitution,which is equal to the marginal cost of the public good. This is all that is required for efficiency. If, by chance, the individualtax price (or unit tax share) is moreover equal to the individualmarginal rate of substitution,the situation is called a "Lindahlsolution", after Erik Lindahl 119581,who first described the analogy to market equilibrium of public expenditure determinationin connectionwith the distributionof tax burden among various groups. 2 I/ The author is gratefulto Martha de Melo for hcr insightfulcomments, but remains responsiblefor the analysis and conclusionsof this draft. 2! Perhaps a more appropriateterm would be thc "Lindahl-Johansen"solution, since it was the latter who popularized Lindahl's approach throughan excellentexposition. See L. Johansen11965]. This approach is related to the benefit principle of taxation because the taxes paid by individualsmay be interpreted as a "price" for the provisionof public expenditures. The optimal amount of public output is then determinedby the conditionthat the tax price must equal the marginal rate of substitutionbetween the public good and the reference private good. This is a powerful result, which has implicationsnot only for the efficient provisionof public expenditures,but also for their distributionalimplications. In a seminal article along tiese lines, Aaron and McGuire [19701analyze the evaluationof public output by individualsand concludethat the beneficiariesof government programs may have diverse perspectivesin their appreciationof public expendituresdepending on their perceived "tax price". In particular, Aaron-McGuireshow that if the unit tax share is less (more) than the marginal rate of substitutionfor an individual,the individualevaluates the public good higher (lower) than the tax it actuallypays to the Government. Therefore, if the actual tax shares are observed, we may determine whether the provisionof public output is higher or lower than the optimal amount desired by individuals. This condition, in turn, determines whether the provision of public services has a positive or negative redistributiveimpact on those individuals. II. PURPOSE AND CONCEPTUALFRAMEWORK OF THE PAPER This paper attempts to shed light on the issue of optimal governmentspending by modifying and extending the Lindahl-Aaron-McGuireapproach to benefit taxation. Its spirit is similar 'Loseveral recent contributionsin this area, which emphasizethe public choice approach to many old questions in public finance.3 In doing so, the analysis introduces elementsfrom the theory of bureaus, and attempts to combine the two approaches into a unified theory of optimal governmentspending. The theory of bureaus (Niskanen [19711,[1975]),recognizes that demand for governmentactivities is expressedthrough the political system, in which "bureaucrats"may play an independentand at the same time powerful role. The basic tenet of this theory is that divergent interests among different groups of society, and, notably, between citizens and bureaucrats, may lead to a discrepancybetween the amount of public services demandedby citizens and that provided by Government.4 In other words, the new theory of public finance recognizesthat the outcomeof governmentactivities depends on both the demand and the supply of goods and services to be provided by the public sector. In recent years particular attention has been paid to the institutionsof upply, i.e. the government bureaucracy, and on how their decisions may influencethe total amount of public services. This perspective is particularlyrelevant for countries that are emerging from a long system of commandeconomy, such as the countries of Central and Eastern Europe. The role of Government under the previous regimes was overwhelming,and it is not clear to what