Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Multi page


Published on

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

Multi page

  1. 1. J$ 16o0}POLICY RESEARCH WORKING PAPER 1607Population Aging As _hinas po ulation age,-China must acdressand Pension Systems macroecononicssuesaffecting its ol.-age sociaiReform Options for China securitysysten. rotjust thedesignand m,-inagemeneotpension fundsF. Desmond McCarthyKangbin ZhengThe World BankOffice of the Senior Vice PresidentDevelopment EconomicsMay 1996PublicDisclosureAuthorizedPublicDisclosureAuthorizedPublicDisclosureAuthorizedPublicDisclosureAuthorized
  2. 2. IPOLICY RESEARCHWORKINGPAPER1607Summary findingsUsingan integrated simulation model, McCarthy and Reai income growth is needed to cope with povertyZheng estimate the scope and speed of population-aging among the elderly, especiallyin developing countries. Toin China, the cost of supporting the old, and the impact establish an adequate, efficient, and equitable socialof different reform options and pension arrangements. security system, China must maintain long-termAmong their conclusions: socioeconomic stability and sustainable growth.The scope and speed of popuiation-aging in China China could improve the labor market by removingmake the present pension system financially management rigidities, facilitating human resourceunsustainable, even assuming that GDP grows steadilv in development, making labor markets more competitive,the long term. Moving the retirement age back would improving the household registration system, improvingprovide a temporary fix for the current pay-as-you-go incentives, and rewarding hard and innovative work. Topension system but would be politicallv viable only reduce unemployment, China can create more jobwhere there isgreat demand for labor. oppor-tunities in nontraditional sectors, especially itsPension funds could be made more sustainable by underdeveloped service industries. To shift jobs to theincreasing GDP growth, raising contribution rates, or nonagricultural sector, it can develop medium-size cities.gradually reducing benefit rates. Blutthe financial costs And to cushion the impact of demographic shocks, Chinaand social obstacles of those refornmoptions must be should preserve traditional values and maintain family-carefullyassessed. Ftullyfunded, privately managed community support.pension schemes might be feasible,but require a sound Drawing onIexperience in Europe and Latin America,regulatory framework and institutional infrastructure, China should move toward a transparent andincluding financial markets that provido adequate siavings decentralized system with (1) a fully funded, portable,instruments and insurance optOls. defined-benefit pension plan, designed to meet basicPension reform is a long term, niuitidimensional needs, and (2) occupational pension plans or personalproblem involving economilic,social, political, and savingsaccounirsto satisfy demand for maintaining orCultural factors. Governments should not focus only on improving living standards.taxes and transfers to redistribute incone to and amongthe elderly.This paper - a product of the Office of the Senior VicePresident, Development Economics - ispart of a larger effort tostudy strategic and policy issues for socially sustainable development in developing countries. Copies of the paper areavailable free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Mila Divino, roomN6-056, telephone 202-47/3-3739, fax 202-522-1157, Internet address mdivino@q? 1996. (47 pages)The Policv Research WorkingPaperSeries disseminates the findings of uork in progressto encourage the exchange of ideas aboutdevelopment issues.An objectiveof the seriesistoget the findingsout quickly,even if thepresentationsare lessthanfully polished. Thepapers carry,the names of the authorsand should be usedand cited accordirngly.Thefindings,interpretations,and conclusionsare theauthors oun and should not be attributed to the World Bank, itsExecutive Boardof Directors, or any of its member countries.ProdJK d !lEt-i PolicyRcsca- h Dissemination C((intc
  3. 3. Population Aging and Pension SystemsF. Desmond McCarthyandKangbin ZhengThe World BankOfficeof the SeniorVicePresidentDevelopment Economics JEL ClassificationNos: H55, 138Thispaperispartofalargerefforttostudystrategicandpolicyissuesforsociallysustainabledevelopmentin developing countries. The authors thank Estelle James, Cheikh Kane, Anita Schwartz, andparticipantsofan internationaleconomicsseminarat MITforhelpfulsuggestionsanddiscmssions.
  4. 4. II
  5. 5. ContentsI. INTRODUCTION III. CHALLENGESAHEAD 4Traditional Values 7Demographic Transition and Population Policy 8Structural Adjustment and Rural-Urban Balance 14Sustainability of Socio-economic Transition 15III.A SIMPLESIMULATIONMODELFOR PENSIONREFORMS 18The Pension Fund Balance 19Demographic Dynamics and Labor Market 21Economic and Wage Growth 22Government Behavior under Different Pension Arrangements 22Pay-As-You-Go Scheme 23Defined Contribution System 24Defined Benefit System 25Private Pension Fund Accounts System 25A Multi-pillar Pension System 26IV. PENSIONSYSTEMREFORMIN CHINA 27The Current Pension System in China 27The Need for Pension Reforms: The Benchmark Case 31Base Year Paramneters 31Key Assumptions 31Pressure from Population Aging 33The Base Case Scenario 35Policy Simulations for Reform Options 37Delaying Retirement Age 37Change of GDP Growth Path 39Changing the Contribution Rate and the Replacement Rate 41Fully Funded Individual Accounts 41V. PRELIMINARYCONCLUSIONS 43REFERENCE 46i
  6. 6. List of FiguresFigure 1 DevelopmentDiamond 1992 6Figure 2 Population Aging 10Figure 3 Population Growth 1985-2150 11Figure 4 AverageAge of Population 1985-2150 12Figure 5 Evolution of Age Structure 13Figure 6 Links Among Pension Performance and Socio-EconomicDevelopment 18Figure 7 PublicPension System Development 1978-1993 30Figure 8 Population Dependency Rate With Different Sets of Retirement Age 38ii
  7. 7. Population Aging and Pension Systems1. IntroductionWhilebeing proud of its civilization as one of the worlds oldest, China,thoughrelativelyyouthful at present, is getting older rapidly and massively. Strict familyplanningpractice has sharplyreduced fertilitysince early 1980s. Better livingconditions andmedical services have cut down mortality and extended life expectancyin both urban andrural areas. Consequently,the percentage of older persons in the total populationincreases continuously, andthe median age of the population rises steadily. Moreover, theaging process is expected to accelerate in the coming decades as the babyboomers bornbetween early 60s and mid-70s, currently numbering more than 250 million(twice Japanspresent total population), approach retirement age around 2030. By then as manyas 22percent of Chinesenationalswillbe more than 60 years old,jumping from9.7 percent in1995. During the same period, the ratio of working-age persons to the aged is estimatedto be halved from 9.7:1to 4.2:1. Without a properly designedsocial security systemandstrategically formulatedpolicy responses, it would be very difficultfor Chinato cope withthis coming old-age challenge.The pressure generated from these demographic dynamicsfor Chinainthe 2030sshare some similaritieswith those facing Western developed countries and Japanin thenext decade. However, China is a low-income country, with per capita incomeof US$490 in 1993 beingless than 2 percent of that in the United States at the current marketexchange rate, or close to 10 percent when measured at purchasingpower parity.Apparently, Chinas quick population aging process and lower levelof developmentmeanthat income resourcesavailableto support non-productive aged people willbe far lessI The WorldBanksWorldDevelopmentReport 1995, page220,presentsa crosscountrycomparisonofPPP estimatesof GNP. HereChinas per capita GNPin 1993is estimatedat US2,330ascomparedwiththatofIndiaat US$1,220.1
  8. 8. than in the developedworld. Even with the Confucian tradition of filialpiety and respectfor the elderly,Chinawith relativelyfewer young will soon finditself in difficultiestoprovide adequate support to its rapidly increasing old-age population.Faced by these serious challengesfor harmonizing economic growth with socialsustainability, the Chinesegovernment has begun to take measuresto reform its currentlyfragmented social securitysystem. Policy options on pension systemreformshave beenstudied, and several experimentsfor social security decentralizationhave been carriedoutat provincial and municipallevels. State and collective enterprises began pooling pensionfunds in order to ease unevenburden of retirement costs. However, manylong-termstrategic issues on the interdependencebetween demographic transitionand economicgrowth remain to be systematicallyinvestigated. For instance,what is the realfinancialburden the population agingprocess creates, given the fact that both the incomegrowthrate and the average privatesavings rate in China rank among the worlds highestin recentyears and will probablycontinueto be so in the near future? What are the macroeconomicimplicationsifthe current pay-as-as-go type of pension scheme were not reformed? Howto deal with the insolvencyof pensionfunds associated with state-owned enterprises?What roles shallindividualworkers, enterprises, and differentlevelsof governmentsplayto establish sufficient,efficient,and equitable social security systems? How can pensionfunds be most effectivelymanaged, privately or publicly? How are differentschemessociallyand financiallysustainable? What are the politicalimplicationsassociatedwithpension system reformssinceany adjustments on this front will inevitablynecessitate inter-generational transfers andintra-generational redistribution among differentincomegroups? How muchcan Chinalearn from experiences of both developedand developingcountries?This paper seeks to make a contribution to policy thinking on these importantquestions. The theme of the paper is that issues around the old-age social security systemin most developingcountries, such as China, need to be addressed ina broadermacroeconomic sense rather than over-stressing on design and managementof the pension2
  9. 9. funds pe se. The ultimatemeans to endure and improve living standards for the oldpopulation is to maintainlong-term stabilityand promote sustainableincomegrowth,while reforming micro-incentivestructure to encourage private savings. Workers,pensioners, enterprises, and differentlevels of governments need to cooperate to improvelabor productivity, create newjob opportunities, preserve traditional andfamilyvalues,liberalizeand develop financialmarkets, and rationalize government interventionsinthesocial security system.There exist extensivediscussions, as cited in World Bank (1994), on pensionschemes and social security systems. However, only a few are directlyconcernedwithpolicy applications in China2 . Using a simple simulation model similarto that of Schieberand Shoven (1994) and the World Banks population forecasting for China(World Bank1995), this paper attempts to quantitatively illustrate: (a) the scope andthe speedofpopulation aging and its associatedcost of supporting the old in the future, (b) whether atypical pay-as-you-go systemis sustainable and how large the cost if no reformswereconducted, and (c) systemsensitivitieswith respect to alternative policy choices.Although other kinds of benefitsand compensations, such as health insuranceanddisabilitybenefits, are integralparts of the social security system,the focus here ismaintained on pensions for retired persons.The rest of the paper is organized as follows. Section II analyzesmajorchallenges China willface duringits demographic and socio-economic transition. SectionIm presents a simple model built around a pay-as-you-go systemwhichintegrates demandfor old age supports with supply of funds available from workers contribution. SectionIV focuses on the sustainabilitysensitivitiesof different policy alternatives. Section Vconcludes this paper with someremarks and suggests directions for furtherpolicy studies.2 A partiallistofstudiesonpensionsystemreforminChinaincludesAhmadandHussian(1991),Barkan(1990),DixionandMacarov(1992),Hu(1994),Hussain(1993,1994),HussianandLi(1989),Lee(1993),Tian(1995),WorldBank(1990,1994),andYue(1985).3
  10. 10. 11. ChallengesAheadLiving on about 7 percent of the arable land of the earth, Chinas population of 1.2billionaccounts for about 21 percent of the worlds total and surpasses that of India, thesecond largest in the world, by more than 30 percent. Significantsocio-economicprogresses, as shown by the indicators in Table I interms of both developmentdynamicsand international comparison, have been made in the past decades. These achievementsare especially apparent within the past 15years, when market reforms andintegration intothe world trade and financesystems resulted in noticeable improvement inproductionefficiency and incomegrowth. Withinthe period between 1980 and 1994, real GDP grewat a rate close to 10 percent per annumon average, the number of persons livinginabsolute poverty was halved, and the internationaltrade volume rose by fiveto sixtimesto more than US$200 billion. Thisremarkable growth was largelyfinancedby domesticsavings, although foreign financialassistance through investmentand lending alsoplayedan important role. As exemplifiedinthe 1992 Development Diamond presented inFigure -1, China compares favorablywith most low-income economiesin all four dimensions.While fundamentaltransformation and economic growth will probablycontinue ata fast pace in the first decades of the twenty first century, a number of concernsfor longterm sustainabilityare emerging. The old-age security systemis one of more pressing.Out of many social, political,and economic factors that can generate significantimpactonthe social security systemin the near future, the followingfour intertwinedfactors arecentral for policy formationto avert a potential old age crisis:(a) how to promote growth while preserving traditional values?(b) how to adjust population policiesto cope with the demographicdevelopment?(c) how to divert labor force concentration from agriculture to industriesandserviceswhile maintainingrural-urban balances? and(d) how to mitigatethe social cost during the transitionfrom a centralplanningregime towards a competitive market economy?4
  11. 11. Table IBasic Indicators of Social and EconomicDevelopment(1993unlessotherwisestated)OtherLow- Middle-China India Income Income WorldEconomies" EconomiesPopulation million 1178 898 1015 1597 5501Urban % 29 26 27 60 37GrowthRate(80-93) % 1.4 2.0 2.5 1.7 L7Total FertilityRate 2.0 3.7 5.5 3.0 3.2LifeExpectancy years 69 61 56 68 66AdultIlliteracy % 27 52 49 17 33PrimaryNetEnrollment % 96 57Accessto SafeWater % 71 75GNPper capita USS 490 300 300 2480 4420GDPGrowthRate (80-93) % 9.6 5.2 2.9 2.1 2.9Avg.InflationRate (80-93) % 7.0 8.7 27.1 90.1 19.6Exports US$bil. 91.7 21.5 43.2 648.2 3701.5GrowthRate (80-93) % 11.5 7.0 1.4Imports US$bil. 103.1 22.8 62.9 724.6 3778.7As Shareof GNPGrossDomesticInvestment % 41 24 17 23 22GrossDomesticSavings % 40 24 10 22 22Agriculture % 19 31 37Industry % 48 27 22Service % 33 41 42Source:WorldBank(1995):WorldDevelopmentReport1995:WorkersinanIntegratedWorld1/ excludingChinaandIndia.5
  12. 12. Figure 1Development Diamand: 1992Low-IncomeEconomiesexcludingChinaand India = 100.0180.0 -GDP per capita160.014 020.0-Accessto Safe WaterLife Expectancy/ / -~~China- - - Low-IncomeEconomiesGrossPrimaryErollExcludingChina& IndiaGross Primary Enrollment
  13. 13. Traditional ValuesFor thousandsof years,traditionalvalues inChinahaveplacedstrongemphasisonthe roleof the familyas aunit of production and consumptioningeneral,andas abasisforraisingthe young andsupportingthe oldinparticular. Three or four generationsusuallyliveunder the sameroof to pool resourcesand sharerisks. Whengrandparentsinordinaryfamiliesbecometoo oldto generateincomedirectlyfrom workingoutside,they stayat homenot onlytojust restbut to take care of young grandchildrenand do houseworkinorderto reduceburdens of the rniddlegeneration. Thissystemworked quitewell,ofteninthe absenceof anyformalmarketwork arrangements. The virtuesof diligence,frugalityandthriftareglorified.Accumulationof socialexperienceand financialwealth, largelybequest-motivated,by the oldthroughtheir life-longhardwork wins them respect fromthe society. Thusthe welfareofolder membershasbeenan essentialresponsibilityof families,as a rewardtothe altruisticspiritof the oldergeneration.Thistraditionalnaturalharmonybetweengenerationswithinthe familyand amongmembersof a communityis now being disruptedas the economymodernizesandopens to extemalinfluences.The familiesthemselvesaregettingsmaller,especiallyinurbanareas becauseof housingproblems. Whilea higher percentageof familymembersarelivinglonger,traditionalextendedfamilylivingarrangementsarebeingstressedbythe needfor somemembersto becomemoremobileto respond to new economicincentives.Understandablyinanincreasinglysecularsocietymore activitiesare valuedthroughmarketinteractions,andthereis often moreemphasison selfgratificationand less on commnunalresponsibilities.In China,significantimportanceneeds to be attachedto preservingthetraditionofrespectingandsupportingthe oldon familyand communitybasis. Currently, Chinas bigpopulation is mostlyrural with low income. About 10 percent of the population are stillliving in absolute poverty with an income less than half a dollar a day. In addition, 27percent of adults are illiterate. For persons 60 years or older, about one inten women andtwo in ten men can actuallyread and write. Socialized old-age securitysystems throughoperations of pensionfunds and insurance arrangements are unlikelyto play a major roleuntil later inthe next century, because of Chinas current developmentlevelof economy7
  14. 14. and infrastructure. Nursery homes,if built to house Chinas 360 Millionoldpersons inthe2030s, would be very costly andlabor-intensive, and maynot necessarilyprovide betterservices to the elderlythan within a familyatmosphere. Hence, how to preserve andpromote traditional familyvalues shallbe a key policy issue in designingold-age securitysystems in China.Demographic Transition and Population PolicyAnother challengecomesfromthe inertia of the demographictransition. Chinaspopulation, unevenlydistributedacross regions, is projected to grow at an averageannualrate of less than 0.9 percent between 1995 and 2030. Thisgrowth rate is onlya half ofthe 1.8 percent rate between 1970 and 1980 and substantiallylower than 1.4 percent ratebetween 1980 and 1993. But the size of the incremental population is enormousbecauseof the much larger populationbase. Close to a net of 430 millionpeople will be added toChinas population withinthe next 35 years, which is equivalentto fivetimes of thecurrent total population inGermany. More than half of thisincrease can be attributed toextensionof lifeexpectancy,whichis forecast to advancefrom69 yearsat birthin 1995to 75yearsin2030. By 2030, closeto 30 percent of the worlds incrementalold-age population,as compared with that in 1990,will be accumulated in China. The normalpopulationagingprocess seenin mostcountrieshasbeen particularlyrapid inChinaduringa longperiodofsteadygrowth, when overalleconomicconditionsimprove andaccessto betterhealthfacilitiesbecomeswider. In France, 140 years elapsed before the proportion of the populationover60 doubled to 18percent in 1976. It will take Chinajust 34 years. The old-agedependency ratio willrise from 15 in 1995 to 38 in2030, while the averageage of thepopulation increasesby 8.3 years from 29.4 in 1995 to 37.7 in 2030. Table 2 belowsummarizes some trendsof the demographic development in China. Figures 2 to 5 plotsome indicative demographictrends of China based on the population projectionsof theWorld Bank.8
  15. 15. Table 2BasicIndicators of Demographic DevelopmentOtherLow Middle-China India Income Income WorldEconomies EconomiesAnnual PopulationGrowth1970-1980 % 1.8 2.2 2.5 1.9 1.81980-1993 % 1.4 2.0 2.5 1.7 1.71993-2000 % 0.9 1.8 2.9 1.4 1.5LaborForce: 1993As% of totalpopulation % 60.0 38.0 37.9ParticipationRate % 91.5 67.4 83.9Engagedin Agriculture % 61.0 63.2InfantMortalityRate1970 0/00 69 137 135 74 12721993 0/oo 30 80 89 39 42TotalFertilityRate1970 0/00 5.8 5.5 6.5 4.5 4.81993 0/00 2.0 3.7 5.5 3.0 3.22000 0/00 1.9 3.2 4.9 2.7 2.9CrudeBirthRate1970 0/00 33 39 45 31 321993 0/00 19 29 40 23 25AgeDependencyRatio:1993 52 77 121 81 73UrbanPopulationGrowth1970-1980 % 3.0 3.7 4.2 3.3 2.61980-1993 % 4.3 3.0 4.2 2.8 2.6Source:WorldBank(1995):WorldDevelopmentReport1995:WorkersinanIntegratedWorld9
  16. 16. Figure 2China: Population Ageing45.0 39.020.0 E¶~37.035.0 15.033.02710.0 631.0 M29.0 MlFemale 5.027.0 Ae 5adu25.0 II I I I i Ii 4 -LO o~ to O) LO o~ U Co La o o LO o n Coco Ca 0) o 0 0o _ _ N N c4) uL rt- 0 N Loo) 0) 0) o 0 0 0 0 O 0 0 0 o- o _ -I%-N N1 0I Nli N N N N N N 0 NYear
  17. 17. _~~~~; iA Z, A 8 8 8 8Cs~ ~~~~~~~~198 5 , 0 . , , A,l1985 I I1990 .1 / / > ,9951990 X 31995 gOOOJ ~~~~~~~~1995t2000 20002005 // 200 w 02005 ~~~~~ ~~~~~~~~~~~2005-2010 t201-2010 S n m / c 20152015 -4 2025 a2020 -1 17i 0 .~~~~~~~~CD20sl § /8m205 lllilti(2025 - _ _ _2030 22 050 (is 2050 cm2050- 00---~-2075 a.0I 05 i2100~~~ ~~~~~~~~~C 207--2125 2122150 21/50__ __
  18. 18. Figure 4China: Average Age of Population 1985-215044424038j363432 AeaeA30 Ml-4-Female28YearChina: Youth and Elderly Support Ratio 1985-2150100 -.1Total Dependencyl J 80- __60 -40 /-_Youth: 0-19 years oldt 20 --Elderly: 60 years and older0ff2 n 88 0 0 S S S S 0 0,_N_ _N N N N N N s> N NYear
  19. 19. Figure 5China:Evolutionof AgeStructure1980 70 7 2 1995 70- ;Male _Female___~~~~~------- ..... --_--2030 tC9= _2075 ... 4 __ _ ~~~~~~ ~~ ~~~Malce,leFemaleE__~~ ~~~~~__7.5 5.0 2.5 0.0 2.5 5.0 7.5 7.5 5.0 2.5 0.0 2.5 5.0 7.5
  20. 20. This accelerated aging process has also been heavilyinfluencedby very aggressivebirth control policies. Especiallythe "one couple-one child" policy has been carriedoutsince the late 1970s inresponse to the baby boom from 1960sto early 1970s, andin orderto move rapidly to a balancedsimple reproduction level. Admittedly,this "one-child"policy has been successfulto reduce the number of the new born, as evidencedby a sharpdecline of the total fertilityrate from 5.8 per woman in 1970 to 2.0 in 1993 andof thecrude birth rate from33 in 1970to 19 in 1993. However, it creates irregularwaves ofpopulation cohorts and imposespressures for frequent adjustmentsin education systems,employment settlements, and pension systems for retirees. In the worst case,whenpersons born under the "one-child" policy reach the age of child-bearing,they may enterinto a 1-2-4 structure, i.e., the middle-aged couple have to support their young child aswell as four old grandparents. Unfortunately, the population age structure mayreflectpast irregularities for severalgenerations before these echo effectsdisappear. Trade-offsbetween short-term benefitsand long-term costs have to be taken into considerationwhenpopulation policies are adjusted.Structural Adjustment and Rural-Urban BalanceRural-urban balanceis another importantaspect of the old-agesecuritysystemevolutioninChina. Asinmostcountries,the share of the populationlivinginurbanareasincreasesduringthe courseof industrialdevelopmentand economicgrowth. Drivenprimarilyby youthfulmigrationfromthe countrysideto big cities,urbanizationchangessocio-economicprofilesof the work forceasworkers shiftfrom predominantlyself-sufficientagriculturalengagementto highlymobileindustrialpursuits and services. Oldpersons,especiallyoldwomen,inurbanareasusuallyhaverelativedisadvantagesincompetingwithbettereducatedandmore flexibleyoungerworkers inincomegeneration, andthus tendto havea decreasedparticipationinthe laborforceanda greater need for support. Afterafastgrowthrate of 3.0percentper annumduring1970s,urbanpopulationin Chinaincreasedat anacceleratedrate of4.3 percentperannumduringtheperiod from 1980 to 1993. Thisrate ismorethanthree timesthe overallnationalpopulationgrowth rate of 1.4 percentinthe sameperiodandsignificantly14
  21. 21. higherthan the correspondingworldaverage of 2.6 percent. Consequently,the shareof urbanpopulationinthe nationaltotal rose sharplyfrom 17 percentin 1970to 29 percentin 1993.However, the currentlevelof urbanizationinChinais stillverylow as comparedwiththat of60 percentinmiddle-incomecountriesand78 percent inhigh-incomecountriesin 1993.Urbanizationstimulateschangesinthe familystructure andkinshipnetworkswithbothbeneficialand adverseimpactforthe well-beingof the elderly.These changeshaveparticularimplicationsfor oldage securityarrangementsin China. On one hand,youngfamilymemberslivinginurbanareasmaynot beableto provideinmmediateanddirectcare fortheirparentsandgrandparentswho mayremaininruralareas. The deliveryof servicesto the ruralelderlyisoften logisticallydifficultandcostlywhen socializedinstitutionsandthe basicinfrastructurearenot yetin place. On the otherhand,the youngergenerationinurbanareaswithbetterjobopportunitiesmayhaveenhancedfinancialcapacitiesto support theirparentsandcanaffordtopay for higherqualityservices,suchas oldage healthcare. GivenChinascurrentlowlevelofurbanization,one approachto addressingthisrural-urbanbalancemightbe to use economicincentivesto reducepopulationconcentrationin large citieswith a populationmorethan onemillionbut developsmallandmediumcitieswith a populationbetween 100to 500thousands.The elderlycanthus livetogetherwithor withina short distancefromtheyoungermembersofthe family. Policyeffortshaveto be formnulatedwell inadvancesincethe magnitudeandthespeedwillbeunprecedented inhistory: To absorb a halfof about400 millionincrementalpopulationbetween 1995and2030 and reach anurbanizationlevelof 35 percent,Chinaneedsto buildup around 1,000citieswithan averagesize of 200,000personswithinjust 35 years!Sustainability of Socio-economic TransitionBottlenecksfor establishingawell-functioningold age securitysystemin Chinacanalsopresentthemselvesintermsof socialand politicalsustainabilityof economictransition. Asthe economyexperiencessystemictransformationsandbecomesmoremarketorientedandmoreopento theoutsideworld, it undergoesvarioussocial,political,andinstitutionalchangesas well. Economicgrowth togetherwithpensionreformswillchangethe15
  22. 22. originalpattem of incomedistributionamongdifferentinterestgroups. Peoplesexpectationsanddecisionson consumptionandsavings,on labor supplyandwage demand,andon inter-andintra-generationalincomeredistributionhave to be reshapedto better reflectthe economicand socialrealities. To minimizedisruptionassociatedwith the transitionaldistortions,governmentshaveto formulatestrategiesand policiesto deepenreformsandrnitigatethepotentialsocialcosts of reforms. Legal and regulatory frameworks have to be set up togovem design and managementof pension and social security systems andto specifythesharing of responsibilitiesfor financingretirement pensions among state, enterprises, andindividualemployees. The labormarketneedsto be moreflexibleand labormoremobile.Financialmarkets needtobe developedto stimulateefficientresourceallocation.Incentivesareneeded to ensureadequatenationalsavingsandcapitalaccumulation. Administrativeinterventionsof governmentsinsupportingthe elderlyneedsto be rationalized.Withoutsociallysustainableeconomicgrowth, socialsecurityreformsare unlikelyto besuccessful.Meanwhile,an appropriateoldage support systemcanbe supportiveto adesirablegrowth process, generatinga numberof positiveextemalities. Theseincludestimulatingcapitalformationandlabormobility,andencouragingproductiveinvestments.One ofthe mostpressingissuesinChina at present is to integrate pension reforms with restructuring ofstate-owned enterprises (SOEs). As in many other historically socialistcountries, SOEsinChina directly provide their employeeswith housing, pension, child care, healthcare, andother benefits to cover nearlyall livingneeds throughout their lifetime"from cradletograve". In some of the older enterprises, the pensioner-employee ratio has reached a levelof one-to-one. In a planned regime,profits were taken away and resources were allocatedaccording to plan, hence the pensionburden on individualenterprises was not a seriousconcem. With marketization, individualenterprises have hard budget constraintsimposedand become directly responsiblefor profits and losses. Pension expendituretums out tobe a bigburden too heavy to carry forward for many SOEs. Even when the core businessof an enterprise is profitable,it may show losses because of its large pensionoutflows;andthus its capacity to operate, borrow, and expand suffers. A large numberof pensionershave not been able to get what was promised to themjust because of the accident of their16
  23. 23. being assigned inthe past to an enterprise that is not making enough moneynow. Thissystem, insufficientinsome cases and unfair in other cases, seriously impedeslabormobilityand productivity andleads to delays for enterprise reforms. Bankruptcy of anSOE raises difficultissues of how to honor the pension commitmentsand other socialwelfare obligationsto its retirees. Therefore, it is very important and urgent to formulateand implement a strategy to develop a market-based pension system and a robust socialsafety net, freeing enterprises of direct welfare responsibilitiesand thereby helpingreformthe enterprise system. In order to de-link the enterprise from provision of allretirementincome and to sharethe costs of retirement programs among enterprises, the centralgovernment has recently encouraged local governments to experiment in setting uppension pools. Thisto a certain degree has relieved some enterprises of full directresponsibilityfor their workers retirement pensions since resources and risksare pooledamong enterprises. However, governments need to accelerate their programsto widenpension, unemployment andhealth insurance pools and improve financiallinkagesbetweencontributions and benefits. Meanwhile, the pooled pension schemes need to enlargetheircoverage, especiallyto includeemployeesin collective (township and village)enterprises,foreignjoint ventures, and individualprivate enterprises, who are becomingincreasinglymore numerous innumber and more important in economic contributions.These four areasof concernsare closelyintertwinedwiththe overallissueof howbestto addressthe old ageincomesecurityinChina,when its populationis agingat a remarkablepace. The keypointhereis that a socialsecuritysystemshallnot beviewedas a purelyaccountingor even economicmatter. Pensionreformsneed to take into considerationtheculturalbackground,respecttraditionalvalues,formulatelong-termvisionson demographictrendsand economicgrowth,and adjustmacroeconomicpoliciesto smooththissocialtransition.17
  24. 24. Ill. A SimpleSimulationModelfor PensionReformsTo quantitativelyanalyze policy implication of pension reforms, a mathematicalmodel is a necessary though not sufficient tool. Ideally, a comprehensivedynamicgeneralequilibriummodel might be better to reflect interdependence of economic factors inpolicyreforms. However, such a model would impose more data and resource requirements.This section describesa simplesimulation model, which is composed of four blocksillustrating interrelations among (a) a representative pensionfund, (b) demographic trends,(c) development of the economy and labor market, and (d) government behaviorunderdifferent pension arrangements. Figure 6 below presents key inter-linksamong thesefactors.Figure 6 Links Among Pension Performance and Socio-Economic DevelopmentGovemrment Culture and~~A ~Behavior TraditionDemographic > Pension _ EconomicDynamics 7n nanceDevelopment | LaborMarket | PensionConditions ArrangementsLike the RMSM-X model used in the World Bank for mediumterm projectionsand policy consistencycheck, the model proposed here is useful to examinethe financialsustainabilityof pensionfunds and to show impacts of external changes and policyalteration. The model, however, has not explicitlyendogenalizedlinkagesbetween thepension accounts and the affectingfactors, such as savingsthrough pensionschemes and18
  25. 25. GDP growth, andbetween labor supply and wage-pernsionsetting. In addition,thetreatment of uncertaintyis done indirectly, though evaluatingthe effectivenessof pensionreform options encounters numerous uncertainties. To a certain extent, both demographicdynamics and economicdevelopment are stochastic processes affectedby manyunpredictable factors. With some information on prior probability distributionsof theserandom factors, the basic structure of the model can be used to generate bounds ofconfidence intervals,such as in the scenario analysisthrough defining"the best" and "theworst" possiblesituations. Another approach of addressingthese uncertaintiesis throughstochastic numeric simulations(e.g. using Monte Cairo techniques).Political economy of pension reforms is also difficultto model. Sincepensionreforms necessarilyredistribute income among differentgenerations and among differentincome groups withinone generation. Pareto optimalityis difficultto achieveduringthereform period. These mayhave serious implications for social stabilityand longer-termsustainableincomegrowthThePension Fund BalanceDuring the planningperiod t E (to, to+1, to+2,.., to+T}, the balanceof arepresentative pensionfund in a specific region at time t, denoted by B,, canbe written asB, = (1+r,)B,, +I, - O, - AE,, for givenB,1lwhere the increment is the differencebetween the current income (as the sum of theinvestment returns on the previous balance B,_1 at rate r, and the contributioninflowIt) andthe current expenditure(as the sum of total pension payments 0, and the administrativeexpenses AE,). Thisincrement can be negative when the income cannot fullycover theexpenditure, and thus the fund decumulates. Let se {M, F) be sex indexfor male andfemale, respectively. GivenA1 for the average first timeworking age andA2 for theretirement age, the total contribution to the fund can be calculatedas19
  26. 26. I,= z ZC,Wa,N,S5srM.F} a-AIwhere c, is the pension-related payroll tax rate, W- is the nominalaveragewage forworkers at age a andN,-5is the number of workers covered by the pensionscheme. Thesummations are conducted over both age and sex. Meanwhile the total pensionpaymentscan be calculated asAs0,= NsE(M.F) a=A -+lwhere A3 is the maximumage and H, is a basis for pension calculation,which can belinked to the pensioners personal wage history, past contributionsto the fund, the wageprofile of current workers, inflation,or some other social or policyfactors. Thereplacement rate, Ot,specifiesthe benefit a pensioner willreceive as a portion of thepension basis. Thus, in a functionalform,Ha,s = a` 3 (w, i,)where w and i are wageand inflation matrices. In some cases, H can be zero if the personis not eligible for pensionbenefits. The administrative cost is affected by manyfactors,includingthe fund balance(portfolio management, cash flows, and some fixedcost. Forsimplicity,, the administrativecost in our simulation is taken as a portion of the average ofthe total income andtotal payments of the fund balance, i.e.,AEt =ei +2where e,captures the cost effectivenessof the fund management.20
  27. 27. DemographicDynamicsandLaborMarketAt time t, the magnitude and age structure of the population canbe presented as asimple first order Markov chain,pa,s = (I -ma5)Ia1 +NI 3 , foraE int(O,A35]where Pt"is the populationof sex s at age a, mt, is the mortality rate, andNIL is the netimmigrants into the region.49pOs = ( o_ ) a,sarrapa=20where rn,° is the infantmortality rate for the first year after birth, andf, is the fertilityrateof a woman at age a, whichdepends critically on the population policy, socialcustom, andcurrent and expected economicconditions including the cost of raisinga child. The old-age dependence ratio, d, is defined{M,F) a= A2d =A4. .-IsEJM.F} a=AIThe size of the laborforce can be written asLa,s = a,spa.:where I,-is the labor participationrate across ages and sexes. Here, the participationrates for the young andthe old are set to zero, namely, I = I1f=0 for a E int[O,Al) andacint[A2 , A3]. The total employmentE, is the product of the labor force andtheemployment rate, 1-u,", namely,E` = (1- U)La 5 for as int[Aj, A2)21
  28. 28. The pension schemecoverage can be presented byI{a.sEa.s for A, a<Alt(E-m)NL"j", for AX< a<Awhere ct" is the contributioncoverage ratio.Economic and WageGrowthThe economy is assumedto follow a growth pattern, which is affectedby a numberof factors includingpopulationgrowth, labor productivity improvement,technological andmanagement innovations,and govermnent policies. For simplicity,we onlyemphasizethelinkage between the increaseof real wage, w,a.,with the growth of per capita output, gt,i.e.,where w( ) is a functionalform and the real wage in the base year is setto be 1.For givenprice levelinthe base year P0 and a path of inflationrates, Tr;the nominalwagethen can be calculated asWa,5= Wa,JP n lr=GovernmentBehaviorunderDifferentPensionArrangementsThe governmentmay have multiple objectives in pension designand reforms. Thesimplest approach is to assumethat the government attempts to maximizea well-definedsocial welfare function,whichtakes into consideration (indirect)utilityfunctionsof allindividualsin the society. Hence generally, the govermnents behavior shallbe derivedfrom the followingmaximizationframework22
  29. 29. Max SWF = SWF(Vi, jEP)subject to all the relevant constraints. Here P is the union of all categories of people inthesociety and Vi is the expected life-timeindirect utility of an individualj, whichmainlydepends on the individualsincome stream (wages when he/she works and pensionbenefitafter he/she retires), variationsof price levels, his/her own time preference parameters, andprobability distributions on future uncertainties. Namely,Vi = E{v[(Wt, OtH); x,]}Policy instruments availableto the government vary with differentkinds of pensionarrangements. Four out of many commonlyadopted pension schemes, namelyaPay-as-you-go scheme, a definedbenefitscheme, a defined contribution scheme,and a privatefund-type scheme, are discussed here. Through some aggregation, policyimplicationsofreforming to a multi-pillarsystemas advocated by the World Bank can also be simulated.For the purpose of policy simulationshere, uniformity is assumed across individualssothat the whole society canbe treated as a unit. When the policy focus turns from thesolvency of pensionfiundsto incomeredistribution, the model needs considerablemodification.Pay-As-You-Go SchemeStrictly speaking,this schemerequires that the pension fund be balancedin eachand every single year during the whole planning period. In practice, pensionsare usuallyindexed to wages and financedby a pay-roll tax. In this case, B, = 0 for all t, i.e.,I, = O, + AE,23
  30. 30. where the total incomeof the pension scheme cover exactly the sum of the total pensionpayments and the administrativecost at each period t. Given the demographicdynamics,the pension-wage linkage, and the administrative efficiency,the behavior of thegovernment can be derived as3Asl+e E, ,H, Na,,* ~~2JEJM,F) a=A2+lMin c = 2 A2 for all t,Mm1 1-2 E Wa;yaCC 2 Wr~~z~a.sNt.sE{CM.F) a=AIwhere [cm , c] is the feasibleregion for the contribution rate. Under the Pay-As-You-Go system,there willbe no explicit or implicitdebts generated from pensionfundinsolvency accumulatingthrough time. But the required minimumcontribution ratechanges from year to year, and could increase to a level sociallyunbearablewhen thepopulation ages.Defined Contribution SystemWithinthe definedcontributionsystem,the contributionrate isusuallyfixedat acertainlevelc=ct. The governmentcanchooseto set a uniformreplacementrate (00,) to the highestpossiblewhilethe pensionfund is maintainedwithpositivebalancesduringthe planningperiod.ThistranslatesintoMax {(}0[Omi,nmax]where [e"", O] is the feasible region for the benefit rate. This maximizationis subject toBt> 0 and that all other relations specifiedabove hold. The disadvantageof this kindofpension system is that when the ratio between the numbers of the pensionersand workersbecomes larger over time, the benefitreceived by the pensioners will declinesignificantly.24
  31. 31. Defined Benefit SystemA defined benefitpension system specifies the pensionpath a retiree willreceiveduring the period after he or she retires. The government minimizesthe requiredcontribution rate, whichcan be assumed uniform through time for simplicity,becausedoing so increases the disposable income of the workers without scarifyingpensionerswelfare. Therefore, policy simulationscan be based on all the accounting and economicrelations specifiedabove and the following optimization framework:Min {C}, subject to B, 2 0,[c; ]where [c , c] is the feasibleregion for the contribution rate. The potential problemwith this systemis that even with a very moderate replacement rate, the burden for thecontributing workers could be too heavy as the population ages.Private Pension Fund Accounts SystemOne critical argumentfor publiclymanaged social security systemis rooted inpaternalistic concerns-the beliefthat many individuals,if left to themselves,could bemyopic in their consumption-savingsbehavior and thus would not save enoughto sustaintheir living standards after retirement. Therefore, everyone in the society needsto bemandated to save at least a minimumportion of his or herincome before the retirementage. A private account can be established to record the savings. Within thissystemaworker pays a certain percentage of his wage to the fund when he works and receiveapension after he retires according to the amount of his total savings and investmentreturns.Let c be the portion of wage a worker sets aside for his retirement, then his lifetime savingsvalues at the time of retirement, A2, willbeA2AAl= (CW-AEr)AI (1 + r,25t=A, T=I25
  32. 32. where AE, is the personal transaction cost of maintainingthe account and r, is the interestrate. If the possibilityof receiving and giving bequest is ignored, namely,BA =0and the pension he derives from his personal fund is an annuity fullyindexedwith theinflation,thenB, =(1+r,)B,l -o07(1+ Jrr) fort E int[A2+1 A31].r=A21+lwhere 0 is the real annuityvalued at time of retirement. The workers problem is then tofind an appropriate c for a given 0 or to derive an acceptable 0 for a givenc. Thegovernment can affectindividualworkers choice through its macroeconomicpolicieswhich exercisedirect impact on wage levels, interest rates, real and expectedlevels ofinflation, andthe transaction cost of maintainingprivate pension funds.A Multi-pillar Pension SystemRecentworkby theWorld Bank(1994) suggeststhat a combinationof multiplepillarsof oldage incomesupport maybe most suitable. There shouldbe someminimumpensionguaranteedby the stateto allpensioners. Thiswould then becomplementedby a secondlegwhichwould allowfor each personto contributeto an individualaccount. Theideais that theformerwouldhelpto protectthe poorer membersof societybut alsothosewho arenot ableorchoosenot to savefor retirement. The latterwould allowfor each personto savewithappropriatesafeguardsandachievecertain.ivingstandardsfortheir old age. In orderto meetthese needs,variousindividualpensionoptionswarrant carefulconsiderationbut it is alsoimportantto weighthe overallimplications.This mightbe doneinthe contextofthe fouroptionsnotedabove.26
  33. 33. IV. PensionSystemReformin ChinaThis section turns to analyze policy options for pension systemreforms inChina.First, some aspects of the evolution of the Chinese pension systems since 1978 arereviewed. This helpsto calibrate parameters essential for simulatingpolicy reforms basedon the simulationmodelpresented inthe last section. A benchmarkcase is then workedout to show that the current pension arrangements are unlikelyto remainfinanciallysustainablefor more thanr-frty years if no reforms are undertaken. Then some policysimulationsare conducted to numericallyillustrate implicationsof differentreform options.The Current Pension System in ChinaOut of a total labor force of more than 600 million,existingpension schemescoverabout 150 million,mostlyemployeesof state enterprises,governmentstaff,plusa portionofurban collectives.Theseschemes,manyof which arenot financiallyhealthyat present,leaveuncovered about75 percentoflabor force, mainlyrural peasants,urbancontractworkers, andemployeesof individualfirms. Permanent workers pensions are publiclymanagedandbasically financedby payroll taxes on a pay-as-you-go basis. Enterprises share a part ofthe financialburden through contributing a portion of permanent employees current wageto the pensionpool. Retirement age is relatively young, 60 for men and 55 for women.Many workers retire early for health reasons or disability,in most cases receivingfullbenefits. Contributionrates vary across regions and sectors and may applyto differentbases. Financialand actuariallinkages between contributions andbenefits are not alwaysclear. Pension funds are usually deposited in local bank accounts that are often less-than-fullyindexed with inflation. Local interest in keeping financialresources within thelocalitiesdo not alwayslead to investments with highest possiblereturns. Over time,theaging of Chinas populationmeans that pension expenditures as currentlystructured willincrease dramaticallyfrom 12.7% of the total wages in 1987 to almost 20% by 1997, to25% by 2010 andto over 48% by 2030. Table3 belowpresentssomestatisticsthatquantitativelysummarizedevelopmentof Chinas formalpensionsystemsbetween 1978and1993.27
  34. 34. Table 3China: Pension and Wage Statistics 1978-1993Pensioners Employees Employee Average Pension Average Wa e Pension Wage Ratio GNP CPINumber Growth Number Growth Pensioner NominalI Real 78 Growth Nominal I Real 78 Growth Average Total Growth Changemillion % million % Ratio Yuan % Yaun % __ %1978 3.1 95.0 30.3 551.0 551.0 615.0 615.0 89.6 3.0 11.7 0.71979 6.0 89.8 99.7 4.9 16.7 714.0 700.0 27.0 668.0 654.9 6.5 106.9 6.4 7.6 2.01980 8.2 36.9 104.4 4.8 12.8 714.0 660.4 (5.7) 762.0 704.8 7.6 93.7 7.3 7.9 6.01981 9.5 16.4 109.4 4.7 11.5 706.0 637.7 (3.4) 772.0 697.3 (1.1) 91.5 7.9 4.4 2.41982 11.1 17.2 112.8 3.1 10.1 709.0 628.4 (1.4) 798.0 707.3 1.4 88.8 8.8 8.8 1.91983 12.9 16.1 115.2 2.1 8.9 726.0 634.0 0.9 826.0 721.3 2.0 87.9 9.9 10.4 1.51984 14.8 14.4 118.9 3.3 8.0 766.0 650.7 2.6 974.0 827.4 14.7 78.6 9.8 14.7 2.81985 16.4 10.8 123.6 3.9 7.5 935.0 730.0 12.2 1,148.0 896.3 8.3 81.4 10.8 12.8 8.81986 18.1 10.3 128.1 3.6 7.1 983.0 724.1 (0.8) 1,329.0 978.9 9.2 74.0 10.4 8.1 6.01987 19.7 9.0 132.1 3.2 6.7 1083.0 743.5 2.7 1,459.0 1,001.6 2.3 74.2 11.1 10.9 7.31988 21.2 7.7 136.1 3.0 6.4 1322.0 765.8 3.0 1,747.0 1,012.0 1.0 75.7 11.8 11.0 18.51989 22.0 3.8 137.4 1.0 6.2 1450.0 713.1 (6.9) 1,935.0 951.6 (6.0) 74.9 12.0 4.0 17.81990 23.0 4.5 140.6 2.3 6.1 1726.0 831.3 16.6 2,140.0 1,030.7 8.3 80.7 13.2 5.8 2.11991 24.3 5.7 145.1 3.2 6.0 1936.0 906.2 9.0 2,340.0 1,095.3 6.3 82.7 13.9 6.1 2.91992 26.0 6.8 147.9 2.0 5.7 2260.0 1003.7 10.8 2,711.0 1,204.0 9.9 83.4 14.6 13.6 5.41993 27.8 7.0 148.5 0.4 5.3 2779.0 1090.2 8.6 3,371.0 1,322.5 9.8 82.4 15.4 13.4 13.2Source: Chinese Statistics Press -- Yearbook of Labor Statistics of China 1994
  35. 35. During thesesixteenyearsbetween 1978 and 1993,pension-coveredemploymentgrewby 56.3 percentwhilethe retireessupportedby the pensionsystemsincreasedby closeto eighttimes,leadingthe employee-pensionerratio decliningdramatically from30.3 in 1978to only5.3 in 1993. Thisratio willcontinueto declinewhen morepeoplewho joinedthelaborforcein1950sand early 1960sreachthe retirementage. Whilethe pension-wageratio on averageremainedrelativelystablearound80 percent inrecentyears,the total pensionexpenditureincreasedremarkablyfrom3.0 percentof the total wage billin 1978to 10.8percentin 1985,andthen to 15.4 percentin 1993. Thisoccurred because the pensionschemewasof a defined-benefittype and thenumberof pensionersincreasedsharplyfrom3.1 millionin 1978to 16.4millionin 1985,andthento 27.8 millionin 1993. People outsideofthe publicpensioncoveragerelymainlyon privatesavingsandfamilysupports for their oldageincomesecurity.This explains,at leastpartially,whyChinas gross domesticsavings(GDS) rate canbe as highas 35 to 40 percent.Real growth of GNPper capita,averagewage, andaveragepensionwerenot closelycorrelated. In 1988and 1989,when GNP grew by 11.0percentand 4.0 percent,respectively,the realaveragewage roseby only 1.0percent in 1988andactuallydeclinedby 6.0percentin1989. High inflation- 18.5percentin 1988and 17.8percentin 1989- completelyerodedmoderategainsinnomninalwage. Correspondingly,the real averagepensionrose slightlyby3.0 percentin 1988anddroppedby 6.9 percent in 1989. Thismaybe oneof the reasonsleadingto socialunrestinmanyregionsin 1989. To investigatehow realwagegrowthislinkedto GDP growth andinflationrate, we run a regressionusing data presentedinTable3andget the followingresult: = 0.8786gt- 0.4060irt, R2= 0.525,F = 7.183WIl (5.94) (-2.383)where the numbers inbrackets are t-statistics, showing that the regressionis significantatlevela=5%. Hence, the real wage gains about 0.88 percent if GDP grows by one morepercent, while an increase ininflationof one percent is associated with a 0.4 percenterosion of the workers gain. Figure 7 shows historical paths of GDP growth, pensionandwage movements, and changes innumbers of contributors and pensioners,respectively.29
  36. 36. Numberof Pnsioners in millions InGrowth Ratesin -wo o > oo1978- v o *E197S - -l i1979- t ___^ 1979 -1981 1980 .198201983 Z~~~~~~~~~~~~19840 1984 j199384o t: w1985 01985 CDI~~~~~~~19861I986CrQ 1987119871988~~~~~~~~~~~~~~~~~~~~181989 -1989- El 0I-~~~~~~~~ 0 -~~~~~~~~99199019019911992 ~~~~~~~~~~~~~~~~~~~~~1992-19931993 0 A - -NINumber of Emnployeesper 100pmsaasers
  37. 37. TheNeed forPensionReforms:TheBenchmarkCaseThe model specifiedabove is utilized to simulate policy sensitivitiesof reformingpension systems in China for 80 years between 1995 and 2075. The basic demographicinformation is extracted from the World Bank Population Projections (WorldBank,1995). Key parameters are calibratedusing historical data published by Chineseauthroties. Essential assumptionsare spelt out based on reasonablejudgment of Chineseeconomic outlook in the next century. For simplicity,many indicators are kept uniformthrough time, but these can be changed as deemed appropriate.Base Year ParametersIn the base year 1995,the initialbalance of all pension funds in urban areas, whichare treated as a singlefund hereafter, remains in the neighborhood of 100billionYuan or1.8 percent of the total GDP. The average annual wages for a male anda femaleworkerin 1995 are estimated to be around five thousand Yuan. A simplereplacementratiobetween the average pension and the average wage is used, which stands at 75 percent in1995. Thus the averagepensions for a male, who retires at age of 60, and a femalepensioner, who retires at 55, in 1995 are estimated to be around 3.75 thousand Yuan.Key AssumptionsTable 4 summarizessome key assumptions for the simulation exercises.RealGDP is assumed to grow at an average rate of 10 percent per annum until 2000 and thenat a lower rate of 7.5 percent between 2001 and 2049. Since2050, GDP growth stablizesat 5% per annum. The annual inflationrate declines from 10 percent during 1996-2000to7.5 percent between 2001 to 2049 and then to 5 percent starting from 2050. The growthof real wage depends on the GDP growth and inflation path, accordingto the formulaspecifiedinthe regression equation presented in the last section. Nominal wages are fullyindexed to inflation31
  38. 38. Table 4China:Major Assumptionsfor PublicPensionSystem Analysis(The Base Case)1995 1996 1997 1998 1999 2000 2010 2025 2050 2075GDP Growth (%) 10 10 10 10 10 10 7.5 7.5 5 5Inflation Rate (X) 10 10 10 10 10 10 5 5 3.5 3.5RealWageGrowth (%) 4.7 4.7 4.7 4.7 4.7 4.7 4.6 4.6 3.0 3.0ContributionRate(X) 27 27 27 27 27 27 27 27 27 27ReplacementRate(X) 75 75 75 75 75 75 75 75 75 75AdministrativeCost Ratio(X) 3 3 3 3 3 3 3 3 3 3RealRateof Retumson FundReserves 2 2 2 2 2 2 2 2 2 2PriceIndex I 1.10 1.21 1.33 1.46 1.61 2.62 5.45 13.85 32.73RealWage(Yuan 000) 5.0 5.2 5.5 5.7 6.0 6.3 9.8 19.2 43.1 89.6NominalWage(Yuan000) 5.0 5.8 6.6 7.6 8.8 10.1 25.8 104.7 597.0 2,934.0RealGDP (Y. trillion) 5.5 6.1 6.7 7.3 8.1 8.9 18.3 54.0 205.8 696.8NominalGDP (Y. trillion) 5.5 6.7 8.1 9.7 11.8 14.3 47.9 294.6 2,849.8 22,806.4
  39. 39. During the next 80 years, the replacement rate is averaged at 75 percent, while thecontribution rate is averaged 27 percent under the selected pension arrangements.Theadministrative cost is assumed to be 3 percent of the average of pensionfund receipts andpayments. The real rate of returns from investment of the pensionfund balanceisassumed to be 2 percent across time.The urbanizationspeedis taken to be 0.3 percentper annum,leadingthe urbanpopulation as a sharein the total national population increasinggraduallyfrom 29.6percent in 1995to 38.4 percent in2030 and to 50 percent in 2075. Thiscompares with 54percent inlow-middle-incomeeconomies on average in 1993. The urban laborparticipation rate averages 90 percent, and the unemploymentrate is kept around 3percent.Pressure fromPopulation AgingUnder these assumptionsfor the benchmark case, the pressure frompopulationaging is apparent. Table5 below presents a projected population structure and pensioncoverage for Chinafrom 1995 to 2075. The population dependencyrate, in terms of thenumber of persons older than the retirement age per 100 working age persons, increasesfrom about 21 in 1995to 44 around 2025 and then rises again to nearly70 in 2075. Thesystem dependencyratio, in terms of the number of pensioners per 100 contributors,decreases initiallyfrom 21 in 1995 to 15 around 2000 when the pensioncoverage rate forthe working age populationextends from 25 percent to 38.5 percent, as projected. Thisimpliesmore contributors willjoin the pension pool inthe near future. However, thissystem dependencyrate quickly rebounds to 35.1 in 2025 and approachesthe populationdependence rate in2075, as the population structure and the pensionsystemmature. If theyoung dependents are also taken into consideration, the total dependencyrate will be inthe neighborhoodof 100 around 2025.33
  40. 40. Table5China:Projectionof PopulationStructureand PublicPensionSystem Coverage(The BaseCase)1995 1996 1997 1998 1999 2000 2010 2025 2050 2075Population (Thousand) 1,218,134 1,230,499 1,244,183 1,259,230 1,275,670 1,293,548 1,420,314 1,597,166 1,763,793 1,821,789Young (below age 20) 424,807 427,741 430,939 434,418 438,185 442,255 444,272 436,160 436,950 438,226Working 656,096 662,532 669,943 678,345 687,760 698,218 766,834 807,431 810,387 814,603SVstem Covered 164,025 186,701 208,267 228,990 249,105 268,813 416,853 507,362 527,935 532,052Not Covered 492,071 475,831 461,676 449,355 438,655 429,404 349,981 300,069 282,453 282,551Old 137,231 140,224 143,301 146,467 149,725 153,076 209,208 353,575 516,455 568,960System Covered 34,308 35,057 36,092 37,388 38,924 40,678 75,676 178,327 312,402 360,693Not Covered 102,923 105,169 107,209 109,079 110,801 112,398 133,532 175,248 204,053 208,267An Percentage of Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0Young 34.9 34.8 34.6 34.5 34.3 34.2 31.3 27.3 24.8 24.1Working 53.9 53.8 53.8 53.9 53.9 54.0 54.0 50.6 45.9 44.7Covered 13.5 15.2 16.7 18.2 19.5 20.8 29.3 31.8 29.9 29.2Not Covered 40.4 38.7 37.1 35.7 34.4 33.2 24.6 18.8 16.0 15.5Old 11.3 11.4 11.5 11.6 11.7 11.8 14.7 22.1 29.3 31.2Covered 2.8 2.8 2.9 3.0 3.1 3.1 5.3 11.2 17.7 19.8Not Covered 8.4 8.5 8.6 8.7 8.7 8.7 9.4 11.0 11.6 11.4System Coverage Rate 1%1Young 25.0 28.2 31.1 33.8 36.2 38.5 54.4 62.8 65.1 65.3Old 25.0 25.0 25.2 25.5 26.0 26.6 36.2 50.4 60.5 63.4Dependency Rates 1%)Population 20.9 21.2 21.4 21.6 21.8 21.9 27.3 43.8 63.7 69.8System 20.9 18.8 17.3 16.3 15.6 15.1 18.2 35.1 59.2 67.8Young 64.7 64.6 64.3 64.0 63.7 63.3 57.9 54.0 53.9 53.8Total 85.7 85.7 85.7 85.6 85.5 85.3 85.2 97.8 117.6 123.6Source: Word Bank Population ProjectionS and authorS estimates.
  41. 41. The Base Case ScenarioIn the current pay-as-you-go pension system continues in to futurewithout anyreform measures beingundertaken, Chinas pension systemwill not be financiallysustainable. Table 6 below presents simulation results, with four indicatorssuggeststheinsolvency of the simulatedpensionfund. First, the funds current balanceturns negativein 2025, meaningthat total receipts of the fund are not sufficientto coverthe duepayments and administrativecost. Second, the accumulated balance of the fund,whichmay peak at more than 30 percent of current GDP around 2010 when babyboomers arethe backbone of the working force and the pension coverage expands quickly,becomesnegative before 2050. Third, on a counter-factual basis, the pension cost rate is defined asthe contribution rate required to obtain zero current balance for a giventotal pensionpayment. This cost rate in Chinarises from 12 percent in 2000 to 28 percent in2025 andthen to 54 percent in2075, doublingthe assumed contribution rate. A pensioncontribution rate higherthan 30 percent is not likelyto be sociallyacceptable. Finally,theaffordable replacementrate can maintain above the assumed benefit rate until2025. Afterthat the affordable benefitrate drops to 32.7 percent, about a half the assumedbenefitrate.Clearly, financialsustainabilityis only one out of manydemands for reformingthepension system inChina. However, the strength of this demand is remarkablystrong.Since pension policiesimpactover a long period, small incrementalchangeson annualbasis can accumulate into enormous effects across time. As shown inTable6, ifthepension systemwere not modified,the negative accumulative balance of the pensionfundin 2075 can reach an unbearablelevelof 41 percent of current GDP. Or alternatively,China can only affordto support half of its pensioners, who infact havecontributedto thepension fund throughout their working life.35
  42. 42. Table 6China:Pension Fund Accounts under Pay-As-You-GoScheme(The Base Case)_ _ _ _ _ _ _1995 1996 1997 1998 1999 2000 2010 2025 2050 2075With Reserves (Y. Billion)Revenue Collection 221.4 290.4 373.1 472.6 592.2 736.2 2,904.6 14,345.2 85,098.9 421,487.9Per-,:an Payments 128.7 151.4 179.6 214.3 257.1 309.5 1,464.7 14,005.7 139,879.8 793,718.1Administrative Cost 5.3 6.6 8.3 10.3 12.7 15.7 65.5 425.3 3,374.7 18,228.1Current Balance 88 132 185 248 322 411 1,374 (86) (58,156) (390,458)As % of GDP 1.6 2.0 2.3 2.5 2.7 2.9 2.9 (0.0) (2.0) (1.7)Investment Returns 12 29 55 91 141 895 5,355 (20,208) (465,750)Accumulative Balance 100 244 459 762 1,176 1,728 15,059 81,763 (445,787) (9,324,392)As % of GDP 1.8 3.7 5.7 7.8 10.0 12.1 31.4 27.8 (15.6) (40.9)NPV 100 222 379 572 803 1,073 5,740 14,992 (32,187) (284,880)With Zero Current BalanceCost Rate (%) 16.17 14.95 13.80 13.00 12.44 12.05 14.46 27.99 47.13 53.99Affordable Benefit Rate (%) 125.3 139.5 151.2 160.5 167.7 173.2 144.3 74.5 44.3 38.7Supportable Coverage Rate for the Old 41.8 46.5 50.8 54.6 58.1 61.4 69.6 50.1 35.7 32.7Needed Coverage Rate for the Working Age 15.0 15.1 15.4 15.8 16.2 16.7 28.2 63.2 110.3 126.7No. of Supportable Pensioners 57,304 65,226 72,760 80,000 87,027 93,912 145,632 177,252 184,439 185,877No. of Needed Contributors 98,203 100,346 103,309 107,019 111,415 116,436 216,613 510,440 894,213 1,032,440An Compared with Benchmark In %Contribution Rate 59.9 55.4 51.1 48.2 46.1 44.6 53.5 103.7 174.5 200.0Benefit Rate 167.0 186.1 201.6 214.0 223.6 230.9 192.4 99.4 59.0 51.5In 1995 YuanReal Wage 5.00 5.24 5.48 5.74 6.01 6.30 9.84 19.20 43.11 89.64Real Pension 3.75 3.93 4.11 4.31 4.51 4.72 7.38 14.40 32.33 67.23As % of 1995Real Wage 100.00 104.73 109.68 114.86 120.29 125.97 196.75 384.03 862.11 1,792.83Real Pension 100.00 104.73 109.68 114.86 120.29 125.97 196.75 384.03 862.11 1,792.83
  43. 43. Policy Simulationsfor Reform OptionsApparently,pensionrefonnscannotbeavoidedinChina.Theissuesturntowhattodoandwhen.Thissubsectionsimulatesimplicationsofvariousreformoptionsforpensionsystems.Withthe simulationmodel,policiesinfourbroaderareascanbesimulated:(a) macroeconomicdevelopment,suchas realGDPgrowth,inflation,andunemployment;(b) labormarketcondition,includinglaborparticipation,realwagegrowth,andchangeof retirementage;(c) demographicdynamics,mainlychangesof mortalityrates,migration,urbanization,andlifeexpectancy;and(d) pensionfundmanagement,suchas effectivecoveragerate, contributionrate,compliancerate,eligibilityconditions,changesofbenefitsinnominalandrealterms,realreturnsfrominvestmentof fundsurplus,administrationcosts,adjustmentoflinkagebetweenrealwageandrealpension,andinflationindexingmechanism.Out of manypossiblecombinationsof policyoptions,the impactof threechangesare illustrated.Thefirstsimulationshowsthe effectsof delayingretirementage. Thesecondshowsthe possibleconsequencesof differentGDPgrowthpaths. Thelastsimulationcomparestrade-offsbetweenraisingcontributionratesandlowingbenefitrates.DelayingRetirement AgeThemost directimpactof delayingthe retirementage isto changethe systemdependencyrate of thepensionsystemthroughenlargingthe laborparticipationrateforoldpeople. Theimpactcanbe dramatic,sincethe effectisdoubledwhenthenumberofpensionersisreducedwhilethe numberof contributorsis increased.Thusoldageproblemis convertedintothe (un)employmentproblemwithpressureson realwagegrowth. Withoutendogenalizingthe relationshipsamongGDPgrowth,labordemandandsupply,andrealwagechanges,it isdifficultto predictthe potentialoutcome.Here,onlypopulationdependencerates areplottedinFigure8 forthreesetsof retirementage.37
  44. 44. Figure 8Population DependencyRate With Different Sets Retirement Age70.060.0 M60F55-4-M60 F6050.0 -- |BaeM65F60 |BaseC40.0 Raising the retirement age30. - by 5 years for female30.020.0aRaising the retirement10.0 age by 5 years forboth male and female_ I II ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~II I I004O 0 00 0mc oC o 0g O m- -4 - s C~4 es s ese
  45. 45. Thekeypolicyissuehereishowthe societyas a wholeabsorbsthissegmentof oldage labor. ForChina,the unemploymentand under-employmentintheurbanareasarebecomingmoreapparent.Transitioninagricultureis releasinga largeamountof surpluslabor,closeto 100millioninrecentyears. Thegovernmentneedstoweighthecostagainstthe benefitof dealingwitheitherthe pensionproblemor the unemploymentproblem. Thelattermaybe morepoliticallysensitiveandmoreuncertaininfinancialterms. Withoutexcessdemandinthe labormarket,it wouldnot bea desirableoptiontoincreasethe retirementage.Change of GDP Growth PathPromotingGDPgrowthenlargesthe pie to be sharedbythe oldandthe young.Withoutlongrunstableincomegrowth,neitherpensionersnorcontributorscanexpecttoimprovetheirlivingstandardssteadily.Better growthperformancemeanshigherrealwage,leadsto largercontributionrevenueforthe pensionfund,andleavessomeleewayfor adjustmentofbenefitrate. Table7 comparesthree caseswiththebasecase,fordifferentGDPgrowthpathsandpension-wageindexation.Thefirstthingtonoticehereisthat, as showninthe lastlineof Table7, one percentagepointchange(fasteror slower)inthe realGDPgrowthrate candoubleor halfthe realwageandpension,as comparedtothat inthe basecasescenario. Themagnitudeof thiseffectdependson thelengthofpensionplanningperiod(80yearshere). Even allowingthe pensionersbenefitratetoshrinkovertime,asshowninthe lastcolumnof Table7, theirwelfaremeasuredinrealpensionincomeissignificantlyimprovedwhenthe GDPgrowsfasterthaninthe basecase.Secondly,it isnoticeablethatevena combinationof fastergrowthandlessthanfullwageindexationwillnotnecessarilymakesthe pensionfundfinanciallysolvent.Highergrowthleadsto higherwage,and thushigherpension. However,thiscombinationdelaysthe firstyear forthe currentandaccumulativebalanceof the pensionfundto benegativeby4yearsand9 years,respectively.Finally,it is worthwhileto lookat thebreak-evencontributionandbenefitratesof thepensionfunds,whichareassumedto beuniformacrossyearsandtoberequiredto achievea zero accumulativebalanceatthe endofthe39
  46. 46. Table7China: Impact of GDP Growth on Pension Fund SolvencyBase GDP GrowthCase Slowerby 1% Fasterby 1%WageIndexation 100% 100% 100% 90%Break-evenContributionRate 39.6 37.5 41.5 31.8As % of the Assumed 146.8 138.9 153.7 117.8Break-evenReplacementRate 51.1 54.0 48.7 63.7As % of the Assumed 68.1 72.0 65.0 84.9The First YearforNegativeCurrentBalance 2025 2025 2025 2029NegativeFund Reserves 2041 2043 2039 2052Fund Reserveas % of CurrentGDP2030 18.1 22.7 14.4 21.82050 -15.6 -13.7 -16.3 12075 -40.9 -47.9 -34.5 -11.4Real Wage/Pensionas % of 1995 RealPension2030 479.9 357.2 643.2 536.32050 646.6 546.2 1,369.4 1060.12075 1,792.8 907.5 3,521.7 2480.4
  47. 47. planning period. One percent faster GDP growth plus 90 percent wage indexationreduces the break-even contributionrate from 40 to 32 percent, and increasesthe break-even replacement rate from 51 to 64 percent.Changing the ContributionRate and the Replacement RateAnother commonlyrecommended policy option is to raise the contributionrate orto reduce the replacementrate, or both. While knowing that this option is politicallydifficult,we conduct a numericalsimulation to examine the financialimplicationof thepolicy recommendation. The outcome is presented as a sensitivitytrade-off table, shownin Table 8. Starting fromthe base case point where the contribution rate is 27 percentand the replacement rate is 75 percent. Raising the contribution rate to 35 percent orreducing the replacementrate to 60 percent alone will reduce the negative pensionbalancelevelas percentage of current GDP at the end of 2075 from 41 percent to around 15percent. Moving towards the lower left corner is the direction for pension solvency,butobjections from both pensioners(for a lower replacement rate) and the contributors (for ahigher contribution rate). To achievea near-zero balance at the end of the planningperiod, one needs to set the contribution rate at 32 percent and the replacementrate at 60percent, or the contributionrate at 33 percent and the replacement rate at 62.5 percent.Fully Funded Individual AccountsThe individualpension accounts specified in the model are fullyfundedbydefinition. They are self-balancedifthe uncertainties on investmentreturns and lifeexpectancy are carefullyhandled, through pooling, for instance. The relevantpolicy issueshere are the positive linkagebetween the contribution rate and the benefitrate. Thegovernment needs to make sure most people in the society save sufficientfor their ownold age expenditure. For China,the private savings rate is usually around 35 percent.Therefore, establishingpersonal notionalpension accounts can be a feasibleand effectivereform option in China.41
  48. 48. Table8Sensitivityof the Pension Fund Balancewith respect to Changes of Contribution Rate and ReplacementRate(in % of Current GDP as of the end of 2075)Contribution Replacement RateRate 60.0 62.5 65.0 67.5 70.0 72.5 75.0 77.5 80.0 82.5 85.021 -34.6 -38.9 -43.2 -47.5 -51.7 -56.0 -60.3 -64.6 -68.8 -73.1 -77.422 -31.4 -35.7 -40.0 -44.2 -48.5 -52.8 -57.1 -61.3 -65.6 -69.9 -74.123 -28.2 -32.4 -36.7 -41.0 -45.3 -49.5 -53.8 -58.1 -62.4 -66.6 -70.924 -24.9 -29.2 -33.5 -37.8 -42.0 -46.3 -50.6 -54.9 -59.1 -63.4 -67.725 -21.7 -26.0 -30.3 -34.5 -38.8 -43.1 -47.4 -51.6 -55.9 -60.2 -64.526 -18.5 -22.7 -27.0 -31.3 -35.6 -39.8 -44.1 -48.4 -52.7 -56.9 -61.227 :JLZ :195 -23.8 28.1 3l23 -36f6 A40.9 -45.2 494 -53.7-9.28 -12.0 -16.3 -20.6 -24.8 -29.1 -33.4 -37.7 -41.9 -46.2 -50.5 -54.829 -8.8 -13.0 -17.3 -21.6 -25.9 -30.1 -34.4 -38.7 -43.0 -47.2 -51.530 -5.5 -9.8 -14.1 -18.4 -22.6 -26.9 -31.2 -35.5 -39.7 -44.0 -48.331 -2.3 -6.6 -10.9 -15.1 -19.4 -23.7 -28.0 -32.2 -36.5 -40.8 -45.132 0.9 -3.3 -7.6 -11.9 -16.2 -20.4 -24.7 -29.0 -33.3 -37.5 -41.833 4.2 -0.1 -4.4 -8.7 -12.9 -17.2 -21.5 -25.8 -30.0 -34.3 -38.634 7.4 3.1 -1.2 -5.4 -9.7 -14.0 -18.3 -22.5 -26.8 -31.1 -35.435 10.6 6.4 2.1 -2.2 -6.5 -10.7 -15.0 -19.3 -23.6 -27.8 -32.1
  49. 49. V. PreliminaryConclusionsPopulation aging and pension system reforms are serious challengesto bothdeveloped and developingcountries inthe coming decades. Theoretical thinkingandpolicy debates in recent years, including Arrau and Schmidt-Hebbel(1994), Diamond(1995), Iyer (1994), Rein (1994), Sayeed (1984), the World Bank (1994), havecontributed significantinsightsin findingsustainable solutions for providingincomesupport to the increasinglynumber of elderly. Thispaper proposes anintegratedframework linking pension reformsto macroeconomic development, demographicdynamics, and labor market conditions. With basic assumptionson the futuredevelopment trends, the model specified in the paper can be utilized to simulatepotentialimpact of different reform options under different pension arrangements. Thesesimulationsgive decision-makers some quantitative indication on possibleoutcomes apolicy package can generate.Based on the World Banks population projection for China, numericalresultsfrom the simulationexercises conducted in the paper suggest some importantpolicymessages. Notably, the scope and speed of population aging in Chinamakes the presentpension systemfinanciallyunsustainable, even though the GDP is assumedto growsteadilyover a long period. Pension reforms are inevitable. For the current pay-as-you-go pensionsystem, increasingretirement age would provide a temporary fixand wouldonlybe politicallyviablewhen there is excess labor demand. Pension fund sustainabilitycan be improved by increasingGDP growth, raising contribution rates or graduallyreducing benefit rates. But implementation of these reform options need carefulassessmentof financialcosts and social obstacles. Establishingfully fundedandprivatelymanaged pension schemes can be a feasibleand effective alternative,but sound regulatoryframework and institutionalinfrastructure need to be establishedto facilitatethe transition.43
  50. 50. More generally,the following considerations should be usefulfor formulatinglonger-term pension reform strategies.First, pensionsystemreform is a long-term problem of multi-dimension,involvingnot only economic, but also social, political, and cultural factors. The governmentsshould not over-concentrate on the old age income security system pe se, relyingexclusivelyon taxes and transfers to achieve the goals of inter-generationalor intra-generational incomeredistribution. A more strategic approach with broader policyoptions needs to be explored. Country characteristics, mostlyimportantlythe cultural andtraditional values as well as expectations and reactions of differentinterest groups, have tobe taken into consideration..Second, real incomegrowth is the ultimate means to cope withpoverty amongtheold, especiallyin developingcountries. An important policy objectivein Chinafor thenext fiftyto one hundredyears should still be to maintainlong-term socio-economicstability and sustainablegrowth. This is clearly a precondition toward establishingasufficient,efficient,and equitable social security systemin China.Third, internationalexperience should be studied to avoid repeatingthe samemistakes others made before. Countries in Europe and Latin Americahave experimentedwith differentkinds of reform options to improve their pension systems. Their experiencesshow that in order to get out of the lurking population aging trap, Chinashould movetowards a systemthat combines two elements. One is a fullyfunded and portable, definedbenefit pension plandesigned to meet basic needs. The other is occupationalpensionplans or personal savingsaccounts to satisfy demand for maintainingor improvinglivingstandards.Fourth, specificaction plans need to be worked out for long-termtargets ofestablishinga sustainablesocial security system. Real economic growth can be reinforcedthrough improvementof the labor market; for instanceby removingmanagementrigidities,44
  51. 51. facilitating humanresource development, strengthening competitivenessin labor markets,reforming the householdregistration system, bettering the incentive structure, andrewarding hard and innovativework. More job opportunities can be created innon-traditional sectors, especiallyservice industries which are still under-developedinChina atpresent, so that the unemploymentrate can be lowered. Whilerapid urbanizationmaycreate some troubles, development of medium-size cities can be a feasibleway to convertsome of rural populationmainlyengaged in agricultural activitiesto urban populationlessdependent on land. Preserving traditional values and maintainingfamily-communitysupport -- the informalsystemfor income security of the aged with no governmentandlittle market involvement-- willcushion some impact of demographic shocks. Financialmarkets should be liberalizedand developed so that people can have adequate savingsinstruments and insuranceoptions, and private savings can be stimulated andput intoproductive investment. With proper inflation control, workers can buildup financialassets now to pay for expenditure later. Management of pensionfunds needsto betransparent and decentralizedto reduce both corruption and administrativecosts. Thebalance of pensionfunds should be invested efficientlyto generate prudent returns.However, manykey issues need further exploration, includingamongothers, thecost of switchingfrom one pension system to another more desirablesystem,theendogenous choiceof aworker with options to select from differentpensionarrangements, the linkagebetween pension systems and labor supply,and theinterdependence amongsavings through pension arrangements and through otherchannels. The modelingtechnique needs improvement, for instance,in the direction ofgenerational accounting(see Auerbach and Kotlikoff, 1985, Auerbach. KotlikoffandHagemann, 1989,Boll,Raffelhuschen and Walliser, 1994). Meanwhile empiricalstudiesshould be extendedto cover regional disparities and sectoral differences.45
  52. 52. ReferenceAhmnad,E. andA.Hussies,1991,SocialSecurityin China:a HistoricalPerspective,inE.Ahead,et al.eds., SocialSecurityinDevelopingCountries,OxfordUniversityPressArrau, P. andK. Schmidt-Hebbel,1994,PensionSystemandReform:CountryExperiencesandResearchIssues,RevistadeAnalisisEconomico9(1):3-20Auerbach,A.J.andL.J. Kotlikoff,1985,SimulatingAlternativeSocialSecurityResponsestoDemographicTransition,NationalTax Journal38, 153-168Auerbach,A.J.,L.J.Kotlikoff,Hagemann,R.P. andG.Nicoletti,1989,TheEconomicDynamicsof anAgeingPopulation:TheCaseof Four OECDCountries,OECDEconomicStudies12:97-130Boll,S., B. RaffelhuschenandJ. Walliser,1994,SocialSecurityandIntergenerationalRedistribution:AGenerationalAccountingPerspective,PublicChoice81(1-2):79-100Diamond,P.A., 1995,EconomicSupportinOld Age,Paperpresentedat theAnnualBankConferenceonDevelopmentEconomics,TheWorldBank,WashingtonD.C.Diamond,P.A.andJ.A.Mirrlees,1978,AModelof SocialInsurancewithVariableRetirement,JournalofPublicEconomics10(3):295-336Dixon,J.E. andD. Macarov,1992,SocialWelfarein SocialistCountries,London:RoutledgeHu, Z., 1994,SocialProtection,labormarketrigidity,andenterpriserestructuringinChina,IMFpaperonpolicyanalysisandassessment,PPAA/94/22Hussain,A., 1993,Reformof theChineseSocialSecuritySystem,LondonSchoolof EconomicsandPoliticalScience,DevelopmentEconomicsResearchProgrammeWorkingPaper27Hussain,A., 1994,SocialSecurityinpresentday ChinaandIts Reform,London:ChinaProgram,LondonSchoolof EconomicsNo.29Iyer,S.N., 1993,PensionReforminDevelopingCountries,InternationalLaborReview132(2):187-207Lee, C., 1993,ChinasTransitiontowardsthe Market:"Socialization"of theSafetyNet,ChinaEconomicReview4(2):169-80Rein,M., 1994,SolidaritybetweenGenerations:A FiveCountryStudyof theSocialProcessofAgeing,IntemationalInstitutefor AppliedSystemsAnalysis,No:94-019Sayeed,A., 1984,ASurveyofPensionReformRecommendations,inConklin,D.W.Ed.PensionsTodayandTomorrow,pages17-6146
  53. 53. SchieberS.J.andJ.B. Shoven,1994,TheConsequencesof PopulationAgingonPrivatePensionFundSavingandAssetMarkets,NBER WorkingPaper4665Tian,X., 1995,PopulationControlina MarketEconomy,SocialSciencesinChina16:74-83WorldBank,1995,PopulationProjectionfor China,WorldBankDatabaseSTARSWorldBank,1994,AvertingtheOldAgeCrisis:Policiesto Protectthe OldandPromoteGrowth,OxfordUniversityPressWorldBank,1990,China:ReformingSocialSecurityina SocialistEconomy,Yue,G., 1985,Employment,WageandSocialSecurityinChina,InternationalLaborReview124:411-2247
  54. 54. II
  55. 55. I
  56. 56. Policy Research Working Paper SeriesContactTitle Author Date for paperWPS1585 Public Finances and Economic Luca Barbone March 1996 C. PelegrinTransition Hana Polackova 85087WPS1586 Did External Barriers Cause the Azita Amjadi March 1996 S. LipscombMarginalization of Sub-Saharan Ulrich Reincke 33718Africa in World Trade? Alexander YeatsWPS1587 Payments and Finance Problems Constantine April 1996 M. de la Puentein the Commonwealth of Michalopoulos 31206Independent StatesWPS1588 Social Insurance in the Transition Deborah Mabbett April 1996 L. Bielyto a Market Economy: Theoretical 36280Issues with Application to MoldovaWPS1589 The Analysis of Emerging Policy Sudarshan Gooptu April 1996 R. VoIssues in Development Finance 31047WPS1590 The Latvian Banking Crisis: Alex Fleming April 1996 S. CoffeyLessons Learned Samuel Talley 32535WPS1591 Competition in Network Industries Michael Klein April 1996 S. Vivas82809WPS1592 The Stockmarkets as a Source of Cherian Samuel April 1996 C. SamuelFinance: A Comparison of U.S. and 30802Indian FirmsWPS1593 Rural Finance for Growth and Poverty Saeed Qureshi April 1996 C. AnbiahAlleviation ljaz Nabi 81275Rashid FaruqeeWPS1594 Why Have Some Indian States Done Gaurav Datt April 1996 P. SaderBetter Than Others at Reducing Martin Ravallion 33902Rural Poverty?WPS1595 In Search of Owners: Lessons of Cheryl W. Gray April 1996 M. GellerExperience with Privatization and 31393Corporate Governance in TransitionEconomiesWPS1596 Agrarian Structure in Poland: The Johan van Zyl April 1996 M. WilliamsMyth of Large-Farm Superiority Bill R. Miller 87297Andrew ParkerWPS1597 Egypt and the Uruguay Round Bernard Hoekman May 1996 F. HatabArvind Subramanian 35835WPS1598 Logistical Constraints on Francois Amiot May 1996 L. LewisInternational Trade in the Maghreb Ovadia Salama 37369
  57. 57. PolicyResearchWorkingPaperSeriesContactTitle Author Date for paperWPS1599 EconomicImplicationsforTurkey GlennW.Harrison May 1996 M. Patenaof aCustomsUnionwiththe ThomasF.Rutherford 39515EuropeanUnion DavidG.TarrWPS1600 IsCommodity-Dependence NanaeYabuki May1996 G. llogonPessimismJustified?Critical TakamasaAkiyama 33732FactorsandGovernmentPoliciesthatCharacterizeDynamicCommoditySectorsWPS1601The DomesticBenefitsof Tropical KennethM.Chomitz May1996 PRDEIForests:ACriticalReview KantaKumari 80513EmphasizingHydrologicalFunctionsWPS1602Program-BasedPollutionControl ShakebAfsah May1996 A. WilliamsManagement:The Indonesian BenoitLaplante 37176PROKASIHProgram NabielMakarimWPS1603 InfrastructureBottlenecks,Private AlexAnas May1996 S.WardProvision,andIndustrialProductivity: KyuSik Lee 31707A Studyof IndonesianandThai MichaelMurrayCitiesWPS1604 Costsof InfrastructureDeficiencies KyuSik Lee May 1996 S. Wardin ManufacturinginIndonesia, Alex Anas 31707Nigeria,andThailand Gi-TaikOhWPS1605WhyManufacturingFirmsProduce KyuSikLee May1996 S. WardSomeElectricityInternally Alex Anas 31707SatyendraVermaMichaelMurrayWPS1606The Benefitsof AlternativePower AlexAnas May1996 S.WardTariffsfor NigeriaandIndonesia KyuSikLee 31707WPS1607PopulationAgingandPension F.DesmondMcCarthy May1996 M.DivinoSystems:ReformOptionsfor China KangbinZheng 33739