the 2003 aging vulnerability index -...
TRANSCRIPT
The 2003 Aging Vulnerability Index
An Assessment of the Capacityof Twelve Developed Countriesto Meet the Aging Challenge
Center for Strategic andInternational Studies andWatson Wyatt Worldwide
W W W. C S I S . O R G ■ W W W. WAT S O N W YAT T. C O M
4
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
THE 2003
AGING VULNERABILITY
INDEX
An Assessment of the Capacity of Twelve Developed Countries
to Meet the Aging Challenge
by Richard Jackson and Neil Howe
The Center for Strategic and International Studies
March 2003
i
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
Executive Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Part I. The Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
The Projection Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Category One: Public-Burden Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Category Two: Fiscal-Room Indicators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Category Three: Benefit-Dependence Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Category Four: Elder-Affluence Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Part II. The Index Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Low Vulnerability Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Medium Vulnerability Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
High Vulnerability Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Technical Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Demographic Scenario. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Economic Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Projection Base and Horizon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Household Income and Taxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Pension and Health Care Benefit Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Fiscal Scenario. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Indicator Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Category and Overall Rankings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Data Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
About the Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
TABLE OF CONTENTS
The rapid aging of the developed countries willpose a major challenge for global prosperityand stability during the first half of the twenty-first century.
Today, there are 30 pension-eligible elders inthe developed world for every 100 working-age adults. By the year 2040, there will be 70.In Italy, Japan, and Spain, the fastest-agingcountries, there will be 100. In other words,there will be as many retirees as workers. Thisrising old-age dependency ratio will translateinto a sharply rising cost rate for pay-as-you-goretirement programs — and a heavy burden onthe budget, on the economy, and on working-age adults in any country that does not takeserious steps to prepare.
Ten years ago, global aging barely registered asa policy issue. Today, with the retirement oflarge postwar baby-boom generations loomingjust over the horizon, it is the focus of growingconcern among political and policy leadersworldwide. From Australia to Sweden, thedeveloped countries are beginning to debate —and enact — major reforms. Yet despite allthe new attention, there exists no satisfactorymeasure of the magnitude of the challenge.
The CSIS Aging Vulnerability Index assesses the“vulnerability” of the developed countries torising old-age dependency costs. In this firstedition, the Index covers twelve countries —Australia, Belgium, Canada, France, Germany,Italy, Japan, the Netherlands, Spain, Sweden,the United Kingdom, and the United States.In future editions, the Index may be expandedto include the rest of the developed world —or, data allowing, selected developing countries.
The Index gives each country an overall rankingand score. The scores show that the twelvecountries fall into three clear groups — a low,a medium, and a high vulnerability group.
Low Vulnerability Group In the 2003 Index, the low vulnerability groupincludes Australia (1), the United Kingdom (2),and the United States (3). These English-speaking countries win the first three placesthanks to their favorable demographics, theirrelatively inexpensive public benefit systems,and their well-developed private alternatives.The group does face some real challenges.The UK, for example, is finding it difficult tokeep costs down without hurting elder livingstandards, while the United States must grapplewith runaway health care spending. Still, allof these countries are in relatively good shape.Australia, in particular, is implementing a far-sighted strategy of mandatory private pensioncoverage that promises excellent results onall fronts.
Medium Vulnerability Group There are six medium vulnerability countries:Canada (4), the English-speaking straggler;Sweden (5) and Germany (7), two continental
EXECUTIVE SUMMARY
iii
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
Aging Vulnerability Index 2003 EditionRankings from Least to Most Vulnerable
Low Vulnerability1. AUSTRALIA2. UNITED KINGDOM3. UNITED STATES
Medium Vulnerability4. CANADA 5. SWEDEN6. JAPAN7. GERMANY8. NETHERLANDS9. BELGIUM
High Vulnerability10. FRANCE11. ITALY12. SPAIN
European countries that have recently enactedmajor benefit reforms; Japan (6), which ranksmuch higher in the Index than its massive agewave might indicate; and the Netherlands (8)and Belgium (9), two countries with generousand unreformed benefit systems. All of themedium vulnerability countries, includingCanada, face more serious demographic chal-lenges than the low vulnerability countries. Anddespite recent reforms, all, including Sweden andGermany, face heavier old-age dependencyburdens.
High Vulnerability Group The high vulnerability group includes threemajor continental European countries that allface a daunting fiscal and economic future:France (10), Italy (11), and Spain (12).Their poor scores can be attributed, in varyingdegrees, to severe demographics, lavish benefitformulas, early retirement, and heavy elderdependence on pay-as-you-go public support.It is unclear whether they can change coursewithout economic and social turmoil. Italy hasscheduled big reductions in future pensionbenefits, but only after grandfathering nearlyeveryone old enough to vote. France and Spainhave yet to initiate major reforms of elderbenefit programs.
The Index is compiled from indicators in fourbasic categories, each dealing with a crucialdimension of the challenge:
■ Public-burden indicators, which track thesheer magnitude of the public spendingburden in each country
■ Fiscal-room indicators, which track eachcountry’s ability to accommodate thegrowth in old-age benefits via highertaxes, cuts in other spending, or publicborrowing
■ Benefit-dependence indicators, which trackhow dependent the elderly are on publicbenefits and thus how politically difficultit may be to reduce their generosity
■ Elder-affluence indicators, which trackthe relative affluence of the old versus theyoung — another trend that could criticallyaffect the future politics of benefit reform
The projections underlying the Index are basedon a “historical trends” scenario, a no-wishful-thinking baseline that assumes a continuationof established demographic and economictrends. According to the projections, publicbenefits to the elderly will reach an averageof 25 percent of GDP in the developed coun-tries by 2040, double today’s level. In Japan,they will reach 27 percent of GDP; in France,they will reach 29 percent; and in Italy andSpain, they will exceed 30 percent. This growthwill throw into question the sustainability oftoday’s retirement systems — and indeed,society’s very ability to provide a decent stan-dard of living for the old without overbur-dening the young.
The Aging Vulnerability Index is the first attemptto develop a comprehensive measure of theold-age dependency challenge that is compa-rable across the developed countries. As such,it must be regarded as experimental. We offerthis first edition of the Index in the hope thatit will stimulate debate and focus attention onthe need for reform. The Index clearly showsthat global aging is pushing much of thedeveloped world toward fiscal and economicmeltdown. There is still time to avert crisis.But time is running short, and the problemis worse than is generally supposed.
iv
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
The authors wish to thank the Richard M.Fairbanks Foundation for believing in theAging Vulnerability Index and supporting theresearch with a generous grant. They are also grateful to Watson Wyatt Worldwide for publishing the resulting report.
In addition, the authors want to single out andthank a number of people for their specialcontributions. Paul Hewitt, Director of theCSIS Global Aging Initiative, developed theoriginal concept. Without him, there would beno Index. Sylvester Schieber, Director of theResearch and Information Center at WatsonWyatt Worldwide, offered much sound advicethat improved the product. Martins Jansons,who helped assemble the Luxembourg IncomeStudy data, deserves special recognition forhis skillful — and tireless — efforts. KeisukeNakashima provided valuable assistance innavigating the Japanese data sources. Craig Romm, the Global Aging Initiative’sindispensable Program Coordinator, helpedkeep everything on track.
The report also benefited from discussions withnumerous experts worldwide. The authorscollectively acknowledge their contributionsto the report while retaining responsibilityfor its shortcomings.
v
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
ACKNOWLEDGMENTS
We live in an era defined by many challenges,from global warming to global terrorism. Butnone is as certain as global aging. And noneis as likely to have as large and enduring aneffect — on the size and shape of governmentbudgets, on the future growth in living stan-dards, and on the stability of the globaleconomy and even the world order.
The whole world is aging, and the developedcountries are leading the way. For most ofhistory, the elderly — here defined as adultsaged 60 and over — comprised only a tinyfraction of the population, never more than fivepercent in any country. Today in the developedcountries, they comprise 20 percent. Fortyyears from now, the share will reach roughly35 percent. And that’s just the average. InJapan and some of the fast-aging countries ofcontinental Europe, where the median age isexpected to exceed 50, the share will beapproaching 50 percent (see Figure 1).
Longer lives may be a great personal boon,but they also pose a great collective challenge.Over the next half century, global aging willdivert a rising share of society’s economicresources from young to old. Much of thistransfer will occur through public benefitsystems. All of the developed countries haveuniversal or near-universal public pension andhealth care benefit programs that tax currentworkers to pay benefits to current retirees. In many countries, the typical elder is almostentirely dependent on a government check.Global aging will throw into question the sus-tainability of today’s pay-as-you-go retirementsystems — and indeed, society’s very abilityto provide a decent standard of living for theold without overburdening the young.
Ten years ago, global aging barely registered as apolicy issue. Today, with the retirement of largepostwar baby-boom generations looming justover the horizon, it is the focus of growing con-cern among leaders worldwide. From Australia
to Sweden, the developed countries are begin-ning to debate — and enact — major reforms.
Yet despite all the attention, there exists nosatisfactory measure of the magnitude of thechallenge to guide leaders in their policychoices. Not all national governments makelong-term cost projections, and those that dorarely include all benefit programs. In any case,the national projections are not comparableacross countries. Recent multi-country projec-tions by the Organization for Economic Co-operation and Development (OECD) havebegun to address the comparability problem,but they assume a variety of fiscally favorabledevelopments — more babies, more workingwomen, slower growth in health care spending,and slower growth in life expectancy — thatmay greatly understate future costs.
1
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
INTRODUCTION
20002040
US
Australia
Canada
UK
France
Netherlands
Belgium
Germany
Sweden
Japan
Spain
Italy
F I G U R E 1 : Number of Elderly, as a Percent of the Population
16.3%26.0%
16.6%30.1%
17.0%33.3%
20.8%33.9%
20.6%35.0%
18.5%35.5%
22.2%36.1%
23.5%37.4%
22.8%37.5%
23.9%44.7%
22.0%45.5%
24.4%46.2%
Elderly are aged 60 and over
2
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
The CSIS Aging Vulnerability Index offers amore complete and realistic assessment of the“vulnerability” of the developed countries torising old-age dependency costs. The projec-tions underlying the Index are based on a“historical trends” demographic and economicscenario, a no-wishful-thinking baseline thatassumes a continuation of established trendsin fertility, longevity, labor-force participation,and productivity growth. The Index’s coreindicators quantify the magnitude of thepublic-benefit burden. In addition to trackingtotal old-age benefits as a share of GDP, theIndex calculates the after-tax cost of benefitsto the elderly as a share of the income of thenonelderly. This innovative “net-transfer”indicator directly measures the impact of currentsocial policy toward the old on the livingstandard of the young.
The sheer magnitude of the future public-benefit burden, of course, is not the only cir-cumstance affecting the vulnerability of thedeveloped countries. The overall fiscal, eco-nomic, and social environment also matters.The Index therefore goes a step further andlooks at a variety of broader factors that maymitigate or aggravate the public burden.
These additional indicators are grouped intothree categories: fiscal room, benefit depen-dence, and elder affluence. The fiscal-roomindicators take into account the differing “room”that governments may have to raise taxes, cutother spending, or borrow in order to accom-modate the growth in old-age benefits. Withits small public sector, Japan, for instance, mayhave more room to raise taxes than Sweden —while Sweden, with its small public debt, mayhave more room to borrow than Japan.
The benefit-dependence indicators take intoaccount the extent of elder reliance on publicbenefits, the availability of alternative means ofsupport, including earnings and private pen-sions, and the strength of extended families.
The elder-affluence indicators take into accountthe relative economic well-being of society’selderly and nonelderly members. Trends ineach of these areas will influence the ability ofcountries to reduce the public burden by enactingor following through with cost-cutting reforms.
The bottom line is an Index that ranks the majordeveloped countries in terms of their overallvulnerability to the costs of rising old-agedependency. The Index now covers twelvecountries: Australia, Belgium, Canada, France,Germany, Italy, Japan, the Netherlands, Spain,Sweden, the United Kingdom, and theUnited States. In the future, it may be expandedto include the rest of the developed world —or, data allowing, selected developing countries.
Part I of the report briefly describes the pro-jection scenario, then discusses the individualindicators that make up the Index — whythey were selected, how they are calculated,and what they reveal. Part II gathers togetherthe strands of the story. It describes how theindividual indicators are combined into anIndex. It also presents the overall rankings andsummarizes the country findings. A TechnicalAppendix contains a more complete discussionof the model, methodology, and data sourcesused in constructing the Index.
The Aging Vulnerability Index is the first attemptto develop a comprehensive measure of theold-age dependency challenge that is compa-rable across the developed countries. As such,it must be regarded as experimental. We offerthis first edition of the Index in the hope thatit will stimulate debate and focus attention onthe need for reform. The Index clearly showsthat global aging is pushing much of thedeveloped world toward fiscal and economicmeltdown. There is still time to avert crisis.But time is running short, and the problem isworse than is generally supposed.
During the early postwar decades, when thedeveloped countries greatly expanded govern-ment’s role in providing old-age benefits, theydecided to pay for nearly all the benefits on apay-as-you-go basis. The pay-as-you-go modelwas attractive because it allowed early partici-pants to receive benefits far in excess of theircontributions. At the time, the model alsoappeared to be affordable. Because the numberof retirees was small and the number of workerswas growing rapidly, the total cost seemedeasily bearable even after allowing for steadybenefit hikes. What’s more, everyone expectedthis situation to continue indefinitely.
In more recent decades — especially since themid-1970s — these expectations have crum-bled. The developed countries stand on thebrink of a great demographic transformationthat no one predicted when the pay-as-you-gomodel was adopted. It’s called global aging,and it will soon be putting enormous pressureon public budgets and national economies. In afew countries, that pressure is already building.
There are two forces behind global aging. Thefirst force is falling fertility. People are havingfewer babies, which decreases the relative num-ber of young in the population. As recently asthe mid-1960s, every developed country wasat or above the so-called 2.1 replacement rateneeded to maintain a stable population fromone generation to the next. Today every countryis at or below it, some far below it. In Japan,the fertility rate is 1.4; in Germany, it is 1.3;and in Italy, Spain, and much of southern andcentral Europe, it is 1.2. Only the United Statesremains at the replacement rate — barely.
The second force behind global aging is risinglife expectancy. People are living longer, andthis boosts the relative number of old in thepopulation. Over the past fifty years, lifeexpectancy at birth has risen by roughly two
and a half months per year in the developedcountries, for a cumulative gain of about tenyears. Few demographers expect this trendto slow — and a growing number believethat it may accelerate.
“Demography is destiny,” it is sometimes said.When it comes to public budgets, the prophecyis irrefutable. Rising longevity and falling fertilitytranslate directly into a rising ratio of retiredbeneficiaries to taxpaying workers. Overall,CSIS projects that the ratio of “elderly” adultsaged 60 and over to “nonelderly” adults aged15–59 in the developed world will increasefrom 30 percent today to 70 percent by 2040.In Italy, Japan, and Spain, the fastest aging of thetwelve Index countries, this old-age dependencyratio will increase to 100 percent by 2040 —meaning that there will be as many adults overthe age of 60 as under (see Figure 2).* A rising
3
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
PART I. THE INDICATORS
US
Australia
Canada
UK
France
Netherlands
Belgium
Germany
Sweden
Spain
Japan
Italy
F I G U R E 2 : Aged Dependency Ratio*
0.260.47
0.260.56
0.260.63
0.340.65
0.340.69
0.290.69
0.370.71
0.380.72
0.380.74
0.350.99
0.391.00
0.401.03
*As explained below, age 60 was chosen as the threshold between elderly and nonelderly because it is much closer to the typicalage of retirement on public benefits in most countries than the more conventional age 65 threshold.
20002040
*Ratio of elderly adults aged 60 and over to working-ageadults aged 15–59.
4
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
old-age dependency ratio will in turn translateinto a rising cost rate for pay-as-you-go retire-ment programs — and a growing burden onthe budget, on the economy, and on working-age adults.
The purpose of the CSIS Aging VulnerabilityIndex is to measure and compare the magnitudeof this burden in different countries and, byextension, the “vulnerability” of those countriesto demographic aging. The Index is calculatedby combining country scores on eleven sepa-rate indicators. The indicators are grouped intofour categories, each dealing with a distinctdimension of the problem:
■ The first and most important category ispublic burden. It contains three indicatorsthat measure the sheer magnitude of eachcountry’s projected public old-age depen-dency burden.
■ The second category is fiscal room. It con-tains three indicators that measure eachcountry’s ability to accommodate the growthin its public old-age dependency burdenvia higher taxes, cuts in other spending, orgovernment borrowing.
■ The third category is benefit dependence.It contains three indicators that measurehow dependent the elderly in each countryare on public benefits and thus how politi-cally difficult it may be to reduce thosebenefits beneath current law — or even tocarry out reductions in benefits that arealready scheduled to take place.
■ The fourth category is elder affluence. Itcontains two indicators that measure percapita elderly income in each country relativeto nonelderly income — another factor thatcould critically affect the future politics ofreform.
The CSIS Aging Vulnerability Index is constructedas follows. For each of the indicators, we gen-erate an indicator ranking, from one (best) totwelve (worst). We also transform the indicatorresults into an index and generate an indexscore for each country. For each indicator, themean result is set to an index value of 50;results that lie above and below the mean byone standard deviation are set, respectively, toindex values of 100 and zero. The index scoresthus preserve the indicator rankings while alsoreflecting the relative distance of each rankedcountry, positively or negatively, from the “centerof the pack.” (Countries located far from themean can and sometimes do have index scoresgreater than 100 or less than zero.)
For each of the four categories, a category scoreis then calculated as an average of the indicatorindex scores. The category score determines theoverall category rankings. Finally, the categoryscores themselves are averaged as follows: Aweight of one-third is given to the first public-burden category, one-third to the second fiscal-room category, and one-third to the third andfourth “policy-climate” categories combined.A country’s final combined average for the fourcategories determines its ranking in the over-all Index.
Each of the indicators tells an interesting storyand is worth examining in detail. Before look-ing at the results, however, it is essential tounderstand the basic scenario and assumptionsthat underlie all of the projections.
The Projection ScenarioThe projection scenario extends from 2000, thebase year, through 2040. The 2040 end yearwas chosen because the “demographic transi-tion” in most countries will by then be largelycomplete. Even after 2040, rising longevity willcontinue to push the cost of retirement bene-fits steadily and indefinitely upward. But theera of swiftest growth, which in most countriesaccompanies the full retirement of a postwar
5
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
baby boom, will occur sometime between themid-2010s and the mid-2030s. A country thatcan successfully navigate the fiscal rapids overthe next four decades will, presumably, be quiteprepared to manage the gentler current there-after. A country that fails to meet the challengeby 2040 will be far more concerned with confronting the destructive legacy of that failure(from high tax levels to runaway debt) thanwith managing any new demographic devel-opments after 2040.
All of the projections assume a continuationof historical demographic and economic trends.Fertility is assumed to remain constant, mortalityto continue declining at its historical pace, andproductivity and real wages to continue risingin line with their average growth rate over thelast twenty-five years. Rates of labor-force participation by age group are also assumed toremain unchanged, with the two exceptionsnoted below. In general, so is the compositionof national income by type of factor income —“asset income” from capital and “earningsincome” from labor — although, as popula-tions age, the distribution of each type betweenage groups will of course shift.
A few exceptions to this “historical trends” ruleare permitted. The scenario allows for an increasein labor-force participation among middle-agedwomen as a new generation of working womennow in their twenties and thirties bring newhabits with them into their forties and fifties.It also allows for a modest rise in labor-forceparticipation among the elderly in a numberof countries, notably Germany and Italy, whererecent pension reforms have begun to changework incentives.
As for policy, the projections assume a contin-uation of current old-age benefit provisions.Public pension programs continue to functionaccording to the same benefit rules that are inplace in the base year. Exceptions (importantin some countries) are allowed only where
future changes in benefit rules have alreadybeen explicitly enacted into law. Health carebenefits are a special case, since health carespending per beneficiary typically grows rapidlyyet unpredictably even with no change in ben-efit rules. Here, the assumption is that growthin per-beneficiary health care spending willinitially follow the historical trend in eachcountry, but that growth in all countries willgradually converge, by 2040, to the rate ofgrowth in per capita GDP plus one percent.In most countries, this implies that cost growthwill be slower in the future than it has beenin the past. While there is little evidence thatcosts are yet slowing, all experts agree that theymust do so eventually to avoid impossibleoutcomes — such as total health care spend-ing crowding out every other category ofconsumption.
The Index makes two other major policyassumptions. The first is debt neutrality: Inthe first year of the projection, each countryis assumed to move to a general governmentdeficit which, when continued unchanged asa share of GDP in all future years, will keep netgovernment debt unchanged as a share of GDP.While unrealistic as a short-term projection,debt neutrality is a standard assumption inlong-term budget models. Year to year, thebudget balance shows a great deal of variability,usually rising and falling with the businesscycle. To suppose that the balance in any sin-gle base year will continue indefinitely is anarbitrary and often an unsustainable assump-tion. Each country is assumed to achieve itsnew budget balance by an increase (or decrease)in taxes combined with an equal decrease (or increase) in government spending.
We considered — but decided against — mak-ing an exception for countries that are proposinga long-term policy of budget surpluses as apartial solution to the aging challenge. Thehistorical failure of governments throughoutthe developed world to validate retirement
6
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
“trust-fund” savings by running sustainedgeneral government surpluses raises seriousquestions about the feasibility of this approach.Among the twelve Index countries, only one —Canada — has announced plans to pursue sucha policy. While Canada may succeed whereothers have failed, the projected level of“prefunding” is small relative to the projectedsize of its dependency burden. Even if factoredinto the Index, it would not alter Canada’soverall ranking.
The other assumption is that, once debt neu-trality is achieved, nonbenefit spending willremain constant as a share of GDP and taxeswill be raised (or lowered) in each future yearin accordance with the projected change inbenefit spending. In every country in almostevery year, of course, this means that taxes mustbe raised. This “rising tax” assumption is onlyrelaxed for two of the fiscal-room indicators,where the object is to assess alternative meansof paying for the growth in benefit costs.
Throughout the Index, the “elderly” are definedas persons aged 60 and over, the “nonelderly”as persons aged 15–59. The income (and taxation) of each age group refers to the incomeof all individuals within that age group, withthe single exception of spouses of heads ofhouseholds, who are considered to belong tothe same age group as the head of household.Other than in the case of spouses, income tohouseholds containing both elderly and non-elderly adults is split between the two groups.The age 60 threshold between elderly and non-elderly may strike some people as a bit early,since in today’s world most 60-year-olds seemrelatively “young.” The threshold, however, isnot meant to indicate anything about health,capacity, or vigor. Age 60 was chosen becauseit is close to the typical age of retirement onpublic benefits in most countries — muchcloser, in fact, than age 65. Before age 60, themajority of people are net contributors. Afterage 60, the majority become net beneficiaries.
The data underlying the projection scenariocome from a wide variety of sources, the mostimportant of which are discussed in theTechnical Appendix. The basic demographicdata come from the UN. Most of the fiscal andmacroeconomic data are derived from standardOECD sources, from the National Accounts tothe Social Expenditure Database. The detailedhousehold data, from income by age to livingarrangements, are derived from the LuxembourgIncome Study for every country except Japan,where data from the Ministry of Health, Labor,and Welfare are used.
Let us now turn to the indicators. The public-burden indicators are presented first, since theyfocus most directly on the core challenge —the projected magnitude of the old-age depen-dency burden. The fiscal-room indicators,which focus on how easily each country canaccommodate the growth in that burden, arepresented second. The benefit-dependence andelder-affluence indicators, which focus on howeasily each country can restrain the growth, arepresented last.
Category One: Public-Burden IndicatorsThe public-burden category includes threeindicators:
■ Public benefits to the elderly in 2040 as a percent of GDP (the “benefit-level” indicator)
■ The growth from 2000 to 2040 in publicbenefits to the elderly as a percent of GDP(the “benefit-growth” indicator)
■ Public benefits to the elderly in 2040 as apercent of the income of the nonelderly(the “net-transfer” indicator)
The first two indicators track the claim thatold-age benefits will place on society’s totaleconomic resources. The third measures thedirect transfer burden that old-age benefits
7
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
will place on the income of the working-agepopulation. Together, they provide a compre-hensive measure of the public burden ofpopulation aging.
Let’s start with the benefit-level indicator. CSISprojects that public benefits to the elderly in thetwelve Index countries will grow to an averageof 24.8 percent of GDP by 2040, double today’slevel. Of this total, 13.5 percent of GDP willbe going to pensions, 8.7 percent to health carebenefits, and 2.6 percent to miscellaneousprograms, from unemployment insurance tohousing subsidies and food stamps.
The average projection masks a wide range ofoutcomes (see Figure 3). Total old-age benefitsin Australia, the UK, and the United States areprojected to grow to between 17 and 20 percentof GDP by 2040. In Belgium, Canada, France,Germany, Japan, the Netherlands, and Sweden,they are projected to grow to between 23 and29 percent. And in Italy and Spain, they areprojected to exceed 30 percent. The differencesare due in part to demographics, and in part tothe varying generosity of benefit systems, espe-cially pensions. The lower-burden English-speaking counties both spend less per capita onold-age benefits and are due to age less. Thehigher-burden countries of continental Europehave the most lavish benefit formulas, the earliestretirement ages, and the fastest-aging popula-tions. Japan is a special case. It faces a massiveage wave, but has a relatively stingy benefitsystem — which is what keeps total Japanesebenefit spending from shooting off the charts.
There are some important differences inrankings between the benefit-level and benefit-growth indicators (see Figure 4). A fewcountries, notably Canada and the UnitedStates, score better on level than on growth.In the case of Canada, the difference is dra-matic: a ranking of four versus a ranking ofnine. The reason lies mainly in Canada’sdemographics. With a fertility rate of only 1.6,
20002040
Australia
UK
US
Canada
Sweden
Belgium
Germany
Netherlands
Japan
France
Italy
Spain
F I G U R E 3 : Public Benefits to the Elderly, as a Percent of GDP
9.0%16.6%
12.1%17.6%
9.4%20.3%
8.9%22.9%
13.3%23.0%
12.7%24.5%
15.1%25.5%
11.7%25.6%
11.8%27.0%
15.8%29.3%
17.3%32.0%
12.6%33.1%
5.5%
7.6%
9.7%
10.4%
11.0%
11.8%
13.6%
13.9%
14.0%
14.7%
15.2%
20.5%
UK
Australia
Sweden
Germany
US
Belgium
France
Netherlands
Canada
Italy
Japan
Spain
F I G U R E 4 : Growth in Public Benefits to the Elderly from 2000 to2040, as a Percent of GDP
Elderly are aged 60 and over
Elderly are aged 60 and over
8
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
its aging trajectory is much closer to that ofcontinental Europe than to the other English-speaking countries. In the case of the UnitedStates, the reason for the poorer growth scoreis government health care benefit spending,which is projected to rise faster in the UnitedStates than in other countries.
There are also a number of countries that scorebetter on growth than on level. Germany, Italy,and Sweden all enacted reforms in the 1990sthat are scheduled to cut average pensionbenefits relative to average wages over the nextfew decades. These countries spend a lot onold-age benefits today and will spend evenmore tomorrow. But total spending will growless than demographics alone would suggest.There is also one low-burden country — theUK — that has enacted deep cuts in pensionbenefits. It still faces rising elder health care
benefit costs. Alone among the twelve Indexcountries, however, it faces virtually no growthin pension costs.
Most countries will begin to experience severefiscal pressure at about the same time. Almosteverywhere, old-age benefits are projected torise relatively slowly over the next decade, aperiod during which large postwar baby boomswill still be in the workforce. This window ofopportunity closes abruptly in the early 2010s.Thereafter, benefit spending rises steeply for twodecades before slowing or plateauing duringthe 2030s. Japan and Spain are exceptions tothis pattern. In Japan, whose massive age waveis breaking early, spending is already risingrapidly. In Spain, spending doesn’t begin toramp up until the 2010s, but once it does, itcontinues to grow rapidly throughout the entireprojection period and beyond.
So what’s more important — level or growth?The question is probably unfair, since eachindicator adds a different perspective to theIndex. The absolute spending level as a shareof GDP is clearly the simplest measure of thetotal resource burden. From this perspective,the English-speaking countries are substan-tially better off. Yet the rise in spending is alsoimportant, since some societies may be insti-tutionally and culturally better equipped tohandle high levels of public-benefit spendingthan others. From this perspective, the roadahead may not be so smooth for Canada andthe United States.
Thus far, we have focused on the projectedclaims that public old-age benefits will place onsociety’s total economic resources. To isolatespecifically the burden on the working-agepopulation, the third indicator tracks totalpublic benefits to the elderly as a share of thetotal income of the nonelderly. Here bothbenefits and income are measured after taxes.Benefits to the elderly are measured “net” —that is, they are reduced by an estimate of the
Elderly are aged 60 and over
UK
Australia
US
Sweden
Canada
Belgium
Netherlands
Germany
Japan
France
Italy
Spain
14.0%19.3%
12.4%22.1%
12.8%26.8%
17.7%27.3%
12.9%29.8%
17.2%31.5%
16.5%35.3%
20.8%35.4%
15.4%35.9%
20.3%36.1%
21.6%37.2%
17.2%42.7%
20002040
F I G U R E 5 : Net Public Benefits to the Elderly, as a Percent of After-tax Nonelderly Income*
*Public benefits to the elderly (aged 60 and over), net of taxes, as a percent of the income of the nonelderly(aged 15–59), measured after taxes for the rest of gov-ernment but before taxes for benefits to the elderly.
9
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
share of benefits paid for by taxes on the elderlythemselves. The income of the nonelderly ismeasured after all rest-of-government taxes —that is, after taking into account all taxes onthe nonelderly except those required to payfor net benefits to the elderly.
By 2040, at the low end, net transfers to theelderly will be consuming 19 percent of theafter-tax income of the nonelderly in the UKand 22 percent in Australia. In every othercountry, net transfers by 2040 will be consum-ing between 27 and 37 percent of nonelderlyincome, except for Spain, where they will weighin at 43 percent (see Figure 5). In no countrytoday do net transfers consume more than 22percent of nonelderly income. Note that theseshares do not refer to worker payroll, but tothe total household income from all sourcesof all nonelderly adults.
The ranking of countries by the net-transferindicator corresponds quite closely to theranking by total benefit spending as a shareof GDP. Since the net-transfer indicator takesinto account differences in the distribution of
the tax burden by age, we had expected to findmore variation in the rankings. But apparentlythese tax differences do not play a decisive role.In a few countries like Sweden, the elderly payfor a somewhat larger share of their owntransfers through taxes. Compared with othercountries, Sweden relies relatively less onpayroll taxation to fund old-age benefits andhas relatively few senior tax breaks. It ranksbetter on the net-transfer indicator than on thebenefit-level indicator. In a few other countries,the tax code leans the other way. In Spain, forinstance, relatively heavy reliance on payrolltaxation pushes up the net-transfer burdeneven faster than the overall benefit level.
Figure 6 summarizes the results for the public-burden category. Australia comes in first on thebenefit-level indicator, the UK on the benefit-growth and net-transfer indicators. Spain comesin last on all three. Overall, the UK receivesthe best category score — Spain the worst.
*Elderly are aged 60 and over; net transfers are public benefits to the elderly, net of taxes, as a percent of the income of thenonelderly (aged 15–59), measured after taxes for the rest of government but before taxes for benefits to the elderly.
16.6%
17.6%
20.3%
22.9%
23.0%
24.5%
25.5%
25.6%
27.0%
29.3%
32.0%
33.1%
1 UK -32
2 Australia -20
3 US 18
4 Sweden 22
5 Belgium 47
6 Canada 47
7 Germany 53
8 Netherlands 68
9 Japan 80
10 France 81
11 Italy 97
12 Spain 139
Category Rank Score
F I G U R E 6 : Public-Burden Category*
Benefit-Level Indicator
Public Benefits to the Elderly in2040, as a Percent of GDP
Benefit-Growth Indicator
Growth in Public Benefits to theElderly from 2000 to 2040, as aPercent of GDP
Net-Transfer Indicator
Net Benefits to the Elderly in2040, as a Percent of After-taxNonelderly Income
1 Australia
2 UK
3 US
4 Canada
5 Sweden
6 Belgium
7 Germany
8 Netherlands
9 Japan
10 France
11 Italy
12 Spain
5.5%
7.6%
9.7%
10.4%
11.0%
11.8%
13.6%
13.9%
14.0%
14.7%
15.2%
20.5%
19.3%
22.1%
26.8%
27.3%
29.8%
31.5%
35.3%
35.4%
35.9%
36.1%
37.2%
42.7%
1 UK
2 Australia
3 Sweden
4 Germany
5 US
6 Belgium
7 France
8 Netherlands
9 Canada
10 Italy
11 Japan
12 Spain
1 UK
2 Australia
3 US
4 Sweden
5 Canada
6 Belgium
7 Netherlands
8 Germany
9 Japan
10 France
11 Italy
12 Spain
10
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
Category Two: Fiscal-Room IndicatorsA country’s vulnerability to rising old-agedependency costs depends not only on themagnitude of the burden, but also on the fiscalroom available to accommodate the burden.There are three ways in which countries canadjust to higher old-age benefit spending: payfor the growth by raising taxes, pay for thegrowth by cannibalizing other public programs,or pay for the growth by borrowing.
The fiscal-room category includes three indi-cators that evaluate the feasibility of thesestrategies:
■ Total taxes as a percent of GDP in 2040(the “tax-room” indicator, which assumes thatbenefit growth is paid for by raising taxes)
■ Total benefits to the elderly in 2040 as a percent of total government outlays (the “budget-room” indicator, which assumesthat benefit growth is paid for by cuts inother government spending)
■ The year that the net government debtreaches150 percent of GDP (the “borrowing-room” indicator, which assumes that benefitgrowth is paid for by government borrowing)
Of the three indicators, the results for the tax-room indicator are the most straightforward.Countries with the largest projected old-agedependency burdens tend to end up with thelargest tax burdens. Since the overall tax burdenalso depends on the overall size of the publicsector, however, there are some exceptions. A fewcountries with outsized public sectors scoremuch worse on tax room than they do on thepublic-burden indicators. Sweden, for example,ranks fourth in its public-burden category score,but twelfth on tax room; Belgium ranks fifth inpublic burden, but tenth on tax room. For a fewother countries with large old-age dependencyburdens but relatively small public sectors,the reverse is true. Spain, for example, rankstwelfth in its public-burden category score,but eighth on tax room; Japan ranks ninth inpublic burden, but fourth on tax room.
What’s most striking, however, is that the taxoption would lead to extraordinary tax burdensin almost all countries. In every country exceptAustralia, total taxes would consume at least 40percent of GDP by 2040; in Belgium, Canada,Germany, Italy, the Netherlands, and Spain, theywould consume at least 50 percent; and in Franceand Sweden, they would consume more than 60percent (see Figure 7). Some European countriesmay literally find it impossible to raise taxesenough to pay for the full cost of the age wave.At some point, rather than generate new revenue,higher tax rates may simply slow the economy,exacerbate unemployment, and push moreworkers into a growing gray economy.
20002040
Australia
UK
US
Japan
Canada
Germany
Italy
Spain
Netherlands
Belgium
France
Sweden
F I G U R E 7 : Total Taxes as a Percent of GDP, Assuming Tax HikesPay for All Growth in Public Benefits
32.5%39.2%
38.4%41.7%
33.2%44.1%
28.9%46.5%
38.4%50.2%
44.4%55.2%
44.0%56.5%
38.2%57.2%
46.9%58.4%
49.5%59.8%
48.3%62.2%
57.4%63.0%
Scale begins at 25.0 percent
41.9%
11
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
The obvious alternative to letting benefits growoutward is to let them grow inward. Thebudget-room indicator looks at what wouldhappen if, instead of raising taxes, governmentssimply allowed old-age dependency costs tocrowd out other spending “dollar for dollar.”By 2040, benefits to the elderly would accountfor over 40 percent of total spending in ten ofthe twelve countries, over 50 percent in sixof the twelve, and over 70 percent in one:Spain. In the country with the largest sharetoday — Italy — they account for just 38percent (see Figure 8).
There are important differences between thebudget-room and tax-room rankings. Countrieswith large public sectors but relatively smallold-age dependency burdens — Belgium andSweden — score much better on budget roomthan tax room. The presumed opportunityimplied by the indicator is that such countriesmay have a lot of ineffective governmentspending that can be cut without much cost tosociety. Countries with relatively small publicsectors but fast-growing burdens — Japan andthe United States — score worse. They may beable to accommodate relatively little growth inold-age spending without crowding out vitalpublic services.
In theory, governments could also finance thegrowth in old-age benefits by borrowing. Thedebt-room indicator tracks the number of yearsthat they could do so before exceeding thecritical net-debt-to-GDP ratio of 150 percent,roughly the highest level any modern developedcountry has attained in peacetime. The higher acountry’s initial debt and the steeper the growthin its old-age benefit spending, the sooner itwill exceed the threshold. Japan, where benefitsare growing steeply even in the near term,breaches it first, in 2020. Italy and Belgium,which have the highest initial net debt, breachit second and third, in 2021 and 2022. With theexception of the UK, every country breachesit by 2040 (see Figure 9).
20002040
Sweden
UK
Australia
Belgium
Canada
Germany
Netherlands
France
US
Italy
Japan
Spain
F I G U R E 8 : Public Benefits to the Elderly as a Percent of TotalGovernment Outlays, Assuming Cuts in OtherSpending Pay for All Growth in Public Benefits
25.3%37.2%
31.6%38.1%
27.2%41.2%
27.5%44.6%
23.4%48.6%
32.9%49.0%
28.0%51.7%
32.5%52.9%
30.9%52.9%
37.9%60.3%
31.9%65.8%
32.7%72.1%
30.9%
10.8%
1.0%
43.5%
38.2%
42.1%
53.0%
42.1%
98.9%
96.5%
58.5%
UK
Australia
Sweden
Germany
Spain
Netherlands
US
Canada
France
Belgium
Italy
Japan
2046
2038
2037
2033
2029
2027
2026
2024
2024
2022
2021
2020
F I G U R E 9 : Net Government Debt in 2001 and Year Net DebtReaches 150 Percent of GDP, Assuming Borrowing Pays for All Growth in Public Benefits
Elderly are aged 60 and over
12
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
It’s worth recalling that the borrowing-roomindicator, like the other fiscal-room indicators,is designed to isolate the impact of risingbenefit costs. Apart from borrowing to payfor the growth in public benefit programs,each country is still assumed to pursue a debt-neutral fiscal policy in the rest of the budget.Even so, the net-debt-to-GDP ratio in eightof the twelve countries — Belgium, Canada,France, Italy, Japan, the Netherlands, Spain,and the United States — goes on to pass 300percent of GDP by 2040. For most countries,borrowing simply isn’t a long-term option.
Figure 10 summarizes the results for the fiscal-room category. Australia comes in first on taxroom, Sweden last. Sweden comes in first onbudget room, Spain last. The UK comes in firston borrowing room, Japan last. Once again,the UK receives the best overall score, Spainthe worst.
Category Three: Benefit-DependenceIndicatorsUp to now, we have assumed that old-agebenefit programs will continue to function
exactly as current law dictates. Yet in reality,every country is bound to enact many policychanges over the next forty years. And in somecountries, these changes may have a dramaticimpact on total benefit spending — eithercutting it beneath what current law now author-izes (easing the dependency burden), or hikingit above what current law now authorizes(worsening the dependency burden). In thiscontext, hiking benefit spending does notnecessarily mean making benefits more gener-ous than they are today. In several countries,the mere failure to follow through with futurebenefit cuts that are already scheduled bycurrent law would have a dramatic impact.Italy, Germany, Sweden, and the UK are allcountries whose future dependency burdenwill be much larger than current projectionsindicate if scheduled benefit cuts do notmaterialize.
Is there any way to assess the likelihood thatfuture policy changes will lean one way or theother? Clearly, one factor that will facilitate orobstruct reform is the degree of elder depen-dence on public benefits. Part of this depen-
39.2%
41.7%
44.1%
46.5%
50.2%
55.2%
56.5%
57.2%
58.4%
59.8%
62.2%
63.0%
1 UK -28
2 Australia -12
3 Sweden 32
4 US 41
5 Germany 43
6 Canada 51
7 Netherlands 66
8 Belgium 68
9 Japan 78
10 France 82
11 Italy 88
12 Spain 91
Category Rank Score
F I G U R E 1 0 : Fiscal-Room Category
Tax-Room Indicator
Total Taxes as a Percent of GDP in2040, Assuming Tax Hikes Pay forAll Growth in Public Benefits
Budget-Room Indicator
Public Benefits to the Elderly as aPercent of Total GovernmentOutlays in 2040, Assuming Cuts inOther Spending Pay for All Growthin Public Benefits
Borrowing-Room Indicator
Year Net Government DebtReaches 150 Percent of GDP,Assuming Borrowing Pays for All Growth in Public Benefits
1 Australia
2 UK
3 US
4 Japan
5 Canada
6 Germany
7 Italy
8 Spain
9 Netherlands
10 Belgium
11 France
12 Sweden
37.2%
38.1%
41.2%
44.6%
48.6%
49.0%
51.7%
52.9%
52.9%
60.3%
65.8%
72.1%
1 Sweden
2 UK
3 Australia
4 Belgium
5 Canada
6 Germany
7 Netherlands
8 France
9 US
10 Italy
11 Japan
12 Spain
1 UK 2046
2 Australia 2038
3 Sweden 2037
4 Germany 2033
5 Spain 2029
6 Netherlands 2027
7 US 2026
8 Canada 2024
9 France 2024
10 Belgium 2022
11 Italy 2021
12 Japan 2020
13
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
dence can be measured by looking at benefitincome as a share of total household income.Another part can be measured by assessing thestrength of informal sources of nonincomesupport, the most important of which is theextended family. A related issue is the extentto which reductions in public benefits wouldpush elders into poverty, a concern to whichmost of today’s developed societies appear tobe very sensitive.
Accordingly, the benefit-dependence categoryincludes three indicators:
■ Public benefits as a percent of after-taxelderly income in 2040 (the “benefit-share”indicator)
■ The percent of the elderly who live with theiradult children (the “family-ties” indicator)
■ The percent of the elderly who would dropbelow the poverty line if public benefits werecut by 10 percent (the “poverty-impact”indicator)
Dependence on government benefits varies agreat deal among the developed countries. It isrelatively low in the English-speaking world,and in two countries — Australia and theUK — it is projected to drop steadily in thefuture. In 2040, according to CSIS projections,public benefits will make up 33 percent of theafter-tax income of the elderly in Australia,37 percent in the United States, 39 percent inthe UK, and 46 percent in Canada. In everycontinental European country, the share exceeds50 percent today and will still exceed 50 per-cent in 2040, suggesting that the countriesthat most need to cut benefits may find it themost difficult to do so. Japan is an exceptionamong high-burden countries. The publicbenefit share there is relatively low today andwill still be low in 2040: just 38 percent,about what it will be in the United States(see Figure 11).
The benefit-share indicator probably under-estimates the absolute level of elder dependence.It uses a cash measure of income that excludeshealth care benefits. Including health carebenefits, public transfers are a significantlylarger share of income. The indicator results,moreover, are averages for all elders, includingthe affluent. Currently, public benefits make upa larger share of the income of “typical” eldersin the third quintile of the income distributionthan they do for the average elder — roughly55 percent of total income in the United Statesand between 70 and 85 percent of total incomein Europe. While we did not project publicbenefits by income level, it is safe to say thatthe much greater dependence of middle- andlow-income seniors is likely to persist in mostcountries. Australia is a likely exception. In fact,dependence among middle-income Australianelders is due to fall dramatically as the coun-try’s new, mandatory private pension system
20002040
Australia
US
Japan
UK
Canada
Sweden
Netherlands
Italy
Belgium
Germany
Spain
France
F I G U R E 1 1 : Public Benefits to the Elderly, as a Percent of After-tax Elderly Income*
43.4%32.7%
34.9%36.6%
34.9%38.4%
50.3%39.3%
42.2%45.5%
56.5%54.6%
54.0%55.2%
58.8%55.6%
54.7%56.2%
61.4%58.4%
64.0%66.9%
67.3%69.5%
Scale begins at 25.0 percent
* Income excludes government health care benefits.
Elderly are aged 60 and over
14
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
matures, elder incomes rise, and fewer retireesqualify for means-tested public pensions.
What about private sources of income? TheIndex tracks three broad income types: earningsincome, including self-employment income;asset income other than private pensions; andprivate pension income. The last categoryincludes all types of employer-provided retire-ment payments and personal pensions, bothlump-sum and annuitized. The only restrictionis that the arrangements be funded, whetherexternally through “autonomous” pension fundsor internally through book reserves. Whilegeneral asset income includes the return onmuch savings that is in effect retirement savings,the Index tracks private pension income sepa-rately because it represents a formal retirementprovision that is often intended as a substitutefor public benefits.
The countries with the lowest public-benefitdependence are also, of course, the countries
where elders rely more heavily on privatesources of income. In Japan and the English-speaking world, a significant share of elderscontinue to work. Earnings now make up 41percent of elderly income in Japan, 26 percentin the United States, 23 percent in Australia,18 percent in Canada, and 16 percent in theUK. With the exception of Sweden, continentalEurope lags well behind. In Spain, earningsaccount for 10 percent of income; in France,8 percent; in the Netherlands, just 7 percent.
It is also important to keep in mind thatearnings among the elderly, even in theEnglish-speaking countries, are highly skewedby income. Except in Japan, Sweden, and theUnited States, earnings make up less than 10percent of the income of middle-income elders.By and large, it is affluent elders who work —or perhaps more accurately, working elderswho are affluent.
Income from assets is generally more importantthan income from earnings, especially formiddle-income elders. Even in Japan, with itstradition of working seniors, only a minorityof the elderly have jobs. Most elderly every-where, however, have at least some savings.Asset income now comprises about 20 to 25percent of income for the average elder in mostcountries and about 15 to 20 percent of incomefor middle-income elders. The relative impor-tance of asset income is projected to decline inthe future, but not because total income fromsavings will be declining. Households will besaving more in the form of private pensions —and the projections assume that households willoffset some of the growth in pension savingsby reducing nonpension savings.
The growth in private pensions is due bothto the maturation of existing systems and, inthe case of Australia, Germany, Sweden, andthe UK, to recent reforms that have added orexpanded a second, funded “private” tier topublic benefit systems. CSIS projections show
20002040
Australia
Japan
UK
Canada
Netherlands
US
Sweden
Italy
Belgium
Germany
Spain
France
F I G U R E 1 2 : Private Pension Benefits, as a Percent of GDP*
2.5%11.1%
3.1%9.2%
3.6%8.9%
3.2%8.5%
3.3%8.2%
3.6%7.4%
2.2%5.3%
1.3%4.3%
1.7%3.6%
0.9%3.4%
0.2%3.1%
0.1%0.3%
*Includes all employer pensions, personal pensions, andseverance pay schemes.
15
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
that by 2040 private pension benefits as a shareof GDP will grow by at least 100 percent inevery country — and by roughly 200 percenton average across all countries (see Figure 12).This is much faster than the growth in theelderly as a share of the adult population,which is roughly 70 percent on average acrossall countries. The relative importance of privatepensions in elderly income will thus increasenearly everywhere. A few countries that nowhave underdeveloped private systems —Germany, Italy, and Spain — will see especiallyrapid growth. Since the growth is from a smallbase, however, private pensions in thesecountries will still be a small share of elderlyincome in 2040, about 10 percent. The onlycountries where private pensions will exceed20 percent of elderly income — Australia,Canada, Japan, the Netherlands, the UK, andthe United States — are countries where privateprovision is already relatively important today.
Along with their own income, elders may alsobe able to rely on the income of younger mem-bers of their extended families. Until now, allof the index measures have excluded the incomeof the nonelderly from calculations of elderlyincome, even if they live in the same house-hold as elders. As a result, we can now lookat the availability of private support from theyoung as a separate factor which, all otherthings being equal, may make it easier for somecountries to cope with their rising old-agedependency burden. Ideally, the Index shouldlook at all forms of family support, includingcash transfers and unpaid caregiving. It wouldalso net transfers from the old to the young,especially bequests, against transfers from theyoung to the old. As a first approximation ofthe importance of family support, the family-tiesindicator simply looks at the prevalence ofmultigenerational living.
Specifically, the indicator measures the share ofall elderly who now live in extended familieswith their adult children, whether the parents
live in their grown children’s household or —what is much more common in most countries— the grown children live in their parents’household. Japan, where fully 50 percent of allelders live with their grown children, headsthe list. In Italy and Spain, the two southernEuropean countries in the Index, the share ofelders living in multigenerational homes is alsoquite large: 42 and 41 percent, respectively.In the next runner up, Canada, the share isjust 20 percent. All the other countries are inthe teens or single digits (see Figure 13).
Multigenerational living constitutes a compen-sating advantage for countries like Italy, Japan,and Spain. It not only allows relatively poorelders to live with their more affluent adultchildren, it also allows relatively poor youngadults to live with their more affluent parents.It mitigates the old-age dependency burdennot just by providing an extra source of sup-port for the old, but by providing a form of“trickle down” support for the young as well.The share of elders living with adult childrenhas been declining over the past few decadesin all countries and may continue to do so inthe future. But unless the relative propensityfor multigenerational living changes, thosecountries that score well today will continueto enjoy an advantage.
50.4%
42.1%
40.5%
20.3%
19.5%
15.6%
15.1%
13.7%
13.0%
12.4%
9.1%
5.0%
*Data refer to latest year available, generally in mid-1990s.
Japan
Italy
Spain
Canada
Australia
Belgium
US
Germany
UK
France
Netherlands
Sweden
F I G U R E 1 3 : Percent of the Elderly Living with Their Adult Children*
Elderly are aged 60 and over
16
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
The final benefit-dependence indicator looksat the percentage increase in the number ofelderly beneath the poverty line that wouldbe caused by an immediate 10 percent cut inpublic benefits. The standard internationaldefinition of poverty is used here: A person is“poor” if he or she lives in a household withan income that is less than 50 percent of themedian income for all households in eachcountry. Performance on this indicator isdetermined, first, by the distribution of elderlyincome around the poverty threshold and,second, by the degree of dependence on publicbenefits among elder households around thepoverty threshold. Here income does includethe income of nonelderly household membersin order to get a complete household picture.
The results for this indicator are surprising —and instructive (see Figure 14). One mightexpect that the continental European countries,with their expansive welfare states, would scorebest on a poverty indicator. In fact, they scoreworst. One reason is the greater dependenceon public benefits among elder households inthese countries, which means that any givenpercentage cut in benefits translates into a largerpercentage cut in total household income.
Another reason is precisely the success enjoyedby Europe’s expansive welfare states in liftingthe incomes of so many elder households abovethe poverty line — but often not far above thepoverty line. Even a small cut in benefits coulddump many of these households back intopoverty, while in other countries they may neverhave been lifted out of poverty to begin with.
In the countries that do best on this indicator— Japan, Italy, and the United States —either the share of public benefits in elderlyincome is relatively low, the extent of multi-generational living is relatively high, or both.Although the poverty-impact indicator couldnot be projected into the future, the resultsfor today provide a valuable indication ofpossible obstacles to reform.
Figure 15 summarizes the results for thebenefit-dependence category. Australia comesin first on benefit share, France last. Japancomes in first on family ties, Sweden last. Japanalso comes in first on the poverty-impactindicator, Germany last. Overall, Japan receivesthe best score, France the worst.
Category Four: Elder-Affluence IndicatorsThe size and direction of future policy changesmay also be influenced by how societies per-ceive the relative affluence of the old and young.Analysis of voter attitudes toward governmentspending often suggests that political supportfor public benefits to the elderly is stronglyassociated with the perception that the elderlyare much poorer on average than youngeradults. If elder affluence is high or rising in thefuture, it may facilitate efforts to reduce publicold-age dependency costs. If the old remainrelatively poor despite a rising burden on theyoung, it may be harder to enact benefit cuts —or to follow through on cuts that are alreadyscheduled but not yet implemented.
The elder-affluence category, like the benefit-dependence category, is an effort to monitor
2.0%
2.9%
2.9%
3.9%
3.9%
4.0%
4.1%
4.5%
4.8%
5.2%
5.2%
5.7%
*Poverty threshold is 50 percent of the median income for all householdsin each country; income excludes government health care benefits; data refer to latest year available, generally in mid-1990s.
Japan
Italy
US
Spain
Canada
Australia
UK
Sweden
Netherlands
France
Belgium
Germany
F I G U R E 1 4 : Percent of the Elderly Pushed into Poverty by a 10 Percent Cut in Public Benefits*
Elderly are aged 60 and over
17
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
the social and political environment for reform.It includes two indicators:
■ The ratio of the per capita income of the elderly to the per capita income of thenonelderly in 2040 (the “elder-affluence-level” indicator)
■ The percentage change in that ratiobetween now and 2040 (the “elder-affluence-trend” indicator)
The elder-affluence indicators refer to after-taxincome. The income of both the elderly andnonelderly includes the average cash value ofgovernment health care benefits. For each agegroup, however, income excludes the income ofpersons of the other age group (except spouses)living in the same household.
The level and trend indicators each offer animportant and independent perspective. Thelevel indicator is critical if one assumes thatsociety compares the living standard of theold and the young directly against each otheraccording to an absolute dollar-value metric
that translates into equivalent size of home,model of car, length of vacation, and so forth.The trend indicator is preferable, on theother hand, if one assumes that society eval-uates the relative affluence of the old andyoung against some customary standard ofgenerational fairness that may have nothingto do with dollar-value equivalence and maybe different for every culture.
As it turns out, the rankings for the two indica-tors generally overlap — that is, the countrieswhere the elderly are projected to be relativelyaffluent in 2040 are also the countries whereelder affluence is trending upwards, and viceversa. Canada and the United States are nearthe top of both rankings. In both countries, theelderly start out well off relative to the youngtoday and keep getting better off. The reasonfor the upward trend: minimal scheduled cutsin current-law pensions combined with largelong-term increases in per capita health carebenefits. These results suggest that, over thelong run, worries about elder destitution areunlikely to deter cost-saving reform (seeFigure 16).
32.7%
36.6%
38.4%
39.3%
45.5%
54.6%
55.2%
55.6%
56.2%
58.4%
66.9%
69.5%
1 Japan -32
2 Italy 14
3 US 19
4 Australia 26
5 Canada 42
6 UK 45
7 Spain 47
8 Sweden 80
9 Netherlands 81
10 Belgium 82
11 Germany 93
12 France 103
Category Rank Score
F I G U R E 1 5 : Benefit-Dependence Category*
Benefit-Share Indicator
Public Benefits to the Elderly in2040, as a Percent of After-taxElderly Income
Family-Ties Indicator
Percent of the Elderly Living withTheir Adult Children: Most RecentYear Available
Poverty-Impact Indicator
Percent of the Elderly Pushed into Poverty by a 10 Percent Cut in Public Benefits: MostRecent Year Available
1 Australia
2 US
3 Japan
4 UK
5 Canada
6 Sweden
7 Netherlands
8 Italy
9 Belgium
10 Germany
11 Spain
12 France
50.4%
42.1%
40.5%
20.3%
19.5%
15.6%
15.1%
13.7%
13.0%
12.4%
9.1%
5.0%
2.0%
2.9%
2.9%
3.9%
3.9%
4.0%
4.1%
4.5%
4.8%
5.2%
5.2%
5.7%
1 Japan
2 Italy
3 Spain
4 Canada
5 Australia
6 Belgium
7 US
8 Germany
9 UK
10 France
11 Netherlands
12 Sweden
1 Japan
2 Italy
3 US
4 Spain
5 Canada
6 Australia
7 UK
8 Sweden
9 Netherlands
10 France
11 Belgium
12 Germany
*Elderly are aged 60 and over; income excludes governement health care benefits; poverty threshold is 50 percent of the medianincome for all households in each country.
18
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
Italy, Sweden, and the UK, on the other hand,are near the bottom of both rankings. Thereason for the downward trend: large scheduledcuts in current-law pension benefits and, inthe case of Sweden, relatively slow growth inhealth care benefits. The very low rankings ofthese countries suggest that they may find itpolitically difficult simply to follow throughwith already scheduled benefit cuts. The UKmay shrug at such a warning, since its long-term cost outlook is so favorable. Italy andSweden cannot afford to be complacent, sinceboth must contemplate cost reductions beyondthose already scheduled.
Figure 17 summarizes the results for the elder-affluence category. The United States comes infirst on the affluence-level indicator, Swedenlast. Japan comes in first on the affluence-trendindicator, Italy last. Overall, the United Statesreceives the best score, Italy the worst.
1.77
1.54
1.51
1.45
1.44
1.36
1.34
1.19
1.18
1.15
1.10
1.04
1 US -29
2 Canada 14
3 Japan 17
4 Netherlands 22
5 Germany 28
6 Australia 34
7 Spain 49
8 France 58
9 Belgium 67
10 Sweden 100
11 UK 117
12 Italy 121
Category Rank Score
F I G U R E 1 7 : Elder-Affluence Category*
Affluence-Level Indicator
Per Capita Ratio of the After-taxIncome of the Elderly to theNonelderly in 2040
Affluence-Trend Indicator
Percentage Change in the Affluence-Level Indicator from2000 to 2040
1 US
2 Canada
3 Netherlands
4 Australia
5 Germany
6 France
7 Japan
8 Belgium
9 Spain
10 Italy
11 UK
12 Sweden
25.9%
23.9%
19.7%
16.5%
15.3%
14.4%
12.0%
10.9%
5.5%
3.0%
-8.1%
-12.9%
1 Japan
2 US
3 Spain
4 Canada
5 Germany
6 Netherlands
7 Australia
8 Belgium
9 France
10 Sweden
11 UK
12 Italy
*Elderly are aged 60 and over; nonelderly are aged 15–59; income includes the cash value ofgovernment health care benefits.
20002040
US
Canada
Netherlands
Australia
Germany
France
Japan
Belgium
Spain
Italy
UK
Sweden
F I G U R E 1 6 : Per Capita Ratio of the After-tax Income of theElderly to the Nonelderly*
1.431.77
1.321.54
1.321.51
1.301.45
1.251.44
1.291.36
1.071.34
1.071.19
0.981.18
1.321.15
1.201.10
1.011.04
*Elderly are aged 60 and over; nonelderly are aged 15–59;income includes the cash value of government healthcare benefits.
19
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
It’s time to gather together the strands of thestory — and present the final results.
As already explained, the Index is constructed asfollows. For each of the indicators, we generatean indicator ranking, from one (best) to twelve(worst). We also transform the indicator resultsinto an index and generate an index score foreach country. For each indicator, the meanresult is set to an index value of 50; results thatlie above and below the mean by one standarddeviation are set, respectively, to index values of100 and zero. The index scores thus preservethe indicator rankings while also reflecting therelative distance of each ranked country, posi-tively or negatively, from the “center of thepack.” (Countries located far from the meancan and sometimes do have index scoresgreater than 100 or less than zero.)
For each of the four categories, a category scoreis then calculated as an average of the indicatorindex scores. The category score determines theoverall category rankings. Finally, the categoryscores themselves are averaged as follows: Aweight of one-third is given to the first public-burden category, one-third to the second fiscal-room category, and one-third to the third andfourth “policy-climate” categories combined.A country’s final combined average for thefour categories determines its ranking in theoverall Index.
A glance at the index scores reveals that thecountries clearly sort themselves into threegroups. Australia, the UK, and the United Statesare well ahead of the next runners-up. At theother end, France, Italy, and Spain lag the restby an equally wide margin. The other six areclustered in the middle. This suggests a three-fold division of the Index into low vulnerability,medium vulnerability, and high vulnerabilitygroups.
This three-fold division is in fact much morerobust than the rankings themselves. Relatively
small changes in assumptions can change theordering of countries within the three cat-egories. Germany, Japan, and Sweden easilychange places, since their scores are virtuallyidentical. The same is true of Belgium and theNetherlands. Large changes in several indica-tors, however, would be required to push acountry from one group to another. Absentsuch changes, the three low vulnerabilitycountries always end up on top, and the threehigh vulnerability countries always end up onthe bottom.
Low Vulnerability CountriesRank Country Index Score
1 Australia –12 United Kingdom +73 United States +18
In the 2003 Index, the low vulnerability groupincludes Australia (1), the United Kingdom(2), and the United States (3). These English-speaking countries win the first three placesthanks to their favorable demographics, theirrelatively inexpensive public benefit systems,and their well-developed private alternatives.The group does face some real challenges.The UK, for example, is finding it difficult tokeep costs down without hurting elder livingstandards, while the United States must grapplewith runaway health care spending. Still, allof these countries are in relatively good shape.Australia, in particular, is implementing a far-sighted strategy of mandatory private pensioncoverage that promises excellent results onall fronts.
Medium Vulnerability CountriesRank Country Index Score
4 Canada + 425 Sweden + 486 Japan + 507 Germany + 528 Netherlands + 629 Belgium + 63
PART II. THE INDEX RESULTS
20
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
There are six medium vulnerability countries:Canada (4), the English-speaking straggler;Sweden (5) and Germany (7), two continentalEuropean countries that have recently enactedmajor benefit reforms; Japan (6), which ranksmuch higher in the index than its massive agewave might indicate; and the Netherlands (8)and Belgium (9), two countries with generousand unreformed benefit systems. All of themedium vulnerability countries, includingCanada, face more serious demographic chal-lenges than the low vulnerability countries.And despite recent reforms, all, includingSweden and Germany, face heavier old-agedependency burdens.
High Vulnerability Countries Rank Country Index Score
10 France + 8111 Italy + 8412 Spain + 93
The high vulnerability group includes threemajor continental European countries that allface a daunting fiscal and economic future:France (10), Italy (11), and Spain (12).Their poor scores can be attributed, in varying
degrees, to severe demographics, lavish bene-fit formulas, early retirement, and heavy elderdependence on pay-as-you-go public support.It is unclear whether they can change coursewithout economic and social turmoil. Italy hasscheduled big reductions in future pensionbenefits, but only after grandfathering nearlyeveryone old enough to vote. France andSpain have yet to initiate major reforms ofelder benefit programs.
Although countries that rank near each otherin the Index typically share key characteristicsin common, they sometimes differ in criticaland even surprising respects. So while theoverall rankings show the basic elevation ofthe terrain, the individual indicators revealplenty of cross-cutting ridges and canyons.
The UK, for example, does best overall in thepublic-burden category, but next-to-worst inelder affluence. Sweden scores at the bottomon tax room — but near the very top onbudget and borrowing room. Japan faces assevere a demographic challenge as any coun-try, but climbs up in the Index because itsmodest public benefits, high rates of elder
F I G U R E 1 8 : Overall Index*
Public Fiscal Benefit ElderOVERALL INDEX Burden Room Dependence Affluence
Country Rank Score Rank Score Rank Score Rank Score Rank Score
Australia 1 -1 2 -20 2 -12 4 26 6 34
UK 2 7 1 -32 1 -28 6 45 11 117
US 3 18 3 18 4 41 3 19 1 -29
Canada 4 42 6 47 6 51 5 42 2 14
Sweden 5 48 4 22 3 32 8 80 10 100
Japan 6 50 9 80 9 78 1 -32 3 17
Germany 7 52 7 53 5 43 11 93 5 28
Netherlands 8 62 8 68 7 66 9 81 4 22
Belgium 9 63 5 47 8 68 10 82 9 67
France 10 81 10 81 10 82 12 103 8 58
Italy 11 84 11 97 11 88 2 14 12 121
Spain 12 93 12 139 12 91 7 47 7 49
*The public-burden and fiscal-room categories are each weighted one-third; the benefit-dependenceand elder-affluence categories are weighted a combined one-third.
21
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
employment, and strong family ties largelyinsulate elders from the pain of cost-cuttingreform. Italy has scheduled deep cuts in futurebenefits that raise its public-burden ranking —but only at the expense of impoverishing itsfuture elderly. The Netherlands has continentalEurope’s largest private pension system, yetnonetheless scores poorly on both publicburden and benefit dependence. Figure 18shows all of this graphically by lining up eachcountry’s category rankings side-by-side.
Let’s conclude by taking a quick look at thetwelve countries — and what the Index revealsabout where each is heading and whether itis likely to change course.
Low Vulnerability CountriesNumber 1: AustraliaAustralia is either number one or two on sixdifferent indicators and scores especially well(number two) on the two most critical categoryrankings: public burden and fiscal room.Australia starts out with favorable demograph-ics, a lean public sector, and a low-cost, largelymeans-tested old-age safety net. In the early1990s, it began to phase in a very large systemof mandatory and funded retirement accounts(called the Superannuation Guarantee). In theyears and decades to come, the “Super” willprovide increasingly generous support forelders while actually reducing the pay-as-you-go burden on working-age adults. By 2040,the payout from funded private pensions willreach 11 percent of GDP, far exceeding thepayout of any other country. Australia over-takes its only close competitor, the UK, due toits higher ranking on every benefit-dependenceand elderly-affluence indicator. Alone amongthe twelve, Australia is meeting and winningthe old-age dependency challenge with a long-term strategy that has no obvious shortcomings.
Number 2: United KingdomThe UK takes first place in both of the corepublic-burden and fiscal-room categories. Like
Australia, the UK is helped by its relative youthand favorable demographic future. Its publicpensions are not projected to grow at all as ashare of GDP, due to systematic reductions inper capita benefits initiated under Prime MinisterThatcher and continued under Major and Blair.As for health care benefits, spending on eldersunder its famously frugal national health caresystem is projected to grow only modestly.The UK is trying to expand worker participa-tion in private occupational pensions andpersonal accounts. But unlike Australia, theparticipation is not mandatory, leaving largegaps in coverage. As a result, the UK earns onlyan average ranking in the benefit-dependencecategory and a next-to-last-place ranking inelder affluence. Overall, the vulnerabilityoutlook for the UK is excellent. Yet the Indexalso reveals some doubts about its ability tostay the course.
Number 3: United StatesOf all the developed countries, the United Statesfaces the most favorable demographic future.With the highest fertility rate and one of thehighest immigration rates, it will be far andaway the youngest of the twelve Index countriesby 2040. Given this demographic advantage,together with a modest Social Security benefitformula, a relatively high rate of elder employ-ment, and a well-developed private pensionsystem, the United States could easily haveranked number one in the Index. But it did not,mostly due to the high projected growth ingovernment health care benefits. The UnitedStates ranks third on the benefit-level andnet-transfer indicators, but fifth on the benefit-growth indicator, lower than two of Europe’sbiggest welfare states, Sweden and Germany.Also, unlike Australia or the UK, the UnitedStates has no overall policy in place to expandits private pension coverage in the future.The growing relative affluence of US elders,however, augurs well for possible cost-cuttingreforms down the road.
22
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
Medium Vulnerability CountriesNumber 4: CanadaAt number four, Canada comes in last amongthe English-speaking countries. To begin with,Canada faces a less favorable demographicfuture than the others. It also starts out witha larger public sector, no plans for future cutsin public pensions, and a projected rate ofhealth care benefit growth that is nearly equalto that of the United States. As a result, the costof Canada’s public old-age benefits as a shareof GDP is due to grow to two and a half timesthe current cost by 2040 — giving it an evenworse ranking on benefit growth than theUnited States (ninth place, just between theNetherlands and Italy). As shown by the tax-room indicator, the Canadian public sectorwill weigh in at 50 percent of GDP by 2040,6 percent more than the United States and 11percent more than Australia. Even so, Canadadoes better on benefit level and tax room thanany country in continental Europe. As in theUnited States, moreover, the rising affluenceof its elders could favor future reform efforts.
Number 5: SwedenRemarkably, it is Sweden — the very symbolof welfare state excess — that ranks highestamong the countries of continental Europe.To be sure, Sweden’s overall score is pulleddown by the vast size of its public sector(which gives it a last-place rank on tax room).On the other hand, Sweden did enact in 1999a major and effective program of old-age benefitreform. It scheduled significant future cuts inits pay-as-you-go public pensions and addeda supplementary system of mandatory andfunded personal accounts. It earns a numberthree rank on the benefit-growth indicator anda number four rank in the entire public-burdencategory. It also ranks number one on budgetroom and number three on borrowing room.There are caution lights. Swedish elders lackstrong family ties (last place), are very vulnerableon both affluence indicators, and are projectedto remain highly dependent on public benefits.
Number 6: JapanNow aging faster than any other developedcountry, Japan is projected to have one of thehighest old-age dependency ratios in the world.It finishes ninth on the benefit-level and net-transfer indicators, eleventh on benefit growthand budget room, and twelfth on borrowingroom. So how does Japan manage to rise to themiddle of the overall rankings, with an indexscore of exactly 50? For one thing, Japan’spublic pensions are relatively stingy and itspublic sector is relatively small. This keeps theoverall cost of old-age benefits from rising evenhigher than it does, while allowing Japan afourth-place finish on tax room. Over the longterm, moreover, the social and political reformenvironment in Japan appears surprisinglyfavorable. Japan ranks number three in theelder-affluence category. And it ranks numberone in the benefit-dependence category, witha first-place rank on family ties, a first-placerank on poverty impact, and (due to very highrates of elder employment and a broad-basedprivate pension system) a third-place rank onbenefit share. No other country with compa-rable demographics and a comparable publicburden enjoys as many offsetting advantages.
Number 7: GermanyLike Sweden, Germany recently enacted a majorreform (the so-called Riester Reform) that com-bines scheduled future cuts in public pensionbenefits with an effort to expand private pen-sion coverage. But Germany’s reform does lesson both counts, a fact reflected in its pooreroverall ranking. Like many older Europeancountries, it scores better (number four) onbenefit growth than it does on benefit level(number seven). It has an average ranking inthe fiscal-room category (number five), doingworse than Sweden on all indicators except taxroom. In the elder-affluence category, Germanyis again about average. But in the benefit-dependence category, it ranks second-to-last —with a number eight ranking on family ties, a number ten ranking on benefit share, and a
23
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
dead-last ranking on poverty impact. This poorshowing calls into question Germany’s abilityeven to follow through on the modest benefitreductions it has already enacted — much lessto initiate any new cost-cutting reform.
Number 8: NetherlandsThe Netherlands is a prime example of anaffluent and aging welfare state that lingers in a condition of policy paralysis. With remark-able uniformity, nearly every indicator for theNetherlands reflects its lackluster eighth-placeoverall ranking. All three of its public-burdenindicators, all three of its fiscal-room indicators,and two out of three of its benefit-dependenceindicators come in between number six andnumber nine. The Netherlands does enjoy onelarge economic advantage: a mature and well-funded private pension system that is the envyof the rest of continental Europe. But its publicpension benefits are so generous, its retirementage so early, and its family ties so weak thatelders in the Netherlands are left just as vul-nerable and dependent on public benefits aselsewhere. The Netherlands’ relatively highrank in the elder-affluence category reflectsthis unique combination of generous publicand generous private benefits — and offerssome hope for future reform.
Number 9: BelgiumLike the Netherlands, Belgium epitomizes thetroubled outlook for the aging societies ofcontinental Europe. Its fertility rate is low, itspublic sector is large, its public pensions aregenerous, its retirement age is early, and itselders are heavily dependent on public benefits.Unlike Germany and Sweden, Belgium has yetto enact a major retirement reform package.In some respects, it does better than average.On the benefit-level and net-transfer indicators,it actually comes out ahead of every othercountry in continental Europe except Sweden.But in other respects, it does worse — forexample, on tax room and borrowing room,where it ranks number ten due to its long
history of fiscal excess. Clearly this is a country,like most of continental Europe, that has yetto face up to its demographic future.
High Vulnerability CountriesNumber 10: FranceFrance is a curious case. Its demographics areno more severe than Germany’s and Sweden’s —or indeed, the UK’s and Canada’s. Yet Francefinishes in tenth place in the public-burdencategory, tenth place in the fiscal-room cate-gory, and twelfth place in the benefit-depen-dence category. Japan’s stingy public benefitshelp to offset its demographic disadvantage.France is just the opposite. Its lavish publicpensions overwhelm its milder demographics— and catapult it into the high vulnerabilitygroup. Even more serious than the sheer mag-nitude of projected dependency costs is theextreme degree of elder dependence on publicbenefits. Private pensions make up a smallershare of elder income in France (less than 1percent) than in any other country; earningsmake up a smaller share (8 percent) than in anycountry except the Netherlands. Its last-placeranking on benefit share suggests that cost-cutting reforms will meet with widespreadresistance — as indeed, they already have inrecent years. But such resistance may not bethe main obstacle when, perhaps uniquely inFrance, it is difficult to find policy leaderswho will even discuss the need for reform.
Number 11: ItalyBy 2040, Italy will have one elder for everyworking-age adult — putting it in a dead heatwith Japan and Spain for the developed world’sworst demographics. On top of that, Italy’spublic pension spending as a share of GDP is already the highest in the Index countries,while the Italian elderly are among the leastlikely to be employed or to receive a privatepension. Yet what’s most worrisome is thatItaly’s overall eleventh-place ranking in theIndex comes despite a series of pension reformsenacted in the 1990s (the “Amato” and “Dini”
24
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
reforms) that are scheduled to make steep cutsin future benefits. Without these reforms, Italywould surely be in last place. The open ques-tion is whether Italy is likely to make good onits reform promises. Its second-place rankingin the benefit-dependence category (with familyties second only to Japan) gives some hope.Yet Italy comes in last in the elder-affluencecategory, a reflection of how seriously futurebenefit cuts threaten elder living standards.
Number 12: SpainSpain might be described as Italy without thereforms. These two countries face a very simi-lar (and dire) demographic future. They bothhave generous public pension systems thatsupport elders who have little employment orprivate pension income to fall back on. UnlikeItaly, however, Spain’s pension generosity is dueto continue unchecked into the indefinitefuture. Spain finishes in last place on all threepublic-burden indicators, and on the net-transferindicator it does so by a wide margin. Asidefrom its runaway pension projections, however,Spain hardly fits the stereotype of the bloatedwelfare state. Its public sector is still relativelysmall. Its net debt is modest. And to judge byits average rankings in the elder-affluence andpublic-dependence categories (it has the third-strongest family ties after Italy and Japan),reform may not be beyond its grasp.
25
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
The appendix describes the projection model,the methodology, and the most importantdata sources used in constructing the CSISAging Vulnerability Index.
1. Demographic ScenarioAll of the projections underlying the Indexassume a continuation of historical trends infertility, longevity, and immigration. This dis-tinguishes them from projections by the OECD,which assume a rebound in fertility, slowerlongevity gains, and large changes in netimmigration.1
The CSIS demographic scenario is based on theUN’s latest (2000 Revision) “constant-fertility”scenario, which assumes that rates of fertility(in all countries) and net immigration (in mostcountries) will continue at their 1995–2000averages.2 The UN scenario, however, alsoassumes that the historical rate of improvementin mortality will slow. We adjust the UN pro-jections to reflect a continuation of the histor-ical trend in mortality. Specifically, we assumethat age-specific mortality rates will continue todecline at their long-term postwar (1950–94)average in each country. This assumptionraises projected longevity relative to the UNscenario in every country except Germany, withthe biggest adjustments in Australia, Canada,Italy, and Japan (see Table A).3
It’s hard to find any evidence that a reboundin fertility is likely any time soon. All of thelong-term trends that have depressed fertilityover the past few decades, from more workingwomen to more effective contraception, remainin force. In a number of countries, to be sure,
age-specific fertility rates are now rising amongwomen in their thirties, suggesting that somewomen have merely been postponing havingchildren rather than reducing the total numberthey plan to have. The impact of this timingshift, however, is relatively small — and in anycase, it is being offset by continuing declinesin fertility among younger women. In mostof today’s lowest fertility countries, moreover,there is no evidence of a timing shift at all. InItaly and Spain (as well as Austria and Greece),fertility rates are declining or flat across everyage bracket and have been so for decades.
The case for a longevity slowdown is no moreconvincing. Demographers who expect oneargue that life expectancy cannot keep rising,since medical progress will eventually pusheverybody up against the “natural limit” to thehuman life span. This thesis, however, implies
TECHNICAL APPENDIX
1See Fiscal Implications of Ageing: Projections of Age-Related Spending, Economics Department Working Papers no. 305(OECD; 2001).
2World Population Prospects: The 2000 Revision, 2 volumes and CD-ROM (UN, Population Division; 2001).
3These calculations are based on historical trends projections for the G-7 countries prepared by Shripad Tuljapurkar of MorningView Research and published in Paul Hewitt and Sylvester Schieber, “Demographic Risk in Industrial Societies,” World Economics,I:4 (October–December 2000). The adjustments for the other five countries in the Index are estimated based on the G-7 average.
Total Fertility Rate Longevity at Birth
UN CSIS
History & Projection History Projection Projection
1995–2050 2000 2050 2050
Australia 1.8 79.5 83.0 86.7
Belgium 1.5 78.4 83.8 85.5
Canada 1.6 78.4 82.8 86.6
France 1.7 78.8 84.0 87.0
Germany 1.3 77.8 83.4 82.9
Italy 1.2 78.8 82.5 86.5
Japan 1.4 80.8 88.0 91.9
Netherlands 1.5 78.2 82.2 85.3
Spain 1.2 78.5 82.6 85.6
Sweden 1.5 79.7 84.6 86.9
UK 1.7 77.6 83.0 84.5
US 2.0 76.8 82.6 83.2
T A B L E A : Demographic Assumptions
26
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
a number of consequences that are not borneout by observation. If there is a limit to thehuman life span, mortality improvements forthe oldest elderly age brackets should beslowing relative to those for younger elderlyage brackets. At the same time, variations inlife expectancy should be narrowing as morepeople bunch up against the limit. None of thisappears to be happening. Over the last coupleof decades, the data show little tendency formortality improvements to slow at advancedages. Nor are variations in life expectancydiminishing, whether one looks at the data bycountry, by region, by income, or by education.Everywhere, people are living much longer.Yet everywhere, some groups continue to livemuch longer than others.
What happens to immigration over the nextfew decades, of course, is anyone’s guess. It isboth more volatile than the other two demo-graphic variables and more directly controlledby policy. In the past, immigration in manycountries has risen and fallen sharply, some-times over the span of just a few years. It maydo so again. Without a crystal ball, however,the most prudent assumption is that currentlaw and current practice will continue.
2. Economic ScenarioThe projections assume a continuation ofhistorical economic trends. In particular, age-specific rates of labor-force participation areassumed to remain constant at their 2000 levelsin each country, unemployment rates areassumed to remain constant at their 1990saverages, and productivity and real wages areassumed to grow at roughly their average his-torical rates over the past twenty-five years.
The projections make two exceptions to thelabor-force participation rule. The first allowsfor a “cohort effect.” In three countries in theIndex — Italy, the Netherlands, and Spain —
women in their twenties and thirties now workat significantly higher rates than women intheir forties and fifties. In our projections, we assume that women aged 40–55 in thesecountries will eventually work at the samerate as women aged 20–39. The secondexception allows for a response to pensionreforms that have scaled back early retire-ment options and improved work incentives.We base this adjustment on OECD estimatesthat recent policy changes will lead to a 10to 15 percent increase in participation ratesamong men aged 55–64 in Australia, Belgium,Germany, and the Netherlands — and to asomewhat larger rise in Italy.4
A word about productivity and wage growth isin order. CSIS follows the OECD in assumingthat that long-term productivity growth rateswill average between 1.6 and 1.8 percent peryear, roughly the historical record in mostcountries over the past twenty-five years.This assumption may be optimistic given thepotentially negative impact of population agingon both private and public savings. Slowergrowth would mean a somewhat higher old-age dependency burden, whether measured asa share of GDP or of nonelderly income. Theproductivity growth assumption, however, isnot decisive, since initial pension benefits inmost countries are indexed to wages, and sogrow (with or without a lag) along with per-worker GDP.
3. Projection Base and HorizonThe base for the Index is 2000, the latest yearfor which most demographic and economicdata were available. Data series not availablefor 2000 were trended to 2000 based on seriesthat were available. The projections are runthrough the year 2040. As already explained,this horizon was chosen because the demo-graphic transition in most countries will bythen be largely complete. Even after 2040,
4Fiscal Implications of Ageing, op. cit.
27
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
rising longevity will continue to push the costof retirement benefits steadily and indefinitelyupward. But the era of swiftest growth, whichin most countries accompanies the full retire-ment of a postwar baby boom, will occursometime between the mid-2010s and themid-2030s.
4. Household Income and TaxesThe model uses a two-step approach to cal-culate household income in the base year.We first derived totals for broad categories ofincome from economy-wide data on the house-hold and government sectors. We then allocatedthe totals to the elderly and nonelderly basedon household survey data. This approach allowsus to take into account society’s total economicresources. It also makes our measure of incomeconsistent with our GDP-based projections ofgovernment benefits, taxes, and outlays.
a. Income by TypeThe economy-wide data on household incomeare mainly derived from the OECD NationalAccounts and OECD Social ExpenditureDatabase.5 The model divides all income intofour broad categories: “earnings,” or incomefrom labor; “asset income,” or income fromcapital except for private pensions; privatepensions; and “public benefits.” Public bene-fits include all government transfers, both cashand in-kind, and are divided into three sub-components: public pensions, health carebenefits, and other benefits.
Earnings are equal to total employee compen-sation, including the employer share of payrolltaxes and employer contributions to privatepension and welfare plans, plus self-employ-ment income. Employee compensation comesdirectly from the national accounts; self-employ-ment income was estimated based on nationalaccounts employment data. Asset income
includes all financial returns to households,except returns to funded private pension plans.It is equal to household property and entre-preneurial income in the national accounts,minus withdrawals from private quasi-corpo-rate enterprises and an estimate of privatepension benefits. This estimate is based onOECD and national sources (see below). Asmall category of miscellaneous private trans-fers is included in asset income.
The totals for public benefits come directly fromthe OECD Social Expenditure Database. Publicpensions equal old-age cash benefits plus survivors benefits and include governmentemployee pensions and special early retire-ment programs. Health care benefits are equalto OECD’s “health” category, plus one-half of“services for elderly and disabled people.” One-half is our rough estimate of the share of thisbenefit category that constitutes spending onnursing homes, home care, and other long-termcustodial services. Other benefits include allother programs, from unemployment insuranceand disability benefits to housing subsidiesand food stamps.
b. Income by Age The model divides the totals for each source ofhousehold income between two age groups: theelderly (aged 60 and over) and the nonelderly(aged 15–59). The threshold for “elderly” wasset at age 60 rather than the conventional age65 because it is much closer to the typicalretirement age in most countries. The incomeof each age group refers to the income of allindividuals within that age group, with thesingle exception of spouses of heads of house-hold, who are always considered to be of thesame age group as the head of household.Other than in the case of spouses, income tohouseholds containing both elderly and non-elderly adults is split between the two groups.
5We rely primarily on the 1999 edition of the National Accounts, the most recent to include a complete set of householdaccounts. The data were updated to 2000 based on broad aggregates available in the 2002 National Accounts. All data ongovernment benefits are from the 2001 edition of the Social Expenditure Database.
28
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
The division of income by age was based ondata from the Luxembourg Income Study (LIS)for all countries except Japan, where we useddata from the Ministry of Health, Labor, andWelfare’s 2001 Comprehensive Survey of LivingConditions. Income ratios for the elderly andnonelderly were calculated for total income andthe following components: wages, self-employ-ment income, private pensions, asset income,public pensions, and other public benefits.Health care benefits, which are not counted inhousehold surveys, were allocated using dataon per capita spending by age from OECD’sHealth Data 2001 database.
Asset income presented a special problem.The national accounts concept of asset incomeused in the Index is broader than the house-hold survey concept. In addition to interest,dividends, and rents actually received byhouseholds, it encompasses all financialreturns that indirectly accrue to householdsin instruments such as annuities and lifeinsurance policies. The types of asset incomecounted in household surveys, moreover, aremore heavily skewed toward the elderly. Wetherefore allocated capital income as follows.We estimated, in each country, the share oftotal capital income accounted for by house-hold-survey-type asset income and allocatedit according to the household survey age-groupshares for asset income. We then allocated thebalance according to each age group’s share intotal income.
We also used LIS data to calculate income bysource for each quintile of the elderly incomedistribution. (The available Ministry of Health,Labor, and Welfare data did not allow this forJapan.) The LIS quintile data were then nor-malized to our model’s income totals. Althoughwe could not project this quintile distribution,it provides a valuable additional perspective.
c. After-tax IncomeThe model calculates the total tax burden borneby the elderly and the nonelderly in the baseyear using the same two-step methodologyemployed for household income. Aggregatedata for total taxes by type are first derivedfrom the national accounts. The totals are thenallocated to the elderly and nonelderly basedon household survey data.
All taxes are divided into three broad categories:payroll taxes, direct taxes, and indirect taxes.We were able to allocate direct taxes betweenthe elderly and nonelderly based on income taxdata from the LIS and Japan’s ComprehensiveSurvey of Living Conditions. Payroll taxes andindirect taxes were allocated based on each agegroup’s share of earnings and total income,respectively. Note that our model makes thestandard economic assumption that all taxes areultimately borne by households. We thereforegross up pretax household income by indirecttaxes and corporate taxes.
5. Pension and Health Care BenefitProjectionsExcluding the impact of growth in fundedprivate pensions, the model assumes that thecomposition of national income by type of factorincome (“asset income” from capital and “earn-ings income” from labor) remains unchanged inthe future. As populations age, the distributionof each type between age groups will of courseshift. The ratios between age groups of percapita earnings income and per capita assetincome, however, remain unchanged, as do theper capita ratios of taxes by type of income.
Within this framework, public transfers andprivate pension benefits are projected as follows.
29
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
a. Public PensionsThe CSIS projections of public pension spend-ing assume current law as of the year 2000.They are based on the latest official projectionsby the OECD, except for Japan, where theyare based on national projections.6 The OECDprojections, which omit some programs insome countries, have been normalized to OECDSocial Expenditure Database totals. Moreimportant, they have been adjusted to reflectthe CSIS historical trends demographic andeconomic scenario described above in sections1 and 2. On the demographic side, the OECDassumes rising fertility, slower growth inlongevity, and, in some countries, significantshifts in net immigration. CSIS assumes a con-tinuation of historical trends. On the economicside, the OECD assumes both falling unem-ployment and a greater rise in female laborparticipation than CSIS does. All adjustmentsto the OECD pension projections were madebased on a sensitivity analysis published alongwith the projections.
b. Health Care BenefitsWe base our health care projections on twocritical assumptions. The first is that currentper-beneficiary ratios of government healthcare benefit spending on the old to spendingon the young will remain unchanged in thefuture. This represents a compromise betweentwo competing models of aging and health:the “compression of morbidity” model, whichassumes that rising longevity will be accom-panied by a falling incidence of morbidity atolder ages, and the “failure of success” model,which assumes the opposite.
The second assumption is that rates of growthin total per capita health care spending willinitially follow their long-term historical trendin each country, but that they will graduallyconverge, by 2040, to the rate of growth in per
capita GDP plus 1 percent. Per capita GDPplus 1 percent is roughly the average historicalgrowth rate for all the developed countriesover the past twenty-five years. For severalcountries (Belgium, France, Italy, Spain, andthe United States) it represents a significantslowdown; for two countries (the Netherlandsand Sweden) it represents a significant acceler-ation. Historical growth rates were calculatedbased on data from OECD’s Health Data 2001database (see Table B).
We decided against a simple projection of his-torical growth rates for two reasons. On theone hand, health care spending growth musteventually slow in high growth countries orelse crowd all other consumption out of GDP.On the other hand, as affluence and expecta-tions rise, governments in countries that havetraditionally rationed health care are finding itharder — not easier — to control costs. Theassumption of convergence at the average growth
6The OECD public pension projections are published in Fiscal Implications of Ageing, op. cit. We do not utilize the OECD projection for Japan because it includes unlegislated “fiat” savings. Our projection is based instead on the 1999 ActuarialValuation of the Employees’ Pension Insurance and the National Pension (Japanese Ministry of Health, Labor, and Welfare; 1999).
Per Capita
Per Capita Spending Growth in
Spending Ratio: Excess of Growth
Old to Young* in Per Capita GDP**
Australia 4.1 0.9%
Belgium 3.2 1.5%
Canada 4.9 1.1%
France 3.0 1.5%
Germany 2.7 0.7%
Italy 3.2 1.4%
Japan 4.9 1.2%
Netherlands 3.9 0.5%
Spain 3.2 1.6%
Sweden 2.8 0.1%
UK 3.4 0.9%
US 7.8 2.0%
T A B L E B : Health Care Benefit Assumptions
*Ratio of public health care spending on persons aged 65and over to persons aged 0–64 in most recent year available.
**Growth rate in total per capita health care spending(1975–1998) minus the growth in per capita GDP.
30
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
rate thus seems reasonable. Other assumptions,of course, are possible. We modeled an alter-native scenario in which rates of per capitaspending growth in all countries converge toGDP per capita by 2040, rather than to GDPplus 1 percent. Under this scenario, projectedelder health care benefit spending in 2040 wasroughly 1 to 2 percent of GDP lower in everycountry.
c. Private Pensions The Index defines private pensions quite broadly.They include most types of formal private-sector retirement provision, whether they areoccupational or personal, whether they resultin a lump-sum severance payment or a lifetimeannuity, provided that the benefits are funded.The definition of funding encompasses bothexternal funding and book-reserve funding.However, we exclude all purely pay-as-you-goarrangements, such as France’s ARRCO andAGIRC. If publicly mandated (which is effec-tively the case in France), these pay-as-you-goplans are included in public pension benefits.
Remarkably, there is no standard source for dataon private pension provision in the developedcountries. We base our estimates of base-yearbenefits on OECD data on private social wel-fare expenditures, adjusted to include largeomissions in Japan and small omissions inFrance and Spain.7 Our projections of futurebenefits build on the base-year numbers asfollows:
(1) We make a “cohort adjustment” to reflectthe fact that, even apart from policy changes,rates of participation in some countries arerising rapidly among younger workers. Theadjustment affects two countries: Italy andSpain. (2) We make an “earnings maturationadjustment” to take into account the fact that,even assuming fixed participation rates, currentper capita benefit levels do not reflect ultimate
benefit levels because the average current retireeis receiving a benefit based on less than a fullcareer. The adjustment affects all countries.(3) We make a “policy adjustment” to take intoaccount important recent pension reforms infour countries: Australia (the “Super”), Germany(the “Riester” accounts), Sweden (2.5-percent-of-payroll personal accounts), and the UK(miscellaneous efforts to expand private pen-sion coverage). (4) Finally, we make a “demo-graphic adjustment” to reflect the changing agecomposition of the population. This adjustmentaffects all countries.
The model assumes that increases in pensionsavings will be partially offset by declines inother forms of household savings. The offsetis assumed to be one-third. The other two-thirds of new pension savings will result innew national savings and new GDP. Since wedo not use a general equilibrium model, how-ever, we assume no further impact on nationalaccounts or factor prices. One way to thinkabout it is to imagine that all of the extranational savings will flow abroad, mainly to thedeveloping countries, where it will generateincome from foreign assets. A number of econ-omists have hypothesized that a large capitalflow from the developed to the developingcountries would be an expected consequenceof global aging.
6. Fiscal ScenarioAs already explained, the model assumes thateach country pursues a long-term policy of “debtneutrality.” In the first year of the projection,each country is assumed to move to a generalgovernment deficit which, when continuedunchanged as a share of GDP in all future years,will keep net government debt unchanged as a share of GDP at its 2001 level. Table Cshows the budget balance that achieves thisresult. The calculations assume an inflation rateof 2 percent and a long-term real interest rate
7Willem Adema, Net Social Welfare Expenditure, 2nd edition, in Labor Market and Social Policy Occasional Papers no. 52(OECD; 2001).
31
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
equal, in each country, to real GDP growthplus 1 percent.
Once debt neutrality is achieved, the modelassumes that nonbenefit spending will remainconstant as a share of GDP and that taxes willbe raised (or lowered) in each future year inaccordance with the projected change in bene-fit spending. In every country in almost everyyear, of course, this means that taxes must beraised. This “rising tax” assumption is onlyrelaxed for two of the fiscal-room indicators,where the object is to assess alternative meansof paying for the growth in benefit costs.
In apportioning the future tax changes betweenpayroll taxes, direct taxes, and indirect taxes,the model follows a few simple rules. We firstassume that payroll taxes will be raised suchthat they pay for the same proportion of totalpublic transfers in the future that they do today.Additional taxes are then divided between directand indirect taxes in proportion to their sharesin total taxation today. We considered pro-jecting each country’s specific tax rules, butdeferred any attempt to a subsequent editionof the Index. The reason is that the task wouldbe very complex. Consider that just one ben-efit program in the United States, Medicare,is financed by a combination of payroll taxes,user premiums, and general revenues — all ofwhich are projected to grow at different ratesover time.
7. Indicator DefinitionsBenefit Level: Total public benefits to theelderly in 2040 as a percent of GDP.
Benefit Growth: The change from 2000 to2040 in public benefits to the elderly as apercent of GDP.
Net Transfer: Public benefits to the elderly in 2040 as a percent of the income of thenonelderly. Both benefits and income aremeasured after taxes. Benefits to the elderly
are measured “net” — that is, they are reducedby an estimate of the share of benefits paid forby taxes on the elderly themselves. The incomeof the nonelderly is measured after all rest-of-government taxes — that is, after taking intoaccount all taxes on the nonelderly exceptthose required to pay for net benefits to theelderly. The indicator assumes that publicold-age benefits are first paid for by payrolltaxes, then, to the extent that payroll taxes areinsufficient, that they are paid for by direct andindirect taxes in proportion to their shares intotal taxation.
Tax Room: Total taxes in 2040 at all levels ofgovernment as a percent of GDP.
Budget Room: Total benefits to the elderlyas a percent of total government outlays in2040, assuming the special “budget room”scenario. Under this scenario, the rising costof old-age benefits is paid for by cannibaliz-ing other spending. Except for the initialdebt-neutrality adjustment, this indicatorassumes no change in taxes.
Long-term Net Government Real GDP
Budget Balance Debt in 2001 Growth Rate:
(Percent of GDP) (Percent of GDP) 2000–50
Australia -0.4% 10.8% 2.0%
Belgium -3.3% 98.9% 1.3%
Canada -1.9% 53.0% 1.6%
France -1.4% 42.1% 1.3%
Germany -1.3% 43.5% 0.9%
Italy -2.5% 96.5% 0.6%
Japan -1.5% 58.5% 0.6%
Netherlands -1.4% 42.1% 1.3%
Spain -1.0% 38.2% 0.7%
Sweden 0.0% 1.0% 1.1%
UK -1.1% 30.9% 1.5%
US -1.8% 41.9% 2.3%
T A B L E C : Long-term Budget Balance under Debt Neutrality*
*Assumes productivity growth rate of 1.75 percent, inflation rate of 2 percent, and real interest rate equal to GDP growth plus 1 percent.
32
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
Debt Room: Year that the net governmentdebt reaches 150 percent of GDP, assumingthe special “debt room” scenario. Under thisscenario, the rising cost of old-age benefits ispaid for by government borrowing. Exceptfor the initial debt-neutrality adjustment, thisindicator assumes no change in taxes. The cal-culations assume an inflation rate of 2 percentand a long-term real interest rate equal, in eachcountry, to real GDP growth plus 1 percent.
Benefit Share: Public benefits to the elderlyas a share of the after-tax income of the elderlyin 2040. Benefits here exclude governmenthealth care benefits. Calculating this indicatorrequired additional assumptions about taxincidence. We assume that payroll taxes areborne by labor income and that direct andindirect taxes are borne proportionally by allother types of income. Clearly, reality is morecomplicated. In many countries, income taxesfall more heavily on labor than on capital.Pensions and other cash benefits are sometimestax-favored or even entirely tax-exempt. Infuture editions of the Index, we may try tomodel some of this institutional complexity.
Family Ties: Percent of the elderly living withtheir adult children aged 18 or older, whetherin the household of the elderly person or thehousehold of the child. As throughout theIndex, nonelderly spouses of the elderly arecounted as elderly. The data on living arrange-ments were derived from LIS, except for Japan,where they are from the Ministry of Health,Labor, and Welfare, and Sweden, where theyare from the OECD. Data refer to the mostrecent year available, generally in the mid-1990s.
Poverty Impact: Percent of the elderly pushedinto poverty by a 10 percent cut in publicbenefits. Poverty is defined as 50 percent ofthe median for all households in each country,the standard international definition. Unlikethe other indicators, elderly income hereincludes the income of nonelderly household
members. It excludes, however, governmenthealth care benefits. For all countries exceptJapan, the indicator was calculated using LISdata. For Japan, it was calculated using datafrom the Ministry of Health, Labor, and Welfare.Data refer to the most recent year available,generally in the mid-1990s.
Affluence Level: Per capita ratio of the after-tax income of the elderly to the after-tax incomeof the nonelderly in 2040. Income includes theper capita cash value of government healthcare benefits.
Affluence Growth: Percentage change in theaffluence-level indicator between 2000 and2040.
8. Category and Overall RankingsAs already explained, the Index is constructedas follows. For each of the indicators, wegenerate an indicator ranking, from one (best)to twelve (worst). We also transform the indi-cator results into an index and generate anindex score for each country. For each indicator,the mean result is set to an index value of 50;results that lie above and below the mean byone standard deviation are set, respectively,to index values of 100 and zero. The indexscores thus preserve the indicator rankingswhile also reflecting the relative distance ofeach ranked country, positively or negatively,from the “center of the pack.”
For each of the four categories, a category scoreis then calculated as an average of the indica-tor index scores. The category score determinesthe overall category rankings. Finally, the cat-egory scores themselves are averaged as fol-lows: A weight of one-third is given to thefirst public-burden category, one-third to thesecond fiscal-room category, and one-third tothe third and fourth “policy-climate” categoriescombined. A country’s final combined averagefor the four categories determines its rankingin the overall Index.
33
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
T A B L E 1 : Aged Dependency Ratio*
*Ratio of elderly adults aged 60 and over to working-age adults aged 15–59.
2000 2010 2020 2030 2040
Australia 0.26 0.32 0.41 0.50 0.56
Belgium 0.37 0.41 0.51 0.64 0.71
Canada 0.26 0.33 0.46 0.58 0.63
France 0.34 0.40 0.50 0.61 0.69
Germany 0.38 0.42 0.51 0.69 0.72
Italy 0.40 0.47 0.57 0.79 1.03
Japan 0.39 0.55 0.66 0.78 1.00
Netherlands 0.29 0.36 0.47 0.64 0.69
Spain 0.35 0.40 0.49 0.69 0.99
Sweden 0.38 0.46 0.55 0.67 0.74
UK 0.34 0.39 0.47 0.61 0.65
US 0.26 0.30 0.39 0.46 0.47
DATA APPENDIX
34
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
T A B L E 2 : Public Benefits to the Elderly (Aged 60 and Over), as a Percent of GDP
2000 2010 2020 2030 2040
Australia
Total Benefits 9.0% 10.2% 12.7% 14.9% 16.6%
Public Pensions 4.2% 4.3% 5.2% 5.8% 6.1%
Health Benefits 2.7% 3.4% 4.6% 5.9% 7.0%
Other Benefits 2.2% 2.5% 2.9% 3.2% 3.4%
Belgium
Total Benefits 12.7% 13.9% 17.3% 21.9% 24.5%
Public Pensions 8.8% 9.1% 11.0% 13.7% 15.0%
Health Benefits 3.0% 3.7% 5.0% 6.6% 7.8%
Other Benefits 1.0% 1.1% 1.3% 1.6% 1.7%
Canada
Total Benefits 8.9% 11.2% 15.5% 20.2% 22.9%
Public Pensions 4.9% 6.0% 8.0% 10.5% 11.7%
Health Benefits 3.3% 4.4% 6.4% 8.3% 9.8%
Other Benefits 0.7% 0.8% 1.1% 1.3% 1.4%
France
Total Benefits 15.8% 18.4% 22.9% 26.9% 29.3%
Public Pensions 11.3% 12.8% 15.5% 17.5% 18.2%
Health Benefits 3.3% 4.3% 5.9% 7.6% 9.2%
Other Benefits 1.1% 1.3% 1.5% 1.8% 2.0%
Germany
Total Benefits 15.1% 15.4% 18.1% 23.2% 25.5%
Public Pensions 10.3% 9.9% 11.2% 14.2% 15.4%
Health Benefits 3.8% 4.4% 5.5% 7.3% 8.4%
Other Benefits 1.0% 1.1% 1.3% 1.7% 1.8%
Italy
Total Benefits 17.3% 19.4% 22.8% 28.0% 32.0%
Public Pensions 12.1% 13.1% 15.2% 18.2% 20.0%
Health Benefits 3.0% 3.9% 5.1% 7.0% 9.0%
Other Benefits 2.2% 2.4% 2.5% 2.8% 3.0%
2000 2010 2020 2030 2040
Japan
Total Benefits 11.8% 16.4% 19.6% 22.1% 27.0%
Public Pensions 6.9% 9.7% 11.3% 12.1% 14.5%
Health Benefits 3.8% 5.4% 6.9% 8.6% 10.9%
Other Benefits 1.1% 1.3% 1.4% 1.5% 1.6%
Netherlands
Total Benefits 11.7% 14.1% 18.1% 22.8% 25.6%
Public Pensions 6.5% 7.7% 9.9% 12.3% 13.8%
Health Benefits 3.0% 3.9% 5.1% 6.7% 7.8%
Other Benefits 2.2% 2.6% 3.1% 3.8% 4.0%
Spain
Total Benefits 12.6% 13.8% 17.0% 23.9% 33.1%
Public Pensions 7.7% 7.9% 9.7% 14.1% 20.3%
Health Benefits 2.7% 3.5% 4.7% 6.7% 9.3%
Other Benefits 2.2% 2.3% 2.6% 3.1% 3.6%
Sweden
Total Benefits 13.3% 15.2% 17.8% 20.9% 23.0%
Public Pensions 7.3% 8.0% 9.3% 10.6% 11.3%
Health Benefits 3.9% 4.7% 5.6% 6.8% 8.0%
Other Benefits 2.1% 2.5% 2.9% 3.4% 3.6%
UK
Total Benefits 12.1% 12.7% 13.9% 16.6% 17.6%
Public Pensions 6.5% 6.2% 6.1% 6.6% 6.5%
Health Benefits 2.8% 3.4% 4.3% 5.7% 6.6%
Other Benefits 2.9% 3.2% 3.6% 4.4% 4.6%
US
Total Benefits 9.4% 11.0% 15.2% 18.8% 20.3%
Public Pensions 5.2% 5.5% 7.4% 8.8% 9.1%
Health Benefits 3.7% 4.9% 7.1% 9.2% 10.5%
Other Benefits 0.5% 0.5% 0.7% 0.8% 0.8%
*Figures are unweighted averages.
All Countries* 2000 2010 2020 2030 2040
Total Benefits 12.5% 14.3% 17.6% 21.7% 24.8%
Public Pensions 7.6% 8.4% 10.0% 12.0% 13.5%
Health Benefits 3.3% 4.2% 5.5% 7.2% 8.7%
Other Benefits 1.6% 1.8% 2.1% 2.4% 2.6%
35
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
T A B L E 5 : Public Benefits to the Elderly (Aged 60 and Over) as a Percent ofTotal Government Outlays, AssumingCuts in Other Spending Pay for AllGrowth in Public Benefits
2000 2010 2020 2030 2040
Australia 27.2% 29.8% 34.8% 38.7% 41.2%
Belgium 27.5% 28.2% 33.7% 41.1% 44.6%
Canada 23.4% 27.1% 35.6% 44.4% 48.6%
France 32.5% 37.0% 44.3% 50.1% 52.9%
Germany 32.9% 33.3% 37.4% 46.2% 49.0%
Italy 37.9% 41.0% 46.1% 54.5% 60.3%
Japan 31.9% 46.5% 52.6% 56.3% 65.8%
Netherlands 28.0% 32.4% 39.5% 47.8% 51.7%
Spain 32.7% 34.4% 40.3% 54.1% 72.1%
Sweden 25.3% 27.2% 30.6% 34.7% 37.2%
UK 31.6% 31.5% 32.8% 37.4% 38.1%
US 30.9% 33.7% 43.8% 51.1% 52.9%
T A B L E 3 : Net Public Benefits to the Elderly, as a Percent of After-tax Nonelderly Income*
*Public benefits to the elderly (aged 60 and over), netof taxes, as a percent of the income of the nonelderly(aged 15–59), measured after taxes for the rest ofgovernment but before taxes for benefits to the elderly.
2000 2010 2020 2030 2040
Australia 12.4% 13.9% 17.2% 19.9% 22.1%
Belgium 17.2% 18.1% 22.5% 28.2% 31.5%
Canada 12.9% 15.4% 20.8% 26.5% 29.8%
France 20.3% 23.4% 28.9% 33.4% 36.1%
Germany 20.8% 21.4% 25.1% 32.2% 35.4%
Italy 21.6% 23.8% 27.6% 33.2% 37.2%
Japan 15.4% 22.3% 26.5% 29.8% 35.9%
Netherlands 16.5% 19.6% 25.0% 31.5% 35.3%
Spain 17.2% 18.5% 22.7% 31.4% 42.7%
Sweden 17.7% 19.2% 22.0% 25.1% 27.3%
UK 14.0% 14.4% 15.7% 18.3% 19.3%
US 12.8% 14.7% 20.2% 24.8% 26.8%
T A B L E 4 : Total Taxes as a Percent of GDP, Assuming Tax Hikes Pay for All Growth in Public Benefits
2000 2010 2020 2030 2040
Australia 32.5% 33.3% 35.5% 37.4% 39.2%
Belgium 49.5% 49.1% 52.5% 57.0% 59.8%
Canada 38.4% 38.6% 42.7% 47.3% 50.2%
France 48.3% 51.2% 55.7% 59.7% 62.2%
Germany 44.4% 45.4% 47.9% 52.6% 55.2%
Italy 44.0% 45.4% 48.6% 53.1% 56.5%
Japan 28.9% 36.0% 39.3% 41.8% 46.5%
Netherlands 46.9% 48.1% 51.5% 55.5% 58.4%
Spain 38.2% 38.9% 42.1% 48.7% 57.2%
Sweden 57.4% 56.3% 58.4% 61.0% 63.0%
UK 38.4% 37.8% 38.7% 40.7% 41.7%
US 33.2% 33.9% 38.3% 42.1% 44.1%
36
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
T A B L E 6 : After-tax Income of the Elderly (Aged 60 and Over) by Source, as a Percent of Total After-tax Elderly Income*
2000 2010 2020 2030 2040
Australia
Earnings 23.3% 23.7% 23.5% 23.3% 22.9%
Private Pensions 13.2% 17.7% 23.1% 28.4% 33.1%
Asset Income 20.2% 18.8% 16.0% 13.5% 11.3%
Public Benefits 43.4% 39.9% 37.4% 34.8% 32.7%
Belgium
Earnings 10.9% 11.4% 11.1% 10.3% 9.9%
Private Pensions 8.5% 9.2% 10.1% 10.9% 11.5%
Asset Income 25.9% 26.1% 24.9% 23.2% 22.5%
Public Benefits 54.7% 53.3% 53.9% 55.6% 56.2%
Canada
Earnings 17.6% 17.7% 17.6% 17.0% 16.4%
Private Pensions 19.0% 20.7% 23.6% 25.5% 26.6%
Asset Income 21.3% 19.3% 15.8% 12.7% 11.4%
Public Benefits 42.2% 42.4% 42.9% 44.8% 45.5%
France
Earnings 8.0% 7.7% 7.2% 6.7% 6.5%
Private Pensions 0.5% 0.6% 0.7% 0.8% 0.8%
Asset Income 24.1% 23.8% 23.0% 22.8% 23.1%
Public Benefits 67.3% 67.9% 69.2% 69.7% 69.5%
Germany
Earnings 13.5% 14.1% 13.8% 12.9% 11.9%
Private Pensions 4.6% 6.1% 7.8% 9.9% 11.1%
Asset Income 20.5% 21.3% 20.8% 19.6% 18.6%
Public Benefits 61.4% 58.6% 57.7% 57.6% 58.4%
Italy
Earnings 12.2% 12.5% 12.2% 12.3% 12.4%
Private Pensions 4.1% 4.9% 5.7% 7.4% 9.3%
Asset Income 24.9% 24.9% 24.1% 23.3% 22.7%
Public Benefits 58.8% 57.8% 58.0% 57.0% 55.6%
2000 2010 2020 2030 2040
Japan
Earnings 40.5% 37.2% 35.6% 35.0% 32.4%
Private Pensions 12.5% 14.7% 16.3% 18.3% 21.1%
Asset Income 12.0% 10.8% 10.0% 9.4% 8.1%
Public Benefits 34.9% 37.3% 38.2% 37.3% 38.4%
Netherlands
Earnings 7.4% 7.4% 7.0% 6.6% 6.1%
Private Pensions 18.7% 20.1% 21.6% 23.9% 24.4%
Asset Income 19.9% 18.9% 17.1% 15.3% 14.3%
Public Benefits 54.0% 53.7% 54.4% 54.2% 55.2%
Spain
Earnings 9.5% 9.9% 9.4% 8.0% 6.3%
Private Pensions 1.2% 2.5% 4.1% 5.8% 8.0%
Asset Income 25.2% 25.5% 24.2% 21.5% 18.7%
Public Benefits 64.0% 62.1% 62.3% 64.7% 66.9%
Sweden
Earnings 19.7% 20.6% 20.2% 19.9% 19.4%
Private Pensions 11.0% 12.5% 14.1% 16.1% 17.6%
Asset Income 12.8% 11.9% 10.6% 9.2% 8.3%
Public Benefits 56.5% 55.0% 55.1% 54.8% 54.6%
UK
Earnings 16.3% 17.2% 18.1% 18.9% 19.0%
Private Pensions 14.0% 16.5% 19.8% 24.1% 26.4%
Asset Income 19.4% 19.2% 18.4% 16.1% 15.2%
Public Benefits 50.3% 47.0% 43.7% 40.9% 39.3%
US
Earnings 25.5% 25.9% 24.8% 23.8% 23.1%
Private Pensions 16.7% 18.2% 20.5% 22.3% 23.0%
Asset Income 22.9% 22.2% 19.7% 18.0% 17.6%
Public Benefits 34.9% 33.9% 35.3% 36.1% 36.6%
*Income excludes government health care benefits.
37
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
2000 2010 2020 2030 2040
Australia
Without 43.4% 39.9% 37.4% 34.8% 32.7%
With 52.8% 50.8% 49.4% 48.1% 47.2%
Belgium
Without 54.7% 53.3% 53.9% 55.6% 56.2%
With 62.0% 61.9% 63.4% 65.3% 66.6%
Canada
Without 42.2% 42.4% 42.9% 44.8% 45.5%
With 55.4% 56.6% 58.1% 60.4% 62.0%
France
Without 67.3% 67.9% 69.2% 69.7% 69.5%
With 72.9% 74.2% 76.0% 77.2% 78.0%
Germany
Without 61.4% 58.6% 57.7% 57.6% 58.4%
With 68.3% 66.9% 66.7% 67.1% 68.3%
Italy
Without 58.8% 57.8% 58.0% 57.0% 55.6%
With 63.8% 63.8% 64.8% 65.1% 65.0%
2000 2010 2020 2030 2040
Japan
Without 34.9% 37.3% 38.2% 37.3% 38.4%
With 44.7% 47.7% 49.6% 50.3% 52.3%
Netherlands
Without 54.0% 53.7% 54.4% 54.2% 55.2%
With 62.0% 62.2% 63.2% 63.5% 64.8%
Spain
Without 64.0% 62.1% 62.3% 64.7% 66.9%
With 70.1% 69.6% 70.6% 72.9% 74.9%
Sweden
Without 56.5% 55.0% 55.1% 54.8% 54.6%
With 66.9% 66.1% 66.5% 66.8% 67.6%
UK
Without 50.3% 47.0% 43.7% 40.9% 39.3%
With 57.3% 55.4% 53.6% 52.2% 51.9%
US
Without 34.9% 33.9% 35.3% 36.1% 36.6%
With 48.2% 49.6% 52.4% 54.7% 56.6%
T A B L E 7 : Public Benefits to the Elderly (Aged 60 and Over), as a Percent of After-tax Elderly Income: With and Without Health Care Benefits
T A B L E 8 : After-tax Income of the Elderly (Aged 60 and Over) by Source and Quintile in 2000, as a Percent of Total After-tax Elderly Income*
1st 2nd 3rd 4th 5th
Australia
Earnings 0.1% 1.8% 2.3% 15.3% 43.6%
Private Pensions 1.3% 4.6% 6.2% 20.5% 15.4%
Asset Income 10.4% 11.7% 9.9% 21.1% 26.6%
Public Benefits 88.1% 81.9% 81.6% 43.1% 14.3%
Belgium
Earnings 0.5% 0.7% 1.5% 5.5% 19.5%
Private Pensions 0.0% 0.7% 5.1% 11.7% 10.5%
Asset Income 13.8% 11.2% 18.4% 21.8% 34.6%
Public Benefits 85.7% 87.4% 75.0% 61.0% 35.4%
Canada
Earnings 0.4% 2.6% 6.6% 12.8% 30.6%
Private Pensions 2.0% 9.6% 16.0% 25.6% 21.6%
Asset Income 7.6% 11.9% 15.5% 19.7% 28.9%
Public Benefits 90.0% 75.9% 62.0% 41.9% 18.8%
France
Earnings 1.2% 1.8% 3.4% 7.2% 12.5%
Private Pensions 0.1% 0.3% 0.4% 0.7% 0.6%
Asset Income 13.6% 16.9% 18.0% 20.9% 32.0%
Public Benefits 85.0% 81.0% 78.2% 71.2% 54.8%
Germany
Earnings 1.1% 2.2% 3.0% 7.1% 26.4%
Private Pensions 3.9% 2.6% 2.5% 3.7% 6.7%
Asset Income 3.7% 5.5% 10.2% 21.4% 30.1%
Public Benefits 91.4% 89.7% 84.3% 67.8% 36.9%
Italy
Earnings 0.5% 1.6% 4.3% 10.7% 18.9%
Private Pensions 0.6% 2.2% 3.1% 2.8% 5.8%
Asset Income 8.1% 8.5% 10.0% 18.9% 38.4%
Public Benefits 90.7% 87.7% 82.7% 67.5% 36.8%
1st 2nd 3rd 4th 5th
Japan
Earnings NA NA NA NA NA
Private Pensions NA NA NA NA NA
Asset Income NA NA NA NA NA
Public Benefits NA NA NA NA NA
Netherlands
Earnings 0.6% 0.5% 1.5% 6.2% 14.1%
Private Pensions 2.9% 6.4% 12.8% 22.6% 26.3%
Asset Income 3.0% 6.6% 11.5% 17.3% 33.0%
Public Benefits 93.4% 86.5% 74.2% 53.8% 26.7%
Spain
Earnings 0.8% 1.7% 3.7% 9.5% 16.1%
Private Pensions 0.3% 0.5% 0.8% 1.4% 1.7%
Asset Income 11.5% 15.1% 18.8% 25.9% 33.1%
Public Benefits 87.4% 82.7% 76.7% 63.1% 49.2%
Sweden
Earnings 1.5% 1.8% 9.8% 13.3% 36.4%
Private Pensions 3.4% 6.2% 8.8% 11.4% 14.7%
Asset Income 11.7% 9.0% 11.1% 11.9% 15.2%
Public Benefits 83.4% 82.9% 70.2% 63.5% 33.6%
UK
Earnings 0.3% 1.3% 3.1% 12.7% 31.0%
Private Pensions 3.0% 7.0% 11.3% 18.6% 17.1%
Asset Income 5.2% 6.9% 10.8% 18.0% 30.3%
Public Benefits 91.5% 84.8% 74.8% 50.7% 21.6%
US
Earnings 2.4% 5.7% 10.6% 22.2% 36.0%
Private Pensions 6.2% 12.6% 20.9% 22.4% 14.6%
Asset Income 6.1% 11.2% 14.3% 17.9% 30.9%
Public Benefits 85.3% 70.5% 54.2% 37.5% 18.5%
38
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
*Income excludes government health care benefits.
39
Th
e 20
03 A
gin
g V
uln
erab
ility
Ind
ex
CS
IS a
nd
Wat
son
Wya
tt W
orl
dw
ide
T A B L E 9 : Private Pension Benefits, as a Percentof GDP*
*Includes all employer pensions, personal pensions,and severance pay schemes.
*Elderly are aged 60 and over; nonelderly are aged15–59; income includes the cash value of governmenthealth care benefits.
2000 2010 2020 2030 2040
Australia 2.5% 3.8% 6.0% 8.6% 11.1%
Belgium 1.7% 1.9% 2.5% 3.2% 3.6%
Canada 3.2% 4.0% 5.8% 7.6% 8.5%
France 0.1% 0.1% 0.2% 0.2% 0.3%
Germany 0.9% 1.2% 1.8% 2.9% 3.4%
Italy 1.3% 1.7% 2.1% 3.2% 4.3%
Japan 3.1% 4.6% 5.7% 7.0% 9.2%
Netherlands 3.3% 4.1% 5.4% 7.4% 8.2%
Spain 0.2% 0.5% 0.9% 1.7% 3.1%
Sweden 2.2% 2.7% 3.5% 4.5% 5.3%
UK 3.6% 4.3% 5.6% 7.8% 8.9%
US 3.6% 4.2% 5.8% 7.0% 7.4%
T A B L E 1 0 : Per Capita Ratio of the After-tax Income of the Elderly to the Nonelderly*
2000 2010 2020 2030 2040
Australia 1.30 1.29 1.33 1.38 1.45
Belgium 1.07 1.04 1.07 1.13 1.19
Canada 1.32 1.33 1.36 1.45 1.54
France 1.29 1.31 1.35 1.36 1.36
Germany 1.25 1.22 1.25 1.33 1.44
Italy 1.32 1.26 1.25 1.19 1.15
Japan 1.07 1.15 1.21 1.24 1.34
Netherlands 1.32 1.31 1.36 1.41 1.51
Spain 0.98 0.95 0.97 1.05 1.18
Sweden 1.01 0.99 1.00 1.01 1.04
UK 1.20 1.15 1.11 1.09 1.10
US 1.43 1.46 1.56 1.67 1.77
40
CS
IS an
d W
atson
Wyatt W
orld
wid
e
Th
e 2003 Ag
ing
Vu
lnerab
ility Ind
ex
Richard Jackson writes on public policyissues arising from the aging of America’s andthe world’s population. He is currently anadjunct fellow at CSIS, where he is affiliatedwith the Global Aging Initiative, an adjunctfellow at the Hudson Institute, and a senioradvisor to the Concord Coalition. He is theauthor of numerous policy studies, includingThe Global Retirement Crisis (CSIS and Citigroup;2002) and Germany and the Challenge of GlobalAging (CSIS and Nationwide Global; 2003).Jackson regularly speaks on long-term demo-graphic and economic issues and is oftenquoted in the press. He holds a B.A. in classicsfrom SUNY at Albany and a Ph.D. in historyfrom Yale University. He lives in Alexandria,Virginia, with his wife Perrine and their sons,Benjamin and Brian.
Neil Howe is a historian, economist, anddemographer who writes and speaks frequentlyon long-term fiscal policy and generations inAmerican history. He is a senior advisor tothe Concord Coalition, senior policy advisorto the Blackstone Group, and on the editorialboard of the Public Policy and Aging Report.His coauthored books include On BorrowedTime (1988), Generations (1991), 13th Gen(1993), The Fourth Turning (1997), andMillennials Rising (2000). He is a cofounderof LifeCourse Associates, a marketing, HR,and strategic planning consultancy servingcorporate, government, and nonprofit clients.He holds graduate degrees in history andeconomics from Yale University. He lives inGreat Falls, Virginia, with his wife Simona andtwo children, Giorgia and Nathaniel.
ABOUT THE AUTHORS
ABOUT CS I S
The Center for Strategic and International Studies (CSIS)
is a public policy research institution dedicated to analysis
and policy impact. CSIS is the only institution of its kind
that maintains resident experts on all the world's major
geographical regions. It also covers key functional areas,
such as international finance, U.S. trade and economic
policy, national and international security issues, energy,
and telecommunications.
Founded in 1962 and located in Washington, D.C., CSIS is
a private, tax-exempt institution. Its research is nonpartisan
and nonproprietary. Sam Nunn chairs its Board of Trustees.
The Center's staff of 90 policy experts, 80 support staff,
and 70 interns is committed to generating strategic analysis,
exploring contingencies, analyzing policy options, and
making policy recommendations. Its goal is to inform and
shape selected policy decisions in government and the pri-
vate sector by providing long-range, anticipatory, and inte-
grated thinking over a wide range of policy issues.
ABOUT WATSON WYATT WORLDWIDE
Watson Wyatt is a global consulting firm focused on
human capital and financial management. We specialize
in three areas: employee benefits, technology solutions,
and human capital strategies.
We combine human capital and financial expertise to
deliver business solutions that drive shareholder value.
Our consulting is grounded in the most extensive
research in the business. This translates into unique
perspectives and forward-thinking solutions.
Watson Wyatt has more than 6,300 associates in 30
countries. Corporate offices are in Reigate, England,
and Washington, D.C., USA.
CENTER FOR STRATEGIC AND
INTERNATIONAL STUDIES
1800 K Street, NW
Suite 400
Washington, D.C., 20006
202/887-0200
WATSON WYATT CORPORATE OFFICES
Washington, D.C., U.S.A.
1/800-388-9868
Reigate, England
44-1737-241144
WWW.CSIS.ORG
WWW.WATSONWYATT.COM
Copyright © 2003
The Center for Strategic and
International Studies and
Watson Wyatt Worldwide.
All rights reserved. Catalog #W-637