Transcript 184 kb PowerPoint presentation

Issues in Group vs. Individual Provision
Lawrence Thompson
Senior Fellow
The Urban Institute
1
Outline
1.
2.
Objectives of a pension system
Role and structure of different pillars in a multi-pillar
system
3. Issues affecting design of second pillar:
a) Structure and size of other components
b) Distribution of system risks
c) Degree of worker choice
d) Operating costs
4. Concluding Observations
•
2
Objectives of a National Pension System
1.
Primary Objectives:
a) Assure retired/disabled a minimum income
b) Assure minimal degree of earnings replacement
2. Secondary Objectives:
a) Increase national savings
b) Promote financial market development
c) Help restructure failing (state) enterprises
d) Minimize adverse side effects
e) Secure high returns on contributions
3
Three Common Elements (Pillars)
of a Retirement Income System
1. Redistribution program(s)
a. Defined benefit, pay-as-you-go, usually government operated
2. Mandatory earnings or contribution-related program(s)
a. Defined benefit and/or defined contribution
b. Pay-as-you-go and/or advance funded
c. Government and/or private operation
3. Voluntary programs
a. Advance funded, private operation
b. Increasingly defined contribution
c. Government oversight and subsidies
5
Issues Affecting Second Pillar Design:
Size and Structure of Other Components
1.
Level of minimum income guarantee
– Establishes floor
2. Size of total mandatory component
– Limits the size of the 2nd pillar
3. Structure of minimum income guarantee
– Affects incentive structure
– Also influences size of 2nd pillar
6
Level of Minimum:
Ratio of Minimum Pension to Average Wage
Austria
35%
Netherlands
24%
Belgium
40
Norway
34
Canada
24*
Poland
33
Chile
25
Portugal
27
Denmark
31
Spain
27
France
42
Switzerland
18
Japan
23
United Kingdom
27
Kazakhstan
25
United States
22
35 year worker, 1997
* 29%, including income tested supplement
7
Structure of Minimum:
Two Approaches
1.
Fill the Gap
– Acts as a supplement to top up a second component
benefit
– Found in Chile, Sweden, U.S.
2. Flat minimum
– Provides floor that second component is built upon
– Found in Russia, China, U.K., Netherlands, Denmark
8
Fill the Gap Minimum
(Chile, U.S.)
Minimum Pension
Preretirement Earnings
9
Fill the Gap Minimum
(Sweden)
Preretirement Earnings
10
Flat Minimum
Basic Pension
Preretirement Earnings
11
Size of 2nd component also depends on
structure of the minimum
(Fill The Gap Minimum)
Minimum Pension
Earnings Related Pension
50%
25%25%
Preretirement Earnings
12
Size of 2nd component also depends on
structure of the minimum
(Flat Minimum)
25%
Basic Pension
Preretirement Earnings
25%
13
Comparison of Approaches
Flat grant:
1. Uniform coverage in dignified manner
2. Fewer disincentives (at least from benefit side!)
3. Less room for earnings-related component
4. More budget resources
5. Easier to administer
•
Fill the gap minimum:
1. Fewer budget resources to finance
2. More room for earnings-related component
3. More difficult to administer, particularly if means-tested
4. Compliance, savings and labor market disincentives
5. Contingent liabilities
14
Issues Affecting Second Pillar Design:
Size and Distribution of System Risks
1.
Of particular concern in pension programs:
a. Few alternative income sources at point of eligibility
b. Long lead time for individual supplementation
2. Major sources of risk
a. Demographic
b. Financial
c. Political/Institutional
15
Demographic Risk
Two elements:
1. Change in mortality among aged
– Affects both funded and unfunded schemes
2. Change in birth rate (or net immigration)
– Affects unfunded schemes directly through change
in beneficiary/worker ratio
– Affects funded schemes indirectly through
changes in relative factor prices
16
Incidence of Unanticipated Demographic
Change
1.
Defined contribution scheme:
– Impact entirely on retired worker
2. Defined benefit scheme:
– Initial impact on plan sponsor (or state)
– Ultimate impact likely to be shared between
contributors and beneficiaries, at least initially
17
Financial Risk
1.
2.
Of concern mainly in funded plans*.
Two elements of financial risk:
– Change in ratio of mean returns to mean wage growth
rate
– Variation in annual mean returns and growth rates
around known mean
* cost of unfunded plans can be sensitive to changes in real wage growth rate
18
Impact of Change in Mean Return and Wage
Growth on Replacement Rate at Retirement
Average Annual Wage Increase
Average Net Investment Returns
1%
2%
3%
3%
.42
.35
.30
4%
.55
.46
.38
.32
5%
.72
.59
.49
.41
.35
.77
.63
.52
.44
.82
.67
.56
6%
7%
4%
35 year work history; 10% contribution rate; 15 year annuity
5%
20
Impact of Variations around the Mean
Analytical approach:
1. Examine impact of variation in wage growth, investment returns, and
annuity interest rate
2. Monte Carlo experiment with actual mean and standard deviation of
annual wage growth and annual average investment returns, 1953-95
3. Examine impact of four structures:
a. Traditional defined benefit -- sponsor absorbs all risks
b. TIAA model -- wages (and contributions) vary; sponsor
smoothes other variations
c. Swedish model -- wages and accumulation returns vary; sponsor
smoothes annuity interest variation
d. Chile model -- worker assumes all risks
21
Impact of Second Component Structure on
Replacement Rate Variation, U.S. Data
Mean
Standard St Dev as
Deviation % of Mean
1. Traditional Defined Benefit
(sponsor absorbs all variation)
.55
0
0%
2. TIAA Model (sponsor smoothes
investment & annuity variation)
.56
.05
9%
3. Swedish Model (sponsor
smoothes annuity variation)
.55
.21
38%
4. Chile Model (worker bears all
risks)
.57
.26
45%
Portfolio of 50% equities and 50% 10 year government bonds; actual average returns 1953-95;
35 year work history; 10% contribution rate; 15 year annuity; administrative costs 1% of assets
22
International Comparison:
Mean & Standard Deviation, Chile Model
Country
Mean
Standard
Deviation
St Dev as
% of Mean
Germany
.40
.15
37%
Japan
.25
.07
26%
United Kingdom
.38
.21
55%
United States
.57
.26
45%
“World Fund*”
.62
.25
40%
Portfolio of 50% equities and 50% 10-year government bonds;
Administrative costs of 1% (of assets) per year
*international average of returns; U.S. wage growth
23
Incidence of Financial Market Risk
1.
2.
Risk of change in mean growth or return rate:
a. Defined contribution scheme: Risk born entirely by retiring
worker
b. Defined benefit scheme:
– Initial impact on plan sponsor (or state plan)
– Ultimate impact likely to be shared between contributors and
beneficiaries
Risk associated with variance in growth rates & returns:
a. Defined benefit scheme: no risk
b. Defined contribution scheme:
– Risk born entirely by retiring worker
– Size of risk depends on plan design
25
Political/Institutional Risks
1.
2.
Unsustainable benefit promises
Decision gridlock
a. Cutting back defined benefits (France)
b. Resolving fiscal imbalances (Argentina)
3. Benefit adjustments to meet budget targets
4. Asset mismanagement
a. Political interference
b. Private mismanagement / regulatory failure
5. Underestimated transition costs
26
Structure Affects Type of Political Risk
Funded, Defined Contrib.
PAYG, DB
Public
Private
Public
Management Management
Management
Unsustainable promises

Decision Gridlock

Asset Mismanagement

Regulatory Failure

Transition cost

27
Issues Affecting 2nd Pillar Design:
Worker Choice
1.
Type of choice:
a. Portfolio composition
b. Asset manager
c. Plan managers
2. Reason for providing choice:
a. Worker preferences
b. Political insulation
c. Improve quality of service
28
Issues Affecting 2nd Pillar Design:
Administrative Costs
1.
Reduce investment returns
a. Limit (eliminate?) ability to lower contributions
through advance funding
2. Vary with:
a. Degree of individual choice
b. Degree of centralization of administration
c. Level of service offered
3. May serve as insurance against certain forms of political
risk
29
Inflation Adjusted Market Returns, 1953-1995
Equity
returns
Bond
Returns
Mixed
Portfolio
Wage
Growth
8.1%
3.8%
6.6%
5.0%
1.6%
Germany
7.4
3.9
6.3
4.8
1.5%
United Kingdom
7.8
1.8
5.6
3.6
2.0%
United States
8.2
2.2
5.6
1.0
4.6%
Japan
Gap
30
Administrative Costs Under Different Models
(Percent of Assets)
Accumulation
Total
Australia
0.5 - 1.9
1.0 - 2.4
Latin America
0.6 - 1.4
1.1 - 1.9
Poland
0.8 – 1.0
?
UK – personal pensions
1.2 – 1.3
1.7 – 1.8
1.0
1.5
0.8 – 1.9
1.3 – 2.5
– stakeholder pensions
US – 401(k) plans
– mutual funds
1.4
– TIAA
– thrift savings plan
0.3
0.1
?
Switzerland
0.4
Denmark
0.3
Sweden
0.2 – 0.4
?
31
Second Pillar Model Variation
1.
2.
3.
Defined benefit or defined contribution
Advance funded or pay-as-you-go
Centralization or decentralization of key functions
a. Collecting contributions
b. Keeping individual account records
c. Paying benefits
4. Degree of worker choice
5. Type of political risk minimized
32
Comparative Strength
of Different Funded Account Models
The approaches
1. Latin American
2. United Kingdom
3. Switzerland
4. Sweden
5. Thrift plan
6. Singapore
7. (Canada)
Political
Insulation





Worker
Choice
Low
Cost



 





35
Concluding Observations I
1.
2.
Risks in pension system:
a. All models have risks, but nature and size varies from one model
to another
b. Some forms of financial risk are unique to defined contribution
plans
Individualization of pension provision:
a. Increases financial risk (variation around the mean)
b. Shifts more of demographic and financial risks to individual
retired workers
c. Tends to increase administrative costs
d. Allows greater choice (to what end?)
37
Concluding Observations II
3.
4.
Structure of second pillar influenced by:
a. Relationship between minimum benefit and ceiling on mandated
benefit
b. Political traditions
1) Pattern of political risks in each society
2) Governance traditions
c. Preferences for:
1) Providing choice in the mandatory system
2) Spreading risks among general population
High administrative costs are:
a. Either political insurance policy or (in part) dead weight loss
b. Hard to reduce (costs are incomes)
38