Measuring Pension Wealth

1
Measuring Pension Wealth

• Prepared for Luxembourg Wealth Study
• Perugia, January 27-29, 2001
• Authors
• Agar Brugiavini (University of Venice)
• Karen Maser (Statistics Canada)
• Annika Sundén (Swedish Social Insurance Agency)

2
Our job

• Not asking for a pension wealth number in all
  Surveys
• On the other hand augmented wealth view is for
  statistical convenience because we measure
  objects that we see. But we see objects which are
  recorded transactions or recorded contracts
• Sooner or later we bump into the research
  question

3
Our job

• .Even simple analysis of cross-sectional
  distributions of wealth will prompt the issue of
  what is the use of this wealth (i.e. the reason
  for accumulation dynamic dimension).
• Around this table some seminal papers on the
  inequality versus uncertainty explanation of the
  cross-sectional distribution

4
Our job

• Economic theory would argue that contingent
  contract would remove dominated assets. Hence
  much of the measured wealth would disappear.
• Of course households do not have access to such a
  wonderful world, they use the available and known
  instruments.
• But there is a lot of hidden wealth, we mean
  wealth that is not directly recorded.
• Question where do you stop?

5
Our job

• A stongy life-cycle believer would argue that
  pension wealth is the bulk of wealth.
• Hence we argue that AT LEAST you should keep an
  eye on this augmented wealth.
• A sizeable welfare state would remove some of the
  need for accumulating wealth.
• Cross-country comparisons should take account of
  this, particularly for pensions.

6
Our job

• These arguments also point to the important
  substitutabilities of measured wealth and
  augmented wealth (see tax priviliges).
• Hence portfolio composition affected.

7
A few key questions about pension wealth

• How does it differ from other wealth?
• Earmarked for retirement
• Cannot be bequeathed
• Degree of liquidity
• Risks
• How much of household wealth is it?
• Sizable fraction of household wealth
• US almost 70 (public earnings related and
  private pensions for households 55-64)
• Canada 40 (private pensions only, family units
  55-64)
• Italy 48 (public earnings related pensions
  only)
• Without pensions, wealth picture incomplete

8
Some key questions requiring measure of pension
wealth

• Examination of retirement behaviour and labour
  supply
• Impact on other savings
• Assessment of retirement income adequacy

9
Defining pension wealththe basic formula

• For an individual, pension wealth is the present
  discounted value of future pension rights, taking
  account of mortality prospects.
• For a worker of age a, retiring at age
  , this is defined as the expected present value
  of future pension benefits
• Here S is the age of certain death,
  , with
• denoting the pure time discount factor and
• the conditional survival probability at age s for
  an individual alive at age a, and
• the pension expected at age in
  case of retirement at age h.

10
Defining pension wealth which programs are
included here?
Focus of this paper is on programs that are shaded
11
Basic types of pensions

• Defined contribution (DC)
• Pension depends on contributions and rate of
  return on those contributions
• Pension wealth account balance
• Relatively easy to value
• Defined benefit (DB)
• Benefit defined by formula involving years of
  service and frequently earnings
• e.g., 2 of earnings averaged over last 5 years
• Estimating pension wealth from these plan more
  complex requires information on both individual
  and plan

12
Issues regarding valuation of pensions

• Which unit (individual, family, household)?
• To value need information about individual need
  to collect at that level
• For analysis often better to use family or
  household because value often includes survivor
  benefits
• Whose pensions should be valued?
• Those expecting a future benefit (current or
  previous members of plan)
• Those now receiving a benefit

13
Issues regarding valuation of pensions (continued)

• Timing of valuation Current time or time of
  retirement?
• Time of retirement fully accounts for
  eligibility conditions need to make assumptions
  about retirement age
• Current time more comparable with basis of
  valuation of other assets
• Best timing may depend on research question
• Make variables available so different estimates
  can be produced

14
Issues regarding valuation of pensions (continued)

• Pension reform
• Large scale reforms have taken place in several
  countries
• Important to value benefits taking reforms into
  account
• Disability benefits
• Important to consider them for two reasons
• Part of package worker is paying for
• People receiving them can be regarded as retired

15
Issues regarding valuation of pensions (continued)

• Survivor benefits
• Many pension programs provide survivor benefits
  impacts on value of pension and on unit of
  analysis
• Need to ensure no double-counting
• Risks
• Type of plan in DC plans risk that market
  conditions may affect adequacy of benefit
• Funding not all pension programs fully funded
• Difficult to account for risk in measuring
  pension wealth should be considered in
  accompanying analysis

16
Issues regarding valuation of pensions (continued)

• Source of information
• Several can be required
• Self reports from household survey for
  information about individual (unless available
  from admin sources)
• Plan information (employer, government), for
  provisions of the pension plan/program
  (individuals often not aware of this information)
• Administrative sources in some countries this
  source can provide information about plan
  participants
• External sources for assumptions about longevity,
  price increases, interest rates, etc..
• Can affect cost and complexity of survey

17
Issues regarding valuation of pensions (continued)

• Integration of occupational and public plans
• Affects value of occupational plans
• Should be considered in valuation methodology
• Retirement age and early retirement
• Assumptions frequently required impact on value
  of benefit important for many research questions
• Variables used to generate estimate should be
  provided to allow use of different ages if
  required

18
Valuation methodology

• For DB plans
• Estimate annual pension to be paid at retirement
  
• Estimate value of benefit from retirement to
  death
• Discount value to current age
• For DC plans
• Consider contributions and rate of return
• For those currently receiving benefits
• Use reported pension to value benefit

19
Defining the data set
SOURCE HS Household survey AR Admin
records PI Plan information OS Other source NR
Not required
for occupational plans
20
Defining the data set
SOURCE HS Household survey AR Admin
records PI Plan information OS Other source NR
Not required
for occupational plans
21
Conclusions

• Recognized that pension wealth difficult to
  measure, requires assumptions be made and can add
  to survey cost
• However, measure needed to understand fiscal,
  economic and social implications of providing
  pensions
• Therefore, it is suggested that household
  surveys should be encouraged to include
  information necessary to measure pension wealth
• Variables needed to value pension wealth should
  also be made available to researchers so they can
  generate alternative estimates based on other
  assumptions
• This paper should be regarded as the first step
  in the discussion of the measurement of pension
  wealth