Welfare and Generational Equity in Sustainable Unfunded Pension Systems

1
Welfare and Generational Equity in Sustainable
Unfunded Pension Systems

• Alan J. Auerbach
• Ronald Lee
• UC Berkeley

2
Overview

• Need to reform existing PAYGO public pension
  schemes, due to lack of stability and
  transparency
• Shift to funded systems confronts economic and
  political problems of transition
• An alternative Notional Defined Contribution
  (NDC) plans PAYGO systems, but with some
  potential advantages (transparency, stability)

3
NDC Systems

• Two phases accumulation and retirement
• Accumulation phase accumulate an imaginary
  stock of Notional Pension Wealth (NPW) based on
  annual contributions
• Retirement phase annuitize NPW using same
  assumed rate of return, based on life table

4
NDC Systems

• How is assumed rate of return determined?
• This is the key decision with respect to
  potential stability
• Sweden bases return on wage growth (g) adjusted
  for annual mortality probability
• Could base return on growth of covered wages (ng)

5
Previous Auerbach-Lee Paper

• Studied the Swedish NDC plan, in terms of
  stability
• With stochastic wages, interest rates, birth
  rates and mortality (based on US data), how
  likely is debt to explode over time?
• Swedish system is stable downward (no excessive
  debt build-up), but only because it also includes
  a balancing mechanism that reduces rate of
  return when trouble near but doesnt avoid asset
  accumulation

6
Previous Auerbach-Lee Paper

• Could avoid accumulation of assets (and pay a
  higher average rate of return) by making balance
  mechanism symmetric, also raising growth rate of
  accounts and benefits when system assets too high
• Could make Swedish system more stable by basing
  growth rate of accounts and annuities on growth
  rate of wage bill (ng) rather than growth rate
  of wage rate (g)

7
This Paper

• Look at risk-sharing and welfare properties of
  different types of fiscally stable systems
• Actual Swedish system and symmetric variants of
  it from our earlier paper
• New German system
• Versions of the US system forced to be stable by
  annual tax or benefit adjustments

8
The Systems

• All systems based on simplified US OASI system
  for a representative individual per cohort
• 10.6 percent payroll tax
• work until age 67, with retirement at 67
• US system variants, all PAYGO, with tax profiles
  based on US data, and benefit profiles generated
  by simplified version of benefit formula assuming
  retirement at 67

9
The Systems

• Three US system variants
• Benefit Adjust raise or lower benefits each
  year so that taxes benefits
• Tax Adjust raise or lower taxes each year so
  that taxes benefits scale so that average tax
  rate 10.6 percent (since actual US system not
  viable)
• 50-50 Adjust divide annual adjustment equally
  between taxes and benefits

10
The Systems

• Swedish system variants
• All with tax rate fixed at 10.6 percent
• Actual Swedish system
• Notional Pension Wealth accumulates at rate g and
  is annuitized at age 67, with annuity rate of
  return also based on g
• Brake mechanism that temporarily lowers benefits
  by setting gross return to (1g)b when a measure
  of assets/liabilities, b, falls below 1

11
The Systems

• Three Swedish system variants
• All with symmetric brake
• Two based on g, one based on ng
• One with full brake, reducing gross rate of
  return by a factor (1- b) two with dampened
  brake, reducing gross rate of return by a factor
  0.5(1-b)

12
The Systems

• German system
• Strictly PAYGO
• Benefits the same for all cohorts at a given time
• Benefits grow according to

• Taxes adjusted as a residual to ensure balance
• System scaled so that taxes average 10.6 percent

13
Evaluation Criteria

• Internal Rate of Return (IRR)
• Net Present Value relative to PV of earnings
  (NPV)
• Expected Utility Approximation (EU)
• Horizontal Equity (HE)
• Social Welfare Function, taking transition
  generations into account (SWF)

14
Social Welfare Measures
15
Social Welfare Measures
16
Social Welfare Measures
17
Social Welfare Measures
18
Conclusions

• Swedish system provides most stability, but
  generally not as good as other systems under
  welfare measures
• This is particularly so when transition is taken
  into account, because the stability is provided
  by a buffer stock accumulated at the expense of
  early generations
• Basing NDC plan on g rather than ng may be
  better for welfare, even if not for stability
• Smaller fluctuations when brake is not in effects
  seem to outweigh more frequent application of the
  brake (with associated fluctuations)
• Systems that spread risk broadly over generations
  (US 50-50, NDC) do best