Summary
Lee and Skinner (1999) assess whether the aging baby boom will cause fiscal catastrophe for the federal government, focusing on Social Security and Medicare. They argue that Social Security Administration (SSA) mortality projections are far too pessimistic relative to international evidence, that the standard 75-year actuarial fix (2.2 percentage point [pp] payroll tax increase) leaves a 75% probability of trust fund exhaustion, and that disability trends are mostly positive with limited Medicare cost implications from longer life. Medical technology, not aging per se, is the dominant risk for long-run Medicare spending. Published in Journal of Economic Perspectives 13(1): 117–140.
Key Claims
- Old-age dependency ratio roughly doubles by 2070, with wide uncertainty: Lee-Tuljapurkar stochastic projection gives a median old-age dependency ratio (OADR) of ~0.47 by 2070, but 95% confidence interval (CI) = 0.26–0.68. Fertility is the primary uncertainty driver; immigration effects on solvency are small relative to fertility and mortality.
- SSA mortality projections are too pessimistic: SSA's 2065 life expectancy (LE) forecast = 81.2 years; Lee-Carter (1992) point forecast = 86 years. At ages 60–64, SSA projects male mortality decline at only half the average rate observed in the UK, France, Sweden, Netherlands, and Japan over 1975–1989. At ages 75–79, SSA projects only one-third of the observed international average. Per SSA's own projections, the US won't attain Japan's current (1996) LE of 80.4 until 2051.
- Pay-as-you-go balanced budget tax rates under different mortality scenarios: If life expectancy rises to the SSA middle projection (LE=81 by 2070), payroll tax must rise from 12% (2000) to 20% (2070). Under Lee-Carter point forecast (LE=87): 24%. Under LE=90: 27%. Under LE=100: 32%. The 8 pp gap between the SSA scenario and the Lee-Carter scenario is substantial.
- The 2.2 pp fix is deeply unreliable: The long-term actuarial balance fix of +2.2 pp "balances" the system in expected-value terms over 75 years, but leaves a 75% probability of trust fund exhaustion before 2070 in the Lee-Tuljapurkar stochastic simulation. A +4 pp increase still leaves a 22% chance of bankruptcy by 2070. An immediate +2 pp increase leaves the system losing money rapidly in 2071 (costs exceed revenue by 40%).
- Longer life doesn't cost much more in Medicare: People reaching age 90 spend ~$63k in lifetime Medicare vs. ~$55k for those reaching 79 (Lubitz et al. 1995). An 8% increase in post-65 lifespan raises Medicare spending by only ~2%, because end-of-life costs are merely postponed, not multiplied.
- Disability trends reduce Medicare costs: The ~14% decline in age-adjusted disability rates 1982–94 (Manton et al.) would reduce Medicare spending by ~6% (−0.5%/yr over 12 years). Sustained for 55 years, this would reduce 2052 Medicare spending by 32% relative to baseline.
- Medical technology is the dominant Medicare risk: 1994–96 slowdown in health care spending growth (1%/yr vs. 6.6%/yr in 1990) may be temporary. Health Care Financing Administration (HCFA) 1992 projections: national health spending to reach 32% of GDP by 2030. Technology that expands expensive interventions to larger patient pools (implantable defibrillators at $88k, cardiac surgery eligibility broadening) is the primary driver of long-run cost growth.
- Geographic variation as cost-saving opportunity: Miami Medicare spending ($7,874/capita) vs. Minneapolis ($3,722/capita) despite similar health outcomes (Skinner and Fisher 1997). Benchmarking high-cost regions to Minneapolis-level spending would restore Medicare trust fund solvency through 2005.
- Policy conclusion: adaptive, cause-contingent responses. How the government should respond to trust fund stress depends on its cause. Stress from poor economic growth → cut benefits to protect working generations. Stress from unexpectedly long, healthy lives → raise the retirement age so longer-lived generations pay more.
Concepts Introduced or Extended
Entities Mentioned
Quotes
"Even with an immediate 2 percentage point increase in the payroll tax rate, Lee and Tuljapurkar find there would still be a 75 percent probability of trust fund exhaustion before 2070."
"In our view, the central Social Security Administration forecasts of mortality decline are far too low."
"A pure 8 percent increase in lifespan past age 65 was only a 2 percent increase in Medicare spending."
My Take
This is one of the clearest accessible expositions of the demographic and fiscal stakes of the baby boom aging. Its key contributions are: (1) the systematic international comparison showing SSA's mortality decline projections are roughly half the observed rates in peer countries, which is the cleanest pre-Soneji-King statement of the SSA pessimism argument; (2) the stochastic simulation results showing that the standard "fix" has a 75% bankruptcy probability; and (3) the Medicare cost analysis showing that longer life per se is not the fiscal threat — it's technology and the near-death cost concentration that matters. The paper was written at the height of the 1990s budget surplus era and its tone reflects genuine uncertainty about whether the problem is real. The disability discussion is now partially outdated by subsequent findings (disease-disability decoupling, Lakdawalla et al. 2004 on rising young-adult disability). But the mortality and solvency analysis remains the most readable introduction to these issues.