Social Security Progressivity

Definition

Social Security progressivity refers to the degree to which the combined Old-Age and Survivors Insurance (OASI) and disability insurance (DI) programs redistribute lifetime income from high earners to low earners. The system has two built-in progressive features (a higher-replacement-rate benefit formula for lower earners; DI reaching low-socioeconomic-status (SES) individuals before retirement) and two anti-progressive offsets (a payroll tax cap; differential mortality causing higher-SES individuals to collect benefits for more years). On net, the individual-level OASI retirement program is modestly progressive; at the household level, spousal benefits make it appear regressive without DI. The combined OASI+DI system is progressive at both levels.

Key Ideas

The Primary Insurance Amount (PIA) formula replaces $90\%$ of earnings in the first Average Indexed Monthly Earnings (AIME) bend point, $32\%$ in the second, and $15\%$ above — giving lower earners a much higher replacement rate. This is the principal progressive design feature.
The payroll tax is capped ( $\$128{,}400$ in 2018), so very high earners pay a lower average payroll tax rate — a regressive feature that partially offsets the benefit formula's progressivity.
Differential mortality is the most consequential offset: lower earners live shorter lives (by $5{+}$ years at age $65$ for the 1941 cohort; Waldron 2007) and therefore collect progressive benefits for fewer years. Bosworth and Burke (2014) estimate this offsets approximately half of overall Social Security (SS) system progressivity.
Spousal and survivor benefits disproportionately benefit higher-income households, making the net OASI household-level distribution regressive. Including DI restores progressivity at the household level.
As the SES mortality gap has widened over successive birth cohorts, Social Security has become less progressive in effective lifetime redistribution — even as the benefit formula has remained unchanged.

How It Works

The Progressive Formula vs. Differential Mortality

The benefit formula sets PIA as a piecewise linear function of AIME with a steep initial slope ( $90\%$ ) that flattens sharply. A worker with very low career earnings receives benefits equaling nearly 90 cents per dollar of average monthly earnings; a high earner receives closer to 15 cents per additional dollar. This generates a downward-sloping benefit-to-tax ratio curve across the earnings distribution (Meyerson and Sabelhaus 2006, Congressional Budget Office (CBO)).

But the slope of that curve depends on how many years each group collects. Under equal life expectancy at 65, the curve would be much steeper — lower earners would collect the higher replacement rate for as many years as high earners. Under actual differential mortality, the slope flattens substantially: lower earners receive higher per-year benefits but for shorter durations. For the 1960s birth cohort, the CBO analysis (Meyerson and Sabelhaus 2006) shows the benefit-to-tax ratio is already near flat for the bottom third of the earnings distribution — the lifetime transfer to low earners is far smaller than the annual benefit formula implies.

The Cohort Deterioration (Auerbach et al. 2017 / National Academy of Sciences [NAS])

Comparing the 1930 and 1960 birth cohorts illustrates the historical trajectory:

Earnings quintile	1930 cohort lifetime SS benefits	1960 cohort	Change
Lowest	$\$126{,}000$	$\$122{,}000$	$-\$4{,}000$ (fell)
2nd	$\$144{,}000$	$\$153{,}000$	$+\$9{,}000$
3rd	$\$166{,}000$	$\$209{,}000$	$+\$43{,}000$
4th	$\$191{,}000$	$\$260{,}000$	$+\$69{,}000$
Highest	$\$229{,}000$	$\$295{,}000$	$+\$66{,}000$

(2009 dollars, men.) Between birth cohorts spanning $30$ years, the top quintile gained $\$66{,}000$ in lifetime SS benefits while the bottom quintile lost $\$4{,}000$ in absolute real terms. This is the direct consequence of the widening life expectancy (LE) gap by SES operating through the benefit collection period.

DI as a Progressive Counter-Force

DI is a progressive element of the combined system for two reasons:

DI benefits reach individuals in the pre-retirement years, before the differential-mortality offset has accumulated — low-SES beneficiaries receive transfers at ages when they still have years of life ahead.
DI beneficiaries are overwhelmingly low-income by definition (inability to perform Substantial Gainful Activity). The program's transfers therefore flow to the bottom of the income distribution, not to a SES-neutral pool.

Including DI in progressivity analyses reverses the household-level regressivity of OASI alone: Steuerle, Carasso, and Cohen (2004a) and Gustman and Steinmeier (2001) both find the combined OASI+DI system is progressive at the household level. This is a key structural difference between the two programs — OASI's redistributive design is eroded by differential mortality at the retirement stage, while DI's transfers hit before that stage.

Why It Matters

Reform debates: Proposals to raise the Social Security retirement age, typically justified by rising aggregate LE, are regressive in distributional terms — LE gains have been concentrated at the top of the SES distribution. Raising the retirement age costs lower earners (who gained little LE) the same number of benefit years as higher earners (who gained substantially more).
DI's distinctive role: The finding that DI restores combined-system progressivity provides a distributional argument for maintaining or expanding DI generosity — it is the program element that counteracts differential mortality's regressive effect on lifetime OASI transfers.
Fiscal design vs. distributional design: The SS payroll tax cap reduces progressivity at the tax side. The progressive benefit formula attempts to restore it at the benefit side. Differential mortality then erodes the benefit-side progressivity. This layered structure means distributional evaluations must track all three simultaneously, not just the benefit formula.

Open Questions

How much of the cohort deterioration in bottom-quintile lifetime SS benefits would be reversed by narrowing the SES mortality gap vs. modifying the benefit formula?
Does the progressivity offset from differential mortality vary by race and ethnicity, given substantial race $\times$ SES mortality disparities documented by Chetty et al. (2016) and Olshansky et al. (2012)?
How does rising educational attainment (reducing the share in the bottom education tier) interact with the SES LE gap to affect future progressivity trajectories?