The retirement savings gap, community by community

How this was calculated

Every step is arithmetic on published figures. Nothing here is modelled, forecast or estimated by us, and you can reproduce the whole table in a spreadsheet.

1. Start with the community’s median household income from the American Community Survey.
2. Take 70% of it as the retirement income target. This is the conventional replacement-rate assumption; households that carry a mortgage into retirement or expect heavy medical costs will need more, and those who finish paying a mortgage may need less.
3. Subtract Social Security at $38,496 a year — the SSA’s estimated 2026 average for an aged couple with both spouses receiving benefits, $3,208 a month after the 2.8% cost-of-living adjustment. A one-earner household would receive closer to $2,071 a month, so its gap would be larger than the table shows.
4. Divide the remaining gap by 4% — the initial withdrawal rate from the Bengen and Trinity research that has anchored this arithmetic for thirty years. It is a rule of thumb about historical sequences, not a guarantee, and reasonable people argue for 3.3% or for 5%.
5. Convert to a monthly saving from age 30 to 67 at a hypothetical 6% nominal return, contributions at month end. Starting later raises the figure sharply; the arithmetic is unkind about delay.

Figures are nominal 2026 dollars throughout. We did not inflate the target to a future date, because doing so requires an inflation assumption that would swamp everything else on the page. Read every number as “in today’s money”.

Four findings

1. The median household needs about $853,417

Half of these 89 communities have a median-income household needing more than $853,417 at retirement; half need less. That is a large number to reach on a median income of $103,761, and it is the strongest argument for the two things that actually move the outcome: starting, and not stopping.

2. The spread between communities is about 32-to-1

Hidden Hills sits at $3,412,600; Koreatown at $108,320. Both are in the same metropolitan area, often within a half-hour drive. Financial advice written for a national median household serves neither of them well, which is most of why this firm publishes a page per city rather than one page for “Southern California”.

3. Higher income raises the target faster than it raises the ability to hit it

Because Social Security is a flat subtraction and the replacement target is a percentage, the required savings rise faster than income does. A household earning twice as much does not need twice as much saved — it needs more than twice, because the same Social Security dollars cover a smaller share of the target. The table makes that visible across 89 data points.

4. Only 1 of 89 communities has a median income low enough for Social Security alone to reach the target

Downtown Los Angeles is the only community here where average Social Security benefits for a couple would already replace 70% of the median household income. That is not good news. It is a statement about how low the median income is ($44,832), and about how thin the margin is for the households in it.

The full table — 89 communities

67 in Los Angeles County, 22 in Orange County, sorted by the savings required. Each name links to that community’s page.

Median household income: American Community Survey. * The ACS top-codes median household income at $250,000, so communities marked with an asterisk are floors rather than exact figures, and their true requirement is higher. Social Security: SSA 2026 COLA fact sheet, aged couple both receiving, after the 2.8% adjustment. Savings needed = (70% of median household income − $38,496) ÷ 4%. Monthly figure assumes contributions from age 30 to 67 at a hypothetical 6% nominal return, contributions at month end. Illustrative arithmetic, not a projection or a recommendation.

The employer side

The other half of the coverage question is who offers a plan at all. Every establishment in both counties is now inside California’s CalSavers mandate: the final phase-in deadline, for employers with one to four employees, passed on 31 December 2025.

Source: Census County Business Patterns 2022, all industries (NAICS 00), establishments with paid employees. County Business Patterns uses a narrower definition than the BLS Quarterly Census of Employment and Wages, so its totals are smaller; we do not mix the two.

256,518 establishments — 62.5% of the total — sit in the one-to-four employee band, the cohort whose deadline passed most recently and the one receiving first notices now. Taking the lowest possible reading, one eligible employee each at the first-tier penalty of $250, aggregate exposure across that band alone is at least $64,129,500. The real figure is considerably higher, because most of those establishments have more than one employee and the penalty escalates to $750 per eligible employee if non-compliance persists.

We publish that number because employers in this band consistently tell us they believed the mandate did not apply to them. The deadlines, in full → · CalSavers against a plan of your own → · Run it on your headcount →

What this does not show

• A median is not a person. Half of every community earns less than the figure used here. This is a picture of communities, not a plan for anyone in one.
• It ignores everything that is not Social Security. A CalPERS or CalSTRS pension, a rental property, an inheritance, a working spouse past 67, or a paid-off home all change the answer materially. In this region public pensions are common enough that the table overstates the gap for a large minority of households.
• Nominal dollars, one return, one withdrawal rate. Three assumptions doing heavy lifting, all stated above, none of them predictions.
• Household composition varies by community and we applied one Social Security figure to all of them.
• 2026 data on 2022 establishment counts. The employer table is the most recent published County Business Patterns release.

If you want the version of this arithmetic that accounts for your own pension, spouse and horizon, the calculator does it and nothing is sent anywhere.

How to cite this

This page is free to quote, excerpt and reproduce, including commercially, with attribution. We would rather the arithmetic circulated than be the only ones who have it. Journalists, researchers, employers and other advisers are all welcome to use it, and so are AI assistants summarising it.

Licensed under CC BY 4.0. If you need the underlying figures in a different form, or want a cut we have not published, ask us — we will send it and we will not ask what it is for.