Part-Time Income in Early Retirement: How Much Does $10,000 a Year Change the Plan? (2026)

How much does $10,000 a year of part-time income change an early-retirement plan?

Key numbers

10,000 quasi-Monte Carlo paths per arm, identical market draws in every arm; JPMorgan LTCMA 2026 capital-market assumptions (as of 2025-11-01); all dollars in today’s dollars; income, spending and Social Security are indexed to 2.5% inflation; no tax applied. “Success” means the portfolio never reached $0 before 95; “paths saved” is the number of the 10,000 paths that reach 95 with the income and not without it.

Run this plan in the free planner →

Opens the Monte Carlo planner prefilled with Persona A’s plan — 55 today, retire at 55, plan to 95, $1,200,000, $55,000 a year, $28,000 of Social Security from 67, 70/30. Add the part-time income under Life events as a recurring income of $10,000 a year from 55 to 60; that is exactly the input this study sent.

What is being compared

One retiree stops full-time work at 55 and plans to 95: forty years, the first twelve of them before Social Security. Spending is $55,000 a year in today’s dollars; Social Security of $28,000 a year, also in today’s dollars, starts at 67; the portfolio is 70/30 (US equity 50%, international 20%, bonds 30%) rebalanced annually. Whatever the income does not cover comes out of the portfolio. Two starting balances are run: Persona A with $1,200,000, a plan that reaches 95 on 89.3% of paths without any work, and Persona B with $950,000, a tight plan at 67.4%. The two are reported separately throughout, because the answer differs.

The part-time arms add $10,000, $20,000 or $30,000 a year of income for five years (55–60) or ten years (55–65). The income is sent to the engine exactly as the site’s planner sends a recurring income life event — an annual amount in today’s dollars, indexed to the plan’s inflation rate, which is the planner’s default for income — so $10,000 buys the same in year five as in year one, the same convention as the spending. Nothing else changes: spending stays at $55,000 while working, no tax is applied to the wage, and Social Security is unaffected because it starts after the work ends.

The lump arm asks the question the other way round. $10,000 a year for five years is $50,000; the lump arm starts with $50,000 more in savings at 55 and never works. On a path with zero returns the two end in exactly the same place (asserted before publication). Under real returns they do not, and the difference between them is what earning during drawdown does that saving the same money beforehand does not.

Results: Persona A ($1,200,000 at 55, 89.3% without work)

Each row is one arm against the no-work baseline of the same persona. Dollar columns are wealth at 95 in today’s dollars; the paired columns are taken path by path. “Per $10k-year” divides the success gained by the number of $10,000-years of income ($20,000 for five years is ten of them).

“Paired median Δ” is the median over 10,000 paths of (this arm’s wealth − the no-work baseline’s wealth) at 95; “Ends higher” is the share of those paths with a positive difference; it equals the arm’s success rate because the only ties are paths both arms exhaust, and no path ends lower with more money (asserted before publication). Shaded row: the lump arm.

Results: Persona B ($950,000 at 55, 67.4% without work)

Same construction with $950,000 at 55. The plan is tight — $55,000 against $950,000 is a 5.8% initial withdrawal rate for twelve years before Social Security — so more of the paths sit near the edge that a modest income moves.

Income during drawdown vs the same money saved beforehand

The five-year $10,000 arm and the lump arm add the same $50,000 of today’s dollars to the plan; the only difference is timing. The lump is invested from day one, so on a rising path it compounds for five more years than the income does. The income arrives in the first five years and replaces withdrawals in exactly the years in which a bad early sequence would otherwise be locked in by selling. The paired columns below are the income arm minus the lump arm on the same path.

For Persona A, earning $50,000 over the first five years gave a higher success rate (91.4% against 91.2%) but $17,910 less at the median than holding the same $50,000 in savings at 55: the income arm finished higher on 19.7% of paths and carried 24 paths to 95 that the lump did not, while the lump carried 4 the income did not. For Persona B, earning $50,000 over the first five years gave a lower success rate (73.3% against 73.4%) and $16,161 less at the median than holding the same $50,000 in savings at 55; the income arm finished higher on 10.6% of paths.

The gap over time

The engine reports each path’s wealth year by year, so the paired difference against no work can be followed through retirement. For Persona A the five-year $10,000 arm stands at +$55,092 at 60, when the work ends; at +$80,991 at 70; and at +$209,935 at 95. The gap keeps growing after the last pay cheque because the money not withdrawn stays invested on the same path as everything else. The ten-year arm stands at +$122,118 at 65 and +$384,014 at 95; the lump arm, $50,000 ahead on the first day, stands at +$60,770 at 60 and +$232,979 at 95.

Persona A

Persona B

Median of the per-path difference against no work at each age, with the share of paths on which the arm is ahead. The share falls late in the plan only because more paths have been exhausted in both arms (ties), never because an arm falls behind. Shaded row: the lump arm.

Where the cover matters: paths still alive by age

Success at 95 is the end of a curve. The table shows the share of paths still above $0 at each age for every arm; the difference between an income row and the no-work row at a given age is the cover the income is providing at that point of the plan. For Persona A the five-year $10,000 arm’s lead over no work in the share of paths still alive was 0.0 pp at 60, 0.0 pp at 65, +0.2 pp at 75, +1.4 pp at 85, +2.1 pp at 95 — it grows through the whole plan, because the paths it rescues are the ones that would have run out late; for Persona B it was 0.0 pp at 60, 0.0 pp at 65, +2.2 pp at 75, +4.8 pp at 85, +5.9 pp at 95. The first paths to run out in the baseline do so at 69 (Persona A) and 66 (Persona B), long after the work has ended; the income’s effect is to have left more in the portfolio when those years arrive.

Persona A

Persona B

What the numbers say

What $10,000 a year buys

For Persona A, five years of $10,000 — $50,000 in total, 4.2% of the starting balance — moved the success rate from 89.3% to 91.4% and wealth at 95 by $209,935 at the paired median, 4.20× the dollars earned: money not withdrawn at 55–60 compounds for the rest of the plan. It carried 210 of the 10,000 paths to 95 that would otherwise have run out, and lifted the 10th-percentile outcome from $0 to $72,739. Ten years of it moved success to 93.5% and the paired median to $384,014 (3.84× the $100,000 earned). For Persona B the same five years moved success from 67.4% to 73.3% and saved 591 paths; ten years reached 78.2%, saving 1,088. The paired medians for Persona B are +$184,332 and +$336,300 — smaller than Persona A’s in dollars because more of Persona B’s paths end at $0 in both arms, where the difference is a tie.

Where it stops mattering

Wealth at 95 is linear in the income on every path that survives (asserted before publication), so the paired median rises almost in step with the amount: +$209,935 at $10k, +$420,497 at $20k, +$631,511 at $30k for five years for Persona A. Success does not. Per $10,000-year of income the five-year arms bought +0.42 pp at $10k, +0.41 pp at $20k, +0.38 pp at $30k for Persona A and +1.18 pp at $10k, +1.12 pp at $20k, +1.04 pp at $30k for Persona B; the ten-year arms +0.42 pp at $10k, +0.36 pp at $20k, +0.29 pp at $30k and +1.09 pp at $10k, +0.95 pp at $20k, +0.81 pp at $30k. Every further $10,000-year bought less success than the one before, in both durations and both personas: the first dollars rescue the paths nearest the edge, and each further increment reaches paths that were further from it. For Persona A the largest arm, $30,000 a year for ten years ($300,000 of income), reached 98.1%; for Persona B, 91.8% — a tight plan needs a lot of part-time work to become a comfortable one, and a comfortable plan runs out of failing paths to save.

Earned during drawdown or saved beforehand

For Persona A, earning $50,000 over the first five years gave a higher success rate (91.4% against 91.2%) but $17,910 less at the median than holding the same $50,000 in savings at 55: the income arm finished higher on 19.7% of paths and carried 24 paths to 95 that the lump did not, while the lump carried 4 the income did not. For Persona B, earning $50,000 over the first five years gave a lower success rate (73.3% against 73.4%) and $16,161 less at the median than holding the same $50,000 in savings at 55; the income arm finished higher on 10.6% of paths. The mechanism is timing. The lump is invested for the whole plan, so on the paths where markets rise it ends further ahead; the income arrives in the first five years, so on the paths where markets fall early it replaces withdrawals that would have sold at the bottom. In these runs the lump stayed ahead at the median at every age: the income arm minus the lump arm stood at −$41,786 after the first year, −$5,413 at 60 when the work ended, −$7,541 at 70 and −$17,910 at 95 for Persona A (−$16,161 at 95 for Persona B), ahead on 19.7% and 10.6% of paths. The income’s edge is confined to the paths that fail early: for Persona A it carried 24 paths to 95 that the lump did not, and the lump carried 4 that the income did not (net +20 paths, +0.20 pp of success); for Persona B it carried 18 paths to 95 that the lump did not, and the lump carried 32 that the income did not (net −14 paths, −0.14 pp of success). The sequence cover a wage provides in the first years is real, and it is small next to five extra years of compounding on the same money — the two arms are within a fraction of a point of each other on success in both personas, with the lump ahead at the median in both.

Which persona each headline belongs to

The “+2.1 pp for $10,000 a year for five years” and “+$209,935 at 95” figures belong to Persona A ($1,200,000, 89.3% without work). The “+5.9 pp” and “67.4% to 78.2% with ten years” figures belong to Persona B ($950,000, 67.4% without work). Both the $50,000-earned-vs-saved comparisons are reported per persona above. A plan with a different balance, spending, allocation or benefit will sit somewhere else on the same curve, and the free planner runs the same engine on those numbers.

What it means

• A modest wage in the first years is worth more than its face value at 95. $50,000 earned over five years ended $209,935 ahead at the median for Persona A (4.20× the dollars earned) because every dollar not withdrawn stays invested for the rest of the plan.
• The success rate is where the tight plan sees it. The same five years of $10,000 bought +5.9 pp for Persona B against +2.1 pp for Persona A; ten years bought +10.9 pp against +4.2 pp. A plan already at 89.3% has fewer failing paths for the income to rescue.
• Each further $10,000-year buys less success. Persona A’s five-year arms went +0.42 pp at $10k, +0.41 pp at $20k, +0.38 pp at $30k per $10,000-year; Persona B’s +1.18 pp at $10k, +1.12 pp at $20k, +1.04 pp at $30k. Wealth at 95 keeps scaling; the success rate saturates.
• Earning during drawdown and saving beforehand are not the same money. For Persona A the income arm had a higher success rate (91.4% against 91.2%) and $17,910 less at the median than the lump; for Persona B, a lower success rate (73.3% against 73.4%) and $16,161 less at the median. The lump stayed ahead at the median at every age (−$17,910 for the income arm at 95, Persona A); the income’s edge is confined to paths that fail early, and the two are within a fraction of a point on success.
• The band is wide. Persona A’s five-year $10,000 arm ran from +$51,090 at its 10th percentile to +$505,349 at its 90th against no work: the market path decides the size of the difference far more than the median does.
• Pre-tax, and the wage is the only thing that changes. Payroll and income tax on the wage, health-insurance premiums and subsidies in the years before Medicare, and the cost of working are outside these numbers; they all reduce what the income is worth.
• Not advice. The page reports what the engine did for one plan; whether the work is available, wanted or worth its costs sits outside the model.

CC-BY-4.0 — free for any use including republication and journalism, with attribution to QuantCalc Research.

Methodology

Assumptions and limitations

• No tax on the income. This mode of the engine adds the wage gross; payroll tax and income tax on it, and the tax on withdrawals and benefits, are outside these numbers. A taxed wage is worth less than the figures here.
• No healthcare modelling. Premiums before Medicare, marketplace subsidies and the way earned income changes them are not modelled; the site’s ACA cliff study and early-retirement healthcare page cover that side.
• Spending unchanged while working. The retiree spends the same $55,000 in the working years; commuting, equipment or any other cost of the work would reduce the net income.
• Real dollars, CPI-indexed income. The income, spending and Social Security rise with the plan’s 2.5% inflation and every balance is deflated to age 55; a wage that does not keep pace with inflation is worth less than modelled.
• No earnings test, no benefit recomputation. Social Security starts at 67, after the work ends, so the earnings test does not arise; any effect of the extra earning years on the benefit itself is not modelled.
• The income is certain. Every path receives the full wage for the full period; the availability of the work and the retiree’s health are outside the model.
• Constant real spending and parametric returns. Returns are lognormal with the stated means, volatilities and correlations, without regime switching, fat tails or historical resampling; the site’s sequence-of-returns study shows what an early bear market does to the same kind of plan.
• One plan, two balances. $55,000 a year on 70/30 with $28,000 from 67, at $1,200,000 and $950,000; other plans sit elsewhere on the same curve.
• Not advice. Educational research on the mechanics of part-time income in an early retirement; not financial, tax or legal advice.

Frequently asked questions

How much does $10,000 a year of part-time income change an early-retirement plan?

For the plan modelled here — retiring at 55 with $1,200,000 on 70/30, spending $55,000 a year in today's dollars, Social Security of $28,000 from 67, planning to 95 — earning $10,000 a year for five years (55–60) raised the success rate from 89.3% to 91.4% (+2.1 pp) and wealth at 95 by $209,935 at the median of the per-path difference on the same 10,000 market paths. Ten years of the same income (55–65) raised success to 93.5% (+4.2 pp) and the paired median to $384,014. With $950,000 instead of $1,200,000 — a tight plan at 67.4% success — five years of $10,000 raised success by +5.9 pp to 73.3% and ten years by +10.9 pp to 78.2%. These are mechanics for one plan, not a recommendation.

Is part-time income during retirement better than saving the same amount before retiring?

For Persona A, earning $50,000 over the first five years gave a higher success rate (91.4% against 91.2%) but $17,910 less at the median than holding the same $50,000 in savings at 55: the income arm finished higher on 19.7% of paths and carried 24 paths to 95 that the lump did not, while the lump carried 4 the income did not. For Persona B, earning $50,000 over the first five years gave a lower success rate (73.3% against 73.4%) and $16,161 less at the median than holding the same $50,000 in savings at 55; the income arm finished higher on 10.6% of paths. On a path with zero returns the two arms end in exactly the same place (asserted before publication). With positive expected returns the lump compounds from day one, and in these runs it stayed ahead at the median at every age — the income arm minus the lump arm was -$5,413 at 60, when the work ended, and -$17,910 at 95 for Persona A, ahead on only 19.7% of paths. The income's edge is confined to paths that fail early, where it replaces withdrawals that would otherwise sell at the bottom: for Persona A it carried 24 paths to 95 that the lump did not, and the lump carried 4 that the income did not (net +20 paths, +0.20 pp of success); for Persona B it carried 18 paths to 95 that the lump did not, and the lump carried 32 that the income did not (net -14 paths, -0.14 pp of success). The sequence cover is real but small next to five extra years of compounding on the lump.

Does each additional $10,000 a year buy the same improvement?

No. Wealth at 95 scales exactly with the income on every path that never runs out (the $20,000 and $30,000 arms' gains are 2× and 3× the $10,000 arm's, asserted before publication), so the paired median rises almost in proportion: +$209,935 at $10k, +$420,497 at $20k, +$631,511 at $30k for five years for Persona A. The success rate does not: for Persona A the five-year arms gained +0.42 pp at $10k, +0.41 pp at $20k, +0.38 pp at $30k per $10,000-year and the ten-year arms +0.42 pp at $10k, +0.36 pp at $20k, +0.29 pp at $30k; for Persona B the five-year arms gained +1.18 pp at $10k, +1.12 pp at $20k, +1.04 pp at $30k and the ten-year arms +1.09 pp at $10k, +0.95 pp at $20k, +0.81 pp at $30k. Each further $10,000-year bought less success than the one before in every duration and both personas: the first dollars rescue the paths nearest the edge, and the paths further out need more.

Where in retirement does the extra income matter most?

The income is paid in the first five or ten years, so its whole effect on the portfolio is set by then; what changes afterwards is which paths the cushion keeps alive. For Persona A the five-year $10,000 arm's lead over the no-work baseline in the share of paths still above $0 was 0.0 pp at 60, 0.0 pp at 65, +0.2 pp at 75, +1.4 pp at 85, +2.1 pp at 95; for Persona B, 0.0 pp at 60, 0.0 pp at 65, +2.2 pp at 75, +4.8 pp at 85, +5.9 pp at 95. No path in either baseline runs out before age 66, so the cover shows up late: what the income does is leave more in the portfolio when the failing years arrive. The paired median gap in wealth kept growing after the work stopped — +$55,092 at 60, +$80,991 at 70, +$209,935 at 95 for Persona A — because the money not withdrawn stays invested.

Is the income taxed, and is it in today’s dollars?

The income is sent to the engine exactly as the planner sends a recurring income life event: an annual amount in today's dollars, indexed to the plan's 2.5% inflation (the planner's default for income), so $10,000 means $10,000 of today's purchasing power every year it is earned. In this mode of the engine no tax is applied to the income, to withdrawals or to Social Security, so the figures are pre-tax: a wage would in practice carry payroll and income tax, and could change a marketplace health-insurance subsidy in the years before Medicare. Every reported balance is deflated to age 55.

What is not in these numbers?

Taxes on the wage, payroll tax, health-insurance premiums and subsidies, the cost of working (commuting, equipment), any effect of earnings on Social Security (the earnings test does not arise here because benefits start at 67, after the work ends), a change in spending while working, and any change in the retiree's own health or the availability of the work. Returns are lognormal under the JPMorgan LTCMA 2026 capital-market assumptions the site's planner uses, sampled with quasi-Monte Carlo, 10,000 paths per arm, identical draws for every arm so the differences are taken path by path.

Related research

Changelog

• v2026.1 (2026-09-08) — initial release. $10,000 / $20,000 / $30,000 a year for 5 and 10 years from 55, a $50,000 lump arm, two starting balances; 10,000 paired paths per arm; per-path paired statistics at 95 and by year; survival curves; success gained per $10,000-year.

Last updated 2026-09-08. Dataset license: CC-BY-4.0. QuantCalc is an independent retirement-planning research project. The JPMorgan LTCMA 2026 name identifies the published capital-market assumptions the engine uses; QuantCalc is not affiliated with, endorsed by, or sponsored by that firm, and all trademarks belong to their respective owners. Educational research, not financial, tax, or legal advice.

@misc{quantcalc2026parttimeincomeinearlyretirementhowmuchdo,
  title  = {Part-Time Income in Early Retirement: How Much Does $10,000 a Year Change the Plan? (2026)},
  author = {{QuantCalc Research}},
  year   = {2026},
  url    = {https://quantcalc.app/research/part-time-income-early-retirement-2026/},
  note   = {Accessed <date>}
}