A worker starts saving at age 25, puts away the equivalent of $1,000/year (in constant purchasing power) into an S&P 500 index fund, and keeps it up for 40 years. How much real wealth does she have at age 65? It depends almost entirely on the calendar year she happened to start.
This post runs the experiment for every starting cohort from 1950 to 1985 (a 40-year career ending by 2024 at the latest), and a shorter 10-year version for every starting cohort from 1950 to 2015. Each cohort contributes the same real $1,000 per year, so the only thing that varies is the sequence of returns and inflation they lived through. Final net worth is reported in 2025-equivalent dollars.
Data sources
- S&P 500 annual total returns (price + reinvested dividends), 1928–2025: Damodaran “Historical Returns on Stocks, Bonds and Bills”.
- Annual CPI inflation, FRED
CPIAUCNS(republished in the same Damodaran workbook).
All simulation and plotting code lives in dca.py; the raw historical
series is in historical_returns.csv.
Method
Each cohort contributes $1,000 in real (constant) dollars at the start of every career year. Working in real terms is equivalent to inflating the contribution and deflating the final value: the real return each year is
$$r_{real} = \frac{1 + r_{nominal}}{1 + \pi} - 1$$
and the final real net worth after $N$ years is
$$W_N = \sum_{k=0}^{N-1} 1000 \cdot \prod_{j=k}^{N-1} (1 + r_{real,, Y+j})$$
where $Y$ is the cohort’s first career year. The dashed line on each chart marks the total amount contributed in real dollars ($40,000 for a 40-year career, $10,000 for a 10-year career) — anything above it is real investment gain.
40-year careers: starting 1950 through 1985
start_year end_year final_real_net_worth multiple_of_contributed real_cagr
1950 1989 $189,159 4.73x 7.85%
1951 1990 $154,975 3.87x 7.04%
1952 1991 $177,909 4.45x 7.25%
1953 1992 $169,718 4.24x 6.94%
1954 1993 $167,035 4.18x 7.18%
1955 1994 $150,050 3.75x 5.99%
1956 1995 $188,369 4.71x 6.03%
1957 1996 $212,505 5.31x 6.37%
1958 1997 $263,940 6.60x 7.46%
1959 1998 $312,148 7.80x 7.16%
1960 1999 $349,918 8.75x 7.34%
1961 2000 $293,800 7.35x 7.03%
1962 2001 $242,773 6.07x 6.04%
1963 2002 $177,857 4.45x 5.60%
1964 2003 $214,206 5.36x 5.71%
1965 2004 $220,912 5.52x 5.52%
1966 2005 $216,254 5.41x 5.30%
1967 2006 $236,052 5.90x 5.98%
1968 2007 $229,885 5.75x 5.53%
1969 2008 $140,890 3.52x 4.19%
1970 2009 $167,621 4.19x 5.11%
1971 2010 $182,466 4.56x 5.48%
1972 2011 $173,543 4.34x 5.19%
1973 2012 $190,184 4.75x 5.17%
1974 2013 $239,115 5.98x 6.50%
1975 2014 $256,554 6.41x 7.94%
1976 2015 $237,869 5.95x 7.29%
1977 2016 $243,321 6.08x 7.08%
1978 2017 $272,570 6.81x 7.92%
1979 2018 $237,260 5.93x 7.82%
1980 2019 $279,597 6.99x 8.37%
1981 2020 $297,732 7.44x 8.35%
1982 2021 $328,832 8.22x 9.21%
1983 2022 $227,797 5.69x 8.10%
1984 2023 $251,565 6.29x 8.19%
1985 2024 $278,227 6.96x 8.66%
What stands out
- Best cohort: the worker who started in 1982 retires with the largest inflation-adjusted nest egg — they caught the 1982–2000 bull market entirely in the accumulation phase, with the biggest contributions invested earliest.
- Worst cohort: the worker who started in 1969 does the worst — they retired in 2008 right into the GFC and their career covered the 1970s stagflation. They still came out ahead, but only by about 3.5× their real contributions ($141k on $40k contributed), vs. 8.2× for the lucky cohort.
- Range: the best 40-year cohort accumulates roughly 2.3× more real wealth than the worst, off the same real saving rate. Same discipline, same index, same plan — pure timing.
- The middle is wide too. Cohorts that retired into 2000 or 2008 (started in 1961 and 1969 respectively) gave back a chunk of their accumulated wealth in the last year of their career — a reminder that “average historical return” is not the same as the return you actually get.
10-year horizons: starting 1950 through 2015
A 10-year window is short enough that the sequence of returns dominates: there are not enough years for compounding to wash out a bad decade. The dispersion across cohorts gets much wider, and several cohorts end up with less real wealth than they contributed.
start_year end_year final_real_net_worth multiple_of_contributed real_cagr
1950 1959 $24,865 2.49x 16.86%
1951 1960 $20,900 2.09x 14.31%
1952 1961 $22,758 2.28x 15.17%
1953 1962 $17,685 1.77x 12.16%
1954 1963 $18,738 1.87x 14.51%
1955 1964 $18,288 1.83x 11.26%
1956 1965 $18,064 1.81x 9.27%
1957 1966 $14,477 1.45x 7.31%
1958 1967 $16,163 1.62x 10.83%
1959 1968 $15,204 1.52x 7.67%
1960 1969 $12,191 1.22x 5.09%
1961 1970 $11,328 1.13x 5.00%
1962 1971 $11,835 1.18x 3.65%
1963 1972 $13,096 1.31x 6.21%
1964 1973 $9,672 0.97x 1.79%
1965 1974 $6,252 0.63x -3.74%
1966 1975 $8,415 0.84x -2.29%
1967 1976 $10,181 1.02x 0.74%
1968 1977 $8,808 0.88x -2.44%
1969 1978 $8,819 0.88x -3.21%
1970 1979 $9,518 0.95x -1.35%
1971 1980 $11,292 1.13x 0.42%
1972 1981 $9,842 0.98x -1.91%
1973 1982 $11,618 1.16x -1.81%
1974 1983 $13,892 1.39x 2.22%
1975 1984 $13,935 1.39x 6.78%
1976 1985 $16,446 1.64x 6.64%
1977 1986 $18,219 1.82x 6.56%
1978 1987 $17,559 1.76x 8.18%
1979 1988 $18,264 1.83x 9.62%
1980 1989 $21,054 2.11x 11.65%
1981 1990 $17,397 1.74x 8.91%
1982 1991 $20,279 2.03x 12.99%
1983 1992 $18,685 1.87x 11.82%
1984 1993 $17,798 1.78x 10.74%
1985 1994 $15,812 1.58x 10.37%
1986 1995 $18,907 1.89x 10.99%
1987 1996 $20,267 2.03x 11.14%
1988 1997 $24,071 2.41x 14.02%
1989 1998 $26,976 2.70x 15.44%
1990 1999 $27,987 2.80x 14.69%
1991 2000 $22,041 2.20x 14.26%
1992 2001 $16,708 1.67x 10.05%
1993 2002 $11,512 1.15x 6.63%
1994 2003 $13,368 1.34x 8.39%
1995 2004 $13,010 1.30x 9.29%
1996 2005 $11,737 1.17x 6.30%
1997 2006 $12,284 1.23x 5.75%
1998 2007 $11,690 1.17x 3.08%
1999 2008 $7,185 0.72x -3.78%
2000 2009 $9,201 0.92x -3.39%
2001 2010 $10,739 1.07x -0.93%
2002 2011 $10,738 1.07x 0.39%
2003 2012 $12,186 1.22x 4.51%
2004 2013 $15,143 1.51x 4.85%
2005 2014 $16,379 1.64x 5.37%
2006 2015 $15,792 1.58x 5.30%
2007 2016 $16,553 1.66x 4.99%
2008 2017 $18,967 1.90x 6.70%
2009 2018 $16,968 1.70x 10.98%
2010 2019 $19,414 1.94x 11.48%
2011 2020 $20,316 2.03x 11.81%
2012 2021 $21,920 2.19x 13.96%
2013 2022 $14,802 1.48x 9.58%
2014 2023 $16,228 1.62x 8.87%
2015 2024 $18,070 1.81x 9.69%
What stands out
- Underwater cohorts. Decades starting in 1964–1972 (stagflation) and in 1999–2000 (dot-com bust into GFC) finish below $10,000 of real contributions — the only thing that saved them was that they were still working, not retiring. The 1965 cohort retired with $6,252 in real dollars after contributing $10,000 — a 37% real loss over a decade.
- Best 10-year window: starting in 1988–1990 (catching the dot-com run-up) or 1950–1952 (the post-war bull) more than doubled the contributions in real terms over a single decade.
- Real CAGR ranges from roughly –4%/yr to +17%/yr depending on which decade you got. Compound that over a career and you get the dispersion we saw above.
Takeaways
- Sequence-of-returns risk is real and large, even for buy-and-hold index investors. The same disciplined $1,000/year real saver retires with anywhere from roughly $140k to $330k in 2025 dollars depending purely on which year they happened to start their career. Holding the index is necessary; it is not sufficient to guarantee a particular outcome.
- 40 years smooths a lot, but not everything. The 40-year real CAGRs range from ~4%/yr (cohorts who started just before the 1970s stagflation) to ~9%/yr (cohorts who started just before the long 1982–2000 bull). Compounded over a career that 5pp gap is a factor-of-2+ difference in retirement wealth.
- 10-year windows can lose money in real terms. Several decade-long cohorts end up with less purchasing power than they contributed. This is why target-date funds and bond ladders exist for people who are 10 years from retirement — the equity premium is an asymptotic statement, not a guarantee over short horizons.
- The lesson is not “time the market” — it’s “extend the horizon.” You cannot pick your starting year, but you can pick how long you stay invested. Going from 10 years to 40 years compresses the dispersion of outcomes dramatically (and shifts the whole distribution up).
Sources
- Aswath Damodaran — Historical Returns on Stocks, Bonds and Bills (NYU Stern)
- FRED — CPI for All Urban Consumers: All Items (CPIAUCNS)
- Shiller, Robert J. — Online Data (Irrational Exuberance) (cross-check)
What if the contribution is 10% of median household income?
The $1,000-real-per-year saver is a useful unit because every cohort puts in the same amount and we can read variation in outcomes as pure variation in returns. But real median household income isn’t constant — it grew from about $36,600 in 1950 to about $88,500 in 2024 (real 2025 dollars, spliced household / family series; see method note below). A worker saving a fixed fraction of household income contributes more in real terms in later career years and in later cohorts.
For this section each cohort contributes 10% of that year’s real median household income. Two things change vs. the $1,000-flat version:
- Within a career, real contributions rise over time — so when real household income grew (most cohorts before 2000), the bigger contributions land in the later career years, where they compound for fewer years. That dampens the final wealth gain a bit relative to a flat real contribution of the same average size.
- Across cohorts, later starters benefit from higher real income, so their total real contributions are larger.
Source for median household income: FRED MEHOINUSA672N (real median household income, 2023 CPI-U-RS dollars) for 1984–2024, spliced via the 1984 HH/Family ratio (~0.85) onto FRED MEFAINUSA672N (real median family income) for 1953–1983, held flat at 1953 for 1950–1952, then inflated to ~2025 dollars using the same CPI series as the rest of the analysis. The raw spliced series lives in median_household_income.csv.
40-year careers — 10% of median household income
start_year end_year total_real_contributed final_real_net_worth multiple_of_contributed real_cagr
1950 1989 $217,026 $877,345 4.04x 7.85%
1951 1990 $220,280 $739,155 3.36x 7.04%
1952 1991 $223,332 $872,218 3.91x 7.25%
1953 1992 $226,329 $855,181 3.78x 6.94%
1954 1993 $229,293 $864,925 3.77x 7.18%
1955 1994 $232,432 $804,013 3.46x 5.99%
1956 1995 $235,548 $1,032,883 4.39x 6.03%
1957 1996 $238,521 $1,184,871 4.97x 6.37%
1958 1997 $241,615 $1,497,138 6.20x 7.46%
1959 1998 $244,978 $1,809,104 7.38x 7.16%
1960 1999 $248,310 $2,058,603 8.29x 7.34%
1961 2000 $251,528 $1,752,595 6.97x 7.03%
1962 2001 $254,576 $1,469,832 5.77x 6.04%
1963 2002 $257,437 $1,089,866 4.23x 5.60%
1964 2003 $260,148 $1,330,141 5.11x 5.71%
1965 2004 $262,654 $1,386,126 5.28x 5.52%
1966 2005 $265,083 $1,368,527 5.16x 5.30%
1967 2006 $267,326 $1,503,927 5.63x 5.98%
1968 2007 $269,580 $1,475,411 5.47x 5.53%
1969 2008 $271,316 $908,723 3.35x 4.19%
1970 2009 $272,729 $1,085,197 3.98x 5.11%
1971 2010 $273,982 $1,186,270 4.33x 5.48%
1972 2011 $275,131 $1,133,274 4.12x 5.19%
1973 2012 $276,005 $1,245,036 4.51x 5.17%
1974 2013 $277,013 $1,568,042 5.66x 6.50%
1975 2014 $278,088 $1,687,886 6.07x 7.94%
1976 2015 $279,671 $1,576,086 5.64x 7.29%
1977 2016 $281,337 $1,619,576 5.76x 7.08%
1978 2017 $283,110 $1,821,897 6.44x 7.92%
1979 2018 $284,788 $1,590,289 5.58x 7.82%
1980 2019 $286,962 $1,878,707 6.55x 8.37%
1981 2020 $289,179 $2,011,853 6.96x 8.35%
1982 2021 $291,542 $2,239,962 7.68x 9.21%
1983 2022 $293,799 $1,570,355 5.34x 8.10%
1984 2023 $296,352 $1,755,883 5.92x 8.19%
1985 2024 $298,816 $1,961,048 6.56x 8.66%
10-year horizons — 10% of median household income
start_year end_year total_real_contributed final_real_net_worth multiple_of_contributed real_cagr
1950 1959 $38,445 $93,397 2.43x 16.86%
1951 1960 $39,152 $79,552 2.03x 14.31%
1952 1961 $39,905 $88,075 2.21x 15.17%
1953 1962 $40,783 $69,808 1.71x 12.16%
1954 1963 $41,817 $75,949 1.82x 14.51%
1955 1964 $43,147 $77,296 1.79x 11.26%
1956 1965 $44,434 $78,676 1.77x 9.27%
1957 1966 $45,747 $64,589 1.41x 7.31%
1958 1967 $47,146 $74,282 1.58x 10.83%
1959 1968 $48,807 $72,655 1.49x 7.67%
1960 1969 $50,525 $60,216 1.19x 5.09%
1961 1970 $52,133 $57,891 1.11x 5.00%
1962 1971 $53,693 $62,690 1.17x 3.65%
1963 1972 $55,402 $71,706 1.29x 6.21%
1964 1973 $57,091 $54,795 0.96x 1.79%
1965 1974 $58,426 $36,398 0.62x -3.74%
1966 1975 $59,454 $50,010 0.84x -2.29%
1967 1976 $60,404 $61,482 1.02x 0.74%
1968 1977 $61,277 $53,994 0.88x -2.44%
1969 1978 $62,093 $54,771 0.88x -3.21%
1970 1979 $62,727 $59,708 0.95x -1.35%
1971 1980 $63,154 $71,293 1.13x 0.42%
1972 1981 $63,404 $62,373 0.98x -1.91%
1973 1982 $63,300 $73,490 1.16x -1.81%
1974 1983 $63,102 $87,648 1.39x 2.22%
1975 1984 $63,264 $88,108 1.39x 6.78%
1976 1985 $63,661 $104,735 1.65x 6.64%
1977 1986 $64,108 $116,618 1.82x 6.56%
1978 1987 $64,606 $112,961 1.75x 8.18%
1979 1988 $64,956 $117,751 1.81x 9.62%
1980 1989 $65,330 $135,864 2.08x 11.65%
1981 1990 $65,841 $112,822 1.71x 8.91%
1982 1991 $66,330 $132,568 2.00x 12.99%
1983 1992 $66,844 $123,661 1.85x 11.82%
1984 1993 $67,284 $119,206 1.77x 10.74%
1985 1994 $67,596 $106,763 1.58x 10.37%
1986 1995 $67,998 $128,744 1.89x 10.99%
1987 1996 $68,261 $138,451 2.03x 11.14%
1988 1997 $68,586 $164,856 2.40x 14.02%
1989 1998 $69,122 $185,301 2.68x 15.44%
1990 1999 $69,729 $192,464 2.76x 14.69%
1991 2000 $70,400 $152,057 2.16x 14.26%
1992 2001 $71,148 $116,362 1.64x 10.05%
1993 2002 $71,891 $81,111 1.13x 6.63%
1994 2003 $72,671 $95,651 1.32x 8.39%
1995 2004 $73,369 $94,443 1.29x 9.29%
1996 2005 $73,969 $86,316 1.17x 6.30%
1997 2006 $74,552 $91,315 1.22x 5.75%
1998 2007 $75,111 $87,675 1.17x 3.08%
1999 2008 $75,145 $53,929 0.72x -3.78%
2000 2009 $74,943 $68,854 0.92x -3.39%
2001 2010 $74,587 $79,996 1.07x -0.93%
2002 2011 $74,248 $79,652 1.07x 0.39%
2003 2012 $73,969 $90,074 1.22x 4.51%
2004 2013 $73,956 $111,903 1.51x 4.85%
2005 2014 $73,859 $120,924 1.64x 5.37%
2006 2015 $74,042 $116,691 1.58x 5.30%
2007 2016 $74,415 $122,571 1.65x 4.99%
2008 2017 $74,806 $140,678 1.88x 6.70%
2009 2018 $75,566 $126,531 1.67x 10.98%
2010 2019 $76,960 $146,673 1.91x 11.48%
2011 2020 $78,351 $155,848 1.99x 11.81%
2012 2021 $79,816 $171,259 2.15x 13.96%
2013 2022 $81,095 $118,061 1.46x 9.58%
2014 2023 $82,441 $132,140 1.60x 8.87%
2015 2024 $83,992 $150,394 1.79x 9.69%
What changes vs. the flat-$1,000 version
- Absolute wealth scales up by ~7×. A 10% savings rate on median household income averages roughly $6,000–$9,000/year real, vs. $1,000 flat. The shape of the cohort curve is similar, but the y-axis is now in millions rather than hundreds of thousands.
- Multiple of contributions compresses slightly. Because the contribution grew in real terms within most cohorts’ careers (real median household income roughly doubled from 1953 to 2024), the largest contributions arrived late and had fewer years to compound. So the same return path produces a smaller final-wealth multiple of contributed dollars than the flat-real version above.
- Cohort ranking is essentially unchanged. The relative dispersion across
cohorts comes from the return sequence, which is identical to the flat
version. Best and worst cohorts move by a year or two but the picture is
the same: the 1982 starter still wins (
$2.24M real), the 1969 starter still loses ($0.91M real).
The analysis above is generated by sp500_dca.ipynb, with the simulation and plotting code in dca.py in the same directory. View on GitHub to inspect or run.