How different is a lifetime-earnings ranking from an average-pay ranking?

The order does change

76.4% of the 3,780 companies move more than 100 places between the two rankings. They are different lists.

But tax is not what moves it

Splitting the comparison isolates the two effects: average pay → lifetime gross is age adjustment only; lifetime gross → lifetime after tax is tax and social insurance only.

Going from gross to after-tax barely moves anything (rank correlation 1.000, median gap 2 places), while going from average pay to lifetime gross moves 191. Nearly all of the reordering is the age adjustment.

Progressive tax works. It just does not reorder.

Tax is doing something — the share kept falls as pay rises.

The share kept ranges from 61.5% to 79.5%. It still does not reorder because the relationship is monotonic: higher-paying companies keep less, but never enough less to overtake anyone.

What does move it — average age

The estimate scales an industry age-wage curve to each company's average pay. At the same average pay, a company with a younger workforce gets a larger multiplier and so a higher lifetime figure.

Does the estimate match real graduate pay?

After the correction, the figure at 22–24 is the industry level itself (the company premium is zero at 22). We checked that level against a separate government table — starting pay for new graduates, converting the monthly figure to an annual one on this site's usual assumption that bonus is 19% of annual pay.

Across 3,780 companies the estimate comes to 0.90–1.14× the published graduate figure (median 0.98×). Before the correction the median was 1.16× and the maximum 5.26×.

What this does not show. Since the corrected figure at 22 is by construction exactly the industry level, this only demonstrates that two government tables agree. How the premium rises between 22 and the company's average age remains our own assumption, and this check cannot validate it.

Before you rely on this

• Lifetime earnings are estimated. A company discloses one average salary figure and nothing about its pay progression, so the "age adjustment effect" measured here is partly a property of the estimation method itself.
• The shape of the premium is our assumption. Applying the multiplier flat across all ages implied ¥13.3m for a 22–24 year old at Mitsubishi Corporation — implausible. It now rises from the industry level at 22 to the company's multiplier at its average age.
• 38 years at one employer, age 22 to 60. No job changes, no re-employment, and retirement payouts are excluded.
• Tax assumptions: bonus at 19% of annual pay, no dependants, health insurance at 9.9% plus the 0.23% child-care support levy (FY2026).