Each year the simulator applies growth, dividends, and salary to beginning-of-year wealth, deducts living expenses, then applies four taxes: a flat-rate wealth tax on the portfolio, ordinary income tax on dividends and salary, and long-term capital gains on assets liquidated to fund the wealth-tax bill itself.
The cap-gains line is the subtle one. To raise W·t in cash, the portfolio must sell enough to cover both the wealth tax and the gains tax owed on the sale. That sale triggers more gains, which require another sale — a geometric cascade. Assuming a near-zero cost basis (typical for founder equity or long-held positions, and conservative against the household), the cascade resolves in closed form:
Capital-gains tax = W · t · ( c / (1 − c) )
where W = beginning wealth
t = wealth-tax rate
c = long-term capital-gains rate
At the default settings (t = 5%, c = 40%), the cascade adds 3.33% of wealth per year on top of the headline 5% wealth tax — meaningfully more than a single-pass Schedule D estimate would suggest.
Breakeven growth, derivation
Setting net annual change to zero and solving for g:
ΔW = W·g + W·d(1−i) + S(1−i) − L − W·t/(1−c) = 0
g = t/(1 − c) − d(1 − i) + ( L − S(1 − i) ) / W
For a $1B portfolio at the default rates, this works out to roughly 7.88%. The simulator solves the same equation numerically over the full ten-year horizon (since wealth itself moves), which is what the "Breakeven growth" stat reports.
What this model does not attempt
Tiered or graduated wealth-tax brackets, dynamic capital-gains realization timing, charitable offsets, basis step-up at death, state-residency arbitrage, and behavioral responses (relocation, asset re-characterization, leverage) are all out of scope. The point is to isolate the arithmetic of a single-rate wealth tax sitting on top of the existing income- and gains-tax system.
This is a thought tool, not tax advice. Anyone making real decisions should consult an actual professional.