Part 1 of 2. A fortune is not a pile of idle goods — and a flat tax rate on capital hides a bite that grows with the horizon, under every design on the table, and under the one we already have.
Ok, I see your point now. Your argument is internally consistent. But it measures from a baseline the post never defends. "Anything that increases your wealth is income" is the Haig-Simons definition — a definition, not a theorem. Definitions choose tax bases; they cannot tell you which base is right. Once you define income to include accruals, deferral becomes a zero-interest loan, realization becomes a subsidy, and a loss year becomes forgiveness — all true relative to that baseline. Whether that baseline is the neutral one is the entire question. The arithmetic says it is not: an annual tax on accruals is a tax on waiting, and a flat rate on waiting compounds. Today's 23.8%, applied annually to a 5% real return, removes 37% of a working life's deferred consumption and two-thirds of a century's. Measured from the consumption baseline, the picture inverts exactly — accrual taxation is not the calling-in of a subsidized loan; it is a demand for prepayment on meals not yet eaten, with the compounding wedge as the interest charged. Both frames are coherent. Adjudicating between them is what the optimal-tax literature exists to do — and note that the benchmark result of that literature is that the correct rate on capital is zero, with the later corrections finding positive optima that are contested in size and, in every serious model, priced against the wedge as a first-order cost. Zero is still on the table. It has never left the table.
On one issue, I would go further than you — conditionally. The convexity you describe — appreciate and the loan grows, die and it is forgiven — is real and indefensible. But its source is step-up at death and buy-borrow-die, not realization itself. So what I say is this: if capital is to be taxed at all, then real gains, indexed, taxed exactly once at sale, dominates both the system we have and the mark-to-market machinery — no annual valuations of private companies, no bills on unsold assets, no refunds mailed to billionaires in crash years, no taxing of phantom inflation. Whether the rate on that best-designed base should be 15%, 5%, or zero is exactly what the models fight about, and the most reasonable reading of that fight is: low, possibly all the way down to zero.
On arbitrage: taxing every accretion identically does close the relabeling door — but arbitrage-freeness alone proves too much, since a 100% tax on everything is also arbitrage-free. The relabeling problem puts a floor under the capital rate; it does not raise the floor to the labor rate, and how low that floor can sit is a question of legal drafting, not economics. Draft well and the floor may sit very near the ground if not on it.
Not sure I understand the point. If you mean that in many cases, we socialize risks (such as investment banks, commercial banks and so on) sure: in that case the bearer of the risk is the taxpayer. If you mean private capital financing productive enterprise, then I don't understand what you mean.
Anything that increases your wealth is income which means the form of what causes your wealth change is irrelevant. How you chose to utilize your wealth at any point in time is your discretion - money in the bank or new car in the garage. This would ensure the tax system is arbitrage free so people are not incentivized to non economically justified activity to shift income that leads to ordinary income into capital. Under this framework, taxing capital only at “recognition” is a zero interest loan from the taxpayer. What is even better is the convexity of these loans - if the asset continues to appreciate the taxpayer gets more loans at zero interest, if the asset depreciates the loan is forgiven. As you point out this interest free loan becomes completely dominant in the long term as zero interest loans would.
Ok, I see your point now. Your argument is internally consistent. But it measures from a baseline the post never defends. "Anything that increases your wealth is income" is the Haig-Simons definition — a definition, not a theorem. Definitions choose tax bases; they cannot tell you which base is right. Once you define income to include accruals, deferral becomes a zero-interest loan, realization becomes a subsidy, and a loss year becomes forgiveness — all true relative to that baseline. Whether that baseline is the neutral one is the entire question. The arithmetic says it is not: an annual tax on accruals is a tax on waiting, and a flat rate on waiting compounds. Today's 23.8%, applied annually to a 5% real return, removes 37% of a working life's deferred consumption and two-thirds of a century's. Measured from the consumption baseline, the picture inverts exactly — accrual taxation is not the calling-in of a subsidized loan; it is a demand for prepayment on meals not yet eaten, with the compounding wedge as the interest charged. Both frames are coherent. Adjudicating between them is what the optimal-tax literature exists to do — and note that the benchmark result of that literature is that the correct rate on capital is zero, with the later corrections finding positive optima that are contested in size and, in every serious model, priced against the wedge as a first-order cost. Zero is still on the table. It has never left the table.
On one issue, I would go further than you — conditionally. The convexity you describe — appreciate and the loan grows, die and it is forgiven — is real and indefensible. But its source is step-up at death and buy-borrow-die, not realization itself. So what I say is this: if capital is to be taxed at all, then real gains, indexed, taxed exactly once at sale, dominates both the system we have and the mark-to-market machinery — no annual valuations of private companies, no bills on unsold assets, no refunds mailed to billionaires in crash years, no taxing of phantom inflation. Whether the rate on that best-designed base should be 15%, 5%, or zero is exactly what the models fight about, and the most reasonable reading of that fight is: low, possibly all the way down to zero.
On arbitrage: taxing every accretion identically does close the relabeling door — but arbitrage-freeness alone proves too much, since a 100% tax on everything is also arbitrage-free. The relabeling problem puts a floor under the capital rate; it does not raise the floor to the labor rate, and how low that floor can sit is a question of legal drafting, not economics. Draft well and the floor may sit very near the ground if not on it.
The whole game is tax free compounding - run the numbers on Buffett. Letting Capital earn returns on zero interest loans from the taxpayer is absurd
Not sure I understand the point. If you mean that in many cases, we socialize risks (such as investment banks, commercial banks and so on) sure: in that case the bearer of the risk is the taxpayer. If you mean private capital financing productive enterprise, then I don't understand what you mean.
Anything that increases your wealth is income which means the form of what causes your wealth change is irrelevant. How you chose to utilize your wealth at any point in time is your discretion - money in the bank or new car in the garage. This would ensure the tax system is arbitrage free so people are not incentivized to non economically justified activity to shift income that leads to ordinary income into capital. Under this framework, taxing capital only at “recognition” is a zero interest loan from the taxpayer. What is even better is the convexity of these loans - if the asset continues to appreciate the taxpayer gets more loans at zero interest, if the asset depreciates the loan is forgiven. As you point out this interest free loan becomes completely dominant in the long term as zero interest loans would.