Capital gains taxes and the lock-in effect that keeps money frozen in place

Picture a toll booth on a road, with one odd rule. You pay only when you leave. Every driver on the road knows this, and so every driver has an excellent reason never to leave. The road fills with cars that are not going anywhere, while better destinations sit empty. A traffic engineer would call that a design failure. The Internal Revenue Code calls it the capital gains tax.

This post explains the lock-in effect, a feature of the capital gains tax that most people have never heard of and that costs the economy more than the headline rate suggests. It is an explainer, so we define first, then work the arithmetic, then argue.

Terms

A capital asset is property held for investment or personal use: stocks, bonds, real estate, a business, a painting. A capital gain is the profit when you sell it for more than you paid. Your basis is, roughly, what you paid for it. The gain is the sale price minus the basis.

Realization is the event that triggers the tax: the sale. A gain that has not been sold is an unrealized gain. It may be large on paper, and the tax code ignores it. Taxes on capital gains are owed only when the gain is realized.

That is a sensible rule in one respect. You should not owe tax on money you have not received and may never receive, since prices fall as well as rise. The side effect is the subject of the rest of this post.

The lock-in effect, with arithmetic

Because the tax falls only on sale, the owner of an appreciated asset faces a choice whenever a better investment comes along. Sell the old asset and move the money, and pay tax now. Or keep the old asset and pay nothing.

Here is a hypothetical, with round numbers and an assumed 20 percent rate on long-term gains. It is not a statement of the law's actual rate, which differs by income and includes other taxes. An investor holds stock worth $1,000,000 that she bought for $100,000. Her unrealized gain is $900,000.

If she sells, the tax is 20 percent of $900,000, or $180,000. She has $820,000 left to reinvest. If she holds, she keeps the full $1,000,000 working.

Now suppose a better opportunity appears. How much better must it be for her to bother? The new investment starts with $820,000 and must grow to match what the old one would be worth. Divide $1,000,000 by $820,000 and you get about 1.22. The new investment must grow by roughly 22 percent just to be worth what the old one is worth today. A modestly better opportunity, even a decidedly better one, does not clear that hurdle. So the rational investor stays put.

That is the lock-in effect. The tax does not merely collect money from those who sell. It discourages the selling, so that capital sits where it is, even when it would do more good somewhere else. The economy loses the difference between what the capital earns where it is and what it would earn where it could go. A free market is supposed to move capital to its best use. A tax on the move slows that process.

There is a second, quieter effect. An owner who holds an asset until death passes the asset to heirs with a stepped-up basis: the heirs' basis becomes the value at death, and the gain accumulated during the owner's life is never taxed at all. The result is that for older owners the tax rate on holding is zero and the rate on selling is the full rate. That is an enormous incentive to hold until death. We take up the estate side of this in the death tax, the family farm and the case for ending a tax on dying. For now notice that the lock-in effect is strongest exactly where the gains are largest, and where the owners are oldest.

The double layer

A second argument concerns corporate stock. When a corporation earns a profit, it first pays the corporate income tax, 21 percent at the federal level since 2018. What remains belongs to the shareholders. If the company retains the money and the share price rises, the shareholder pays capital gains tax when she sells. If it pays a dividend, the shareholder pays tax on the dividend.

Two layers of tax fall on the same earnings. In a round hypothetical, $100 of corporate profit becomes $79 after the 21 percent corporate tax. A 20 percent tax on the $79 leaves $63.20. The combined burden is $36.80 on $100, or 36.8 percent. Compare that with income taxed in only one layer. The capital gains tax is therefore partly a second tax on income that has already been taxed once, and this is one reason many economists have favored a lower rate on gains than on wages. We explain who bears the corporate layer in why the corporate income tax is paid by people, and the only question is who.

Taxing inflation, not profit

Now the argument that tends to surprise people. The tax code measures a gain in dollars. It does not adjust the basis for inflation. Over a long holding period, a good part of a nominal gain may simply reflect a decline in the purchasing power of the dollar, not any real increase in wealth.

A labelled hypothetical makes it plain. Suppose an investor buys stock for $100,000. Over many years, prices in the economy double, so that it takes $200,000 to buy what $100,000 bought at the start. The stock keeps pace with prices and is sold for $200,000. The investor has made no real profit at all. She can buy exactly what she could have bought if she had never invested.

The tax code sees a $100,000 gain. At a 20 percent rate, the tax is $20,000. She keeps $180,000. In terms of the original purchasing power, $180,000 buys what $90,000 would have bought at the start. She began with $100,000 of purchasing power and ended with $90,000. She paid a tax of $20,000 on a profit of zero, and ended up poorer in real terms. The tax rate on her real gain was not 20 percent. It was infinite: a positive tax on a zero gain.

Defenders of the system sometimes point out that the rate is lower than the rate on wages, which partly compensates. It is a fair point for some holders and some periods, and not for others. The compensation is a rough rule of thumb. The inflation distortion varies with the holding period and with the inflation rate, and it is worst for long-term holders in high-inflation years. A tax that depends on the accident of how fast prices rose is not a principled tax. Indexing the basis for inflation would fix this, and proposals for it have come up repeatedly. It has not been enacted for the broad run of assets. For why inflation itself is a tax in another sense, see inflation hurts the poor most, and the arithmetic of a fixed paycheck.

A short history of the rate

The rate on capital gains has moved around a great deal, and the history is instructive.

Through the 1970s the top rates on gains rose until the effective maximum, once the minimum tax and other provisions were counted, stood at about 49 percent by 1978, by one Treasury account. In 1978 Congress passed the Revenue Act of 1978, which included an amendment by Representative William Steiger of Wisconsin to cut the capital gains rate, over the objection of the Carter administration. The top rate fell to 28 percent. In the years after the cut, realizations of gains and the venture capital industry both grew, a fact its supporters cite as evidence that the earlier rate had been choking transactions.

The Tax Reform Act of 1986 raised the top rate on gains to 28 percent (33 percent for some taxpayers in phase-out ranges), treating gains like other income, in exchange for lower rates elsewhere. Then the Taxpayer Relief Act of 1997 cut the top rate on long-term gains to 20 percent. Supporters point to rising revenue from the tax in the years that followed.

I want to be careful here, because this is a place where partisans overclaim. Rising revenue after a rate cut does not prove that the cut caused the rise, since the late 1990s saw a booming stock market, and a boom raises gains no matter what the rate is. A fair reading is narrower: for a tax on voluntary sales of assets, a lower rate tends to unlock some transactions that a higher rate would suppress, so the revenue loss from a cut is smaller than a static estimate says, and in some ranges the revenue may even rise. There is a rate beyond which more rate produces less revenue, and many economists believe that the revenue-maximizing rate for gains is not very high. Where exactly that point lies is disputed. This is a case where the dynamic effect is genuine, and the honest discussion of when it is and is not is in where tax cuts pay for themselves and where they honestly do not.

The steelman

The best case against cutting the capital gains rate is a fairness argument.

Most capital gains go to the wealthiest households. Taxing the income of someone who lives off investments at a lower rate than the income of someone who lives off a paycheck is a gift to the rich. The gains of a billionaire's stock holdings can go untaxed for decades, and then vanish at death through the step-up. The system rewards those who already have capital.

It deserves a serious answer. I will give it in numbered points.

1. The step-up in basis is a real problem, and the answer is not to raise rates on everyone. If gains vanish at death, the remedy is to end that rule or to tax gains at death with protections for farms and small firms. That fixes the worst abuse directly. Raising the rate on sale, by contrast, increases lock-in and so worsens the very problem that the objection describes.

1. Capital is the source of wages. The factory, the software, the tools and the training that make a worker productive were paid for by investment. A tax that reduces the reward for investing reduces investment, and workers pay for it in lower wages over time. The wealthy are the visible owners, but the people who feel the loss of investment are the ones who would have worked at the next plant.

1. Gains are partly inflation and partly double taxation. The lower rate on gains is a crude offset for those two features. If you want to equalize the rate on wages and gains, you must also index the basis and integrate the corporate tax, or the result is a rate on real, after-corporate-tax gains that is far higher than on wages.

1. Realization produces a voluntary tax. Nobody is forced to sell. A government that raises the rate on gains finds that the base shrinks, since the sale is a choice. The higher the rate, the more people choose not to realize. A tax on a voluntary event raises less than it appears.

A fair person can accept some of these points and not others. What a fair person cannot do is pretend that the choice is between taxing the rich and not taxing them. The choice is between different designs, with different effects on investment, mobility of capital and the rest of the economy.

What the Founders thought of capital

Alexander Hamilton's argument for a national bank and a funded debt rested on the idea that capital in the hands of enterprising people is the engine of national wealth, and that a government that treats capital well attracts more of it. His Report on Public Credit, which we discuss in Hamilton's 1790 Report on Public Credit and what honest borrowing requires, was built on that premise. A tax design that traps capital in place works against the premise. The Constitution's grant of the taxing power in Article I, Section 8 leaves Congress free to design the capital gains tax as it likes. The question is only whether it does so wisely.

Reforms worth debating

• Index the basis for inflation, so that only real gains are taxed.
• End or reform the step-up in basis at death, or replace it with a carryover basis with exemptions for family farms and small firms, so the incentive to hold until death disappears.
• Lower the rate, or phase it down with the holding period, so that the longer an asset is held the lower the rate, which reduces lock-in directly.
• Integrate the corporate and individual layers, so that corporate income is taxed once.

Each of those has costs and a political price, and none of them is free. But all are aimed at the same goal: a tax that raises revenue without freezing capital in place or taxing the inflation that the government itself caused.

The case in brief

1. Capital gains are taxed only when an asset is sold, so owners with large unrealized gains have a strong reason never to sell, a distortion called the lock-in effect.
2. In a labelled example with a 20 percent rate, an investor holding $1,000,000 of stock with a $100,000 basis would need a new investment to earn about 22 percent more just to break even after selling.
3. Corporate earnings are taxed twice, once at the corporate level and again as gains or dividends, which in the round example makes a combined 36.8 percent on $100 of profit.
4. Because the basis is not indexed, part of the tax falls on inflation, so an investor whose stock merely kept pace with doubled prices still owes tax on a gain that is zero in real terms.
5. The Steiger amendment of 1978 and the 1997 rate cut show that a lower rate can unlock sales, though rising revenue also reflected booming markets, and better designs would index the basis and fix the step-up.

What you can do: ask your representative whether they support indexing the basis of capital assets for inflation, and look up the holding-period rules in the IRS instructions for Schedule D before you sell anything.