Taxing Unrealized Capital Gains and Interest Rate Policy

First read Tyler on the practical difficulties implementing a tax on unrealized capital gains!

I have a different argument that I rarely see discussed. A significant fraction of what we call capital gains is due to variation in the discount rate rather than variation in income. Take the simplest Gordon model of stocks P=D/r where D is the annual dividend and r is the discount rate. If D=100 and r=.1, for example, then the stock is worth 100/.1=$1000. Now suppose people become more patient and the discount rate falls to .05 then P=$100/.05=$2000. The stock price doubles, a massive capital gain. But notice that income hasn’t gone up at all. It’s still D per year. Income hasn’t gone up and lifetime consumption possibilities haven’t gone up for someone who doesn’t sell (but recall this is a tax on unrealized gains. If there is a sale then tax the realized gain.) Ultimately, we want to tax consumption so we should not be taxing “capital gains” which reflect changes in discount rates rather than changes in income or consumption possibilities.

Taxing unrealized capital gains also connects interest rate policy even more tightly with fiscal policy. Need a tax boost? Lower interest rates! Fed policy already influences taxes but this adds another lever for political business cycles. More generally, interest rate volatility now adds to fiscal volatility. When we exited zero interest rate policy, for example, banks had huge capital losses. As rates fall, capital gains increase. Do we really want to add the tax system to this? 

If we generalize the Gordon model to P=D/(r-g) where g is the growth rate of dividends then we can see that another cause of increased capital gains, an increase in g. It’s not obvious that we should tax unrealized changes in asset values due to increases in the growth rate of dividends. On the one hand, this is more income-like but it’s expectational. It’s taxing the chickens before the eggs have hatched.

The one clear increase in income which should be taxed is increases in D. An unrealized capital gains tax would do that but at the expense of also taxing changes in asset values due to changes in r and g which should not be taxed.

Now add the point I mentioned to Tyler, which is that taxing unrealized gains divorces the entrepreneur from the firm at a time when the “marriage” is likely at its most productive. Not good. Taxing unrealized gains might not even be a good idea from the point of view of the tax collector. Does the IRS want to tax X now or a much larger figure later? If the IRS taxes entrepreneurship too early it can reduce total discounted tax revenues.

Bottom line: I don’t see how taxing unrealized capital gains is a well thought out policy. Eliminate the stepped up basis, declare victory and go home.

Addendum: Aguiar, Moll, and Scheuer make some similar points but embedded in a fully GE framework. Ben Moll also points me to earlier pieces by Frank Paish 1940, Nicholas Kaldor 1955 and John Whalley 1979.