Stock gains are not income!

EconLog and Marginal Revolution consider the question of whether capital gains count as savings. Arnold Kling says this is equivalent to asking whether capital gains count as income. Well, it seems to me that there is a clear and simple argument that for stocks, capital gains are not income.

The price of a stock is (generally, and approximately) equal to the present value of its future stream of dividends. Hence if the price goes up, the expected stream of dividends has gone up. Hence if you count the price gain as income, you are *double-counting* that income - counting it once as income now, and counting it again when it comes in as part of the dividend stream. While Kling is right that "a re-valuation of assets produces no output" it does indicate an expected increase of output in the future. But if we're going to count that future income, we can't count it again in the present.

Is the same true for other capital assets? For assets which generate a stream of cash (say, a rental property), the same argument holds - count the income as it arrives, not in the current asset value. Other assets are more complicated, and the answer is not as clear to me.

Share this

Don - while you have given

Don - while you have given up your right to those future dividends, they still exist and someone is getting them. If you count the sellers realized capital gains as income, and the buyers dividends as income, that is double-counting.

Remember we are talking about net national income here, so if you have sold to someone else in the country, how much has output changed? It has changed by the jump in dividend stream, not the jump in dividend stream plus the jump in value.

Brandon - why would stock price fall from a predicted dividend? Consider a world where the interest rate is 10%. A stock is expected to pay out $1/year forever. The stock is thus valued now at $10. It's price does not change on its yearly payout. If new information comes in that tells you the stock will pay out $1.10/year forever, then its price will jump to $11 - and remain there, regardless of how many $1.10 dividends it pays out (unless new information comes in).

This is a good example to illustrate my point. If this share of stock is the only wealth in our little country, what is national income? It is $1.10/year, not 1$ in the year the price jumped, AND $1.10/year.

If the market value of your

If the market value of your assets has increased, that's income. There's no double-counting, because when the dividend payout finally comes, it's cancelled out by the resulting fall in stock price.

Patri, The price of a stock

Patri,

The price of a stock is (generally, and approximately) equal to the present value of its future stream of dividends. Hence if the price goes up, the expected stream of dividends has gone up. Hence if you count the price gain as income, you are double-counting that income - counting it once as income now, and counting it again when it comes in as part of the dividend stream. While Kling is right that “a re-valuation of assets produces no output” it does indicate an expected increase of output in the future. But if we’re going to count that future income, we can’t count it again in the present.

The question of double-counting disappears if you only count REALIZED capital gains as income. If you have an actual realized capital gain that results from the sale of the holding, one on which you are responsible for taxes, then future dividends are no longer of any importance to you as you have given up your right to them.

Regards, Don

"The price of a stock is

"The price of a stock is (generally, and approximately) equal to the present value of its future stream of dividends."

That's only a theory in a finance textbook, you'd be hard pressed to find any stocks in today's markets priced on that basis (except perhaps some oil and gas trusts or REITS).

>The price of a stock is

>The price of a stock is (generally, and approximately) equal to the
>present value of its future stream of dividends.

Uh, could you explain all the companies that don't have dividends and yet still have positive stock prices?

Patri, ... while you have

Patri,

... while you have given up your right to those future dividends, they still exist and someone is getting them. If you count the sellers realized capital gains as income, and the buyers dividends as income, that is double-counting.

If you don't sell a stock, but rather just mark it to market daily, then every day that the stock goes up you have a positive income, and every day that it goes down, you have a negative income. If everything else is equal, a stock that goes ex-dividend, meaning that a buyer of the stock will no longer receive the current dividend payment, will normally fall in value by an amount that approximates the value of the current dividend payment. This means that a holder of the stock that receives a dividend income payment will also experience a negative income from the fall in the value of the stock.

Overall, as dividends are paid one by one, the market value of the stock falls in accordance. This means that you can either realize the discounted present value of any unpaid future dividends, or a single actual dividend plus the discounted present value of all the rest of the future dividends. I don't see any double counting from this.

Regards, Don

Stormy - I think the

Stormy - I think the keywords here are "present value" and "future". In short holders of these stocks believe there is a present value for a future stream of dividends. as with all things financial, they could be mistaken.

Stormy, a company that does

Stormy, a company that does not currently pay dividends, or is even profitable, will presumably be profitable in the future and in the future pay dividends, and it is the present value of those future dividends that approximate the stock price.

Brandon - why would stock

Brandon - why would stock price fall from a predicted dividend?

Not from a predicted dividend. From a paid dividend.

Consider a world where the interest rate is 10%. A stock is expected to pay out $1/year forever. The stock is thus valued now at $10. It’s price does not change on its yearly payout.

A stock that pays a $1 dividend once a year starting tomorrow is worth nearly $1 more than a stock that pays a $1 dividend once a year starting yesterday.

While you have given up your right to those future dividends, they still exist and someone is getting them. If you count the sellers realized capital gains as income, and the buyers dividends as income, that is double-counting.

If I buy a stock at $10, and it goes up to $11 due to (let's say accurate) expectations of greater future earnings, only current holders of the stock realize any economic profit from the increase. I get an extra dollar when I sell the stock, but the buyer has to pay an extra dollar to get the extra dime per year.

I've been investing in the

I've been investing in the stock market since 1997. My returns beat those of most mutual funds out there, though as Patri noted in recent posts, this is probably due mostly to luck. I have never received a dividend. I almost never use the dividend yield to make any buying/selling decisions. Somehow I doubt that the >95% of tech companies that don't give out dividends ever plan to.

I think dividends are one reason among many that people buy equities. Among others -

- speculation, believing that the price at a point later in time will be higher than the time of buying, regardless of the dividend paid
- store of value, believing that equities will preserve their value better than cash or other investments. An example would a person who sees a danger of inflation and invests in commodity-related companies. (Which I think is one reason for the rise of such stocks for the past 4-5 years or so.)
- they're in love with the stock. Irrational, but people do it.

Would there be anything wrong with replacing the phrase "equal to the present value of its future stream of dividends" with the broader phrase "equal to the present value of its future stream of returns"? I realize that it's a bit circular (what are those returns based on?) but I don't see dividends being the main determinant of share price.

I hate to correct Arnold

I hate to correct Arnold Kling, but conflating "savings" and "income" adds to the confusion. They're not the same thing.

Income is money that flows into your wallet, and is either used for consumption (spending), or is saved (i.e., put away for future consumption/spending.)

In this context, it is clear that unrealized stock growth does count as an increase in savings, but it does not count as income until such time as it is converted to cash, at which point it is either spent or put back into savings.

In any single period, savings growth = (income - spending) + capital gains.

A capital gain does not become income until it is realized. Some capital assets have regular, periodic realizations, such as coupons on a bond, rent from real estate and dividends on stocks. They also have a realization when they are sold.

IMO, any asset that can be converted into income in the future should be counted as savings, but the tricky part is figuring out how big those income streams will be, and when they will be realized. and while we're looking at the overall evaluation of "human welfare", we should not overlook the use value of depreciating assets. Foe example, my car will be worth less next year than this year, but the benefit I derive from owning the car is probably greater than the benefit I would derive from placing the money I spent on my car into savings.

But that's a whole other article.

Suppose I don’t sell the

Suppose I don’t sell the stock. If we count the increased capital value as profit now and the future dividends, don’t you agree we are counting twice?

Three times. But two of them cancel each other out. The seller has an extra dollar in income, and that's canceled out by the fact that the buyer has an extra dollar in expenses. So you're right, capital gains don't immediately increase the total wealth of the economy. But they do increase the total wealth of the individual.

Brandon: A stock that pays a

Brandon:

A stock that pays a $1 dividend once a year starting tomorrow is worth nearly $1 more than a stock that pays a $1 dividend once a year starting yesterday.

You are right, the stock would fluctuate, since the present value of the dividend stream changes on a yearly cycle.

If I buy a stock at $10, and it goes up to $11 due to (let's say accurate) expectations of greater future earnings, only current holders of the stock realize any economic profit from the increase. I get an extra dollar when I sell the stock, but the buyer has to pay an extra dollar to get the extra dime per year.

Yes, I agree that only current holders get an economic profit. But I don't see how it changes my point about double-counting. Suppose I don't sell the stock. If we count the increased capital value as profit now and the future dividends, don't you agree we are counting twice?

Yes, dividends is of course

Yes, dividends is of course a simplification, and to be more accurate you should include any return to the shareholders, such as a stock buy-back, one-time cash settlement, etc.

The entire argument is based

The entire argument is based on the false assumption the stock prices are determined by dividend payouts. Stock prices are based on what someone is willing to pay for them (i.e. what the market will bear), not anything else. Although some investors will look at dividend yield as a way of calculating what they are willing to pay for it, there are many other measurements of a stock's value which are more commonly used. Many companies do not pay dividends at all, and I highly doubt that any smart investor buying those companies' stocks are doing so on the anticipation that they will someday start.

Replacing "dividends" with

Replacing "dividends" with "returns" would be a nice way of taking into account the Modigliani-Miller theorem, which says among other things that dividend payouts and stock buybacks are equivalent from the shareholder's point of view. IIRC that theorem has been a factor in reducing the frequency with which companies pay dividends, though the recent exemption of dividends from taxation may change that back.

Seems like you really need some way of taking that into account. Suppose that (in a world w/o tax distortions) a company decides to just buy back its own stock periodically, increasing the value of the remaining stock, and never pay any dividends. Clearly that represents some income flowing to all shareholders. Right?

Patri, I think your initial

Patri,

I think your initial conclusion (capital gains are not income) is correct, but using dividend discounting took us off on a tangent.

Gordon's dividend model (P=D/(k-g) is a good place to start for stock valuation, but it has been widely discounted (couldn't resist) as too simplistic and unrealistic. For example, if g and k are close, price will be too high; if g>k, price is negative. As pointed by other posters, there are many stocks that do not pay dividends, but have a market price. One of the reasons high tech firms don't pay dividend is that their growth rate is so high it makes more sense to reinvest.

Aside from the question of how stock prices are obtained, dividends are not capital gains because nothing has been sold. Dividends are income and are taxed as such. Likewise with bond income, rental fees, etc. although they fall into different income categories. If capital gains were income, we would all have to submit asset values every year to the IRS.

Patri, you are neglecting

Patri, you are neglecting the notion of a horizon when calculating the present value of a future revenue stream.

Say you had an asset that would pay $10 per year for 10 years. For some given discount rate, at year zero the asset would be worth the ten payments, each discounted from the future time of payment to the present. At year one, there would only be nine future payments, so the asset price would go down, but as the holder of the asset you would have gotten one payment.

If you sold the asset to Bob, then at year two Bob would see a capital loss from the asset losing value due to there being only eight payments left, but he would have gained another payment. If Bob holds it to maturity, the asset will have lost all of its value (a capital loss) but Bob will have gotten all of the payments (income). From Bob's perspective, he paid some amount X for the asset, lost all of it, and gained $80 from holding it to maturity. After taking into account the time value of money, Bob came out even. Counting the capital loss as negative income and the payments as income isn't double-counting.

As Brandon says, the value of a stock goes down every time there is a dividend paid. The reason that you can think of a $10 stock paying $1 every year into perpetuity without the stock price ever falling is that there is essentially a rolling horizon. The market price isn't really based on the assumption that it will pay $1/yr forever; with every year into the future that you discount back to the present, there is also an increase in the discount rate owing to the increased uncertainty about the future. Maybe thirty years from now the dividends will stop, or the company will have gone out of business. That uncertainly establishes a virtual horizon.

Every day that goes by extends that horizon by one day. It also brings you one day closer to the next dividend payout, which increases the value of the stock, until the dividend is paid out and the value falls again. The daily increase in the value of the stock is a capital gain, which counts as income. On the day of the payout, the stock loses value, which is a capital loss, which counts as negative income. But the payout itself is positive income.

There's no double-counting.

As an aside, the question of when or whether you realize your capital gains is irrelevant. Asset appreciation is income whether you move your money out of the asset and into cash or whether you hold it.