David R. Henderson  

Did the Dot.Com Boom Boost GDP Much?

Gilens vs. the Political Exter... Ed Crane: An Appreciation...

In my August 26 post, "U.S. Federal Budget Cuts in the 1990s," I pointed out that U.S. government spending fell from 21.8 percent of GDP in 1990 to 18.4 percent of GDP in 2000, a 3.4-percentage-point drop. I focused on the numerator--government spending--but I did point out that the denominator, GDP, had done nicely too.

One commenter argued that my reasoning was misleading because of the dot.com boom in the last half of the 1990s. Presumably, his point was that the dot.com boom boosted GDP. I'm guessing that he's right. But I don't think it boosted GDP as much as many people think. And the reason has to do with the nature of GDP.

The major effect of the dot.com boom was to boost capital gains. How do capital gains show up in GDP? They don't. GDP = C + I + G, where C = spending on consumption, I = spending on investment, and G = government spending on goods and services. Do you see capital gains in there? I don't either. GDP does not include capital gains.

Of course, one can argue that capital gains made people wealthier and that they used that wealth both to invest in real plant and equipment (thus boosting I) and to buy consumer goods and services (thus boosting C.) But that has to be the argument. Capital gains per se are irrelevant.

Comments and Sharing

COMMENTS (18 to date)
OneEyedMan writes:

Some capital gains is profits not paid out as dividends or interest. That should be counted as GDP under the GDI measure.
Y = r*L + w*L, r*L is capital income.

That is, if corporate profits went up a lot that should be counted in GDP regardless of if that actually gets paid out or retained.

Jonathan writes:

I think this line of explanation made the issue more opaque. The last two paragraphs don't provide a clear explanation of why capital gains income is not a part of GDP.

David could have written: "How does labor income show up in GDP? It doesn't. GDP = C + I + G… Do you see labor income in there?… Of course, one can argue that they used that income to invest in real plant and equipment (thus boosting I) and to buy consumer goods and services (thus boosting C.) But that has to be the argument. Labor income per se is irrelevant."

This is misleading. It's taking the expenditure-approach definition of GDP to imply that the income-approach definition of GDP isn't relevant. This is obviously wrong.

If capital gains didn't show up in GDP, it must be because the dot-com-era capital gains were unrealized, so asset wealth on paper never actually became income. Maybe this is what David was getting at, but the last two paragraphs of the post don't convey that to me.

Capital gains income appears to have been pretty large in the late 1990s; for example, take a glance at Figure 3 of Atkinson, Piketty, and Saez (JEL 2011) or http://taxfoundation.org/article/great-recession-and-volatility-sources-personal-income

Nathan Smith writes:

To Jonathan: No, labor income *is* included in GDP. Directly. So is capital income, that is, actual profits.

I think the dot.com boom boosted GDP through the Tobin's Q mechanism. When asset prices rise above the replacement cost of capital, it's time to start making more capital. Investment was booming in the late 1990s and has been relatively weaker ever since.

(I guess I mean the high stockmarket boosted GDP, maybe not the dot.com boom per se.)

Jon Murphy writes:

Jonathan and Nathan:

Question for both of you:

Let's say a person earns income (doesn't matter the source). He takes the cash and sits on it. Does not spend a penny.

How will his income show up in GDP? (Hint: it won't).

Income (and its source) is irrelevant. It is not until you do something with that income does it get recorded in GDP.

OneEyedMan writes:

What definition are you working off of?

happyjuggler0 writes:

I suspect that Deirdre McCloskey would say that you don't need capital gains per se; instead an increase in the *dignity* of bourgeois dotcommers could have (did?) encouraged increased entrepreneurship in the dot.com field, resulting in the alleged increase in GDP due to the dot.com boomlet.

Brandon Berg writes:

Federal spending during the 90s was almost perfectly flat after adjusting for inflation and population growth. This was the only time in living memory that Federal spending had remained flat for anywhere near that length of time. Spending as a percentage of GDP would have fallen somewhat even with very weak real GDP growth, though obviously it wouldn't have fallen as far.

Ghislain writes:


Let's say a person earns income (doesn't matter the source). He takes the cash and sits on it. Does not spend a penny.
How will his income show up in GDP? (Hint: it won't).

The source of the gain matters: If the source of his gains are "just" increases in stock price, it does not count in GDP. But if his gains are someone else expenses, it does count in GDP.

And if he then buys something, it counts a second time...

David R. Henderson writes:

You're right that I should have used the income method. My conclusion is correct, though. Even using the income method, and even if the person realizes his capital gains as income, the part of income that is realized capital gains is not counted in GDP.

Brandon Berg writes:

Another way of looking at it is that GDP stands for Gross Domestic Product, and capital gains aren't product. If I buy a stock at $30 and sell it at $60, the sale produces nothing.

Now, in a healthy stock market, the capital gains are going to reflect investment. The stock appreciates because the firm made real investments, which does contribute to GDP via the I component. Including the capital gains in GDP would be double-counting.

Moreover, during a speculative boom, the gains may not reflect real investment. If I buy a stock for $30 today, and the next day it rides a wave of irrational exuberance up to $60, there's nothing real happening there. Including these capital gains in GDP wouldn't be double-counting; it would be counting something that ought not be counted at all.

I'm oversimplifying, of course. In the real world, stock prices are driven by a complex mix of real investment, consumer behavior, investor sentiment, and spread of information. But the basic point---that contributions to capital gains which should be included in GDP are counted elsewhere---remains.

I think focusing on cap gains is missing the point. The dot.com boom boosted employment and thus growth (output). Anyone remember Paul Krugman denying the economy could grow 4% annually, just before it did so four years in a row?

The Budget Reform Act of 1990 had constrained the growth in Federal spending (thank you, Phil Gramm), otherwise the politicians would have been happy to spend even more when we had the surprising boom of the late 1990s.

According to this site, 1990 GDP was $5,800B. With Federal spending of $1,253B, that comes to 21.6%.

In 2000, the figure for GDP is, $9,952B, which is an increase of about 72% from 1990. Spending went up to $1,789B, an increase of about 43%.

Scott Gustafson writes:

The Information processing equipment and software component of real gross private domestic investment grew from $138.4B in '95Q1 to $403.4B in '01Q1. It then fell to $369.6B in '02Q1 before continuing its upward march.

That's in increase in capital expenditures of $270B in 6 years. Said another way, real expenditures tripled in 6 years.

That's a big increase, but how much is it compared to total government expenditure growth?

Mr. Econotarian writes:

Keep in mind that Federal government is just one type of government spending. States & localities have raised their spending on schools (pensions) and health care dramatically over the last 25 years.

Costard writes:
"The major effect of the dot.com boom was to boost capital gains."

It makes no sense to talk about the "effect" of something you haven't yet defined. If the rise in capital gains was not the cause but the effect, then what caused it? Please clarify what you're talking about -- and then try to argue its effect on GDP.

Capital spending rose in '92 and remained high until the crash. There's no other period like it in recent history. Who disputes that the tech revolution was a - and almost certainly the - driving force behind this massed investment? Even staid industries were forced to adjust to the internet age... to position themselves for the "new" economy. Call it the tech boom, or the dot.com bubble, or whatever you like. But it's clear that we're talking about something more than high valuations on the Nasdaq.

"How do capital gains show up in GDP?"

How does debt show up in GDP? It doesn't - and yet you can't discuss an economic boom without also discussing debt. And similarly, you can't discuss debt without discussing assets. Investors borrowed on margin in the 90's, and took out second mortgages during the housing bubble, and in both cases, (unrealized) capital gains provided the basis for increased consumption and investment. To call this irrelevant with respect to GDP, is to say that you don't at all understand the role of leverage in the business cycle.

Shayne Cook writes:

Mr. Econotarian:

The "U.S. government" spending (as percentage of GDP) Dr. Henderson refers to in his post includes state and local spending. See bea.gov Table 1.1.5.
But I would emphasize that these percentages/spending levels are nominal, and by definition (GDP), only reflect (G)overnment spending for the purchases of new goods/services - not transfer payments.

D. Henderson:

From the bea.gov Table 1.1.5, I derive slightly different percentages of nominal (G)overnment spending percentages for 1990 and 2000 - 20.37% and 17.39%, respectively. May I ask the source of your numbers?

One other item: In your August 26th post, you cite reduced public debt interest payments as part of the explanation for reduced (G)overnment percentage-of-GDP 1990 to 2000. It's my understanding that interest payments on the public debt are regarded as transfer payment, not purchase of new goods/services, and thereby are not counted in GDP. Am I mis-informed on this?

David R. Henderson writes:

@Shayne Cook,
I thought it was clear, but I guess it wasn't. By U.S. government spending, I mean spending of the U.S. government, that is, the federal government, not all governments in the United States. And federal government spending does include interest payments. You're right that they're not in the G part of GDP (C + I + G), but my discussion is about the ratio of all federal government spending to GDP. So, yes, federal spending on interest payments is part of U.S. government spending.

Shayne Cook writes:

Dr. Henderson:

Gotcha - thanks.

Mr. Econotarian:

Per David's response, I mis-informed you.

Comments for this entry have been closed
Return to top