Gross Domestic Product and Gross Domestic Income

Everyone should read Daniel Gross's "Economic View" column this forthcoming Sunday: he's trying to make sense of the fact that estimates of National Product show productivity growth reverting to its average post-1995 pace, while estimates of National Income continue to show more rapid productivity growth. It's an important topic, and it's bound to be a good column.

Meanwhile, here's a short squib from Business Week on the issue:

On Aug. 9, the Bureau of Labor Statistics said that in the second quarter, the most widely followed measure of productivity, output per hour in the nonfarm business sector, grew at a 2.2.% annual rate from the first quarter. Over the past year, productivity increased just 2.3%, down sharply from the 5% yearly pace seen at the end of 2003.... [B]ut the BLS actually calculates a second broad measure of productivity, one that shows a more robust trend... covers only the nonfinancial corporate sector... shows productivity in the first quarter grew 5.4% from a year ago... faster than the 4.5% yearly clip recorded at the end of 2003....

Neither set of data is necessarily better.... They just use different measures... gross domestic product... [and] gross domestic income. Theoretically, the value of products and the income they generate should be the same....

What raises this debate into the realm of policy importance is that, in the past, Greenspan has tracked this second set of productivity data closely....

In the nonfarm business measure, smaller gains in efficiency are no longer offsetting increases in labor compensation. As a result, it now costs businesses far more to make and sell one unit of their product. Last quarter, nonfarm unit labor costs... were up 4.3% from a year ago, the fastest yearly clip in nearly five years.... But within the nonfinancial [corporate] sector... unit labor costs are up by just 1.4% from a year ago.... Which is right?.... [O]ne advantage of the nonfinancial corporate numbers is that... they give more of an "apples and apples" comparison between compensation and output.... [T]he surprising strength of profits so far this year suggests that the more sanguine reading of unit labor costs from the nonfinancial sector may well be closer to the truth....

And here's a graph showing the ratio of (measured) national product to (measured) national income:

From 1995 to 2000 the ratio of product to income shrank as measures of product grew more slowly than measures of income. This was important: for five years measures of productivity on the income side grew 0.8% per year faster than measures of productivity on the product side. Then for three years measures of product grew 0.9% per year faster than measures of income. And in the past four quarters it looks like it has turned around again, with income growing 0.6% per year faster than product.

It's a data issue I don't understand. It's a vitally important data issue--especially if you have a view of the world that gives some weight to what's going on with unit labor costs as a look into the guts of the inflation-unemployment relationship and to the current level of the natural rate of unemployment.

I am not a happy camper.

Comments

[Movie Guy]

I trust that you will address this subject again in a later post.

While I am confused, it is my understanding and expectation that per unit profits are down primarily due to absortion of additional production costs (energy, for example) as opposed to passing such on to consumers. Yet, corporate profits are supposedly at or near record high levels.

Does this imply that most of the corporate profits are from outsourcing, foreign-based inbound goods, and similar foreign-based export operations? Or increased domestic unit volume sales?

If U.S. domestic operations' per unit profit volumes are lean, and if increased sales volumes are not offsetting the per unit profit losses, what is the source of the overall corporate profit gains? Or are the unit sales volumes so much larger that per unit losses are offset?

And what rollup sources are available to compare foreign-sourced to domestic-sourced profits margins on U.S. goods and services?

The energy costs are being absorbed from what I have read.

[pi]

Some random thoughts:

As the graph indicates, these kind of divergences between GDP and GDI growth rates aren't new.

The BEA says to put your faith in GDP over GDI, but it's hard to think that GDP should get all of the weight.

One reason the BEA places more faith in GDP than GDI is that we don't have very good data on compliance with the tax code. The BEA has fudge factors to boost up income to account for the assumed cheating, but who knows how accurate they are? For a long time the IRS avoided doing one of its compliance studies because of the bad press they generate. How would you like to undergo a root canal audit just so the IRS and BEA can learn how well the average American complies with the tax code?

The changing trend in the statistical discrepancy could have some connection to the surprising strong corporate income tax receipts.

[Mary McKinney aka Academic Coach]

Can you please explain more about your thoughts about this to someone who knows absolutely nothing about economics.

From a completely naive viewpoint, it sounds like bad news to me if our incomes are rising faster than our productivity.

[spencer]

I do not think you are correct that output for nonfinancial corps uses national income. I beleive it uses adjusted gdp just like they do for the other sectors.

The section on productivity for the nonfinancial corporation actually has a profits per unit data series that is very good -- but the report on it lags the other reports for a quarter so the data is not so timely. But overall, the single most important determinate of profits is the spread between prices and unit labor costs. Thru the end of last year this was very high, but since the start of this year it has turned negative, implying a sharp slowdown in profits growth.

Remember this is for the overall economy. so when you are looking at one industry having poor profits because
they are paying a lot for oil the energy sector is having high profits exactly because the other sector is suffering. If you look at the s&P data on profits by sector( www.spglobal.com ) you will find that profits growth in the energy sector are much higher then the average while profits growth in the other sectors are significantly weaker.

[spencer]

Nonfinancial corp data does use income data rather then
gdp data. Something i did not know. that is how they can get profits out of the data that they could not do using gdp data.

[cw]

I thought that one of the differences between the product and income data is that income includes overseas income from US owned facilities. Increases in productivity at plants abroad won't have much effect on the labor-inflation relationship within the US.

[dilbert dogbert]

I read a comment that said the productivity numbers were being jobbed to make the SS panic story work better. I don't think so as the SS panic people don't give a Sh*t about numbers.

[Ragout]

Mary,

You're missing the point that output and income are the same thing. Brad's point is that we have two measures of the same thing (GDP and NI), but they show different trends in recent years. One reason that this could happen is that they are measure from different underlying data sources (NI from tax data and GDP from sales data). One of these measures is presumably better than the other, but which one?

Besides measurement problems (e.g., tax evasion, which would affect NI but not GDP), there are also conceptual differences Brad doesn't mention. GDP and NI aren't really measuring the exact same thing. The BEA puts out a table comparing GDP and NI:

http://bea.gov/bea/dn/nipaweb/SelectTable.asp?Popular=Y

The conceptual differences between GDP and NI are foreign profits and depreciation. Looking at the foreign profits component of the difference, it looks pretty small. So, it seems to me like the obvious theories are (1) tax evasion, as mentioned above, and (2) problems in the measurement of depreciation.

Depreciation is huge, over 10% of GDP, and is pretty obviously hard to measure.

[Auros]

"Economic View" column? I know about the journal "The Economist's Voice" and the blog "Economist's View", and I know Gross has the Moneybox column at Slate -- where is this "Economic View" column to which Brad refers?

[Auros]

Oh, wait, it's in the NYT, isn't it?

In any case, I'm trying to make sense of what this data might indicate, at face value... Could this be related to that old debate about stock valuations, where the right-wing blogs were trying to say that corporations could capture more of the benefits of productivity (which would reduce productivity gains reflected in GNI), cutting down on wage increases and boosting returns?

[anne]

Mary and Ragout

Paul Krugman mentioned such a measurement difference several years ago. Possibly I can find the note I made. I take all sorts of notes, finding them can be another matter :)

The problem here is as important as hoping productivity growth is as robust as the Federal Reserve has for several years assumed. This problem worries me some.

[Gerard MacDonell]

The BW article is based on data that are already defunct. Since the Q1 NFC productivity data were produced, the 4-q swing in the measured gap between the expenditure and income based measure of national output has been cut in half, from 150 bps to 75 bps. Assuming this was ever an issue, it is now half as big.

[BB]

Why is the y-axis on your plot in Billions of chained dollars? Should it not be dimensionless? I'm not an economist and might need some explaining.

[Ragout]

Oh, I just realized Brad's ratios are all close to 1, so he must be making an apples-to-apples comparison (GDP to GNI or NNP to NNI). So the two figures differ only due to measurement problems (like tax evasion) and everything I wrote about depreciation above was just blather. Oops.

[nate]

Barry Ritholtz and Kudlow have two perspectives on the stock market.

per fuzzy recollection... for commpanies in stock markets (S&P 500?):
...cash flow increases (ex energy) are up 8% year-over-year (Kudlow points out),
...earnings (ex energy) are up 4% year-over-year (Ritzholtz points out)

Do you think this could at all figure into this discussion and graph? I am totally not sure. Do companies book "accruals" that lower income but not real cash flows?

[nate]

for companies, why do cash flow increases exceed earnings increases?

[Tim Lundeen]

I don't know enough about how the numbers are generated to argue the case strongly, but I wonder if this change in the ratios is a byproduct of the deflationary Fed policy in the late 90s, which they started fixing in 2001. If you look at the price of gold as a proxy for inflation/deflation bias in Fed policy, it tracks the ratio of product to income pretty nicely. (See the 10-year gold chart at http://www.kitco.com/charts/livegold.html for example.)

My speculation is that deflation/reflation affected product prices more quickly than incomes, thus generating this effect in the product/income ratios.