Industrial Production as a share of real GDP
I regularly see people comment about the decline in the importance of industrial production but I never see any data on how much it has declined. Finally, I saw a libertarian blogger talk about as a point in his argument but he also revealed that he had not idea what had happened to industrial production relative to the economy.
The Federal Reserve actually publishes the value of industrial production as part of its monthly industrial production report, so it is quite easy to get the raw data to construct a series of industrial production as a share of GDP. It has fallen from 33% of real GDP in 1972 — as far back as the data goes — to 23% currently. The trend is for the share to fall 2.4% annually.
Manufacturing employment has had a similar fall, from 31% to 13.7% last quarter. The employment has seen a 2.4% annual decline.
For a while I agreed that there was an error in using this data to calculate industrial production as a share of GDP. But after more consideration and checking with the Fed I have concluded that the original chart is correct. There is no conceptional difference between the Fed data and the real GDP accounts.
If you look at the BLS employment data you’ll see that manufacturing employment has been falling since the early 20th century. If nothing else, the introduction of small electric motors, more complex machines and sensor technology killed so many jobs that economists were sure the US was doomed in the 1930s.
It has been that way since the mid-70’s. I have said it again and again. You can’t have in a capitalist system, national production and high living standards. They simply don’t mesh.
The Rage, we agree when it started, but its the politicians on both sides of the isle that caused all the problems. Through failing to have fair trade in stead of free trade.
Capitalist system only works when well regulated in the country it is meant to serve.
You’re only telling half the story.
Industrial production is just off it’s all-time high, which was just last December:
https://research.stlouisfed.org/fred2/series/INDPRO
Similarly, our agricultural output is at or near all-time highs:
http://www.ers.usda.gov/media/1875389/err189.pdf
But both industrial and agricultural employment, and share of GDP, are declining. Why? Productivity improvements.
In the early 20th century, improvements in agricultural productivity (tractors over mules) freed up labor to move into factory work. In the later 20th century, improvements in industrial productivity (robots over people) freed up labor to move into the service sector.
We can only eat so much (though more than one would have thought fifty years ago). Though our houses are twice the size they were fifty years ago (with smaller families, too), there is only so much stuff we can put in our houses.
And so, as productivity improvements have allowed us to movedfrom the basics (food and shelter), to nice to haves (automobiles, appliances, computers, and a McMansion), to luxuries (vacations, lawn care, and pedicures), our economy has moved from primarily agriculture, to primarily industry, to primarily services.
Wishful thinking there, Beene. I’d love to put all the politicians on an isle somewhere! 😀
We can call it Fletcher Island:
Warren, enjoyed the song.
But, thought you banished only those who you has some hope of their reform. The rest you made more permeate there exit. :~)
Politicians — no hope for reform. Thus the last lines:
is everyone in?
are you having a nice time?
now the final solution can be applied
Perhaps some folks get caught up in the weeds of all the data. Or perhaps they just choose not to see the data or perhaps they just cannot connect the dots of what all the data is trying to tell them. I like to look at data from many sources to formulate opinions and conclusions, not just the Bureau of Liar Statistics. Perhaps one should look at what others have to say like Martin Ford ,Michael Hudson, Joseph Stigilitz , Paul Craig Roberts, Tom Hartmann, Gerald Celente and many others. Go see and read up at ProsperousAmerica.org for greater credibility of analysis, opinions and justice on America’s manufacturing and labor share. The globalization of everything along with Frankenstien- predatory capitalism greed is killing the middle class and any resurgence of manufacturing in America. We most definitely need the balanced trade agenda along with the variable rate tariff that will bring a more fair trade platform that includes currency manipulation provisions to protect against unfair merchantilism.
“And so, as productivity improvements have allowed us to movedfrom the basics (food and shelter), to nice to haves (automobiles, appliances, computers, and a McMansion), to luxuries (vacations, lawn care, and pedicures), our economy has moved from primarily agriculture, to primarily industry, to primarily services.”
So true. What you left out is the impact on people. Fewer and fewer people make the transition from basics to luxuries and those people end up in the low paying service industries and continue at the basic (food and shelter) level. The result is a diminished middle class and extreme wealth inequality.
Spencer,
By dividing the *gross value* of industrial production by GDP you are greatly overstating industrial production’s share of GDP.
The correct way to measure this is to divide the *value added* of industrial production by GDP. This is because gross value includes intermediate inputs, and by including intermediate inputs in industrial production’s share of GDP you are essentially counting the value added of other sectors of the economy.
Unfortunately monthly and quarterly data is not available for that, but the BEA does provide annual data here:
http://www.bea.gov/industry/gdpbyind_data.htm
The GDP by industry data (which starts in 1947) reveals that industrial production (manufacturing, mining and utilities) as a share of GDP peaked at 32.4% in 1953 and had already fallen to 25.6% by 1972. It is 16.3% of GDP as of last year.
P.S. Quarterly data is available for national income by industry (excluding capital adjustment) . On this basis the historical shares are similar.
https://research.stlouisfed.org/fred2/graph/?graph_id=136614&updated=8776
Mark:
For this manufacturing throughput analyst and others, what are you defining as “intermediate inputs?”
I have to disagree with that assessment, Jerry. I would say that MORE people are making that transition. The thing is, we have changed the definition. Once, central air conditioning was a luxury. Now, people think it is a basic necessity. What was once a typical house is now considered a hovel. Our poor are better fed, better educated, healthier, and more comfortable than the middle class was fifty years ago.
Warren, the BEA publishes a table where they break GDP down by major product–goods, services and structures.
This table shows that the goods sector of the economy has actually increased from some 22% of GDP in the early 1950s to about 33% now. Services share has not shown a strong tendency to increase or decrease from the low 60s —
it was 61% in the early 1950s and is now 33% and it’s high was 66% in the mid 1970s. What has changed is structures that has fallen from some 3% in the early 1950s to some 0.3% now.
Structures includes housing, but it also includes a lot of capital like roads,airports,canals, schools, hospitals as well as commercial building for offices and plants.
You are absolutely right that because of productivity that the basics like
food take a smaller share of the economy and that our standard of living has grown because of this. But on the other hand much of our higher standard of living is taken in the consumption of manufactured goods like electronics. So your comment on shifting labor from manufacturing to services is not as clear cut as you think. Remember, much of out increase in goods consumption is filled by imports, not domestic goods
so it does not show up in domestic manufacturing. The entire story is just much more complicated then you seem to think.
Warren,
I grew up in a middle class family 50 years ago. Believe me, we had it much better than today’s poor.
50 years ago a Sargent could pay rent, wife stayed home, raise his children, own a car, while paying for house somewhere else to retire in.
Mark — you are absolutely correct.
Unfortunately, the correct data is only available back to 2005.
As soon as more data is available I will publish the correct data.
I will also add a note to the original post that I used incorrect data.
Thanks
Mark your email address seems to be incorrect.
My email to you bounced.
Spencer,
Sorry about the email. I clicked without checking to see if autofill had filled correctly. This is the correct address.
Run75441,
Let me answer that question by quoting the BEA:
“Intermediate inputs of an industry are the goods and services (including energy, raw materials, semi-finished goods, and services that are purchased from all sources) that are used in the production process to produce other goods or services rather than for final consumption. It equals the industry’s gross output (consisting of sales or receipts and other operating income, commodity taxes, and inventory change) less value added (consisting of compensation of employees, taxes on production and imports less subsidies, and gross operating surplus).”
http://www.bea.gov/faq/index.cfm?faq_id=185
Mark:
We would look at cost of goods sold plus any other costs added to it and not included in COGS. There would be no double dipping at that point which it appears to be what the BEA is making a point of discussing. Two of three inputs I can have an impact on which are Labor and Materials. Overhead is mostly regulated by law and therefore I can not change it directly. I am nosey and I hope you do not mind. My economics days are long gone and I mostly work to improve costs and time.
I got the industrial production data from the Fed rather than BEA and thought that the term Gross Value meant that the double counting had been eliminated and that the data was comparable to the GDP data.
Obviously I made a mistake.
Mark, I have been thinking about your comments and have been in touch with the Feds. I am reaching the conclusion that my original chart of production as a share of GDP was correct.
You suggested I use the value added series, but that is not what GDP measures. It measures final sales adjusted for trade and inventories to calculate GDP. So GDP does include the labor, capital and intermediate inputs used in building final demand. So it is a very different concept than the “value added” measure you suggested I use. Value added is just part of what goes into defining final demand for goods and services.
Finally, the economists at the Fed think that my use of their data is correct. They said:
It is true that the values presented in Table 9 are not exactly comparable to GDP because they do not include retail and wholesale margins. However, there should be no double counting in the statistics in Table 9 because the output being measured is only for final products and nonindustrial supplies, categories of goods which are consumed outside of the industrial sector. The output of products included in the materials category, which comprises goods that are consumed in the industrial sector, is excluded from Table 9.
Spencer (and Run 75441),
There are three approaches to computing GDP: 1) Value Added, 2) Expenditure and 3) Income. In principle all three measures should be the same. Only the US and Japan produce a separate income measure. The US calls theirs GDI, and Japan calls theirs “GDP by the income approach”. But in every country the value added measure and the expenditure measure of GDP are identical and are called GDP.
“Gross value” is *very roughly* equivalent to final sales which corresponds to the expenditures approach to GDP. As a consequence it necessarily includes the intermediate inputs needed to produce a given good or services. The BEA produces GDP by industry measures using the value added (which I linked to above) and the the income approach (also I also linked to above as “national income”.) But the BEA does not produce GDP by industry measures using the expenditures approach.
The bottom line is, the value added approach *is comparable* to GDP precisely because it is one of the definitions of GDP. Gross value *is not comparable* to GDP. Thus you are comparing apples to oranges when apples to apples measures are available.
Mark:
To machine a part for sale, I am going to start with a chunk of steel which I bought. To this steel, I will perform a series of operations on various machines (cut, machine[cnc], debur, and package). Each one of those operations will require Direct Labor performed. At the end of the process, I have a part ready to ship. To this part, I proportionately burden it with the Overhead of Operations (indirect labor, energy costs, etc.). This is my COGS. To this I add margin and profit. I would think this would approximate Gross Value and this should equal Final Sale unless one discounts.
Are we speaking the same language here as Value added? When I cost model parts for companies, this is what I am looking at practically.
Dr. Sadowski:
I trotted out my example as a way of communicating with you to see if I am speaking the same language as you are and in a practical manner. After 44 years of doing the magic I perform on the shop floor of old factories taking them from losers to winners, I find my points may not come across as clearly as possible and I have to explain them. I do not claim to be an economist as I am not of the same caliber as you, Robert, or even Spencer for that matter; but, I am very good at what I do and companies typically are thrilled to have me on staff.
The definition of “value added” is not clear and neither are the other approaches. It may be I might disagree; but for now, I am attempting to understand whether the approach by economics is the same as what I understand to be practical.
Intermediate products have to be included in GDP. That is why the BEA measures gross output of final demand to calculate GDP. The Fed
estimate of the value of industrial production is also an estimate of
final demand so my calculation is a measure of final demand divided by
a measure of final demand — it is apples to apples.
Dividing value added in manufacturing by GDP would be a comparison of two different concepts — an apples to oranges comparison.
Spencer:
Is this for me?
revisiting this thread today, since i happen to be looking at the Industrial Production data for October…you guys seem to have resolved this, but i’ll add that looking at table 9 it’s fairly clear from the heading that it’s not really “gross” in that it’s the gross value of “Final Products and Nonindustrial Supplies” http://www.federalreserve.gov/releases/G17/Current/table9.htm
the value of your intermediate inputs (semi-finished, primary and crude) is included in table 10: http://www.federalreserve.gov/releases/G17/Current/table10.htm