In previous articles, we have covered the overall wealth and income of American households, and discussed the class breakdown of reported adjusted gross income. Here I want to present the basic story on the sector breakdown of US production-side GDP and its history over the last 10 years. The primary moral is that services account for most value added in the US, while the term "services" conceals a wide array of productive activities.
I hope the article gives important perspective on issues like the scale of the US energy sector, whether we should be deeply concerned about changes in manufacturing share, the overall exposure of the US economy to changing commodity prices, and similar matters. I will address further issues of capital intensity and employee compensation by sector in a future article, and preparing that conversation is another purpose of this one.
First, a word about what we are measuring - value added. GDP is a balanced accounting relationship, with a production side and a use side. Value added tells us where on the production side of GDP income is earned, not where it is spent on the consumption side. Each industry has gross overall sales, but a portion of those simply fund the purchase of that industry's inputs from other sectors of the economy.
Those sector relationships are tracked in what are called "Use-Make" tables, and the gross outflows from such tables are subtracted from each sector's gross receipts to arrive at the value added within that sector. In turn, all value added can be assigned to 3 factor income shares - compensation of employees within the sector, net taxes paid by the sector's business organizations (only), and gross operating profit of those businesses.
Some qualifiers are needed to each of those three shares of value added. Compensation of employees is larger than wages and salaries, because it includes deferred compensation in pension plans, benefits, and the like. It measures the actual cost to businesses of employing everyone in the sector. Net taxes fully account for any direct subsidies the sector receives from government, but do not account for the taxes paid by employees in the sector as individuals, nor the taxes paid by business owners on dividends received and the like.
Net tax rates vary widely from business to business, far more than the tax code alone might suggest, because businesses run at very different levels of profitability, may have prior losses to write off, have differing levels of depreciation deductions and so forth. Last, the gross operating profit line item contains no charges for cost of capital, interest paid or anything of the sort.
In some industries, very large amounts of capital are used and paying for that capital is the majority of the sector's receipts (e.g. real estate), in others, most of those receipts end up paid to its employees. I will explore the variation in such matters across major sectors at length in a future article.
My primary source is the BEA, and its tables can be found here:
The XLSX file linked off the first section on the left is the primary source for the analysis that follows. You can also browse these tables and other related tables in the National Income and Product Accounts (NIPA) by following the links on the right side of that page. The figures here are annual and the most recent year covered is 2014. In my tables below, I will focus on 2014, 2009, and 2004 to give a picture of the present state, at the recession lows, and 10 years ago across the recession period. Some of my time series plots will go back to the end of 1997, which is where the BEA's data table begins. Let us start with a top down view of our value added measure and its breakdown over that time frame.
Notice that the green section showing employee compensation is a bit more than half the total and that relationship is basically stable. The 2009 recession shows only a minor drop in the top line, with the milder 2000 recession a period of sideways movement for the employee line. Keep in mind that all figures presented here are in nominal dollars. Inflation over this period was mild, 2.5% in the first half and 1.5% or so more recently, but not 0.
There is some tendency for the employee pay portion to lag in recoveries by about 1 year, reflecting higher unemployment at the cycle lows. The profit share is growing marginally, but by less than 0.1% per year. The profit share was 38.5% in 1997 vs. 40.0% in 2014, while the pay shares were 54.8% and 53.3% respectively.
Now I give the sector breakdowns. The BEA extends this as far as 71 separate industries by dividing line items further, but I choose to report some of those at a higher node in the breakdown to keep the number of sectors reasonable. Where necessary, I will report below on the major components within some of the line items, to avoid any confusion those selections might cause. I report the actual value added in each sector in billions of dollars for 2004, 2009, and 2014, with the share of overall GDP in each year as a percentage, and in the last 2 columns the annual rate of growth seen for that sector over the 10 and 5-year periods ending 2014, also as a percentage.
Here is the table:
The top line gives the overall totals so its shares are always 100%. Notice that any sector showing a growth rate below those of the overall total will lose share, while to increase in share, a sector must show growth faster than the overall average. Those rates are 3.77% over the last 5 and 3.52% over the last 10 years. We can immediately see the rise of the "mining" sector as a move from 1.37% of value added in 2004 to 2.62%.
The largest portion of that is oil and gas extraction, with another large component for oilfield services. Note that the energy sector as a whole also has entries under "manufacturing" for "non-durable goods" - that is where refining is found, as distinct from primary production. While the move to the 2014 peak is significant as to its rate, the overall growth of the mining sector is 1.25% of GDP spread over 10 years.
Here is the breakdown of mining sector value added since 1997 -
Notice that the employee pay line grows much slower than the top line, while also being far less volatile. The profit portion swings with the oil price, and will show sharp declines in 2015 when that data becomes available. As we will see in a following article, oil and gas extraction is a textbook case of a few skilled workers each using very large amounts of capital, and earning high incomes as a result. But most of the factor input in this sector is capital rather than labor. The capital earns most of the rewards when times are good, but is quite exposed to the sector's strong cyclicality.
For comparison, notice the "construction" line item, which shows a $10 billion decline from 2004 to 2009, and a loss of almost 1% of GDP share from then to now. That is the lowest 10-year rate in the table, 1.23%, thus trailing the overall top line total by 2.25% per year. Durable goods is also somewhat weak over the 10-year horizon, but faster than the overall total since the recession lows. The auto sector is the main component of that swing.
Here is the chart for the factor breakdown of construction value added:
The boom and bust of the housing bubble is clearly apparent. Unlike the mining sector, here capital and labor fully share the sector's volatility. Also notice that the weak period here lasted for 4 years, and the recovery underway still has not quite reached the previous peak (even in nominal terms).
Is the US economy becoming purely a matter of financial circulation rather than making things, as some journalist level treatments suggest? There isn't much sign of it. The finance line item is just over 7% of value added, vs. 7.25% back in 2004. That includes banking, securities brokerage, and insurance. Notice also that this line item is basically the same size as the healthcare services sector, and that each has different periods of faster than average growth. The factor share plot for the finance sector looks similar to the overall economy. Notice, too, the top line these days is above $1.2 trillion in value added annually.
It can be helpful to group a few of those line items into broader categories. The top 3 after the overall total might be called a primary sector, from farming and resource extraction to utilities, that simply provide raw materials for the higher stages of production. Together those are only 5.5% of GDP, including the "utilities" portion. The next 3, construction plus durable and non-durable manufacturing, might be called the "make" sector, and combined are 15.9% of GDP. Everything below those line items are "services" of one form or another.
Trade and transportation are about as big as the "make" sector, at 14.7% of GDP. Moving existing product from place to place, distributing it to those who value it more than its producers, is a major source of valued added in the US economy. Such activities consistently earn about as much as creating the items in the first place, and if "construction" is backed out of the "make" portion as not creating things that are moved or distributed in this manner, more than half of it.
Next, we can group the finance sector proper with the real estate line item, creating the accounting category "FIRE" (financial, insurance, and real estate), and this broader finance total rises to 20% of GDP. Notice that most of that is contributed by the real estate portion, not the banking or insurance or Wall Street parts. Rent of existing buildings falls into this category, and it is easily the most capital-intensive sector in the table. Basically, in that "real estate" sector, it is capital not workers that is earning nearly all of that piece of production side GDP.
The last 2 items in the table can be grouped as "government," and account for 13.1% of value added. Note that taxation and transfers are a larger portion of GDP than this, since those track redistribution. Here we have only the salaries paid to government employees, plus smaller line items for government-controlled enterprises. More than half the federal line item is the military, and more than half of the state and local line items is for primary and secondary education. In other words, people who do work for a living.
Notice that the smaller, separate line item above for "educational services" is for private sector education only. Notice also that the present combined share of the government line items is 13.1% of GDP, down from 13.4% 10 years ago. It is the "transfer" or welfare state portions of government that have grown in that period; the working pieces have not quite kept pace with the overall economy.
There are 7 other "services" line items amounting to 30.8% of GDP that are not in the aggregations above. Business services is the biggest of them, with healthcare almost equally large. The "information" line item includes broadcasting, publishing, and software, and thus includes much of what we call the "tech" sector these days (business services also includes software related services), while "tech" hardware is a minor portion of the durable goods manufacturing line item.
Together the service sectors outside of finance, transportation and trade are more than 11 times the weight of the mining sector, and twice the weight of all manufacturing. Do they get coverage in the financial press commensurate with their actual economic weight?
I hope this is interesting, and questions and comments are welcome.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: Long US stocks, long US real estate, long US banks