The gross domestic product is a measure of our nation's output, the wealth that we create. What is your share? If you work for a living, it isn't very much. After all, you have to share it with 300,000,000 other Americans. On the other hand, productivity has been rising. You are working smarter and more efficiently. How much of that has shown up in your paycheck?
It isn't that hard to find out. The GDP and population have been measured fairly accurately since the late 19th century. Wages have been measured, but there are many ways to measure wages. Are we talking about hourly wages, weekly wages, or annual wages? How do we take unemployment into account? Luckily, we don't have to account for inflation. Wages are paid in the same dollars that are used to measure the GDP.
Let's take a look at the numbers, and take a look at your share of the GDP.
Thanks to the glory of the modern internet I was able to get all sorts of numbers from the Census Bureau, the Bureau of Labor Statistics (BLS), and the Bureau of Economic Analysis (BEA). This should lend this post an aura of accuracy and authentiticy, but do remember, I am not an economist. I just play one on the internet.
The first chart I produced shows the value of a full time annual wage against an individual's share of the GDP. Ages ago, back when there were labor economists, one of the masters, Lebergott, prepared a set of full time annual wage statistics from the start of the century into the 1960s. This series has been adopted by a number of government agencies and has been extended to the current day. GDP statistics and population statistics were from the BLS and the Census respectively. It was pretty easy to divide the annual wage by the per capita GDP.
The first thing you can see in this chart is that the share started out around 1.5. That is, a full time job would likely pay about 1.5 times your per capita share of the GDP. This fraction rose to nearly 2.5 during the Great Depression. This isn't surprising. The GDP was down, and full time work was hard to find what with 25% unemployment. If you had been one of the lucky ones with a full time job, you would have been doing pretty well. The ratio drifted back down to 1.5 and stayed near that value into the 1960s. Then it began to fall into the 1980s, reaching about 0.9. That's a 40% drop! It has continued to drift downwards towards 0.75. That's nearly a 50% drop.
Clearly, something in our economy changed starting in the 1960s, but this drop might be an artifact related to full time work, or the particular measure I chose. I decided to see if other wage measures showed a similar decrease. For example, how have weekly earnings held up?
The chart below is based on the BLS series of weekly earnings, which unfortunately only goes back to 1964. Still, we can see the decline in the value of work. In 1964, 52 weeks or work would get you more than 1.4 times your share of the GDP. This fell quickly through the 80s, and has slowly descended to 0.67 since. Once again, work is only half as valuable as it was 45 years ago, whether measured by the year or the week.
What has caused this drop? To get some insight, I tried to look at wages another way.
So far we have been looking at things from a micro-economic view, that of a single worker who has seen the value of his or her labor falling even as the nation has prospered as a whole. To get a bigger picture, I took a macro-economic view and looked at aggregate personal income, a major component of the GDP. The personal income data from the BEA breaks aggregate personal income into its major components. I plotted these separately in the chart below as fractions of the GDP, so we can track any notable changes.
You can see the six major components of personal income shown in six lurid colors. (The chart is cumulative, so the components are the vertical distances between the lines, and the top line is the grand total).
To start with, we can see that wages have comprised about 50% of personal income, and while they may have slipped a bit below that, the value of job benefits has been rising. This includes health insurance, unemployment insurance premiums, pension contributions and other "bennies". If we look at wages plus benefits, we actually see a modest rise. While each worker's wage has been falling with respect to the GDP, that portion of the GDP which is paid as wages has fallen nowhere near 50%.
Let's look at the other components. Maybe there is a story there.
Proprietor income has actually been falling. There are fewer small farms and mom and pop stores. Modern stores over a certain size are usually incorporated, so their income is corporate income, and those who would have been their proprietors are now paid wages. Rental income is also decreasing. Fewer people are renting their properties as a means of making a living.
As we might expect, interest and dividends (paid to individuals as opposed to corporations) have been increasing as a share of the GDP. The chart shows a squeeze in this component during World War II, but this component has grown since then. Finally, we see transfer payments growing almost monotonically, through the Great Depression, and still growing today. Transfer payments include Social Security and various social welfare payments, so it isn't surprising to see a rise in the mid-60s and a subsequent rise in the mid-80s, when the children of the 1920s post World War I baby boom started to retire.
If the data is to be believed, overall personal income has actually increased as a share of the GDP. This happened in the 1960s, and seems to have been related to the rise in transfer payments. In fact, if you look at the green interest and dividends payments line, you can track the sum total of personal income without transfer payments included. It falls from over 80% in the 1930s to perhaps 75% where it stayed until around 1980. Then it returned to about 80% of the GDP. Meanwhile, transfer payments continued to rise, bringing overall personal income to 90% of the GDP by 1980 where it has stayed.
You can read all sorts of things into this chart. You can see the Great Depression and the Great Society. You can see demographic change and societal change. What we don't see is wages dropping by 50% as their share of the GDP. Even if we ignore benefits, wages have dropped by perhaps 10%, if that much.
That was an interesting chart, but there was no answer there. Maybe we should be looking at the workforce. Perhaps the workforce now comprises a larger fraction of the population than it did before the 1960s. Imagine only 25% of the population working. If this percentage rose to 50%, there would be twice as many people working. If the wage pool remained constant, we'd expect each worker's wages to fall by 50% as the working population rose by 100%.
As it turns out, the workforce did grow starting in the 1970s, much of it because women were entering and re-entering the workforce. The chart below was based on BLS statistics and shows, cumulatively, the male and female components of the workforce.
From the 1950s through the early 1970s, about 40% of the population was in the workforce. That means that each worker was supporting himself or herself and an additional 1.5 people. Starting in the 1970s there was a dramatic change and by the 1990s each worker supported himself or herself and only one additional person. This is good news for those worried about Social Security. We've gotten by with a much smaller percentage of the population in the workforce, and some might argue that we were in better shape economically back then.
Unfortunately, this change can only account for some of the decline in the value of work. A 25% increase in the workforce should mean a 20% decline in wages, if total wages are unchanged. If we combine this 20% with the 10% decrease in the wage component of personal income, and assuming we ignore benefits, we can perhaps account for a 28% decline in the value of wages. (80% times 90% is 72%.) We are still nowhere near 50%.
I decided to try another analysis. I decided to focus on a maximal case to put some kind of boundary on the problem.
Suppose everyone in the workforce could get full time work at the average annual wage. How would this potential level of wages possibly earned compare with the actual wages paid?
If the ratio is greater than one, we are in an impossible situation. Some workers have to be unemployed or working part time. There is not enough money being paid for all of them to work full time. If the ratio is less than one, that means that everyone could be working full time and there would still be money being paid as wages left over. Let's look at the chart below.
I broke out the data for men and women. Partly because it was easy, and partly because the data covers the period when women were entering the workforce in earnest, the 1970s and 1980s. This should let us cross check with the previous chart and perhaps gain some differential insight.
This is a cumulative plot, so the upper plot line represents this potential wages earned versus actual wages paid ratio. This ratio starts out well above one in 1948. It took a drop in the mid-1960s and then a big drop in the mid-1980s, dropping below one. Since then it has continued a slow slide, leveling in the 2000s to a bit over 0.8. As we have seen, the money being paid as wages did not change all that much as a share of the GDP, but now, working people seemed to have a much smaller claim on it. Even if everyone were working full time and being paid for such, there would be a lot of money left over. This has only been the case since the mid-1980s, but it was far from the case in the 1960s. Clearly, something has changed in our economy.
One interesting thing to notice is how little the plot changes for women, even though they made a big move into the workforce in the 1970s and 1980s. As wages went down, women entered the workforce and kept their potential share of wages nearly constant. For men it was different. They continued to work, but they just got paid less and less. In fact, their claim on wages dropped nearly in half, from about 0.9 to about 0.45.
In 1948, paying every worker a full time wage would have cost nearly 25% more than all the wages actually paid. By 2007, it would have taken only 82%. That leaves a lot of money on the table, especially given that full time work is not always available.
Where did all the extra money go?
One answer is in the footnotes. I took another look at the Bureau of Labor Statistics web site and looked more closely. When I see numbers, I tend to get excited and can't wait to dump them into my computer, but numbers don't mean anything without a good look at the fine print.
To start with, according to the BLS, the earnings, annual and weekly, were only for certain classes of employees. In particular, they were for non-supervisory employees in certain industries, and they accounted for only 80% of the workforce. Obviously, a share of personal income is going to the other 20%. One word leaps out of the text, "supervisory". The 80% does not include managers, they are part of the remaining 20%. Needless to say most management salaries are quite modest, but upper level executives have seen their salaries rising dramatically.
I may be jumping the gun here. It is possible that some other wage earning component has been increasing its share of earnings at the expense of the vast majority of workers. Salaries for rock stars, professional athletes, actors and other "stars" have been rising as well at that of upper management, but these "stars" negotiate their own salaries in an adversarial process. Upper level executives simply set their own, a micro-economic fact that seems to be having macro-economic consequences.
It is interesting to consider that we could completely eliminate unemployment and involuntary part time work simply by cutting executive and other high end salaries, but that seems to be what the macro-economic analysis shows.