Could the rise of women entering the workforce play a role? Trying to think of major shifts that start in the late 70s and then normalize in the 1990s, and that's one:
Not just women, but an overall positive supply shock of labor:
1. Baby boomer generation reaching working age
2. More women in the workforce
3. More immigration
Also the composition effect of the workforce being younger. All considered, it's kind of a miracle the workforce grew so much without wages going down.
I was thinking the same thing about women joining the workforce. That and the movement of factories away from the Midwest to the South for lower wages of non-unionized labor
The end of the rapid rise in women's labor force participation was right around the end of the wage stagnation, but it was rising rapidly before the stagnation started, so not a perfect explanation.
Yeah I don't think it's PERFECT but it's a very logical answer and it at least deserves to be in the conversation, but is probably not included for ideological reasons
Which is also why the falling all-important white birth rate is creating such butt hurt for conservative America. Women don't want to work a full-time job AND work full-time as a mother. Externalizing the costs of population replacement is something those primitive EU Socialists seem to understand in ways that our own technocratic class never will.
Such an incredibly striking chart, and then Noah just didn't follow up on it! Especially since it so clearly supports the "a single man used to be able to support a family" narrative - average wages for men went down by 10% from 1980 to 1995, and are only up 16% overall in the past 45 years!
"The term is also commonly used to describe the false belief that increasing labour productivity, automation, immigration, or women's participation in the workforce causes an increase in unemployment. "
We're not discussing an increase in unemployment. We're discussing whether a supply shock could lead to stagnating wages. I don't think the fallacy applies (and I'm not sure I agree it is always a fallacy.)
The principle also applies to wages. New entrants to the labor market (be it women, immigrants, whatever) do not depress wages for existing workers or cause an increase in unemployment because the wages the new entrants earn increase demand at the same time that their employment increases labor supply. So we have a “growing of the pie” rather than a one-sided supply shock. The fallacy is assuming an increase in the supply of workers operates in the same way as an increase in the supply of goods or services. It does not, because the supply of workers is tied to the demand side in this way.
So you’re saying that if there are 20 jobs for lawn care in my town, the addition of 40 new laborers from Guatemala seeking to find work in lawn care will lead to a demand for more lawn care? Despite the fact that nothing has increased the number of lawns?
Demand will rise or stay constant, no one will lose their job, and wages will stay the same or rise?
“So you’re saying that if there are 20 jobs for lawn care in my town, the addition of 40 new laborers from Guatemala seeking to find work in lawn care will lead to a demand for more lawn care?”
There will be a demand for more managers of lawn care workers and logistics services for lawn care (trucks, trailers, equipment) since the labor component will now be quite affordable, the demand for lawn care may be going up.
The number of lawns will not stay constant. The Guatemalans will buy houses, get their own lawns, and the demand for lawn care will increase accordingly. Alternatively, if Guatemalans are not too keen on lawns, they may spend their money on different goods and services, increasing the demand for labor in other sectors. Different occupations may rise and fall in pay, but there will not be general wage compression.
This is where it’s silly to argue from economic la-La land. The Guatemalans won’t buy houses with lawns, because they are poor. Poor people do not buy houses and hire others to do their lawns. That’s like saying there will never be competition for butler jobs because the new laborers seeking to be butlers will need butlers of their own, keeping demand constant.
To take a simple example, suppose Jack has a job making widgets. His wife Jill gets a job making widgets too. This increases the supply of widget makers, putting downward pressure on Jack’s wages. But Jill puts her new salary towards buying widgets. This increases the demand for widgets, putting offsetting upward pressure on Jack’s wages.
In the example, widgets are a stand-in for the production of goods and services in general. If women all piled into one line of work while spending their money broadly, then yes, the men who were already in that line of work would have their wages compressed. But that isn’t what happened. Women work in a wide variety of jobs.
Expansion of the labor force does not compress wages. The USA has a much, much bigger labor force than Portugal, because it has many more people. But because American workers also spend their money, American wages are not lower than Portuguese wages.
I suspect that the difference has to do with women!
The US Census Bureau has a nice graph showing inflation-adjusted earnings by gender back to 1959 (link is https://www2.census.gov/library/publications/2024/demo/p60-282.pdf, on page 10 in the text and page 19 in the PDF viewer). It looks like male wages stagnated in the 70s and are still stagnant, while women's wages been consistently increasing over the past 65 years. Earlier in that PDF (page 3 in text, page 9 on the PDF viewer), there's a graph that shows that wage stagnation was relatively consistent across races.
I'm a little suspicious of this story, because it almost seems too convenient. But at the same time, the financial emancipation of women fits the bill almost exactly. It's a relatively simple supply-side labor anti-shock (what's the opposite of a shock?) that caused male wages to stagnate as female wages rose to catch up (which they have yet to do). And as you've pointed out, there aren't any other larger-scale macroeconomic factors to look for here. None of them would explain why wage trends differ by gender anyway!
Your story is also supported by how technology allowed traditionally male jobs in agriculture, manufacturing, and other areas to be broken up into different tasks that could be assigned to different workers, including women. Much of supply chain, invoice payment, inventory management, billing became separate tasks from actually making stuff once tracking of inputs and outputs was done with PCs and databases. As tasks like these got factored out, some employment demand shifted away from the guys who used to do it all end-to-end. By the end of the 60s, mainframe database technology had got good enough for this job refactoring, and the arrival of PCs in the 70s made it economical for even smaller enterprises. As another example, contrast how printing was done up to the 1960s (https://en.wikipedia.org/wiki/Linotype_machine), with how it changed with the advent of laser printers (https://en.wikipedia.org/wiki/Laser_printing) in the early 1970s and desktop publishing a bit later.
The period of stagnation closely corresponds to the period between my birth and my graduation from college. The obvious explanation is that the cost of raising me was a huge drag on the economy, but once I became a giver rather than a taker, things turned around.
I have a problem with this post. You make no distinction between the wages of the college educated and the non-college educated. The gap between these wages has been widening steadily since 1978. If you look at overall wages in the United States, the picture is not too bad, but for the non-college educated it has been very daunting. Because this gap between the two groups does not play a role in the minds of too many economists, no wonder the non-college educated have turned to Trump as a false messiah. See:
Working-class men really got shafted, but it happened in waves. The first wave of deindustrialization (factories moving out of the cities) hit the black working class pretty hard (which was then followed by the crack epidemic and the rise in crime). The second wave of deindustrialization hurt the Rust Belt in the 1970s, but the China Shock really did in the white working class. Of course there are other factors to consider: the poverty trap of means-tested welfare programs, deunionization, immigration, etc. But one way or another, it's been slim pickings for the blue-collar crowd.
One thing that puzzles me about that first wave of deindustrialization: economically it didn't make sense to locate factories in city centers after roughly the 1920s, because single-floor electrically-powered factories (which take up too much land for a central location to make sense) were far more efficient that the old-fashioned multi-floor mills that ruled the pre-electrical era when factory machines needed to be mechanically coupled to a basement steam engine.
So why did factories only _actually_ leave the centers of US cities after the race riots of the '60s?
Technically, deindustrialization hit before the race riots of the 1960s, partly facilitated by the construction of interstate highways and the rise of urban sprawl. People like to talk a lot about how "welfare broke the black family," but in reality the black nuclear family began deteriorating way before LBJ even proposed the War on Poverty and the other Great Society programs; the Moynihan Report suggests that the decline started as early as the 1940s, despite the material progress black Americans secured through the New Deal and wartime economic policy. Eventually, that plague spread to white working-class families, and is exacerbated by the fact that American politicians insist on fighting poverty with means-tested benefits instead of generous universal programs that aren't subject to poverty cliffs and other perverse incentives.
Factories couldn’t leave our cities until we finished building out our transportation system—the 40,000 miles of Interstate highways that was essentially complete by the early eighties.
But he did not directly compare that category to the college educated or supervisory employees. The comparison would have shown a great contrast in increases in compensation.
Yes, because that was not the topic of the article. It was not about inequality. It was about wage stagnation for production and non-supervisory employees from 1973 to 1994 and the possible causes of it.
Saying, “that's not the topic” is like a doctor looking at one arm and ignoring the rest of a patient, which is something you do very regularly in our conversations regarding your Substack.
Noah Smith noted that median compensation and average compensation began to diverge in 1974, which is when the compensation of college educated people began to greatly diverge from the non-college educated, He noted it, but he did not comment on it. I find that disappointing.
Another issue that he failed to address is why the wages of non-supervisory men have only gone from $21.63 per hour to $25 per hour in 45 years. That's a measly 15% raise after 45 years of work!
That's a reality that is clear in today in a graph that he posted.
You can defend yourself by insisting on staring only at the arm, but I'm going to look at the adjacent body parts as well and call on Noah Smith to look at them too.
Yes, you always trying to change the subject away from what the author actually wrote in an article to what you want to talk about is a regular theme. This is likely a defense mechanism to avoid dealing with facts that conflict with your world view.
At some point you are going to have to realize that the author of each article gets to define the topic. If you are not interested in that topic, then do not comment.
You are also obviously changing the subject so you can promote your latest post on Trump (which is also not the topic of Noah's article).
I do have a persistent interest in the fact that the wages of the non college educated have risen little if any in the last 45 years. I believe this is a great factor in the failure of the Democratic Party to address the true needs of the majority of the population. If the Democrats ever took their eyes off racial inequality and turned it to the stagnant wages of the non-college educated, which includes every racial group, then they might be able to do something worthwhile for the country.
I believe that Noah Smith is distracting us from this big and important fact by his argument that these wages were stagnant but now have begun to rise again SLOWLY. Well, the rise is unimpressive, and it has left millions of people miserable and ready to look for anyone who will recognize their problem, and this is has been Trump.
My interest is always in looking at the big picture that affects the most people as opposed to a small and misleading portion of that big picture. I will ALWAYS ask the question, “How does this relate to the big picture?”
I wonder about a sociological explanation: the erosion of norms about acceptable differences between management and worker pay. We all know the basic stats on this. I've never accepted the claim that hypercompetition for middle and upper managers (in a more pure economic sense) means they need be paid more and more to recruit and retain these super talented people. That and pure rent seeking....
Was there ever such a norm? I hear people complaining that it's higher than it used to be, but I don't see any evidence that people were targeting such a limit back then. Also this ratio is more dependent on stock performance than anything else since it is always calculated on large publicly traded companies. And there are very few of such companies, so I don't think there is really any way to get meaningful data from the compensation of 500 rich people out of the 300 million or so in the country.
There was more wage/salary compression back then, so-to-speak, but this wasn't exactly airtight: executives back then had generous corporate expense accounts that more than made up for the limits in salaried compensation.
That being said, channelling the excess compensation of high-income earners directly into the consumption of goods and services may be a better alternative than letting them use their excess funds to gobble up appreciating financial/real estate assets.
The timing of the post-1993 or so boom looks like it was kicked off by the internet. Maybe the lull before that was caused by the electricity boom having, ahem, run out of gas. There definitely was a feeling in the 90s that household appliances hadn't improved much in the previous two decades, with the microwave being the big innovation and things like remote controls and cordless phones being the fairly weak tea next in line inventions. HDTV was constantly promised but never delivered, and the available channels going from 4 to 50 thanks to cable was considered a big improvement. The rates of growth we've become accustomed to are historically extremely abnormal so the question isn't so much why they have lulls but how they manage to keep going.
I was there, so I can say that there was a giant, huge, very big beautiful Internet boom post 1993. First the computers were installed in offices - I have vivid memories of setting aside my entire Friday mornings to back up a system on floppy disks, remember those? - which had to be inserted and removed in Very Precise Order or your backup would be ruined. We had the office tech and then AOL and CompuServe and Prodigy moved computers into homes, along with the easier availability of desktops (laptops weren’t a thing until the late 90’s and they were so bulky they required a dedicated rolling cart for them, so they were not even really “lap”tops!).
But those were indeed glory days for anyone who wanted a job, and it also helped along the renaissance of cities as places people wanted to live in and not escape from.
It’s been a while but I remember reading that once you take benefits into account, wage growth throughout that whole period looks much stronger, and that especially the slowdown in the early 2000s was mostly healthcare costs eating up wage growth
Mixed causes seem the most likely answer. Increased imports from China would have at least a temporary impact, and women entering the labor force would result in increased supply. Regarding the last, it wasn't until 1972, with the Supreme Court case Baird v. Eisenstadt, that birth control became legal for everyone, including unmarried individuals. The impact of that decision modified the labor mix, too.
I think the Fed's response to the 1973 energy crisis was inappropriate. The oil price shock was a supply price-induced inflation, not a result of excess consumer demand. The Fed treated the oil price shock in the same way it treated excess consumer demand. The Fed raised interest rates to stifle inflation. The difficulty with that approach is that it contravened what was needed. Businesses needed new investment to adapt to higher energy and feedstock prices. (Remember that oil and gas are not only fuels but also inputs to many production processes.) The Fed's actions delayed for years the shift in manufacturing needed to adapt to both higher and more volatile energy prices. As a partial example, the automobile industry was atrociously slow in addressing Japan's and Europe's competition in higher fuel economy cars, although some of that may be due to ineptitude.
One unanswered question is that the period was also characterized by an increased penetration of information technology. Minicomputers (remember DEC?) came to dominate, and the first PCs landed on employees' desks. Did that counteract some of the reduced productivity trends, but the data isn't sufficiently granular to detect that?
I entered the workforce with my first "real" job in 1972, though I had been self-supporting for at least ten years prior. I have expressed the opinion for many years now that the stagnation of the American economy beginning in the 1970's was the direct result of the confluence of two underreported (or under-considered) economic policy changes. These were the rise of "the Chicago School" economic theories and the adoption of the Harvard Business School's Management By Objectives (MBO). The combination of trickle-down and MBO presaged an era of declining growth in wages, sure, but much more visibly and easier to track, was the huge decline in employee benefits, business lunches, airline perks (despite the introduction of points) and concommitently the increase of "red tape", not in government, but in business, where the increased reporting requirements severely affected morale and job satisfaction, if not the productivity (which increased) of sales, supervisory, and lower to middle management workers. It was in this era that middle-management was hollowed out, by the mass firings of regional managers and the imposition of (previously managerial) reporting requirements on frontline workers. This was further exacerbated by the growth (some forced) in the use of personal computers. I could go on and on about how the confluence of all of these major milestones in American business created the situation which you have ascribed to more memorable events.
Average hourly earnings data does not include medical care benefits provided by corporations. Although including such benefits would not completely change the story about wage stagnation, it might change the visual pattern of the wage graph and support or undermine other explanations.
When I read this post, my first thought was “anyone who has ever listened to any of Bruce Springsteen’s back catalogue knows he built his career around this!” Just off the top of my head, The River, Downbound Train, My Hometown, the desperate guy in Atlantic City… those were all quintessential 80’s. The farm crisis was 80’s. (Those of GenX vintage, or older, remember Farm Aid?) So you are right, it’s been going on a LOT longer than post-Great Recession. We even griped about it and blamed Reagan.
I wonder how much shareholder compensation had to do with it (making sure profits went to stock holders), how much was merging and consolidation, how much was the baby boom glutting the workforce, and how much was the inevitability of Europe and Japan shaking off their post war malaise and rebuilding? The US lucked out in being the only industrialized country left standing after WWII, so we had 20 years of advantage there, and once Europe and Japan were able to rebuild and catch up, we had competition.
Looking back, the big difference between then and the post-Great-recession landscape was that housing was still affordable for most working people, even if it meant sharing a down at heel flat with roommates (I moved into a flat that had a mouse problem, despite the resident cat. I brought a young cat with great hunting instincts. She liked to leave them on the floor for people to step on, but we had no more mice in very short order). I think a lot of us have roommate stories, shabby starter housing stories, etc. and weren’t forced to live with our parents well into our 20’s. Nor was there carefully curated social media to give us the comparison heebie-jeebies; one *could* read Town and Country if that was one’s jam, but otherwise, the Joneses were that group of other roommates in the equally shabby apartment upstairs.
I am puzzled that none of the analysis looks at the size of the labour force (including 'illegal' immigration if this is relevant to these basic stats) over this period, its rate of change and any sociological issues (eg, more women working, all working longer before retirement etc). It is relevant: one can have a relatively constant unemployment percentage with a rapidly growing force, but it's unlikely without wage stagnation or slowdown.
Fascinating analysis. Could some cultural vibe shift explain part of the stagnation? In 1974 there were problems with quality control in the auto manufacturing sector. GM had cars piling up on a lot in Van Nuys, California due to poor workmanship/design/construction/manufacturing capabilities. The guiding ethic was "just get them off the line and we'll sort out the quality problems later". They didn't. Production throughput superseded all other concerns like quality. Honda saw the opportunity and took it. Budweiser had lunchtime restaurants in the parking lots of auto manufacturing firms in Detroit. At lunch it was apparently common to consume a couple of beers before returning to the assembly line. How does the consumption of alcohol contribute to quality construction? The Three Martini Lunch? That was management's adoption of the same concept. In addition, there was the elevation of the acceptability of consumer debt. The populace gradually changed from austerity and avoiding financial bondage which was a reaction to the Great Depression to a full-throated acceptance of acquisition of goods now for the promise of payment later. Of course, there would be interest to be paid, but I don't care, I want it now. The Viet Nam War had created distrust of the government and working for "The Man" was beginning to gnaw at corporate credibility. This was an era when young people grew distrustful of and rejected authority, consequently adopting the attitude of looking out for yourself first. Boomers might have to accept some responsibility for what is happening today. I am from that era and was able to watch it firsthand. Distrust of society and its motivations is a long term trend which antedates the current political dilemma.
The unemployment rate was generally higher during the period of wage stagnation. Maybe the economy was working at below capacity, hence lower demand for labor.
Would love to see an international comparison of the same time series. Is it the same basic picture with slightly different timing or totally different patterns? That might narrow down the explanations a little.
Could the rise of women entering the workforce play a role? Trying to think of major shifts that start in the late 70s and then normalize in the 1990s, and that's one:
https://equitablegrowth.org/womens-history-month-u-s-womens-labor-force-participation/
Figure 2
Not just women, but an overall positive supply shock of labor:
1. Baby boomer generation reaching working age
2. More women in the workforce
3. More immigration
Also the composition effect of the workforce being younger. All considered, it's kind of a miracle the workforce grew so much without wages going down.
Good point. Immigration reform/liberalization occurs at almost exactly the same time as the women's rights movement. And, indeed, throw in the boomers
Yes, the working age population grew a lot in during this period.
https://fred.stlouisfed.org/graph/?g=1J7yr
And labor force participation rate grew consistently in this period too:
https://fred.stlouisfed.org/graph/?g=1J7Bd
I was thinking the same thing about women joining the workforce. That and the movement of factories away from the Midwest to the South for lower wages of non-unionized labor
I was about to post the same. There's a better chart here https://www.bls.gov/cps/demographics/women-labor-force.htm
The end of the rapid rise in women's labor force participation was right around the end of the wage stagnation, but it was rising rapidly before the stagnation started, so not a perfect explanation.
Yeah I don't think it's PERFECT but it's a very logical answer and it at least deserves to be in the conversation, but is probably not included for ideological reasons
He did post a chart with men's and women's wage right there cleverly hiding in plain sight. But it wasn't explicitly referenced, that is true.
Yeah, and there are a lot more problems for liberals than just that if they admit that supply and demand works in the labor market.
Which is also why the falling all-important white birth rate is creating such butt hurt for conservative America. Women don't want to work a full-time job AND work full-time as a mother. Externalizing the costs of population replacement is something those primitive EU Socialists seem to understand in ways that our own technocratic class never will.
Would have been a good idea to look at EU birth rates before writing that.
Yup. US birth rate is still higher. Despite little/no child care assistance and maternal leave.
Yes, because what better paid jobs were women substituting for men for in the 1950s-1970s. I don’t think I buy that explanation.
You don't need to replace any highly paid jobs to affect the median
That doesn't explain the fall in men's real wages.
Such an incredibly striking chart, and then Noah just didn't follow up on it! Especially since it so clearly supports the "a single man used to be able to support a family" narrative - average wages for men went down by 10% from 1980 to 1995, and are only up 16% overall in the past 45 years!
0.3% growth per year is terrible
What jobs do we think women substituted for men in doing?
I was curious about this too and why it wasn't addressed
Dang—you beat me to it by a few minutes.
Same.
Great minds and whatnot
No - this is the “Lump of Labor Fallacy”: https://en.m.wikipedia.org/wiki/Lump_of_labour_fallacy
"The term is also commonly used to describe the false belief that increasing labour productivity, automation, immigration, or women's participation in the workforce causes an increase in unemployment. "
We're not discussing an increase in unemployment. We're discussing whether a supply shock could lead to stagnating wages. I don't think the fallacy applies (and I'm not sure I agree it is always a fallacy.)
The principle also applies to wages. New entrants to the labor market (be it women, immigrants, whatever) do not depress wages for existing workers or cause an increase in unemployment because the wages the new entrants earn increase demand at the same time that their employment increases labor supply. So we have a “growing of the pie” rather than a one-sided supply shock. The fallacy is assuming an increase in the supply of workers operates in the same way as an increase in the supply of goods or services. It does not, because the supply of workers is tied to the demand side in this way.
So you’re saying that if there are 20 jobs for lawn care in my town, the addition of 40 new laborers from Guatemala seeking to find work in lawn care will lead to a demand for more lawn care? Despite the fact that nothing has increased the number of lawns?
Demand will rise or stay constant, no one will lose their job, and wages will stay the same or rise?
“So you’re saying that if there are 20 jobs for lawn care in my town, the addition of 40 new laborers from Guatemala seeking to find work in lawn care will lead to a demand for more lawn care?”
There will be a demand for more managers of lawn care workers and logistics services for lawn care (trucks, trailers, equipment) since the labor component will now be quite affordable, the demand for lawn care may be going up.
So you’re saying that wages would indeed decline for yard work?
The number of lawns will not stay constant. The Guatemalans will buy houses, get their own lawns, and the demand for lawn care will increase accordingly. Alternatively, if Guatemalans are not too keen on lawns, they may spend their money on different goods and services, increasing the demand for labor in other sectors. Different occupations may rise and fall in pay, but there will not be general wage compression.
This is where it’s silly to argue from economic la-La land. The Guatemalans won’t buy houses with lawns, because they are poor. Poor people do not buy houses and hire others to do their lawns. That’s like saying there will never be competition for butler jobs because the new laborers seeking to be butlers will need butlers of their own, keeping demand constant.
To take a simple example, suppose Jack has a job making widgets. His wife Jill gets a job making widgets too. This increases the supply of widget makers, putting downward pressure on Jack’s wages. But Jill puts her new salary towards buying widgets. This increases the demand for widgets, putting offsetting upward pressure on Jack’s wages.
Does Jill put 100% of her wages into buying widgets? If not, this doesn't work.
In the example, widgets are a stand-in for the production of goods and services in general. If women all piled into one line of work while spending their money broadly, then yes, the men who were already in that line of work would have their wages compressed. But that isn’t what happened. Women work in a wide variety of jobs.
Expansion of the labor force does not compress wages. The USA has a much, much bigger labor force than Portugal, because it has many more people. But because American workers also spend their money, American wages are not lower than Portuguese wages.
I suspect that the difference has to do with women!
The US Census Bureau has a nice graph showing inflation-adjusted earnings by gender back to 1959 (link is https://www2.census.gov/library/publications/2024/demo/p60-282.pdf, on page 10 in the text and page 19 in the PDF viewer). It looks like male wages stagnated in the 70s and are still stagnant, while women's wages been consistently increasing over the past 65 years. Earlier in that PDF (page 3 in text, page 9 on the PDF viewer), there's a graph that shows that wage stagnation was relatively consistent across races.
Brookings marks the 1970s as a particularly important point for women in the labor force (https://www.brookings.edu/articles/the-history-of-womens-work-and-wages-and-how-it-has-created-success-for-us-all/). For example, women couldn't independently own credit cards until 1974 (https://womenshistory.si.edu/blog/voices-independence-four-oral-histories-about-building-womens-economic-power).
I'm a little suspicious of this story, because it almost seems too convenient. But at the same time, the financial emancipation of women fits the bill almost exactly. It's a relatively simple supply-side labor anti-shock (what's the opposite of a shock?) that caused male wages to stagnate as female wages rose to catch up (which they have yet to do). And as you've pointed out, there aren't any other larger-scale macroeconomic factors to look for here. None of them would explain why wage trends differ by gender anyway!
The expression you're looking for is probably "positive shock".
https://en.m.wikipedia.org/wiki/Lump_of_labour_fallacy
Definitely the case in the long run; in the short run, anything can happen!
Your story is also supported by how technology allowed traditionally male jobs in agriculture, manufacturing, and other areas to be broken up into different tasks that could be assigned to different workers, including women. Much of supply chain, invoice payment, inventory management, billing became separate tasks from actually making stuff once tracking of inputs and outputs was done with PCs and databases. As tasks like these got factored out, some employment demand shifted away from the guys who used to do it all end-to-end. By the end of the 60s, mainframe database technology had got good enough for this job refactoring, and the arrival of PCs in the 70s made it economical for even smaller enterprises. As another example, contrast how printing was done up to the 1960s (https://en.wikipedia.org/wiki/Linotype_machine), with how it changed with the advent of laser printers (https://en.wikipedia.org/wiki/Laser_printing) in the early 1970s and desktop publishing a bit later.
I’m sorry. That was all my fault.
The period of stagnation closely corresponds to the period between my birth and my graduation from college. The obvious explanation is that the cost of raising me was a huge drag on the economy, but once I became a giver rather than a taker, things turned around.
I have a problem with this post. You make no distinction between the wages of the college educated and the non-college educated. The gap between these wages has been widening steadily since 1978. If you look at overall wages in the United States, the picture is not too bad, but for the non-college educated it has been very daunting. Because this gap between the two groups does not play a role in the minds of too many economists, no wonder the non-college educated have turned to Trump as a false messiah. See:
https://kathleenweber.substack.com/p/how-to-fight-trump-part-one
Working-class men really got shafted, but it happened in waves. The first wave of deindustrialization (factories moving out of the cities) hit the black working class pretty hard (which was then followed by the crack epidemic and the rise in crime). The second wave of deindustrialization hurt the Rust Belt in the 1970s, but the China Shock really did in the white working class. Of course there are other factors to consider: the poverty trap of means-tested welfare programs, deunionization, immigration, etc. But one way or another, it's been slim pickings for the blue-collar crowd.
One thing that puzzles me about that first wave of deindustrialization: economically it didn't make sense to locate factories in city centers after roughly the 1920s, because single-floor electrically-powered factories (which take up too much land for a central location to make sense) were far more efficient that the old-fashioned multi-floor mills that ruled the pre-electrical era when factory machines needed to be mechanically coupled to a basement steam engine.
So why did factories only _actually_ leave the centers of US cities after the race riots of the '60s?
Technically, deindustrialization hit before the race riots of the 1960s, partly facilitated by the construction of interstate highways and the rise of urban sprawl. People like to talk a lot about how "welfare broke the black family," but in reality the black nuclear family began deteriorating way before LBJ even proposed the War on Poverty and the other Great Society programs; the Moynihan Report suggests that the decline started as early as the 1940s, despite the material progress black Americans secured through the New Deal and wartime economic policy. Eventually, that plague spread to white working-class families, and is exacerbated by the fact that American politicians insist on fighting poverty with means-tested benefits instead of generous universal programs that aren't subject to poverty cliffs and other perverse incentives.
Factories couldn’t leave our cities until we finished building out our transportation system—the 40,000 miles of Interstate highways that was essentially complete by the early eighties.
Yes, he did, but in a slightly different way. The dependent variable is “production and non-supervisory employees.”
But he did not directly compare that category to the college educated or supervisory employees. The comparison would have shown a great contrast in increases in compensation.
Yes, because that was not the topic of the article. It was not about inequality. It was about wage stagnation for production and non-supervisory employees from 1973 to 1994 and the possible causes of it.
Saying, “that's not the topic” is like a doctor looking at one arm and ignoring the rest of a patient, which is something you do very regularly in our conversations regarding your Substack.
Noah Smith noted that median compensation and average compensation began to diverge in 1974, which is when the compensation of college educated people began to greatly diverge from the non-college educated, He noted it, but he did not comment on it. I find that disappointing.
Another issue that he failed to address is why the wages of non-supervisory men have only gone from $21.63 per hour to $25 per hour in 45 years. That's a measly 15% raise after 45 years of work!
That's a reality that is clear in today in a graph that he posted.
You can defend yourself by insisting on staring only at the arm, but I'm going to look at the adjacent body parts as well and call on Noah Smith to look at them too.
Yes, you always trying to change the subject away from what the author actually wrote in an article to what you want to talk about is a regular theme. This is likely a defense mechanism to avoid dealing with facts that conflict with your world view.
At some point you are going to have to realize that the author of each article gets to define the topic. If you are not interested in that topic, then do not comment.
You are also obviously changing the subject so you can promote your latest post on Trump (which is also not the topic of Noah's article).
I do have a persistent interest in the fact that the wages of the non college educated have risen little if any in the last 45 years. I believe this is a great factor in the failure of the Democratic Party to address the true needs of the majority of the population. If the Democrats ever took their eyes off racial inequality and turned it to the stagnant wages of the non-college educated, which includes every racial group, then they might be able to do something worthwhile for the country.
I believe that Noah Smith is distracting us from this big and important fact by his argument that these wages were stagnant but now have begun to rise again SLOWLY. Well, the rise is unimpressive, and it has left millions of people miserable and ready to look for anyone who will recognize their problem, and this is has been Trump.
My interest is always in looking at the big picture that affects the most people as opposed to a small and misleading portion of that big picture. I will ALWAYS ask the question, “How does this relate to the big picture?”
I wonder about a sociological explanation: the erosion of norms about acceptable differences between management and worker pay. We all know the basic stats on this. I've never accepted the claim that hypercompetition for middle and upper managers (in a more pure economic sense) means they need be paid more and more to recruit and retain these super talented people. That and pure rent seeking....
Is this an example of "the best way to rob a bank is to own one"?
not the banks my parents and grandparents relied upon, but yeah.
Was there ever such a norm? I hear people complaining that it's higher than it used to be, but I don't see any evidence that people were targeting such a limit back then. Also this ratio is more dependent on stock performance than anything else since it is always calculated on large publicly traded companies. And there are very few of such companies, so I don't think there is really any way to get meaningful data from the compensation of 500 rich people out of the 300 million or so in the country.
There was more wage/salary compression back then, so-to-speak, but this wasn't exactly airtight: executives back then had generous corporate expense accounts that more than made up for the limits in salaried compensation.
That being said, channelling the excess compensation of high-income earners directly into the consumption of goods and services may be a better alternative than letting them use their excess funds to gobble up appreciating financial/real estate assets.
The timing of the post-1993 or so boom looks like it was kicked off by the internet. Maybe the lull before that was caused by the electricity boom having, ahem, run out of gas. There definitely was a feeling in the 90s that household appliances hadn't improved much in the previous two decades, with the microwave being the big innovation and things like remote controls and cordless phones being the fairly weak tea next in line inventions. HDTV was constantly promised but never delivered, and the available channels going from 4 to 50 thanks to cable was considered a big improvement. The rates of growth we've become accustomed to are historically extremely abnormal so the question isn't so much why they have lulls but how they manage to keep going.
I was there, so I can say that there was a giant, huge, very big beautiful Internet boom post 1993. First the computers were installed in offices - I have vivid memories of setting aside my entire Friday mornings to back up a system on floppy disks, remember those? - which had to be inserted and removed in Very Precise Order or your backup would be ruined. We had the office tech and then AOL and CompuServe and Prodigy moved computers into homes, along with the easier availability of desktops (laptops weren’t a thing until the late 90’s and they were so bulky they required a dedicated rolling cart for them, so they were not even really “lap”tops!).
But those were indeed glory days for anyone who wanted a job, and it also helped along the renaissance of cities as places people wanted to live in and not escape from.
It’s been a while but I remember reading that once you take benefits into account, wage growth throughout that whole period looks much stronger, and that especially the slowdown in the early 2000s was mostly healthcare costs eating up wage growth
Not too familiar with the data but couldn’t there also be a demographic effect from boomers and a greater share of women entering the workforce?
Mixed causes seem the most likely answer. Increased imports from China would have at least a temporary impact, and women entering the labor force would result in increased supply. Regarding the last, it wasn't until 1972, with the Supreme Court case Baird v. Eisenstadt, that birth control became legal for everyone, including unmarried individuals. The impact of that decision modified the labor mix, too.
I think the Fed's response to the 1973 energy crisis was inappropriate. The oil price shock was a supply price-induced inflation, not a result of excess consumer demand. The Fed treated the oil price shock in the same way it treated excess consumer demand. The Fed raised interest rates to stifle inflation. The difficulty with that approach is that it contravened what was needed. Businesses needed new investment to adapt to higher energy and feedstock prices. (Remember that oil and gas are not only fuels but also inputs to many production processes.) The Fed's actions delayed for years the shift in manufacturing needed to adapt to both higher and more volatile energy prices. As a partial example, the automobile industry was atrociously slow in addressing Japan's and Europe's competition in higher fuel economy cars, although some of that may be due to ineptitude.
One unanswered question is that the period was also characterized by an increased penetration of information technology. Minicomputers (remember DEC?) came to dominate, and the first PCs landed on employees' desks. Did that counteract some of the reduced productivity trends, but the data isn't sufficiently granular to detect that?
I entered the workforce with my first "real" job in 1972, though I had been self-supporting for at least ten years prior. I have expressed the opinion for many years now that the stagnation of the American economy beginning in the 1970's was the direct result of the confluence of two underreported (or under-considered) economic policy changes. These were the rise of "the Chicago School" economic theories and the adoption of the Harvard Business School's Management By Objectives (MBO). The combination of trickle-down and MBO presaged an era of declining growth in wages, sure, but much more visibly and easier to track, was the huge decline in employee benefits, business lunches, airline perks (despite the introduction of points) and concommitently the increase of "red tape", not in government, but in business, where the increased reporting requirements severely affected morale and job satisfaction, if not the productivity (which increased) of sales, supervisory, and lower to middle management workers. It was in this era that middle-management was hollowed out, by the mass firings of regional managers and the imposition of (previously managerial) reporting requirements on frontline workers. This was further exacerbated by the growth (some forced) in the use of personal computers. I could go on and on about how the confluence of all of these major milestones in American business created the situation which you have ascribed to more memorable events.
Average hourly earnings data does not include medical care benefits provided by corporations. Although including such benefits would not completely change the story about wage stagnation, it might change the visual pattern of the wage graph and support or undermine other explanations.
When I read this post, my first thought was “anyone who has ever listened to any of Bruce Springsteen’s back catalogue knows he built his career around this!” Just off the top of my head, The River, Downbound Train, My Hometown, the desperate guy in Atlantic City… those were all quintessential 80’s. The farm crisis was 80’s. (Those of GenX vintage, or older, remember Farm Aid?) So you are right, it’s been going on a LOT longer than post-Great Recession. We even griped about it and blamed Reagan.
I wonder how much shareholder compensation had to do with it (making sure profits went to stock holders), how much was merging and consolidation, how much was the baby boom glutting the workforce, and how much was the inevitability of Europe and Japan shaking off their post war malaise and rebuilding? The US lucked out in being the only industrialized country left standing after WWII, so we had 20 years of advantage there, and once Europe and Japan were able to rebuild and catch up, we had competition.
Looking back, the big difference between then and the post-Great-recession landscape was that housing was still affordable for most working people, even if it meant sharing a down at heel flat with roommates (I moved into a flat that had a mouse problem, despite the resident cat. I brought a young cat with great hunting instincts. She liked to leave them on the floor for people to step on, but we had no more mice in very short order). I think a lot of us have roommate stories, shabby starter housing stories, etc. and weren’t forced to live with our parents well into our 20’s. Nor was there carefully curated social media to give us the comparison heebie-jeebies; one *could* read Town and Country if that was one’s jam, but otherwise, the Joneses were that group of other roommates in the equally shabby apartment upstairs.
I am puzzled that none of the analysis looks at the size of the labour force (including 'illegal' immigration if this is relevant to these basic stats) over this period, its rate of change and any sociological issues (eg, more women working, all working longer before retirement etc). It is relevant: one can have a relatively constant unemployment percentage with a rapidly growing force, but it's unlikely without wage stagnation or slowdown.
Fascinating analysis. Could some cultural vibe shift explain part of the stagnation? In 1974 there were problems with quality control in the auto manufacturing sector. GM had cars piling up on a lot in Van Nuys, California due to poor workmanship/design/construction/manufacturing capabilities. The guiding ethic was "just get them off the line and we'll sort out the quality problems later". They didn't. Production throughput superseded all other concerns like quality. Honda saw the opportunity and took it. Budweiser had lunchtime restaurants in the parking lots of auto manufacturing firms in Detroit. At lunch it was apparently common to consume a couple of beers before returning to the assembly line. How does the consumption of alcohol contribute to quality construction? The Three Martini Lunch? That was management's adoption of the same concept. In addition, there was the elevation of the acceptability of consumer debt. The populace gradually changed from austerity and avoiding financial bondage which was a reaction to the Great Depression to a full-throated acceptance of acquisition of goods now for the promise of payment later. Of course, there would be interest to be paid, but I don't care, I want it now. The Viet Nam War had created distrust of the government and working for "The Man" was beginning to gnaw at corporate credibility. This was an era when young people grew distrustful of and rejected authority, consequently adopting the attitude of looking out for yourself first. Boomers might have to accept some responsibility for what is happening today. I am from that era and was able to watch it firsthand. Distrust of society and its motivations is a long term trend which antedates the current political dilemma.
The unemployment rate was generally higher during the period of wage stagnation. Maybe the economy was working at below capacity, hence lower demand for labor.
Would love to see an international comparison of the same time series. Is it the same basic picture with slightly different timing or totally different patterns? That might narrow down the explanations a little.