# Productivity vs real earnings in the US — what happened ca 1974?

Somewhat related to the most recent US elections, I've been researching the whole "white working class" situation, and one strange anomaly has popped up.

When one looks at a graph of productivity (for the US) vs real wages, there is a marked "disturbance in the force" around 1974.

(11/18/18 note: A nearly identical graph appears in Scientific American, Nov 2018, p 61, in an article titled "A Rigged Economy".)

Is there any agreement among economists as to what caused wage growth to plateau like this?

Update:

I studied the article that luchonacho suggested, by Bivens & Mishel (2015), and found it interesting, but it didn't directly address my question. The article was basically a justification for the numbers presented in the above graph and similar presentations, in particular arguing against skeptics that the productivity numbers were real -- that productivity hadn't leveled off with wages. I'm not able to evaluate these arguments to any significant degree, but at least on the face of it their justification seems sound.

However, that article pointed me to an earlier article by Bivens, et al (2014) which does attempt to identify the factors that caused the change of trajectory of productivity. I'm still evaluating this article, but it can be noted that:

1. This was during the Nixon administration
2. This was during a period of fairly serious "stagflation"
3. Nixon ordered his wage/price freeze in August of 1971
4. The Arab oil embargo began in 1973
5. The US went off the gold standard in 1971
6. The US trade balance went negative in 1972 and has remained mostly negative ever since
7. Union membership, which had been declining since the 50s, began a steeper slide in the 70s
8. OSHA was founded in 1971
9. And, of course, the CEO-to-worker compensation ratio climbed slowly but steadily through the 70s (after being relatively flat for the entire postwar period), then accelerated rapidly in the late 80s

Which of these factors (or several other ones too "minor" to mention here) is relevant to my question is hard to guess at this point. The issue is confused by the fact that, though the "knee" in the curve appears to be 1973, that was a period of significant economic turmoil (probably the worst since WWII and before the Bush-era difficulties), so it's not unexpected that the lines would "jiggle" a bit, and thus it's hard to place a date on the actual change of the nature of the economy that caused this phenomenon.

I will continue investigating, and I would appreciate any (constructive ;) ) input.

Further update:

Finished a first review of the Bivens (2014) article, and it didn't have anything real solid to offer. Several contributing factors were mentioned, but most didn't kick in until the 80s, or later, and the general drift of the article was to propose "fixes" to current conditions rather than explain the triggers that caused the phenomenon in the first place. The one factor that might have applied early is private sector union membership, as it took a nose-dive in the early 70s. (This was made up, in part, by a rise in public sector membership.) And some argument can be made that tax policies had some part in things, though the evidence for this is weak.

However, while reviewing the article I went off on a few tangents, exploring statistics that the article did not directly address. I (a bit to my surprise) found some seeming correlations between the "real wage" fall off and the growth of money supply, the negative balance of payments, and the rise in national debt, all of which changed course in the early 70s, resulting in permanent deviations from historical patterns.

It seems strange to me that this apparent correlation is not (so far as I know) widely discussed, as one would think that "budget hawks" would zero in on any apparent connection between these factors and wage levels. (Of course, they might not have noticed, as the same factors appear to be making CEOs quite wealthy.)

One thing I haven't found, but which I suspect, is that some changes in the rules for corporate governance may have been a factor.

Still searching.

Tentative conclusion:

In 1971 President Nixon supposedly said "I am now a Keynesian in economics", after Milton Friedman had previously stated "We are all Keynesians now". The main issue at play here was the idea of intentionally running a federal budget deficit to "spur" the economy.

There is no single event to point to, but over about 10 years, from 1970 to 1980 tax rates were lowered significantly for some segments of the population -- basically the 1%ers and corporations, and large amounts of Treasury bonds were issued to make up the resulting deficits. (Prior to Nixon's pronouncement Republicans, in particular, were deficit hawks, but the change in viewpoint gave them "permission" to lower taxes without cutting spending, and similarly allowed Democrats to raise spending without raising taxes.)

The result is that the federal budget has been (with the exception of 1998-2001) running (ever increasing) deficits.

And, close as my Ouija board can decipher, this has the effect of drawing in a lot of foreign investment which, oddly, upsets the balance of trade in the wrong direction (the balance of trade has been negative since about 1975).

This imbalance is commonly blamed on "free trade policy", but basic economics argues against this -- especially with "free trade" the value of the dollar should self-adjust to the point where exports and imports balance.

Of note is the fact that about 31% of the federal debt is held by offshore entities, and (using slightly stale numbers, and assuming that the ratio is relatively steady over time) that amounts to about \$370B of the annual deficit, as compared to the roughly \$500B trade imbalance. So one can argue that deficit spending accounts for the lion's share of the trade imbalance.

And an important point is that the top income tax bracket (the one that had tax rates reduced from 70% to 35% between 1981 and 1988) accounts for (rather roughly calculated) about \$500B in tax revenue, and presumably would produce much more (let's say another$300B) if taxed at the pre-1981 rate. (And, to the extent that this increased revenue did not materialize, it's likely that the incredible expansion of the CEO/worker pay ratio would be reversed, as companies judged it more useful to put the money somewhere other than into more taxes.)

So my conclusion is that a "Keynesian" tax policy (accompanied by an ever-stronger tendency to under-tax the rich) is what is mostly responsible for the plateauing of real wages, and restoring pre-1981 upper bracket tax rates would likely alleviate much of the problem.

What say you?

(I'll note that I also suspect a change in corporate governance somehow fits in here, but I haven't researched that very much yet.)

Update 5 June 2017

My wife, knowing I'd been interested in this topic, jotted down a reference she'd heard on the radio for a book called The CEO Pay Machine, by Steven Clifford. Though I've not attempted to verify any of his assertions the book is published by Penguin/Blue Rider, so I assume it's reasonably honest and accurate. (It also has some pretty good end notes and a decent index. However, Clifford's writing style is rather plodding and not well-organized.)

Anyway, Clifford claims that, ca 1980, corporate boards were in effect sold a new bill of goods by Michael Jensen and Milton Rock. These guys put forward a philosophy of corporate management which emphasized giving CEOs "incentives" of various types, vs a simple salary and straight-forward stock options. They also promoted the idea of comparing CEOs to their peers, in terms of pay, and targeting some Nth percentile of their peer group pay.

These proposals seemed innocent enough, and many corporate boards picked them up (also accounting firms and various consultant groups, as they saw that they could profit from assisting in their implementation).

In theory "pay for performance" seemed good, but it quickly got out of hand. Obviously, when calculating CEO base pay for your company, you don't target the 50th percentile of "peers" but go for 60, 75, even 90, because clearly you expect your CEO to be exceptional.

(Consider, for a minute, what it means if you have a bunch of companies in a "peer group" and, each time CEO pay comes up for a vote in one of the companies, they award the 70th percentile of the group. It's a spiral, and not a downward one.)

And various stock bonus schemes strongly encouraged CEOs to play games to manipulate stock prices. Plus, as it turned out, boards were really lousy at withholding bonuses when they weren't really earned.

Then, in 1993, Clinton and the Democratic Congress pushed through a tax increase bill that had one tiny loophole -- stock options for CEOs would be exempted from corporate taxes. This resulted in an explosion of stock option awards, and overall CEO pay jumped from 100x the average worker to 300x almost overnight.

This seems to fairly completely account for the effects seen, in terms of escalating CEO pay and a portion of the lost "working class" jobs/wages. Unfortunately, the way out of this quagmire is not obvious. Clifford offers some suggestions, but getting anything through (CEO-owned) Congress will be difficult, to say the least.

Update 30 May 2018

The New Yorker published, in it's May 14, 2018 issue, a review of the book Can Democracy Survive Global Capitalism (Norton) by Robert Kuttner. (The review is written by Caleb Crain.)

As is the case for many articles in that magazine, the review borders on impenetrable, but it does discuss factors surrounding my issue at considerable length. Based on the review, the book glorifies the economic system, both US and international, that existed after the Bretton Woods agreement (1944), and prior to 1973. According to Kuttner, 1973 marked "the end of the postwar social contract." To quote Crain, "Politicians began snipping away restraints on investors and financiers, and the economy returned to spasming and sputtering. Between 1973 and 1992, per-capita income growth in the developed world fell to half of what it had been between 1950 and 1973." Income inequality increased, the median real income of "working class" Americans fell. And, significantly, "faith in democracy slipped."

Kuttner/Crain discuss a number of things that went on beginning around 1973 (including the fallout of the Arab oil embargo), but, from the philosophical standpoint they hang the economic turn-about on the return of laissez-faire philosophy to the political sphere. Eg, in January, 1974, the US removed the constraints on sending capital abroad, and in 1978 the Supreme Court overturned most state laws against usury. A litany of effects basically gutted Keynesianism as it had previously existed, and the US economy (and the world's) lost its balance.

(I'll note that this point of view is about 180 degrees different from the Nixon/Friedman "We are all Keynesians" view. I gather Kuttner regards the Nixon/Friedman postures as false.)

• Spend a lot of time looking at the rise of automation, which also took rise in the 1980s on an industrially disruptive level. At an increasing rate we are seeing the redistribution of wealth from labor owners to owners of capital -enriching the already rich. Those workers that are able to keep their jobs see their wages rise but on average companies have let go dozens of employees. A Brookings Institute report on technological innovation from 2015, stated that within a generation at current trends, it is possible that at any given time a quarter of middle aged men will be out of work. – Ian Brigmann Mar 1 '17 at 19:22
• @IanBrigmann - Except that "the rise of automation" has been at work for 200 years. There's nothing really unique about the last 40 years. I agree that there is a redistribution of wealth to the "one percent", but that isn't explained by automation. – Hot Licks Mar 1 '17 at 20:58
• Automation alongside the internet. – Ian Brigmann Mar 2 '17 at 16:45
• @IanBrigmann - The internet actually brings very little to the table. It provides faster communications, and opens up some new markets (on-line videos, etc), but doesn't really shift any paradigms. "Online stores" are really just a new twist on the Sears catalog. Likewise many other online activities that simply replace the US Mail. The projection that the unemployment rates of lower wage and older workers will increase (given current policies) is likely true, but automation is more an excuse than an explanation. – Hot Licks Mar 2 '17 at 18:06
• Automation is an excuse rather than an explanation? You must really be joking. – Ian Brigmann Apr 3 '17 at 18:11

Here is one explanation, albeit focusing on the wider economy, and comparing median wages. Still the method applies:

This is taken from Bivens and Mishel (2015). In essence, they decompose labour productivity into its different components (see technical appendix, page 25). These components are three:

• Labour share: how much of total product (they use Net Domestic Product, NDP) is paid to workers

• Terms of trade: discrepancies between NDP deflator and CPI. Recall real wages are deflated with CPI, whereas GDP price index include other stuff like price of investment, terms of trade, etc.

• Inequality of pay: a rough measure of the difference between mean and median pay. If wages are equally distributed, these two are equivalent.

If you are interested on mean wages, then the third component does not apply. The other two do. Thus, a lot of the problem is due to prices issues. As the authors note:

That is, workers have suffered worsening terms of trade, in which the prices of things they buy (i.e., consumer goods and services) have risen faster than the prices of items they produce (consumer goods but also capital goods). Thus, if workers consumed investment goods such as machine tools as well as groceries, their real wage growth would have been better and more in line with productivity growth. We sometimes refer to this terms-of-trade wedge as the difference between “consumer” and “producer” price trends. (p.6)

• I have noticed that many charts, such as yours above, begin at around 1975, thereby omitting the anomaly. Is there any specific reason why this might be the case? – Hot Licks Feb 24 '17 at 16:45
• Figure A in the document I refer to has the long term picture. They are not omitting the earlier period, but simply focusing on the relevant one, for the sake of the argument. In other words, they focus on that particular period because that is the period where the anomaly exists. – luchonacho Feb 24 '17 at 17:02
• OK, thanks! That article appears to be dead on to my question. But it's going take some study to understand. – Hot Licks Feb 24 '17 at 17:21

Part of the stagnating real wage anomaly could be explained by the increase in labour supply that occurred in the US labour market in 1975 due to the participation rate rising. The US participation rate jumped from around 60% in 1970 to 64% in 1980. As the labour supply increased, the price of labour decreased due to the competition created by the additional supply. As the Marginal Productivity of Labour is diminishing with quantity, additional labour supply would cause the marginal productivity of labour (MPL) to fall and thus the value of nominal wages (assuming wage is a function of price and MPL, Wage=Price*MPL) to also fall.

In terms of stagnating real wages, that is, the number of goods and services a household can consume, the graph again contradicts what we would expect to see from rising productivity. That is, higher productivity should result in a reduction in the price of goods and services. To explain this, the relative decrease in the nominal wage (as discussed above) would have to be equal to or greater than the reduction in the price of goods and services caused by increases in productivity. This solution appears ambiguous at best as it is unlikely that nominal wages would fall significantly from a 4% increase in labour supply.

An analysis of labour supply in the US, during and following the 1970s can be used to explain some of the stagnation in the value of real wages but cannot be used conclusively to explain all of the lack of real wage growth.

http://gonzoecon.com/2012/01/december-unemployment-rate/

• But higher productivity only produces lowered prices (and increased wages) if the corporations (and their owners/CEOs) aren't siphoning off their gains somewhere along the line. – Hot Licks Sep 12 '18 at 11:34
• Fair point, we can only assume that higher productivity would have these beneficial effects if competitive markets force firms to pass on the benefits of productivity. With the US increasingly being exposed to international competition firms should have more incentive to reward productivity, though historically this has been problematic. – Michael Elliott Sep 13 '18 at 4:21
• Certainly the benefits of the recent corporate/1% tax cuts are not reaching workers (and consumers) to any great degree. – Hot Licks Sep 13 '18 at 11:45
• Perhaps not directly, even if we assume the worst case scenario where the 1% simply bank profits made through tax cuts then their will be some beneficial side effects. As banks receive more deposits their supply of money shall increase. This will enable them to loan money at lower interest rates. A lower cost of money should increase investment and consumption and therefore stimulate growth in the private sector (and therefore reduce unemployment). – Michael Elliott Sep 14 '18 at 1:03
• Spoken like a true 1%er. – Hot Licks Sep 14 '18 at 1:21

The house is taking a percentage right off the top. In 1979, the top 1% of the population's share of U.S. income, including capital gains, was hovering around 9-10%, with the top 0.1% having about 2.5-4.0%. In 2013, those percentages were 20% and 9.5%, respectively. In 1979, the bottom 50% in the U.S. earned 20% of all income. As of 2013, that share had dropped to 13%. All income group segments of the U.S. population (by quintiles) experienced real pre-tax and after-tax income gains from 1979-2013. The "rising tide lifted all boats," though not evenly. Capital and equity owners have been taking excessive profits because they can. Labor has been left in the dust: because employees became satisfied.

Policies that made it happen: Taxation is less progressive; unions power has decreased or at least have not grown. See Martin Ford's 2015 book Rise of the Robots. It is just starting, and this trend is consistent for the last thirty years. Slowly but surely, jobs that used to afford a living for families have been disappearing in favor of machines, systems and outsourcing. That in itself is not a bad thing. BUT humans are not able to adapt quickly enough, and labor is slowly being cheapened compared to whiz-bang machines and systems.

• Please try to keep the answer more analytical and less opionated – Guy Louzon Oct 24 '18 at 7:33

Benefit costs have been increasing. One hypothesis is that the overall compensation increases have been going to benefit increases rather than wage increases. I.e. if my employer is paying \$5000 a year more for my healthcare, it may be unwilling to give me \$5000 cash in addition.

Another hypothesis is that inflation undercounts quality improvements. If my house is 50% bigger than my parents' house and costs 50% more, is that 50% inflation? Or zero? Or something else?

Also consider some things that are ubiquitous now.

• Television in every bedroom. I grew up with just one in the house.
• Everyone owns their own computer. My family didn't own a single computer until I was around ten. My parents each have their own now.
• Mobile phones. I was the first member of my family to get one as an adult. Many families now have one for each person, including the kids.
• DVD (or Blu-Ray) player. My family got its first VHS player after the computer.
• Microwave. Another appliance that appeared in my house when I was older.
• Digital Video Recorder. Not ubiquitous yet, but much more common.
• Internet. My family got dial up years after the computer. Now everyone has broadband.
• Cable. Once you already have broadband, adding cable TV is cheap. Particularly if you also have a landline.
• Cars are much more advanced now. E.g. higher fuel economy.
• Prepared meals have replaced garden produce. As a kid, my family made its own tomato sauce from garden-picked produce. Now, I buy that in a jar or even get a whole frozen dinner that is already cooked and just needs heated.

How is it that we have so much more stuff on less money?

The cause of the wage and productivity growth gap varies a lot by each sector so it's useful to look at that data. If we look at a breakdown of the last 30 years, we can probably try to understand things in a more nuanced fashion:

The erosion of labor unions (directly and indirectly) probably accounts for some of the failure to translate productivity into wage growth but it's important to not forget labor productivity is synonymous with labor efficiency.

1. In the retail, wholesale, and transportation and warehousing sectors, in the past large amounts of time used to be spent checking inventories. Today this needs to happen much less often and therefore makes labor more efficient. Software automatically calculates what inventory needs to be restocked and optimizes the process by which it's done. Labor has been reduced to flesh drones that are utilized much more efficiently.
2. In the wholesale sector, in the past workers had simple 8 hour schedules. Today, Amazon workers only work for while there are packages that need to be processed so labor is fully utilized.
3. In manufacturing, it's just in time manufacturing processes and better automation.
4. In the information sector, email and instant messaging increases efficiency in communication. Googling answers to common problems saves workers hundreds of millions to billion hours a year easily.

There are other factors certainly but the biggest law passed 50 years ago that's contributed most to the rise of productivity without wage growth is Moore's Law. In the end, software has made us all much more efficient commodities.