Hacker News Reader

Comment by rtisdale | Hacker News Reader

These articles are almost always based off the wonderfully misleading U-3 statistic instead of the U-6.

https://www.investopedia.com/articles/investing/080415/true-...

The U-6 unemployment rate is nearly double that of the U-3 and is currently at 7.4 percent

https://www.macrotrends.net/1377/u6-unemployment-rate

There haven't been any major gains to wages since the 80's except for high wage workers, who have seen a nearly 35% increase.

Page 4 for table.

https://fas.org/sgp/crs/misc/R45090.pdf

If the unemployment rate is so low why have we seen so little gain to our wages in the past 40 years?

The issues are greater than just faffing about with the unemployment rate but learning to spot this sort of trickery and call it out is important.

Here is a great article on Forbes that goes into some of the reasons why this is the case.

https://www.forbes.com/sites/teresaghilarducci/2018/07/18/wh...

View on HN

No, U-3 is not misleading. See the discussion at https://news.ycombinator.com/item?id=17475481 The point of metrics is to make them comparable. Comparing apples to apples, the US unemployment rate is 3.9% vs the EU average at 7.1%. If you want to use U-6 instead, you can make an orange to orange comparison and get 7.4% for the US and roughly 14% for the EU.

Given that EU wages have increased much more than US wages we can conclude that unemployment has very little to do with wage increases. And since empirical evidence trumps conjectures, we throw the so called "law" of supply and demand into the garbage bin.

And since empirical evidence trumps conjectures, we throw the so called "law" of supply and demand into the garbage bin.

Would you mind expanding on this point? It wasn't supported by your other assertions, and I'm curious how such a counterintuitive conclusion might be true.

I interpreted it as follows:

If unemployment is indeed down (and he asserts that it is) then the supply of extra workers is down also.

The salaries, following the supply/demand ratio, should be up. And they are not.

Since they're not, the law of supply/demand dictating 'prices' does not work for worker's compensation.

What people call "salaries" does not include employer cost of benefits. For instance, US health care spending grew 4.3% in 2016 [0]. Since many employers cover health care in the US, a growth in health care expenditures on the employer side can be seen as a increase in salary. And since health care benefits is a disproportionate amount for low or middle income workers, this will hamper their wage growth the most in relative terms.

In other words, if 25% of your "pay" is health care, and health care is growing at 4%, your effective "salary" can be said to be growing at 1% as long as your premiums and coverage stay the same.

[0] https://www.cms.gov/Research-Statistics-Data-and-Systems/Sta...

I think the meager growth in health care spending is not nearly big enough to cover the discrepancy between the increase in salaries predicted by extrapolating market forces and what the salaries actually are.

Not to mention, most low paying jobs don't provide health care.

This is something that can easily be referenced. Benefits make up around 30% of workers wages with health insurance around 8%. If health care expenditures grow at 4%, that's 0.32% wage growth from health care costs. You may consider that insiginificant but I don't.

https://www.bls.gov/news.release/ecec.nr0.htm

> If health care expenditures grow at 4%, that's 0.32% wage growth from health care costs.

So if you earn $100,000 each year you get a raise of $320 due to health care? To be honest with you, that seems pretty insignificant. And it's also sort of a feedback loop - the quality of care isn't increasing - the system is only costing more because the system can charge more. All that's doing is accounting for "healthcare inflation", and if a cost of living raise is a significant raise, something is wrong.

Excellent point!

Healthcare truly should be the #1 policy discussion in the US, due to the magnitude of its economic impact and how the current status quo is not supported by economists on either side of the aisle.

I wish that someone drafted some economic-concensus presidential mandates for trump to irresponsibly sign before his term is over and someone argues for socialization of healthcare.

Supply and demand can be affected by externalities. In the labor market, those externalities include legal discrepancies between individuals and companies as well as international trade. With inappropriately cheap shipping costs local labor is competing against foreign labor. So the model would have to take into account wage levels abroad. Ie until US wages sink to Brazilian, Chinese, and Indian levels no broad wage increases are likely locally (as a strong interpretation).

Other things to keep in mind: - Labor market is bigger than a single country - Other costs may affect wages (if profits are squeezed, increased competition, rising costs of goods or benefits, etc) - Government policies impact wages (law of unintended consequences)

> The salaries, following the supply/demand ratio, should be up. And they are not.

Supply and demand says that offers will go up. And they have. After all, if you are unemployed, employers see your value as being less than the amount you are willing to work for (which can be no lower than minimum wage). If they could hire you for $0, or a negative amount, they most certainly would. There is always some price point where it is worthwhile.

Those rising offers is what has allowed those previously unemployed to find jobs with agreeable compensation and join the employed where previously offers were not high enough to reach an agreement (by choice or due to the law).

> Since they're not, the law of supply/demand dictating 'prices' does not work for worker's compensation.

The law of supply and demand is being observed exactly as expected here.

Or there are a lot of hidden unemployed that are not captured by the statistics.

A law is a universal truth. If you find that empirical evidence disproves it, clearly it is not universal, and you cannot call it a law without suffixing it with exceptions.

Following from this, supply and demand is a tendency, not a law.

Also, the poster has not applied the law. He is saying that because the labor supply in the US is low, wages should be up. However, the US can outsource functions, which means that the labor supply is bigger than the US.

Anyone who'd read even an introductory economics text would know that such laws are defined "ceteris paribus", meaning all other things being equal. Like in physics: the fact that a law that assumes something is a perfect sphere doesn't work perfectly when something is not a perfect sphere doesn't disprove the law, it just shows that it alone isn't sufficient to model complex phenomena.

The law of supply and demand is a law, but many people don't understand that it is a law in a very basic economic model with major assumptions. If you remove the assumptions in real life, then of course the law will no longer always hold.

I've been reading on the philosophy of science for a while. According to Karl Popper, what distinguishes science from pseudo-science such as Astrology is that the former is able to produce falsifiable theories. Is economy science or pseudo-science? If it is, then its theories needs to be falsifiable.

But apparently you can't falsify the Law of Supply and Demand (LSD). When the world again and again doesn't work in accordance with what the LSD predicts, its adherents always comes with excuses as to why it didn't work in this particular case. But there are huge labor shortages across the whole US, but wages still doesn't increase! So you throw the theory into the garbage bin.

> But apparently you can't falsify the Law of Supply and Demand (LSD). When the world again and again doesn't work in accordance with what the LSD predicts, its adherents always comes with excuses as to why it didn't work in this particular case

It's weird that you dismiss "more advanced models that explain the behavior we observe and provide better predictions" as "excuses".

Yes, it's more convenient to argue against a straw man (the most basic formulation of "supply and demand" that doesn't even cover what's taught in introductory economics classes), but it doesn't mean you'll produce valid criticism from it.

>No, U-3 is not misleading.

The U-3 is exactly what it's defined as. In that sense, it isn't misleading.

It is misleading in the sense that it does not capture what unemployment actually is as it's a very narrow definition of unemployment.

>Given that EU wages have increased much more than US wages we can conclude that unemployment has very little to do with wage increases.

We can conclude from this that it may not be the only factor affecting wages, though I never made that claim to begin with.

You're dismissing supply and demand a little too quickly with a broad stroke.

When it comes to workforce, the supply is directly related to the supply of people able to do a job. Across the entire country, a job that requires little to no training to do will enjoy a much larger supply of workers than a job that requires more skill and experience.

Then there are other factors that essentially cap wages by limiting the cost of a service, such as insurance. Auto insurance, for example, largely dictates the compensation level for repairs and consistently gets tighter and tighter. This means that no matter how effective you are as a mechanic, the business paying you has a limit on how it defines your value because it can't opt to simply charge more for your work.

This applies to other industries as well.

If wages are the "price" in the supply/demand formula, price fixing has to be taken into account.

> Given that EU wages have increased much more than US wages

Where exactly? Surely not in western Europe, it's stagnating since the 80s. In the former soviet bloc for sure I agree.

Please see the chart on this page: https://www.businessinsider.com/france-vs-us-real-wages-2015... Yes, wages have begun to stagnate in the Eurozone in the last 10 years, but it is a completely different situation from the total robbery American workers have been put through!

This shows that although wages in Europe rose faster, they're still higher in the US: http://www.pewglobal.org/2017/04/24/middle-class-fortunes-in....

To me, that implies the possibility of there being a "wage ceiling" for people of a given skill level as the other western nations appear to be on a trajectory that will top out around the same level at which US wages have stagnated. The existence of such a ceiling does make sense given that the distribution of innate ability in a given population is on a bell curve. Even if we assuming a perfect allocation of resources (capital, education, etc), there will eventually be a point at which an individual can be made to be more efficient in performing a task and without groundbreaking advances in technology that drastically change the amount of value an individual can generate at their peak level of efficiency, there will be no economic force acting to increase their wage.

There are plenty of inefficiencies in the current American model of resource distribution, hence the continued slow growth of wages as they get worked out over time, but it looks like we may have exhausted all the "easy" problems given the length of time this trend of slow growth has been in effect.

Great. Now give me an inflation ceiling.

> Given that EU wages have increased much more than US wages we can conclude that unemployment has very little to do with wage increases.

No, we cannot. Unemployment and wage increases could be perfectly positively correlated and still be completely consistent with EU wages rising faster in the US when the EU unemployment rate is twice the US rate.

All we can conclude is that the EU unemployment rate to wage increase function is not the same as the US function, so that a given level of unemployment does not result in the same rate of wage increases in the EU and the US.

There are huge social safety net differences between the EU and the US that might affect how urgent it is for an unemployed person to find a job, and so could affect what employers have to offer to get workers.

In the US, being out of work is a major disaster for many people. If you are in a state that has not expanded Medicaid, you may not be able to afford even the cheapest health insurance for more than a short time with no money coming in if, like many Americans, you do not have much in savings.

> All we can conclude is that the EU unemployment rate to wage increase function is not the same as the US function

If the difference between the outputs has held steady over time...which I dont think is the case.

Start adding other variables into the equation and I have no doubt you can make it work. I also have little confidence in the predictive power of the resulting functions.

I dont think unemployment is UNRELATED to wages, but it seems obvious that the straightforward econ 101 equation has been broken by some factor. I'd rather see people try to figure out what is really going on than assuming the broken equation must actually be right, we just need to depend on it MORE.

One of the primary tasks of a govt is stability, but the recent US trend has been focused on short term stability over long term. I've seen a few (different) theories for what is going on (and am far too ignorant to have a clue myself), but I've not seen ACTION indicationing any concerns are being seropusly considered. I'm sure a chunk of the Fed is concerned, but what are we doing to bleed off the inequality pressure?

> Given that EU wages have increased much more than US wages we can conclude that unemployment has very little to do with wage increases. And since empirical evidence trumps conjectures, we throw the so called "law" of supply and demand into the garbage bin.

One could argue that the reason wages increased in the EU is because it restricts the supply of workers. For example, strong unions could explain that.

Its like you are saying a helium balloon float and conclude gravity has been disproven.

>we throw the so called "law" of supply and demand into the garbage bin

This is like throwing away the law of conservation of energy because you found a perpetual motion machine instead of checking inside the machine for batteries.

For example, are we considering how the impact of workers too scared or financially unstable to leave their current job impacts supply? Or how outsourcing jobs, hiring people who can't legally work, and other such changes impact supply?

We could just fire 3.2% of the nation's lowest wage earners so unemployment equals Europe's and the average wage would instantly be higher.

...Also just pay them more.

If pay gets raised, the unemployment number goes up. If you believe in supply and demand.

For those that don't believe in supply and demand, why are they so stingy that they only want a $15 minimum wage?

There are plenty of good reasons to use alternative unemployment metrics. The one that's usually used (the "official one") is outdated, and mostly useful for estimating the type of unemployment that impacts running costs of unemployment assistance programs. It isn't the best for telling us how many people are working, or as a general labor market health indicator.

That said, it isn't a conspiracy. U6 is probably better as a general indicator. I personally prefer employment/population for each age brackets, fwiw. But... since these correlate, the trend is the same. U6 is always about double u3. Both are lower than they had been. The "structural" threshold for u6 is also double u3.

Even more interesting is the long-term trend of long-term discouraged workers. This doesn't seem to be recovering anytime soon, if you believe SGS - http://www.shadowstats.com/alternate_data/unemployment-chart...

The shadowstats metric is an estimate based on "proprietary modeling". In other words, it's made up with no real methodological backing. The people running shadowstats are pushing a political agenda, not doing real economics.

> The seasonally-adjusted SGS Alternate Unemployment Rate reflects current unemployment reporting methodology adjusted for SGS-estimated long-term discouraged workers, who were defined out of official existence in 1994. That estimate is added to the BLS estimate of U-6 unemployment, which includes short-term discouraged workers.

You can argue the accuracy of the numbers or estimates, but my understanding it is just supposed to take into account long-term discouraged workers, which you can decide whether they matter or not, which are not reflected in official numbers since 1994 according to SGS.

> You can argue the accuracy of the numbers or estimates, but my understanding it is just supposed to take into account long-term discouraged workers

Sure, and my estimate (based on very sophisticated modeling I won't give you any details about) is that the number of long-term discouraged workers is exactly 42.

Here's the entirety of SGS's stated methodology for their estimate by the way:

> The ShadowStats number—a broad unemployment measure more in line with common experience—is my estimate. The approximation of the ShadowStats “long-term discouraged worker” category—those otherwise largely defined out of statistical existence in 1994—reflects proprietary modeling based on a variety of private and public surveying over the last two-plus decades.

SGS's estimates don't even pass basic plausibility tests either. For example, in May 2015 the SGS metric was 23%. Using the published U6 for the same month, we can solve for SGS's estimate for the number of long term discouraged workers: 26,114,100. That's 4x the BLS's estimate of people who wanted a job but did not not have one (6,536,000) and 7x the number who wanted a job and had not looked at least once in the last year (3,712,000) for the same time period.

What's their model? What "public and private" surveys are they taking into account? What makes up the discrepancy between their estimates and the BLS's numbers? We don't know because SGS doesn't tell us, but I think we can both make an educated guess; their numbers are pulled out of their underwear.

True. In the long term, these alternate definitions don't correlate.

Long term "discouragement" is one that comes up relatively often because it can relate to political hot topics.

Even more interesting (to me) are some long term trends that are outside of most unemployment debates. Eg

(1) The long term implications are more about demography than economics, societal age structures and such.

(2)The biggest mover of the last couple of generations has been women's employment. The employee pool basically skyrocketed. Employment/unemployment rates have not been overly impacted by this rapid increase in labour supply. This is still happening. The pool of experienced, female 58 year old lawyers is impacted by the norms/trends if the 80s.

I think this shows how much flexibility the employment system has. Aggregating encourages us to think of demand for labour in macro (kind of Keynesian) terms. I think the truth is that demand for marginal labour (the difference between 4% and 8% in traditional/official unemployment measures) is different to the general demand.

In other words, some portion of the population (different individuals at different times) is harder to employ. That doesn't mean it will look struggle to employ a different pool. If we had another gender to liberate, I think the market could probably absorb it regardless of without impacting unemployment rates very much.

> There haven't been any major gains to wages since the 80's except for high wage workers, who have seen a nearly 35% increase.

Chart employment costs in addition. Notice the huge divergence between the growth in what employers are paying and what employees are getting in wages. Figure out where that money is going and you’ll have the real answer. Though you may not like it.

Link, or way to do that?

Well for one just look at the 5-7% annual increase in employer healthcare costs for the last couple decades, and despair :/. It's like 3-4x as expensive as it was 20 years ago for employers to offer health insurance.

And yet, employers prefer the employer-provided health insurance model, because it increases the friction of switching/quitting jobs.

> And yet, employers prefer the employer-provided health insurance model, because it increases the friction of switching/quitting jobs.

Where do people get this idea from?

No, employers do not prefer the employer-provided model. Especially at scale (employers with more than about 1000 employees), it's dramatically more expensive for them than it would be if they didn't have to operate health plans themselves, and instead could pay employees more to purchase comparable health insurance independently. However, employers aren't the culprit, because they can't change that part of the system. The only choice an employer can make is what kind of health insurance to offer.

There's one backdoor - employers can offer a cafeteria plan, which is what a few employers (like Netflix) do, but in most states, there simply are no good plans (at any price) for health insurance that are comparable to what an employer can provide, so this option is rarely used.

This isn't exactly what you want, but it should give you the idea: https://fred.stlouisfed.org/graph/?g=lfQt

The cost per employee for benefits is definitely a key reason why full time / full benefits wages have stayed relatively level. It's been hidden as well since folks don't normally understand how much their employer has to pay for their benefits. The book Catastrophic Care by Goldhill has a good take on this in my opinion.

That said - there is a reason to be optimistic about all of this. If this demand pressures pushes to increase wages it might provide enough of an incentive to create room for real reform here, as well as transparency around what is happening with benefits that could spur longer-term cost control fixes.

My issue with U-3 isn't that it's misleading in the statistical sense, but that it says absolutely nothing about the quality of jobs that people have. If every software engineer on HN was fired tomorrow and given a minimum wage job at McDonald's, U-3 would be exactly the same. It's not really a meaningful statistic anymore when talking about actual economic health of workers unless you pair it with several other metrics.

What I'd like to see is a metric of Quality Jobs that we can compare over time. Something like % of people working at least 35 hours a week at a single job that provides basic benefits and a living wage in the local market for a single worker.

But wouldn't comparing U6 to U3 be misleading? U6 has always been a double of U3, more or less. Right now U6 looks like U6 in the late 90s.

If you want to talk about the economy’s ability to generate jobs, then U3 is the best metric. U6 contains everybody who chooses not to look for employment for whatever personal reason, whether they’re discouraged or simply just don’t really need or want a job.

It's about what each of these statistics measure.

As another user mentioned, the U3 is a very narrow definition that is largely useful for measuring the cost of unemployment benefits, it does not give a true idea of the real unemployment rate.

This article goes into the differences and their relevance

https://www.investopedia.com/articles/investing/080415/true-...

>The U-3 unemployment rate is a comparatively narrow technical measure that leaves out a whole swath of out-of-work people who are willing and able to take a job but who don't fit the narrow BLS definition of "unemployed." For example, a stonemason who wants to work but who has become discouraged by a lack of opportunity in the midst of a deep economic recession would not be included in U-3 unemployment. A marketing executive who is laid off at age 57 and stops scheduling new job interviews due to her experience of age discrimination would not be included in U-3 unemployment. A person who only works one six-hour shift per week because no full-time jobs are available in his area would not be included in U-3 unemployment.

Exactly. The other gotcha with US unemployment is that you need to add about 2% to compensate for the incredibly large number of people stuck in prisons, compared to other countries. Approximately 0.7% of the population is behind bars, and most of them are working age.

theyre not unemployed because theyre not in the workforce

Well they should be counted. How many had to resort to crime because of no employment opportunities? That’s the real question and I would surmise that there is a percent greater than 0 that would not have stolen or robbed a bank if they had the given job opportunity.

I have heard a bunch of people going into credit card stealing because it pays more than min wage.

U-6 is following similar trend as other Unemployment indicators. It was relatively high in the last decade.

What is your point? It is relatively high but falling.

> There haven't been any major gains to wages

How about unit labor cost?

unit labor cost is the ratio of total labor compensation per hour worked to output per hour worked (labor productivity).

Reply on news.ycombinator.com