Sunday, October 23, 2011

Can Aging Population Explain Income Stagnation?

Since 1970, the number of retired workers receiving Social Security has increased by 159% from 13.35 million to 34.59 million.  During the same period, the number of active employed workers increased by less than half that amount, by 77% from 78.7 million to 139 million.  The fact that retired workers have increased so significantly relative to active workers since 1970 can be explained by advances in medical care that have increased life expectancy by 20 years since 1930.  

The top chart shows that the ratio of active workers to retired workers decreased by almost 32% since 1970, from 5.90-to-1 in 1970 to 4.02-to-1 by 2010.  The bottom chart shows the inverse - retired workers as a percent of active workers - which has increased from 17% in 1980 to 24.8% in 2010.    

So what? Well, perhaps this demographic trend of an aging U.S. population explains why real median household income has stagnated in recent decades, according to Census Bureau data. Relative to active workers, we now have significantly more retirees, many of whom might have significant assets (a house with no mortgage, stocks and bonds, mutual funds, etc.) that make them very wealthy, but in retirement would be receiving relatively low annual incomes compared to when they were working full-time.

With an increase in retirees, we have thousands or millions more Americans over time moving from top household income quintiles while working, into lower income quintiles when in retirement,  bringing down average or median household income as measured by the Census Bureau.

Therefore, the growing ratio of retirees to active workers makes it appear that household income is stagnating, when in fact the median household income of active workers (or real compensation per hour, see chart below) could be increasing, and the average wealth per household could also be increasing.  

Q: Could the aging of the U.S. population make it appear that we're in a period of "Great Stagnation" based on stagnating household income, when in reality the finding of stagnation is only a "statistical artifact" or "spurious finding" due to a significant demographic shift over time?

Comments welcome.  

Update: The chart below shows no stagnation over time in "Real Compensation per Hour," which has increased by almost 30% since 1990 and almost 14% since 2000 (thanks to Peak Trader).  Perhaps another factor in the "non-stagnation story" is that fringe benefits have increased over time relative to money income, and that increase in compensation is not captured by the Census Bureau when it calculates household income. In that case, household money income could be stagnating at the same time that total household compensation is increasing.    

The non-stagnation of real compensation per hour.


At 10/23/2011 7:31 PM, Blogger Stephen Purpura said...

The BLS breakout data that are usually reported are for: full time workers who receive wages, salaries, commissions, tips, payment in kind, or piece rates. The group includes employees in both the private and public sectors but, for the purposes of the earnings series, excludes all self-employed persons, regardless of whether or not their businesses are incorporated. Full time workers are people who usually work 35 hours or more per week at their sole or principal job."

At 10/23/2011 7:36 PM, Blogger Stephen Purpura said...

This comment has been removed by the author.

At 10/23/2011 7:36 PM, Blogger Stephen Purpura said...

An example BLS report is available here:

And it shows that, in constant 1982 - 1984 dollars, that the median weekly income has gone from $337 (in 2002) to $335 (in 2011).


For this data series, I do not know if there have been refinements in population gathering during the periods.

At 10/23/2011 7:44 PM, Blogger PeakTrader said...

Real Compensation Per Hour is way up since the mid-'90s:

At 10/23/2011 8:25 PM, Blogger Mary Robinson said...

Peak Trader,
It looks like
is not in constant dollars? In which case, inflation needs to be factored in.

At 10/23/2011 9:03 PM, Blogger Mark J. Perry said...

It's REAL compensation per hour, meaning that it's been adjusted for inflation.

At 10/23/2011 9:27 PM, Blogger Stephen Purpura said...

Nonfarm business includes financial sector workers (I don't think that's wrong, but it is skewed from median worker by the high end). Also, hours per worker are dropping (nonfinancials they are at the lowest point since 1995).

At 10/23/2011 10:48 PM, Blogger Hydra said...

One thing we know is that income is NOT stagnating for the top 10% or the top 1%.

Maybe that is skewing the data.

At 10/23/2011 11:29 PM, Blogger juandos said...

"One thing we know is that income is NOT stagnating for the top 10% or the top 1%"...

Ooh! Ooh! Jealousy rears its ugly head among the questionably useful...

At 10/24/2011 12:15 AM, Blogger Hydra said...

Who said anything about jealousy? I just made a suggestion about data.

I have enough not to need jealousy, but not so much as to fear the government will take it: they have bigger fish to fry.

At 10/24/2011 12:18 AM, Blogger Hydra said...

Juandos: I notice you are not disavowing my statement: merely continuing your usual pathetic attempts to promote dogma over truth.

At 10/24/2011 12:23 AM, Blogger Hydra said...

Questionably useful. I like that.

Are you now signing up to my suggestion that we should accept the fact that we really do NOT want some people to work? That we might be better of to have some people NOT work?

At 10/24/2011 12:25 AM, Blogger Hydra said...

You do agree that income for the top 10% has not been stagnant, right?

At 10/24/2011 12:27 AM, Blogger B-Daddy said...

An alternate explanation for the causal relationship is that the retiring workers (baby boomers) leaving the workforce are at their peak earning power. However, their are fewer members of the next generation following them (Gen X) so the average is getting skewed towards a younger work force with fewer skills that therefore earn less. Look at the demographic bulge in the current U.S. population profile.

At 10/24/2011 1:48 AM, Anonymous Anonymous said...

Hydra, of course they're not stagnant, they're highly volatile and debt-financed to boot.

At 10/24/2011 6:55 AM, Anonymous Anonymous said...

Changes in demographics are important to this story, but I don't think explain it away. In this case, they don't say what Perry says.

In the first two charts, the current readings are due to current high unemployment. Look at those charts from the end of the early 1980s recessions to 2007 and you don't see much change at all.

In the bottom graph, I guess I have to say it again: if you're plotting a series that grows over time, you MUST use a logarithmic scale to accurately compare growth rates. Here's the same data series graphed logarithmically. While not showing stagnation, the difference in growth rates before and after ~1970 becomes apparent.

At 10/24/2011 7:34 AM, Blogger Frozen in the North said...

MJP that's the biggest problem with Median income. Demographic that in the past have helped the number (growing population and retiree being a small number). However, and this is important, the implication is also that median income IS stragnant, creating an important wealth effect.

Also income (in total) is rising, but its distribution is increasingly skewed -- as the data for social mobility seems to confirm (those on top account for a greater proportion in the increase in income).

Proof, as if any was needed, that understanding where (and if) the problem lies. Breaking down income growth by decile shows in fact, that income as accrued mostly to the top economic strata.

At 10/24/2011 8:48 AM, Blogger Jon said...

Family working hours have risen dramatically over the same time period, so I don't think additional retirement explains the disparity.

Per hour wages since 1970 are down.

At 10/24/2011 1:07 PM, Anonymous Anonymous said...

Looking at wages alone isn't sufficient; if you're trying to understand compensation, you have to include non-wage compensation as well, which has risen substantially primarily due to rising health insurance costs.

Anybody who points to wage data rather than total compensation data is either trying to fool you or has a few things to learn.

At 10/24/2011 2:05 PM, Blogger rollo tomassi said...

Screw numbers on figuring wages. Use a comparison to the CURRENCY value since the 70's. THERE'S where the secret lies.

At 10/24/2011 3:04 PM, Blogger houston15 said...

Don't forget that the size of the average household has declined over the last 30 years. Fewer breadwinners per household.

At 10/24/2011 4:22 PM, Blogger PeakTrader said...

Frozen in the North says: "Breaking down income growth by decile shows in fact, that income as accrued mostly to the top economic strata."

That doesn't explain where the gains in real compensation took place.

Even if the steep rise in real compensation took place in higher-wage jobs, it likely means tens of millions of better-paying jobs were created in the Information Revolution.

Some people view that as a negative, because some people decided to miss the train and stay at the station.

At 10/24/2011 4:38 PM, Blogger PeakTrader said...

Voxrationalis says: the difference in growth rates before and after ~1970 becomes apparent.

The steeper rise in U.S. real compensation between 1945 and 1970 was the result of WWII.

The U.S. economy was intact, while the world's major economies were rebuilding.

Perhaps, the U.S. reaching peak oil production in 1970 had something to do with it too.

At 10/24/2011 5:13 PM, Blogger VangelV said...

It's REAL compensation per hour, meaning that it's been adjusted for inflation.

It is only legitimate if the inflation rate is measured properly. I see no evidence that it is. Note that the chart does not apply the post-Boskin methodology to all of the data. If it applied the pre-Boskin methodology it would show a decline over the past two decades.

At 10/24/2011 5:37 PM, Blogger PeakTrader said...

I wanted to know about compensation at a typical U.S. tech firm. So, I checked EBAY:

eBay Compensation and Benefits

•Medical, dental and vision insurance from the date of hire.
•Life Insurance, AD&D, short term and long term disability.
•Flexible spending accounts,
Employee Assistance program
•Business Travel Accident insurance
•PTO starts at 16 days per year, with an additional day added for each year of service up to 20 days per year.
•Holidays - 10 holidays per year, plus one floating holiday of your choice
•Sabbatical After five years of service with eBay, you are eligible for 4 weeks of time off with pay.
•401k plan with participation eligibility on hire, 100% vesting in all contributions (including employer contributions) and a 100% company match up to $2000 per year.
•Employee Stock Purchase Plan
•Employee Referral Program – Up to $1000
•Charitable Contribution & Gift Matching

eBay also offers a lot of additional perks, including the following:

•Work/Life Balance Toolkit
•eBay Development Workshops
•Tuition Reimbursement
•Adoption Assistance Program
•Ergonomic Consultations
•Pet Insurance
•Onsite Conveniences including: complimentary beverages and snacks, ATM’s, onsite dry cleaning, massage, auto detailing, oil change, hair cut and dental cleaning – just to name a few.

At 10/24/2011 5:47 PM, Blogger PeakTrader said...

Here's an interesting tidbit:

The very first item sold on eBay was a broken laser pointer, to a collector of broken laser pointers.

At 10/25/2011 9:18 AM, Anonymous Anonymous said...

PeakTrader wrote: "The steeper rise in U.S. real compensation between 1945 and 1970 was the result of WWII."

This might fly if real GDP growth also experienced the flattening of the compensation curve. It didn't.

At 10/25/2011 9:27 AM, Blogger Jon said...

vox, your point on compensation makes sense, but the problem may be that it's just difficult to show. Since health care costs are a non-income benefit they in a sense can be left out of the equation because their increase only means the compensation is constant as adjusted for inflation. Not exactly, but close. In fact health care expenses are now coming out of income as co pays increase.

The other compensation to consider is pension benefits vs 401k. Pensions have all but disappeared at this point and have been replaced by the 401k. The matching contribution sometimes is present and sometimes gets yanked depending on the economy. All told I expect this is a further reduction in compensation, but I don't have numbers to back that up.

At 10/25/2011 4:02 PM, Blogger Your humble blogger said...

Senor Krugman 'splains all here:

At 10/25/2011 4:39 PM, Blogger PeakTrader said...

Voxrationalis, your chart looks like real GDP grew faster from 1945-1970.

Anyway, how does your chart prove workers were over-compensated or under-compensated after WWII?

All it really shows is a long-run growth trend, i.e. too much growth is inflationary and too little growth is recessionary.

At 10/25/2011 4:57 PM, Blogger Jet Beagle said...

B-Daddy:" the retiring workers (baby boomers) leaving the workforce are at their peak earning power."

That's good logic, and a good explanation for explaining trends we will see in the future. But I don't think many Boomers have retired so far.

If you look at the bulge in the chart you referenced, you can see that the peak Boomer bulge was aged 45 to 54 in 2010. That group has not yet started to retire. The group behind the Boomers - those born 1965 to 1969 - is actually larger than the early pre-1950 Boomers who are now retiring.

At 10/25/2011 5:19 PM, Blogger Hydra said...

Hydra, of course they're not stagnant, they're highly volatile and debt-financed to boot.


Since when do you count income that is debt financed as income?

What has volatility got to do with the general trend being upward.

Are you really claiming that the top ten (or one) percent have not generally increased their share of total income earned over the past 20 years? And that they have not also increased their share of the total wealth?

At 10/25/2011 7:43 PM, Anonymous Anonymous said...

Hydra, since when do you not count income that is debt-financed as income? More debt simply means that the income is more sensitive to fluctuations, as we saw with all the investment bank failures recently, but it's still income. If you read the article, you'll see what volatility has to do with the trend: it means that they're rocketing all over the place, with only 27% of the top 400 ever making the list more than once in the last 17 years. That means new people are constantly working into the top 1% and others are constantly falling out. That's why when you make idiotic claims that the top 1% have increased their share over the decades, that's not true, because who's in the top 1% is constantly changing over that time period. Of course, Sergey Brin or Jeff Bezos or some hedge fund manager du jour may rocket into the top 1% and make a couple billion more than the rich who formerly occupied that slot, but they generally don't stay there long. That's a sign of a healthy, competitive market, where competition constantly creates new winners.

And as I've pointed out many times before, new technology is fundamentally changing the nature of the market, where many fields that were winner-take-a-lot are now newly and rabidly competitive. Katie Couric made $15 million/year because there were only three broadcast anchor slots and a lot of old people still cling to the broadcast news. But as these seniors die off one by one- they're the only ones still buying print newspapers and keeping those dinosaurs alive, btw- the competition online for everybody else is rabid: Brian Williams or Scott Pelley will be replaced by hundreds of such anchors online, each getting paid a fraction of what the broadcast anchors made. This is what's killing off the "rock star" musicians and newspapers already and the trend is only beginning to pick up steam. :)

At 10/25/2011 10:59 PM, Anonymous Anonymous said...

PeakTrader wrote: "VR, your chart looks like real GDP grew faster from 1945-1970."

True. But the point is that compensation started falling around 1970 relative to GDP. Let's make 1970Q1 our arbitrary marker. From 1947-1970, real GDP grew at 3.88% per year, and real compensation grew at 2.64% per year (about 2/3 the rate of GDP growth). From 1970-2008, real GDP grew at 3.05% while real compensation grew at just 1.1% (about 1/3 the rate of GDP growth). If the post-WWII period explained the ~60% drop in compensation growth starting around 1970, there should also have been a fairly comparable drop in GDP growth starting at the same time. Didn't happen.

"Anyway, how does your chart prove workers were over-compensated or under-compensated after WWII?"

I was never trying to make such an argument.


Post a Comment

<< Home