David R. Henderson  

How Real Wage Increases Have Been Understated

PRINT
Charles Murray Watch... Foreclosure Paperwork...

Over the years, In discussing the alleged decline in real U.S. median wages, I've pointed out that there are two important ways in which the growth in real wages has been understated:
(1) The inflation adjustment used to compare wages over time is the Consumer Price Index. As Michael Boskin has shown, the CPI overstates the increase in the cost of living.
(2) The wage data typically exclude non-monetary benefits, one of the main ones of which is health insurance. Of course, health insurance has been getting more expensive but one reason is that we're getting more for it.

But I had an interesting conversation with one of my favorite liberal economists ("liberal" in the statist sense of that word), Ken Judd, at Hoover a couple of weeks ago. Ken grew up on a farm in Wisconsin and worked 7 days a week from a fairly early age: milking cows, etc. This was in all types of weather: cold, heat, rain, snow, etc. But now, he pointed out, so many jobs are so much more comfortable: workers in manufacturing plants who have air conditioning, etc. This, he noted, is an increase in real wages.

Moreover, there's been a secular decline in fatality rates on most jobs. That doesn't get captured in wage data. In fact, all else equal, wages are lower when jobs get safer.


Comments and Sharing


CATEGORIES: Labor Market



COMMENTS (14 to date)
Mr. Econotarian writes:

The real data you need is right here Real Compensation per Hour.

Looks like it has been going up, and at an all time high in 2010Q3 (although it took a bit of a dive in the last few quarters).

Let's Be Free writes:

Yep, and most of the time the data are taxable incomes, which fails to recognize that the middle class has benefited from growing use of tax shelters, first IRAs, then 401 Ks and things like tax sheltered education and medical accounts. It's obvious that standards of living have steadily increased which can only happen if, in real terms, incomes have also increased.

Michael Stack writes:

Hmmm, that graph tells me a different tale. From 1950 to 1970, a twenty year period, real wages per hour increased about 75%. From 1970 to 2010, a 40 year period, twice larger than the earlier period, wages increased less than 50%.

Clearly wages did not stagnate entirely but the rate of increase slowed dramatically.

While there are corrections that maybe ought to be made to some of the reports the bigger picture still comes through - wages aren't increasing at nearly the rate they once did.

Montani Semper Liberi writes:

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

MG writes:

I am amazed at how we are all spinning our analytical wheels on the issue of inequality. The points made here are real and agreed by most -- an overstated inflation adjustment and exclusion of non-wage compensation items can never lead to an unbiased answer. They have been also made many times before. I hope we can continue to have this general discussion, but not again on the basis of studies that continues to perpetuate errors such as those here noted. Moreover, I really think we must move to focusing mostly on consumption inequality, since bottomline consumption would adjust address these omissions and would better capture additional biases such as differences (usually counterbalancing) in after-tax , after-public and private transfers (from welfare to financial aid).

On a separate note, we should further study the effect of some outcome-adjusted means of consumption that did not presume that because you spend 10x on gourmet meal you have consumed 10x a meal. (At a minimum, we should deflate consumption based on the actual basket of goods each quintile consumes -- if CPI overstates the lower quintile basket's it way understates the uppper one's).

David R. Henderson writes:

@MG,
I’m not sure whom you’re addressing. I wasn’t discussing inequality. I was discussing a factor in real wages that doesn’t show up in the wage data and, by the way, also wouldn’t show up in the consumption data.

MG writes:

David, sorry for the impression, since I appreciate that your blog entry was narrowly focused on the claim/nature of stagnant real wages (which almost always leads to discussions on income inequality trends, etc). Ironically, I was reacting to the fact that economists like you would have to spend time repeatedly reminding people about well known significant distortions that almost always bias the analysis in the direction of stagnant real wages. I think by now the adjustments you mentioned should be taken as a starting point. But if you are not frustrated, may be I should not be either! I also wished wished we could spend more time analyzing other explanatory factors such as the nature of the jobs being compared (safety, as you point out, e.g), inflation rates that adjust for the low earner as well as the high earner consumption baskets, etc.

Ken B writes:
And make no mistake about it, these struggles that I'm talking about, these are not new. For decades now -- decades -- middle-class folks have been squeezed from all sides. See, the cost of things like gas, groceries, tuition -- prices continuing to rise, but people’s paychecks just haven’t kept up.

Michelle Obama, Friday

Mark Brady writes:

"But now, [Ken Judd] pointed out, so many jobs are so much more comfortable: workers in manufacturing plants who have air conditioning, etc. This, he noted, is an increase in real wages."

"Moreover, there's been a secular decline in fatality rates on most jobs. That doesn't get captured in wage data. In fact, all else equal, wages are lower when jobs get safer."

How do improvements in working conditions and the secular decline in fatality rates on most jobs that reduce the disutility of labor count as increases in real wages?

Ken B writes:

@Mark Brady:
I hire you -- the marginal worker -- at $50 an hour to dodge traffic. (I like to watch traffic dodging.) Each day I reduce the wage by $1 until I find the minumum wage I must pay you for the job. We have found the market wage.

Now I let you dodge not traffic but large nerf balls. I bet if I start my wage reductions again I'll get takers. The wage for the nerf job is lower. The difference is surely mostly due to increased safety.

Mark Brady writes:

@Ken B:

If I understand you aright, I don't have a problem with your analysis and examples. However, I suggest it doesn't refute my point.

Strictly speaking, free lunches and health benefits raise the real wage rate (movement along the supply curve of labor). Improvements in working conditions and more breaks reduce the disutility of labor and cause the supply curve of labor to shift to the right (with a consequent fall in the real wage rate). That's how I would teach this to my students.

VangelV writes:

(1) The inflation adjustment used to compare wages over time is the Consumer Price Index. As Michael Boskin has shown, the CPI overstates the increase in the cost of living.

But Boskin is not right. First of all, all of us have a unique basket of goods and experience a different price change from our fellow citizens. Second, there is an incentive to underestimate CPI, which is what Boskin has done. The methodological changes in the 1990s were quite severe and made CPI look lower than it should have been. If we apply that same methodology to the price changes in the 1970s most of the inflation would simply disappear from the BLS reports.

For the life of me I can't figure out why intelligent people tend to have trouble understanding what is really going on. Will you wait until gold goes to $5,000 and silver goes over $100 an ounce before you finally start to look at the issue more carefully? Boskin has a reputation to defend so he will keep pushing the understated CPI message. But we don't and can afford to look at reality as it is rather than as we want it to be.

David R. Henderson writes:

Mark Brady is correct.
What I should have said is that the real wage data understate the improvement in well-being of workers.

Andy writes:

One factor that gets ignored is commute time. It's never counted as part of your hours worked, even though it should be, and since commuting makes people much unhappier than work we should probably consider it even more negatively.

I'm not sure what the trend has been over time. For people driving, there has been a modest increase in travel time. However more people can now telecommute than before. So the net is unclear.

Comments for this entry have been closed
Return to top