Incomes Up, Wages Down

"Census Shows a Modest Rise in U.S. Income" says the headline-writer. Reporter Amy Goodnough has a rather more interesting story:

The nation’s median household income grew modestly in 2006, the Census Bureau reported yesterday, even as the percentage of people without health insurance hit a high.

Experts said the rise in income was mainly a reflection of an increase in the number of family members entering the workplace or working longer hours. Average wages for men and women actually declined for the third consecutive year.

Sounds like a recession is just what the doctor didn't order....

UPDATE: Republicans are really psyched about this declining wages economy:

Some Republicans seized on the new data as evidence that Bush administration policies had been good for people’s pocketbooks. In a statement, President Bush said the news was a sign that Congress should not raise taxes. The data, he said, confirmed “that more of our citizens are doing better in this economy, with continued rising incomes and more Americans pulling themselves out of poverty.”

The poverty thing, seriously, is good news, but note that "the poverty rate fell in 2006 for the first time this decade." So if Bush wants to be judged by the impact of his policies on the poverty rate he, well, sucks.

Comments

The nation’s median household income

Average wages for men and women actually declined

Hmmmmm...

Median =/= average
income =/= wages

Where is the statistic about median wages for men and women? Where is the statistic about income for men and women? Oh, I get it, let's cherry pick something to make things look bad?

Why, oh why, can't we have better economics writers?

Posted by Al | August 29, 2007 1:08 PM

Why, oh why can't we have trolls that don't post irrelevant tangents, non-sequiturs, non-responsive information, talking points culled from elsewhere, or horrifying leaps of logic?

Silly rabbit, because then they wouldn't be trolls.

Posted by Jeffrey Davis | August 29, 2007 1:18 PM

So if Bush wants to be judged by the impact of his policies on the poverty rate he, well, sucks.

Yes, but it's unfair to judge him by this negative statistic, on which his policies had little impact. Republicans can only be judged on the good things that happen, even coinsidentally, while in office. Please shelve your negativity until January 2009, or perhaps January 2013.

Posted by jlw | August 29, 2007 1:28 PM

Uh, median is most certainly not the average.

Posted by Kurzleg | August 29, 2007 1:29 PM

Yes, but it's unfair to judge him by this negative statistic, on which his policies had little impact. Republicans can only be judged on the good things that happen, even coinsidentally, while in office. Please shelve your negativity until January 2009, or perhaps January 2013.

Right. We're currently up to the John Tyler administration in terms of economic accountability.

Posted by Jeffrey Davis | August 29, 2007 1:32 PM

Al has a point here ... there is something rather odd going on with the stats.

However, doesn't it just serve to amplify MY's point?

Anyway, unless the hourly productivity rate is going down, why are people having to work more hours to get the same (or when all is said and done, higher) wages? Either they are producing less per hour or there is some form of deflation going on (i.e. the same amount of goods/services are worth less money -- even though when it comes to me visiting the grocery store, the opposite seems to be the case) ... or those extra goods and services they are producing are being siphoned off somewhere.

I wonder, it's always hard to make these comparisons given changes in product availability, technology changes and what not, but how are people living now compared to a year or two ago?

If people are producing, on average, the same amount of goods/services per hour, yet having to work more time to have the same standard of living (or maybe even one that's a little greater), they are producing more goods/services to exchange for the same weighted number of goods/services they had before: so they are getting the short end of some deal. But the median person cannot be getting the short end of the stick without someone getting the long end, eh?

So who's getting the long end? And is that really a just outcome?

And where are the Bible-thumpers who are so keen to sniff our underpants' drawer: seems to me the Bible cares more about just distribution of wealth (and "justice, justice shall you pursue") than it does about teh hawt sex, doesn't it?

Households may be making, at median, nominally more income than before, but do they really

Posted by DAS | August 29, 2007 1:37 PM

No, average is not the same as median (if you take "average" to equal "mean," total/N). So what? They report different statistics for different things. For household income, they use median; for wages, mean. Chances are they'll rise and fall together unless something really weird is going on (for example, if changes were limited to tails in the distribution). Matt's point -- that the economy isn't doing so hot -- still stands.

Posted by Matt Stevens | August 29, 2007 1:54 PM

Interestingly, I took a look at the document I guess the Times writer looked at, and it seems to me that the statistic cited is "median earnings of full time year round workers", not average wages. I don't know why it is that economics writers can't seem to grasp the difference between median and average. Are they all idiots?

In any case, here are the years during the last two andministrations in which median earnings decreased relative to the year before:

Men: 2006, 2005, 2004, 2001, 2000, 1996, 1995, 1994, 1993

Women: 2006, 2005, 2004, 2003, 1999, 1995, 1994, 1993

Now, someone please explain me why men's and women's median wages were decreasing in 1993, 1994, and 1995? Were they years leading up to a recession, as Matthew implies for now?

And why women's median wages decreased in 1999 and men's median wages decreased in 2000? Weren't those boom years?

Posted by Al | August 29, 2007 1:56 PM

Just to add to what Kurzleg said,..., average, median, mode, standard deviation,..., it's fairly easy to look them up so you understand what they mean (start with wikipedia.org, unless you're worried about the liberal bias, ... if so, check out your local library, unless it's a haven for commies,...).

Do anyone know if the poverty statistic is actually "real"? I don't know the source and I'm not saying it isn't "real", but the conservatives have a habit of redefining the criteria to make it look better. "You know, if we move some money over here..., change this word to 'can' instead of 'can't',..., and then we lower the income level that defines poverty from $A to$B, then poof, poverty rates go down by naught point eight percent..." Just wonderin'.

Posted by rihilism | August 29, 2007 2:12 PM

Bush said this data is a reason that we should not raise taxes. Has Bush ever stated any circumstance in which raising taxes would be a good idea? If the answer is no, why are we even citing such statements in the newspaper, as they have no credible validity?

Posted by Martin | August 29, 2007 2:16 PM

Al,

Matthew is not implying a recession; he's saying that the economy isn't so hot. Noting how Matthew mentioned in this post that recent increaeses in household income are a function more of rising working hours than rising wages, I think Matthew is implying that most Americans are not receiving the spoils of prosperity that our economy has generated in recent years. Considering that this a theme of many previous Matthew posts, it's not very logical for you to conclude that Matthew is talking about a recession.

Why did median wages for men & women decrease from 1993-1995? If you remember, Al, the recession from 1990-92 prompted structural changes in the US economy. Employment in the manufacturing sector fell dramatically, and most jobs lost in that sector were never replaced. Since jobs in the manufacturing sector tended to be heavily unionized, these lost jobs were high-wage ones as well. This restructuring would explain median earnings for both men & women would decrease.

Structural changes in the manufacturing sector (including a massive investment in information technology and automation) will probably explain the decrease for men in 2000 also, as well as the fact the manufacturing sector began experiencing a recession as early as 1999. (I distinctly remember that the decline in the manufacturing sector was an issue mentioned in the 2000 presidential campaign.)

I'm not sure why women saw their wages decrease in 1999. Could this have been prompted by the increase of former welfare recipients into the workforce in the wake of welfare reform? Did welfare-to-work create a massive influx by women into low-paying service sector jobs?

Posted by eltoro70 | August 29, 2007 2:36 PM

First thing, i'll stipulate that it's folly to read too much into year-on-year changes in figures like these. It's the trend that counts, and the trend since 2001 has been one of pretty good growth in output, minus weak growth in employment and wages, equals an unparalleled share of GDP going to retained corporate earnings & fairly dramatic concentration of the modest personal income growth that was left over in the hands of people at the very top of the income scale.

But note a couple things about the poverty figures. First, the decline in the poverty RATE was small but significant; the decline in the NUMBER of people in poverty was miniscule and didn't reach the level of significance. Second, the decline in the rate of poverty from 2005 to 2006 seems to have been focused almost entirely on the elderly, especially those who are female and living alone. There was no significant decrease in poverty among children or working-age adults.

For both of those reasosn, it's iffy to speak of people "pulling themselves out of poverty" this past year. Because net/net, it's quite possible that nobody really did.

Pretty weak results from 2006, the fifth year of an expansion, and the best year to date for the purported "Bush boom" in terms of labor market & output aggregates. That's right, these figures are from what looks to be the peak year of this cycle. 2007's report will probably be even worse! So pardon me for my

Posted by Tom | August 29, 2007 2:54 PM

if you're only going to look at one stat, real median household income is the one to look at, but no single stat is dispositive (the distinction, btw, between "income" and "wages" is that "income" includes things like cap gains and dividends; for most households, those are somewhere between de minimis and nonexistent).

that said, better an improved median household income than not, regardless of its source, but the underlying trends (not all of which are even george bush's fault!) over the bush years remain the same: weak job growth; continuation of the pattern eltoro70 describes of "higher" wage jobs being lost and "lower" wage (and often, benefit-free) jobs being created; rise of household debt as households desperately try to maintain living standards in the face of shoddy income growth; increased concentration of income in the upper 2%, 1%, and most importantly, .1% of households; tremendous returns to capital and limited returns to labor; huge rise in the national debt (meaning a deferred tax hike); and, of course, the continued decline in the number of people with health insurance as our job-based system gradually breaks down.

and i've gotten this far without mentioning how increased house values and associated MEW (morgage equity withdrawal) have further disguised the problem: indeed, we have the lowest level of household equity on record, and i'm sure i don't need to point out that number of ARMs waiting to reset in the coming 12-24 months and the problems that will cause in light of all of the above. (as a side note, with the decline in MEW as standards tighten and house values fall, we're seeing a rise in credit-card defaults:

http://calculatedrisk.blogspot.com/2007/08/credit-card-defaults-on-rise.html)

and, of course, bush wants to borrow more from the chinese central bank for iraq. sounds like a great time for a tax cut!

PS. Martin, no, there are no circumstances under which george bush thinks a tax hike is justified: he's a free lunch man 24-7.

Posted by howard | August 29, 2007 3:07 PM

btw, speaking of poverty, i just happened to notice two great posts over at prof delong's site about global poverty in 1900 and american living standards in 1900, which provide some interesting perspective:

http://delong.typepad.com/sdj/2007/08/slouching-tow-1.html (global)

http://delong.typepad.com/sdj/2007/08/hoisted-from-ar.html (american)

Posted by howard | August 29, 2007 3:16 PM

"Uh, median is most certainly not the average."-Posted by Kurzleg

This is not precisely true. While what we usually consider to be the definition of average is the arithmetic mean, it is not so. Average is an imprecise term that can be used for any method of coming up with a typical value. Median is perfectly valid as a value for what is average. It may confuse some people who do not know the real meaning of the word "average", but it is accurate (if not precise). Mean, mode and median are all averages.

Posted by Njorl | August 29, 2007 4:10 PM

"Bush said this data is a reason that we should not raise taxes. Has Bush ever stated any circumstance in which raising taxes would be a good idea?"

I believe he thinks it would be warranted to fund efforts to combat a plague of porcine aviation.

Posted by Njorl | August 29, 2007 4:14 PM

The poverty rate in 2006 was lower than in all but the final two years of the Clinton administration. So, really, get a bit of perspective.

Now, on the other hand, if you really thought 1998 was a crappy year, it would be good to know that.

Tom focuses, interestingly enough, on the increased population in the country. I'm not sure why numbers rather than rates would be interesting to anyone in this context. But howard makes the same sort of error, in reverse: the number of people with health insurance actually increased last year, as did the number of people without.

Posted by Thomas | August 29, 2007 4:26 PM

Thomas, wrt to insurance, you're right, i phrased it poorly (and, frankly, i'd have to dig in too much to verify if the number of people with health insurance actually did increase, since there are roughly 2.2M more people without health insurance this year than last, which, iirc, is roughly the annual population increase), but if we turn to proportions, as you want, we discover:

The number of people without health insurance coverage rose from 44.8 million (15.3 percent) in 2005 to 47 million (15.8 percent) in 2006.

http://www.census.gov/Press-Release/www/releases/archives/income_wealth/010583.html

as for poverty, i think the correct perspective is exactly the one that has been previously cited here: after years of improvement in the proportion of people regarded as living in poverty, the bush years saw a reversal. the fact that last year was a marginal improvement doesn't change the bottom line: a higher proportion of americans now live in poverty than did when bush took office.

i mean, hell, the proportion of people living in poverty is way lower now than it was 40 years ago: does that mean we shouldn't care any more?

Posted by howard | August 29, 2007 4:38 PM

actually, thomas, i realized that there was an easy way in to approximate; if 44.8M constituted 15.3% of the 2005 population, then the 2005 population, in round numbes, was 292M.

and if 47M constitutes 15.8% of the 2006 population, then the 2006 population, in round numbers, was 297M.

so you're right: the number of people with health insurance is up, even though the proportion is down (although i'm amazed at a population growth of nearly 5M, which doesn't seem right, so perhaps there's a further twist here that i'm not seeing right now).

Posted by howard | August 29, 2007 4:48 PM

eltoro70:

Matthew is not implying a recession; he's saying that the economy isn't so hot.

I dunno; Matthew mentioned a recession in the post. I can't exactly figure out what Matthew was saying in that sentence, but it looked to me as though he was predicting a recession.

Why did median wages for men & women decrease from 1993-1995? If you remember, Al, the recession from 1990-92 prompted structural changes in the US economy.

It very well may be true that '90-'91 recession "prompted structural changes in the US economy". Same may be true about the '01 recession. My only point is that, if we are going to judge the economy on whether median wages of men and women decreased, then we can simply say that the economy now is as bad as it was back in the halcyon days of 1995. And I'm sure we can all remember how awful the economy was then, right?

Posted by Al | August 29, 2007 5:02 PM

i agree, al, that matthew was very unclear in his final sentence, but jeez, i'm predicting a recession. big deal: amazingly enough, every business cycle we've ever had in america has ended with a recession, so there's nothing special about prediciting one.

as the old wall street saw goes, i can tell you what's going to happen, but not when, or when something's going to happen, but not what. in this case, the what is a recession; the when is what's uncertain.

but matthew's deeper point isn't the issue of when you can find a comparable moment when some aspect of median wages by gender did something comparable to something current. as i noted earlier, no single economic stat is dispositive (other, i suppose, than that 2 consecutive quarters of negative gdp means a recession), and it's not like all we know about 1995 is something about median wages.

we know all kinds of things, and based on knowing all kinds of things, the economy was looking much better in 1995 than it does today.

which takes us back to matthew's point: we've had a 23 quarters into an expansion, which would bring us to more like some time in 1998, and we'd had lots of gains all across the economy in that span of time. after this 23 quarters of expansion, not so much....

Posted by howard | August 29, 2007 5:53 PM

btw, al, here's a nice chart just posted over at angry bear about male and female median wages over time:

http://angrybear.blogspot.com/2007/08/spencer-on-real-median-income-for-male.html

Posted by howard | August 29, 2007 6:23 PM

er, i said "wages" but i meant to say "income." so easy to slip up!

Posted by howard | August 29, 2007 6:25 PM

howard, the twist you're missing might be illegal immigration, which drives both the population number and the uninsured number.

The poverty rate in 1998 was 12.7%. The "gains all across the economy" hadn't reduced the poverty rate at that point to the level reached in 2006 (12.3%). The poverty rate in 1995, a year you remember fondly, was 13.8%. The highest level--highest--in the Bush years was 12.7%.

Posted by Thomas | August 29, 2007 6:26 PM

Thomas, i'm not sure really what your point is.

here's what i am sure of: the rate in poverty has, over the last quarter century, correlated pretty strongly to the economy itself: it tends to fall during expansions and rise during recessions.

so, for example, the rate hit a high in the early '80s, then fell consistently until the early '90s recession, when it went back up, and then started to fall again in '93 and continued to fall to a low in the late '90s, then went back up as we moved into recession and...

continued going up until now, despite 23 quarters of expansion. so if you want to applaud he absolute level of poverty, go right ahead, but it doesn't change the underlying dynamic i outlined.

after all, the poverty rate when clinton took office was (roughly, i'm just eyeballing off the census bureau graph that i'll link to in a moment, although if you've got a link to the hard numbers, i'd appreciate seeing it) 15%; the poverty rate when clinton left office was roughly 11.5%. that number then went as high as 12.6% before falling back to 12.3% last year.

if you want to see that as a positive for the economy of the 21st century, i can't stop you, but i'm not going to agree with you, either....

Posted by howard | August 29, 2007 6:42 PM

oh, right, the link:

ttp://www.census.gov/prod/2007pubs/p60-233.pdf

PS. i don't have the time to root around in the census material, but are you sure the census bureau includes illegal immigrants in its numbers?

Posted by howard | August 29, 2007 6:43 PM

er, i said "wages" but i meant to say "income." so easy to slip up!

Ah, but "income is up, wages are down" is exactly the point we're talking about!

In any case, if you want to know why median male and female wages are down, I'll tell you. It is because that is a misleading description of the relevant statistic. As I posted above, the census statistic we are discussing is: "median earnings of full time year round workers". The only counts the earnings of workers who are employed on a full time basis, year round. In other words, it excludes persons who are employed part time, and people who are unemployed for all or part of the year. New full time year round jobs for those who previously didn't have access to full time year round jobs are likely to be at lower than median wages - employment is already very high among those who have the capability to earn above-median wages. This then brings median earnings among full time year round workers down.

In other words, lets take a universe of 5 people. In year one, three people have full time year round jobs, making $70k, $50k, and $30k; the other two people are unemployed. The median earnings of full time, year round workers is $50k, as that is the median of the people who actually have jobs. In year two, the two previously unemployed people get jobs for $25k each, but the other three people have the exact same jobs at the exact same salaries. Hey, guess what, the median earnings of full time, year round workers dropped from $50k to $30k! Even what really happened is that earnings were the same for the people who previously had jobs but the economy was able to add two new jobs. The seems to me to be good news, not bad, even though the statistic you are looking at went down.

If you look at the numbers of people in the US who fit the relevant category from year to year, that seems to me to be what has happened here - the economy is adding a bunch of new jobs and is able to employ previously unemployable people, even though at lower wages.

Posted by Al | August 29, 2007 7:03 PM

Poverty rates have recently tended to lag recessions and expansions. The early 90s recession began in 1990 and ended March 1991 (according to NBER). However, poverty rates in 1991, 1992, and 1993, only moving down beginning in 1994. The peak in poverty rate came in 1993, at 15.1%. The most recent recession began in March 2001 and ended in November 2001. The peak poverty rate came in 2004, at 12.7%. So the timing is similar, but the rate is considerably lower, compared to the early 1990s experience.

Your assertion that the rate "continued going up" until now is flat wrong. The rate declined last year as well, though not by a statistically significant amount.

I think you've simply misread the evidence here. As for the meaning of it, I think you draw the wrong lessons as well. It is true that we care about trend lines, but we also care about absolute levels. We want unemployment rates to fall, for example, but we look at not just the trend line (is unemployment rate rising or falling), but also at the absolute level (is the rate low relatively speaking?). I think the same is true with the poverty rate. Given that the trend experience is similar to the early 1990s and the absolute level is lower than all but 1999 and 2000 out of the last 20 years, I'm drawing different conclusions from those you draw.

Posted by Thomas | August 29, 2007 9:19 PM

Al, in terms of "wages" and "income," as i noted, for most households, the distinction is de minimis; i just hate to get it wrong through inadvertent typing.

as for your insight, it would be a wow were it not for the fact that it's been true every year. for it to make sense, you'd have to show us that there was something in the composition of jobs this year was that was different than last year to account for the difference you seem to care so much about.

and you need a better model; your universe has much too great a rate of change (that is, there's something like, what is it, 120M americans with jobs. a couple million or so leave the labor force each year; a couple million or so join the labor force each year; another several million get laid off and another bunch change jobs voluntarily, but put it all together and it's still a relatively minor number of jobholders when we're looking at median incomes).

Thomas, i said that poverty correlated pretty strongly to the economy; i didn't say there was a day-to-day relationship. of course there are lags.

as for 2005, well look, i told you that i've been eyeballing the chart, and that chart, plus the press release (which stated that 2006 was the first drop in poverty level) suggested to me that the poverty rate was still going up in 2005. if it levelled off, so be it.

because the real point (and thomas, i'll ask you again: can you provide a link to your data source? i've hunted around, and the best i can find is poverty data broken by sex, which for my purposes will do fine, but i'd like to see the numbers you're using just to file away for future discussions of this ilk).

according to the census bureau poverty of people by sex, 1966 - 2006 chart (http://www.census.gov/hhes/www/poverty/histpov/hstpov7.html), when clinton took office, the male poverty rate was 12.9%; when he left office it was 9.9%, a 23% improvement; the female poverty rate was 16.6 and ended at 12.6, a 24% improvement.

So under Bush, we've moved from a male rate of 9.9% to a male rate of 11%, a 10% worsening; the female rate from 12.6% to 13.6%, an 8% worsening.

and you want us to give more credit to the "worsening" than to the "improving," which is certainly not a conclusion that i would draw.

and the issue at hand was not, in fact, the absolute level of poverty (see matthew's update); it was whether bush should be boasting about the improvement in the poverty rate now when his overall record is one of worsening. so this is the precise metric we should be reviewing.

the discussion of absolute level is for another time.

Posted by howard | August 29, 2007 11:08 PM

'Ah, but "income is up, wages are down" is exactly the point we're talking about!'

No it isn't.

Household income is up, personal income is down. More precisely, income per earner is down, per capita income is up, because a higher percentage of people are earning. Wages were not mentioned in the latest press release. Wage information will probably come out later. I could find wage information no more recent than 2005.

Posted by Njorl | August 30, 2007 2:10 PM