This time the LA Times gets into the mix, claiming that 60 million Americans live on less than $7 a day. Like many awful statistics, this was generated by a sort of Rube Goldberg machine; the LA Times got the figure from a guy with a website, who got it from the World Socialist Workers Website, which took it from an article on reported taxable income in the New York Times. In this case, though, I don't see how this could have happened. The Census bureau is the obvious place that you go to see how many Americans are living on how little income. Random people with websites are not. Not that the Times article is much better; it compares reported taxable income to the poverty line, even though it grudgingly concedes that benefits aren't included, and even though we know that the poor tend to strongly underreport income because reporting income means losing their EITC and health benefits.
« Suckers for a good story | Main | Oh, Dr. Boli, I love you »
Department of awful statistics
15 Jan 2008 06:34 am
TrackBack
TrackBack URL for this entry:
http://meganmcardle.theatlantic.com/mt/mt-tb.cgi/17926
Comments (80)
That Times stat is pretty crazy, though. It does say that IRS figures indicate there was only $7 a day of reported income for the bottom 60 million people, including taxpayers and their dependents. How exactly does that happen? I mean, I know a lot of self-employed people manage to report very low net incomes that don't reflect what they'd be reporting if they had a job at somebody else's company. And presumably people at the bottom end are working for cash a lot of the time, and not reporting it. But does it really get you down that low? Apparently, I guess.
This is just the most glaring of many examples of media people screwing around with statistics to support their claim that America is a sea of inequity--that the rich are getting richer and the rest are getting left behind.
The problem is that the claim is not true. So they have to do some impressive gymnatstics to make it look true. Paul Krugman is nototrious for this sort of thing. So he uses household income statistics without noting that households have gotten smaller and uses median wages, eschewing averages and ignoring benefits.
Krugman & co. also choose to ignore the fact that the CPI does not take sufficient account of quality improvements, so that it overstates inflation--making the claim of "stagnating real wages" just false.
Plenty of commenters here have been guilty of the same sort of sophistry. I won't mention any names, but one (with dazzling SAT scores) has commented above.
OK, I followed the whole 'Soldiers Committing Crimes' with a jaundiced eye because of my more than 20 years of soldiering (with multiple deployments in not so pleasant places) I understood why we might want to know why some soldiers committed crimes they might not have otherwise.
But. This is beyond the pale (from Pajamas Media):
Lisa Richardson of the Los Angeles Times did not return calls. David Cay Johnston of the New York Times did not return email. William Shanley later called me back and explained, “No, I did not look at the U.S. Census Bureau figures. I used the World Socialist Web Site as my source. They source the New York Times. Certainly, no one questions this as an official source.” He later called back to amend his statement: “You are right. In my piece, I said it was [information from] the U.S. Census Bureau. It was from the IRS.”,
Obviously Shanley is a hack, but the NYT and LAT? Christ...
It is absolutely true there is a huge gap between the MSM and people serving in the military. I believe that; else the glaring errors when referring to uniformed officials and such basic things as rank and position wouldn't be so common.
But these errors of basic math twisted in order to form a pre-conceived bottomline make me wish for the continued decline of these newspapers. A pity. Perhaps these bottom 60 million included my dependent son whose effective income was zero? In any event, the ignorance driving such conclusions suggests that the days of Jayson Blair are not yet over.
Though the Johnston article doesn't offer any help whatsoever in figuring this out, the IRS statistic being used is AGI, not taxable income. Tom Maguire and his commenters pick over the 2007 version (2005 data) of this annually recurring article here.
The averages for the lowest quintile are skewed by the presence of nearly 2 million filers who report an average income of close to negative $50,000-- mainly capital losses, I'm guessing.
Johnston's saving grace is that he appears to be an assiduous ego-googler with a willingness to turn up on blogs and defend his work. You have a good chance of getting some answers.
Paul, it couldn't be capital losses included in AGI...capital losses are deductible only to the extent of capital gains, plus another $3,000 per year. More than likely, negative AGI's of that magnitude stem from business operations losses - sole proprietorships and pass-through losses from S-corps, partnerships, etc.
the poor tend to strongly underreport income because reporting income means losing their EITC and health benefits.
And don't forget that the appeal of working a cash pay job, like waiter or bartender, is that your official reported wage is tiny. I'm sure there's one or two waiters out there logging exactly how much they made in tips and filing proper returns on it every year, but I've never met one of them.
So yeah, "income reported to the IRS" is not really a great benchmark for someone's standard of living.
Even the New York Times would have trouble competing with this LA Times editorial for absurdity:
"...our healthcare system leaves millions without access to doctors."
That's just not true. There are many without insurace, but they do have access to doctors.
rwe
What's so bad about using median wages? Using averages can result in a massive skewing of data. I mean, neither are perfect, and both have their benefits in certain areas, but I think that the median is more useful here.
As you get further and further up into the upper classes, there are fewer instances of salaries, and they have a much larger variance between them (lots of people make 50,000 a year, I'm not so sure how many make 1,000,000 a year, or even 250,000 a year). As a result, those numbers can begin to look like outliers, skewing your "average" data. However, your median data isn't influenced by outliers, giving what statisticans may sometimes view as a truer value.
Now, I'll admit that one statistic to describe ALL of our income data is not perfect. I personally think it would be more useful to use the median values in the lower, middle, upper-middle, and upper classes. It means a few more data points, but a much clearer view of each class, without allowing outliers (either low or high) to skew either the statistic for all of the income data, or even each classes statistic.
Could well be, Michael; it's been a few years since I did any harvesting, and my recollection was that the entire loss gets reported in the income section and the limitation gets taken care of in Schedule D. Memory's probably playing me false again.
This is an actual (loud) conversation I overheard on the bus recently:
"I've been a musician my whole life. It's a cash business, you know. I went to the Social Security Office and the lady said, Mr., you've never reported any taxes. You'll be getting $85 a month."
"What's so bad about using median wages?"-stephanie
Stephanie, medians are useful but so are averages. The problem is when people ignore averages and look only at medians and use that to characterize the distribution. A simple numercial example shows the fallacy. If there are 5 people, two with incomes of 10,000 one with an income of 20,000 and two with incomes of 30,000, the average and median are both 20,000. Then suppose the bottom two see their incomes rise to 19,000 in five years, the top two see incomes rise to 39,000 and the middle one sees no change. The median has not changed, but the average has changed dramatically (rising by 36%). Clearly the average here is telling us something important, but Paul Krugman could come along and say "hey, look, median incomes are stagnating."
Also, as I explained above, another problem with using median wages is that that ignores benefits. If my wage stays the same, but my medical benefits improve substantially, I'm certainly better off. So it's just sophistical to ignore benefits.
***
Really Megan McArdle needs to do something about her spam filters. I had more criticisms of that Times editorial, but they have been "received and held for approval by the blog owner." That sounds like something out of 1984. "The blog owner is watching you."
Well, let's see ...
-Count all the dependents under 16, and most over that up to ~23. They're not making any money, and $0/day
rwe -
Sorry - I wasn't defending Krugman's use of ONLY median statistics (I'm a stat student, not a Krugman pal), I was questioning your seeming criticism of using medians at all.
I definitely agree that you should check both, especially when ONE appears to be stagnating (and I didn't say anything at all about benefits, but it seems to me that at least between the very upper class and the very lower classes, which both influence the average more than the median, that benefits - in dollars - would look sort of similar - it feels like it makes sense that someone in a lower class has a lower dollar value of benefits - but I don't know, so I didn't mention it).
Also, though, let's take your example, but expand it so that you have 2 people with an income of 10,000, 1 person with an income of 75,000, and 2 people with an income of 140,000 a year.
Again, both the median and the average are 75,000. But, let's say then the lower end sees an increase to 20,000, and those at 140,000 see an increase of...well, whatever you want. Sure, the median will look like it's stagnating, while the average will go up.
But, if you have a large enough sample size, wouldn't it be possible to prove that, indeed, the middle class wages ARE stagnating (and averages would seemingly cover that up)? Comparing the difference of middle to low, and then high to middle before and after the "raises" would allow you to do that, right? And to compare the "middle" of the middle class to the "middle" of the lower class, etc, wouldn't it be more useful to use the median so that the lowest incomes in the lower class and the highest incomes in the lower class don't affect the "middle" so much?
I guess I'm making a pretty big assumption that incomes (without benefits) in a particular class would behave sort of loosely normal-ish, but probably not perfectly, and I'm sure with some tails at the end of the distribution. (This must be the part where I start sounding like an idiot.) In which case, you'd maybe want to function on the stuff in the middle, and leave the ends out, which would point you to using the median, no?
I've known 5 people who collected unemployment. Every single one of them worked side jobs or under the table at the same time.
I can't imagine it's any different for many unemployed workers who feel they are going to work the system after "getting screwed by their employer and the government for so long".
I've known 5 people who collected unemployment. Every single one of them worked side jobs or under the table at the same time.
A friend quit school, moved to NYC, worked in restaurants for a few years, and then returned to WI to finish school, while collecting unemployment from NY (don't sure if he was employed tho). Of course he was one of those end-welfare types, but at least he was fully aware of the irony.
If there are 5 people, two with incomes of 10,000 one with an income of 20,000 and two with incomes of 30,000, the average and median are both 20,000. Then suppose the bottom two see their incomes rise to 19,000 in five years, the top two see incomes rise to 39,000 and the middle one sees no change. The median has not changed, but the average has changed dramatically (rising by 36%). Clearly the average here is telling us something important, but Paul Krugman could come along and say "hey, look, median incomes are stagnating."
Good point there, rwe! I'm sure the reason median incomes have been stagnant for the past 6 years is that one guy who happens to keep earning the exact median income of all 300 million Americans, and whose income never seems to change. If only that one guy would get off his ass and earn some extra bucks! Sure is weird how nobody who started out below him seems to have surpassed his income yet, though. You'd think with tens of millions of wage earners, it'd be impossible for something like that to happen, wouldn't you?
Accusations of "innumeracy" are getting flung around here a lot lately. They don't seem to be getting flung in the right directions, though.
Stephanie, your post is interesting. I agree that a careful and honest researcher can figure out what's really going on, using medians, averages, gini coefficients and every other relevant statstics to undersand the distribution.
And many discussions (including Krugman's) ignore a crucial fact: that there is substantial income mobility. The people the bottom of the income distribution in 1996 are not generally the same ones who are there now. So, as economist Thomas Sowell reported:
Anyone who follows the media has probably heard many times that the rich are getting richer, the poor are getting poorer, and incomes of the population in general are stagnating. Moreover, those who say such things can produce many statistics, including data from the Census Bureau, which seem to indicate that.On the other hand, income tax data recently released by the Internal Revenue Service seem to show the exact opposite: People in the bottom fifth of income-tax filers in 1996 had their incomes increase by 91 percent by 2005.-Tom Sowell
Sowell shows the fallacy here of ignoring income mobility. When you actually follow what happened to the people at the bottom, you find that their incomes have risen significantly.
But, if you have a large enough sample size, wouldn't it be possible to prove that, indeed, the middle class wages ARE stagnating (and averages would seemingly cover that up)? Comparing the difference of middle to low, and then high to middle before and after the "raises" would allow you to do that, right? And to compare the "middle" of the middle class to the "middle" of the lower class, etc, wouldn't it be more useful to use the median so that the lowest incomes in the lower class and the highest incomes in the lower class don't affect the "middle" so much?
Exactly right. What you're talking about is skew, and is quite common in descriptive statistics:
http://www.cvgs.k12.va.us/DIGSTATS/main/descriptv/d_skewd.html
I teach statistics myself on occasion, and am very interested in the teaching of the subject (it really should be offered in high school as a requirement; in fact I'd place it over algebra), so here's another piece which is more informative:
http://www.amstat.org/publications/jse/v13n2/vonhippel.html
Oh, and btw, rwe's example is not one that occurs that frequently, and when it does it's usually taken to be an indicator that something is seriously amiss. It's in a class of what are called bimodal distributions and is usually the result of mixing two or more distinct populations.
-Count all the dependents under 16, and most over that up to ~23. They're not making any money, and $0/day
That's not it, unless the journalist is simply lying, which makes this a pointless exercise. The article says the 60 million people are the bottom quintile of tax filers and their dependents, not everyone else's dependents.
As for the rest of your cites: you have to pay taxes on unemployment benefits, so there's no reason not to count people getting unemployment; anyway benefits are set to 50% of your last job, which should be a lot more than $7 a day. There's no reason not to include "contract workers between contracts" either, especially if they were "between contracts" for a whole year. The figure obviously doesn't count illegal immigrants unless they're taxpayers, in which case why shouldn't they be counted? And obviously one wouldn't want to exclude "the homeless" since they're among the poor, which is exactly the population we're supposedly trying to count; though in fact one imagines they're undercounted in this stat since few homeless people file tax returns.
Since we're discussing statistical fallacies, in the post of mine that was caught up in the spam filters, I was taking issue with the LA Times assertion that poverty has been spreading.
As we contemplate the coming election, we seek a president who understands that the unchecked spread of poverty is not a natural condition but rather a failure of government, and that action is required.
I don't know how they arrived at that conclusion, but it is a common refrain in the media. So I checked the facts. According to the Census Bureau, the poverty rate for all families was 9.8% in 2006, 11.0% in 1996 and 10.9% in 1986. How in the world, then, could anyone conclude that poverty has been spreading unchecked?
I don't know, but as my pal brooksfoe's comments make clear, people don't like to let facts get in the way of a narrative they find appealing (the rich are getting richer etc...). I wouldn't accuse him of "innumeracy" mind you (don't forget his 1600 SAT score), but rather of confirmation bias.
people don't like to let facts get in the way of a narrative they find appealing (the rich are getting richer etc...)
I'd like to see you try to argue that the rich aren't getting richer, honeybunch.
Sowell is a hack. Anyone who has the slightest bit of familiarity with the punditry knows this. Further, median wages have been stagnant for a long time, and that's another fact everyone knows. Finally:
When you actually follow what happened to the people at the bottom, you find that their incomes have risen significantly.
You mean those 'studies' that include college kids, which find that, ten or twenty years after they graduate, they have moved into a higher income bracket from when they were in school? Here's a relatively informed discussion:
http://www.huppi.com/kangaroo/L-mobility.htm
And a quote:
But it gets worse. Instead of comparing the group’s incomes to each other, the study compared them to the nation at large. This introduces a second bias: the fact that incomes rise with age. This was especially important for this study, since the taxpayers it identified in the bottom quintile in 1979 had a median age of only twenty-two. Labor economist Kevin Murphy explains the study’s results this way: "This isn't your classic income mobility. This is the guy who works in the college bookstore and has a real job by his early thirties." (6)
And people wonder why I dismiss these people as nuts and cranks (I usually go with dishonest as a second or third bet; fifteen years ago, I would have had dishonesty reversed with incompetence in terms of egregiousness. What a difference kids make!)
Brooksfoe, my friend, its the "etc..." part I was doubting, namely that "the poor are getting poorer."
I think the rest of the statement (replaced by ... ) is that the rest are getting left behind (as was stated above).
The rich may be getting richer, but so are the poor, apparently.
EI
rwe -
"People in the bottom fifth of income-tax filers in 1996 had their incomes increase by 91 percent by 2005.-Sowell"
So that would mean, taken at face value the way you present it, anyway, that the lower class should be getting smaller right? I mean, logically, if the bottom 20% saw an increase of 91%, either people are moving from the lower class to the lower-middle class, or, at the VERY least, those in the lowest 20% are closer to the next 20% (so there is less variance between the lower class incomes).
Somehow, that seems illogical to me - when someone wins, generally, someone else loses (even if it's not an equivalent loss). So whether that means that someone ELSE in the second fifth from the bottom of the income range LOST their job (offsetting the statistical gains of the person whose income rose 91%), or something else, income mobility works in the opposite way than what you're mentioning also.
I don't think anyone would ever argue that there is NO income mobility, but rather, that income mobility has some kind of equilibrium. Not everyone can be in the middle class or upper class, so when people move up, chances are other people move down.
Studying whether or not the middle class incomes are stagnating says nothing about whether a specific person tripled his income by changing jobs or whatever. What it is saying more, I think is:
If Job A is in the lower class, Job B in the middle class, and Job C in the upper class, over time, what $$ is offered for those jobs. Is it increasing, decreasing, or stagnant?
It does NOT take into account whether a specific pool of individuals was able to improve their circumstances. Think of it as looking at jobs, not people.
Sowell is talking about a group of people. Those talking about stagnating wages are not necessarily talking about a specific person - they're referring to what a job is worth in the marketplace, and what they're saying is that a middle class job worth is not as positively dynamic as a higher class job worth.
Just because the same people happen to be in those jobs doesn't make the wage statistic about people and their situations.
I think both statistics are important, but they are different, and don't necessarily support each other.
We'll stop treating movement between income quantiles as class mobility as soon as you guys stop treating income quantiles as classes. How's that?
Johnston does deserve some credit for joining the comments thread over there- I don't think he was googling himself, since Annie Jacobson emailed him to get his response.
But the statistic is utterly meaningless- the most glaring problems are that it doesn't include Social Security income and does include college and high school students working part-time jobs.
I suspect the biggest portion of the statistic is Social Security recipients. Indeed, there are 30 million retired workers alone on Social Security, plus another 6 million on disability, and 6.5 million who are family members of deceased retired workers, all of whom average around $1000 per month in benefits that are for the most part not taxable. So right off the bat, probably around 35-40 million of his 60 million are actually receiving between $25 and $40 a day in "earned" benefits that are not included in AGI.
Then of course you have the 15 million or so college students, of which a fair percentage work part-time summer jobs, and perhaps a few million more high school students with part-time jobs. Being pretty conservative, that probably gives you another 5 million out of the 60 million figure.
Then of course you have your 2 million filers that reported an average income of around negative $50,000 as Paul Zrimsek pointed out, most of whom are probably quite well off, but recorded losses on the year for various reasons (capital losses, business start-up costs, high college tuition expenses, retired divorcees with high alimony payments, etc.). But I don't want to rely too much on this number- I'm not certain it is entirely individual income tax returns (some portion would seem to be corporate returns, which Johnston didn't include).
When you add in dependent children for all of these groups, especially for filers receiving disability payments, suddenly you've explained about 90-95% of the 60 million figure he cites. Very few of whom you'd call "poor" and none of whom are "living on $7 a day."
...And that's before you even get to welfare beneficiaries.
Point is that almost every person included in the "$7 a day" statistic is someone who is not only unemployed or underemployed, but is also someone who is not even seeking full employment.
The rich may be getting richer, but so are the poor, apparently.
Leaving aside whether that is true or not, the argument is that the rich are getting richer MUCH faster than the poor, and that that is in fact a problem which needs to be addressed.
We'll stop treating movement between income quantiles as class mobility as soon as you guys stop treating income quantiles as classes. How's that?
How about you simply stop treating the movement of a specific individual over time as class mobility and talk about groups of people instead?
Like people in the real world do. Or is that too much to ask?
Sowell is talking about a group of people. Those talking about stagnating wages are not necessarily talking about a specific person - they're referring to what a job is worth in the marketplace, and what they're saying is that a middle class job worth is not as positively dynamic as a higher class job worth.Just because the same people happen to be in those jobs doesn't make the wage statistic about people and their situations.
I think both statistics are important, but they are different, and don't necessarily support each other.
This is exactly right. Most people start out their economic life in bottom quintile, no matter how wealthy their family happens to be. So if Joe Doaks makes 5K/yr working part-time in college, and by the age of 35 is making $40K/yr, he's demonstrating 'class mobility'. Even if his parents were making $60K/yr at the age of 35.
Sowell is talking about a group of people. Those talking about stagnating wages are not necessarily talking about a specific person - they're referring to what a job is worth in the marketplace, and what they're saying is that a middle class job worth is not as positively dynamic as a higher class job worth.-stephanie
Stephanie, if I understand you correctly, you're pointing out that income inequality has been on the rise. I don't deny that. It has been rising steadily since 1950. There is a lot of economic debate about why that has happened--the most common view is that technological changes have increased the rewards for education.
I just object to people muddying the distinction between absolute and relative measures of well-being and pretending that because those at the bottom are relatively worse off than they used to be, they are also absolutely worse off. A link I gave above shows that the poverty rate was 18.5% in 1959 (the first year it was measured) and only 9.8% by 2006. So, while inequality rose in that time, the percentage living in poverty fell.
I have some work to just now, but I'll catch up with the discussion later. It has been interesting, including the criticisms of my posts. SoV should calm down though.
To sum up: income inequality has risen in recent decades, income mobility has stayed about the same, and people generally are better off than they were.
In fact, here is a table that demonstrates the lessening of mobility:
Percentages of families making transitions to and from middle class
(5-year period before and after 1980) (2)Before After
Transition 1980 1980
----------------------------------------------
Middle income to low income 8.5% 9.8%
Middle income to high income 5.8 6.8
Low income to middle income 35.1 24.6
High income to middle income 30.8 27.6
Now we're all agreed that an 18-year-old making $5000 shouldn't be considered the equivalent of a 40-year-old making $5000, we can stop wasting time on NYT articles which fail to distinguish between them.
rwe -
I wasn't arguing with all of your ideas. Just your blanket criticism of medians, and your use of a person-based statistic to counter a job-based statistic.
And I wasn't even trying to make a statement about inequality (I think we all know it's there) - I was just trying to explain why Sowell's statistic doesn't support an argument about wages for certain jobs. I guess I took more of an issue in HOW you made your points, not what points you made.
This was good - always fun for me as a student to post something and realize how much that gets written back is still over my head.
What Mark said. The original NYT article said that 12 million of the 60 million were dependent children filing their own returns, most likely (based on my knowledge of such people) because they either had interest income on their college funds or had to file returns to get back withholding from their part-time or summer jobs.
I'm sure Social Security recipients are a big chunk of the rest, and I wouldn't be surprised if dependent adults other than spouses figure prominently too. Not everyone is trying to work full time.
That's before you even get to transfer payments and efforts to hide income.
Now we're all agreed that an 18-year-old making $5000 shouldn't be considered the equivalent of a 40-year-old making $5000, we can stop wasting time on NYT articles which fail to distinguish between them.Posted by Paul Zrimsek |
Sure. Just as soon as you concede that income mobility has decreased over time. Like the facts and figures say.
Not hard to do.
Johnston's saving grace is that he appears to be an assiduous ego-googler with a willingness to turn up on blogs and defend his work.
This reminds me of the time David Cay Johnston spammed me. I had posted in a Usenet thread about his book Perfectly Legal, and he found the thread and mailed me (and presumably everyone else on the thread) a solicitation to buy a copy. Which I thought was kind of cool, because I'd never gotten spam written by a Pulitzer-winning journalist before.
I have some work to just now, but I'll catch up with the discussion later. It has been interesting, including the criticisms of my posts. SoV should calm down though.
You interpret substantive criticism using facts and figures to rebut your assertions as 'not calm'.
Interesting. It does explain a bit, though.
If movement mobility is defined as movement of individuals between income quantiles, my impression is that it has been decreasing. But I've just been told that we shouldn't define mobility that way, and I don't know what definition we're supposed to be using instead.
Well, I figured someone would challenge the claim that income mobility is unchanged, so here is a link:
http://gregmankiw.blogspot.com/2007/11/economic-mobility-is-unchanged.html
But maybe Mankiw is a "hack" too. Anyway, if you follow the links Mankiw gives you'll find an NBER paper that argues that overall income mobility has been stable, though men have seen their income mobility decline.
MC- Thank you for agreeing with me.
But I wanted to point one quick thing out- the 12 million children number is not individual filers, but dependents of individual filers. In fact the numbers Johnston cites are for 26 million filers who reported about an average of $15 per day in income each. However, those 26 million filers included couples filing jointly and families with dependents. When you add up the spouses and dependent children, you get the 60 million figure, and an average of $7 per day per person.
Point is that it's not 12 million children working part time jobs filing income taxes, but 12 million children whose parents fit into the bottom 20% of AGI.
Er, no. You've been told that it's groups of people moving between quintiles. Probabilities. That sort of thing. That ten years later 30% or 40% or whatever of all nineteen-year-olds have moved out of the bottom quintile is not surprising. And is in no way evidence for class mobility.
You really don't see this? What is the question you have with regards to this?
rwe:
According to the Census Bureau, the poverty rate for all families was 9.8% in 2006, 11.0% in 1996 and 10.9% in 1986. How in the world, then, could anyone conclude that poverty has been spreading unchecked?
Also according to the Census Bureau, the (estimated) population of the United States was 299M in 2006, 265M in 1996, and 240M in 1986. Using your percentages (which I'm not sure are correct, I found other tables on the census.gov site showing higher percentages) that works out to 29.3M in poverty in 2006, 26.4M in 1996, and 26.16M in 1986. Looks like that number is going up, not down. And yes I know that total population was also going up and that therefore in percentage terms the number in poverty was decreasing; but in real terms, there are more people in poverty than before. 2.9M more in the past ten years, by your numbers. That's progress?
"Unchecked" is a loaded term and I'm not sure I'd use it either. At the same time, I'm not going to blind myself to the facts in front of my face. Despite what you might think.
liberalrob,
Leaving aside whether that is true or not, the argument is that the rich are getting richer MUCH faster than the poor, and that that is in fact a problem which needs to be addressed.
Why is it a problem, and how do you propose to address it?
SoV:
You mean those 'studies' that include college kids, which find that, ten or twenty years after they graduate, they have moved into a higher income bracket from when they were in school?
Sure, but that's the whole point, isn't it? A lot of the people in the bottom income quintile aren't really "poor" in the common sense of the word. They're students. Or retirees living off savings. Or business owners with highly variable incomes.
A corollary to this: The median income, and also the average income of the lower quintile, are probably dragged down by increased college enrollment.
I guess what I don't get is who these "groups" are (not age cohorts, evidently), or how using them rather than individuals as the unit of analysis is supposed to change anything.
Let's not forget the other half of the bargain! I'm quite willing to agree that the guy who made $5000 in college and makes $40,000 now was really middle-class all along-- but in return I demand that people who write articles about the poor proles scraping by on $5000 acknowledge that many of those are really middle-class people who haven't gotten their age-related bump yet.
And of course (one more time!) if we're going to talk about mobility at all, we'll need to know under what circumstances we are entitled to say that people who've made the transition from $5000 to $40,000-- whether singly or in bunches-- have benefited from mobility.
Stephanie wrote: Somehow, that seems illogical to me - when someone wins, generally, someone else loses (even if it's not an equivalent loss).
Unfortunately, that's basically the zero-sum fallacy in a nutshell (which is sometimes arrived at by abusing the meaning of pie charts). There is no inherent reason why one person's gain should become another person's loss in an economy, because the overall size of the economy can change.
rwe:
I just object to people muddying the distinction between absolute and relative measures of well-being and pretending that because those at the bottom are relatively worse off than they used to be, they are also absolutely worse off.
I think this is the essence of your problem and why you get so much flak (from me at least). You analyze things in absolute terms and declare that everything's just peachy; but those people who are relatively worse off say hey wait a minute, that may be but those guys over there are WAY better off, what's up with that? Why is it OK to rig the system so they get a HUGE benefit and I get a dinky one? (And one which I may not even really understand I'm getting.)
You are dealing with society and people, not just mathematical equations and statistical charts. If your policy doesn't result in palpable, "real" improvement in people's lives, it isn't going to matter that they're slightly better off in absolute terms. Compared to the relative inequality (and the absolute one too, for that matter), it's meaningless. So as in your analysis of the relatively improved poverty picture above, you see improvement and others see more people in poverty. Maybe it's irrational, but you're going to have to deal with it.
liberalrob wrote: And yes I know that total population was also going up and that therefore in percentage terms the number in poverty was decreasing; but in real terms, there are more people in poverty than before. 2.9M more in the past ten years, by your numbers. That's progress?
It isn't anything but meaningless numbers until analytically reduced to the proportion that are systematically impoverished (problem) versus first- and early second-gen immigrants who entered poor, but will shortly work their way into a higher income bracket.
liberalrob: that's not a real strong argument there.
I believe the poverty rate among children has been steadily increasing, however.
rwe, I've seen stats that income mobility is lower than it used to be. But your note that the NBER finds income mobility among men has dropped while overall mobility is unchanged is certainly interesting; it suggests that mobility, like overall income levels, is only being preserved by the increasing percentage of women in the labor force. Since that percentage is pretty much maxed out, one would expect to see mobility drop now.
But in any case, Stephanie was sort of right above: increased mobility means more people moving from poor to rich, and from rich to poor. If the bottom quintile's income level isn't rising, that means that when middle-income people drop to low-income levels, they're dropping back to being just as poor as low-income people were in 1973, when a more equitable distribution of growth would have them only dropping to the level of middle-income people in 1973.
Scent,
Further, median wages have been stagnant for a long time, and that's another fact everyone knows.
Ah yes, "everyone knows." Riiiight.
Wages are only one component of employment compensation. They don't necessarily even include all cash employment compensation (commissions, bonuses, profit-sharing, tips, etc.). As a share of national income, employment compensation has remained relatively constant for decades.
Wages are also only one type of money income. The Census Bureau lists around 15 additional types. Wage data also omits the effects of taxes and cash and non-cash government benefits.
Bottom line: wage data alone tells you little about true incomes.
In addition, income is a poor measure of standard of living or material condition. Consumption-based measures are much more useful.
Also according to the Census Bureau, the (estimated) population of the United States was 299M in 2006, 265M in 1996, and 240M in 1986. Using your percentages (which I'm not sure are correct, I found other tables on the census.gov site showing higher percentages) that works out to 29.3M in poverty in 2006, 26.4M in 1996, and 26.16M in 1986. Looks like that number is going up, not down. And yes I know that total population was also going up and that therefore in percentage terms the number in poverty was decreasing; but in real terms, there are more people in poverty than before. 2.9M more in the past ten years, by your numbers. That's progress?
No, it's the rate that is the usual measure, for obvious reasons. To use reductio absurdum, a poverty percentage of .01% in a population of one trillion is still 100 million, far more than your figures give. That doesn't make sense.
liberalrob,
The increase in poverty is being almost completely driven by low skilled
immigration. This is important to note when evaluating the US government's effort to mitigate poverty.
liberalrob wrote: Why is it OK to rig the system so they get a HUGE benefit and I get a dinky one? (And one which I may not even really understand I'm getting.)
Well, Rob, it's like this: We mice don't want you screwing up our 10 million year The Answer To It All computer system, see? That's why we've rigged it so that the proportions of rich and poor people will always be maintained in balance. And that's why your exemplary determination, ambition, and general attempts to make yourself blind stinking rich have all failed; your hard work means nothing because we rigged the system, doncha know.
Good luck with navigating your way through it! But now that you're in on our secret, don't keep working hard to advance; we recommend lottery tickets.
Why is it a problem, and how do you propose to address it?
It's a problem because many people perceive it to be one. The reasons why are amply detailed on many blogs and websites. John Edwards' campaign website has one such explanation; try looking up something called "two Americas."
I don't propose to address it, as I'm not running for office. I simply note its presence. As to how I might hypothetically propose to address it, again those policies I would advocate have been amply detailed many places. Step one would be repeal of the Bush tax cut for the wealthy. You can take it from there, I'm sure.
Sure, but that's the whole point, isn't it? A lot of the people in the bottom income quintile aren't really "poor" in the common sense of the word. They're students. Or retirees living off savings. Or business owners with highly variable incomes.
Actually, I made my first comments there about the idiocy of citing Sowell, and his quoted statistic that 91% of people in the bottom quintile leave it. Here's rwe's 'contribution':
Sowell shows the fallacy here of ignoring income mobility. When you actually follow what happened to the people at the bottom, you find that their incomes have risen significantly.
I'm merely pointing out this Just Ain't So.
Here is an article about the material living conditions of America's "poor." The data all comes from official government sources. As you can see, the typical American household officially classified as being in "poverty" actually enjoys a pretty good standard of living. This is not to say that there is not a small number of Americans who really do live in conditions of severe material deprivation (the homeless, for example). But in general, America's "poor" have it pretty good in this respect.
liberalrob,
It's a problem because many people perceive it to be one. The reasons why are amply detailed on many blogs and websites.
Why do you perceive it to be one? What is the nature of the problem?
I don't propose to address it, as I'm not running for office. I simply note its presence.
No, you didn't simply note its presence. You said it "needs to be addressed."
As to how I might hypothetically propose to address it, again those policies I would advocate have been amply detailed many places. Step one would be repeal of the Bush tax cut for the wealthy. You can take it from there, I'm sure.
Repealing the Bush tax cuts for the wealthy would more-or-less return us to the tax regime as it was during the Clinton years. But even during those years, the rich still got richer much faster than the poor did. Clinton presided over a huge increase in economic inequality, at least as it is conventionally measured. So your proposal isn't likely to do much even to slow the rise in inequality, let alone reverse it.
Mark -
Are you sure? I read the article a couple of times and couldn't quite make it out. There are 26 million tax filers involved and 60 million individuals covered by those returns, but it doesn't say "households" in connection with the IRS data. Census data, yes, but the IRS seems to be going by tax return. It's a pretty confused presentation in the Times, and I can't find the IRS report itself to see what it says.
Anyway, the mom + dad + kids model may not be the most typical household at these ranges anyway. What do you do with grandma (on Social Security) + mom (working part time while in school) + kid? Or a group of grad school housemates, one of whom just had a baby? Or an extended family group with multiple wage earners (who can't file together) living in the same house?
The 16-year-old with a summer job probably still uses his parents' address, but the 21-year-old may not even if his parents (or other relatives) still send him a lot of money. The article doesn't say that taxpayers who check the box about being claimed as a dependent were excluded from the analysis, so I assume they are still in there SOMEWHERE. Maybe they are counted with the 48 million adults instead, but who can really tell with a presentation like this?
that's not a real strong argument there.
My argument is about perception vs. reality. rwe takes issue with the LA Times statement that "we seek a president who understands that the unchecked spread of poverty is not a natural condition" and presents a fairly stable percentage of people in poverty as evidence that poverty is in fact in check. I'm pointing out that many people's perception is that poverty is spreading; they don't care that the percentage reported by the Census Bureau is stable or dropping, they see 2.9M more people in poverty (or more importantly increasing visibility of the poor where they live) and conclude that for them at least poverty seems to be spreading. bp may (or may not) be correct that low skilled immigration is the reason for that perception, but the bottom line for the Presidential race is that poverty is perceived to be spreading. In that narrow sense, the LA Times statement was accurate.
The rest of the thing does seem to be pretty off-base. I suppose the fact that it was an editorial and not a "hard news" story might give them some cover, but even I know better than to use the "World Socialist Workers" web site as a source. Megan's right about that.
I suspect it won't be much longer before the NYTimes publishes a story on how the numbers of those in the lowest quintile has been stuck at 20% of the population for decades.
SOV,
So, is Sowell correct or not? You seem to be claiming that he is wrong to include the young in his calculations. If that is correct, then I must object to your including the newly retired (who often move down in earned income), the newly semi-retired, and all new immigrants, legal and otherwise.
Also, I just loved the table that uses the five years before and after 1980. Nice choice.
Liberalrob,
As for the comparisons of the absolute numbers of those below the poverty level, I think it sufficient to simply point out that even Scent of Violets couldn't quite swallow that one.
No, you didn't simply note its presence. You said it "needs to be addressed."
No, I said "the argument" was that it needed to be addressed. Not just MY argument, although I do in fact agree.
Repealing the Bush tax cuts for the wealthy would more-or-less return us to the tax regime as it was during the Clinton years. But even during those years, the rich still got richer much faster than the poor did.
There's this concept called "being off topic," and I'm afraid if I go much further down this road with you on your quest to find more things to laugh at me about we'll be truly far off topic indeed. This argument has been rehashed over and over on many websites and blogs (including this one) and I'm not going to go any further than that. I believe that massive disparities in wealth are bad for society, I believe that that disparity has increased significantly over the past 7 years, and I believe that it is a problem that needs to be addressed. When an appropriate topic comes along again, I'll explore my beliefs with you (even though your only object is to ridicule me). Until then, no more.
MC-
It is definitely a confused presentation, and I can see how you came to the conclusion about the 12 million dependent taxpayers. But the numbers only add up if you include the children as dependents of taxpayers rather than taxpayers themselves.
The thing is that you have to remember that the phrase "dependent children" has a specific meaning to the IRS- they are children for whom the parents can claim a credit on their return.
It has been a while, but if I remember correctly, I filed one tax return in which I was still claimed as a dependent on my parent's return. The net effect was that I was not allowed a personal deduction. I have no idea how people like this affect the numbers under discussion.
So, is Sowell correct or not? You seem to be claiming that he is wrong to include the young in his calculations. If that is correct, then I must object to your including the newly retired (who often move down in earned income), the newly semi-retired, and all new immigrants, legal and otherwise.
Sigh. What is it with these conservatives who just can't seem to take the plain meaning and infer something deeper instead? Besides the fact that the inferred meaning is always used as a slam that is, where the plain meaning often can't be.
No, Yancey, I said nothing about could lead you to conclude that.
And yes, as a matter of fact, since I do include adults of all ages, I would as a matter of fact include the newly retired.
Where have I said otherwise? Or is this just another attempt at some lame emotive slam?
Given what the rest of the article is like, I wouldn't be surprised if the NYT got it all wrong. There could be dependents of both kinds within the 12 million, for all we really know. And I still want to know where the kids with summer jobs are. To the extent they aren't in the 12 million, they should be in the 48. Perhaps the IRS really did exclude them, but the article doesn't let us draw that conclusion.
I thought the IRS term of art was "qualifying child."
It's also worth noting that the NYT article says that the bottom 60,000,000 people live on an average income of $7/day. Putting aside the fact that this is obviously wrong (that's far below the poverty line, and the poverty rate is closer to 10% than 20%), it still doesn't mean that 60,000,000 people live on incomes of less than $7/day. If the average for that group is $7 per day, it's likely that around half make more.
A corollary to this: The median income, and also the average income of the lower quintile, are probably dragged down by increased college enrollment.
And the increased numbers of men in prison. In fact, could this also explain why mobility for men has decreased? Also, more women go to college than men. (Today's Slate has an article that says the ratio is 4 female to 3 male graduates.)
If there is one thing that will make it hard to get a middle class job, it's a felony record. I know that my employer, a large aerospace firm, won't hire felons for most jobs - they can't get any form of security clearance.
So brooksfoe had stellar SATs? Feh. Come back when you have a 1600 on the GRE.
"we know that the poor tend to strongly underreport income because reporting income means losing their EITC and health benefits."
Do you have any data here, or is this just something you "know"? What does "tend to strongly underreport" mean? Certainly, people who exceed the income limits have an incentive to underreport, but those of us who aren't poor have an incentive to "tend to strongly underreport" our income because that leads to lower taxes. Is the likelihood of getting caught filing phony reports to qualify for the EITFC lower than the likelihood of getting caught cheating on your income taxes? I wonder.
SOV,
Then I have to simply take you at your word that you are objecting to the fact that he actually counts the people who move from one quintile to another, which you explicitly state is to be expected for 19 year-olds. If we are going to talk about income mobility at all, then we must discuss actual people, and their personal experiences throughout their lives, otherwise we have to discount all the people that deliberately choose their quintile status (the people I asked you about in my previous question), and, even more importantly, if we are not talking about the experieces of actual people any longer, then why should anyone care about this other than as an abstraction?
Your's and other's stated thesis- that incomes in the lowest quintiles has a falling relative value to the incomes at higher quintiles. What reasonable guesses could we make about this situation? There are several, but one of the major ones not mentioned in this thread yet is that the population is getting older (and has been since about the time that median wages started stagnating), and the more experienced a worker is, the more likely he is to be paid more. Because income quintiles are defined as percentages of the total population, the aging boomers, the large majority of which are in their prime earning years, are now increasing the average age of the higher quintiles relative to the lower quintiles which are still largely populated by workers just starting their working lives and, increasingly over the last 20 years, new immigrants both legal and illegal. Thus it is little surprise that the top half of the working population is drawing a larger income relative to the bottom half.
And note that I have not even mentioned the fact that older people, having had the actual benefit of more time, will derive larger quantities of income from accumulated investments. This is an effect that has grown over the course of time due to the increased productivity of the overall economy.
All of this, of course, is part of Sowell's point- the quintiles of 1975 are not composed of the same people, or even the same types of people as the ones of today.
The underreporting of income is most likely to happen with those who work more on a cash basis. Are such people more likely to be poor or wealthy?
I wasn't going to respond to this, but I'm bored, so...
squeak:
Well, Rob, it's like this: We mice don't want you screwing up our 10 million year The Answer To It All computer system, see? That's why we've rigged it so that the proportions of rich and poor people will always be maintained in balance.
Your efforts have been a miserable failure, then, because that balance is nowhere to be seen.
You're the funniest of the three, though, so you get to be Curly. rwe is Moe, and Mixner is Larry. Auditions for Shemp and Curly Joe remain open.
Word of the day: kleptocracy.
Your efforts have been a miserable failure, then, because that balance is nowhere to be seen.
Nonsense! We successfully kept you down in spite of your unceasing best efforts, didn't we? And now...Get his brain!
Alan Vanneman wrote: Is the likelihood of getting caught filing phony reports to qualify for the EITFC lower than the likelihood of getting caught cheating on your income taxes? I wonder.
Here's a simple hypothetical for you: You are a poor person. Fortunately, you do have some entreupreneurial talent, and after a few big farm paydays in the summer months, you decide to buy an old truck, a lawnmower, a leaf rake, a box of garbage bags, and a pair of gloves. Congratulations: you're now a self-employed grounds-maintenance worker! After a few weeks of saving up money, you can invest in selection of chemicals and a few more implements, and add "Lanscaper" to your title, as well.
Your customers reimbuse you in cash or checks, the latter of which can be cashed for a few bucks at the hole-in-the-wall Check, Cash, & Pawn. The checks are lost in the blizzard of deposits made by that agency, and the cash never passes through a hands of any financial operator at all. And only your commercial customers ever file I-9s. Some of them. Maybe. Or maybe some of the self-employed ones decide the extra paperwork is more hassle than it's worth, and simply pay you cash out of pocket.
Question: Assuming you are graced with much less than an exemplary amount of civic virtue, how much of that income can you potentially non-report without the IRS ever having a hope of discovering that you're not reporting it? (Shoot, if you play it right, they may never even find out that your business exists...especially if you work some other low-wage job and run this business on the side.)
That is more in line with the kind of scenario we are discussing here.
"And only your commercial customers ever file I-9s. Some of them. Maybe. Or maybe some of the self-employed ones decide the extra paperwork is more hassle than it's worth, and simply pay you cash out of pocket."
Don't forget tips either. I'm not sure how many waitresses and bartenders there are in this great country, but the vast majority are underreporting their income by a significant degree. Not that there's anything wrong with that.
Then I have to simply take you at your word that you are objecting to the fact that he actually counts the people who move from one quintile to another, which you explicitly state is to be expected for 19 year-olds. If we are going to talk about income mobility at all, then we must discuss actual people, and their personal experiences throughout their lives, otherwise we have to discount all the people that deliberately choose their quintile status (the people I asked you about in my previous question), and, even more importantly, if we are not talking about the experieces of actual people any longer, then why should anyone care about this other than as an abstraction?
Sigh. You're working much too hard at this. The po
Theme of the week:
Reporters who can't count.
Posted by heedless | January 15, 2008 7:48 AM