Lies, damned lies and statistics.
I do not know who predicted this NYT headline, come Judgement Day, “World to End Tomorrow, Poor to Suffer Most,” but their faith in the Gray Lady must once again have been restored. In an article purporting to show that the average US income for most people had dropped in the first five years of the Bush presidency (as contrasted to the final throes of the dot com bubble) in an article whose headline had to be later softened up, Times journalist David Cay Johnson Johnston managed to choose his data points rather carefully.
So carefully indeed, as to paint a rather misleading conclusion. To his credit, Johnston shows up in comments at Bizzyblog to defend his result, if not his method. Points on for playing the game, but still – bad form.
I don’t know what’s more disturbing: That the paper thinks the rest of us are too stupid to check their numbers, or that – based on the earnest approbation such news receives in partisan quarters – the fact that they may be right.
We have a term we use for people in the Navy who attempt such cavilling – they are called “sea lawyers”, and the term is by no means an honorific. Plebe midshipman are taught what label the Honor Concept attaches to any attempt to use language which, while it may be technically accurate, is intended to foster a false impression.
This, we were told, is known as a “lie.”
Update: Mr. Johnston has acknowledged certain flaws in his methodology even while letting others – like the 2000 bubble and 9/11 effects – go unremarked upon. Other complications to the method as presented, including the fact that he chose to use pre-tax incomes as a proxy for disposable income, are ignored entirely as are the implications of changing tax policy from 2000 – 2005. After all, since we’re taxed on AGI, it’s in all of our selfish interest (gross income remaining constant) to reduce our AGI insofar as possible. Finally, mean (as opposed to median) incomes are a terribly ineffective way of judging how “most” of us are doing, tax adjusted or otherwise, since the data can cause skews and “fat tails” rendering the measurement useless.
One way to have avoided the impact of tax policies and capital gains would have been to use the common GDP per capita metric as found here. It’s instructional:
Although per capita GDP did in fact decline marginally from 2000 to 2001 – the last full year of the previous regime’s economic influence – they increased markedly thereafter (these are measured in constant 2000 dollars). Disposable income never decreased on a per capita basis. I wonder why this data wasn’t presented?
Perhaps they didn’t quite tell the right story.
Huh.
You know, an alternate theory to truth-shading or agenda-pushing presents itself at this point – it’s quite possible that the folks over at the NYT simply don’t understand the math.
I think I understand the refutation of this article however, where I live, a true blue state, property taxes have skyrocketed over the last ten years as well as oil fuel, gasoline and insurance of all stripes. Wages have not risen to meet these fixed costs and people are bailing out left and right…
I don’t happen to base this loss of buying power on the Federal gov’t. Quite the contrary, I base it on the Dems that run the state, county and local gov’t. My county collects a cool $40 every time I fill my heating oil tank . I find it obscene to tax heating oil and I have told my legislator that on several occasions… I actually got an email back where she said “I feel your pain!” Pain, you stupid woman? Pain is sitting in my house during an ice storm wearing a coat and ski hat knowing that I pay over $5,000/year to heat the place (heat, in this sense, means preventing the pipes from freezing). Pain is the “one light policy” we employ to keep the electric company at bay (a state gov’t sponsored monopoly, I might add, that seems to be above the law of regulation as more than half of my electric bill is now comprised of an amorphous “energy surcharge” that has never been confirmed by our state legislature…
The school taxes have increased at an average of 10%/year for the last decade and every year we hear how “underpaid” school teachers are for their 185 days of service per year with complete medical care with no personal contribution and a pension!
This is all on the backs of working stiffs like my husband that bring basic dollars into the local economy and are subject to having a cardboard box shoved into their hands the better to clean their desks out at a moments notice.
By the time my state and local gov’t gets done taking it off the top, we pay almost no federal tax. If half the neighborhood weren’t already for sale, we would leave…
Babs;
Sorry to hear it, but, here in “Doity Joisey” it’s the same. Saw a poster on the back of a bus that proclaimed that it cost the Newark School system $750,000 for each ordinary high school diploma. I fear for us all.
You know Pixie… This thread hits a raw nerve. When my oldest was taking HS Calculus the teacher announced that two days prior to the 8 school day spring break he would not be in attendence because he was getting an early start on the holiday he was taking at his condo in Aspen… This guy was making over $100,000/year as a HS math teacher along with health insurance and pension! I thought I would go nuts as the kid wasn’t doing very well in the class. I once called this same bastard to talk about my son’s performance in his class and he told me that “if I got to the school within the next 15 minutes he could give me 10 minutes of his lunch hour.” I wanted to tear his throat out.
Heh, it is a funny thing but when the boys were babies I had to supress the idea of killing anyone I thought would harm them…
The HS math teacher sort of fell into the same category and I can’t help but think that during our brief interview some of that hormonal energy bled into the discussion…
The fact is that this elite teacher found room in his schedule to work with my son and things turned out OK.
I still hate him like crazy and hope he never crosses my path again (or I smell him in the wind and get my fur up!)
So just what do you assert I lied about? That is a very serious charge and yet you do not cite above a single fact, you just make assertions (and get my name wrong), while my article relied on my careful analysis of official government data.
We may reasonably disagree about we think is most significant — five-year data affecting everyone as opposed to one-year results, for example. But your post does not assert a disagreement over significance, but takes a quantum leap in asserting that I lied. The data say just what I reported they say.
I take great care to deal with these complex numbers, which virtually no one else reports on, so that readers get a rounded view taking into account perspectives such as time and other relevant factors.
That said, in retrospect I should have given more attention to the one-year data.
I had on my mind two thoughts. 1. Those who criticized me two years ago and one year ago for focusing on one-year changes in the same dataset. 2. Multi-year data is more revealing and affects more people and gives more historical perspective.
By every journalistic convention and standard because of significance both in numbers of people affected and duration the longer term is a better measure and would come first, but, as I said above, in retrospect I should have included a line high up with more on the one-year data change.
Lying is such a serious charge that you should either prove I made up facts (and The Times should fire me forthwith) or you should have the good manners to acknowledge that this is a disagreement about what is most significant out of the innumerable possibilities from the thousands of data points I examined and reported on.
Lex, Well written, Sir. I’m not exactly a kid and I’m not dead, yet. Just don’t tell the coroner. You make a good point. Did you ever notice when people argue or debate, their strategy is whoever can talk the longest or loudest wins? Did you notice, there is no mention of fact or THE TRUTH, not just truths or true statements. There was an old man that I the privilege of knowing until he was over 100 years old. He would often say, “The judge of the debate is not length or volume, but the real judge is a real hag (Lex, what you might call a real ‘sea hag’.) This hag has no forgiveness, she’ll make sure THE TRUTH comes out. She’s a nasty old hag. But there is one thing I want you to do, I want you to remember her name. Her name is ‘History’”. He is gone, but the old hag is still here.
Thank you,
Grumpy
Sorry about mishandling your name, Mr. Johnston.
The “lying” referred to reflects upon the title of the post itself, that there are three kinds of lies: Lies, damned lies and statistics. Not knowing you personally, I take no flings against your personal sense of honor and regret if it seemed that I had done so, even as I acknowledge how easy it would have been to feel personally affronted in your shoes.
Nevertheless, I make the altogther ordinary observation that statistics are often used to paint a story that, while technically true in the most limited sense, seems at least apparently calculated to foster an impression opposite to the general reality. The charts linked above demonstrate that after-tax (i.e. disposable) incomes have risen for all but the top 20% of wage-earners. That, combined with the one-year data issue you mention above and the suspect comparison date of the study as framed – 2000 was, after all, the last of the champagne – leaves those of us not favored with your acquaintance to form the impression that you chose your data to buttress an impression congenial to your pre-existing conclusions.
Not lying exactly, no. But neither entirely forthright, is it?
Gauntlet… And then, thrown!
Looking forward to the back and forth…
Thank you for clearing up my question.
In my world, lack of forthrightness is on a par with lying. So, no, I do not for a moment think I was less than forthright in this or anything else I have ever written. And if my piece was in any way less than forthright or shaded for some venal reaso
There is no place here for testi-lying, as the cops call it.
After 40 years, top prizes, being cited in textbooks and having written stories that have withstood scrutiny by grand juries, government investigating committees at all levels and other attacks, I feel confident on my judgments on what is news and on all of my facts. That doe snot mean I am perfect as like everyone else I make mistakes — and I own up to them, including volunteering corrections of fine points that no one saw or likely would ever figure out because I want to record to be right because that is the onloy standard I can live with.
I go where the facts go and in this case it seemed to me that five years of below peak average incomes is long enough to be significant and, indeed, the most significant fact to emerge from the data.
That is the key test of news — is it significant? How many people does it affect?
The five years is from the peak of the last economic expansion, a date widely and commonly used by journalists and others not just in this decade, but in periods past. (Trust me, I have spent long hours reading clips going back a century to get a good grasp on past economic reporting in this country as well as on other issues.)
That said, in retrospect I should have given more attention to the one-year change. On that point I think the criticism is valid and next year I will keep that balance in mind.
But I never pull punches whether I am accusing people of committing crimes (and people have gone to prison for long periods of time because of my work) or engaging in misconduct or when I hunted down a killer the police failed to catch.
I have been hit with police night sticks, roughed up by unionists and threatened by people with bottomless funds to go after me and I have never flinched, never changed a word because of that. I go where the facts go and report what of all the facts I gather are, in my judgment, the most significant.
Readers can reasonably disagree on whether I picked the right facts.
But readers should appreciate that I always take care with the facts, do not report numbers that are alarming but misleading (and in this dataset there were plenty of those that emerged from the two dozen or so spreadsheets I prepared before writing my article.)
And I don’t give a hoot about who is in power (as decades of clips with my name on them will show).
Now maybe we can focus not on the messenger but on the facts.
Mr. Johnston:
Should you get in trouble with your employer, merely say that the data was ‘fake, but accurate”, and that “while the facts were wrong, the narrative was right”.
If all else fails, write a nice piece about Comrade Stalin. Maybe you can get a Pulitzer for that.
Personally, I am far more interested in the article’s actual message rather than quibbling over the methodology and data used; and more importantly in finding answers to the question: Why are so many Americans feeling economic stress, and are so gloomy about the economy?
Regardless of what average incomes have been for the past year, 5 years, or a decade, they will certainly fall in a future recession. And given the many current red flag indicators, a slowing of growth if not a true recession certainly looms in our future.
[As an aside, I have noticed frequent critiques of the NYT here, and the resulting and uninhibited great sport of piling on with negative comments. It's therefore refreshing to see Mr. Johnston come on to provide some balance and counterpoint, and to state his case well.]
Re: “focus… on the facts.”
Fair enough. Do you think it’s fair to judge the end-of-era dot.com boom as just another cyclical peak? I mean the tech collapse “vanished” billions of dollars in market capitalization that probably shouldn’t have ever existed, based on the market’s subsequent reaction. It also threw tens of thousands of high paid workers on the streets and culled something like half of all tech start ups from the rolls. Had to happen of course, but it does rather color the value of that year when used as a comparison, not to mention the effects of the terrorist attacks on 9/11.
Wouldn’t you call the 2000-2001 period someting of a statistical outlier? You’re tacitly arguing that this is the equivalent of a random five-year peak-to-peak comparison – a hard sell, especially given the fact that there’s no reason to suspect real income growth after taxes – the kind of money people actually “make ends meet” with – had yet peaked by 2005. Growth remained strong throughout 2006 and we’ve yet to understand what the effects of the sub-prime mortgage driven credit crunch will do in 2007.
Do you really think that Americans are concerned about present day spending power – which can only be sensibly be defined by the combination of after-tax income and benefits – because of their 2005 pre-tax incomes? To me it smacks more than a bit of begging the question, Mr. Johnston.
Wouldn’t a better comparison have been the peak of the 1987 equities market followed by the October crash? Five years after that, George H.W. “Read My Lips” Bush was getting creamed by the Kid from Hope for failing to have the “vision thing” on the economy. You could have easily – and more accurately – written, “Bush economic policies quickly generated new highs in American spending power despite market collapse, terror attacks.”
This isn’t just a matter of reasonable people disagreeing on what the right numbers are to pick, it’s about the context in which those numbers are presented. That’s the analysis work people like you get paid to make, and the distinction is not merely technical.
Unemployment has been at frictional levels for year over year even as inflation remained low, tens of thousands of people moved out of poverty and the economy has grown for quarter after quarter. That may not suit the redistributionist policies of those at Citizens for Tax Justice, but it’s true nonetheless.
There’s probably a good reason why Americans are down on the prospects of the general economy, even as they’re overwhelmingly satisfied with their own situations.
Each of us should look within ourselves to see if we’ve contributed to that cognitive dissonance.
There. I’m done. How about you?
fliterman,
I’ll give you my own opinion on the issue which is, like all opinions, mine and mine alone, etc, ad nauseum.
The impression that scads of people feel worried about the economy is a direct result of folks like Mr. Johnston telling them they should feel worried. Mr. Johnston is paid to tell a story, and like it or not, his story will most often (if not always) reflect the biases and opinions of his editor and the owner of his paper.
That paper, the New York Times, has a clear agenda at odds with the current administration, conservatives of all stripes, and most things American. It has a demonstrable bias against the American Armed Forces, which is easily seen in the constant (over 100) stories about Abu Graib, even though that unfortunate event was simply a side show of the big show, and a pretty insignificant one at that. Insignificant, that is, until the NYT decided it could damage the current administration by repeated airings of the tale in various permutations.
But I digress.
I could give a tinker’s damn about Mr. Johnston’s awards, experiences, personal history, et al. All those put together and $5.00 will buy him a cup of Starbucks. If he likes that stuff. I personally believe it is overpriced and over-hyped, much like his employer. Everything I need to know about him may be learned through his writing.
The problem that I have is that his writing is in the NYT, and it pains me to have to review that daily in order to read his works, as there is so very little to like about it, (the NYT)let alone believe without multiple corroborating sources.
If the NYT was as vociferous in it’s fact-checking and objectivity as it is in foisting it’s political views and biases upon the general public, it might one day rise to the level of a good news source.
Until then, it shall be best remembered as the “grey lady” which, as we all know, is the pallor associated with death. In this case a not wholly inappropriate moniker.
The above is my own personal opinion, and in no way should be taken to reflect in any manner the views of the management, owners, or any one else associated with or visiting this blog.
Respects,
Mr. Johnston- I accept your assertions that you are a man of great inellect and spotless integrity.
Why, then, do you insist on associating with an institution with a propensity to reveal classified information which harms our national security; prints “the truth” while carefully omitting facts that conflict with their agenda driven story line; and has become the recognized epitome of the biased liberal media?
I have not read your article, and will not, considering the source which printed it. I mistrust everything in the “gray lady” as possibly false, fabricated, manipulated, distorted, selectively plucked or planted.
Since you are better than that, you obviously do not belong there.
I have great respect for Lex’s intellect, judgement, and honor, so I think I will accept his characterization of the piece as accurate.
But fliterman, the data is the message and the process is the way in which the data emerges. Everything else is mere packaging to make the message comprehensible. To appreciate the message more than the underlying data is to commit oneself either to an act of faith or else to preference. I’ve no principled objection to doing so, but such things should not be passed off as analysis.
The WSJ article I linked immediately above gives a good answer why people are glum even as they’re happy with their own prospects. In the name of “balance,” any good economic news – and there has been plenty – must be balanced by the 5% “yes, but” club. Fair enough as far as it goes, but if 95% of economists believe one thing (impossible, I know, but bear with me) and 5% believe its opposite, it is hardly reasonable to maintain that the two positions are roughly equal. Yet that is what we’re treated to.
And with respect, most of Mr. Johnston’s rebuttal above seem not so much to me to be a case “well-stated,” so much as an argument by assertion and a pro forma and wounded appeal to authority. I can tip my hat to him for making the effort, especially in these our lowly digs, but merely to claim for oneself the tribune’s mantle while rebutting points not put forth in argument does not in itself seem so very compelling.
The NYT does not attract our attentions merely for existing, we do not arise in the morning and look for reasons to tilt at them – they daily are thrust upon us.
Oh God Lex!
The NYT are “thrust upon us?’…
Not me. They spent one year dropping it at the end of my driveway like so much litter. I still didn’t read it…
Shipmates,
All of this reminds me of one of my favorite passages from one of my favotie plays.
From:
Henry IV Part I: Act 3
Scene 1, the Archdeacon’s house.
…..
GLENDOWER
Cousin, of many men
I do not bear these crossings. Give me leave
To tell you once again that at my birth
The front of heaven was full of fiery shapes,
The goats ran from the mountains, and the herds
Were strangely clamorous to the frighted fields.
These signs have mark’d me extraordinary;
And all the courses of my life do show
I am not in the roll of common men.
Where is he living, clipp’d in with the sea
That chides the banks of England, Scotland, Wales,
Which calls me pupil, or hath read to me?
And bring him out that is but woman’s son
Can trace me in the tedious ways of art
And hold me pace in deep experiments.
HOTSPUR
I think there’s no man speaks better Welsh.
I’ll to dinner.
…..
GLENDOWER
I can call spirits from the vasty deep.
HOTSPUR
Why, so can I, or so can any man;
But will they come when you do call for them?
GLENDOWER
Why, I can teach you, cousin, to command
The devil.
HOTSPUR
And I can teach thee, coz, to shame the devil
By telling truth: tell truth and shame the devil.
If thou have power to raise him, bring him hither,
And I’ll be sworn I have power to shame him hence.
O, while you live, tell truth and shame the devil!
One only needs to replace the name of Glendower with the New York Times and that of Hotspur with the common man to see the similarities…
Respects,
Web Reconnaissance for 08/22/2007…
A short recon of whats out there that might draw your attention, updated throughout the day…so check back often….
God, but how I love this! My thanks to you all, even the NYT, (although I no longer get it since my parrot died).
My thanks to you especially, Lex, for stirring the pot.
lex #15
Rarely are data alone the message. To be telling, data must be selected, organized, processed, analyzed, and yes even interpreted to have meaning. Still data are most often then used to support a message, rather than convey it. And factual but limited selected data from the same source can indeed support conflicting theses, whether intended or not. Just like the Saxe’s The Blind Men and the Elephant.
Indeed in the WSJ article you linked, that author (while ironically using the media to critique the media) conveniently selected only a few positive data points to support his message while he conspicuously, by their omission, avoided several major data points that would severely weaken his message ?
lex #15
Rarely are data alone the message. To be telling, data must be selected, organized, processed, analyzed, and yes even interpreted to have meaning. Still data are most often then used to support a message, rather than convey it. And factual but limited selected data from the same source can indeed support conflicting theses, whether intended or not. Just like the Saxe’s The Blind Men and the Elephant.
Indeed in the WSJ article you linked, that author (while ironically using the media to critique the media) conveniently selected only a few positive data points to support his message while he conspicuously, by their omission, avoided several major data points that would severely weaken his message – something I believe you accused Mr. Johnston of.
Your interest is using AGI rather than personal or disposable income unfortunately would skew the data, since unlike personal income, AGI includes capital gains and losses, but it also excludes the income of those many who do not file tax returns. Nor does it include nontaxable benefit and transfer payments like VA disability, Medicare, or a portion of Social Security payments.
Likewise, using Per Capita GDP would be a very poor metric for personal income since there is no direct relationship between the two.
Whatever data Mr. Johnston selected, and whatever added or replacement data may be used, I believe the message is the same. While the economy and personal income on the surface may have been growing nominally, there are some underlying problems.
Essentially, the rich have been getting richer, and the far less rich have essentially stayed poor. This is better explained by the well known Gini Index which shows that for three decades, the gap between America’s rich and poor has been ever widening.
And while for now a growing GDP and rising incomes (either mean or median) do raise almost all boats (but certainly not farm income…), unlike a rising tide, not all our boats are rising equally.
Income inequality is a seriously growing problem. And surprisingly it is a greater problem now for the US than many other advanced economies, as can be seen in the Gini graph. But while Bush tax cuts and policies may have exacerbated this widening gap, it also occurred under the Clinton administration.
It is a growing problem that will eventually have to be addressed. And widening income inequality is only one of many serious and growing economic problems that beg to be addressed…. And better addressed smartly, sooner rather than later for all our aggregate welfare.
He lost me when he mentioned “quantum leap”, followed by “complex numbers”. Fortunately, he missed using “decimation”.
Cheers
Per capita GDP is equally valid as a measure of economic activity as is Mr. Johnston’s pre-tax AGI. Certainly it runs the risk of biasing the average via the extremes, but consider the opposite case: Suppose President A in 2010 sponsors a confiscatory tax policy in the name of “fairness” whose effect it would be to place a flat tax of 95% on all incomes. In 2014, President B replaces that flat tax with a 5% rate. Even if the median AGI – pre or post tax – under President A was twice the median under President B, everyone would have more disposable income under the second regime. That’s why it was disingenous of Mr. Johnston to use pre-tax data and minimal deviations and declare that to be the cause of current economic melancholia.
As for the rest, you have hit upon the core of the liberal/conservative economic divide. You are uncomfortable that, while the lower quintiles are experiencing economic growth, the pace of that growth is overtaken by those at the top. Your remedy for this “growing income inequality” I can only presume is some form of redistributionist tax policy. I would argue for an equality of opportunity rather than an assurance of outcome.
So long as all boats rise – or at least, all those not hopelessly mired in dependency, despondency and deficiency – then we all have access to a rising standard of living, and social pressures are eased. Focusing on the fact that some people are doing better – risk takers usually, quite often the people driving the engine of general economic growth which generates those same rising tides – is to play upon the politics of envy while simultaneously killing the goose that lays the egg.
Our entire society excels because smart people who take bold entreprenuerial risks are rewarded commensurately by a free market. Their success in turn fuels general economic activity, not the smothering hand of government attempting to regulate common outcomes.
Economic activity ≠ personal income.
While GDP is an excellent measure of economic activity and national wealth, it does not measure personal income. And while pre- or after-tax metrics each have their limitations, both are far better measures of income than GDP.
Furthermore, personal income for many can certainly decrease during times of increased economic activity and a growing GDP.
Various tax policies, while they certainly have a significant impact upon personal income, are never the overriding or deciding factor. Therefore I believe ‘hypotheticals’ in the tax arena, while important and instructive, are nevertheless if not tangential, then secondary discussions.
However, you are most correct that we are both discussing the “core of the liberal/conservative economic divide.” However, while I certainly advocate some tax reform, your presumption that I view tax reform as “the remedy” is off target.
I’ll save for later some points/counterpoints in the economic liberal vs. conservative debate. And standing alone, your comments make very good sense. But meanwhile, our currently wonderful GDP obscures many underlying problems that should not be ignored.
It is forecast that the US gap between the nation’s rich and poor if current trends continue, will match that of Mexico’s in the not too distant future. If our much greater GDP remains, that will still compensate for the gap, and other problems.
But if GDP drops as it certainly does cyclically, and if it drops precipitously, as it sometimes has in history, we will have major problems stemming from our past and current economic policies.
In looking at the Gini graph, I (facetiously) wondered if like Mexico, our future recession/depression poor would then cross illegally into Canada, where the income gap would be far less, even given their lesser GDP.
You say, “Our entire society excels because smart people who take bold entreprenuerial risks are rewarded commensurately by a free market,” and I wholeheartely agree! But it is not that simple. There are many other important factors to consider.
[Oh yes, regarding geese and golden eggs & class inequality, it reminded me of this, although it is not "exactly" related:-]
I find it telling that Mr. Johnston never addressed the fact that he only looked at pre-tax incomes rather than after tax incomes. I mean it is after tax incomes that we all have to “get by on” isn’t it? Wouldn’t those numbers be a better indicator of how the American people are doing financially?
Ahh, agenda journalism at it’s best worst…
Jim C
that should have read “agenda journalism at it’s worst…
Jim C