Recent Items

The Dog That Isn’t Barking: Why So Little Pundit Attention to the Caliber of Statistics?

Posted on by

Ah, the halcyon days of early 2007, when economics and finance bloggers would study the clouds on the horizon and debate what they foretold. Maybe I’m not hanging out in the right circles these days but now that financial markets seem to be completely in thrall to central bankers, there isn’t much point in doing fundamental analysis. As a result, from what I can tell, the level of bullshitting among market pundits has risen considerably.

Now I am using “bullshit” in the Harry Frankfurt sense. Frankfurt drew a distinction between lying, which is choosing to deceive someone, while bullshitting has no regard for accuracy. A bullshitter may coincidentally be speaking the truth or be utterly and hopelessly wrong, but the key is the coincidental nature of any relationship to the truth. The bullshitter says what he says and where and what the truth is is irrelevant to him.

For instance, one of the things you’ll hear regularly (more like all the time) is how terrific corporate earnings are. Now on the one hand, corporate earnings have hit the highest proportion of US GDP in recorded history. But when stock touts are talking about corporate earnings, they mean of public companies. S&P 500 earnings peaked in the first quarter of 2012 and have fallen each quarter since then. Third quarter 2012 S&P 500 earnings were 6.3% below the year earlier level.

On the government statistics front, I wonder if the decline in poking and prodding of the official statistics has to do with a counteroffensive against John Williams, who produces ShadowStats. Williams was in vogue in 2007, and has done a very good job of cataloguing the various ways official measurements have changed over time. He also produces his own version of various measures, and that left him open to attack, since some of his approaches (like simply adding 4% to the official Consumer Price Index derived inflation rate) were a bit wanting in precision. He became enough of a burr in the side of the officialdom that statisticians from the Bureau of Economic Advisers were dispatched to some economics conferences to defend their methods (I hardly get around, and they showed up at a conference at which I was speaking). The fact that Williams may have been overly broad in some of his criticisms does not obviate that many of them are generally correct (for instance, most economists will concede that CPI is too low, although a typical estimate is on the order of 0.5%)

In keeping, experts and financial markets largely shrugged off the 0.1% decline in fourth quarter GDP as largely the product of one-offs, mainly a fall in government spending. But comparatively little attention was paid to a component that used to be more closely watched, that of the GDP deflator. The smaller the deflator, the bigger the GDP figure. The deflator was an implausible 0.6%. Using the Personal Consumption Expenditures as the measure of inflation would have lowered GDP to -0.8%. Using CPI would have brought it in a -1.6%. One hedgie buddy further contends that only 1/3 the inputs to GDP are based on actual measurements; the rest is imputed. He argues that the measured 1/3 has not returned to 2007 levels, making him skeptical about the rest (he cited Andrew Hunt, but I have been unable to locate it on Google, so reader confirmation or qualification would be useful).

Similarly, this blogger and Sober Look have taken note of the fact that in each of the last three years, the economy strengthened in the first quarter and then the “recovery” petered out. Regular NC commentor and sometime guest blogger Hugh looked under the hood and has more cheery observations:

….note the second graph with the national data from Markit’s US PMI. The tenor of the article is that there have been similar spring peaks in business expansion in each of the last 3 years. But the seasonally unadjusted data for this year is considerably worse this year than the the previous 3. As I keep pointing out, the unadjusted data is where we and the economy actually are at any given point in time. The adjusted data is a trendline. So it is not surprising that the trend is going up in 2013 because that’s what happened in 2010-2012. That is the adjusted (trend) line is going up in 2013 precisely because that’s what it did in spring 2010-2012. In 2010-2012 the trend was being driven and supported by a spike in the unadjusted numbers. The difference this time is that it isn’t. That is there is no support from the unadjusted data (where the economy is) for the spike in the adjusted trendline. I would consider this quite worrisome.

Now as we indicated, inattentiveness about the real state of the economy isn’t entirely irrational given that the top 1% is doing just fine and the Fed seems determined to keep a floor under asset prices. So why worry when the entire world has become one big carry trade?

One problem is getting bad information signals, or what the folks at Lazard in its heyday called “believing your own PR”. We published an analysis by Paul DeGrauwe yesterday (which some readers took aim at) that mentioned an theory that some economists actually subscribe to, namely, that markets reflect fundamentals. Of course, if they really did, we’d never have bubbles. Yet central bankers (perversely) seem to be looking at firming prices of financial assets as a sign of improving economic confidence, when it looks to be primarily, if not entirely, confidence that Uncle Ben and his friends will not let Mr. Market down. And if you believe Mr. Market is saying nicer things about the real economy than he really is, perversely, Uncle Ben and his friends might start withdrawing life support.* The horror!

_____
* For the record, we’ve long said that the use of ZIRP is a lousy substitute for fiscal stimulus (and “lousy substitute” is too kind, but we don’t need to go into a long form discussion here). But the Fed deciding the economy is getting better when it ins’t much or at all isn’t just reflected in its actions, but also its various communications, which influence all sorts of other policy makers. So the Fed in its role as leader of groupthink means this sort of perverse informational feedback loop is even more pernicious than it might appear to be.

Print Friendly
Twitter11DiggReddit0StumbleUpon0Facebook15LinkedIn0Google+1bufferEmail

31 comments

  1. Lafayette

    QUO VADIS HUMANUS ECONOMICUS?

    Ah, the halcyon days of early 2007, when economics and finance bloggers would study the clouds on the horizon and debate what they foretold.

    Maybe they were barking up the wrong tree? Far too many complaints were made, after the Toxic Waste Mess, that economics had “failed to predict it”. Nowadays, some economists are talking about the “Minsky Moment” to explain what happened.

    Quo vadis?

    To my mind, Toxic Waste was simply a classic failure on the part of finance market-oversight agencies (namely the Fed and the SEC) to do their job correctly. They were both negligent in their oversight duties, having been much too lax. Both or either should have been investigating the underlying nature of the loans that were escaping any competent creditworthiness verifications.

    The originators of the fraudulent loans should have all been given perp-walks. By means of which law enforcement? How about the Truth In Lending Act of 1968? All it took was about ten perp-walk amongst the Main Street people who were originating the loans mediatized across the nation on the nightly news – and, of course, subsequent convictions.

    What can economics do in terms of prediction when a Fed Head, with brainless Ayn Randian notions of human behaviour, utters such inanities as (and I paraphrase) “the markets are run by intelligent people and therefore cannot possibly go fail”? No one ever called Greenspan out on his lack of oversight diligence. In fact, most kissed his a** and hailed him as “the divine Mister G”.

    The finance markets where being manipulated by clever Golden Boys ‘n Girls with one debilitating passion motivating them: Lucre.

    MY POINT?

    Economics (as a profession) should be telling the world how to prevent a repeat of that debacle, based upon what we know from the past. Otherwise, the profession should simply content itself with predictive GDP-models.

    One good thing on the horizon: The Europeans will be instituting the Tobin Tax on financial transactions, which will take a bit of the mojo out of electronic trading.

    1. fresno dan

      Good points.
      It seems to me that economists are unable or unwilling to acknowledge that fraud and corruption can make up a substantial (if not majority) of economic activity. Until economists finger out that the major players have their fingers on the scales, all of their formulas miss the point.

  2. rjs

    this post sure hits one of my sore spots..

    one major problem i have with almost all the reporting on econ data is tha no one recognizes or reports on the margin of error, or as it’s often referred to in technical notes as “the 90% confidence range” – just take last weeks January retail sales; everyone reported them down 0.1% as if it were case in stone; here was my take:

    this report should have given us the first indication of how the year end expiration of the 2% payroll tax cuts, which after income taxes reduced the January take-home pay of most US workers by more than 2%, would affect consumer demand…however, with the large margin of error (±0.5%) in this months report, likely coupled with the difficult december to january seasonal adjustment for this data, it’s generally inconclusive; the footnote informs us that census does not have sufficient statistical evidence to conclude that the actual change is different than zero..

    same thing with unemployment data…the 90% confidence range on non-farm payrolls is ±100,000, and over ±400,000 on the household survey….yet a report showing a gain of 180,000 jobs will be praised as good, and one showing a gain o 130,000 will be seen a week…neither tells us diddly…

  3. rjs

    this post sure hits one of my sore spots..

    one major problem i have with almost all the reporting on econ data is tha no one recognizes or reports on the margin of error, or as it’s often referred to in technical notes as “the 90% confidence range” – just take last weeks January retail sales; everyone reported them down 0.1% as if it were case in stone; here was my take:

    this report should have given us the first indication of how the year end expiration of the 2% payroll tax cuts, which after income taxes reduced the January take-home pay of most US workers by more than 2%, would affect consumer demand…however, with the large margin of error (±0.5%) in this months report, likely coupled with the difficult december to january seasonal adjustment for this data, it’s generally inconclusive; the footnote informs us that census does not have sufficient statistical evidence to conclude that the actual change is different than zero..

    same thing with unemployment data…the 90% confidence range on non-farm payrolls is ±100,000, and over ±400,000 on the household survey….yet a report showing a gain of 180,000 jobs will be praised as good, and one showing a gain o 130,000 will be seen a week…neither tells us diddly.. .

    1. rjs

      the actual retail sales for January were estimated to be $382.9 billion, down from the $469.1 billion sales estimate of December, so you can see what a stretch it is to say with any certainty that seasonally adjusted sales were up 0.1%…

      1. rjs

        all the data on new homes from the census is virtually worthless, yet the major blogs and MSM report it as if it were gospel; here’s the story on new home sales for december:

        …the census reports “Sales of new single-family houses in December 2012 were at a seasonally adjusted annual rate of 369,000…7.3 percent (±15.3%)* below the revised November rate of 398,000″
        virtually all reports on this data focus on just the annualized number and the monthly change without mentioning the large margin of error…the actual census estimate for homes sold in december was a much less impressive 26,000, with 15,000 of those in the South, and 5,000 in the West…census takes that rough estimate of 26,000 and plugs it into a program which compares it to other recent Decembers and computes what annual sales would be if that december sales level were continued through the entire year; so when they say “sales were (±15.3%)* below the revised November rate of 398,000″ they mean they are 90% confident that the seasonally adjusted annual rate of new home sales would fall someplace between 337,106 and 458,894, or that it’s fairly likely that the monthly change in new home sales was somewhere between a gain of 8.0% and a decline of 22.6% at a seasonally adjusted annual rate…the comparison to last December’s home sales is no more useful: “(December 2012′s seasonally adjusted annual rate) is 8.8 percent (±24.8%)* above the December 2011 estimate of 339,000″ that’s almost a range of 50%! about the only new homes sales data from this report that we can bank on is annual estimate for all new homes sold in 2012: “An estimated 367,000 new homes were sold in 2012. This is 19.9 percent (±4.8%) above the 2011 figure of 306,000.” so it’s reasonable to say new home sales for this year were roughly 20% above last year, which was the worst year in the half century of census record keeping..

  4. fresno dan

    I read an artile by Schiff.
    http://www.realclearmarkets.com/articles/2013/01/14/exposing_paul_krugmans_low_inflation_propaganda_100086.html

    Schiff made the claim that health insurance is merely 1% in determining the cost of living and I thought Schiff surely must be taking a view that was out of context.

    “Although medical insurance premiums are an important part of consumers’ medical spending, the direct pricing of health insurance policies is not included in the CPI. As explained below, BLS reassigns most of this spending to the other medical categories (such as Hospitals) that are paid for by insurance. The extreme difficulty distinguishing changes in insurance quality from changes in its price forces the CPI to use this indirect method.”
    http://www.bls.gov/cpi/cpifact4.htm

    I would also note that my deductibles have gone up 50%, as well as the fact that my 60 seconds spent with the physician has been reduced to 30 seconds. on the other hand, the leaches do seem to be more effecitve and efficient at getting more blood out of me… ;)

    1. Ms G

      “Although medical insurance premiums are an important part of consumers’ medical spending, the direct pricing of health insurance policies is not included in the CPI.”

      Sheesh. No wonder CPI has always seemed (even to the ignorant but observant lay observer) to be about (at least) 5%? 10%? 20%? 30%? Cool trick, though, to just not count what is perhaps the biggest inflationary item in any average american’s “budget.” Do they ignore the price of oil, food and school too?

      That would start to explain why the CPI numbers have always seemed like a big joke at the expense of working people!

  5. Thorstein

    A wise programmer once said, “For any large computer system, it must be admitted that nobody truly understands how it all works.” We were fools, those of us who devoted our education and our working lives to understanding in any detail our own small component of a large computer system (or the banking system, or a large corporation, or …). Since nobody could ever really understand how the whole thing would work (or what the consequences would be), the smart money wined and whored and cheated their way through Harvard, never intending to cash in on anything but bullshit. For the most part, these are the “can do” types who have ascended to the command or our Babel. The ones who could do bullshit. Of course these types have always been with us, but when confronted by modern complexity, they do not constrain themselves with logic or conscience.

      1. Ms G

        They’re the ones who always look around the room and, in the manner of a jr. high school athletic coach, like to shout: “Ok folks, let’s roll up our sleeves — all hands on deck…” etc., etc.

  6. from Mexico

    Daniel Ellsberg is without a doubt one of the greatest heroes of American history, because due to his formidable courage, we now know the nuts and bolts of how the public is lied to. In “Lying in Politics,” Hannah Arendt takes a look into what the Pentagon Papers revealed of our ruling overlords.

    Arendt was gobsmacked at “the extravagant lengths to which the commitment to nontruthfulness in politics went on at the highest level of governement,” the “extent to which lying was permitted to proliferate throughout the ranks of all governmental services, military and civilian — the phony body counts, the doctored after-damage reports, the ‘progress’ reports to Washington by subordinates who knew that theier performance would be evaluated by their own reports.”

    And as Arendt goes on to point out, lies are very beguiling: “Lies are often much more plausible, more appealing to reason, than reality, since the liar has the great advantage of knowing beforehand what the audience wishes or expects to hear. He has prepared his story for public consumption with a careful eye to making it credible, whereas reality has the disconcerting haibt of confronting us with the unexpected, for which we were not prepared.”

    There is of course nothing new about lying in politics. It’s been around througout the millenia. But there are a couple of new twists, Arendt says. The first one is the public relations managers who learned their trade from the inventiveness of Madison Avenue. Their psychological premise of human manipulability “has become one of the chief wares that are sold on the market of common and learned opinion.” It’s easy to see how politicians could fall for this spiel that the masses are mere putty in the hands of the engineers of consent, because the public relations managers play such a key role in getting them elected.

    The second one is the technocrats, or what Arendt calls the “problem-solvers.” These are the unelected experts and advisers that surround, for instance, the president. They include his cabinet members — McNamara being the quintessential example — and a vast Washington bureaucracy who “exercise their power chiefly by filtering the information that reaches the President and by interpreting the outside world for him.” They too believe that politics is but a variety of public relations. “What these problem-solvers have in common with down-to-earth liars is the attempt to get rid of facts and the confidence that this should be possible,” as though “a fact is safely removed from the world if only enough people believe in its nonexistence.” Confident “of place, of education and accomplishment,” they lie with abandon. They created a defactualized world where policial goals were set and decisions were made, and “the relation, or, rather, nonrelation, between facts and decision…is perhaps the most momentous, and certaintly the best-guarded, secret that the Pentagon papers revealed.”

    “It is this remoteness from reality that will haunt the reader of the Pentagon papers,” Arendt says. “The bureaucratic model had completely displaced reality: the hard and stubborn facts…were ignored.”

    1. ambrit

      Dear from Mexico;
      Thanks for the reference to Arendts’ book on the Pentagon Papers. I haven’t read it; now I shall hunt it down and do so.
      These observations are equally applicable to the world of retail business. Those at the top are demonstrably out of touch with reality; the mass of complaints retail floor workers deal with on a daily basis have taken on the quality of a chorus in a Greek tragedy. “The gods defend us, hubris has gripped the mighty.”
      It all reminds me chillingly of the Twenties mantra: “The business of America is business.” That didn’t end well. Neither shall this. I’m starting to drift in the direction of Survivalism. Anyone else get that feeling? This being a ‘leading indicator’ type of commentariat, the groupthink from around here would be informative.

    2. McMike

      Indeed, lying is not new.

      What is new is the widespread pervasiveness of it (along with indeed the scientific sophistication of the lying).

      It is the final stage of simulacra. At this point, the only thing that comes out of Washigton is lies. Effectively every single statement that comes out of anyone anywhere at any time is a manufactured falsehood. Lying has pushed out truthfullness to the point that the original no longer exists. Lying is not the exception, it is not even the majority, it is the ONLY thing. As the saying goes, a gaffe is now considered telling the truth.

      What is remarkable for me is how the media and the public has gone along with the transition – the Invasion of the Truth-Snatchers – so we all tune in and point cameras at spokespersons and Pentagon generals and SOTU speeches, night after night, even though it is a certainty that every single word they utter is a constructed falsehood, and that will, sooner or later, likely be contradicted, if it even stands to scrutiny the moment it is uttered.

      This willful charade we participate in, in fact we devote a great deal of energy into validating the artiface, this kabuki, that’s what amazes me.

      I know, Orwell and Huxley saw this coming. No repetitious cycle of self-contradiction, exposure, or scandal interrupts the sacred nightly theater of lies distributed via the news and punditry.

      1. Bunk McNulty

        Sissela Bok’s Lying: Moral Choice in Private and Public Life is still in print and worth reading for her discussion of the ways in which deception is inimical to a functioning society.

  7. RueTheDay

    I’m glad to see someone else highlighting last quarter’s unexpected drop in GDP. I think it’s a bigger deal than the media have let on. The typical explanations about it being due to a reduction in government spending or Hurricane Sandy ring hollow for a number of reasons. Neither the government spending reductions nor Hurricane Sandy were unknowns when the figures were released on Jan 30, yet the consensus forecasts were for 1.0% GDP growth. The GDP slowdown was almost global in nature, hitting virtually every country in Europe as well as Japan and others, and again, to a largely unexpected degree. Did Hurricane Sandy cross the Atlantic or something? Finally, it’s not just GDP, a number of other leading indicators from housing starts to factory orders to industrial production have been coming in below forecast. I wouldn’t quite go so far as to predict another global recession just yet, but there is no doubt that a slowdown has occurred over the past 6 months.

  8. Chris Engel

    The advance figures from Q4 may revise upward in the prelim/final (Just like Q3 did).

    not too worried about GDP unless sequester goes through.

    I’m more worried about unemployment. Even if growth picks up, Okun’s law may not hold given the nasty structure of the current economy that ignores a substantial portion of the market participants’ health/wealth.

  9. Hal Roberts

    Sarbanes – Oxley Act looks good on paper and it seems would go a long way to fixing the problem, if only all the regulators weren’t involved in the markets. Then there would be no conflict of interest. The top 10% see it as cool to turn a blind eye in this late day and time form the SCC to the DMV. It’s all about the retirement funds and false profits. Na say it ain’t so. lol

    1. McMike

      SOX has already been undermined by selective non-enforcement. Between SOX, SEC regs, and RICO, we had all the tools we needed, assuming (naively) they would be used

      1. Ms G

        “undermined.”

        Based on what one has (not) heard, a better word might be “tossed in the trash bin.” Has it even been used one time against any CEO, CFO, etc.?

        Or did some little clause inside the FrankenStatutes of Dodd-Frank and ObamaCare quietly just repeal it?

  10. McMike

    It’s Godwin’s Law meets Moore’s Law.

    The more that the media and punditry proliferates, plus the shorter the attention spans shrink, equals the more likely that information will be merely regurgitated or manufactured from thin air.

    Godwin: As an online discussion grows longer, the probability of a comparison involving Nazis or Hitler approaches 1.

    Moore: Over the history of computing hardware, the number of transistors on integrated circuits doubles approximately every two years.

    Law of Moore-Godwin: B = P/(A*C)

    where: B = bullshit; P = scope of punditry; A = attention spans; C = particpants without conflicts of interest.

    As punditry expands, attention spans shorten, and credible mechanisms for independent scrutiny decline, the proportion of bullshit in the media/internet approaches 1.

    As proof of this, I got the quotes above from Wikipedia, and did nothing to verify them.

    1. jurisV

      Many thanks McMike for advancing the science and math of the original Bullshit Baffles Brains theory — IOW the famous “B-cubed” equation. Actually, no proof (not even internet citations much less Wiki) is needed as it is all around us, and even those of us the lowly masses understand its import.

      Onward to “McMikes Law” …….

  11. Max424

    We have entered the Age of Perpetual GDP Contraction. What does this mean? Well, to lil’ ol’ me, it means terrible things, like, the Age of Slavery is just around the corner.

    But to more important others, it just means the Age of Ever Increasing Corporate Profits* rolls ineluctably onward.

    *Also known as, the Age of the Status Quo

  12. Roland

    Yves, thank you for this post. I am happy that you’re writing about inflation.

    One thing I would find interesting would be differentiated inflation figures for households surveyed within different age and income bands. When we have a class-differentiated society, we also end up with very different consumption patterns and hence different experiences of inflation.

  13. Hugh

    It’s pretty clear most media people never read the tables or try to analyze the data. The press releases, essentially executive summaries, that accompany reports are usually about 1-3 pages in length. For the nightly news, it probably runs something like this, the correspondent either glances through the press release or gets an underling to highlight the important bits, goes to the rolodex to set up a soundbite with one of the usual “experts”, adds in something about Administration reaction, and with that they have enough for a 30-60 second report. A trained monkey or a computer algorithm could do as much. Cable and the newspapers don’t do much better.

    What you don’t get, as rjs above points out, is any idea of how the data is put together, its precision, what is actually being measured, and what the significance of it is.

    Deficiencies in all these areas are institutional in nature. One reason confidence levels are so wide in so much data is that the Census Bureau collects a lot it and it is chronically severely underfunded. It’s hard to see that as an accident. We are a big country with a big economy yet somehow we can’t afford to spend the money to know with any accuracy what is actually going on in the economy.

    I am most familiar with jobs data. It looks structured to not give an accurate view of how we are doing. Did you know for example that uniformed military (like those in prison) don’t exist in terms of the jobs and employment surveys? They aren’t counted in either the working age population (the potential labor force) or in the actual labor force.

    Unemployment is always understated because it only counts those who are active job seekers. If you don’t have a job and haven’t looked for one in the last 4 weeks before the survey, you are no longer employed but out of the labor force.

    Population growth is seldom taken into account. Economists tend to look at jobs needed to return to pre-recession levels ignoring that the population has been growing throughout the period.

    Job quality is never directly addressed.

    The two surveys, establishment and household, do not cover quite the same populations, but are often reported as if they were independent of each other. That is both jobs and unemployment increasing is often reported without comment even though the two increasing are usually a contradiction.

    Again politicians love this because depending upon their party they can praise or decry whichever number best serves their purposes.

    1. McMike

      Indeed. Every time the media celebrates a 100,000 new jobs added to the economy, they inevitably fail to point out this still lags population growth.

      Dean Baker’s blog Beat the Press is a great place to review snarky updates on the endless deficiencies of MSM economic reporting.

  14. JTFaraday

    “So the Fed in its role as leader of groupthink means this sort of perverse informational feedback loop is even more pernicious than it might appear to be.”

    This makes me think of a lot of people’s support for the D-Party throughout the 2000s. They no longer believe what they’re saying, but they find their gums still flapping. I should know because I did this myself for a while. I’m basically a good knee jerk liberal.

    What else does BB have to say? That helicopter already flew. I’m surprised he’s still there. If I were him, I’d get out.

    It’s not like anyone who replaces him is going to come in and immediately contradict him, just as Timmy’s replacement will surely run the country just like he ran NYU. ie., in line with the current consensus, (the way God intended it):

    http://chronicle.com/blogs/bottomline/senator-grassley-grills-treasury-nominee-about-past-job-at-nyu

Comments are closed.