The Dog That Isn’t Barking: Why So Little Pundit Attention to the Caliber of Statistics?
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 cataloging 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 isn’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.
This is a chart showing different inflation measures and the oil price for comparison. All prices have been rebased to 1950
The first thing I want to point out is not how different the measures are, but how similar they are. Normally this type of chart is designed to make some kind of political point, but if you take a step back, note how they all track each other. EJ has pointedly made use of the CPI as an effective inflation measure as far as the minimum wage goes. I think we should seriously consider that he might be right. Look at the PPI in the above chart. Note how it consistently is under the CPI since 1950. I take this to mean that there is a consistent markup on consumer goods above cost. This makes intuitive sense, but it seems like the margin has changed over time. For instance in the 1975-1980 period the gap between PPI and CPI closed up. A good check on the effectiveness of CPI is how closely it tracks PPI.
I will wrote more in a followup comment.