Turbulence vs Chaos: the Buddha's Lessons on Data, Distortions & Decisions
One of the things we try and do around here is find the right data, put it thru the right filters and interpretative frameworks without torturing it to say what we want and then use to reach the conclusions and make the right decisions. Over the last five years that approach has served us very well by and large and hopefully helped those who've read this blog or seen our newsletters or other communications. Beyond that when we don't understand something we try and dig into until we find the right data, a useful explanation and even a framework for thinking about it. The last several posts have all served that "tradition" but rather than get back to our knitting we're going to take another example and then try some big picture lessons from it. The example, which is vitally important for its own sake, is the Employment Data. The important lesson is understanding the data sources and interpretations but the critical lesson is the implied one about letting your thinking be distorted by your pre-dispositions. Which we'll come back to you in a bit. Let's dive in on the Employment data.
There are three OMG look at the data ideas floating around the memosphere (meme+MSM+blogosphere) about the Employment data. Now we'll just say that in the last few years this is not the first such tempest that's blown up from conspiracy theories to mal-adjustment in inflation data to....on and on. EVERY single one of them on being checked out was somebody not looking at the data, charting it out and seeing whether or not it made sense or if they had a workable suggestion for making it better at an affordable cost. NOT a single OMG moment has held up in five years. There are three related ones floating around about the Employment data that we want to test in the process of learning something about what's really going on. One is that the Non-seasonally adjusted data shows job increases so it's not as bad, another is that the two different sources of data - the Establishment survey which talks to businesses directly and has a more limited but more accurate sample and the Household Survey which picks up freelancers, consultants, farmers and other n'er do wells like ourselves have a huge disparity and the third is because this economy "is so different this time" that you can't rely on the employment data as a lagging indicator because it's actually a leading indicator. Putting these to rest is almost as important as understanding the data per se. 
Three Gotchas in the Employment Data
Let's start with looking at the Establishment employment data, both seasonally adjusted and unadjusted, and comparing it to the Household survey data and seeing what we see. As usual and always click to enlarge the chart please! The top chart shows SA and NSA Establishment vs Household Employment. The thin red line is the NSA and notice especially that a) it's very noisy constantly wiggling around the SA data and b) there's a constant gap of around 10 million jobs between the two surveys, which once you adjust for the difference in data sources, turns out over the last several decades to yield the same answer as to the state of Employment. At least according to the Fed in several repeated looks they've taken at the issue. If you get what we just said all three of the OMG shibboleths just died on the facts. We could spend the rest of this blog explaining in detail but will pass on.
The second sub-chart shows new jobs, from which we've dropped the NSA data which is SO noisy that the repeated and useless up & down spikes result in a chart too noisy to read and interpret. By and large over nearly three decades the two surveys yield exactly the identical number of new jobs, though there is a slight discrepancy on this quarterly basis in the last quarter. Which is not out of line with previous discrepancies. The third sub-chart shows YoY% changes - SA and NSA are so close that the red and blue lines overlay and turn purple. Compare that inadvertently created line with the Household survey and guess what - again the YoY changes give us just about the same answer again. Though the Household survey is nosier and therefore it tends to be harder to rely on period to period. The bottomline therefore is that YoY SA data gives us the cleanest, most reliable indicator of trends, patterns and turning points.
Continued ...A Deeper Dive in Granularity
Lest you think we palmed a card lets go to the monthly data, which is as fine as it gets and see if we can keep those mis-interpretations properly out of the way. Lo and behold the top sub-chart comparing SA'd Employment (Establishment) and Civilian Employment (Household) seem to be dancing the same dance to the same music, beat and rhythm. You can see the discrepancy in detail that everybody's excited about on the tail of the second sub-chart where EmpSa shows about -200K jobs and CivEmp shows about -600K. Of course earlier in the year it was running the other way and has before in the last decade. Is this mere noise or some deep mal-adjustment? So far it's not out of line for noise.
Which leads us to the third sub-chart which shows monthly YoY% changes between the two. Wow - what wild divergences over a decade. Why it must be approximately zero percent, or so close as to not make much if any difference for our purposes. Just ignore the fact that for most of the last year they've laid over eac other almost exactly please.
As for the third mis-understood (the old KGB tactical term Maskirova comes to mind, or would except with them it was deliberate instead of a failure to do their due diligence) meme that this time it's different gets put to bed. Employment growth peaked around Jan06 and started slowing down almost immediately. Which we widely warned about at the time and were almost as widely ignored about.
Let's Get Serious About the Real Data
Hopefully having put the monsters, at least the imaginary ones, back in the closet it's time to take a look at the real ones with the right data. The top sub-chart compares YoY changes in real GDP, net of trade putting another little data difficulty to rest, with Employment. Notice that GDP stopped improving in Q104, leading Employment by far but also telling us how delayed the last recovery was and how weak and non-organic it was. Again something we tried to tell people at the time. There is a piece of wonderful news however for those of you who recall our comparisons to Eastern Front major battles. We've been talking about this being Stalingrad - well the good news is we won. The bad news it's time to start the intel work and planning for Kursk.
All of which is reinforced in the second sub-chart which shows four different Employment indicators: Employment, Hours, Unemployment and Private jobs. Same 'ol good news, bad news. The bottom chart comes back to that implicit bad news about organic growth, self-sustaining in other words, and jobs. It shows New Jobs, Net New Jobs and Cumulative Net New Jobs. We've discussed before the notion that it takes about 147K jobs/month (~150K), though estimates range from 120-170, to breakeven on labor force plus productivity growth. New Jobs is just that, Net New is New-450K/Qtr and Cumulative is the running total. Here's the really bad news in two parts. Because of the non-organic nature of the last recovery we entered this downturn still about 1.5 million jobs in the hole. Now we're about -12 million jobs in the hole and it'll keep getting worse. In fact Lakshman Achuthan was on PBS, despite his deserved rep as an optimist, and pointed out it would take at least four years of 5% real growth to get back to breakeven which we're not going to get. In fact in one of those previous reality checks (Refreshing the Economic Outlook: Fundamentals to Business Outlook) we put up this chart showing the various relationships and found that to get 2.5% Employment growth we need 5% real GDP growth. At 2.5% GDP growth we get about 1.8% Employment growth. Whether anybody knows it or not that pretty well settles the V-shape vs. U-shape debate, though not in a way we'd like.
Non-organic Recovery and the Drawn-Out U
Just to drive home some other nails in the coffin we'll put up this composite borrowing from our buddies Bill over at CalculatedRisk and Jake at EconPic (we'll confess to a great deal of guilt since Jake and I started cross-talk his posts are getting longer and longer....deeply informative but we still feel like a corrupting influence). Anyway if you think you've seen Bill's chart in the UL corner in the NYT your right - they must've liked it 'cause he thought it up several months ahead of time. Here's the reinforcing lesson - take the current curve, copy it, flip it around as a mirror image in your mind and ask yourself what you get and where it crosses the zero line. We think we'll be lucky on current course and speed if it bottoms before Month30. Applying the flip transformation that tells us zero is at Month60. That's so depressing we'll let you work out the rest of it for yourselves. Consider Jakes' and my charts as repeating and reinforcing the point in different guises.
The Bodhisattvas of Reality
One of the things about the Buddhists we like and think is especially apt for our purposes is that they don't tell you that some authority somewhere has the final and ultimate answers. Instead they tell you to go and see for yourself. Though they do start and add that they've found some basic principles, some techniques for calming and clearing your mind and making sure your decisions are based on un-distorted data and analysis and they don't kid you it's easy or there are magic answers. That might be a summary description of the worst danger we face - running after simple but distorted answers beause they're easy to pretend to understand or suit our unexamined prejudices.
When the Buddhists talk about enlightenment they talk about a Buddhist Saint who has chosen to remain in the world after reaching Nirvana, which for our purposes we take as always making un-deluded decisions without being led by their emotions. They stay behind with the rest of us poor slobs to serve as examples and coaches. Well we won't go that far here but it's a comparison worth bearing in mind with the last two must-read readings. One is Janet Yellen's take on the economic outlook, which strangely enough, rather closely resembles the one you've just read thru. The other is a superb, and we mean superb, piece by Paul McCulley on explaining why it's U-shaped not V-shaped but why that's consistent with the surge in asset prices. Ahem - needless to say his much clearer and well-written explanation bears some passing resemblence to several of our recent discussions.
The graphic is Ganesha the Hindu God of Wisdom among other things, taken from another earlier post (Ganesh Filter II: Clear-seeing Algorithims for an '08 Plan) and it seems appropriate to quote from an earlier post of ours on the topic of not letting your thinking get distorted:
Seeing thru Maya: Piercing the Veils of Delusion
The Buddhists (& the Hindus) talk about Maya the "Veil of Illusion" and suggest that we don't see the real world underneath the surface. Now for a while, and still, there was a tendency in the West to do some sort of mystical mapping between quantum physics and so forth to suggest that was what's being talked about. Or alternatively it was a mystical point and to be dismissed. In any case the critics, pro and con, took it as arm-waving philosophy.
When you dig into the propositions are much simpler, harder and useful. What they really mean is that we tend to look at the world as prisoners of a whole host of assumptions and presumptions - to see the world as we'd wish it to be instead of how it is. Hence we view the world thru a distorting veil of delusions brought about by our own bad thinking.
The recommended cure is to both learn to see the world and to learn how to think. To see the world the primary tool is to learn how one thing leads to another and that to another and so on in a cause-and-effect chain. Or if it's complex enough web of linkages and feedbacks. In other words Maya is not mystical it's talking about having a bad model and not knowing how, or choosing not to, fix it. Which is the other part of the prescription - learn how to think and make up your own mind. The Buddhists call that "Right Thinking".
Well the common thread running all throughout the four stories, or at least so it seems to me as it finally dawned, was each was the result of distorted thinking, bad models and delusions about what was really going on. Delusions all the worse for having gotten so much intellectual and emotional investment that the commitment to denial was pretty pronounced. The problem with Reality is that sooner or later the tide comes in, water doesn't run up hill, business cycles cycle, profits decline and bad due diligence and poor operating performance get their entropic rewards.
Actually with Prof. Ben's announcement that they're going to do everything they can to prevent a recession, though it's too late to stop it and now the goal is to mitigate it before it cracks all the fault lines we've been talking about (just yesterday and for months now was it :) ), we may have a fifth good example. Watching the markets gyrate back and forth where's the surprise? And what changes about the underlying realities?=======================================================================
Employment, Data & Outlook
Two PBS Newshour vidclips:
- Unemployment Hits 26-Year High Despite Economic Growth The U.S. Labor Department announced that unemployment jumped to over 10 percent on Friday, the highest it's been since 1983. Jeffrey Brown talks to an economist for more.
- Freelancers Struggle As Unemployment Worsens in U.S. As part of his Making Sen$e series on the financial crisis, Paul Solman looks at how freelancers are faring in the declining job market.
US Economy: Unemployment Rate Jumps to 26-Year High The unemployment rate in the U.S. jumped to 10.2 percent in October, the highest level since 1983, casting a pall over the prospects for a sustained recovery and risking further erosion of President Barack Obama’s popularity. Payrolls fell by 190,000 last month, more than forecast by economists, a Labor Department report showed today in Washington. The jobless rate rose from 9.8 percent in September. Factory payrolls dropped by the most in four months, and the average workweek held at a record low. Treasury two-year notes rose on bets the Federal Reserve is more likely to maintain its pledge to keep interest rates near zero. The figures prompted Obama, who signed a bill today extending jobless benefits, to promise fresh measures to help put some of the 15.7 million unemployed Americans back to work. “We will certainly have very bad payroll numbers in November and December,” said Harm Bandholz, an economist at UniCredit Global Research in New York, whose forecast for a 10.1 percent unemployment rate matched the highest among economists surveyed by Bloomberg. “We don’t foresee businesses going on a hiring spree anytime soon.” Payrolls were forecast to drop 175,000 after an initially reported 263,000 decline for September, according to the median estimate of 84 economists surveyed by Bloomberg News. The jobless rate was projected to rise to 9.9 percent.
- White House’s Bernstein Says Ideas Weighed to Boost Job Growth
- Soaring US Unemployment Threatens Path to Economic Recovery
- Employment Report: 190K Jobs Lost, 10.2% Unemployment Rate
- Employment-Population Ratio, Record Part Time Workers, Weak Holiday Hiring
- Unemployment: Stress Tests, Unemployed over 26 Weeks, Diffusion Index
- Broader Unemployment to 17.5%
- Job Losses... Again, Worse than Reported
- Civilian Hours per Week Cliff Dive Continues
- Civilian Hours vs. Real GDP
Economists Seek to Fix a Defect in Data That Overstates the Nation’s Vigor A widening gap between data and reality is distorting the government’s picture of the country’s economic health, overstating growth and productivity in ways that could affect the political debate on issues like trade, wages and job creation. The shortcomings of the data-gathering system came through loud and clear here Friday and Saturday at a first-of-its-kind gathering of economists from academia and government determined to come up with a more accurate statistical picture. The fundamental shortcoming is in the way imports are accounted for. A carburetor bought for $50 in China as a component of an American-made car, for example, more often than not shows up in the statistics as if it were the American-made version valued at, say, $100. The failure to distinguish adequately between what is made in America and what is made abroad falsely inflates the gross domestic product, which sums up all value added within the country. American workers lose their jobs when carburetors they once made are imported instead. The federal data notices the decline in employment but fails to revalue the carburetors or even pinpoint that they are foreign-made. Because it seems as if $100 carburetors are being produced but fewer workers are needed to do so, productivity falsely rises — in the national statistics. The statistical distortions can be significant. At worst, the gross domestic product would have risen at only a 3.3 percent annual rate in the third quarter instead of the 3.5 percent actually reported, according to some experts at the conference. The same gap applies to productivity. And the spread is growing as imports do.
That may help to explain why the recovery from the 2001 recession was a jobless one for many months and why the recovery from this recession is likely to generate few jobs for many months. In addition, more detailed import data would help to explain wage inequality, by linking some low wages more accurately to particular industries exposed to import competition. On another front, many argue that labor productivity is rising faster than the pay of workers who made the greater productivity possible. That argument would be watered down if more accurate data showed that productivity had been overstated. Grappling with these blind spots, nearly all of the 80 experts at the conference, which was sponsored by the Upjohn Institute and the National Academy of Public Administration, agreed that the statistics now published tend to overstate the strength of the economy. That view was shared by those who attended from the Bureau of Economic Analysis, the Bureau of Labor Statistics and the Federal Reserve, all big players in measuring economic performance.
Memosphere Disinformation and Mis-understandings
Seasonally Adjusting Unemployment Floyd Norris points us to one of the foibles of seasonal adjustments: How SA can make things look better or worse at various times of the year. He is referring to the Non-Farm Payrolls data from Friday. Comparing non seasonal apples to apples, the data was somewhat better than usual.
The Glide Path Option The headlines said unemployment, as measured by the "establishment survey," was down by 190,000; and even though that was slightly worse than forecast, market bulls were cheered by the fact that the number was not as bad as last month's. It is an improvement that we are not falling as fast. Well, maybe. What I did not see in many of the stories I read was that the number of unemployed actually soared by 558,000, to 15.7 million, as measured by the household survey. The establishment survey polls larger businesses; the household survey actually calls individual households. Let's look at the real number in the establishment survey. If you don't seasonally adjust the number, the actual change in unemployment for October was 641,000, or about 450,000 more than the seasonally adjusted number.
Real Reality Checks: the Bodhisatvas Speak
Fed's Yellen on the Economic Outlook Lately I’ve been leaning against the view of a “V shaped” recovery. I think that growth will be decent in the second half of 2009, but growth will be sluggish in 2010. San Francisco Fed President Dr. Yellen has a similar view – from her speech this morning: The Outlook for the Economy and Real Estate;
The big issue is how strong the upturn will be. With such enormous reservoirs of slack in the form of high unemployment and idle productive capacity, we need a strong rebound to put unemployed people back to work and get underutilized factories, offices, and stores humming again. Unfortunately, my own forecast envisions a less-than-robust recovery for several reasons. As the impetus from government programs and inventories diminishes in the quarters ahead, private final demand will have to fill the breach. The danger is that demand may grow at too anemic a pace to support vigorous expansion.
The uncomfortable dance between V’ers and U’ers, …. while rich risk asset prices can certainly be viewed as a consensus expectation for a strong recovery, such lofty valuations can also be viewed as a consensus expectation about the Fed’s commitment to erring on the side of being too late, rather than too early, in starting a Fed funds tightening cycle. Indeed, one could actually be agnostic, even antagonistic, about a big-V recovery and still be favourably disposed to risk assets, in the short run. Historically, what pounds risk asset prices is either a recession or unexpected Fed tightening; or worse, both. Right now, it is hard to get wrapped around the axle about recession, since we’ve just had one, which might not even be over. Thus, as long as economic recovery appears underway, even if stoked primarily by (1) policy stimulus and (2) a turn in the inventory cycle, there is no urgent reason for investors to run from risk assets. Put differently, investors can be agnostic about (3) the strength of private demand growth until the one-off forces supporting growth exhaust themselves, as long as they don’t have fear of Fed tightening. In turn, a bull flattening bias of the Treasury curve, with longer-dated rates falling toward the near-zero Fed policy rate, can be viewed as a consensus view that the level of the output/unemployment gap plumbed during the recession is so great that disinflationary forces in goods and services prices, and perhaps even more important, wages, will be in train, even if growth surprises on the upside. Accordingly, Treasury players, like their equity brethren, need not fear the Fed, as there is no economic rationale for an early turn to a tightening process. Thus, both rich risk markets and the lofty Treasury market can be viewed as rational in their own spheres, even if they are seemingly irrational when compared to each other. The tie that binds them, that allows them to co-exist, need not be a common view regarding the prospective strength of the recovery, but rather a common view as to the Fed’s friendly intent and reaction function.
The Shadow Banking System and Hyman Minsky’s Economic Journey”

