31 July 2019

Artificial Energy

Artificial energy
Is wrestling my mind
I've got a strange feeling
I'm going to die before my time
-- Chris Hillman, Roger McGuinn, and Michael Clarke

OK, so a bit of misdirection. As if missive titles in these endeavors ever did that? A piece on Stat News contains this nifty graph:

Healthcare is a prime venue for AI?

So, why might that be? Kind of simple: what passes for much of what is called AI is just minimized squared differences in many ways. Computers are really good at that sort of thing. Given that health diagnosis is all about mapping myriad symptoms to a single cause, doesn't massive correlation analysis have a chance to top human experience and cognition? Of course it can. The thing about health diagnosis is that it might appear to violate these endeavor's prime directive: correlation analysis, esp. time series version, dealing with human venues is a waste of time. Just read the New York Times and Wall Street Journal and professional journals. Right? Well, no. These data are generated by Mother Nature, so it's merely a matter of sufficient understanding of Her rules. Humans are the target of the data generating process, not the source. Which process, in most of the remaining dark corners of medicine, is really dark.

27 July 2019

Boeing Boeing - part the second

Some further and missed thoughts from part the first on the Max problem. Is the Worst Case Scenario likely?

First, as reported variously, MCAS was not designed to correct pitch-up in climb mode. The issue was discovered in flight testing, and MCAS was handed the responsibility.

Second, I let my fingers do the walking through the Yellow Googles, and found that 'let the software fix the hardware' attitude was come in for some scathing rebuttal. From engineers and pilots, of course. Here's one example. I find this particularly telling
Ultimately, Travis also bemoans what he calls "cultural laziness" within the software development community that is creeping into mission-critical systems like flight computers. "By laziness, I mean that less and less thought is being given to getting a design correct, and simple — up-front," he wrote. "What needs to happen, I think, is for liability to accrue where it is generated."

If that sounds like something you've read in these missives, you're right. Being a die-hard RM advocate, which necessarily means 'measure twice cut once', I've long railed against the mode of writing code in the debugger. That works, of a fashion, with purely software systems. Not so much in aeroplanes.
Travis is unequivocal in his assessment of the Boeing 737 MAX. "It's a faulty airframe. You've got to fix the airframe [and] you can't fix the airframe without moving the engines" back and away from their current position.

Clearly, the more time passes without a (chagrined) FAA re-cert of the Max with a software-only fix, the more likely it becomes that FAA has to say to Boeing, "get those fucking engines off the tarmac!!" And the means, at least, a complete wing re-design. Will Boeing bother? Or will they just go back to vending the 737NG? We'll just have to see.

26 July 2019

Well, D'Oh!!

OK, another admission today that neither the pundits nor policy makers nor press get it
Central bankers around the world, including in the US, are also struggling to understand why inflation has remained muted, propelling officials to move toward cutting rates to get ahead of any economic calamity.

Gad. You have to have more moolah chasing output today than you had yesterday to get inflation. We know, without doubt, that CPI/PPI inflation is near non-existent and has been for some years following The Great Recession. We also know, without doubt, that the moolah pushed into the USofA (and global) economy by Washington has gone to the 1% predominantly. We also know, without doubt, that asset inflation has been raging the whole time.

So, here's the deal:


25 July 2019

Boeing Boeing

No, not the silly movie from the 60s starring the 707 (which, by the bye, the US taxpayer paid for because Boeing just took the KC-135 and added windows and seats), but the current issues with the 737Max. Reporting all over the place today that Boeing may stop making any. Until it's approved, again. What none of the usual pundits have asked is, what's the Worst Case Scenario? I'll ask.

First, why is the 737 such a funny looking plane? It's got dwarves' legs, aka landing gear that's obviously shorter than any other commercial aircraft. How come? Turns out that the original plane was intended to be usable at the smallest airports, typically without jetports, those moving articulated passageways from plane to terminal. IOW, airports where passengers have to walk off the plane to the tarmac. And, worse, airports that don't have a fleet of movable staircases. What to do? Boeing decided to make the original plane with internal stairs, just like your average business jet. In order to do that, the fuselage has to be not much farther off the tarmac than a business jet.

If you do the wiki, you can see a picture of the 737-100 with its stair extended. Really.

This design requirement also made it necessary to embed the JT-8D (a low-bypass turbofan type) into the wings. This is an engine with a max diameter not much more than military turbojets. This cozy relationship between fuselage and tarmac has persisted ever since.

And all was well for years. If you've flown a 737 in the last couple of decades, it's had the CFM56 engine (high-ish bypass) with that flat bottom front. This engine is mounted on a pylon hanging well out in front of the wing.

The latest engine which makes the Max special, the CFM LEAP-1B, has even a bigger max diameter than the '56. What to do? The '56 has a fan 60 inches, while the LEAP is 69 inches. Oops. Among other things, Boeing was determined to slide the Max through the existing certification, which is to say no additional training on type-specific simulator. Them dings be expensive. Among other things, the LEAP had to be placed even higher and more forward than the '56. The upshot was that the flight characteristics of the Max differed from previous 737s. What to do? Cheat.

Ok, that's harsh, but what Boeing really did. We know, now, that they slipped in this MCAS software to make the Max behave like previous 737s in conditions where the testing showed discrepancies in behavior. And now we know that didn't work. We also now know that a simple change to the software isn't in the cards. If that kind of fix would fix the problem, we'd be there by now. There be dragons in that plane.

What to do? What's the Worst Case Scenario?

In simple terms, the Max can't be certified as is, MCAS or otherwise. Software can't always rescue hardware. In other words, my suspicion is that Boeing has been lying about the degree of aerodynamic problems the LEAP engine does to the airframe. If that's the answer, Boeing has to go back to the drawing board. At the very least, the landing gear and wings have to be redesigned. From the original plane, the main trucks just tuck into the fuselage, without doors. Longer legs mean the trucks have to be anchored farther out on the wings, which brings all kinds of aero questions. Can the existing wing form support landings where the hinging has to be? Does the existing wing form even have room for trucks farther out? Will the longer legs interfere with engine mounting? That's a really big oops. And so on and on.

Worst Case Scenario.

24 July 2019

The Measure of A Man

The Joy of Quant gets no better than observing when Mother Nature, manifest in the laws of physics, and every now and again, economics, gives the Single Finger Salute to those who try to break the rules. Once you get past calculus, if you do, there is a long list of 'higher mathematics' courses open to you. Most don't actually make much use of calculus, but do strain your memory of algebraic manipulation you last did faithfully in high school. That's the reason such courses are daunting. That and the fact that each makes up its own, arcane, jargon. One of my favorites, at least by name, is measure theory. Herein some examples of measuring the costs of man.

We start by delving into the rich trove of man's perfidy as manifest in 'Law & Order' with the episode 'Mega' wherein Annette O'Toole frames husband Michael McKean for murder. They play a pair of self-help/created religion grifters (aka, Scientology/EST). McKean's character is convicted of the murder, but post-trial the DAs realize that he's likely been set up, and the most likely candidate for that is O'Toole's character. The DAs meet with her to attempt to figure out the problem. She parrys their questions, but does tell them her motive (if she did it!):
There was never enough, enough power, money, attention, women, enough control.
(Why did you speak up in his defense and let him live?)
So he would have to sit very quietly for a very, very long time, and know that now I had it all.
I wanted him to know that, for him, there wasn't going to be any more.

If you see the video, note that O'Toole drags out that last as 'moooooooore', with lips deeply curled. Which brings us to three situations where 'more' is something of an issue.

First, my adventures with medical care on Block Island. We decided to do an additional week this year, Memorial Day week. Normally we visit in spring and fall when the number of Mainlanders isn't much more than the 900 Islanders, so this would be a visit closest (for us) to High Season. Would we like all those crowds? On the whole, the weekend was tolerable and the rest of the week much better as most Mainlanders had been sent back to where they came from. On Memorial Day afternoon, I managed to drive a huge splinter into my foot, which meant medical intervention. Fortunately, the Island has a Medical Center, and I found it staffed. In the end, I was able to get around well enough for the rest of the week. My doctors here found yet more stuff in my foot, so recovery took longer than it might have.

What made the whole situation interesting is that we had come over on the Point Judith ferry on Memorial Saturday, and, as usual, read the latest issue of 'The Block Island Times'. This is a weekly tabloid, and the huge headline read 'Dr. Clark Resigns'. At the time, on the ferry, the drama was amusing, but of no concern to us. Boy howdy, was I wrong about that!! Since then there have been many twists and turns in the saga of Dr. Clark, the Medical Center, and the controlling Block Island Health Services (BIHS). You can follow it on the site, if you wish. Which, of course, I have since I've got some skin in the game. So to speak.

In short order, the clear dysfunction led me to send a letter-to-the-Editor at the Times. He replied asking for some additional information, but I had by then concluded that a Mainlander's view of the issue might not be welcome, so asked him to spike it. But, of course, as more reporting (news, editorials, letters) appeared in yet later issues, I took to the keyboard and composed a longer and more specific missive (more than one draft, as more news poured out). No response from the Editor this time. Didn't surprise me. So, here is the bit relevant to this missive:
None of the reporting since Memorial Saturday has spoken to the principle issue: health care for a community of the size and isolation of Block Island will only work if it's run as a community service. Based on reporting, some number of Islanders consider the Board to be the problem. Others say that the Medical Center should prioritize PCP services. Well,
"We don't physically have enough bodies on the island to make up for the deficit," said Miller. "If you only have 800 to 1,000 people on-island, you will only see them once a year if they're not sick. You can't bring them in for nothing."
-- Times 2/2/14

So, clearly, at one time the conflict between serving Islanders and getting revenue from Fumbling Mainlanders was understood.

The conflict boils down to: who should control? Clearly, the town government provides substantial funds, but, once again based on reporting, has little to no say in the running of BIHS/Medical Center. Why? Islanders must know how the arrangement was consummated in 1989 ceding control. Reading the latest financial report, I didn't see any numbers delineating the split in revenue between Islanders and Fumbling Mainlanders over the year. But I think it fair to guess that we Fumbling Mainlanders manage to get into all sorts of predicaments both at a higher rate and simple count, given the widely asserted number of 20,000 of us per day during High Season. Naturally, this last visit was the only one even near High Season; way too many Fumbling Mainlanders. (Yes, I get the irony.)

Health economics has been a sub-specialty of the field for some time. According to the wiki: "A seminal 1963 article by Kenneth Arrow, often credited with giving rise to health economics as a discipline, drew conceptual distinctions between health and other goods." One of the differences between health and other kinds of production is that is largely fixed cost. In the case of my injury, the direct costs were:
- a disposable scalpel
- some gauze
- some lidocaine
- some latex gloves
- a tetanus booster
- and, eventually, some antibiotic (turned out that more splinter was left in)

So, of course, as I was polishing and editing the fourth iteration of this missive, The Editor publishes the story about the 'slipup' with the multiple Stovers. Which, of course, led me to look up as much history of Mr. Stover as the innterTubes provided. Of particular interest was his later rebuttal of financial shenanigans. As you can see from the latest (calendar 2016, in innterTube territory) BIHS financial report, direct costs of service are a minuscule part of the income statement at $37,249.34 (line 5650-30 Medical Supplies) from a total expenses of $907,309.75; given the number of patients times the consumables in my case, that number is well within the ballpark. The rest, or vastly most, of expense is fixed cost. Note also the %39.000 in depreciation, which is a non-cash item, or should be. With a patient load of 5,126 (page 1), direct cost of service works out to $7.27 per patient. It is likely that fully 5,000 of those visits were Fumbling Mainlanders. As you can see, it's an overwhelmingly fixed cost operation. With such a structure, the only way to reduce *total average cost* per patient is to run as many patients through the system as possible, spreading all those fixed costs as thinly as possible. There aren't enough Islanders to pull that off. Fumbling Mainlanders to the rescue!! You're welcome. If this back of the envelope analysis, dependent as it is on sparse innterTubes available data, comes as a surprise to anyone, well...

In all, likely less direct cost than the co-pay for my insurance. I don't know, and may never, how much my insurance paid the Medical Center, but nearly all of that reimbursement went to pay down the fixed costs of the Medical Center. Unless the staff are paid on a piece work basis ($X per patient, scaled by severity of condition) even labor costs are fixed. Well, until someone gets canned? :) The point, of course, is that Fumbling Mainlanders have to be the largest source of revenue, or second behind the town contribution/memberships. Yet neither Fumbling Mainlanders nor the town have any control over the Medical Center.

Here on the mainland, within some limits, folks can change providers if they become dissatisfied. I suppose (and was likely true before the current regime existed?) Islanders could hop a plane to Westerly in an emergency or take the ferry for scheduled appointments. Having a PCP practice on-Island is clearly of benefit to Islanders. But it can't be funded just by Islanders on an out of pocket fee-for-service model. Early this year I asked my PCP (he has a few others working for him), just for yucks, how many patients he carried. He said 2,000. An HMO just for Block Island?

In sum, health services on the Island is a public good, not like a tee shirt concession. It is a monopoly that is currently unregulated. Staffing is at the whim of, by all reporting, one person. It is funded largely, if not almost entirely, by the town and Fumbling Mainlanders such as myself. Does this make any sense to anyone?

So, how does fixed cost dominant production survive? And the answer is: make it up on volume. The three days of Memorial weekend demonstrated how stupid Fumbling Mainlanders are on the narrow pavements, crossing in front of vehicles and riding bikes and mopeds into traffic. It's a wonder that there's not a death every day during High Season. Or is there...? There can't be 'more' health care for Block Islanders without more Fumbling Mainlanders' insurance coverage being billed. Or just go all socialist by having the Town government take over the Medical Center and BIHS. Or less stringently, putting BIHS under closer regulation. Which seems to be ignored by Islanders. Some times 'more' requires changes.

Which brings us to Second: 5G smartphones. Will there ever be 'more' than 4G/LTE? Not likely, most places. First off, what does 5G mean? The telecoms are baiting and switching like mad over the very definition. 'Real 5G' is in the microwave bands and higher, while 'Fake 5G' just uses an extension to LTE bands. mmWave transmission goes almost no where. Even humidity deflects it. Will there be 'real 5G' outside a few metro areas? Nope. And even in metro areas, inside use will work only with some expensive tweeks. As Barnum said about suckers.

And now, hot off the press, Third: Netflix going all marginal cost in India.
Netflix customers in India watch a higher percentage of videos on their phones than anywhere else in the world, and a greater proportion of Indian customers sign up to the platform via mobile than any other country, Ajay Arora, Netflix's director of product innovation, said Wednesday.

"We want to really broaden the audience for Netflix, want to make it more accessible, and we knew just how mobile-centric India has been," he told reporters in New Delhi.

The push to widen its audience in India, which now has more smartphone users than the entire population of the United States, comes a week after Netflix reported lower-than-expected subscriber growth.

This is exactly the behavior of a producer with high fixed costs: do anything to attract revenue that's greater than the marginal/variable cost of the next widget, thus amortizing a bit more of that fixed cost burden. For the likes of Netflix, that variable cost is indistinguishable from zero.

It's not nice to fool Mother Nature.

15 July 2019

New Gold - part the eighth

Well, that didn't take long. A report today says it all
The other factor is that the US dollar is the world's reserve currency. That status creates persistent demand for the greenback, making it difficult to properly value the currency.

Read the piece.

14 July 2019

Figures Don't Lie, But Liars Figure - part the second

This is now a series. Yahoo!

Today's entry derives from what should be a helpful explanation of what the unemployment really is. What's truly irritating is that the author is described as a 20 year economics reporter for the NYT. Phooey. He either makes rookie mistakes, or he's in on the gag.

His thesis is that the top-line monthly unemployment number doesn't tell the whole story. D'oh!!! Of course it doesn't, which has been discussed in these endeavors more than once. He could simply direct the reader to the BLS site which gives the alternative numbers (in table A-15), to which he makes oblique reference, but never spends a word or two to mention or explain. U-3 is the top-line number, while U-6 is the larger definition. Come on!!!

Moreover he says
Older people typically don't participate in the work force to the degree that younger people do...

While historically true before the Baby Boomers began aging, and supporting their geezer parents both directly and indirectly, it hasn't been true for some time. As you can see, geezers participation has been rising for the simple reason that retirement benefits have largely disappeared for today's Geezer Boomers. While it hasn't reached the level of middle age prime earners, necessity is the mother of retention.

And, naturally, we can't forget The (would be) Manchurian President's biggest lie
I've seen numbers of 24 percent -- I actually saw a number of 42 percent unemployment. Forty-two percent.

Of course, pure bullshit. Almost. If you know where to look (now you do) there is a number from BLS that's in the 40% ballpark. It's the employment/population ratio, which if inverted gets that fantastic number. Of course, it makes no sense to call sick 80 year geezers and infants as unemployed.

Figures don't lie, but liars figure.

13 July 2019

I Still Hate Neil Irwin - part the thirteenth

Good ole Neil creeps ever closer to the problem of the USofA's economic/quant oxymoron. How can we have low unemployment rate, low inflation rate, and low (or negative, at times) long term interest rate? The answer can't be found, in a causative manner, in the data. At least, not the commonly available batch. The data will show the result, and does.

It took the Fed's Powell, not a bleeding heart liberal, to state the obvious: "The Fed's New Message: The Economy Can Get a Lot Better for Workers".
... American workers were due for some catch-up growth in their compensation — after years in which their pay fell as a share of the economy.

Boy howdy! Once again, the sources of inflation
1 - cost push
2 - wage push
3 - demand pull

With wages retrenching for years, how can the highly paid pundits not admit that without a growing wallet in the lower classes, there won't be measurable inflation (by the long held definition of CPI, of course) from wage push? There can't be of course; inflation by definition requires an excess of moolah chasing (a shortage of?) output. OTOH, we've seen ample evidence that there's been significant demand pull inflation in assets. That's why long term inflation, measured by rising bond prices (the inverse of the interest rate) is rampant. All those highly paid CxOs put the TARP and QEs moolah into Treasuries, M&A, and dividends. Not to mention the flickering rate inversion over the last year or so.
"I think that's really the underlying problem. We're getting reasonable wage growth, but we missed all of those years beginning at the beginning of the century."

In other words: events drive the data, not the other way around. At the macro-analysis level, anyway. Having lost badly with the Dot Bomb around 2000, the moneyed class went looking for high-return at low-risk, which, of course, doesn't really exist, and created the mulligatawny stew of ARM and CDO and CDS. Knowing how to game out the events following from the creating, and sowing, of these instruments isn't in the data. It's in the New York Times and Wall Street Journal. If anywhere. One can, to a limited extent, make a gross prediction if one has money flows data. House prices couldn't keep skyrocketing forever, and neither can share prices.

Where's the next witch's brew?

08 July 2019

The Red and The Blue

Regular reader likely knows that one of the major themes of these endeavors is that there are almost no Red states or Blue states, but rather old, male, white, uneducated, rural counties and just the opposite demo in cities. A recent episode showed that, by that metric, there are only three majority Red states. Even from the Right, there's been much lamenting over the Left Behind States. What none of the reporting, including what's about to be referenced, have pointed j'accuse at the very people in these Left Behind States: they always, and continue to, elect Damn Gummints which treat them like chattel. Little education. Little health care. And so on. To quote Pogo: "We have met the enemy and he is us." Those Left Behind, luxuriating in huntin' and fishin' and bar fightin' and proud of it, bitch about not having any future. Wake up, jerk. You did it to yourself. All the opioids that Purdue can churn out won't improve your situation.

Well, two stories today offer up more information.

First, 'The Texas Miracle' proves what the Red/Blue divide really is. It's quite long and detailed. Snicker if you will. I sure did.
Nearly all of the net growth in jobs and new businesses in Texas over the last decade, Labor Department data show, has been concentrated in four large metropolitan areas — Austin, Dallas, Houston, and San Antonio. Those areas accounted for more than four out of every five jobs created in the state since the recession ended.

The article goes on the tell us about a Left Behind town/city, Longview, and the efforts to become less Left Behind.
"We have everything you need in Longview", Mr. Mack said recently, after a luncheon sponsored by the Chamber of Commerce that focused on community health. "We have plentiful land. We have adequate water, which most places don't have. We have everything you need. All we need is for people to realize that we're here now."

Of course, what Mr. Mack and the rest of Longview's booster won't say: if you lose your job at one of their wonderful establishments, you're shit out of luck if your position goes away, since there's little to no similar companies doing what your former employer did. Capitalists like, beyond overt fascist Gummints, company towns. The advantage of Blue cities is just that workers have alternatives. Recent reporting on hiring collusion in Silicon Valley is worth reading up. Maybe not as free and open as the Valley boosters say, but really big cities do have far greater diversity within and across sectors.

Also in recent reporting, Alaska goes total reactionary. "We don't need no education!"
The University of Alaska System is bracing for a 41% cut in funding it receives from the state, after Gov. Mike Dunleavy vetoed a $130 million line item in the state's budget.

Wouldn't want the kids to have useful skills, now would we?

02 July 2019

The Downer Party

All too frequently, there is reporting of a cabal of Debbie Downer Democrats going into 2020 based on the loss to The Manchurian President in 2016. This cabal keeps making noise that Democrats (aka, Hillary) blew it by not pandering to enough Rural Red Reactionaries. Baloney. She lost by 78,900 votes across three states with substantial urban weight.
- Pennsylvania - 78.7%
- Michigan - 74.6%
- Wisconsin - 70.2%

The Manchurian President pandered to, and got out the vote by, the 6,000 year old Earth lunatics. On the other hand, the Democrats simply didn't get out the vote for Hillary. Perhaps she wasn't the most charismatic campaigner (OK, she was lousy), but if the Democrats can't find another 78,900 urban votes...

Of more concern, and I've not yet seen anyone other than my humble self raise the red flag, is Johnson/Stein. The net difference Johnson was approx. the popular vote margin for Hillary. Of more significance is that Stein votes wouldn't swing any state from The Manchurian President to Hillary even if she got all of them, while Johnson out-polled Stein in every state where both were on the ballot. In sum, without either/both The Manchurian President would likely have won the popular vote, since Johnson's a close surrogate of The Manchurian President, while Stein votes were also, on balance, anti-Hillary votes. Numbers matter.

As to losing Red states, all but Maine, Mississippi, Vermont, and West Virginia are majority urban. Betcha didn't know that, right? There aren't as many Red states as you thought. Getting The Manchurian President was mostly the fault of the Democratic leadership, not The Manchurian President's 'superior campaign'. Not even close. The 'shellacing' in 2010 was for the same reason. Some folks just never learn. Get out your base; it's bigger by far, and even in states commonly labeled Red.

Now, on to a related topic. For some time, at least since the 2016 vote count, I've been telling anyone who will listen (and many more who won't) that the White House can't be the primary goal in 2020. That goal is the Senate. At long last someone of the pundits finally says so. Took long enough.