09 February 2016

Pity the Poor Bankster

Well, pundits are finally getting it: they ain't much return in real investment.
Central banks have kept interest rates low to stimulate demand for loans. But loans with low interest rates are often less profitable for banks. As a result, banks may then lend less, which may then reduce the overall impact of low interest rates on the economy.

Unfortunately, but not surprising, the reporter (and, one might presume, his sources) ignores the other sides of the situation. First, corporations can sell bonds directly, thus benefiting from low interest rates. Second, if business generally had some idea of how to grow organically, low rates lowers the necessary return from investment. That was, after all, the notion behind QE (and successors): business will invest more if the internal rate of return of real investment only has to pay back 2 or 3 percent on borrowing. But, of course, corporations packing away trillions in retained profit aren't going to borrow to invest. They have been borrowing to pay dividends, buy back own shares, and buy competitors. Invest? How childish.

08 February 2016

Dee Feat is in Dee Flation, part the thirty first

Well, the shit keeps hitting the fan:
Treasuries Rally on Risk Aversion

U.S. Treasuries are starting the week on a strong note as European equity indices move to multi-year lows and the German 10-year Bund yield declines 5 bps to 0.25%. The S&P 500 is indicating down 1.14% to 1,856.5 and WTI crude is down 2.85% to $30.01/bbl. The U.S. Dollar Index is up 0.19% to 97.22 and gold is up 1.75% to $1,178/troy oz., a fresh multi-month high
Yield Check:
2-yr: -2 bps to 0.70%
5-yr: -3 bps to 1.21%
10-yr: -3 bps to 1.81%
30-yr: -2 bps to 2.65%

briefing.com 8 Feb 2016, 7:38am

07 February 2016

Deductive Reasoning

Among the most interesting, if heinous, ideas of the 1% is that healthcare is just another consumption good, and if "we" make it more expensive, then people will spend less on healthcare. And the rich will get richer, of course. The notion, of course, is nonsense. Only the 1% could ever afford tummy tucks and silicone boobs and such. The rest of us only seek healthcare when something bad happens. There have been studies (you can find them, if you're interested) which find that preventative healthcare makes no difference to overall health of a population. That should be obvious: how many health problems do you know of that show up in an annual (or biannual) physical? Vanishingly few.

One of the core ideas of the save-money-by-making-healthcare-expensive is high deductible plans. Again, the notion that such plans are both less expensive and offer better health is grounded in the notion that the 99% are visiting doctors like corner ice cream shops. Of course not. Sure, there are a few hypochondriacs, but the gross effect is to deny vital healthcare when it's a matter of life or mortality/morbidity.

Today is a report on a study bringing some data to the issue. I expect you'll find it illuminating.
The hospital offering the $3,000 M.R.I. might lose enough business that it will lower its price.

I picked that particular sentence, since The Tyranny of Average Cost™ is generally couched in terms of what happens to the price of MRI scans if far fewer are done. The notion that MRI scans are highly price elastic makes no sense: the cost is almost entirely amortization of the machinery. There's not much else to pay for. Fewer MRIs at $3,000 mean that the MRI scan now costs $4,000 or $5,000, depending on how many are removed from the accounting.

Anyway, a large, unnamed company instituted a high deductible plan, and was followed by some academic researchers.
Amitabh Chandra, an economist at Harvard, and one of the researchers, said he was convinced the study would prove the value of deductibles, at least for well-off and well-educated workers.

He was wrong.
Mr. Chandra said he was no longer convinced that deductibles turned patients into good consumers. "The best case was the theoretical case," he said. "I was all for high-deductible plans before I wrote my paper."

And, of course, the punchline is stated explicitly:
"There's essentially nothing they can do to prevent the likelihood they'll have high-cost health events," [Dr. Peter B. Bach] said.

Well, just die early. Which is the whole point, of course.

04 February 2016

Lucy and the Football

I'm not the most avid sports fan, but the rise of "advanced analytics" has the benefit of amusement. AA is really nothing more than descriptive stat plus a bit of population correlation. Squared differences, variously manipulated. From what I gather, it all began in baseball and Billy Bean decades ago. Now, one can be paid silly amounts of money for running simplistic correlation programs on reams of data. Such a life.

About the only sports TeeVee I'll admit addiction to is "Pardon the Interruption", and largely because Korny and Wilbon are ex-Washington Post. I lived there for rather a while, and deeply regret leaving. That's another episode. Yesterday had, as one might expect, much discussion about the Broncos and Panthers. In particular, whether the Panthers' record is legit or not. Korny noted that the Panthers had a really weak schedule (27th, meaning almost the weakest), thus their 15-1 regular season record is some part smoke and mirrors.

I certainly don't care enough (I don't bet on anything other than MegaMillions) to confirm to what, if any, degree the following scenario applies to the Panthers this year, but just consider.

The issue is strength of schedule, which gets various names in use. Consider a team's (ABC) third game of the season against team PQR. ABC wins. At the end of the season PQR has a record of 9-7, making them a so-so team, neither really good nor really bad; a middling weight. Might even be in the playoffs. From ABC's point of view, its record's SoS weighting would view the PQR third week win as a net positive. But... how well was PQR playing in the third week (I'll assume that ABC, for these purposes, played at a constant level for the year; accounting for ABC's fluctuations just means more arithmetic), which is what really counts in determining the appropriate weight? There are two extremes that lead week three to be viewed as a local minimum or local maximum, from a weighting point of view: PQR's record in the previous and latter two (or one or three or four or...) weeks. On the one hand, PQR could have been blown away in those four games by teams that had losing records both for the year, and the local five week schedule; making a win over a 9-7 team over weighted. On the other, PQR could have blown away teams with winning records, both for the year and the local five week schedule, excepting ABC which beat them; making the win more significant. Details matter. Actual SoS calculations often go to additional levels of opponent schedules, so some folks do care to be accurate.

What matters most, of course, in attempting to predict any single game outcome is standard SWOT analysis, data appropriate to the sport. That's not so much data as anecdotal assessment. There's a reason legal betting odds on sports aren't determined by experts, but by the flow of bets on the participants. The sports books change the odds to move the money flows between the opponents aiming for a 50-50 split, taking a cut from the gross.

31 January 2016

Strange Bedfellow

Greg Mankiw isn't on my list of favorite pundits. Forced to assign him to a list, it would certainly be Plutocrat Panderer. But then he goes and writes this piece. What makes the piece amusing is its backhanded swipe at the puffed up quants.
So if looking at contemporaneous economic conditions is not a reliable way to judge presidents, how should they be graded?
Similarly, a better way to judge presidents is by the policies they pursue, not the outcomes over which they preside. This task is harder than merely looking at unemployment, inflation and the growth of gross domestic product. It requires having a view about what policies are best at fostering prosperity and acknowledging that the experts are often divided on that question.

IOW, if not the proximate data, then what? Well, how well do you reward your friends and punish your enemies? Policy changes.

29 January 2016

Days of Wine and Roses

Since Data Science definition has become at least as lucrative a job as doing Data Science, I suppose I should do my part. Let's start with science, from the Wiki of course.
To be termed scientific, a method of inquiry is commonly based on empirical or measurable evidence subject to specific principles of reasoning. The Oxford English Dictionary defines the scientific method as "a method or procedure that has characterized natural science since the 17th century, consisting in systematic observation, measurement, and experiment, and the formulation, testing, and modification of hypotheses."

A bit of history. Back in the mid-60s, before I ever got involved, there were IBM and the 7 dwarves. If one designed, built, or sold such machines one was almost always a EE, preferably from a top tier engineering school. Likewise, if one programmed such machines. Programming was mostly in each machine's assembler. Among real science majors, the hierarchy started:
1 - math
2 - physics
3 - electrical engineering
X - everything else in any order, since they don't matter to the first 3

This became a problem, since "doing computers" was the au courant field, much as the web and big data and data science and such are today. The problem was that very few could survive a EE curriculum. The demand for some degree that was "computers" was high, but the number who qualified for the then appropriate degrees was small. So, just as sub-prime and ALT-A mortgages were invented to fulfill the demand for MBS, so was the computer science degree. Not smart enough for EE? Want to do computers? OK, you can write programs, just sign here.

The gag continues. We've seen NoSql created just because the average kiddie koder can't grok simple set theory, so toss out ACID and DRI and such and return to the thrilling days of yesteryear with COBOL and VSAM, only now it's java/PHP/C# and xml. Or whatever file type is hot.

Data science is quite the same. Your average quant wannabe can't grok stat or OR or maths, so let's create a new discipline that's lite on the tech, but heavy on the buzzwords. Thus, data science. Many of the skills attributed to data science used to be carried out by admin assistants: data cleaning, data entry, and other drudge tasks. Now, these are the critical skills of the data scientist.

The nature of science is to discover previously unknown aspects of God's world. Humans don't create scientific artifacts, we find them lying around, often hidden under millennia of ignorance. Priestley didn't invent oxygen, just found it lying around, after some experimental effort to isolate it. And so on for the rest of the periodic table. Einstein found relativity by asking a really simple question: what happens if this tram departs from the clock tower as fast as light? Nothing more than that. The resulting maths are, to be fair, a tad intimidating. And he had help with that bit.

What science, then, is there in data science? What previously unknown aspect of God's world has been found by the efforts of data science? None that have crossed my path. Nor will there be.

Old wine in new bottles. Or, as the child observed, he has no clothes.

26 January 2016

"A Man's Got to Know His Limitations" [update]

Today we muse on Apple and Data Science. Apple's quarterly is due after market close today, and more and more tutorials/explanations/paeans to Data Science appear. Let's see what they may have in common.

Apple has been on a growth spurt ever since the iPhone came into being. Those with short memories may have forgotten that it was a sick puppy at birth. Teeny display. Barely 3G capable. Exclusive to the lousiest carrier in the country. But the Apple crowd bought into the reality distortion field meme that the iPhone was somehow different and better. "They" concluded that Apple had made the first smartphone. Not hardly. What Apple had was a lock, for a time, on cap touchscreens. Kind of hard for anyone else to make a phone with one. Skip happily forward to today, and iPhone accounts for about 2/3 of company revenue and more net income (including spend on that "ecosystem").

The nature of today's quarterly is written about by the tonne of virtual ink. Mostly, "channel checking" says that iPhone sales, at best, are flat for the current quarter and next; at worse, both fall a bit. The compromise is that current quarter meets expectations, but guidance for next is down.

But, isn't this just the strong suit of Data Science? Accumulating unstructured data, analysing with arcane Bayesian functions, and confirming the priors? What could be easier? Wait... Cook has, in the past, tut-tut-ed the analysts for extrapolating from their channel checking to conclusions about Apple. Could he be right again, this time?

The stock market is an insiders' game first, last, and always. Corporations are under no obligation to report material issues other than in filings, and, on the whole, each company decides the meaning of "material". Prominently displaying pretty financial results, aka non-GAAP, while burying the legal numbers in footnotes is a common tactic. Shareholders in biopharma companies get blindsided every morning. Nearly all meaningful financial data is quarterly/annual filings. Every blue moon there might be a separate filing on a material issue, but often only good news and mostly not direct financial information. Like, "we told our suppliers to reduce deliveries by 30% this and next quarters". Doesn't happen. Management, or anyone else for that matter (just ask Martha Stewart), isn't supposed to go out and trade company instruments knowing such non-public information. Since Apple is a significant component in mutual funds, hedge funds, indexes, and ETFs one might wonder whether Tim et al is allowed to trade those instruments? Clearly, the NASDAQ has been down because Apple has been down. Some people think so. Research it if you're interested.

If Data Science were all such magic that real stat can't do, wouldn't we already know what Apple will say a bit after 4:00 Eastern?

Well, the verdict is in: the channel check looks mostly right. Moreover, many have pooh-poohed Beijing's remnimbi rumba as no big deal. Not so much, it turns out. Not so many rich Chinese, either.