18 August 2014

Life is A Totem Pole

Having lived, in my first job in Boston and thence in Washington, DC (and having done my grad school in econometrics), the notion of localized inflation was well understood at least as far back as the 1970s. No matter where you work/live, the $$$ you get paid aren't what determine how well you live, but where those $$$ seat you on the local totem pole. I guess it's still news to some folks.


These endeavors have spent a lot of virtual ink over the last year or so making that case that financial quants don't know as much as they think, and that the events which led to The Great Recession and London Whale should be, but clearly haven't been, teaching moments. I've heard tell of various bots and such, but this expose` is truly mindnumbing.

As has been promulgated here, it is events which move data, not the other way 'round. While time series analysis can give one a warm and fuzzy feeling that the way of the world is clearly delineated in all-the-data-to-today, and one can happily predict off the end of this curve, it's not much more than tossing a dart in a pub.

You must read it.

17 August 2014

An Open Letter to Dr. Shiller [update]

In today's Times Business section, Robert Shiller gives his take on stocks, specifically the question: are they priced too high? He titles the piece (dead trees version), "The Mystery of Lofty Elevations". The web title is similar.

Note, of course the use of the word 'mystery'. Are stock prices really too high? As I've been arguing for some time, NO. Those with piles of moolah to 'invest' have traditionally been in bonds, clipping coupons each quarter. Living off the interest, in simple terms. As even Dr. Shiller has to know, short term interest rates (in particular, the Fed Funds rate) are not what determines long term (corporate) interest rates. The value of corporate bonds is determined by the return earnable from new plant and equipment in the hands of 'job creators'. To the extent that 'job creators' choose to not buy plant and equipment, at any rate of interest, then the long term return to moolah holders will go down.

There's also the matter of the supply of loanable funds, aka "The Giant Pool of Money". It's still around, and corporations continue to hold ever larger amounts of idle cash. They'd love, I'm sure, to be given 10% annually by Washington or Bejing, but that ain't gonna happen. Until such time as the 'job creators' decide that physical investment is more lucrative than fiduciary games, the supply of long term funds will outstrip the demand for such. It's that simple. So far, their notion of smart capital allocation is to stuff it into the mattress.

The Captains of Industry have simply run out of ideas. The digital economy is, on the whole, super-cheap to run. Even as we approach the next (and, perhaps, last) barrier to continuance of Moore's Law (for myself, it looks like physics has repealed the law), the cost per cycle continues to diminish. In any case, the web economy amounts to little more than giants and dwarves fighting over advert buyers and advert clickers. Arabs, and their blue eyed Texas brethren, learned the hard way that you can't eat oil. The web economy will eventually realize that advert peddling is just a lot of wheel spinning; much ado about little.
So I've been trying to come up with a theory to explain today's elevated stock prices -- and maybe convince myself that they could remain lofty for some time. One factor to consider is that bond prices are high, too.

It's as simple as I've said: bond prices are high (yields low) because the Captains of Industry aren't smart enough to figure out new ways to turn moolah into machines. Even with money available at historically low rates, and retained earnings at historical highs, the Captains just don't know what to do with it.
When there aren't enough good investing opportunities, people wishing to save more for the future may succeed only in bidding up existing assets even if they think they're overpriced. Call it the "life preserver on the Titanic" theory.

And, in other words, all those Koch Brothers types who bleat that Americans have to save and invest more ignore the plain fact that corporate America hasn't been able to allocate The Giant Pool of Money that's already sloshing around. Push more money into the Pool, and returns will fall still further. Econ 101.

We can expect yet more bleating from the .1% class that capital gains taxes are ruinous. Why? Since they can't get by on the low coupon returns, they must needs turn to stock price appreciation (as we've seen) and share selling for income. And, of course, since only Little People pay taxes, their incomes shouldn't be taxed. Just watch.

Seems some of the sell-side analysts (larger fish in the pundit pond than I, alas) take exception. This shouldn't be too surprising, since they make their obscene incomes flogging stocks. But they do have a salient point: they tout forward P/E, while Shiller's CAPE (where's Superman when you need him) is explicitly backward looking and for a long time at that. The notion that more data is better is generally a good thing, except when there's been an inflection in the data of your model, and worse if the inflection is in the recent past. The evidence is clear that the Captains of Industry are using fiduciary capital in more financial engineering exercises and less in buying plant and machines. No physical investment, no hard returns to capital, no interest earned, no demand for additional capital. The causation is that long term earnings determine short term rates. Central banks can hope to manipulate short term rates, and even set one or two, but they can't change the course of science and engineering (the real kind). Without tech progress, there's only foregone consumption to pay the vig. Looking at 21st century economies through the lens of 19th century experience (even to mid 20th) is folly.

It isn't a fluke that The Giant Pool of Money (larger now than then) was dumped into non-productive real estate. The Captains of Industry had no use for it. They still don't.

13 August 2014

DBA According to Moses

Every now and again I'll run across a posting which is prescriptive in nature. Some more aggressive in tone than others. Today brings a latter day Moses, via the PG site.

I didn't add anything, or more commandments. What piques my interest, enough to spend a few minutes typing, is the nature of his commandments. Fact is, that's always the interesting bit to such posts: what is the battlefield?

For a database oriented site, it was, at one time, about a 50/50 chance that the battlefield was over how to build database-centric applications in the face COBOL/PHP/java + RBAR data opposition. Alas, these days it seems that even the simple-talk, and certainly the PG folks (it is a code + RBAR oriented datastore from the beginning), have been assimilated by the Borg. It's really too bad. Current technology, modulo the Big Data knuckleheads, is exactly what Dr. Codd had in mind when laying out the RM: all is related, all actions happen "at once", and implmentation is left to the vendor. The key notion to the RM, more than any other aspect, is analogous to the difference between sigma notation and matrix notation. RBAR on the one hand, and it-happens-all-at-once on the other. Of course, real hardware is inherently sequential on the metal. Well, in truth, not so much any longer.

Anyway, this Moses posting is about the care and feeding aspects, what was once the only definition, of DBA. Save for a jibe to promote better coding!! Too bad. Dr. Codd's ghost should be haunting more who claim to "do" RDBMS. May be then we'll all find the gonads to fight for what is right? We need a Bravedata, complete with lots o blue paint (that's IBM avatar, of course).

11 August 2014

A Visit in the Hadleyverse

Hadley Wickham is interviewed by Eduardo Arino de la Rubia in conjunction with the useR! meeting in Los Angeles. He warms the cockles of my heart starting at 15:50 of the interview, when they discuss the bipolar (my term, not theirs) nature of R. Yes, R is both an Excel on steroids (neither of them says Excel, but reading between the lines...) and a kinda, sorta language to write programs. Since most of us know Hadley via his packages (and he's using Rcpp more, lately), and this interview is about how he goes about making same, it's impossible to judge how he feels about R as a stat command language for stats, quants, data scientists, and the like. But it is clear that he gets the difference. It's also my inference that he envisions lots o London Whales making their mistakes in R rather than Excel. Whether that's a good thing is another matter. Sometimes brain surgery should only be done by trained neurosurgeons. If The Great Recession taught us nothing else, it's that quant is too often done cavalierly, by those with little to no understanding of what they're doing, yet with an outcomes' agenda. Bad dog.

07 August 2014

Sovaldi Sings Iago

For those who aren't up on Shakespeare and Rossini, Iago is a bad guy. A really bad guy. Sovaldi, according to some (and a growing count, by most accounts), is the archetype of Big Pharma profiteering; taking money from the many to give to the few for little or no real value.

Sovaldi is Gilead's hep/C vaccine, more accurately anti-viral, which costs $1,000/pill or $84,000/treatment. It is claimed to offer higher cure rate, lower side effects, and shorter treatment period.

Unlike breast cancer or prostate cancer, hep/C, by and large, is a Bad Person's Disease; mostly IV drug users. A lifestyle choice. Mostly, not all. It happens that a large number of the infected are also incarcerated, which shouldn't be too surprising. Turns out prisons don't get mandated discounts. Remember those TV cop shows, where the perp turns out to be an ex-con who just can't seem to get it together on the outside, and heists a bodega in order to get back inside? Likely see more of that.
The drug's virtue, that it frequently cures a chronic and sometimes very costly disease, is not always relevant in the prison setting. Hepatitis C can take up to 30 years to turn from active infection to serious liver disease. Therefore, a costly investment in a cure for those prisoners close to their release date will offer no relief to a prison's long-term medical budget, even if it might improve the prisoner's lifetime health.

In other words, this is what happens when public health problems are shunted off to for-profit outfits. Gilead gets all the profit (they did drop more than $11 billion to buy the company which devised it, though). Taxpayers get all the costs. What would Adam Smith (the real one) say? I mean, where's the cost/benefit analysis of making life more comfortable for drug addicts? There are nearly as effective, far less costly, existing therapies. So, what is the marginal benefit? What is the marginal cost?

Here's another case where even the 1% will need Obamacare. If the insurance companies, and other right wing scolds ("$84,000 for drug addicts!!") get their way, then far fewer treatments will happen, and Gilead will, unless stopped by the Damn Gummint, raise the price further. By restricting the use of Sovaldi to those who get hep/C, but not through Bad Behavior, the unit price will have to go up in order to preserve Gilead's entitled profit. How much up? Well, according to this chart from CDC data between 5% and 15% of infections can be considered non-willful. Like that word? Very Biblical, that word. So, the innocents infected will have to pay (or some form of insurer) some multiple of $84,000 to keep Gilead in golden threads. Take away 90% of your customers, and, if you can get away with it, the remaining 10% get truly hosed.

Here's the kicker (which hasn't been stated, so far as I've seen): since 60% (more or less) of those infected are IV drug users, and Sovaldi isn't a vaccine which offers immunity to disease, but an anti-viral, we should expect that those 60% will rotate through with some frequency. Drug use: the gift that keeps on giving. Perhaps, at the end of treatment, patients get an 'S' tattoo, indicating that they've had their once-in-a-lifetime Sovaldi treatment. No second helpings.

This is the prototypical case where an army of forensic accountants could be unleashed to discover how much of Sovaldi's claimed development cost actually was spent on the ground, and how much went to SG&A. I'd love to see that number. Not going to happen.

06 August 2014

Big Fish, Little Fish

The NYT DealB%k has an interesting piece from one of the regulars on the Big Fish eating the Little Fish in the cloistered world of the innterTubes. But Solomon leaves out the lede, as they say in the pubbiz: these Big Fish ain't really conglomerates.
The paradox is that conglomerates outside the tech sector are an endangered species. The 1960s was the age of the conglomerates. ITT, for example, made both weapons and movies, with the idea that smart managers could operate any business and different operations would diversify the business. But that strategy did not work out as planned. The problem was that managers needed to focus on their businesses. If investors wanted to diversify, they could do so by simply investing in the separate companies. And splitting off businesses would discipline managers not to waste extra cash.

What he misses, or simply doesn't want to say, is that Google, et al, of today aren't (structurally) anything like GTE, ITT or the rightfully infamous Gulf+Western of Charley Bluhdorn. Those were real conglomerates, with real divisions (ex-companies) in real, different, businesses. With the possible exception (too soon to tell) of Apple/Beats, the rest are all agglomerations of advert pushing platforms. Despite the appearance of differing users and "products" for those users, the real clients are the advert buyers. And they're (Google, et al) all attempting what the New York Yankees have done for generations: buy up any player who's any good to keep said player from ending up with another team. It takes very little to create an web-based advert platform; WhatsApp was, what, 6 guys? If you're Google, that makes your sphincter kinda tight. 93% of its revenue is from adverts, still (last 10-K). How much of its profit is assigned, I couldn't find, but I'll guess all of it. All of the side projects are just that: on the side.

In any case, these aren't conglomerates, just oligopolists aiming to be the monopolist.