What is step 2?

Via Paul Krugman an article in the Salt Lake City Tribune about a guy with a business model that the Underpants Gnomes look at as incomplete:

Under Jonathan Johnson’s leadership at Overstock.com, the company has socked away $10.9 million in gold and silver and has a three-month food supply for each employee, bracing for a financial crisis that the Republican gubernatorial candidate said is certain to come…
“We expect when there is a financial crisis, there will be a banking holiday,” Johnson, Overstock chairman of the board, told the group. “I don’t know if it will be two days or two weeks or two months. But we have $10 million in gold and silver in denominations small enough … that we can use it for payroll. We want to be able to keep our employees paid [and] safe and our site up and running.”

My wife is the Overstock user in the Mayhew household, but as I understand it, Overstock.com is a fairly typical business to consumer web portal that sells things to people and then ships orders from warehouses to people via either the US Postal Service or a number of private entities. An extremely large percentage of Overstock’s payments from customers are made via credit cards over the web.

Is that an accurate enough description of their business process?

If it is, how does having a Smaugian hoard help Overstock keep their site up and running in a bank holiday scenario?

A bank holiday would only happen if there is extreme doubt about the liquidity of the entire US banking system. It would come as all private short term (1 to 30 day) lending either comes to a halt or has credit card level interest rates attached to it. It would be a massive breakdown of social trust and every single entity that can grab cash is going to hold onto cash for as long as legally (and usually illegally) possible as the lawyers can fight later over contract breaches.

So in that scenario, who is actually going to go on Overstock.com and buy an expandable three piece luggage set?  Sure, Overstock might give their ISP four pounds of gold to keep their connection on as their bank account would be frozen, but their ISP’s bank account would be frozen so the four pounds of gold would be insufficient to keep the ISP’s lights on (which are probably going off shortly anyways as the natural gas power plant is shutting down as their supply is not being delivered unless the power plant can supply physical cash in unmarked, non-sequential twenties instead of a line of credit).

How does having a gold hoard for an online retailer actually help that retailer stay open when the rest of the social and commercial network is shutting down due to the lack of credit?


Kynecting the elections to policy

And now that officially sucks.

Tue Nov 03, 2015 at 5:22 PM PT (David Jarman): To recap, Republican Matt Bevin (whom you might remember from losing his tea-flavored primary challenge to Mitch McConnell in 2014) has won a surprising (to the extent that no poll had given him the lead) victory in Kentucky’s gubernatorial race.

Bevin (R-Sociopath) had run on a promise to repeal as much of PPACA as he could.  That means shutting down Kynect, the Kentucky exchange and turning down the very successful Medicaid expansion.  As Charles Gaba noted, he was promising to take away health insurance from 9% of the state.  And he looks like he won going away and with coattails.

There are two major health policy implications of a Bevins win.  The first is the switch from Kynect to Healthcare.gov.  This is not that big of a deal in and of itself.  It is a different portal and a different set of branding but subsidies flow to people who buy insurance from healthcare.gov after the King case.  The biggest downside is if Kentucky wanted to go the Wyden Waiver route, having their own exchange makes plumbing a Wyden Waiver, even a very conservative Wyden Waiver a whole lot easier.

The big policy change is Medicaid expansion.  Medicaid expansion covered approximately 420,000 Kentuckians as of this morning or 9% of the state. Bevin ran against expansion.  The best case scenario is the hospital groups march into  Bevin’s chief of staff office tomorrow and tell him flat out that their books don’t balance without expansion of some sort and they’ll lay covering fire for an extremely punitive waiver application.  That is the best case scenario and I’ll give it a 10% chance of happening.  The probable case scenario is 420,000 people are fucked as of February 1, 2016 and most of Appalachian Kentucky has medical care and medical financing resembling Third Wold nations again.

Kentucky, a red state, is highly likely to return to being a purple state on the New York TimeUpshot map of uninsurance rates:

NYTIMES Uninsurance Rate

It was a good two years of actually connecting people to health insurance without the death defying worry that a toothache could either be immediately fatal or financially destroying.


Destructive but not purely crazy

Via Lawyers, Guns, and Money:

The House Freedom Caucus seems nice:

Yesterday, Politico published the House Freedom Caucus “questionnaire” which it described as pushing for “House rule changes.” The document does do that. But it also does a lot more. It seeks substantive commitments from the next speaker that would effectively send the entire country into a tailspin…..

The government will run out of money on December 11. Unless additional funding is approved before that date, the government will shut down.

The House Freedom Caucus wants the next speaker to commit to not funding the government at all unless President Obama (and Senate Democrats) agree to defund Obamacare, Planned Parenthood and a host of other priorities. This is essentially the Ted Cruz strategy whichprompted at 16-day shutdown in 2013. This would now be enshrined as the official policy of the Speaker Of The House.

The House Freedom Caucus wants the next speaker to commit to oppose any “omnibus” bill that would keep the government running. Rather, funding for each aspect of government could only be approved by separate bills. This would allow the Republicans to attempt to finance certain favored aspects of government (the military), while shuttering ones they view as largely unnecessary (education, health)

Just as a reminder, the House Freedumb Caucus is roughly 40 Congress critters who hold the balance of power as long as the rest of the House Republicans believe that maintaining in-group norm of only passing major legislation with only Republican votes is worthwhile. They are the power bottoms of the House GOP caucus.

But they are not intrinsically crazy if the goal of the Freedumb Caucus is to both protect their own ass and make it less likely for a Democrat to be elected to the White House in 2016. Extraordinarily destructive and cynical, yes, but they are not completely crazy.

The first goal of assuring their own re-election for most of the HFC is achieved if they can get out of the Republican primary.  Right now the Republican primary electorate is extremely pissed at the “Establishment” Republicans as those Republicans have overpromised and under-delivered (ie they could not repeal Obamacare nor impeach Obama for making them think that he would take away all of their guns due to the Democrats controlling some veto points against maximalist Republican goals).  Being a Republican who the Establishment (2nd DW dimension positive) hates is a good thing in the Republican primary in R+5 or redder districts.  The few HFC in D+0 or bluer districts can only win in waves or due to local idiosyncratic circumstances.  They’re irrelevant as long as the rest of the HFC controls the 218th vote.

This is fairly straightforward playing to the base.

Now the second contention that I’ll make that the HFC demands for a shutdown or default or far more likely large contractionary policies increases the odds of a Republican being elected is pure political science cynicism.

Read more

Meanwhile, on the Republic of Gilead Campaign Trail…

Link from The Hill:

Sen. Ted Cruz (R-Texas) said on Sunday that the concept of birthright citizenship is illogical given America’s struggles with rising illegal immigration.

“I think birthright citizenship as a policy matter doesn’t make sense,” the GOP presidential candidate told host John Dickerson on CBS’s “Face the Nation.”

“We have right now upwards of 12 million people living here illegally,” he said. “It doesn’t make any sense that our law automatically grants citizenship to their children, because what it does is it incentivizes additional illegal immigration.”…

Brave words for a candidate born on foreign soil whose own father’s immigration to America was legally a little iffy. Except that Ted Cruz doesn’t seem to be running for President any more — he’s running for God-King in the Republic of Gilead.

Conventional wisdom is that Cruz is now deploring “illegals” in an attempt to lure the angry racists populists once Trump gets bored. I’m sure those voters would, as Austen might phrase it, make a tidy supplement to the marks already cheering for Cruz’s long con. Which wouldn’t worry me, if there weren’t a lot of desperate women going to get hurt by his ongoing maneuvers to draw Talibangelical attention…

The Washington Post, on the Iowa event last weekend:

… Here in Iowa, 57 percent of Republican caucus-goers identified as evangelical Christians, according to entrance polls. Cruz has been courting faith leaders here, dispatching his father Rafael, a pastor, to speak at events around the state and announcing Friday that he wants to recruit a pastor in each of Iowa’s 99 counties to do faith outreach.

Cruz’s Iowa campaign chairman, Matt Schultz, suggested Cruz was chosen by God. Read more

He’s The Heavy, He’s My Brother

Of all the problems of his own making that plague Kansas GOP Gov. Sam Brownback (like a state economy disintegrating under tax cuts for the rich, vice tax for the poor, and basic government services like public school districts that can’t pay their bills to keep the lights on and can’t pay teachers enough to keep them working) the most interesting problem may be his asshole brother Jim.

Undulating fields of crops and livestock-dotted pastures are the domain of a trigger-happy bully who brags about a political cloak of invincibility keeping him beyond reach of the law in faithfully conservative Linn County.

Adversaries say he has woven a liquor-infused tapestry of fear colored by intimidation, abuse and lies. The saga features stalking, death threats, trespassing, drive-by gunfire, massive explosions, cattle theft, loan defaults, hit-and-run driving and marital strife. Linn County Sheriff’s Department files bulge with complaints about him.

There is trepidation among acquaintances to speak freely, a point accentuated by the number expressing nervousness about reprisal if they were candid. There is genuine fear.

Descriptions of events offered by those willing to speak out converge to reveal a potentially lethal menace. Neighbors allege some in law enforcement responded to cries for help with degrees of indifference or favoritism.

Locals aware of the dynamics shake their head in dismay. In a place where people honor the Second Amendment and revere the self-defense castle doctrine, there is astonishment no one has been gunned down.

Folks in direct path of this prairie hellion pray for an end to what some coined “neighborhood terrorism.”

So far, their nemesis has found no reason to relent.

Not when your name is Jim Brownback and you are a brother to Sam, the most powerful politician in Kansas.

I swear, this story reads like a Jack Reacher novel by Lee Child or an episode of Longmire.  Farm pigs mauled by dogs, a keg of nails scattered on the driveway, drunken shotgun drive-bys,  blowing up stuff in the middle of the night to wake the neighbors, cattle unleashed to devour crop fields, frightened insurance adjusters, and good ol’ boy shitkicker cops that won’t touch the governor’s brother.

Jim Brownback is possibly the second biggest asshole in the entire state, next to his brother.  It’s terrorism by a cartoon bully and nobody can put a leash on this guy.

Least of all the governor of the great state of Kansas.  What Jim is doing to his neighbors on the Kansas plains, Sam is doing to the entire state.

Bullies all around, doing what they do best, has always been the way.

The bigotry of expectations

Paul Waldman in the Plum Line makes a very reasonable argument as to why the US-Iran deal won’t be overturned even if the Republicans win the White House in 2017.  It is a good argument grounded in reality, practicality, institutional design, economics and diplomacy.  It makes sense.  However there is one line in the article that sticks in my craw:

That’s a plan so stupid that it’s hard to imagine even the current GOP presidential candidates carrying it out.

The incentive structure for members of the Republican Party is to compete to out-stupid each other. Collectively this may or may not be the optimal incentive for the party as a whole, but that is the individual level incentive.

Don’t bet against extremely dumb beating out somewhat stupid.

Open thread

A couple bucks per name and voting rights

For some odd reason, this week at work I’ve been pulled back into data cleaning and data structuring.  St. Mungos, a  local hospital is trying to bounce their service data against Mayhew Insurance claims data.  We received several hundred scrolls of names and claims.  None of their data is quite structured like our claims and enrollment data, so we’ve been building a crosswalking system.  For data security purposes, none of the hospital data contains Social Security numbers and for Pain In The Ass purposes, none of their data contains the Mayhew Insurance coverage number.

So that means we have to find the identities of people in their service universe by a combination of birth date, last name, first name, provider, and date of service.  Creating automated queries that match on a significant number of not quite unique keys has allowed for very high confidence matches on 97% of the probable unique individuals.  The remaining 3% are manual look-ups and judgement calls.

For instance, is Harold J. Potter with a DOB of 7/31/80 from the hospital file the same person as Harry Potter DOB 7/31/80, with multiple dates of death in the insurance file the same individual?  Eyeballing the data and making an informed judgement, I am guessing at 95% confidence that this is the same person.  How about Sirius R. Black and S. Regulas Black?  Are they the same person?  Probably, but not definately.  Being slightly more serious, how do you match Markos Moulitsas Zúniga to Kos Moulitsas where the DOBs are off by a day?  Going through the 3% is costing the company at least a dollar per name, and probably closer to $1.50/name.

For the purposes of my current task getting 98% or 99% of the data set matched at high confidence is more than sufficient.  The research plan allowed for a 5% mismatch rate.

Eyeballing the names that are in the unmatchable bucket, there are three notable groups.  The first are recent births where the name is Baby Girl Smith matched against 15 girls born on that date in the shared claims universe.  These are no big deal.  The second group of unmatched names are women whose relationship status has changed multiple times.  My wife would be in this group as the hospital data has her as Jane Doe while the insurance data has her as Jane Mayhew.  The final group are people with names that fall outside of culturally dominant naming conventions.  Names with extra punctuation and names whose ordering structures don’t follow the First/Last or First/Maiden-Last convention fall out very readily.

This is important because error is acceptable even if we are willing to spend money to minimize the error.  Verifying voter rolls based on name and date of birth matches is guaranteed to produce significant unmatched.  There is a good argument that voting rolls should be as up to date as possible, but to do that in a good faith manner requires spending serious money as one person dropped from the rolls who should not have been dropped is one too many.  $5 to $10 per name at 100% confidence would be reasonable estimate of the cost of scrubbing an entire list of votes.  The vast majority of names would be passed at a dime per name, while the last 5% to 1% might cost $2 per name to verify, and the last 1% will cost $50 per name to verify.

Scrubbing an entire state for pennies per name is a guarantee of producing false positives and statistically certain mismatches.  And given who is more likely to have mismatched names between data sets, cynicism should reign.